Handbook of organic materials for electronic and photonic devices [Second edition] 9780081022849, 9780081022856, 2572592592, 0081022840

Citation preview

Handbook of Organic Materials for Electronic and Photonic Devices

This page intentionally left blank

Woodhead Publishing Series in Electronic and Optical Materials

Handbook of Organic Materials for Electronic and Photonic Devices Edited by

Oksana Ostroverkhova

An imprint of Elsevier

Woodhead Publishing is an imprint of Elsevier The Officers’ Mess Business Centre, Royston Road, Duxford, CB22 4QH, United Kingdom 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States The Boulevard, Langford Lane, Kidlington, OX5 1GB, United Kingdom © 2019 Elsevier Ltd. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-08-102284-9 (print) ISBN: 978-0-08-102285-6 (online) For information on all Woodhead publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Matthew Deans Acquisition Editor: Kayla Dos Santos Editorial Project Manager: Peter Adamson Production Project Manager: Maria Bernard Cover Designer: Miles Hitchen Typeset by SPi Global, India

Contents

List of contributors Preface

Part One 1

2

3

Materials

Organic materials for optoelectronic applications: Overview Yasuhiko Shirota, Hiroshi Kageyama 1.1 Introduction 1.2 Photoactive and electroactive organic materials 1.3 Organic materials for use in optoelectronic devices 1.4 Fabrication and performance of organic optoelectronic devices 1.5 Summary and outlook References Key trends in sustainable approaches to the synthesis of semiconducting polymers Assunta Marrocchi, Valeria Trombettoni, Daniele Sciosci, Filippo Campana, Luigi Vaccaro 2.1 Introduction 2.2 Solar cell (OPV) and organic field-effect transistor (OFET) structure and operation 2.3 Nitroxide-mediated radical polymerizations 2.4 Perspective References Further reading Functional blends of organic materials for optoelectronic applications Valerii Sharapov 3.1 Introduction 3.2 Thermodynamics and kinetics of polymer mixing 3.3 Blends for organic solar cells 3.4 Conclusion References

xiii xix

1 3 3 4 9 28 32 33 43

43 45 80 83 83 89 91 91 92 97 106 107

vi

4

5

6

7

Contents

Organic photonic nanostructures Deirdre M. O’Carroll 4.1 Types of organic photonic nanostructures: Particles, wires, fibers 4.2 Photonic properties as a function of size, confinement and molecular ordering 4.3 Nanostructure fabrication 4.4 Photonic functions and applications of organic photonic nanostructures 4.5 Outlook References Molecular engineering of organic and organometallic second-order nonlinear optical materials Christopher M. Ashcroft, Jacqueline M. Cole 5.1 Introduction 5.2 Measuring the SHG response of a molecular chromophore 5.3 Dipolar SHG molecular chromophores 5.4 Octupolar SHG chromophores 5.5 Future prospects for the molecular engineering of NLO chromophores 5.6 Conclusions References Further reading

111 111 115 120 123 128 128 139 139 143 146 158 166 172 172 176

Molecular crystals and thin films for photonics € Mojca Jazbinsek, Peter Gunter 6.1 Introduction 6.2 Second-order NLO organic crystals 6.3 Integrated EO applications 6.4 Terahertz-wave generation and detection with organic crystals 6.5 Conclusions and outlook References

177

Hybrid perovskites for device applications Kyle Frohna, Samuel D. Stranks 7.1 Introduction 7.2 Structural properties 7.3 Deposition methods 7.4 Optoelectronic properties of HOIPs 7.5 Charge carrier properties 7.6 Charge carrier recombination 7.7 Device applications and challenges 7.8 Conclusions References Further reading

211

177 177 189 193 201 202

211 213 218 220 224 227 234 243 244 256

Contents

Part Two 8

9

10

11

vii

Mechanisms

Frenkel exciton dynamics: A theoretical perspective € Oliver Kuhn 8.1 Frenkel excitons in organic materials 8.2 Model Hamiltonian 8.3 Dynamics of excitons 8.4 Spectroscopy 8.5 Applications 8.6 Final remarks Acknowledgments References Strong light-matter interactions and exciton-polaritons in organic materials Arko Graf, Laura Tropf, Jana Zaumseil, Malte C. Gather 9.1 Introduction 9.2 Materials for strong light-matter coupling 9.3 Mixing different excitonic states via strong coupling 9.4 Polariton condensation 9.5 Electrical pumping 9.6 Summary and outlook References Advances in modeling the physics of disordered organic electronic devices Michael C. Heiber, Alexander Wagenpfahl, Carsten Deibel 10.1 Introduction 10.2 Microscopic simulation and modeling 10.3 Macroscopic simulation 10.4 Conclusions and outlook Acknowledgments References Doping in organic semiconductors Hannes Hase, Ingo Salzmann 11.1 Introduction 11.2 Doping organic semiconductors 11.3 Electronic structure upon molecular electrical doping 11.4 Key role of the microstructure in doping OSCs 11.5 Summary and outlook Appendix. Compound abbreviations References

257 259 259 261 264 269 270 276 276 276

281 281 285 291 295 300 303 304

309 309 310 327 338 339 339 349 349 351 358 370 373 375 376

viii

12

13

Contents

Spintronics and magnetic field effects in organic semiconductors and devices Tho Duc Nguyen, Eitan Ehrenfreund, Zeev Valy Vardeny 12.1 Introduction 12.2 Magneto-conductance and magneto-electroluminescence in OLEDs 12.3 Organic spin valves (OSVs) 12.4 Optically detected magnetic resonance in DOO-PPV isotopes 12.5 Time-resolved magneto (M)-photoinduced absorption (PA) in donor-acceptor (DA) copolymers 12.6 Conclusions Acknowledgments References Doping and processing of organic semiconductors for plastic thermoelectrics € Anna I. Hofmann, Renee Kroon, Christian Muller 13.1 Introduction to organic thermoelectrics 13.2 The Seebeck effect and the thermoelectric figure of merit 13.3 Design of thermoelectric generators 13.4 Optimizing the thermoelectric properties of OSCs through doping 13.5 Interplay of processing and the nanostructure of doped organic semiconductors (OSCs) 13.6 Processing of doped organic semiconductors (OSCs): Thin film versus bulk architectures 13.7 Conclusions and outlook Acknowledgments References

Part Three Characterization, structure-property relationships, processing, and stability 14

15

Conductivity measurements of organic materials using field-effect transistors (FETs) and space-charge-limited current (SCLC) techniques Katelyn P. Goetz, Oana D. Jurchescu 14.1 Introduction 14.2 Field-effect transistor measurements 14.3 SCLC measurements 14.4 Outlook References Organic thin-film microstructure characterization Victoria Savikhin, Michael F. Toney 15.1 Introduction 15.2 Fundamentals of X-ray scattering

385 385 387 400 411 417 421 423 423 429 429 430 433 435 438 440 442 445 445

451

453 453 454 470 477 478 489 490 493

Contents

15.3 15.4 15.5 15.6

Characterization on length scales above 10 nm Domain composition characterization Characterizing material packing and behavior Summary: Pathways toward more systematic optimization and development of organic materials Acknowledgments References

16

17

18

19

Surface-enhanced Raman scattering (SERS) as a characterization method for metal-organic interactions Katherine Willets, Kathryn Mayer 16.1 Introduction 16.2 SERS background 16.3 SERS applications 16.4 Active and passive control of SERS signals 16.5 Conclusion References

ix

497 507 513 522 522 522

529 529 531 537 541 546 546

Advances in solution processing of organic materials for devices Katherina Haase, Mike Hambsch, Cecilia Teixeira da Rocha, Jakob Zessin, Stefan C.B. Mannsfeld 17.1 Introduction 17.2 Solution-deposited, carbon-based electrode materials 17.3 Semiconducting device layers by solution processing 17.4 Solution deposition of dopants and electrically active additives 17.5 Advances in coating dielectric materials for organic devices 17.6 Summary References

551

Advances in device fabrication scale-up methods Tatsuo Hasegawa 18.1 Introduction 18.2 Printing of semiconductor layers 18.3 Printing of metal wiring 18.4 Outlook References

579

Device stability in organic optoelectronics Ayse Turak 19.1 Introduction 19.2 Degradation measurements and decay curves 19.3 Degradation mechanisms 19.4 Development of device lifetimes 19.5 Summary and outlook References

599

552 552 557 563 566 570 571

579 584 590 593 594

599 600 604 637 639 640

x

Contents

Part Four 20

21

22

23

Applications

663

Organic photovoltaics (OPVs): Device physics Michael A. Fusella, YunHui L. Lin, Barry P. Rand 20.1 Introduction 20.2 Organic photovoltaics (OPVs) 20.3 Principles of operation 20.4 Device structures 20.5 Directions for progress: Alternative organic photovoltaic (OPV) materials and device concepts 20.6 Conclusions and outlook References

665

Organic light-emitting diodes Paul-Anton Will, Sebastian Reineke 21.1 Introduction 21.2 Fundamentals of OLEDs 21.3 Efficiency considerations 21.4 Applications 21.5 Current research frontiers 21.6 Outlook References

695

Materials and physics of light-emitting electrochemical cells (LECs) Shiyu Hu, Jun Gao 22.1 Introduction 22.2 The physics of LECs 22.3 LEC materials 22.4 PLECs with bipolar electrodes 22.5 Conclusions and outlook References Vertical organic transistors € Lussem, € Bjorn Changmin Keum, Gil Sheleg, Nir Tessler 23.1 Introduction 23.2 Organic permeable base transistors 23.3 Vertical organic field-effect transistors 23.4 Patterned source vertical field-effect transistor 23.5 Vertical organic light-emitting transistors 23.6 Outlook Acknowledgments References

665 666 667 678 679 687 688

695 696 703 711 715 719 719 727 727 732 739 744 749 750 759 759 761 768 771 777 778 779 779

Contents

24

25

26

Vapor sensing using organic, polymer, and nanomaterial field-effect transistors Hui Li, Wei Shi, Jennifer Dailey, Hyun-June Jang, Jian Song, Junsheng Yu, Howard E. Katz 24.1 Introduction 24.2 Active OSCs in vapor sensors 24.3 Morphological and additive enhancements 24.4 Biomaterial incorporation 24.5 Health-related vapor sensors using organic and polymer semiconductors 24.6 Small biomolecule detection 24.7 Conclusions and outlook Acknowledgments References Processing and patterning of conducting polymers for flexible, stretchable, and biomedical electronics Tom Kitto, Come Bodart-Le Guen, Nicolo Rossetti, Fabio Cicoira 25.1 Introduction 25.2 Organic conducting polymer-based transistors 25.3 Processing methods for PEDOT 25.4 Patterning organic devices 25.5 Biomedical applications 25.6 Conclusion and outlook References Further reading Organic electronic memory devices Michael C. Petty 26.1 Introduction 26.2 Memory types 26.3 Resistive memory 26.4 Organic flash memory 26.5 Ferroelectric RAMs 26.6 Spintronics 26.7 Molecular memories 26.8 Biological memories 26.9 Conclusions and outlook References

Index

xi

785

785 788 794 799 801 806 808 808 809 817 817 819 823 828 830 834 834 842 843 843 844 851 857 859 862 863 864 868 870 875

This page intentionally left blank

List of contributors

Christopher M. Ashcroft Cavendish Laboratory, Department of Physics, University of Cambridge, Cambridge, United Kingdom Come Bodart-Le Guen Polytechnique Montr
eal, Montreal, QC, Canada Filippo Campana Laboratory of Green Synthetic Organic Chemistry, Department of Chemistry, Biology and Biotechnology, University of Perugia, Perugia, Italy Fabio Cicoira Polytechnique Montr
eal, Montreal, QC, Canada Jacqueline M. Cole Cavendish Laboratory, Department of Physics; Department of Chemical Engineering and Biotechnology, University of Cambridge, Cambridge; ISIS Neutron and Muon Source, STFC Rutherford Appleton Laboratory, Didcot, Oxfordshire, United Kingdom; Argonne National Laboratory, Argonne, IL, United States Jennifer Dailey Department of Materials Science and Engineering, Johns Hopkins University, Baltimore, MD, United States Carsten Deibel Institut f€ ur Physik, Technische Universit€at Chemnitz, Chemnitz, Germany Eitan Ehrenfreund Physics Department, Technion—Israel Institute of Technology, Haifa, Israel Kyle Frohna Cavendish Laboratory, University of Cambridge, Cambridge, United Kingdom Michael A. Fusella Department of Electrical Engineering, Princeton University, Princeton, NJ, United States Jun Gao Department of Physics, Engineering Physics and Astronomy, Queen’s University, Kingston, ON, Canada Malte C. Gather Organic Semiconductor Centre, SUPA, School of Physics and Astronomy, University of St Andrews, St Andrews, United Kingdom Katelyn P. Goetz University of Heidelberg, Heidelberg, Germany

xiv

List of contributors

Arko Graf Organic Semiconductor Centre, SUPA, School of Physics and Astronomy, University of St Andrews, St Andrews, United Kingdom; Institute for Physical Chemistry, Universit€at Heidelberg, Heidelberg, Germany Peter G€ unter ETH Zurich and Rainbow Photonics AG, Zurich, Switzerland Katherina Haase Department of Electrical and Computer Engineering, Technische Universit€at Dresden, Dresden, Germany Mike Hambsch Department of Electrical and Computer Engineering, Technische Universit€at Dresden, Dresden, Germany Hannes Hase Concordia University, Montr
eal, QC, Canada Tatsuo Hasegawa Department of Applied Physics, The University of Tokyo, Tokyo, Japan Michael C. Heiber Center for Hierarchical Materials Design (CHiMaD), Northwestern University, Evanston, IL, United States Anna I. Hofmann Department of Chemistry and Chemical Engineering, Chalmers University of Technology, G€ oteborg, Sweden Shiyu Hu Department of Physics, Engineering Physics and Astronomy, Queen’s University, Kingston, ON, Canada Hyun-June Jang Department of Materials Science and Engineering, Johns Hopkins University, Baltimore, MD, United States Mojca Jazbinsek Zurich University of Applied Sciences (ZHAW), Zurich, Switzerland Oana D. Jurchescu Wake Forest University, Winston-Salem, NC, United States Hiroshi Kageyama University of the Ryukyus, Nishihara, Japan Howard E. Katz Department of Materials Science and Engineering, Johns Hopkins University, Baltimore, MD, United States Changmin Keum Organic Semiconductor Centre, SUPA, School of Physics and Astronomy, University of St Andrews, St Andrews, United Kingdom Tom Kitto Department of Chemical Engineering, University of Bath, Bath, United Kingdom

List of contributors

xv

Renee Kroon Department of Chemistry and Chemical Engineering, Chalmers University of Technology, G€ oteborg, Sweden Oliver K€ uhn University of Rostock, Institute of Physics, Rostock, Germany Hui Li Department of Materials Science and Engineering, Johns Hopkins University, Baltimore, MD, United States YunHui L. Lin Department of Electrical Engineering, Princeton University, Princeton, NJ, United States Bj€ orn L€ ussem Department of Physics, Kent State University, Kent, OH, United States Stefan C.B. Mannsfeld Department of Electrical and Computer Engineering, Technische Universit€at Dresden, Dresden, Germany Assunta Marrocchi Laboratory of Green Synthetic Organic Chemistry, Department of Chemistry, Biology and Biotechnology, University of Perugia, Perugia, Italy Kathryn Mayer Department of Physics and Astronomy, University of Texas at San Antonio, San Antonio, TX, United States Christian M€ uller Department of Chemistry and Chemical Engineering, Chalmers University of Technology, G€ oteborg, Sweden Tho Duc Nguyen Department of Physics and Astronomy, University of Georgia, Athens, GA, United States Deirdre M. O’Carroll Rutgers University, Department of Materials Science and Engineering; Rutgers University, Department of Chemistry and Chemical Biology, Piscataway, NJ, United States; Trinity College Dublin, School of Physics, Dublin 2, Ireland Michael C. Petty Department of Engineering and Centre for Molecular and Nanoscale Electronics, Durham University, Durham, United Kingdom Barry P. Rand Department of Electrical Engineering; Andlinger Center for Energy and the Environment, Princeton University, Princeton, NJ, United States Sebastian Reineke Dresden Integrated Center for Applied Physics and Photonic Materials (IAPP) and Institute of Applied Physics, Technische Universit€at Dresden, Dresden, Germany

xvi

List of contributors

Nicolo Rossetti Polytechnique Montr
eal, Montreal, QC, Canada Ingo Salzmann Concordia University, Montr
eal, QC, Canada Victoria Savikhin Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, Menlo Park; Electrical Engineering Department, Stanford University, Stanford, CA, United States Daniele Sciosci Laboratory of Green Synthetic Organic Chemistry, Department of Chemistry, Biology and Biotechnology, University of Perugia, Perugia, Italy Valerii Sharapov Department of Chemistry, The University of Chicago, Chicago, IL, United States Gil Sheleg Sara and Moshe Zisapel Nano-Electronic Center, Department of Electrical Engineering, Technion-Israel Institute of Technology, Haifa, Israel Wei Shi Department of Materials Science and Engineering, Johns Hopkins University, Baltimore, MD, United States; State Key Laboratory of Electronic Thin Films and Integrated Devices, School of Optoelectronic Information, University of Electronic Science and Technology of China, Chengdu, People’s Republic of China Yasuhiko Shirota Osaka University, Suita, Japan Jian Song Department of Materials Science and Engineering, Johns Hopkins University, Baltimore, MD, United States Samuel D. Stranks Cavendish Laboratory, University of Cambridge, Cambridge, United Kingdom Cecilia Teixeira da Rocha Department of Electrical and Computer Engineering, Technische Universit€at Dresden, Dresden, Germany Nir Tessler Sara and Moshe Zisapel Nano-Electronic Center, Department of Electrical Engineering, Technion-Israel Institute of Technology, Haifa, Israel Michael F. Toney Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, Menlo Park, CA, United States Valeria Trombettoni Laboratory of Green Synthetic Organic Chemistry, Department of Chemistry, Biology and Biotechnology, University of Perugia, Perugia, Italy Laura Tropf Organic Semiconductor Centre, SUPA, School of Physics and Astronomy, University of St Andrews, St Andrews, United Kingdom

List of contributors

xvii

Ayse Turak Department of Engineering Physics, McMaster University, Hamilton, ON, Canada Luigi Vaccaro Laboratory of Green Synthetic Organic Chemistry, Department of Chemistry, Biology and Biotechnology, University of Perugia, Perugia, Italy Zeev Valy Vardeny Physics and Astronomy Department, University of Utah, Salt Lake City, UT, United States Alexander Wagenpfahl Institut f€ ur Physik, Technische Universit€at Chemnitz, Chemnitz, Germany Paul-Anton Will Dresden Integrated Center for Applied Physics and Photonic Materials (IAPP) and Institute of Applied Physics, Technische Universit€at Dresden, Dresden, Germany Katherine Willets Department of Chemistry, Temple University, Philadelphia, PA, United States Junsheng Yu State Key Laboratory of Electronic Thin Films and Integrated Devices, School of Optoelectronic Information, University of Electronic Science and Technology of China, Chengdu, People’s Republic of China Jana Zaumseil Institute for Physical Chemistry, Universit€at Heidelberg, Heidelberg, Germany Jakob Zessin Department of Electrical and Computer Engineering, Technische Universit€at Dresden, Dresden, Germany

This page intentionally left blank

Preface

Organic optical and optoelectronic materials have attracted increasing attention due to their low cost, easy fabrication, and tunable properties. This has spurred rapid development in the field of organic optoelectronics and photonics, with tremendous progress seen over the past several years, which inspired the second edition of this book. A variety of novel, high-performance materials have been synthesized and characterized; new approaches to the synthesis of green materials have been developed; novel theoretical, computational, and experimental methods that enabled better understanding of the physics behind optical and optoelectronic properties have been established; and a number of exciting applications have been demonstrated. This book aims to provide an overview of recent developments in all these areas (i.e., materials development, mechanisms, characterization and fabrication methods, and applications). Although it is difficult to keep the level of material coverage uniform from chapter to chapter in such a book, the overall level should be sufficiently accessible for researchers who are entering the field of organic optoelectronics and photonics (including graduate students, postdoctoral researchers, and professionals switching fields). However, it also should be advanced enough to serve as a useful desk reference for both academic and industrial researchers with a background in physics, chemistry, materials science, and engineering, who are already working with organic materials and their applications. To help readers deepen their knowledge of a particular topic, each chapter contains an extensive list of references that provide further information on the subject matter discussed therein. The book is organized as follows. Part I overviews novel optoelectronic (Chapters 1–3), optical, and nonlinear optical (NLO) (Chapters 4–6) organic materials, as well as hybrid perovskites (Chapter 7), with examples of structure-property relationships. New directions highlighted in this part include the molecular design of highperformance organic semiconductors (OSCs) (Chapter 1) and NLO materials (Chapter 5), approaches involving more environmentally friendly synthesis (Chapter 2), tuning of optoelectronic properties of OSCs using blends (Chapter 3), molecular assembly design strategies for photonic applications (Chapters 4 and 6), and design of hybrid perovskites for a variety of applications (Chapter 7). Part II of the book starts with a general theoretical description of exciton physics (Chapter 8) and discussion of strong light-matter interactions (Chapter 9) in organic materials. These are followed by descriptions of charge transport (CT) in OSCs and numerical modeling of various organic devices (Chapter 10) and mechanisms of doping in OSCs and resulting properties (Chapter 11). An extensive discussion of various magnetic field effects and spintronic applications of organic materials is provided in Chapter 12, and the part concludes with a description of the mechanisms behind the thermoelectric properties of OSCs (Chapter 13).

xx

Preface

Part III provides descriptions of experimental methods used to characterize organic materials, which are illustrated by examples of physical studies (Chapters 14–16), and also discusses issues pertaining to device fabrication (Chapters 17–18) and stability (Chapter 19). In particular, measurements of charge-carrier mobility and other parameters pertaining to CT are detailed in Chapter 14. A structural characterization of organic thin films, which are critical for optimizing electronic properties of organic devices, is detailed in Chapter 15. The surface-enhanced Raman scattering (SERS) method, as applied to studies of metal-organic interactions, which are important for understanding the contribution of contacts at the metal-organic interfaces to optoelectronic properties, is introduced in Chapter 16. Chapters 17 and 18 review recent advances in the solution processing of organic materials and strategies for scaling up organic devices, respectively. Finally, Chapter 19 discusses bottlenecks in organic device stability and methods for improving that stability. Part IV of the book is dedicated to selected applications of organic optoelectronic materials (Chapters 17–25). The chapters in this part summarize the characteristics of the best available organic devices, such as organic solar cells (Chapter 20), organic light-emitting diodes (OLEDs) (Chapter 21), and electrochemical cells (Chapter 22); vertical transistors (Chapter 23); chemical and biological sensors (Chapters 24 and 25); and electronic memory devices (Chapter 26). As no single book can provide a complete coverage of such a broad field of study, some important topics had to be omitted. For information on these topics, readers are referred to recent comprehensive review papers. Examples of such topics include ultrafast charge-carrier dynamics in organic materials [e.g., Chemical Reviews 117 (16), 10940 (2017)]; analytical description of CT in OSCs [Physica Status Solidi B 251 (3), 457 (2014)]; a wide variety of characterization methods and various aspects of structure-property relationships [Chemical Reviews 116 (22), 13279 (2016)]; and many others. I would like to express my gratitude to all the contributing authors for investing their time and effort into writing as accurate, complete, and up-to-date chapters as possible and for keeping on schedule as much as possible. Many thanks also go to Professors C. Risko, A. Laskarakis, and M. Kemerink, who provided comments on the first edition of this book and valuable suggestions for revision. It has been a great pleasure to work with Peter Jardim, Peter Adamson, Swapna Praveen, and Kayla Dos Santos at Elsevier—many thanks to them, and to the Elsevier production team, for their excellent professionalism in managing this project, ensuring timely communication with authors, and getting things done promptly and with high quality. This book could not have happened without the dedicated effort of all the authors, the editor, and the publisher, working together as a team. Oksana Ostroverkhova, Editor

Part One Materials

This page intentionally left blank

Organic materials for optoelectronic applications: Overview

1

Yasuhiko Shirota*, Hiroshi Kageyama† *Osaka University, Suita, Japan, †University of the Ryukyus, Nishihara, Japan

1.1

Introduction

Electronic, optoelectronic, and photonic devices using inorganic semiconductors, such as silicon transistor-based computers, light-emitting diodes (LEDs), semiconductor laser-based devices, and optical communication devices such as waveguides, have contributed greatly to the arrival of the present ICT society. In recent years, new fields of organic electronics, organic optoelectronics, and organic photonics using organic materials have opened up. These fields of science and technology, which are mostly related to information and energy, are mainly concerned with thin-film, flexible devices using organic materials. Devices using organic materials have advantages over inorganic semiconductor-based devices due to their light weight, flexibility, and and potentially low cost. Among organic electronic and optoelectronic devices, organic photovoltaic devices (OPVs), organic light-emitting diodes (OLEDs), and organic field-effect transistors (OFETs) have been the central subjects of current research and development. In addition, the fabrication of devices using organic materials by solution processes, which is called printed electronics, has been progressing. In particular, extensive research on various kinds of sensors using organic materials has been actively performed toward the realization of the society of the Internet of Things (IoT). Organic electronics, optoelectronics, and photonics constitute interdisciplinary fields that cover physics, chemistry, biology, and materials science. The science and technology of these fields, which can be termed organic functional materials science or organic device science, include wide areas from the molecular design and synthesis of photoactive and electroactive organic materials to the elucidation of their physical and chemical properties, as well as their structures, fabrication, and performance evaluation of devices using synthesized organic materials and the creation of new knowledge underlying the operation of organic devices. Photoactive and electroactive organic materials, including both small molecules and polymers, have been playing a crucial role in the development of organic electronics, optoelectronics and photonics, and for the 21st-century industries related to information and energy. This chapter gives an overview of organic materials, including both molecular materials and polymers, for use in optoelectronic devices, OPVs, and OLEDs, focusing on the recent progress in these fields. The characteristic features and molecular design concepts for molecular materials and polymers, concrete examples of various Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00001-2 © 2019 Elsevier Ltd. All rights reserved.

4

Handbook of Organic Materials for Electronic and Photonic Devices

kinds of photoactive and electroactive organic materials for use in such devices, and their fabrication and performance are discussed. Book and review articles on optoelectronic devices are cited in Chapter 1 of the first edition.

1.2

Photoactive and electroactive organic materials

Organic materials, which are assemblies of organic molecules, are mainly classified into small-molecular-weight organic materials, oligomers, and polymers from the standpoints of molecular weight and chemical structure. These organic materials are generally characterized by light weight, flexibility, ease of molecular design, and potentially low cost compared with inorganic materials. Organic materials for use in electronic and optoelectronic devices are photoactive and electroactive. Such photoactive and electroactive organic materials for use in electronic and optoelectronic devices (e.g., OFETs, OPVs, and OLEDs), are often termed organic semiconductors.

1.2.1 Organic π-electron systems Photoactive and electroactive organic materials are usually composed of π-electron systems. Generally, organic π-electron systems are characterized by their capability of absorbing and emitting light in the wavelength region from ultraviolet to nearinfrared, the capability of generating and transporting charge carriers, and excellent nonlinear optical properties. Photoactive and electroactive organic materials based on π-electron systems include small molecules, oligomers, and polymers. Polymers based on π-electron systems are roughly classified into linear π-conjugated polymers, planar π-conjugated polymers, and polymers having π-electron systems as side-chain groups. Generally, polymers readily form large-area, uniform films with mechanical strength and low surface roughness by solution processing. This film-forming ability of polymers makes them well suited to a number of practical applications. While the properties and functions of linear π-conjugated polymers are quite different from the corresponding small molecular units, polymers havng π-electron systems in the side chain possess essentially the same properties and functions as those of the corresponding small π-electron-based organic molecules. By contrast, molecular materials are crystalline in nature and are not able to form smooth films. Meanwhile, a family of amorphous molecular materials based on π-electron systems does form smooth, uniform, amorphous films.

1.2.2 Control of structures and morphologies Organic materials, including both molecular materials and polymers, exist as either crystals, amorphous glasses, or mesophase materials such as liquid crystals and plastic crystals, although polymers usually contain both crystalline and amorphous phases with varying ratios depending upon the kind of polymer. Control of such material morphology is crucial not only in scientific research, but also as applications for devices.

Organic materials for optoelectronic applications: Overview

5

Small organic molecules usually exist as crystals below their melting temperature. Oligomers with well-defined structures (e.g., oligothiophenes) are also crystalline. It should be noted, however, that a family of small organic molecules that readily form amorphous glasses above room temperature exists. They constitute a family of versatile organic materials (Shirota, 2000, 2005). On the other hand, polymers usually contain both crystalline and amorphous phases, ranging from highly crystalline polymers to fully amorphous polymers. It is noteworthy that certain kinds of small molecules [e.g., CuPc (2), TiOPc (3), and perylene pigments] exhibit polymorphism, taking a few different crystal structures. The crystal structures of materials greatly affect their properties and device performance, as reported with titanyl phthalocyanine, to cite just one example. Polymorphism in these compounds and its effect on their properties are well described in the first edition, in Chapter 1. Control of morphologies to provide the electrical conduction path in OPVs using copper phthalocyanine as a p-type organic semiconductor also has been reported (Peumans et al., 2003; Yang et al., 2005a,b). Room-temperature phosphorescence emission is observed not only for transition metal complexes, but also for metal-free molecular crystals, where intramolecular motions are suppressed by several intermolecular interactions in the crystal lattice, thereby minimizing nonradiative decay from the electronically excited triplet state. The search for room-temperature phosphorescent organic compounds in the crystalline state and the establishment of guidelines for designing room-temperature phosphorescent organic molecules, along with crystal engineering, have been current topics (Mukherjee and Thilagar, 2015; Shimizu et al., 2016). The properties and functions of crystalline oligothiophenes with well-defined structures can be controlled by varying the π-conjugation length. Extensive studies have been made of the synthesis, properties, and functions of oligomers with welldefined structures and varying conjugation lengths. They include oligo(arylene vinylene)s, oligothiophenes, and oligoporphyrins (q.v. first edition, Chapter 1). The correlation between the conjugation length of oligothiophenes and their electrical conductivities has been examined (Shirota, 2000). Usually, π-conjugated polymers contain mostly the crystalline phase. They have difficulty with solubility. To make π-conjugated polymers soluble in organic solvents, alkyl groups are often introduced as substituents. The improved synthetic methods have enabled the synthesis of highly regioregular poly(3-alkylthiophene)s with high degrees of the head-to-tail (HT) structure. They exhibited superior charge-carrier mobility and electric conductivity after doping relative to those of lower regioregular polymers. Highly regioselective poly(3-alkylthiophene)s have been synthesized, and the effect of the regioselectivity on their properties and device performance has been discussed (q.v., first edition, Chapter 1).

1.2.3 Crystalline molecular materials Examples of the representative classes of crystalline molecular materials are given by (1) polycondensed aromatic hydrocarbons, such as anthracene and pentacene (1); (2) metal and metal-free phthalocyanines (2,3); (3) porphyrines; (4) fullerenes and their

6

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 1.1 Examples of representative crystalline molecular materials.

derivatives (4); (5) perylenebisdiimides, such as anthra[200 ,100 ,900 :4,5,6:600 500 1000 :40 ,50 ,60 ] diisoquino[2,1-a:20 ,10 -a0 ]dibenzimidazole-10,21-dione (PTCBI) (5); and oligothiophenes with well-defined structures (6), among other substances (Fig. 1.1).

1.2.4 Amorphous molecular materials Generally, organic small molecules tend to crystallize readily, and hence they usually exist as crystals below their melting temperature. Systematic studies of the creation of small organic molecules that readily form stable, amorphous glasses above room temperature, which are termed amorphous molecular materials, started in the late 1980s (Shirota et al., 1989) and have become very active since the mid-1990s due to the application of these materials in OLEDs. It has now been widely recognized that like polymers, small organic molecules also form stable, amorphous glasses with definite glass-transition temperatures (Tgs) above room temperature if their molecular structures are properly designed, and that like crystals and liquid crystals, amorphous molecular materials constitute a new class of versatile materials that are useful in various applications. New concepts for photo- and electroactive organic materials have been proposed, which include electrically conducting amorphous molecular materials (Higuchi et al., 1991), amorphous molecular resists (Yoshiiwa et al., 1996), photochromic amorphous molecular materials (Shirota et al., 1998), and amorphous molecular materials for OLEDs (Shirota et al., 1994), OPVs, and photorefractive devices. Characteristic features and molecular design of amorphous molecular materials are fully discussed in the first edition, in Chapter 1. It is noteworthy that amorphous molecular glass is in a thermodynamically nonequilibrium state, and hence, it tends to undergo structural relaxation, exhibiting well-defined Tgs. Amorphous molecular materials have nonplanar molecular structures and take different conformers. The incorporation of bulky and heavy substituents and the enlargement of molecular size make glass formation easier and enhance the stability of the glassy state. The introduction of structurally rigid moieties such as biphenyl, terphenyl, carbazole, and truxene increases the Tg (Higuchi et al., 1992; Kuwabara et al., 1994; Shirota,

Organic materials for optoelectronic applications: Overview

7

2000, 2005). A variety of amorphous molecular materials have been created and have found successful application as materials in OLEDs. The representative classes of amorphous molecular materials include π-electron starburst molecules such as the 4,40 ,400 -tris(diphenylamino)triphenylamine (TDATA) family (7), the 1,3,5-tris(diphenylamino)benzene (TDAB) family (8) and 1,3,5-tris (diphenylamino)triphenylbenzene (TDATB) family (9), a family of spiro-linked molecules (10), and a tetraphenylmethane derivatives family (11) (q.v. first edition, Chapter 1). Amorphous molecular materials containing boron, silicon, or phosphorus atoms [e.g., π-electron systems end-capped with a dimesitylboryl group, TMB-TB (12) (Kinoshita and Shirota, 2001); dimethyloligo(diphenylsilylene)s (Imae and Kawakami, 2005); and tetraarylsilanes (13) (Duan et al., 2005), and POAPF (14) (Chien et al., 2009)] also have been created (Fig. 1.2). Eu(III) coordination compounds having phosphine oxide ligands (Hirai et al., 2015) have been reported to be amorphous molecular materials. The question of whether there exists any microscopic ordering in amorphous molecular materials has been a subject of interest. Studies of this subject have revealed that vacuum-deposited amorphous molecular materials with long molecular structures, such as N,N,N0 ,N0 -tetra(biphenyl-4-yl)-[1,10 -biphenyl]-4,40 -diamine, orient horizontally along with the substrate surface, while m-MTDATA is completely amorphous (Lin et al., 2004; Yokoyama et al., 2008; Yokoyama, 2011).

Fig. 1.2 Representative classes of amorphous molecular materials.

8

Handbook of Organic Materials for Electronic and Photonic Devices

1.2.5 Linear π-conjugated polymers Linear π-conjugated polymers have received attention not only as electrically conducting polymers, but also as polymeric semiconductors for use in OPVs and OLEDs and as nonlinear optical materials. The representative classes of linear π-conjugated polymers are given by polyacetyrene (15), poly(p-phenylene)s (16), poly(p-phenylene vinylene)s (17, PPV), poly(9,9-dialkylfluorene)s (18), polythiophenes (19), polypyrroles (20), polyanilines (21), among other substances (Fig. 1.3). They have been transformed into electrically conducting polymers by chemical and electrochemical doping, the nature of which is charge transfer. A variety of linear π-conjugated polymers have been synthesized by means of cross-coupling reactions using transition-metal catalysts to form CdC and Cdheteroatom bonds.

1.2.6 Polymers containing π-electron systems in side chains Polymers containing π-electron systems as side-chain groups have advantages over linear and planar π-conjugated polymers in both solubility and processing. Polymers having π-electron systems as side-chain groups, which are usually amorphous in nature, have the following characteristic features: ease of molecular design, air stability, solution processing, and various properties and functions based on pendant π-electron systems. Various polymers have been designed and synthesized to date. One example is a family of polymers containing pendant oligothiophenes with varying conjugation lengths (22) (Nawa et al., 1993a,b; Imae et al., 1997) (Fig. 1.4). In contrast to linear π-conjugated polymers, polymers having π-electron systems in the side chain possess essentially the same properties and functions as those of the corresponding small organic molecules based on π-electron systems. Thus, polymers containing side-chain aromatic chromophores are characterized by the absorption and emission of light in the wavelength from near-ultraviolet light to visible light, excitation energy migration over pendant π-electron systems, charge carrier generation and transport resulting in manifestations of electrical conduction or photoconduction, and chromogenic phenomena, among others.

Fig. 1.3 Representative linear π-conjugated polymers: polyacetyrene (15), poly(p-phenylene)s (16), poly(p-phenylene vinylene)s (17), poly(9,9-dialkylfluorene)s (18), polythiophenes (19), polypyrroles (20), and polyanilines (21).

Organic materials for optoelectronic applications: Overview

9

Fig. 1.4 Nonconjugated polymers with pendant oligothiophenes.

1.3

Organic materials for use in optoelectronic devices

Organic materials are used for many kinds of electronic devices, such as organic fieldeffect transistors (OFETs), photoreceptors in electrophotography, OPVs, OLEDs, and photorefractive devices. This section deals with organic materials for use in optoelectronic devices, focusing on OPVs and OLEDs. OPVs are applied for organic solar cells and photodetectors. In particular, thin-film organic solar cells have attracted a great deal of attention as potential candidates for next-generation solar cells. OLEDs are now used for various flat-panel displays such as smartphones and televisions owing to their characteristic features: (1) planar emission; (2) excellent quality image due to high brightness, wide view angle, and good contrast resulting from selfemission; (3) portability; (4) suitability for moving-image display owing to fast response; and (5) capability of full-color emission. OLEDs are also expected to be a promising candidate for next-generation solid-state lighting. Both molecular materials, oligomers with well-defined structures and polymers, are used in optoelectronic devices. Generally, polycrystalline materials might prevent smooth, uniform thin-film formation, which can cause damage to devices. On the other hand, grain boundary-free amorphous materials, including both molecular materials and polymers, allow smooth, uniform amorphous thin-film formation. In addition, amorphous materials are suitable for solution processes owing to their good solubility and capability of forming uniform thin films. In general, either crystalline or amorphous materials are used, and the choice of which one depends on the kind of devices. While polycrystalline materials have been widely used in OPVs and OFETs because of their higher charge carrier mobilities than those of amorphous materials, amorphous molecular materials have been applied successfully in OLEDs. In addition, they have been used for solution-processed bulk heterojunction OPVs.

10

Handbook of Organic Materials for Electronic and Photonic Devices

1.3.1 Molecular materials for organic photovoltaic devices (OPVs) OPVs are based on the idea of the dissociation of electron-hole pairs generated from excitons (i.e., electronically excited-state molecules) to yield charge carriers. One photon absorbed by an organic molecule produces at most one pair consisting of a free hole and an electron, which is transported in the donor and acceptor layers, respectively, and collected at each electrode. Therefore, OPV materials require both electron donor (D) and electron acceptor (A) molecules, which are called p-type and n-type organic semiconductors, respectively. The interface between D and A plays a vital role in the photogeneration of charge carriers. Both D and A materials should meet the following requirements: 1. Materials should absorb as much of the sunlight as possible, and hence should have broad spectral sensitivity from the visible to near-infrared wavelength region, together with large absorption coefficients. 2. Materials with suitable highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) energy levels should be chosen as D and A, respectively, so as to reduce as much as possible the excitation energy loss in the electron-transfer process from the exciton and to give high open-circuit voltage (VOC). 3. Materials should have high charge carrier mobilities for efficient charge separation and transport and reducing series resistance.

A variety of molecular materials for OPVs, including both p-type and n-type organic semiconductors, have been developed. A review of molecular materials for organic solar cells has recently appeared (Mishra and B€auerle, 2012).

1.3.1.1 p-Type organic semiconductors Representative classes of p-type organic semiconductors include metal and metal-free phthalocyanines, such as CuPc (2), and pentacene (1). Organic dyes, such as a squaraine derivative (23) and a rhodanine-based molecule (24), have been reported to give relatively high power conversion efficiencies (PCEs) in combination with fullerenes as n-type organic semiconductors. Materials 25 and 26 also have been reported to function as p-type organic semiconductors (q.v. first edition, Chapter 1). Other p-type organic semiconductors include 27 (van der Poll et al., 2012); 28 (Zhou et al., 2011); 29 (Li et al., 2017); a porphyrin derivative containig diketopyrrolopyrrole units, 30 (Gao et al., 2015); a donor-acceptor-acceptor molecule, 31 (Chiu et al., 2012); and a donor-acceptor-donor molecule, 32 (Sun et al., 2012). Amorphous molecular materials also function well as p-type organic semiconductors, such as tris(oligoarylenyl)amines (33) (Kageyama et al., 2009a,b); 34 (Shang et al., 2011); and π-conjugated moieties end-capped with the triarylamine moiety, such as 35 (Kwon et al., 2010), and 36 (Li et al., 2009). The chemical structures of these materials are shown in Fig. 1.5.

1.3.1.2 n-Type organic semiconductors Fullerenes, such as C60 (4) (Sariciftci et al., 1992, 1993) and C70 (37), and their derivatives, including PC61BM (38) (Yu et al., 1995), PC71BM (39) (Wienk et al., 2003), indene-C60 bisadduct (40) (Zhao et al., 2010), indene-C70 bisadduct (41)

Organic materials for optoelectronic applications: Overview

11

(He et al., 2010), and trimetallic nitride endohedral fullerene (42) (Ross et al., 2009), have been proven to function as excellent n-type organic semiconductors for OPVs (Fig. 1.6). Amorphous molecular materials of fullerene derivatives (e.g., 43), also have been developed (Ohno et al., 2001; Zhang et al., 2009). Other than fullerenes, n-type organic semiconductors include perylene bisimides, such as anthra[200 ,100 ,900 :4,5,6:600 500 1000 :40 ,50 ,60 ]diisoquino[2,1-a:20 ,10 -a0 ] dibenzimidazole-10,21-dione (PTCBI, 44) (Tang, 1986; Peumans et al., 2000),

Fig. 1.5 See legend on next page (Continued)

12

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 1.5, Cont’d p-Type organic semiconductors for OPVs.

diketopyrrolopyrrole (45) (Sonar et al., 2010), pentacene derivative (Shu et al., 2011), 3,9-bis(2-methylene-(3-(1,1-dicyanomethylene)-indanone))5,5,11,11-tetrakis(4-hexylphenyl)-dithieno[2,3-d:20 ,30 -d0 ]-s-indaceno[1,2-b:5,6-b0 ] dithiophene (ITIC, 46) (Lin et al., 2015), 3,9-bis(2-methylene-(3-(1, 1-dicyanomethylene)-6,7-difluoro)-indanone))-5,5,11,11-tetrakis(4-hexylphenyl)dithieno[2,3-d:20 ,30 -d0 ]-s-indaceno[1,2-b:5,6-b0 ]dithiophene (IT-4F, 47) (Zhao et al., 2017), and 2,20 -((2Z,20 Z)-((5,50 -(4,4,9,9-tetrakis(4-hexylphenyl)-4,9-dihydros-indaceno[1,2-b:5,6-b0 ]dithiophene-2,7-diyl)bis(4-((2-ethylhexyl)oxy)thiophene5,2-diyl))bis(methanylylidene))bis(3-oxo-2,3-dihydro-1H-indene-2,1-diylidene)) dimalononitrile (IEICO, 48) (Yao et al., 2016) (Fig. 1.7).

Organic materials for optoelectronic applications: Overview

13

Fig. 1.6 Fullerene derivatives as n-type organic semiconductors for OPVs.

Fig. 1.7 n-Type organic semiconductors other than fullerenes for OPVs.

1.3.2 Polymers for use in OPVs Polymer-based OPVs fabricated by solution processes are usually single-layered bulk pn-heterojunction devices consisting of a blend of D and A because of the difficulty of fabricating layered structures. The blend system is composed of either the

14

Handbook of Organic Materials for Electronic and Photonic Devices

combination of electron-donating and -accepting linear π-conjugated polymers to give interpenetrating network polymers (Halls et al., 1995), or the combination of linear π-conjugated polymers as D and small molecules as A to give bulk heterojunction OPVs. Examples of interpenetrating network polymers include the combination of MEHPPV (49) + cyano-substituted PPV derivatives (50, CN-PPV) (Halls et al., 1995); P3HT (51) + P(NDI2OD-T2) (52) (Schubert et al., 2012); P3HT (51) + PNDIBS (53) (Hwang et al., 2012); PTB7-Th (54) + PNDIS-HD (55) (Hwang et al., 2015); PTB7-Th (54) + NDP-V (56) (Guo et al., 2017); and PTzBI-Si (57) + P(NDI2ODT2) (52) (Fan et al., 2017) (Fig. 1.8). The blend systems are composed of the combination of linear π-conjugated polymers as D and small molecules as A, including MEH-PPV (49) or regioregular P3HT (51) (McCullough and Lowe, 1992) as D in combination with small molecules such as fullerenes (C60 and C70) and their derivatives, such as PC61BM (38) and PC71BM (39), as A. The magnitude of bandgap energies (Egs) in linear π-conjugated polymers is an important factor that determines their optical and electronic properties. A variety

Fig. 1.8 π-Conjugated polymers for OPVs.

Organic materials for optoelectronic applications: Overview

15

Fig. 1.9 Buliding blocks for low-bandgap polymers.

of low bandgap-energy π-conjugated polymers have been designed and synthesized for use as D in OPVs with the idea of harvesting photons in the red and near-infrared region of the solar spectrum. Low-bandgap, electron-donating π-conjugated polymers hitherto reported have chemical structures containing both D and A moieties. Such polymers are also effective for reducing the photoenergy loss that inevitably accompanies the process of electron-transfer from D to A in the electronically excited state. Several building blocks for the D and A moieties in low-bandgap polymers are shown in Fig. 1.9. While thienothiophene (TT), benzodithiophene (BDT), cyclopentadithiophene (CDT), and dithienopyran (DTP) moieties constitute building blocks for D, diketopyrrolopyrrole (DPP), benzothiadiazole (BT), and difluorobenzothiadiazole (DFBT) moieties constitute building blocks for A. Fig. 1.10 shows the chemical structures of several representative low-bandgap polymers that act as D in OPVs. On the other hand, wide- or middle-bandgap polymers with absorption bands at 300–500 nm have also been developed. They are used in combination with lowbandgap acceptors with the absorption bands at 500–900 nm. Examples of these polymers are shown in Fig. 1.11. Tuning the interface between the electrode and the active layer is essential for attaining high device performance. Poly(4-styrene sulfonate)-doped poly(3,4ethylenedioxythiophene) (PEDOT:PSS, 66) has been widely used as an anode interfacial material in OPVs. Other interfacial polymers have been developed as well. A blend of 4,40 -bis{[p-(trichlorosilylpropyl)phenyl]phenylamino}biphenyl covalently bonded to the anode and poly{9,9-dioctylfluorene-co-N-[4-(3-methylpropyl) phenyl]diphenylamine} has been reported to act as an efficient hole-extraction/electron-blocking layer (Hains and Marks, 2008). Inorganic salts, such as MoO3 (Shrotriya et al., 2006), WO3 (Chan et al., 2006a), and V2O5 (Li et al., 2006; Shrotriya et al., 2006), also have been used as anode interlayer materials. Regarding cathode interlayer materials, inorganic salts, such as LiF (Brabec et al., 2002), CsF

16

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 1.10 Low-bandgap polymers for OPVs.

Fig. 1.11 Wide- and middle-bandgap polymers for OPV.

( Jiang et al., 2009), Cs2CO3 (Li et al., 2006), metal oxides such as TiOx (Kim et al., 2006; Hayakawa et al., 2007), ZnO (Umeda et al., 2003; White et al., 2006), and MoO3 (Kageyama et al., 2011), have been used. In addition, low-molecular-weight compounds, including bathocuproine (BCP, 67) (Peumans et al., 2000) and 4,7diphenyl-1,10-phenanthroline (BPhen, 68) (Chan et al., 2006a,b), and polymers such as ethoxylated polyethylenimine (PEIE, 69) (Zhou et al., 2012) and PFN (70) (He et al., 2011) have been used as interlayer materials (Fig. 1.12).

Organic materials for optoelectronic applications: Overview

17

Fig. 1.12 Interlayer materials used for OPVs.

1.3.3 Molecular materials for organic light-emitting diodes (OLEDs) Organic electroluminescence (EL) is based on the fluorescence or phosphorescence emission orginating from the electronically excited singlet or triplet state (singlet or triplet exciton) generated by the recombination of holes and electrons in the device. Holes and electrons are injected from the electrodes (the anode and cathode, respectively) by applying an external electric field to the device. For single-layer devices, the material should have the properties of accepting both holes and electrons from the electrodes and transporting them as well as emissive properties. Because of the difficulty of providing one material with these multiple functions, several materials with each function of charge acceptance from the electrodes, charge transport, charge blocking, and emission are employed together in devices. They are called hole-transporting, electron-transporting, chargeblocking, and emitting materials, respectively. The structure of OLEDs consists of multilayers that usually are fabricated by vacuum deposition of these various materials. Generally, materials for use in OLEDs should meet the following requirements: 1. They should form uniform, homogeneous, pinhole-less thin films by either thermal deposition in vacuum or by solution processing. 2. They should have thermal stability with high Tgs. 3. They should be morphologically stable.

Amorphous molecular materials that form smooth, uniform, amorphous films have proved to be well suited for use in OLEDs. A variety of hole- and electrontransporting, hole- and electron-blocking, and emitting amorphous molecular materials have been designed and synthesized.

18

Handbook of Organic Materials for Electronic and Photonic Devices

1.3.3.1 Hole-transporting materials Hole-transporting materials should meet the following requirements. They should have proper HOMO energy levels so they may accept hole carriers from the anode through the hole-injection layer. The injected hole carriers (i.e., the cation-radical species of hole-transporting molecules) should be stable. In other words, the anodic oxidation processes of hole-transporting molecules should be reversible. They should have hole-transporting ability, along with electron-blocking. Hole carriers injected from the anode are then injected into the emitting layer in a stepwise process via the hole-transport layer. A variety of hole-transporting amorphous molecular materials have been developed. Of the hole-transporting materials, those with very low ionization potential can be used for the hole-injection layer in contact with the anode. Representative examples are 4,40 ,400 -tris(3-methylphenyl(phenyl)amino)triphenylamine (m-MTDATA, 71) (Shirota et al., 1989), 4,40 ,400 -tris(1- or 2-naphthylphenylamino) triphenylamine (1- and 2-TNATA, 72) (Shirota et al., 1997), 4,40 ,400 -tri-Ncarbazolyltriphenylamine (TCTA, 73) (Kuwabara et al., 1994), and α-NPD (74) (Van Slyke et al., 1996) (Fig. 1.13). An example of a hole-transporting amorphous molecular material with vey high Tgs is TFATr (75) (Tg: 208°C) (Shirota, 2005). In addition, m-MTDATA (71) and 1- and 2-TNATA (72), which are characterized by very low solid-state ionization potentials of about 5.10 eV, have been used as holeinjection materials for OLEDs (Shirota et al., 1994; Murata et al., 1999). CuPc (2) has also been used as a hole-injection material (Van Slyke et al., 1996). TCTA (73) and α-NPD (74) have been widely used as hole-transporting, amorphous molecular materials for OLEDs.

Fig. 1.13 Hole-transporting amorphous molecular materials for the OLED-TDATA family.

Organic materials for optoelectronic applications: Overview

19

1.3.3.2 Electron-transporting materials Electron-transporting materials should possess the following properties. They should have proper LUMO energy levels so as to accept electron carriers injected from the cathode through the electron-injection layer. Injected electron carriers (i.e., the anionradical species of electron-transporting molecules) should be stable. In other words, the cathodic reduction processes of electron-transporting molecules should be reversible. They should have electron-transporting ability along with hole-blocking. Electron carriers injected from the cathode are then injected into the emitting layer in a stepwise process via the electron-transport layer. Various kinds of electron-transporting amorphous molecular materials have been developed. They possess central cores such as benzene, 1,3,5-triphenylbenzene, 1,3,5triazine, and tetraphenylmethane, to which electron-accepting moieties are attached. Electron-accepting moieties include oxadiazole, pyridine, triazine, silole, dimesitylboryl, and triarylborane moieties. Green-emitting Alq3 has been extensively used as an excellent electrontransporting material in OLEDs. Electron-transporting materials containing oxadiazole, triazole, and phenylbenzimidazole moieties [e.g., TPBI (76) (Gao et al., 1999)] have been developed (Kulkarni et al., 2004; Hughes and Bryce, 2005). Silole derivatives also have been shown to be excellent electron-transporting materials for OLEDs; these include PySPy (77) (Tamao et al., 1996) and PyPySPyPy (78) (Palilis et al., 2003a,b). The electron mobility of PyPySPyPy (78) is 2  104 cm2/Vs at 6.4  105 V/cm, which is more than two orders of magnitude higher than that of Alq3 (Murata et al., 2001). Amorphous molecular materials containing a dimesitylboryl moiety, BMB-nT (n ¼ 2, 3) (79), have been found to be good electron-transporting materials owing to the vacant p-orbital of the boron atom (Noda and Shirota, 1998; M€akinen et al., 2001). Other electron-transporting amorphous molecular materials containing a dimesitylboryl group have been developed as well (Kinoshita and Shirota, 2001; Li et al., 2007; Sun et al., 2011). Like hole-transporting materials with very low ionization potential for use as the hole-injection layer, electron-transporting materials with strong electron-accepting properties, BMB-2T, BMB-3T (79, n ¼ 2, 3), and PySPy (77), have been shown to function as the electron-injection layer that facilitates electron injection from the cathode (Noda and Shirota, 1998; Tamao et al., 1996). In addition, a number of triarylboranes have been developed (Yamaguchi et al., 2000a,b; Kinoshita et al., 2002; Tanaka et al., 2007a). A variety of amorphous molecular materials containing a pyridine moiety, such as B3PyPB (80) (Sasabe and Kido, 2011) and 2,4,6-triaryl substituted 1,3,5-triazines, have been developed as well. The molecular structures of these electron-transporting amorphous molecular materials (7680) are shown in Fig. 1.14.

1.3.3.3 Hole-blocking materials Charge-blocking layers are used to confine both hole and electron charge carriers within the emitting layer by preventing the escape of charge carriers from the emitting layer. Hole-transporting and electron-transporting materials naturally have

20

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 1.14 Electron-transporting amorphous molecular materials for OLEDs.

electron-blocking and hole-blocking properties, respectively. Hole-blocking materials should have great ionization potential to prevent the acceptance of hole carriers escaping from the emitting layer, and weak electron-accepting properties so as to accept electron carries injected from the electron-transporting layer and pass them to the emitting layer. They should not form any exciplex with the emitting materials with electron-donating properties (Kinoshita et al., 2002). Materials that have been used as the hole-blocking layer in OLEDs include BCP (67), 1,3,5-tris(4-fluorobiphenyl-40 -yl)benzene (F-TBB), X-branched oligophenylenes, tetra(β-naphthyl)silane, and others (q.v. first edition, Chapter 1).

1.3.3.4 Emitting materials The emitting layer functions as the recombination center for holes and electrons injected from the anode and cathode, respectively. Therefore, emitting materials should possess a bipolar character, with proper HOMO and LUMO energy levels so as to accept both holes and electrons injected through the hole- and electrontransport layers, respectively. In addition, injected charge carriers (namely, the cation- and anion-radical species of emitting molecules) should be stable; in other words, the anodic oxidation and cathodic reduction processes of emitting molecules should be reversible. Furthermore, emitting materials should have high-luminescence quantum efficiencies. When emitting materials with high quantum efficiencies for either fluorescence or phosphorescence lack good film-forming ability, they can be used as emitting dopants by being dispersed in the host amorphous molecular materials.

1.3.3.5 Fluorescence-emitting materials Electron-transporting materials (e.g., BMB-2T, PyPySPyPy, and 2PSP) have been shown to function as emitting materials with high-fluorescence quantum efficiency. In particular, the fluorescence quantum efficiency for 2PSP as vacuum-deposited solid films has been reported to be 100% (q.v. first edition, Chapter 1).

Organic materials for optoelectronic applications: Overview

21

Fig. 1.15 Bipolar-emitting amorphous molecular materials.

Emitting amorphous molecular materials with bipolar character are given by BMA-nT (81) (Noda et al., 1997a,b, 1999) and FlAMB-nT (82) (Shirota et al., 2000; Doi et al., 2003). The emission color varies according to the conjugation length of the oligothiophene moiety in BMA-nT and FlAMB-nT. Another bipolar-emitting material containing dimesitylboryl and arylamino groups (83) has been developed as well (Pan et al., 2011). BFA-BT, containing diarylamino and benzothiadiazole moieties (84), and BFA-An (85), having diarylamino and diphenylanthracene moieties, function as orange- and blue-emitting materials, respectively (Shirota, 2005) (Fig. 1.15). A triphenylamine-oxadiazole-fluorene triad molecule and its analogs emit light ranging from light blue to green color, depending on the modification of the molecular subunits (Kamtekar et al., 2006).

1.3.3.6 Thermally assisted delayed fluorescence-emitting materials High-performance fluorescence-based OLEDs utilizing thermally activated, delayed fluorescence (TADF) generated by the thermal upconversion from the triplet exciton to the singlet exciton have been developed (Endo et al., 2009). TADF-emitting materials, which are called third-generation OLED emitters (Adachi, 2014), have an advantage over phosphorescent dopant materials, in that they do not require expensive transition metals. Both small molecules (Endo et al., 2011; Uoyama et al., 2012; Zhang et al., 2012, 2014; Lee et al., 2014; Suzuki et al., 2015; Wada et al., 2015; Kaji et al., 2015; Gibson et al., 2016) and polymers (Nikolaenko et al., 2015; Albrecht et al., 2015; Ren et al., 2016; Luo et al., 2016; Lee et al., 2016) that emit TDAF have been developed. Materials that emit TADF should possess both large spin-orbit interaction and a small singlet-triplet energy split so that reverse intersystem crossing should be promoted (Tao et al., 2014). TADF-emitting molecules are designed to contain both electron-donating and electron-accepting moieties to give intramolecular charge-transfer character. The reported TADF-emitting small molecules contain

22

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 1.16 TADF-emitting molecular materials.

1,3,5-triphenyltriazine, cyanobenzene, triarylboron, diphenylsulfone, or diphenylketone moieties as electron-accepting units and carbazole, 2,7-diphenylaminocarbazole, acridan or phenoxazine moieties as electron-donating units. The following compounds have been designed and synthesized: 2-biphenyl-4,6-bis(12phenylindolo[2,3-a]carbazole-11-yl)-1,3,5-triazine (PIC-TRZ, 86) (Endo et al., 2011); 1,2,3,5-tetrakis(carbazol-9-yl)-4,6-dicyanobenzene (4CzIPN, 87) (Uoyama et al., 2012); 4,40 -(9,9-dimethyl-9,10-dihydroacridino)-1,10 -diphenylsulfone (DMAC-DPS, 88) (Zhang et al., 2014), triarylboron-based molecules (PXZ-Mes3B, 89) (Suzuki et al., 2015), 9-[4-(4,6-diphenyl-1,3,5-triazin-2-yl)phenyl]-N,N,N0 , N0 -tetraphenyl-9H-carbazole-3,6-diamine (DACT-II, 90) (Kaji et al., 2015) (Fig. 1.16). These TADF-emitting materials have been used as dopants by being dispersed in the host material. It has been shown that reverse intersystem crossing takes place effectively from the local excited triplet state (3LE) to the excited charge-transfer singlet state (1CT) when vibronic coupling exists between 3LE and the lowest excited charge-transfer triplet state (3CT) (Gibson et al., 2016; Hosokai et al., 2017).

1.3.3.7 Phosphorescence-emitting materials Room-temperature phosphorescent materials hitherto used in OLEDs are transitionmetal complexes. As the transition-metal complexes usually do not have film-forming ability, they have been used as emissive dopants, being dispersed in a host material with good film-forming ability. They include platinum complexes, such as PtOEP (91); iridium complexes, such as Ir(ppy)3 (92); btp2Ir(acac) (93); (ppy)2Ir(acac) (94); FIrpic (95); and osmium complexes, such as 96, [Os(bpy)2L)]2+(PF 6 )2 (97) and Os(fptz)2(PPh2Me)2 (98) (Fig. 1.17) (q.v. first edition, Chapter 1). While PtOEP and osmium complexes are red emitters, iridium complexes release red, green,

Organic materials for optoelectronic applications: Overview

23

Fig. 1.17 Phosphorescent dopants for OLEDs.

and blue emissions depending on the structures of the ligands. Ir(ppy)3 and (ppy)2Ir (acac) are green emitters, and btp2Ir(acac) and FIrpic are red and blue emitters, respectively.

1.3.3.8 Host materials for emissive dopants Emissive dopants in OLEDs require host materials with good film-forming ability. The selection of host materials as the emitting layer is crucial. Materials that function as hosts for emissive dopants usually play a role as a charge recombination center to generate excitons, and then transfer the excitation energy to emissive dopants via singlet-singlet or triplet-triplet energy transfer. Consequently, host materials for emissive dopants should have a bipolar character so as to accept both hole and electron carriers, and their excited singlet and triplet energy levels should be higher than those of the emissive dopants in order for the exothermic energy transfer to take place efficiently and for the reverse energy transfer from the dopant to the host to be prevented. In particular, host materials for blue-phosphorescent dopants are required to possess high triplet-energy levels. Host materials such as CBP (99), TCTA (73), and TDAPB (100) for green-emitting dopants, SimCP (101), and UGH2 (105) for blue-emitting dopants, and Alq3, and TFTPA (102) for red-emitting dopants have been developed (q.v. first edition, Chapter 1). Amorphous molecular materials containing triphenylphosphine oxide, POAPF (14) (Chien et al., 2009), bis{2-[di(phenyl)phosphino]phenyl}ether oxide (DPEPO, 103) (Han et al., 2011), and bis-4-(N-carbazolyl)phenyl)phenylphosphine (BCPO, 104) (Chou and Cheng, 2010) have high triplet-energy levels of 2.995, 2.62, and 3.01 eV, respectively, and function as universal host materials for red-, green-, and blue-emitting materials (Fig. 1.18).

24

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 1.18 Amorphous molecular host materials for phosphorescent dopants.

1.3.4 Polymers for use in OLEDs OLEDs using molecular materials mostly consist of multilayers such as holetransporting, emitting, hole-blocking, and electron-transporting layers so as to attain high brightness, high power, and quantum efficiencies. Such thin-film multilayers are usually fabricated by vacuum deposition of amorphous molecular materials. By contrast, polymer films, which are usually fabricated by solution processing such as spin coating and inkjet printing, hamper the fabrication of layered structures because of the solubility problem; that is, the preparation of the second layer dissolves the first layer that has already formed. This restriction, however, ideally leads to all solution-processed single-layer OLEDs, where the polymer films used should have multiple functions of charge transport and emission.

1.3.4.1 Linear π-conjugated polymers Linear π-conjugated polymers, such as PPV (17) and substituted PPVs (in particular, MEH-PPV (49), PPV copolymers), poly(9,9-dialkylfluorene)s (18), and their copolymers, have been intensively studied for use as materials in OLEDs. Linear π-conjugated polymers possess both charge-transporting and emitting properties, enabling single-layered devices.

Organic materials for optoelectronic applications: Overview

25

1.3.4.2 Nonconjugated side-chain polymers Polymers having π-electron systems as side-chain groups generally possess better solubility and air stability than linear π-conjugated polymers. Vinyl, acrylate, and phosphazene polymers and their copolymers having charge-transporting and/or emissive side chains have been synthesized. Layered OLEDs using these polymers and small molecules by combining the solution process and vacuum deposition have been fabricated. Hole-transporting polymers containing pendant π-electron systems include a vinyl polymer containing an N,N0 -bis(3-methylphenyl)-N,N0 -diphenyl-[1,10 -biphenyl]-4,40 diamine (TPD) moiety as a pendant group; poly{4-(m-tolyl-m-fluorophenylamino)40 -(m-fluorophenyl-p-vinylphenylamino)biphenyl} (106, P3) (Shaheen et al., 1999; Bellmann et al., 1999) and vinyl polymers containing triarylamine moieties as pendant groups; and poly{4-vinyl-40 -[bis(40 -tert-butylbiphenyl-4-yl)amino]biphenyl} (PVBAB, 107) and poly{4-vinyl-40 -[N,N-bis(9,9-dimethylfluoren-2-yl)amino]biphenyl} (PVFAB, 108) with high Tgs of 229°C and 204°C (Mutaguchi et al., 2003). Electron-transporting polymers include vinyl polymers with such side groups as trifluoromethyl-substituted quaterphenyl (109) (Pommerehne et al., 1997) and 2,5diphenyl-1,3,4-oxadiazole (110) (Dailey et al., 2001) (Fig. 1.19). Emitting polymers with bipolar character, poly(2-{4-[4-vinylphenyl(4-methylphenyl)amino]phenyl}-5-dimesitylborylthiophene) (PVPhAMB-1T) with a glass-transition temperature of 194°C (Mutaguchi et al., 2003) and poly[4(7-[4-[N,N-bis(9,9-dimethylfluoren-2-yl)amino]phenyl]-2,1,3-benzothiadiazol-4-yl) phenylethene] (PVFABT, 111) with a glass-transition temperature of 199°C (Nagamatsu et al., 2005) have been developed (Fig. 1.19). PVPhAMB-1T emits green fluorescence with a peak wavelength of 521 nm and a quantum efficiency of 0.58 in

Fig. 1.19 Hole-transporting side-chain polymers, P3, PVBAB, and PVFAB, electrontransporting side-chain polymers (109 and 110), and emitting side cain polymer, PVFABT (111).

26

Handbook of Organic Materials for Electronic and Photonic Devices

solution (Mutaguchi et al., 2003). PVFABT emits orange fluorescence with a peak wavelength of 653 nm and a quantum efficiency of 0.65 in solution (Nagamatsu et al., 2005). Polymers that emit TADF are exemplified by a block copolymer with triazineamine-triazine emitter units (112) (Nikolaenko et al., 2015) and a copolymer of styrene with a vinyl monomer having a pendant TADF unit (113) (Ren et al., 2016). Other TADF-emitting materials include a dendrimer type molecule (114) (Albrecht et al., 2015), copolymers with a backbone of polycarbazole (Luo et al., 2016), and polymers with a backbone consisting of electron-donating and accepting units (115) (Lee et al., 2016) (Fig. 1.20). A triphenyamine-based polysiloxane (116, PTPAMSi) with good amorphous filmforming ability and a sufficiently high triplet-energy level of 2.9 eV has been reported to serve well as a host material for a phosphorescent iridium complex (Sun et al., 2014a). A bipolar alternating copolymer containing phenylcarbazole and triphenylphosphine oxide moieties linked to the backbone of polysiloxane, poly(phenylcarbazole-alt-triphenylphosphine oxide)siloxane (PCzPOMSi, 117) has

Fig. 1.20 TADF-emitting dendrimer and polymers.

Organic materials for optoelectronic applications: Overview

27

Fig. 1.21 Side-chain host polymers, PTPAMSi and PCzPOMSi.

Fig. 1.22 Example of a multifunctional side-chain polymer.

a relatively high glass-transition temperature of 118°C and a sufficiently high tripletenergy level of 3.0 eV (Sun et al., 2014b) (Fig. 1.21). Copolymers with an emissive iridium complex and a hole-transporting carbazolyl moiety as side chains have been synthesized (Fei et al., 2010; Haldi et al., 2008). Toward the fabrication of all solution-processed, single-layer OLEDs, side-chain polymers containing hole-transporting, electron-transporting, and emitting moieties have been designed and synthesized. A new class of emitting polymers composed of a liquid-crystal-forming moiety, an emitting 1,3,4-oxadiazole moiety with electron-transporting character, and a hole-transporting arylamine moiety in the same side chain have been synthesized (Mochizuki et al., 2000a,b, 2003). A single-layer OLED using a liquid-crystalline polymer sandwiched between indium tin oxide (ITO) and MgAg electrodes has been reported to exhibit polarized blue EL emissions, with an emission peak wavelength of 458 nm. The origin of the polarized emission has been attributed to the anisotropic arrangement of the mesogenic chromophore (Kawamoto et al., 2003). A vinyl copolymer having a hole-transporting TPD moiety, an electrontransporting 2-(4-biphenyl)-5-(4-tert-butylphenyl)-1,3,4-oxadiazole (PBD) moiety, and an emitting iridium complex in different side chains (118) (Fig. 1.22) has been reported. A single-layer OLED using this multifunctional side-chain polymer has been reported to give a high external quantum efficiency of up to 11.8% (Suzuki et al., 2005).

28

1.4

Handbook of Organic Materials for Electronic and Photonic Devices

Fabrication and performance of organic optoelectronic devices

1.4.1 OPVs Since the pioneering work of Tang (1986), extensive studies have been performed on thin-film OPVs to improve PCEs. The creation of novel molecular materials and polymers and the implementation of new device structures, such as a planar pnheterojunction structure with the insertion of an exciton-blocking layer (Peumans and Forrest, 2001) or optical spacers (Hayakawa et al., 2007; Park et al., 2009), pn-heterojunction tandem structure (Yakimov and Forrest, 2002), a p-i-n structure (Xue et al., 2005), and a bulk pn-heterojunction structure (Sariciftci et al., 1992; Yu et al., 1995; Halls et al., 1995), have led to remarkable improvement in the PCE. With regard to materials, the development of molecular materials of n-type organic semiconductors, including fullerenes (C60 or C70), fullerene derivatives such as [6,6]phenyl-C61-butyric acid mehyl ester (PC61BM) and [6,6]-phenyl-C71-butyric acid methyl ester (PC71BM)), nonfullerene electron-accepting molecules, such as indacenodithieno[3,2-b]thiophene derivatives, linear π-conjugated polymers, such as poly(2-methoxy-5-(20 -ethylhexyloxy)-1,4-phenylenevinylene) (MEH-PPV) and poly(3-hexylthiophene) (PH3T), and a variety of low bandgap π-conjugated polymers as p-type polymeric semiconductors, is noteworthy. A deeper understanding of device physics (Bredas et al., 2009) has contributed to the attainment of high PCEs as well. An OPV having a p-i-n structure with the insertion of a thin mixed layer (i.e., 10 nm) between the two layers of CuPc (D) and C60 (A) gave a PCE of 5.0% (Xue et al., 2005). Planar pn-heterojunction or bulk-heterojunction OPVs using molecular materials such as a squaraine derivative (23), a rhodanine-based molecule (24), a donor-acceptor-acceptor molecule (31), p-DTS(FBTTh2)2 (27), 28, and DTS(PTTh2)2 (32), in combination with fullerenes (C60 and C70) and fullerene derivatives (PC61BM and PC71BM), exhibited PCEs of 6%–7% (Zhou et al., 2011; Chen et al., 2012; Li et al., 2012; Chiu et al., 2012; van der Poll et al., 2012; Sun et al., 2012). Planar pn-heterojunction OPVs using the amorphous molecular material 33 as D and C60 or C70 as A gave PCEs of 1.5%–2.2% (Kageyama et al., 2009a,b). The material has been reported to exhibit the highest-level hole drift mobilities of about 1.0  102 cm2/Vs among amorphous molecular materials (Ohishi et al., 2004). Recently, PCEs of up to 11% for single-junction and 12% for tandem OPVs under simulated sunlight illumination have been attained by the use of molecular materials. That is, a ternary, single-junction OPV using molecular materials, ITO/PEDOT:PSS (30 nm)/p-DTS(FBTTh2)2 (27) + ZnP (30) + PC71BM (3: 2: 5 w/w)/PFN (70) (5 nm)/ Al (100 nm), showed a JSC, a VOC, an FF, and a PCE of 17.99 mA/cm2, 0.79 V, 0.7719, and 10.97%, respectively (Nian et al., 2017). A double-junction OSC using molecular materials, ITO/CuSCN (35 nm)/DR3TSBDT (29) + PC71BM (1:0.8 w/w) (125 nm)/ ZnO (40 nm)/n-PEDOT:PSS (about 40 nm)/ZnP (30) + PC61BM (1:1 w/w) (125 nm)/PFN (70) (5 nm)/Al (70 nm), exhibited a JSC, a VOC, an FF, and a PCE of 12.05 mA/cm2, 1.625 V, 0.627, and 12.28%, respectively (Li et al., 2017).

Organic materials for optoelectronic applications: Overview

29

Further, growing attention has been paid to bulk-heterojunction OPVs fabricated by the solution process, which have been given increasingly high PCEs by the use of soluble fullerene derivatives such as PC61BM and PC71BM. A solutionprocessed, bulk-heterojunction OPV using MEH-PPV as D and PC61BM as A gave 2.9% under 430-nm monochromatic light illumination (Yu et al., 1995). A bulkheterojunction OPV with the additional layers of 2,3,5,6-tetrafluoro-7,7,8,8tetracyanoquinodimethane (F4-TCNQ)-doped m-MTDATA (71) and rhodamine B-doped dimethyl-perylene-tetracarboxylic-diimide (MPP) exhibited a PCE of about 3.37% (Gebeyehu et al., 2003). Solution-processed bulk-heterojunction OPVs using MEH-PPV as D and the fullerene derivatives as A gave about 3% (Alem et al., 2004). Polymer-based bulk pn-heterojunction OPVs using the combination of an electrondonating π-conjugated polymer, P3HT, and a small molecular fullerene derivative, PC61BM, as an electron acceptor has given PCEs ranging from 5% (Ma et al., 2005) to 6.1% (Kim et al., 2007). Amorphous molecular materials are suitable for the fabrication of solution-processed bulk-heterojunction OPVs owing to their good film-forming ability. A PCE of 4.3% has been attained for solution-processed bulkheterojunction OPV using 34 (Shang et al., 2011). Higher PCEs achieved for bulk-heterojunction OPVs than those for planar pnheterojunction OPVs have been ascribed to the much higher JSC, which stems from the enhanced generation of the hole-electron pair as a result of the larger probability of the photogenerated excitons reaching the D/A interface within their lifetimes. Morphology control to form phase-separated structures to form electric conduction paths in the direction perpendicular to the electrodes is essential (Ma et al., 2005) Significantly, the creation of low-bandgap polymers has led to the improvement in PCEs. Bulk heterojunction OPVs using low-bandgap electron-donating polymers such as PBDTTT-CF (60), PTB7 (58), PTB7-Th (54), and PDTP-DFBT (61), in combination with electron acceptors such as PC61BM and PC71BM and an adequate interfacial layer such as PFN (70), have given PCEs of 7.73% (Chen et al., 2009), 7.4% (Liang et al., 2010), 8.3% (He et al., 2011), and 12.25% (Huang et al., 2017) for single cells. An OPV consisting of double-bulk heterojunction layers, ITO/ZnOx/PTB7-Th (54) + PC71BM (1:1.5 w/w) (200 nm)/PDPP3T (62) + PC71BM (1:1.2 w/w) (50 nm)/ MoO3/Ag, showed a JSC of 23.75 mA/cm2, a VOC of 0.77 V, an FF of 0.67, and a PCE of 12.25%, respectively (Huang et al., 2017). The following two single-junction OPVs using wide- or middle-bandgap polymers in combination with low-bandgap nonfullerene acceptors, ITO/PEDOT:PSS (40 nm)/ PTzBI-Si (57) + P(NDI2OD-T2) (52) (1:0.5 w/w) (140 nm)/cathode interlayer (about 5 nm)/Ag (90 nm) (Fan et al., 2017) and ITO/ZnO (100 nm)/PBDB-T-SF (64) + IT-4F (47) (1:1.5 w/w) (about 100 nm)/MoO3 (10 nm)/Al (100 nm) (Zhao et al., 2017), showed JSCs of 15.57 and 20.88 mA/cm2, VOCs of 0.87 and 0.88 V, FFs of 0.7339 and 0.713, and PCEs of 10.1% and 13.10%, respectively. Double-junction OPVs using wide- or middle-bandgap polymers in combination with the low-bandgap nonfullerene acceptors, ITO/PEDOT:PSS (10 nm)/PBDD4T-2F (63) + PC71BM (90 nm)/ZnO (30 nm)/n-PEDOT:PSS (15 nm)/PBDTTT-E-T (59) + IEICO (48) (95 nm)/cathode interlayer/Al (Yao et al., 2016) and ITO/ZnO (30 nm)/PBDTS-TDZ (65) + ITIC

30

Handbook of Organic Materials for Electronic and Photonic Devices

(46) (80 nm)/PEDOT:PSS (50 nm)/Ag (5 nm)/ZnO (30 nm)/PBDTS-TDZ (65) + ITIC (46) (100 nm)/MoO3 (10 nm)/Ag (100 nm) (Xu et al., 2018), showed JSCs of 10.3 and 9.77 mA/cm2, VOCs of 1.70 and 2.13 V, FFs of 0.61 and 0.641, and PCEs of 10.7% and 13.35%, respectively. Meanwhile, perovskyte solar cells using hybrid materials of inorganic and organic substances have attracted a great deal of attention because of their very high PCEs (Kojima et al., 2009). Organometal halide perovskite compounds with a general chemical structure of ABX3, where A, B, and X stand for methyl ammonium cation, inorganic divalent or trivalent cation (e.g., Pb2+, Sn2+, Bi3+, Sb3+), and negativelycharged halide, respectively, possess excellent photovoltaic properties: (1) strong absorption of visible light owing to a small band gap of about 2.3 eV, (2) weak exciton binding energy of about 0.03 eV, and (3) high charge carrier mobility of 7.5 cm2/Vs for electrons and 12.5 cm2/Vs for holes. In addition, perovskite solar cells can be fabricated by printing methods. The PCEs of perovskite solar cells have been improved quite significantly, from 3.8% in 2009 (Kojima et al., 2009) to 20.2% in 2015 (Yang et al., 2015) and 22.7% in 2017 (NREL Efficiency Chart, 2017). A perovskite solar cell, FTO/dense TiO2 (60 nm)/mesoporous TiO2 (150 nm)/ (NH2CH ¼ NH2PbI3)0.85(CH3NH2PbBr3)0.15/polytriarylamine/Au, showed JSC ¼ 24.7 mA/cm2, VOC ¼ 1.06 V, FF ¼ 0.775, and PCE ¼ 20.2% under the illumination of AM1.5G at an intensity of 100 mW/cm2 (Yang et al., 2015).

1.4.2 OLEDs Since the appearance of the pioneering reports toward the practical use of OLEDs using molecular materials (Tang and Van Slyke, 1987) and polymers (Burroughes et al., 1990), extensive research and development on OLEDs have been conducted with the goal of achieving higher power and quantum efficiencies for EL, longer operational durability, and full-color and white-color emissions. Our understanding of the operational processes (e.g., charge injection from electrodes into the organic layers, charge transport in the organic layer, recombination of holes and electrons to produce the electronically excited singlet and triplet state, excitation energy transfer from host materials to emissive dopants, up-conversion from the excited triplet state to the singlet state, and exciplex formation at the solid/solid interface) has greatly deepened. OLEDs can be classified into fluorescence- and phosphorescence-based devices for the sake of convenience. Hole-transporting, electron-transporting, and chargeblocking materials can be used for both devices.

1.4.2.1 Fluorescence-based OLEDs With regard to fluorescence-based OLEDs, the EL external quantum efficiencies (EQEs) close to or over 5%, which is assumed to be a theoretical limit for the fluorescence-based OLEDs when the light outcoupling efficiency is assumed to be about 20%, has been achieved (Murata et al., 2002; Wu et al., 2004). The fabrication and performance of various kinds of OLEDs based on fluorescence and exciplex emission including the use of fluorescent dopants such as DCM2, Nile Red, coumarin 6, DPP, and ACY as red dopants,

Organic materials for optoelectronic applications: Overview

31

rubrene as a yellow dopant, DMQA as a green dopant, BCzVBi and perylene as blue dopants are fully discussed in Chapter 1 of the first edition.

1.4.2.2 Phosphorescence-based OLEDs Phosphorescence-based OLEDs should give in principle much higher EQEs than the conventional fluorescence-based OLEDs since the statistical probability ratio of the generation of the excited triplet state to the excited singlet state by the recombination of holes and electrons is 3:1.The finding of room-temperature phosphorescent transition metal complexes (e.g., osmium complexes, platinum complexes, and iridium complexes) has led to the development of high-performance phosphorescent OLEDs, evolutionizing the improvement in luminous and quantum efficiencies of OLEDs. The EQEs for phosphorescence-based devices tend to decrease with the increasing injected current because of the occurrence of the triplet-triplet annihilation resulting from the longer lifetimes of phosphorescence relative to those of fluorescence (Baldo et al., 1998). Therefore, the lifetimes of phosphorescent dopants should be short. Among the transition metal complexes, the iridium complexes have been thought to be promising candidates for phosphorescent dopants. Their phosphorescent lifetimes [0.5–1.3 μs for Ir(ppy)3 in CBP (Baldo et al., 1999; Endo et al., 2008) and 1.2 μs for FIrpic in mCB (Endo et al., 2008)] are shorter than those of the Pt complexes (91 μs for PtOEP in polystrylene) (Papkovsky, 1995). The triplet state of the iridium complexes is understood in terms of metal-ligand charge transfer. The structures of ligands significantly affect the triplet-state energy level; hence, emission color can vary based on the structures of ligands (Lamansky et al., 2001). The triplet state of dopants is generally produced by the exothermic triplet-triplet excitation energy transfer from the excited triplet state of the host material to the dopant. It is necessary, therefore, that the energy level of the excited triplet state of the host material should be higher than that of the phosphorescent dopant. That is why host materials with wide optical bandgap energies are needed, particularly for blue-emitting dopants. EQEs of about 19% with power efficiencies of 60–70 lm/W, which exceed the theoretical limit of 15% on the assumption that the light outcoupling efficiency is about 20%, were reported for the green-emitting phosphorescent OLEDs using cyclometalated iridium complexes (Adachi et al., 2001; Ikai et al., 2001). Further significant improvements in EQEs have been achieved. A green OLED using Ir(ppy)3 (92) doped in CBP (99) as the emissive layer and bis-4,6-(3,5di-3-pyridylphenyl)-2-methylpyrimidine as the electron-transport layer has been reported to exhibit high EQEs of 29% at 100 cd/m2 (power efficiency: 133 lm/W) and 26% at 1000 cd/m2 (power efficiency: 107 lm/W) (Tanaka et al., 2007b). A red-light-emitting OLED using Os(fptz)2(PPh2Me)2 (98) doped in a host material with bipolar functionalities, 2,7-bis(diphenylphosphoryl)-9-[4-(N,N-diphenylamino) phenyl]-9-phenylfluorene (POAPF, 14) gave an EQE of 19.9% and power efficiency of 34.5 lm/W (Chien et al., 2009). The phosphorescence-based, white-emitting OLED consisting of multiple layers, ITO/α-NPD (74)/TCTA (73) + Ir(MDQ)2(acac)/TCTA (73)/TPBI (76) + FIrpic (95)/TPBI (76)/TPBI (76) + Ir(ppy)3 (92)/TPBI (76)/Ag, gave a power efficiency of 90 lm/W and an EQE of 34% (Reineke et al., 2009).

32

Handbook of Organic Materials for Electronic and Photonic Devices

1.4.2.3 TADF-based OLEDs For fluorescence-based OLEDs, much higher EQEs should be expected if the upconversion from the electronically excited triplet state to the excited singlet state to produce the electronically excited singlet state is utilized. Since the pioneering report on high-performance, fluorescence-based OLEDs utilizing TADF (Endo et al., 2009), research on TADF-based OLEDs has been actively performed. OLEDs using PIC-TRZ (86) and 4CzIPN (87), ITO/α-NPD/m-CP/PICTRZ (86) (6wt%) + m-CP /BP4mPy/LiF/Al and ITO/α-NPD (35 nm)/4CzIPN (87) (5 wt%) + CBP (99) (15 nm)/TPBI (65 nm)/LiF (0.8 nm)/Al (70 nm), emitted bluegreen and green light with EQEs of 5.3% and 19.3%, respectively (Endo et al., 2011; Uoyama et al., 2012). DMAC-DPS (88) functons as a blue emitter. A device using this material, namely, ITO/α-NPD (30 nm)/TCTA (20 nm)/CzSi (10 nm)/ DMAC-DPS (88) + DPEPO (103) (20 nm)/DPEPO (10 nm)/TPBI (30 nm)/LiF (1 nm)/Al (100 nm) exhibited an EQE of 19.5% (Zhang et al., 2014). An OLED using PXZ-Mes3B (89) as a green emitter, ITO/TAPC (80 nm)/PXZ-Mes3B (89) (16 wt%) + CBP (99) (40 nm)/BAlq (30 nm)/Liq/Al, exhibited an EQE of 22.8% (Suzuki et al., 2015). An OLED using DACT-II (90), ITO/TAPC (100 nm)/DACT-II (90) (9%) + CBP (99) (40 nm)/BAlq (30 nm)/Liq/Al, emitted green light with internal and external quantum efficiencies of 100 and 29.6%, respectively (Kaji et al., 2015). A TADFemitting OLED using m-MTDATA and a triarylboron molecule has been reported to show exciplex emission with a high EQE of 5.4% (Goushi et al., 2012). A block copolymer with triazine-amine-triazine emitter units (112) and a carbazole-based dendrimer type molecule (114) functions as film-forming TADF materials. OLEDs using these materials, ITO/PEDOT:PSS (65 nm)/interlayer (40 nm)/112 (80 nm)/NaF (2 nm)/Al (100 nm)/Ag (100 nm) and ITO/PEDOT:PSS (30 nm)/114 (35 nm)/TPBI (40 nm)/Ca (10 nm)/Al exhibited EQEs of 10% and 3.4%, respectively (Nikolaenko et al., 2015; Albrecht et al., 2015). OLEDs using copolymers 113 and 115 doped into host materials emitted green and yellow light, with EQEs of 20.1% and 9.3%, respectively (Ren et al., 2016; Lee et al., 2016).

1.5

Summary and outlook

This chapter is concerned with a survey of the recent progress in the field of organic optoelectronics, highlighting molecular materials and polymers for use in optoelectronic devices, OPVs, and OLEDs. Molecular materials based on π-electron systems were discussed, taking into account the organization of molecules (i.e., crystals and amorphous glasses). Polymers were discussed with regard to linear π-conjugated polymers and polymers containing side-chain π-electron systems. Steady efforts have continued up to this time, with the aim of improving the PCE for OPVs and the EQE for OLEDs, and significant improvements of PCEs and EQEs have been achieved over time, laying the foundations for the subsequent progress that has been made. Problems still need to be solved, though. Further improvement of PCEs for OPVs, along with their durability, are required to achieve their practical

Organic materials for optoelectronic applications: Overview

33

use as solar cells. With regard to perovskite solar cells, stability and durability of the perovskite materials remain to be achieved. The improvement in the operational durability of blue-emitting phosphorescent dopants is required in order for white-light emitting OLEEs to be put into practical use for lighting. Basic studies on TADF materials have made remarkable progress in recent several years, and OLEDs using TADF are expected to progress further as well toward their practical use. Devices using organic photorefractive materials and organic photonic devices, which have not been discussed here, are also expected to improve. Quantum-dot OPVs, integrated devices, and organic lasers by current injection are also challenging subjects. Research and development on printed electronics toward practical use will be accelerated in view of its application to various kinds of sensors and its contribution to solving the environment and energy issues of the 21st century.

References Adachi, C., 2014. Third-generation organic electroluminescence materials. Jpn. J. Appl. Phys. 53, 060101/1–11. Adachi, C., Baldo, M.A., Thompson, M.E., Forrest, S.R., 2001. Nearly 100% internal phosphorescence efficiency in an organic light emitting device. J. Appl. Phys. 90, 5048–5051. Albrecht, K., Matsuoka, K., Fujita, K., Yamamoto, K., 2015. Carbazole dendrimers as solutionprocessable thermally activated delayed-fluorescence materials. Angew. Chem. Int. Ed. 54, 5677–5682. Alem, S., de Bettignies, R., Nunzi, J.-M., 2004. Efficient polymer-based interpenetrated network photovoltaic cells. Appl. Phys. Lett. 84, 2178–2180. Baldo, M.A., O’Brien, D.F., You, Y., et al., 1998. Highly efficient phosphorescent emission from organic electroluminescent devices. Nature 395, 151–154. Baldo, M.A., Lamansky, S., Thompson, P.E., Forrest, S.R., 1999. Very high-efficiency green organic light-emitting devices based on electrophosphorescence. Appl. Phys. Lett. 75, 4–6. Bellmann, E., Shaheen, S.E., Grubbs, R.H., et al., 1999. Organic two-layer light-emitting diodes based on high-Tg hole-transporting polymers with different redox potentials. Chem. Mater. 11, 399–407. Brabec, C.J., Shaheen, S.E., Winder, C., Sariciftci, N.S., Denk, P., 2002. Effect of LiF metal electrodes on the performance of plastic solar cells. Appl. Phys. Lett. 80, 1288–1290. Bredas, J.-L., Norton, J.E., Cornil, J., Coropceanu, V., 2009. Molecular understanding of organic solar cells: the challenges. Acc. Chem. Res. 42, 1691–1699. Burroughes, J.H., Bradley, D.D.C., Brown, A.R., et al., 1990. Light-emitting diodes based on conjugated polymers. Nature 347, 539–541. Chan, M.Y., Lee, C.S., Lai, S.L., et al., 2006a. Efficient organic photovoltaic devices using a combination of exciton blocking layer and anodic buffer layer. J. Appl. Phys. 100, 094506/1-4. Chan, M.Y., Lai, S.L., Lau, K.M., Lee, C.S., Lee, S.T., 2006b. Application of metal-doped organic layer both as exciton blocker and optical spacer for organic photovoltaic devices. Appl. Phys. Lett. 89, 163515/1-3. Chen, H.-Y., Hou, J., Zhang, S., et al., 2009. Polymer solar cells with enhanced open-circuit voltage and efficiency. Nat. Photonics 3, 649–653. Chen, G., Sasabe, H., Wang, Z., et al., 2012. Co-evaporated bulk heterojunction solar cells with >6.0% efficiency. Adv. Mater. 24, 2768–2773.

34

Handbook of Organic Materials for Electronic and Photonic Devices

Chien, C.-H., Hsu, F.-M., Shu, C.-F., Chi, Y., 2009. Efficient red electrophosphorescence from a fluorene-based bipolar host material. Org. Electron. 10, 871–876. Chiu, S.-W., Lin, L.-Y., Lin, H.-W., et al., 2012. A donor–acceptor–acceptor molecule for vacuum-processed organic solar cells with a power conversion efficiency of 6.4%. Chem. Commun. 48, 1857–1859. Chou, H.-H., Cheng, C.-H., 2010. A highly efficient universal bipolar host for blue, green, and red phosphorescent OLEDs. Adv. Mater. 22, 2468–2471. Dailey, S., Feast, W.J., Peace, R.J., et al., 2001. Synthesis and device characterisation of sidechain polymer electron transport materials for organic semiconductor applications. J. Mater. Chem. 11, 2238–2243. Doi, H., Kinoshita, M., Okumoto, K., Shirota, Y., 2003. A novel class of emitting amorphous molecular materials with bipolar character for electroluminescence. Chem. Mater. 15, 1080–1089. Duan, X., Jiang, Z., Yu, G., et al., 2005. Blue organic electroluminescent device with tetra (β-naphthyl)silane as hole blocking materials. Thin Solid Films 478, 121–124. Endo, A., Suzuki, K., Yoshihara, T., et al., 2008. Measurement of photoluminescence efficiency of Ir(III) phenylpyridine derivatives in solution and solid-state films. Chem. Phys. Lett. 460, 155–157. Endo, A., Ogasawara, M., Takahashi, A., et al., 2009. Thermally activated delayed fluorescence from Sn4+–porphyrin complexes and their application to organic light emitting diodes—a novel mechanism for electroluminescence. Adv. Mater. 21, 4802–4806. Endo, A., Sato, K., Yoshimura, K., et al., 2011. Efficient up-conversion of triplet excitons into a singlet state and its application for organic light emitting diodes. Appl. Phys. Lett. 98, 083302/1-3. Fan, B., Ying, L., Zhu, P., et al., 2017. All-polymer solar cells based on a conjugated polymer containing siloxane-functionalized side chains with efficiency over 10%. Adv. Mater. 29, 1703906/1-7. Fei, T., Cheng, G., Hu, D., et al., 2010. Iridium complex grafted to 3,6-carbazole-alttetraphenylsilane copolymers for blue electrophosphorescence. J. Polym. Sci. A Polym. Chem. 48, 1859–1865. Gao, Z., Lee, C.S., Bello, I., et al., 1999. Bright-blue electroluminescence from a silylsubstituted ter-(phenylene-vinylene) derivative. Appl. Phys. Lett. 74, 865–867. Gao, K., Li, L., Lai, T., et al., 2015. Deep absorbing porphyrin small molecule for highperformance organic solar cells with very low energy losses. J. Am. Chem. Soc. 137, 7282–7285. Gebeyehu, D., Maennig, B., Drechsel, J., Leo, K., Pfeiffer, M., 2003. Bulk-heterojunction photovoltaic devices based on donor-acceptor organic small molecule blends. Sol. Energy Mater. Sol. Cells 79, 81–92. Gibson, J., Monkman, A.P., Penfold, T.J., 2016. The importance of vibronic coupling for efficient reverse intersystem crossing in thermally activated delayed fluorescence molecules. ChemPhysChem 17, 2956–2961. Goushi, K., Yoshida, K., Sato, K., Adachi, C., 2012. Organic light-emitting diodes employing efficient reverse intersystem crossing for triplet-to-singlet state conversion. Nat. Photonics 6, 253–258. Guo, Y., Li, Y., Awartani, O., et al., 2017. Improved performance of all-polymer solar cells enabled by naphthodiperylenetetraimide-based polymer acceptor. Adv. Mater. 29, 1700309/1-6. Hains, A.W., Marks, T.J., 2008. High-efficiency hole extraction/electron-blocking layer to replace poly(3,4-ethylenedioxythiophene):poly(styrene sulfonate) in bulk-heterojunction polymer solar cells. Appl. Phys. Lett. 92, 023504/1-3.

Organic materials for optoelectronic applications: Overview

35

Haldi, A., Kimyonok, A., Domercq, B., et al., 2008. Optimization of orange-emitting electrophosphorescent copolymers for organic light-emitting diodes. Adv. Funct. Mater. 18, 3056–3062. Halls, J.J.M., Walsh, C.A., Greenham, N.C., et al., 1995. Efficient photodiodes from interpenetrating polymer networks. Nature 376, 498–500. Han, C., Zhao, Y., Xu, H., et al., 2011. A simple phosphine-oxide host with a multi-insulating structure: high triplet energy level for efficient blue electrophosphorescence. Chem. Eur. J. 17, 5800–5803. Hayakawa, A., Yoshikawa, O., Fujieda, T., Uehara, K., Yoshikawa, S., 2007. High performance polythiophene/fullerene bulk-heterojunction solar cell with a TiOx hole blocking layer. Appl. Phys. Lett. 90, 163517/1-3. He, Y., Zhao, G., Peng, B., Li, Y., 2010. High-yield synthesis and electrochemical and photovoltaic properties of indene-C70 bisadduct. Adv. Funct. Mater. 20, 3383–3389. He, Z., Zhong, C., Huang, X., et al., 2011. Simultaneous enhancement of open-circuit voltage, short-circuit current density, and fill factor in polymer solar cells. Adv. Mater. 23, 4636–4643. Higuchi, A., Inada, H., Kobata, T., Shirota, Y., 1991. Amorphous molecular materials: synthesis and properties of a novel starburst molecule, 4,40 ,400 -tri(N-phenothiazinyl)triphenylamine. Adv. Mater. 3, 549–550. Higuchi, A., Ohnishi, K., Nomura, S., Inada, H., Shirota, Y., 1992. Tri(biphenyl-4-yl)amine and tri(p-terphenyl-4-yl)amine as a novel class of molecules for amorphous molecular materials. J. Mater. Chem. 2, 1109–1110. Hirai, Y., Nakanishi, T., Kitagawa, Y., et al., 2015. Luminescent coordination glass: remarkable morphological strategy for assembled Eu(III) complexes. Inorg. Chem. 54, 4364–4370. Hosokai, T., Matsuzaki, H., Nakanotani, H., et al., 2017. Evidence and mechanism of efficient thermally activated delayed fluorescence promoted by delocalized excited states. Sci. Adv. 3, e1603282/1–9. Huang, J., Wang, H., Yan, K., et al., 2017. Highly efficient organic solar cells consisting of double bulk heterojunction layers. Adv. Mater. 29, 1606729/1-9. Hughes, G., Bryce, M.R., 2005. Electron-transporting materials for organic electroluminescent and electrophosphorescent devices. J. Mater. Chem. 15, 94–107. Hwang, Y.-J., Ren, G., Murari, N.M., Jenekhe, S.A., 2012. n-Type naphthalene diimide– biselenophene copolymer for all-polymer bulk heterojunction solar cells. Macromolecules 45, 9056–9062. Hwang, Y.-J., Courtright, B.A.E., Ferreira, A.S., Tolbert, S.H., Jenekhe, S.A., 2015. 7.7% Efficient all-polymer solar cells. Adv. Mater. 27, 4578–4584. Ikai, M., Tokito, S., Sakamoto, Y., Suzuki, T., Taga, Y., 2001. Highly efficient phosphorescence from organic light-emitting devices with an exciton-block layer. Appl. Phys. Lett. 79, 156–158. Imae, I., Kawakami, Y., 2005. Novel family of molecular glasses based on silicon-containing compounds. Chem. Lett. 34, 290–291. Imae, I., Nawa, K., Ohsedo, Y., Noma, N., Shirota, Y., 1997. Synthesis of a novel family of electrochemically-doped vinyl polymers containing pendant oligothiophenes and their electrical and electrochromic properties. Macromolecules 30, 380–386. Jiang, X., Xu, H., Yang, L., et al., 2009. Effect of CsF interlayer on the performance of polymer bulk heterojunction solar cells. Sol. Energy Mater. Sol. Cells 93, 650–653. Kageyama, H., Ohishi, H., Tanaka, M., Ohmori, Y., Shirota, Y., 2009a. High performance organic photovoltaic devices using amorphous molecular materials with high chargecarrier drift mobilities. Appl. Phys. Lett. 94, 063304/1-3.

36

Handbook of Organic Materials for Electronic and Photonic Devices

Kageyama, H., Ohishi, H., Tanaka, M., Ohmori, Y., Shirota, Y., 2009b. High-performance organic photovoltaic devices using a new amorphous molecular material with high hole drift mobility, tris[4-(5-phenylthiophen-2-yl)phenyl]amine. Adv. Funct. Mater. 19, 3948–3955. Kageyama, H., Kajii, H., Ohmori, Y., Shirota, Y., 2011. MoO3 as a cathode buffer layer material for the improvement of planar pn-heterojunction organic solar cell performance. Appl. Phys. Express. 4, 032301/1-3. Kaji, H., Suzuki, H., Fukushima, T., et al., 2015. Purely organic electroluminescent material realizing 100% conversion from electricity to light. Nat. Commun. 6, 8476/1–8. Kamtekar, K.T., Wang, C., Bettington, S., et al., 2006. New electroluminescent bipolar compounds for balanced charge-transport and tuneable colour in organic light emitting diodes: triphenylamine-oxadiazole-fluorene triad molecules. J. Mater. Chem. 16, 3823–3835. Kawamoto, M., Mochizuki, H., Shishido, A., et al., 2003. Side-chain polymer liquid crystals containing oxadiazole and amine moieties with carreir-transporting abilities for singlelayer light-emitting diodes. J. Phys. Chem. B 107, 4887–4893. Kim, J.Y., Kim, S.H., Lee, H.-H., et al., 2006. New architecture for high-efficiency polymer photovoltaic cells using solution-based titanium oxide as an optical spacer. Adv. Mater. 18, 572–576. Kim, K., Liu, J., Namboothiry, M.A.G., Carroll, D.L., 2007. Roles of donor and acceptor nanodomains in 6% efficient thermally annealed polymer photovoltaics. Appl. Phys. Lett. 90, 163511/1-3. Kinoshita, M., Shirota, Y., 2001. 1,5-Bis[5-(dimesitylboryl)thiophen-2-yl]benzene and 1,3,5tris[5-(dimesitylboryl)thiophen-2-yl]benzene as a novel family of electron-transporting hole blockers for organic electroluminescent devices. Chem. Lett. 30, 614–615. Kinoshita, M., Kita, H., Shirota, Y., 2002. A novel family of boron-containing hole-blocking amorphous molecular materials for blue-and blue-violet-emitting organic electroluminescent devices. Adv. Funct. Mater. 12, 780–786. Kojima, A., Teshima, K., Shirai, Y., Miyasaka, T., 2009. Organometal halide perovskites as visible-light sensitizers for photovoltaic cells. J. Am. Chem. Soc. 131, 6050–6051. Kulkarni, A.P., Tonzola, C.J., Babel, A., Jenekhe, S.A., 2004. Electron transport materials for organic light-emitting diodes. Chem. Mater. 16, 4556–4573. Kuwabara, Y., Ogawa, H., Inada, H., Noma, N., Shirota, Y., 1994. Thermally stable multilayered organic electroluminescent devices using novel starburst molecules, 4,40 ,400 -tri (N-carbazolyl)triphenylamine (TCTA) and 4,40 ,400 -tris(3-methylphenylphenylamino) triphenylamine (m-MTDATA), as hole-transport materials. Adv. Mater. 6, 677–679. Kwon, J., Lee, W., Kim, J.-Y., et al., 2010. Solution processable donor materials based on thiophene and triphenylamine for bulk heterojunction solar cells. New J. Chem. 34, 744–749. Lamansky, S., Djurovich, P., Murphy, D., et al., 2001. Highly phosphorescent biscyclometalated iridium complexes: synthesis, photophysical characterization, and use in organic light emitting diodes. J. Am. Chem. Soc. 123, 4304–4312. Lee, S.Y., Yasuda, T., Yang, Y.S., Zhang, Q., Adachi, C., 2014. Luminous butterflies: efficient exciton harvesting by benzophenone derivatives for full-color delayed fluorescence oleds. Angew. Chem. 126, 6520–6524. Lee, S.Y., Yasuda, T., Komiyama, H., Lee, J., Adachi, C., 2016. Thermally activated delayed fluorescence polymers for efficient solution-processed organic light-emitting diodes. Adv. Mater. 28, 4019–4024. Li, G., Chu, C.-W., Shrotriya, V., Huang, J., Yang, Y., 2006. Efficient inverted polymer solar cells. Appl. Phys. Lett. 88, 253503.

Organic materials for optoelectronic applications: Overview

37

Li, H., Sundararaman, A., Venkatasubbaiah, K., J€akle, F., 2007. Organoborane acceptorsubstituted polythiophene via side-group borylation. J. Am. Chem. Soc. 129, 5792–5793. Li, W., Du, C., Li, F., et al., 2009. Benzothiadiazole-based linear and star molecules: design, synthesis, and their application in bulk heterojunction organic solar cells. Chem. Mater. 21, 5327–5334. Li, Z., He, G., Wan, X., et al., 2012. Solution processable rhodanine-based small molecule organic photovoltaic cells with a power conversion efficiency of 6.1%. Adv. Energy Mater. 2, 74–77. Li, M., Gao, K., Wan, X., et al., 2017. Solution-processed organic tandem solar cells with power conversion efficiencies >12%. Nat. Photonics 11, 85–90. Liang, Y., Xu, Z., Xia, J., et al., 2010. For the bright future—bulk heterojunction polymer solar cells with power conversion efficiency of 7.4%. Adv. Mater. 22, E135–E138. Lin, H.-W., Lin, C.-L., Chang, H.-H., et al., 2004. Anisotropic optical properties and molecular orientation in vacuum-deposited ter(9,9-diarylfluorene)s thin films using spectroscopic ellipsometry. J. Appl. Phys. 95, 881–886. Lin, Y., Wang, J., Zhang, Z.-G., et al., 2015. An electron acceptor challenging fullerenes for efficient polymer solar cells. Adv. Mater. 27, 1170–1174. Luo, J., Xie, G., Gong, S., Chen, T., Yang, C., 2016. Creating a thermally activated delayed fluorescence channel in a single polymer system to enhance exciton utilization efficiency for bluish-green electroluminescence. Chem. Commun. 52, 2292–2295. Ma, W., Yang, C., Gong, X., Lee, K., Heeger, A.J., 2005. Thermally stable, efficient polymer solar cells with nanoscale control of the interpenetrating network morphology. Adv. Funct. Mater. 15, 1617–1622. M€akinen, A.J., Hill, I.G., Noda, T., Shirota, Y., Kafafi, Z.H., 2001. Photoemission study of frontier orbital alignment as a metal-organic interface as a function of conjugation length of oligothiophene derivatives. Appl. Phys. Lett. 78, 670–672. McCullough, R.D., Lowe, R.D., 1992. Enhanced electrical conductivity in regioselectively synthesized poly(3-alkylthiophenes). J. Chem. Soc. Chem. Commun. 28, 70–72. Mishra, A., B€auerle, P., 2012. Small molecule organic semiconductors on the move: promises for future solar energy technology. Angew. Chem. Int. Ed. 51, 2020–2067. Mochizuki, H., Hasui, T., Kawamoto, M., et al., 2000a. Novel liquid-crystalline and amorphous materials containing oxadiazole and amine moieties for electroluminescent devices. Chem. Commun. 1923–1924. Mochizuki, H., Hasui, T., Shiono, T., et al., 2000b. Emission behavior of molecularly doped electroluminescent device using liquid-crystalline matrix. Appl. Phys. Lett. 77, 1587–1589. Mochizuki, H., Hasui, T., Kawamoto, M., et al., 2003. A novel class of photo- and electroactive polymers containing oxadiazole and amine moieties in a side chain. Macromolecules 36, 3457–3464. Mukherjee, S., Thilagar, P., 2015. Recent advances in purely organic phosphorescent materials. Chem. Commun. 51, 10988–11003. Murata, H., Merritt, C.D., Inada, H., Shirota, Y., Kafafi, Z.H., 1999. Molecular organic lightemitting diodes with temperature-independent quantum efficiency and improved thermal durability. Appl. Phys. Lett. 75, 3252–3254. Murata, H., Malliaras, G.G., Uchida, M., Shen, Y., Kafafi, Z.H., 2001. Non-dispersive and airstable electron transport in an amorphous organic semiconductor. Chem. Phys. Lett. 339, 161–166. Murata, H., Kafafi, Z.H., Uchida, M., 2002. Efficient organic light-emitting diodes with undoped active layers based on silole derivatives. Appl. Phys. Lett. 80, 189–191.

38

Handbook of Organic Materials for Electronic and Photonic Devices

Mutaguchi, D., Okumoto, K., Ohsedo, Y., Moriwaki, K., Shirota, Y., 2003. Development of a new class of hole-transporting and emitting vinyl polymers and their application in organic electroluminescent devices. Org. Electron. 4, 49–59. Nagamatsu, D., Maeda, M., Hashimoto, K., et al., 2005. Development of a novel emitting vinyl polymer and its application in organic electroluminescent devices. J. Soc. Inf. Disp. 13, 399–404. Nawa, K., Miyawaki, K., Imae, I., Noma, N., Shirota, Y., 1993a. Polymers containing pendant oligothiophenes as a novel class of electrochromic materials. J. Mater. Chem. 3, 113–114. Nawa, K., Miyawaki, K., Imae, I., Noma, N., Shirota, Y., 1993b. A novel type of semiconducting polymers: synthesis and properties of electrochemically-doped polymers containing pendant oligothiophenes. Synth. Met. 55–57, 1176–1181. Nian, L., Gao, K., Jiang, Y., et al., 2017. Small-molecule solar cells with simultaneously enhanced short-circuit current and fill factor to achieve 11% efficiency. Adv. Mater. 29, 1700616/1-7. Nikolaenko, A.E., Cass, M., Bourcet, F., Mohamad, D., Roberts, M., 2015. Thermally activated delayed fluorescence in polymers: a new route toward highly efficient solution processable OLEDs. Adv. Mater. 27, 7236–7240. Noda, T., Shirota, Y., 1998. 5,50 -Bis(dimesitylboryl)-2,20 -bithiophene and 5,500 -bis (dimesitylboryl)-2,20 :50 ,200 -terthiophene as a novel family of electron-transporting amorphous molecular materials. J. Am. Chem. Soc. 120, 9714–9715. Noda, T., Imae, I., Noma, N., Shirota, Y., 1997a. 5,500 -Bis{4-[bis(4-methylphenyl)amino]phenyl}2,20 :50 ,200 -terthiophene and 5,50 -bis{4-[bis(4-methylphenyl)amino]phenyl}-2,20 :50 ,200 :500 ,20 quaterthiophene as a novel family of amorphous molecular materials. Adv. Mater. 9, 239–241. Noda, T., Ogawa, H., Noma, N., Shirota, Y., 1997b. A novel family of amorphous molecular materials containing an oligothiophene moiety as color-tunable emitting materials for organic electroluminescent devices. Adv. Mater. 9, 720–722. Noda, T., Ogawa, H., Noma, N., Shirota, Y., 1999. Organic light-emitting diodes using a novel family of amorphous molecular materials containing an oligothiophene moiety as colourtunable emitting materials. J. Mater. Chem. 9, 2177–2181. NREL Efficiency Chart, 2017. https://www.nrel.gov/pv/assets/images/efficiency-chart.png. Ohishi, H., Tanaka, M., Kageyama, H., Shirota, Y., 2004. Amorphous molecular materials with high carrier mobilities: thiophene- and selenophene-containing tris(oligoarylenyl)amines. Chem. Lett. 33, 1266–1267. Ohno, T., Moriwaki, K., Miyata, T., 2001. Intramolecular charge-transfer interaction in a new dyad based on C60 and bis(40 -tert-butylbiphenyl-4-yl)aniline (BBA) donor. Asian J. Org. Chem. 66, 3397–3401. Palilis, L.C., Murata, H., Uchida, M., Kafafi, Z.H., 2003a. High efficiency molecular organic light-emitting diodes based on silole derivatives and their exciplexes. Org. Electron. 4, 113–121. Palilis, L.C., M€akinen, A.J., Uchida, M., Kafafi, Z.H., 2003b. Highly efficient molecular organic light-emitting diodes based on exciplex emission. Appl. Phys. Lett. 82, 2209–2211. Pan, H., Fu, G.-L., Zhao, Y.-H., Zhao, C.-H., 2011. Through-space charge-transfer emitting biphenyls containing a boryl and an amino group at the o,o0 -positions. Org. Lett. 13, 4830–4833. Papkovsky, D.B., 1995. New oxygen sensors and their application to biosensing. Sensors Actuators B 29, 213–218.

Organic materials for optoelectronic applications: Overview

39

Park, S.H., Roy, A., Beaupre, S., et al., 2009. Bulk heterojunction solar cells with internal quantum efficiency approaching 100%. Nat. Photonics 3, 297–303. Peumans, P., Forrest, S.R., 2001. Very-high-efficiency double-heterostructure copper phthalocyanine/C60 photovoltaic cells. Appl. Phys. Lett. 79, 126–128. Peumans, P., Bulovic, V., Forrest, S.R., 2000. Efficient photon harvesting at high optical intensities in ultrathin organic double-heterostructure photovoltaic diodes. Appl. Phys. Lett. 76, 2650–2652. Peumans, P., Uchida, S., Forrest, S.R., 2003. Efficient bulk heterojunction photovoltaic cells using small-molecular-weight organic thin films. Nature 425, 158–162. Pommerehne, J., Selz, A., Book, K., et al., 1997. Electron tunneling in organic bilayer lightemitting diodes with a novel electron-transporting polymer. Macromolecules 30, 8270–8277. Reineke, S., Lindner, F., Schwartz, G., et al., 2009. White organic light-emitting diodes with fluorescent tube efficiency. Nature 459, 234–239. Ren, Z., Nobuyasu, R.S., Dias, F.B., et al., 2016. Pendant homopolymer and copolymers as solution-processable thermally activated delayed fluorescence materials for organic light-emitting diodes. Macromolecules 49, 5452–5460. Ross, R.B., Cardona, C.M., Guldi, D.M., et al., 2009. Endohedral fullerenes for organic photovoltaic devices. Nat. Mater. 8, 208–212. Sariciftci, N.S., Smilowitz, L., Heeger, A.J., Wudl, F., 1992. Photoinduced electron transfer from a conducting polymer to buckminsterfullerene. Science 258, 1474–1476. Sariciftci, N.S., Braun, D., Zhang, C., et al., 1993. Semiconducting polymerbuckminsterfullerene heterojunctions: diodes, photodiodes, and photovoltaic cells. Appl. Phys. Lett. 62, 585–587. Sasabe, H., Kido, J., 2011. Multifunctional materials in high-performance OLEDs: challenges for solid-state lighting. Chem. Mater. 23, 621–630. Schubert, M., Dolfen, D., Frisch, J., et al., 2012. Influence of aggregation on the performance of all-polymer solar cells containing low-bandgap naphthalenediimide copolymers. Adv. Energy Mater. 2, 369–380. Shaheen, S.E., Jabbour, G.E., Kippelen, B., et al., 1999. Organic light-emitting diode with 20 lm/W efficiency using a triphenyldiamine side-group polymer as the hole transport layer. Appl. Phys. Lett. 74, 3212–3214. Shang, H., Fan, H., Liu, Y., et al., 2011. A solution-processable star-shaped molecule for highperformance organic solar cells. Adv. Mater. 23, 1554–1557. Shimizu, M., Shigitani, R., Nakatani, M., et al., 2016. Siloxy group-induced highly efficient room temperature phosphorescence with long lifetime. J. Phys. Chem. C 120, 11631–11639. Shirota, Y., 2000. Organic materials for electronic and optoelectronic devices. J. Mater. Chem. 10, 1–25. Shirota, Y., 2005. Photo- and electroactive amorphous molecular materials - molecular design, synthesis, reactions, properties, and applications. J. Mater. Chem. 15, 75–93. Shirota, Y., Kobata, T., Noma, N., 1989. Starburst molecules for amorphous molecular materials. 4,40 ,400 -tris(N,N-diphenylamino)triphenylamine and 4,40 ,400 -tris[N-(3-methylphenyl)-N-phenylamino]triphenylamine. Chem. Lett. 18, 1145–1148. Shirota, Y., Kuwabara, Y., Inada, H., et al., 1994. Multilayered organic electroluminescent device using a novel starburst molecule, 4,40 ,400 -tris(3-methylphenylphenylamino) triphenylamine, as a hole transport material. Appl. Phys. Lett. 65, 807–809. Shirota, Y., Kuwabara, Y., Okuda, D., et al., 1997. Starburst molecules based on π-electron systems as materials for organic electroluminescent devices. J. Lumin. 72–74, 985–991.

40

Handbook of Organic Materials for Electronic and Photonic Devices

Shirota, Y., Moriwaki, K., Yoshikawa, S., Ujike, T., Nakano, H., 1998. 4-[Di(biphenyl-4-yl) amino]azobenzene and 4,40 -bis[bis(40 -tert-butylbiphenyl-4-yl)amino]azobenzene as a novel family of photochromic amorphous molecular materials. J. Mater. Chem. 8, 2579–2581. Shirota, Y., Kinoshita, M., Noda, T., Okumoto, K., Ohara, T., 2000. A novel class of emitting amorphous molecular materials as bipolar radical formants: 2-{4-[bis(4-methylphenyl)amino]phenyl}-5-(dimesitylboryl)thiophene and 2-{4-[bis(9,9-dimethylfluorenyl)amino] phenyl}-5-(dimesitylboryl)thiophene. J. Am. Chem. Soc. 122, 11021–11022. Shrotriya, V., Li, G., Yao, Y., Chu, C.-W., Yang, Y., 2006. Transition metal oxides as the buffer layer for polymer photovoltaic cells. Appl. Phys. Lett. 88, 073508/1-3. Shu, Y., Lim, Y.-F., Li, Z., et al., 2011. A survey of electron-deficient pentacenes as acceptors in polymer bulk heterojunction solar cells. Chem. Sci. 2, 363–368. Sonar, P., Ng, G.-M., Lin, T.T., Dodabalapur, A., Chen, Z.-K., 2010. Solution processable low bandgap diketopyrrolopyrrole (DPP) based derivatives: novel acceptors for organic solar cells. J. Mater. Chem. 20, 3626–3636. Sun, C., Hudson, Z.M., Helander, M.G., Lu, Z.-H., Wang, S., 2011. A polyboryl-functionalized triazine as an electron transport material for OLEDs. Organometallics 30, 5552–5555. Sun, Y., Welch, G.C., Leong, W.L., et al., 2012. Solution-processed small-molecule solar cells with 6.7% efficiency. Nat. Mater. 11, 44–48. Sun, D., Yang, Z., Sun, X., et al., 2014a. Synthesis of triphenylamine based polysiloxane as a blue phosphorescent host. Polym. Chem. 5, 5046–5052. Sun, D., Fu, Q., Ren, Z., et al., 2014b. Synthesis of well-defined poly(phenylcarbazole-alttriphenylphosphine oxide) siloxane as a bipolar host material for solution-processed deep blue phosphorescent devices. Polym. Chem. 5, 220–226. Suzuki, M., Tokito, S., Sato, F., et al., 2005. Highly efficient polymer light-emitting devices using ambipolar phosphorescent polymers. Appl. Phys. Lett. 86, 103507/1-3. Suzuki, K., Kubo, S., Shizu, K., et al., 2015. Triarylboron-based fluorescent organic lightemitting diodes with external quantum efficiencies exceeding 20%. Angew. Chem. Int. Ed. 54, 15231–15235. Tamao, K., Uchida, M., Izumizawa, T., Furukawa, K., Yamaguchi, S., 1996. Silole derivatives as efficient electron transporting materials. J. Am. Chem. Soc. 118, 11974–11975. Tanaka, D., Takeda, T., Chiba, T., Watanabe, S., Kido, J., 2007a. Novel electron-transport material containing boron atom with a high triplet excited energy level. Chem. Lett. 36, 262–263. Tanaka, D., Sasabe, H., Li, Y.-J., et al., 2007b. Ultra high efficiency green organic light-emitting devices. Jpn. J. Appl. Phys. 46, L10–L12. Tang, C.W., 1986. Two-layer organic photovoltaic cell. Appl. Phys. Lett. 48, 183–185. Tang, C.W., Van Slyke, S.A., 1987. Organic electroluminescent diodes. Appl. Phys. Lett. 51, 913–915. Tao, Y., Yuan, K., Chen, T., et al., 2014. Thermally activated delayed fluorescence materials towards the breakthrough of organoelectronics. Adv. Mater. 26, 7931–7958. Umeda, T., Shirakawa, T., Fujii, A., Yoshino, K., 2003. Improvement of characteristics of organic photovoltaic devices composed of conducting polymer-fullerene systems by introduction of ZnO layer. Jpn. J. Appl. Phys. 42, L1475–L1477. Uoyama, H., Goushi, K., Shizu, K., Nomura, H., Adachi, C., 2012. Highly efficient organic light-emitting diodes from delayed fluorescence. Nature 492, 234–238. van der Poll, T.S., Love, J.A., Nguyen, T.-Q., Bazan, G.C., 2012. Non-basic high-performance molecules for solution-processed organic solar cells. Adv. Mater. 24, 3646–3649.

Organic materials for optoelectronic applications: Overview

41

Van Slyke, S.A., Chen, C.H., Tang, C.W., 1996. Organic electroluminescent devices with improved stability. Appl. Phys. Lett. 69, 2160–2162. Wada, Y., Shizu, K., Kubo, S., et al., 2015. Highly efficient electroluminescence from a solution-processable thermally activated delayed fluorescence emitter. Appl. Phys. Lett. 107, 183303/1–4. White, M.S., Olson, D.C., Shaheen, S.E., Kopidakis, N., Ginley, D.S., 2006. Inverted bulkheterojunction organic photovoltaic device using a solution-derived ZnO underlayer. Appl. Phys. Lett. 89, 143517/1-3. Wienk, M.M., Kroon, J.M., Verhees, W.J.H., et al., 2003. Efficient methano[70]fullerene/ MDMO-PPV bulk heterojunction photovoltaic cells. Angew. Chem. Int. Ed. 42, 3371–3375. Wu, C.-C., Lin, Y.-T., Wong, K.-T., Chen, R.-T., Chien, Y.-Y., 2004. Efficient organic bluelight-emitting devices with double confinement on terfluorenes with ambipolar carrier transport properties. Adv. Mater. 16, 61. Xu, X., Yu, T., Bi, Z., et al., 2018. Realizing over 13% efficiency in green-solvent-processed nonfullerene organic solar cells enabled by 1,3,4-thiadiazole-based wide-bandgap copolymers. Adv. Mater. 30, 1703973/1-8. Xue, J., Rand, B.P., Uchida, S., Forrest, S.R., 2005. A hybrid planar-mixed molecular heterojunction photovoltaic cell. Adv. Mater. 17, 66–71. Yakimov, A., Forrest, S.R., 2002. High photovoltage multiple-heterojunction organic solar cells incorporating interfacial metallic nanoclusters. Appl. Phys. Lett. 80, 1667–1669. Yamaguchi, S., Akiyama, S., Tamao, K., 2000a. Tri-9-anthrylborane and its derivatives: new boron-containing π-electron systems with divergently extended π-conjugation through boron. J. Am. Chem. Soc. 122, 6335–6336. Yamaguchi, S., Shirasaka, T., Tamao, K., 2000b. Tridurylboranes extened by three arylethynyl groups as a new family of boron-based π-electron systems. Org. Lett. 2, 4129–4132. Yang, F., Shtein, M., Forrest, S.R., 2005a. Morphology control and material mixing by hightemperature organic vapor-phase deposition and its application to thin-film solar cells. J. Appl. Phys. 98, 014906/1-10. Yang, F., Shtein, M., Forrest, S.R., 2005b. Controlled growth of a molecular bulk heterojunction photovoltaic cell. Nat. Mater. 4, 37–41. Yang, W.S., Noh, J.H., Jeon, N.J., et al., 2015. High-performance photovoltaic perovskite layers fabricated through intramolecular exchange. Science 348, 1234–1237. Yao, H., Chen, Y., Qin, Y., et al., 2016. Design and synthesis of a low bandgap small molecule acceptor for efficient polymer solar cells. Adv. Mater. 28, 8283–8287. Yokoyama, D., 2011. Molecular orientation in small-molecule organic light-emitting diodes. J. Mater. Chem. 21, 19187–19202. Yokoyama, D., Sakaguchi, A., Suzuki, M., Adachi, C., 2008. Horizontal molecular orientation in vacuum-deposited organic amorphous films of hole and electron transport materials. Appl. Phys. Lett. 93, 173302/1-3. Yoshiiwa, M., Kageyama, H., Shirota, Y., et al., 1996. Novel class of low molecular-weight organic resists for nanometer lithography. Appl. Phys. Lett. 69, 2605–2607. Yu, G., Gao, J., Hummelen, J.C., Wudl, F., Heeger, A.J., 1995. Polymer photovoltaic cells: enhanced efficiencies via a network of internal donor-acceptor heterojunctions. Science 270, 1789–1791. Zhang, Y., Yip, H.-L., Acton, O., et al., 2009. A simple and effective way of achieving highly efficient and thermally stable bulk-heterojunction polymer solar cells using amorphous fullerene derivatives as electron acceptor. Chem. Mater. 21, 2598–2600.

42

Handbook of Organic Materials for Electronic and Photonic Devices

Zhang, Q., Li, J., Shizu, K., et al., 2012. Design of efficient thermally activated delayed fluorescence materials for pure blue organic light emitting diodes. J. Am. Chem. Soc. 134, 14706–14709. Zhang, Q., Li, B., Huang, S., et al., 2014. Efficient blue organic light-emitting diodes employing thermally activated delayed fluorescence. Nat. Photonics 8, 326–332. Zhao, G., He, Y., Li, Y., 2010. 6.5% Efficiency of polymer solar cells based on poly(3hexylthiophene) and indene-C60 bisadduct by device optimization. Adv. Mater. 22, 4355–4358. Zhao, W., Li, S., Yao, H., et al., 2017. Molecular optimization enables over 13% efficiency in organic solar cells. J. Am. Chem. Soc. 139, 7148–7151. Zhou, J., Wan, X., Liu, Y., et al., 2011. A planar small molecule with dithienosilole core for high efficiency solution-processed organic photovoltaic cells. Chem. Mater. 23, 4666–4668. Zhou, Y., Fuentes-Hernandez, C., Shim, J., et al., 2012. A universal method to produce low– work function electrodes for organic electronics. Science 336, 327–332.

Key trends in sustainable approaches to the synthesis of semiconducting polymers

2

Assunta Marrocchi, Valeria Trombettoni, Daniele Sciosci, Filippo Campana, Luigi Vaccaro Laboratory of Green Synthetic Organic Chemistry, Department of Chemistry, Biology and Biotechnology, University of Perugia, Perugia, Italy

2.1

Introduction

The research area of organic electronics has grown exponentially since the understanding that π-conjugated polymers can be successfully implemented in several electronic devices, including organic light-emitting diodes (OLEDs), organic field effect transistors (OFETs), sensors, integrated circuits, solar energy storage, organic photovoltaic cells (OPVs), and RF-ID tags (Lee et al., 2017; Huang et al., 2017; Dou et al., 2015; Guo et al., 2013; Marrocchi et al., 2012). To date, most of the efforts focused on the development of new semiconductor structures, the main challenge being to enable organic electronic devices with optimal performance. On the other hand, only a limited emphasis has been placed on the synthetic methodology employed in and environmental impact associated with their synthesis. These problems include toxicity of the synthetic reagents, catalysts, and by-products to humans, animals, and plants, electrical energy for their production, and transfer from laboratory to large scale production. Typical methodologies to synthesize π-conjugated active polymers in organic electronics are based on traditional metal catalyzed step-growth polycondensation (Carsten et al., 2011) and chain growth polymerization reactions (Schmidt et al., 2014; Liu et al., 2014; Senkovskyy et al., 2012; Yokozawa et al., 2012). Despite their great versatility and wide substrate scope, these reactions have a number of drawbacks, including the number of steps to prepare the monomeric precursors, the instability of several organometallic reagents employed, poor conversion related to unreactive monomers, difficulties in controlling the polymer architecture, and poor atom economy owing to the formation of stoichiometric amounts of toxic by-products. In addition, device-quality organic semiconductors generally require extensive purification processes to remove undesired traces of by-products, metals, or both (Strappaveccia et al., 2015a; Osedach et al., 2013). In the pursuit of a sustainable electronics industry, efforts to implement economically attractive and benign novel approaches to organic semiconductors that minimize the use of solvents and reagents, as well as the number of workup procedures Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00002-4 © 2019 Elsevier Ltd. All rights reserved.

44

Handbook of Organic Materials for Electronic and Photonic Devices

(Marrocchi et al., 2016a), undoubtedly represent important contributions for the development of the field. In this regard, the development of synthetic approaches to semiconductor materials encompassing the principles of green chemistry is mandatory. About 15 years ago, the US Environmental Protection Agency (EPA) defined green chemistry as “innovative chemical technologies that reduce or eliminate the use or generation of hazardous substances in the design, manufacture and use of chemical products” (EPA, n.d.). Using the so-called 12 principles developed by Anastas and Warner (1998) as guidelines, chemists can evaluate and improve current procedures and develop new ones that will have a limited impact on the environment and therefore be more sustainable and economical in the long run. The 12 principles can be summarized as follows: 1. Prevent waste. 2. Design synthetic methods to maximize the incorporation of all materials used in the process into the final product (i.e., atom economy). 3. Design less hazardous chemical synthesis. 4. Design safer chemicals and other products. 5. Use safer solvents/auxiliaries. 6. Increase energy efficiency. 7. Use renewable raw materials/feedstock. 8. Avoid derivatization (e.g., protecting groups, temporary modification of physical/chemical processes), because it requires additional reagents and can generate waste. 9. Use selective catalysts, not stoichiometric reagents. 10. Design chemical products considering their degradation after use. 11. Analyze in real time to prevent pollution. 12. Minimize the potential for accidents, including releases, explosions, and fires.

Particularly, waste production should be reduced using minimal amounts of reagents and by using easily recoverable/reusable heterogeneous catalytic systems to simplify workup procedures and reduce energy and production times (Strappaveccia et al., 2015a,b; Tkachov et al., 2014; Burke and Lipomi, 2013; Bartollini et al., 2013). Particular attention should be directed toward solvents, which are commonly a major input in industrial and lab-based procedures as reaction media, in addition to their use for extraction, separation, and purification. These substances typically comprise the largest fraction of the waste generated (about 90%), as well as presenting significant workplace hazards. For instance, traditional solvents such as dipolar aprotic solvents [e.g., 1-methylpyrrolidone (NMP), dimethylformamide (DMF), and dimethylacetamide (DMA)], pyridine, and chlorinated solvents, which are commonly used as media in condensation polymerizations, increase air pollution; most of them are toxic, flammable, or both. These solvents are not compliant with relevant legislation, ranking high on the list of harmful chemicals (Anderson, 2012; Henderson et al., 2011; ECHA, n.d.). To avoid the issue of traditional solvents, it is important to employ substitutes to conventional solvents while retaining their desirable properties. Highboiling point media are used to reduce workplace hazards for solution polymerization because these reactions can be carried out at elevated temperatures without excessive

Key trends in sustainable approaches to the synthesis of semiconducting polymers

45

pressure buildup. However, a major drawback is related to the high energy cost for the solvent removal from the final products. This chapter will give readers an overview of the efforts in the synthetic approaches to semiconducting polymers that satisfy several criteria relevant to green chemistry. We restrict this discussion to the preparation of semiconducting poly(arylenes) (Fig. 2.1), because of the significant efforts devoted to their sustainable synthesis and their prime importance for the realization of unconventional (i.e., organic) electronic devices, particularly OFETs and OPVs. Briefly, from a structural point of view, the main feature of the poly(arylenes) backbone is the direct connection of the aromatic repeating units by single carbondcarbon bonds. Effective conjugation between the phenylene moieties, therefore, is affected by substituents on the phenylene rings, chain defects, and morphology in the solid state. All these factors confer to each polymer distinctive chemical-physical and electronic properties. Poly(arylenes) are also of great interest due to their persistent rigid rodlike structures, which are useful for the construction of nanostructures (Li et al., 2010). Thus, specifically, this chapter is organized into three main sections. First, we will introduce the basic concepts of OFET/OPV operation. Next, we will focus on the green methodologies for metal-mediated polymerization reactions, including Suzuki cross-coupling, direct (hetero)arylation polycondensation, and click polymerization. Finally, we will refer to the state of the art in metal-free oxidative coupling protocols, with a focus on nitroxide-mediated radical polymerization. Providing an exhaustive list of examples of such chemistries is beyond the scope of this chapter. Instead, we intend to identify key developments in the field that seem particularly suitable for large-scale sustainable synthesis. Limitations and problems associated with some polymerization protocols in terms of material physical properties and performance in OTFT/OPV devices will be illustrated and discussed as well.

2.2

Solar cell (OPV) and organic field-effect transistor (OFET) structure and operation

This section gives a brief survey of the physics and operating characteristics of OFETs and photovoltaic cells. The aim is to provide an essential background for the subsequent sections that describes the results obtained from poly(arylene)-based devices. The most typical OFET-device architecture is that in which a semiconductive thin film is deposited on a dielectric layer with an underlying gate contact (G) (Fig. 2.2, left side) (Saudari et al., 2010). In an OFET, the current flow between the source (S) and drain (D) electrodes (Id) upon the application of a drain-source bias (VD) is modulated by the bias applied between the gate (G) and source electrodes (VG). When Vg ¼ 0 V, Id is minimal, the device is in the off state. When Vg 6¼ 0 V is applied (VSG < 0 V for p-type transistors) the device turns on and charge carriers are accumulated at the semiconductor (SC)-dielectric interface, resulting in a gate-controlled Id. Principal OFET figures of merit include field-effect mobility (μ), current on-to-off ratio (Ion:Ioff), threshold voltage (Vth) defining the average charge carrier drift velocity, current

Fig. 2.1 Representative poly(arylene) semiconductors used for the fabrication of highly efficient OPV and OTFT devices, as reported in the literature.

Key trends in sustainable approaches to the synthesis of semiconducting polymers

47

VD – +

S

VG

D

SC

Dielectric G

Substrate

Fig. 2.2 Schematic representation of a bottom-gate top-contact OFET structure and a conventional OPV cell.

drain-source ratio between the on and off states, and the gate voltage at which the injected carriers are mobile. Fig. 2.2, right side, shows the typical layered configuration used for state-of-the-art bulk heterojunction organic photovoltaic (BHJ OPV) cells (G€unes et al., 2007; Yu et al., 1995). The BHJ cell photoactive layer is composed of a bicontinuous composite of donor (hole transporting, p-type) and acceptor (electron transporting, n-type) semiconductors. In a BHJ OPV device, the light passes through the transparent contact [typically indium tin oxide (ITO)], and is absorbed in the active layer to generate hole-electron pairs (excitons), which dissociate into free charges at the interface between the donor and the acceptor interface. Holes and electrons flow in the donor and acceptor regions, respectively, and are collected at the electrodes, resulting in the generation of electrical power. Principal figures of merit include power conversion efficiency (PCE), short circuit current (Jsc), open-circuit voltage (Voc), and fill factor (FF), which define the ratio between the output device electrical energy versus the input solar energy (Pin), the device current when no reverse bias is applied, the device voltage when no current flows through the cell, and the ratio between maximum power device and Jsc  Voc, respectively. PCEs of photovoltaic cells are calculated using the following equation: PCE ¼ FF  Voc  Jsc/Pin.

2.2.1 Metal-mediated polycondensation reactions This section gives a critical overview of the synthetic approaches used to access semiconducting poly(arylenes) efficiently, based essentially on the formation of CdC bonds by metal-mediated Suzuki-Miyaura cross-coupling, direct arylation via CdH activation technology, or Click-Chemistry.

2.2.1.1 Suzuki-Miyaura cross-coupling polymerization The polymerization based on the Suzuki-Miyaura reaction of dihalo- and bis (organoboron)-functionalized monomers (Suzuki, 1999) is sustainable in some respects, though it typically requires the presence of additives such as a base, water,

48

Handbook of Organic Materials for Electronic and Photonic Devices

and a phase transfer catalyst, which add significant complexity to the reaction setup. This is likely amplified at a large scale, where multiphase reactions may be slower and more difficult to control. Fig. 2.3(top) shows the catalytic cycle for this reaction. Furthermore, the stability of the boron-containing monomers may be an issue because commonly employed boronic acids tend to condense into trimeric cyclic anhydrides and, consequently, several synthetic protocols use an excess of boronic acid to ensure a complete conversion of the electrophilic component of the reaction. Additionally, their purification on a large scale may be challenging, as many of them are waxy solids.

Fig. 2.3 Mechanism of the Suzuki-Miyaura cross-coupling reaction.

Key trends in sustainable approaches to the synthesis of semiconducting polymers

49

The use of boronates solves these problems to a degree, although they are expensive and their use leads to a significant decrease of atom economy, which may make them less appealing. Finally, most of the synthetic approaches to boronic esters still require a lithiation reaction of the precursor, and their purification on a large scale is often challenging. Brouwer et al. (2011) successfully addressed the latter two issues by using bis(pinacolate)diboron (BiPi) to generate in situ the active boronate species for Suzuki polymerizations. Dibromo-substituted, thiophene-based monomers 1–4 (Scheme 2.1) were homopolymerized at 110°C (37%–74%) by employing the relatively expensive 1,10 -bis(diphenylphosphino)ferrocenepalladium(II)dichloride dichloromethane complex (Pd(dppf )Cl2) catalyst (5 mol%), K3PO4 as the base (5 eq.), and DMF as the reaction medium. It is important to recall that the use of DMF, although standard for Suzuki cross-coupling reactions, is clearly incompatible with the current drive toward more sustainable processes due to the toxicity issues previously mentioned. The copolymers P1–P4 exhibited molecular weights up to Mn ¼ 9.46 kDa and polydispersity indexes (PDIs) ranging from 1.80 to 2.86. The nature (and distribution) of the polymer chains’ end-groups correlated with the proposed mechanism for the BiPi method, relying on two palladium-mediated catalytic cycles—that is, Miyaura, which generated the active Br/B terminated monomer, and Suzuki, growing the polymer chain (Fig. 2.3). The authors also compared the BiPi method to the commonly used Stille copolymerization. The average P1 Mn and polydispersity were comparable to those achieved with the BiPi method (Mn ¼ 5.87 vs 5.20 kDa; PDI ¼ 1.61 vs 1.84), whereas the yield was higher (90% vs 77%). Despite the structural similarity to P1, P2 was produced with a lower yield (30%), Mn (8.73 kDa), and PDI (1.85) than under the BiPi conditions (yield ¼ 48%; Mn ¼ 8.92 kDa; PDI ¼ 2.56). The copolymer P3 was prepared in a similar yield of Suzuki polymerization (71% vs 74%), but with a much lower Mn (3.97 vs 9.46 kDa) and a similar PDI (2.20 vs 2.73). This result was ascribed to the stronger electron-withdrawing nature of fluorenes compared to thiophenes, which lead to an increased Suzuki turnover rate compared to the Miyaura one (Fig. 2.3). An opposite behavior was found for P4, which was obtained in higher yields compared than when using BiPi (87% vs 67%) and featured a nearly double Mn (10.09 vs 5.72 kDa), as well as a narrower polydispersity (1.99 vs 2.86). Notably, Mw was generally higher for the BiPi method, indicating that this latter produced longer chains than the Stille cross-coupling approach. Indeed, the BiPi method involves a homopolymerization of the active Br/B terminated monomer, which does not suffer from chain-terminating reactions as much as copolymerizations. It is worth highlighting that this method avoids the synthesis of boron- or tin-containing monomers, thus enabling the preparation of copolymers P1–P4 in fewer steps than through the equivalent standard Suzuki or Stille reactions. A significant advance toward sustainability in Suzuki methodology has been represented by the replacement of boronic acids/esters with trifluoroborates (Lee et al., 2011). Trifluoroborate salts may be easily accessed from readily available and inexpensive starting materials (i.e., they are chemically robust/highly tolerant functional groups that are stable to moisture/air and often are readily recrystallizable.

Scheme 2.1 Synthesis of thiophene-based P1–P4 polymers by Suzuki-Miyaura cross-coupling polymerization.

Key trends in sustainable approaches to the synthesis of semiconducting polymers

51

Additionally, they possess a relatively low molecular weight, and the by-products from cross-coupling reactions would be relatively benign inorganic materials. However, the introduction of the trifluoroborate requires both lithiation and fluorination (Molander and Ellis, 2007), thereby adding a step to those required to synthesize conventional boronic acids/esters. Thus, Lee et al. reported that the successful Suzuki cross-coupling reaction between dipotassium [2-(heptadecan-9-yl)-2H-benzotriazole]-4,7-bis(trifluoroborate; 5) and dibromobenzothiadiazole (6) to yield polymer P5 (Scheme 2.2) was composed of all electron-accepting units. The optimal polycondensation protocol involved a 1:1 comonomer molar ratio, Pd(PPh3)4 as the catalytic system, tetraethyl ammonium hydroxide as the base, LiOH H2O as an additive, and toluene/water as the reaction medium (90°C, 2h); it also afforded a high-Mn polymer P5 (120 kDa , PDI ¼ 1.5). This successful result was ascribed to the superior ability of lithium hydroxide (LiOH) at hydrolyzing the trifluoroborate into boronic acid (Yuen and Hutton, 2005). These findings suggested that dipotassium heteroaryl bis(trifluoroborate) can be used as a viable alternative to boronic acid/esters in Suzuki polycondensation. To expand the scope of the protocol, the statistical copolymer P6 (Mn ¼ 29 kDa, PDI ¼ 1.8) containing 20 mol% of a bithiophene unit was also synthesized under the same reaction conditions (Scheme 2.2). Both polymers P5 and P6 exhibited high decomposition temperatures (onset > 400°C), as indicated by thermogravimetric analysis (TGA). The highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) energy levels for P5 (5.4 and 3.4 eV, respectively) and P6 (5.0 and 3.15 eV, respectively) were measured from ultraviolet photoemission spectroscopy and optical absorption data. Solution-processed OFET devices based on both polymers showed ambipolar behavior, with hole mobilities of 4  105  7  105cm2/Vs and respectable electron mobilities of up

Scheme 2.2 Synthesis of benzothiazole-based copolymers P5 and P6 by Suzuki cross-coupling polymerization.

52

Handbook of Organic Materials for Electronic and Photonic Devices

to 2  102 cm2/Vs (P5) and 8  104 cm2/Vs (P6) (Ion/Ioff of about 1  104 and about 1  102, respectively). Recent advances in the development of greener Suzuki cross-coupling protocols to semiconducting polymers include the use of microwave radiation (Tsami et al., 2008; Nehls et al., 2006). For instance, Nehls et al. (2006) reported microwave-promoted Suzuki coupling between 2,5-dihexyl-1,4-phenylene diboronic acid (7) and 4,40 -didecyl-20 ,50 dibromoterephthalophenone (8) to yield the polymer P7 precursor to ladder-type p-phenylene P8 (Scheme 2.3A). Optimal conditions involved the use of a relatively inexpensive catalytic system Pd(PPh3)2Cl2 (about 4%) in tetrahydrofuran (THF)/4M aq. sodium carbonate (Na2CO3) under a constant microwave power of 150 W (about 130°C) for about 9 min. Under these conditions, polyketone P7 with Mn ¼ 12.6 kDa and a PDI of 1.8 was obtained in 61% yields. Notably, when compared to conventional synthesis (refluxing toluene/2 M aq. Na2CO3 for 1–3 days), the reaction time could be significantly reduced, although polymer molecular weights and yields were comparable. The authors also reported that by increasing the irradiating

Scheme 2.3 Synthesis of copolymers P7–P10 through microwave-promoted Suzuki-Miyaura cross-coupling polymerization.

Key trends in sustainable approaches to the synthesis of semiconducting polymers

53

power (70 ! 100 ! 150 W), a corresponding progressive increase in Mn values (3.7 !4.2 ! 12.6 kDa, respectively) were achieved. In a further work, Nehls et al. (2005) reported the microwave-assisted Suzuki coupling synthesis of the sterically hindered 1,5-naphthylene polyketone precursors to the polymers P9 and P10 in good yields (89% and 78%, respectively) and with a reasonable molecular weight (Mn ¼ 13.6 and 9.0 kDa, respectively), as well as PDI (2.4 and 2.9, respectively; Scheme 2.3B). Satisfactory coupling reactions could be achieved by employing the appropriate monomeric building blocks [i.e., 8/2,6-dialkyloxynaphthalene-1,5-bis(4,4,5,5-tetramethyl-1,3,2-dioxaborolate (9 or 10)] in a 1:1 molar ratio and using Pd(PPh3)2Cl2 (about 4%) as the catalyst. Microwave heating at 115°C (300 W, 10 min) was realized under heterogeneous conditions by using powdered KOH (12 eq.) as the base in dry THF. Interestingly, it has been demonstrated that the polymer Mn values increased almost linearly with the quantity of base. The authors claimed that the interaction between the base and the microwaves resulted in the generation of hot-spots within the reaction mixture, the concentration of which increased as more KOH was added; and that this thermal effect accelerated the cross-coupling reaction. Interestingly, in this study, P9 and P10 were attainable only under microwave conditions. Finally, Zhang et al. (2013) reported a straightforward, energy-saving, microwaveassisted Suzuki polymerization protocol to poly(9,90 -dihexylfluorene; P11; Scheme 2.4). Optimal conditions employed the monomers 2,7-dibromo-9,90 dihexylfluorene (11) and 2,7-bis(4,4,5,5-tetramethyl-1,3,2-dioxoborolan-2-yl)-9,9dihexylfluorene (12), PdCl2(dppf ) (1 mol%) as the catalyst, K3PO4 as the base (about 5 eq.), and THF/water (3:1 v/v) as the reaction medium. At the pulsed microwave power level of c150 W (130°C, 11 min), the polymer P11 was prepared with Mw up to about 40 kDa (PDI ¼ 1.9; 68% yield). Further increasing the power level, reaction temperature/time, or both led to the formation of some insoluble, gel-like, cross-linked polymers, which lowered the yield of soluble chains. Notably, a defect-free P11 polymer structure was achieved, as indicated by in-depth instrumental characterizations. When the polymer P11 was synthesized by a conventionally heated (80°C) analogous protocol, a markedly lower molecular weight was achieved after 48 h (Mn ¼ 13.24 kDa, PDI ¼ 1.51), whereas there was no significant effect on the reaction yields (60.5%). The use of heterogeneous catalysts may enable greener synthetic strategies to π-conjugated polymers via Suzuki cross-coupling (Liu et al., 2012). Indeed, by carrying out a reaction using a heterogeneous catalyst, purification of the desired reaction product can be simplified because of the facile recovery from the reaction mixture by filtration/centrifugation. In addition, in most cases, the need for chromatography may be avoided for metal removal, thus reducing operating costs. Importantly, the use of heterogeneous catalysts prevents metal contamination of the semiconductor, which may substantially affect device performance (see discussion of this point earlier in this chapter). Heterogeneous catalysts are also easier to handle, allowing the formation of inorganic salts to be avoided, and they are recoverable/recyclable and safer to be stored or discarded. Liu et al. (2012) reported for the first time the use of the low-cost, ligand free Pd/C catalyst for Suzuki coupling polymerization, obtaining very good results.

54 Handbook of Organic Materials for Electronic and Photonic Devices

Scheme 2.4 Synthesis of the fluorene-based polymer P12 by microwave-assisted Suzuki cross-coupling polycondensation.

Key trends in sustainable approaches to the synthesis of semiconducting polymers

55

Thus, N-(2-ethylhexyl)-2,7-(bis-40 ,40 ,50 ,50 -tetramethyl-10 ,30 , 20 -dioxaboralan-20 -yi)9H-carbazole (13; Scheme 2.5) was copolymerized at 85°C for 72 h with the diketopyrrolopyrrole (DPP)-based bis-bromo monomers 14 and 15, respectively, using Pd/C as the catalyst (5 mol% Pd), K2CO3 as the base (3 eq.), and THF/DMF/ H2O (1:1:0.5 v/v) as the reaction medium. These conditions led to the copolymers P12 and P13 in good yields (about 70%) and with molecular weights of Mn ¼ 14.5 kDa (PDI ¼ 1.5) and Mn ¼ 13.0 kDa (PDI ¼ 1.8), respectively, after Soxhlet extraction. The insoluble Pd/C catalyst was easily recovered after solvent extractions from the crude P12 and P13 polymers, but the activity of the recovered catalytic system decreased, resulting in low molecular weights. For comparison, the two polymerizations in Scheme 2.5 were investigated under the same reaction conditions, but using homogeneous Pd(PPh3)4. Interestingly, it was found that such a catalytic system afforded the polymers P12 and P13 with lower molecular weights (Mn ¼ 12.0 kDa and Mn ¼ 11.0 kDa, respectively) and yields (65% and 45%, respectively) compared to those obtained with Pd/C. Further, a higher polydispersity (PDI ¼ 2.0, PDI ¼ 2.1, respectively) was observed in the former case. Despite the use of the soluble Pd(PPh3)4 catalyst, atomic absorption spectroscopy measurements revealed that both polymers P12 and P13 contained a Pd level of 3668 and 3350 ppm, respectively, whereas when the heterogeneous Pd/C was used, the Pd level was reduced by about 100 times (34 and 56 ppm, respectively). In the same study, the effect of the residual palladium on the P12 OFET charge transport properties was investigated as well. Thus, the optimal hole mobility of the Pd/Ccatalyzed P12 (about 4  104 cm2/Vs) was found to be significantly higher than that of the Pd(PPh3)4 counterpart (about 1.4  105 cm2/Vs).

Scheme 2.5 The synthesis of diketopyrrolopyrrole-based copolymers P12 and P13 by heterogeneously catalyzed Suzuki cross-coupling polymerization.

56

Handbook of Organic Materials for Electronic and Photonic Devices

2.2.1.2 Direct (hetero)arylation polymerization Transition metal (TM)-catalyzed coupling reactions of (hetero)arene CdH bonds with aryl halides or aryl triflates [i.e., direct (hetero)arylation (Fig. 2.4)] have undergone an exponential development in recent years, and they are becoming a viable and sustainable alternative to traditional cross-coupling polymerizations (Bohra and Wang, 2017; Pouliot et al., 2016; Mercier and Leclerc, 2013; Facchetti et al., 2012). This class of cross-coupling processes does not require the preparation of organometallic reagents and needs fewer reaction steps, as well as generating less waste. However, control of the regioselectivity of this reaction can be controlled only when a given (hetero)arene exhibits preferential C-H reactive positions (Rudenko et al., 2012). Thus, achieving regioselectivity for unbiased substrates is required to greatly broaden the scope of direct (hetero)arylation reactions. Also, some aspects of this reaction are not user-friendly (e.g., strictly anhydrous conditions/inert environment are needed), and superheated solvents requiring pressurized reactors are often employed, which contribute to preventing its widespread use. This section aims at giving a critical overview of the seminal contributions to this field. Thus, in 1999, Sevignon et al. (1999) first employed the direct arylation protocols for the synthesis of semiconducting polymers. Thus, they prepared a series of poly(3alkylthiophene)s (P14 and P15) by polycondensation of the corresponding 2-bromo-3-alkylthiophene derivatives (16 or 17, respectively; Scheme 2.6, left side). The general procedure implied the use of Pd(OAc)2(5 mol%)/nBu4NBr as the catalytic system (Gozzi et al., 1997) and K2CO3 as the base in DMF at 80°C (48 h). Although the polymers were obtained in good yields (about 90%), they featured low molecular weights (Mn about 2.4–3.0 kDa; PDI about 1.5–2) and degree of polymerization (DP about 12–16). On the basis of 1H-NMR magnetic resonance measurements, the authors stated that head-to-tail regioregular (P14) polymer chains were achieved. A regioregularity of about 90% was reported for polymer P15. Subsequently, Ozawa and coworkers (Wakioka et al., 2014, 2013, 2015; Wang et al., 2010) successfully applied this strategy to the synthesis of a variety of conjugated polymers. For the poly(3-alkylthiophene) family in particular, the well-known head-to-tail poly(3-hexylthiophene) (Wang et al., 2010) (P3HT, P16; Scheme 2.6, right) with a high molecular weight (Mn up to 30.6 kDa; PDI ¼ 1.6) and high regioregularity (up to 98%) was prepared in quantitative yields (99%) starting from bromo-3-hexyl-thiophene (18) by employing the relatively expensive Herrmann’s catalyst (19) to generate in situ a palladium catalyst (ca. 1 mol%) with a tertiary phosphine ligand [i.e., tris(o-Me2N)phosphine (20; 2 mol%), featuring a high coordination

Fig. 2.4 Direct arylation for the synthesis of π-conjugated polymers.

Key trends in sustainable approaches to the synthesis of semiconducting polymers

57

Scheme 2.6 Synthesis of poly(3-alkylthiophene)s by Lemaire’s (left, P14 and P15) and Ozawa’s (right, P3HT) protocols.

ability]. The developed catalyst enabled the reaction in THF solvent, which was preferred to DMF or DMA—the reaction medium commonly used for palladiumcatalyzed direct arylation—because it ensured the solubility of P3HT throughout the reaction, though still presenting toxicity and safety issues. Optimal direct arylation/polycondensation was carried out at 120°C for 24 h in the presence of (expensive) Cs2CO3 (1 eq.) as base. Although the Herrmann’s catalyst (19) converted only partially to a catalytically active species, this route is substantially more atom-economical than the conventional Grignard metathesis (GRIM) polymerization to P3HT, which produces stoichiometric amount of MgBr2. On the basis of MALDI-TOF analyses, the authors proposed that homocoupling reactions between CdH and CdH bonds and CdBr and CdBr bonds were operative to a considerable extent at the initial polymerization stage, whereas the cross-coupling reaction between CdBr and CdH bonds dominated at a later stage. In a further contribution (Wakioka et al., 2013), the same group broadened the scope of this protocol by investigating the preparation of the structurally diverse poly [(9,9-dioctylfluorene-2,7-diyl)-(2,3,5,6-tetrafluoro-1,4-phenylene)] (PDOF-TP, P17) (Scheme 2.7A, left) via the polycondensation reaction of 1,2,4,5-tetrafluorobenzene (22) with 2,7-dibromo-9,9-dioctylfluorene (21). The unsatisfactory results obtained prompted the authors to screen a variety of palladium complexes. The use of the relatively expensive catalytic complex Pd2(dba)3CHCl3 (0.5 mol%) with the supporting ligand P(C6H4-o-OMe2)3 (2 mol%), together with pivalic acid (PivOH, 0.5 mol%), and Cs2CO3 as the base, in THF, allowed the synthesis of P25 with an extremely high molecular weight (Mn ¼ 347. 7 kDa, PDI ¼ 2.83) in a nearly quantitative yield (96%) after 24 h at 100°C. Note here that according to the reaction stoichiometry, an excess of expensive Cs2CO3 is necessary for generating pivalates in situ and trapping HBr generated during polycondensation. The UV-vis spectrum of polymer P17 exhibited a strong absorption band at 345 nm.

58

Handbook of Organic Materials for Electronic and Photonic Devices

Scheme 2.7 (A) Comparison between the direct arylation conditions to access fluorene-based copolymer P17 proposed by Ozawa (Wakioka et al., 2013) and Kanbara (Lu et al., 2011) and (B) synthesis of fluorinated conjugated copolymers by direct arylation polycondensation.

It is noteworthy to mention here that the same reaction was previously investigated by Lu et al. (2011) in DMA, and PDOF-TP (P17) with Mn up to 31.5 kDa was obtained (Scheme 2.7A, right). As already mentioned, DMA used in this reaction is a highly polar solvent, which is high in the list of harmful chemicals, being volatile and toxic and flammable. The solvent effect was also examined, but unfortunately, no polymeric product was obtained when using toluene or dichlorobenzene under the optimized conditions. Interestingly, P17’s molecular weight through direct arylation was about 10–100 times higher than that for the same polymer synthesized via the conventional Suzuki cross-coupling (Mn ¼ 3.2 kDa) reaction (Kameshima et al., 2001). Lu et al. (2012) also reported the preparation of several related fluorinated copolymers via direct arylation polycondensation (Scheme 2.7B). For instance, the direct

Key trends in sustainable approaches to the synthesis of semiconducting polymers

59

arylation of 1,2,4,5-tetrafluorobenzene (22) with 3,6-dibromo-N-octadecylcarbazole (23), 2,7-dibromo-N-octadecylcarbazole (24), or 2,8-dibromo-10,10-dioctyl-Nmethylphenasaziline (25) as a comonomer carried out in the presence of Pd(OAc)2 (5 mol%), a phosphine ligand (PtBu2Me/HBF4, 10 mol%), and K2CO3 (2 eq.) in N,N-dimethyl acetamide (DMA) as the reaction medium (100°C, 24 h) successfully afforded the high-Mn polymer P18 (or P19 or P20). The reaction by-products of these polycondensations (KBr, CO2, and H2O) were easily removed from the polymers. Nuclear magnetic resonance (NMR) and matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-MS) analyses suggested the presence of a cross-linked structure for P18 and P19 caused by the relatively high reactivity of carbazole CdH bonds. Indeed, direct arylation likely occurred at these CdH bonds, as well as at the CdH bonds (He et al., 2010) of monomer 22. The same group (Kuwabara et al., 2013) also reported the preparation of several 3,30 ,4,40 -tetramethylbithiophene-based alternating copolymers via direct arylation polycondensation. Examples include the reaction of 3,30 ,4,40 -tetramethylbithiophene (26) with 2,7-dibromo-9,9-dioctylfluorene (21) using Pd(OAc)2 (2 mol%)/PCy-HBF4 (4 mol%) as the catalytic system, and K2CO3/PvOH in DMA afforded after 24 h at 100°C, a high-Mn polymer P21 (Mn ¼ 36.1 kDa, PDI ¼ 2.76) in good yields (82%; Scheme 2.8). Notably, the removal of the phosphine ligand allowed the reduction of the reaction time (3 h) and led to good molecular weight Mn up to 31.8 kDa (PDI ¼ 2.46) in high yields (91%). NMR analyses did not reveal the presence of any branched structures, thereby suggesting that a negligible direct arylation reaction occurred at the aromatic CdH bonds of the fluorene units. Further, the reaction of 3,6-dibromo-N-octadecylcarbazole (23) as a monomer under these phosphine-free conditions proceeded to give the polymer P22 in 99%

Scheme 2.8 Synthesis of bithiophene-based copolymers P21–P25 by direct arylation polycondensation

60

Handbook of Organic Materials for Electronic and Photonic Devices

yield (Mn ¼ 11.7 kDa, PDI ¼ 1.93). On the other hand, the polycondensation of 2,7dibromo-N-octadecylcarbazole (24) afforded the polymer P23 in low yields (47%), although a very high Mn of 32.9 kDa (PDI ¼ 5.36) was achieved. This was attributed to the formation of cross-linking structures due to the high reactivity of the CdH bonds at the carbazole 3- and 6-positions, as mentioned previously, promoting undesired side reactions. Interestingly, when the reaction time was reduced to 1.5 h, polymer P23 was obtained in a 96% yield (Mn ¼ 26.0 kDa, PDI ¼ 2.76), thereby indicating that the direct arylation of the CdH bonds in the 3,30 ,4,40 tetramethylbithiophene moiety proceeded preferentially over those at the CdH bonds of the carbazole moieties in the early stage of the reaction. In addition, the authors investigated the direct arylation polycondensation of 3,30 ,4,40 -tetramethylbithiophene (26) as a monomer with 2,5-di-(2-ethylhexyl)-3,6bis(4-bromophenyl)]pyrrolo[3,4-c]pyrrole-1,4-dione (27) or 6,60 -dibromo-N,N0 -(2ethylhexyl)-isoindigo (28). Under conditions similar to those employed for the synthesis of P21–P23, the reaction between 26 and 27 afforded copolymer P24 with a Mn ¼ 18.1 kDa (PDI ¼ 2.35) in a 96% yield. Copolymerization with brominated isoindigo derivative 28 resulted in copolymer P25, featuring Mn ¼ 11.3 kDa and PDI ¼ 2.76 (82%). No branching was revealed by NMR and MALDI-TOF analyses for P24. On the other hand, P25 featured a branched structure, likely resulting from a direct arylation side reaction at the CdH bonds in the isoindigo moiety. Bulk heterojunction OPV cells based on copolymers P24 and P25 exhibited poor performance characterized by a similar Voc of about 1 V, Jsc in the range of about 1.5–3 mA/cm2, and FF about 30% under simulated standard solar light (AM 1.5 G, 100 mW/cm2), giving a maximum power conversion efficiency of 0.89% and 0.39%, respectively (1:2 weight ratio with PC61BM). The authors attributed the low short circuit current and FF to the methyl groups on the tetramethylthiophene moiety. Indeed, though preventing side reactions as discussed previously, such bulky groups lead to a twisted polymer backbone, which was unfavorable for light absorption and charge carrier transport. Additional studies from Lu et al. (2013) enabled the ruthenium (Ru)-catalyzed the direct arylation of pyrrole derivatives (e.g., 29) with 2,7-dibromo-9,9-dioctylfluorene 21 (Scheme 2.9) bearing a pyrimidinyl-directing group on the nitrogen atom. Experiments involving (Ackermann and Lygin, 2011) the relatively expensive [RuCl2(p-cymene)]2 (5 mol%) as the catalytic system, K2CO3 (6 eq.) as the base, pivalic acid (60 mol%) as an additive, and m-xylene as the reaction medium (140°C) afforded poly[(1-(2-pyrimidinyl)pyrrole-2,5-diyl)-(9,9-dioctylfluorene-2,7-diyl)] P26 with Mn up to 19.8 kDa (86% yield). NMR and MALDI-TOF-MS analyses demonstrated that the direct arylation proceeded selectively at the pyrrole α-positions and provided a well-defined polymer, with no formation of cross-linked structures. Next, the removal of the directing group from the pyrrole unit of polymer P26 was investigated because the steric hindrance of bulky substituents on aromatic repeating units might lead to limited π-conjugation along the main chain of the polymer, which ultimately is known to affect the device’s performance negatively. Thus, the 2-pyrimidinyl group was cleaved by treating P26 with an excess KOH in NMP at 200°C, to afford poly[(1-H-pyrrole-2,5-diyl)-(9,9-

Key trends in sustainable approaches to the synthesis of semiconducting polymers

61

Scheme 2.9 Ru-catalyzed direct arylation polycondensation to copolymer P27.

dioctylfluorene-2,7-diyl)] P27 in 82% yield. Importantly, GPC/1H NMR/MALDITOF-MASS data indicated minimal polymer decomposition and a nearly quantitative cleavage of the pyrimidinyl substituent. Further, by comparing the optical properties, it was evident that polymer P27 exhibited maximum red-shifted absorption relative to P35, both in solution and in the solid state (Δλmax ¼ 26 and 41 nm, respectively). This was ascribed to the enhanced coplanarity along the P27 main chain. Kuwabara et al. (2014) also focused on 3,4-ethylenedioxythiophene (EDOT) monomer 30 for the direct arylation polycondensation with 2,7-dibromo-9,9dioctylfluorene 21 (Scheme 2.10). Thus, the authors synthesized poly[(3,4-ethylenedioxythiophene-2,5-diyl)-alt-(9,90 -dioctylfluorene-2,7-diyl] P28 (L-PEDOTF, Mn ¼ 39.0 kDa and PDI ¼ 2.24) in a 85% yield by employing a 0.3-M monomer concentration, Pd(OAc)2 as the catalyst (2 mol%), potassium 1-adamantane carboxylate as the base, and DMA as the reaction medium (100°C, 6 h). Interestingly, by using microwave heating and changing the base to potassium pivalate while reducing the catalyst loading (1 mol%), the reaction temperature (80°C), and the monomer concentration (0.1 M), it was possible to obtain P29 with Mn ¼ 147 kDa (H-PEDOTF) after 30 min of reaction (85% yield). Importantly, the MALDI-TOF data for P29 obtained under conventional heating (L-PEDOTF) revealed the presence of Br groups at the chain ends due to an incomplete reaction, whereas the polymer synthesized using a microwave-assisted protocol did not contain terminal halogens, indicating a highly efficient reaction. Additionally, this polymerization occurs without branching caused by side reactions. Next, the authors compared direct arylation to the commonly used SuzukiMiyaura cross-coupling copolymerization (Scheme 2.10). To this end, polymer P30 (S-PEDOTF) was prepared (85% yield) using 2,5-dibromo-3,4ethylenedioxythiophene (31) and 9,9-dioctylfluorene-2,7-diboronic acid bis(1,3propanediol) ester (32), the X-phos-containing system 33 as the catalyst, K3PO4 as the base, and THF/water as the reaction medium, at 25°C (48 h). The average

62

Handbook of Organic Materials for Electronic and Photonic Devices

Scheme 2.10 Synthesis of EDOT-based polymers via direct arylation polymerization (P28, P29) and via Suzuki cross-coupling (P30).

S-PEDOTF Mn and PDI were 17.1 kDa and 2.08, respectively. Residual Br groups at the chain ends were observed by MALDI-TOF-MS measurements, similarly to P28 prepared via conventionally heated direct arylation (L-PEDOTF, as discussed earlier). The absorption spectra, the optical band gap (about 2.5 eV), and the HOMO energy (about 5.2 eV) of the polymers H-PEDOTF, L-PEDOTF, and S-PEDOTF were almost identical. To investigate the effect of various synthetic methods on the purity of the polymers P28–P30, inductively coupled plasma-atomic emission spectrometric (ICP-AES) measurements were performed. Interestingly, it was found that HPEDOFT and L-PEDOFT, which were synthesized using 1 mol% of catalyst, contained about 2000 ppm of palladium, whereas about two times the amount of palladium was observed in S-PEDOFT, for the preparation of which 5% mol of catalyst was needed. Further, the latter contained 470 ppm of phosphorous, which originated from the phosphine ligand of the catalyst. All these findings confirmed the importance of the synthetic method in terms of purity of the resulting polymer, along with the advantages of microwave heating. Within the same study, the effect of the polymer purity on its charge transport properties also was investigated. Thus, OFETs were fabricated using P28–P30 as prepared by microwave and thermal direct arylation, as well as palladium-catalyzed Suzuki protocol. The optimal hole mobility μh of H-PEDOFT (1.1  103 cm2/V/s) and Ion/Ioff ratio (about 104) was found to be slightly superior to that for L-PEDOFT (8.5  104 cm2/Vs, about 102), and remarkably higher than that for the

Key trends in sustainable approaches to the synthesis of semiconducting polymers

63

S-PEDOFT counterpart (3.1  105 cm2/Vs, about 102). The authors claimed that the high polymer purity, along with the high molecular weight of H-PEDOFT, contributed to the good OFET performance. Similarly, bulk heterojunction OPV cells based on copolymer H-PEDOFT performed the best, with a Voc of 0.83 V, a Jsc of about 9.4 mA/cm2, and a FF ¼ 0.52 under simulated standard solar light (AM1.5 G, 100 mW/cm2), giving a maximum power conversion efficiency of 4.08% (1:4 weight ratio with PC71BM as acceptor). On the other hand, OPV cells based on L-PEDOFT/PC71BM and S-PEDOFT/ PC71BM blends (1:4 wt/wt ratio) exhibited power conversion efficiencies of 2.55% (Voc ¼ 0.78 V, Jsc about 7.6 mA/cm2, FF ¼ 0.42) and 0.48% (Voc about 0.6 V, Jsc about 2.6 mA/cm2, FF about 0.3), respectively. Because extreme morphological variations of the BHJ layers were not observed, ruling out the influence of molecular weight, the authors mainly attributed the observed PCE difference within this polymer series to variation in impurities. Indeed, it was demonstrated a significantly higher trap-assisted photogenerated carrier recombination in S-PEDOTF-based OPV compared to L-PEDOTF- and H-PEDOFT-based OPVs. Berrouard et al. (2012) investigated the direct arylation polycondensation reaction of a thieno[3,4-c]pyrrole-4,6-dione (TPD)-based polymer (P31; Scheme 2.11A). Indeed, π-conjugated polymers containing TPD moieties are known to exhibit excellent performance in both solar cells (PCEs up to 7.3%) and field effect transistors (hole mobilities up to 0.6 cm2/Vs) (Guo et al., 2011). Note also that the imide group in TPD may act as an orienting/activating group for the CdH bonds at the 2- and 20 -positions (Thirunavukkarasu et al., 2010). Thus, the authors synthesized polymer P31 by employing 5-(2-hexyldecyl)-5Hthieno[3,4-c-pyrrole-4,6-dione 34/5,50 -dibromo-4,40 -dioctyl-2,20 -bithiophene (35) comonomers (1:1 molar ratio, 0.25 M), Herrmann’s catalyst (19) as a palladium source (4%), tris(o-Me2N)phosphine (20; 8 mol%) as the ligand, Cs2CO3 (2 eq.), and THF as the solvent (120°C, 22 h) under pressure. These conditions, which are optimal, led to P31, with a high Mn (56 kDa, PDI ¼ 2.6), as well as high yield (96%). For comparison, P31 was also prepared by Stille polymerization between 1,3-dibromo-5-(2-hexyldecyl)-5H-thieno[3,4-c]pyrrole-4,6-dione (36) and 4,40 dioctyl-5,50 -bis(trimethylstannyl)-2,20 -bithiophene (37; 1:1 molar ratio, about 0.2M) (Yuan et al., 2010), employing a 2 mol% [Pd2(dba)3] coupling catalyst, P(o-tol)3 (16 mol%), and chlorobenzene as the reaction medium (130°C, 48 h). The copolymer P31 was obtained with a moderate molecular weight Mn ¼ 9 kDa (PDI ¼ 1.5, 71% yield), which was ascribed to the loss of some functional groups during the cross-coupling reaction. The UV-vis absorption spectra of the two polymer samples exhibited similar features, with an absorption maximum at 464 nm and 474 nm (CHCl3) for the Stille- and the direct arylation-polymerized P31, respectively. Their solid-state UV-vis absorption spectra featured a bandgap of about 1.75 eV, but the direct arylation-polymerized P31 exhibited a red-shifted absorption maximum (about 18 nm), which was explained by the different molecular weights, solid-state morphology, or both. The 1H-NMR spectra were comparable to those reported for analogous polymers, only differing in the alkyl-side chain area. Differential scanning calorimetry revealed a slightly greater crystallinity, as well as solid-state stability for

64

Handbook of Organic Materials for Electronic and Photonic Devices

Scheme 2.11 (A) Comparison of the direct arylation and Stille polymerization approach to access the TBD-based polymer P31; (B) synthesis of conjugated TPD-based copolymers P32– P34 by direct arylation polycondensation.

P31 obtained via a direct arylation polycondensation protocol, likely owing to its higher molecular weight. Finally, X-ray diffraction (XRD) patterns showed analogous features for both polymers, indicating they were similarly organized in the solid state. Jo et al. (2012) also prepared two related TPD-based copolymers P32 and P33 via direct arylation polycondensation of 5-(9-heptadecanyl)-4H-thieno[3,4-c]-pyrrole4,6-dione (38) with 5,50 -dibromo-4,40 -didodecyl-2,20 -bithiophene (39) and 5,50 dibromo-4,40 -didodecyl-2,20 :50 ,20 -terthiophene (40), respectively (about 93% yields; Scheme 2.11B). These reactions were carried out under the same conditions used for the synthesis of P31 in order to achieve an Mn as high as 50 kDa (PDI ¼ 2) for copolymer P32, whereas P33 featured Mn ¼ 41 kDa (PDI ¼ 2.3). The photovoltaic properties of P32 (or P33)/PC71BM were investigated in a BHJ OPV structure consisting of ITO/PEDOT:PSS/P32 (or P33): PC71BM/TiOx/Al. The P32-based devices demonstrated a short-circuit current density of 1.03 mA/cm2, a Voc of 0.95 V, a FF of 0.38, and a PCE of 0.38% (1:1.5 weight ratio with PC71BM). When 3 vol% additive (i.e., 1-chloronaphthalene, CN) was used to process the photoactive layer, a PCE of 1.90% was achieved. The performance increase was attributed to improved thin-film P32/PC71BM film morphology, which led to enhanced Voc

Key trends in sustainable approaches to the synthesis of semiconducting polymers

65

(0.90 V) and FF (0.42) values. On the other hand, OPVs based on P33 were found by optimizing the donor:acceptor ratio (1:1 wt/wt), to show an outstanding PCE of about 6% (Voc of about 0.9 V, Jsc of about 10.5 mA/cm2, and FF of about 0.6) for the time. The authors claimed that such superior performance arose from the side-chain distance along the length of conjugated backbone between P32 and P33 composed of bithiophene- and terthiophene-TPD repeat unit, respectively. Indeed, a wide sidechain gap was previously demonstrated to ensure conformational interdigitation, thereby enabling a better packing of the polymer main chains (Koppe et al., 2007). In a subsequent contribution, the same group (Wang et al., 2013) reported the dithienylselenophene-thienopyrrolodione copolymer P34 (Scheme 2.11B) as an analog of P33, where a selenium replaced the central ring sulfur atom of the terthiophene unit. The reaction conditions were modified by adding pivalic acid in the catalytic system. Thus, the coupling protocol involved 2,5-bis(4-dodecyl-5-bromothien-2-yl) selenophene (41) and 5-(9-heptadecanyl)-4H-thieno[3,4-c]-pyrrole-4,6-dione (38) as comonomers (1:1 molar ratio, 0.2 M concentration), Hermann’s catalyst as a palladium source (4 mol%), tris(2-methoxyphenyl)phosphine as the ligand (8 mol%), Cs2CO3 as the base (2.7 eq.), and pivalic acid (30 mol%) in THF (120°C, 22 h). The copolymer P34 was obtained in 94% yields with a Mn as high as 36.0 kDa (PDI ¼ 1.97). Interestingly, BHJ solar cells fabricated using PC71BM as the electron acceptor (1:1 wt/wt) exhibited PCEs up to 5.80% (Voc ¼ 0.88 V, Jsc ¼ 10.74 mA/cm2, and FF ¼ 0.62) when Clevious PVP was employed as the hole-transport interlayer. Processing additives were not found to be beneficial in this blend system. The Leclerc group reported other interesting alternating copolymers (Mercier et al., 2013; Beaupre et al., 2012) consisting of a series of furo[3,4-c]pyrrole-4,6dione-based systems P35 (Scheme 2.12), where the sulfur atom in the heterocycle of TPD was replaced by an oxygen atom. For instance, direct arylation between 5-(octyl)furo[3,4-c]pyrrole-4,6-dione (42) and diverse dithieno[3,2-b:20 ,30 -d]silole

Scheme 2.12 Synthesis of the furo[3,4-c]pyrrole-4,6-dione-based copolymer P35 through direct arylation polymerization.

66

Handbook of Organic Materials for Electronic and Photonic Devices

monomers (43–47) were prepared using Herrman’s catalyst (19; 4 mol%), tris (o-methoxyphenyl)phosphine (8 mol%), Cs2CO3 (2.4 eq.), and pivalic acid (40 mol %) in toluene (120°C; Scheme 2.12). Good average molecular weights in the 12–16-kDa range were achieved (PDIs ¼ 1.3–2), with the maximum achieved for P35b. Note that P35b also featured the longest absorption maximum; indeed, it exhibits straight alkyl chains on the silicon atom, thereby suggesting a greater solid-state ordering. BHJ solar cells fabricated with these polymers using PC71BM as the electron acceptor (1:2 wt/wt) exhibited power conversion efficiencies under 2%, with the largest measurable value for P35b (Voc ¼ 0.76, Jsc ¼ 7.8 mA/cm2, FF ¼ 0.35). The poor results were attributed to the lack of planarity in the polymer backbone, which led to low Jsc values. Grenier et al. (2013) also explored direct arylation polycondensation to prepare n-type copolymers P36 and P37 (Scheme 2.13) using the brominated isoindigo derivative (42) and 5-octylthieno[3,4-c]pyrrole-4,6-dione (43) or 5,50 -dioctyl-1,10 -4Hbithieno[3,4-c]pyrrole-4,40 ,6,60 (5H, 50 H)-tetrone (44) as electron-withdrawing monomers, respectively. Under the conditions already used for P31 (Scheme 2.11, as discussed previously), copolymer P36 with Mn ¼ 24 kDa and a PDI of 2.2 was obtained in a 77% yield, whereas P37 (80% yield) featured an Mn value of 18 kDa (PDI ¼ 2.2). The synthesis of P36 and P37 was also performed using standard Suzuki conditions between 6,60 -(N, N0 -2-hexyldecyl)-pinacolatodiboronisoindigo (45) and 1,3-dibromo-TPD (or dibromo-TPD dimer). Only the catalytic system Pd2dba3/tri(o-tolyl)phosphine/tetraethylammonium hydroxide yielded some polymeric product P36 (Mn ¼ 13 kDa, PDI ¼ 1.3; 16% yield), but copolymer P37 could not be isolated. Furthermore, the Suzuki-/direct arylation-polymerized P36, as well as P37, exhibited similar thin-film UV-vis spectra, with absorption maxima located in the range of 611–618 nm and optical band gaps of about 1.7 eV. The HOMO/LUMO energy levels of all polymers were relatively low (i.e., about 6.1 and 4.1 eV, respectively), suggesting they might be good acceptor semiconductors for all-polymer solar cells. Indeed, such energy levels are almost as low as PC61BM (about 6.1 and 4.3 eV, respectively). Finally, the authors investigated OFET properties of the direct arylation-polymerized P36 and P37, which exhibited electron mobilities (about 103 cm2/Vs) comparable to that of PC61BM. Note that in a later contribution, Grenier et al. (2017) reported an interesting alternative protocol for the direct (hetero)arylation to afford P36, involving biphasic water/ toluene conditions, a low-cost base (K2CO3), and PdCl2(PPh3)2 catalytic systems that were revealed to be tolerable to oxygen to some extent, thereby enabling the reaction in the presence of air. A phase-transfer agent (i.e., tetrabutyl ammonium bromide, or TBAB) was also required to avoid premature precipitation of the polymer; however, this leads to an increase in the complexity of the reaction mixture. Complete P36 characterization demonstrated that this protocol provided comparable properties than the very best values published to date (Scheme 2.13). Pouliot et al. (2013) also investigated the use of the DPP monomer (46) as a dibromoarene coupling component (Scheme 2.14). Indeed, DPP is a well-studied,

Key trends in sustainable approaches to the synthesis of semiconducting polymers

67

Scheme 2.13 Synthesis of the isoindigo conjugated copolymers P35 and P36 by direct arylation.

electron-poor monomer that has been incorporated into many π-conjugated polymers exhibiting extremely high charge carrier mobilities, which is a highly desirable feature for organic electronics applications (Li et al., 2016). Thus, polymers P38 and P39 were prepared by employing 5-octylthieno[3,4-c]pyrrole-4,6-dione (43) and 5,50 -dioctyl-1,10 -4H-bithieno[3,4-c]pyrrole-4,40 ,6,60 (5H, 50 H)-tetrone (44) as comonomers, respectively, under reaction conditions analogous to those for the synthesis of P36 (as previously discussed), except for the reaction medium (toluene). Copolymers P38 and P39 (86%–91% yields) reached Mn values of 21 kDa (PDI ¼ 2.1) and 28 kDa (PDI ¼ 3.0), respectively, without evidence of cross-linking reactions. For comparison, the boronate DPP derivative (47) was copolymerized with dibrominated TPD using previously published Suzuki cross-coupling conditions (Mohebbi et al., 2011). Interestingly, P38 was obtained with a low Mn (7 kDa) and a narrow PDI, suggesting that for this polymer, direct arylation is more effective for achieving high molecular weights while reducing synthetic steps. The copolymers P38 and P39 were found to exhibit a broad solid-state absorption between 600–900 nm, resulting in low band gaps (1.31–1.38 eV). Preliminary OFET

68

Handbook of Organic Materials for Electronic and Photonic Devices

Scheme 2.14 Synthesis of the DPP-based copolymers P38 and P39 by direct arylation polymerization.

measurements indicated electron mobilities of about 103 cm2/Vs for both polymers in agreement with their low-lying LUMO energies (< 4 eV). Almost simultaneously, Sherf’s group (Kowalski et al., 2012) and Horie’s group (Chang et al., 2012) independently reported the preparation of poly[4,4-bis (2-ethylhexyl)-4H-cyclopenta[2,1-b;3,4-b0 ]dithiophene2,6-diyl-alt-2,1,3-benzothiadiazole-4,7-diyl] (PCPDTBT, P40) by reacting 4,7-dibromo-2,1-3 benzothiadiazole (6) with 4,4-bis(2-ethylhexyl)4H-cyclopenta[2,1-b; 3,4-b0 ]dithiophene (48; Scheme 2.15). The Scherf group (Kowalski et al., 2012) optimized a ligand-free, direct arylation polycondensation scheme employing Pd(OAc)2 (4 mol%) and 1.5 eq. of K2CO3 in polar (toxic) DMA (110°C, 24 h) without the addition of any carboxylic acids. The copolymer P40 was obtained in 70% yields, with a Mn ¼ 40.3 kDa (PDI ¼ 3.48). Horie’s optimal route (Chang et al., 2012) differed in that it used polar (toxic) N-methylpyrrolidone as the reaction medium, pivalic acid in the catalytic system (30 mol%), and an excess of K2CO3 (2.5 eq), at a lower temperature (80°C). The resulting polymer P40 (98% yields) featured an Mn of 37.7 kDa (PDI ¼ 4.60) before Soxhlet extraction (70% yield). Interestingly, the average molecular-weight values of

Key trends in sustainable approaches to the synthesis of semiconducting polymers

69

Scheme 2.15 Synthesis of PCPDTBT via direct arylation polymerization.

polymers from both direct arylation strategies were higher than those achieved by conventional Stille (Mn ¼ 28 kDa, PDI ¼ 1.5) (Zhu et al., 2007) or Suzuki polymerization methodologies (Mn ¼ 15.3 kDa, PDI ¼ 2.05) (Horie et al., 2010). However, in both of these cases, branched as well as cross-linked polymers formed, as indicated by 1H-NMR and MALDI-TOF-mass spectrometry analyses. BHJ solar cells with PC61(or 71)BM as the acceptor component were fabricated with P40 (PCPDTBT) as the donor synthesized by Chang and coworkers, using 1,8octanedithiol as the processing additive. The highest PCEs were measured for the sample with the largest Mn (PCE up to 3.98%, Mn ¼ 70 kDa), greater than the performance obtained for the Suzuki-coupled polymer (PCEs up to 3.74%). Abdo et al. (2013) used 4,7-bis(3,30 /4,40 -hexylthiophene-2-yl)benzo[c][2,1,3] thiadiazoles (48) as building blocks to synthesize the low-bandgap π-conjugated copolymer P41–P43 series (80%–95% yields) via microwave-assisted direct arylation protocols using an equimolar amount of 3,4-ethylenedioxythiophene (EDOT; 30), bis-EDOT (50), and 2,3-dimethyl-thieno[3,4-b]pyrazine (51; about 12 mM; Scheme 2.16). The polymerization reactions were performed in DMF in the presence of KOAc (6 eq.), Bu4NBr (2 eq.), and a catalytic amount of Pd(OAc)2 (0.2 eq.) under microwave heating (150°C, about 30 min). Good Mn values of 15.8–22.2 kDa were achieved (PDIs ¼ 1.50–1.83). Note that under thermal conditions (100°C, 72 h), only oligomeric P41 featuring R ¼ H, R0 ¼ hexyl could be prepared, with a maximum Mn ¼ 3.9 kDa (PDI ¼ 3.82). The authors claimed that 1H-NMR and UV-vis spectral data of all the direct arylation-polymerized P41–P43 were identical to those of the corresponding polymers prepared via the Stille method. An interesting advance in direct heteroarylation methodology was recently reported by Hayashi et al. (2016), who explored heterogeneous catalysis in directarylation polycondensation. Particularly, the authors succeeded in affording P3HT

70

Handbook of Organic Materials for Electronic and Photonic Devices

Scheme 2.16 Synthesis of the low-bandgap copolymers P41–P43 through direct arylation.

(90% yield) with a high molecular weight and high regioregularity (Mn of about 16.3 kDa, PDI ¼ 3.17, HT 97%) by employing ligand-free heterogeneous catalysis (Hayashi et al., 2015). Optimal direct arylation of 2-bromo-3-hexyl-thiophene (18) was carried out at 120°C for 24 h, employing K2CO3 as the base (2 eq.), PivOH (1 eq.), and 10% Pd/C (2.5 mol% Pd) as the catalyst in NMP (Scheme 2.17). Notably, when P3HT was synthesized by using a homogeneous Pd(OAc)2 catalyst, a marked drop in yield was observed (10%), along with a large amount of insoluble products (>55%), as a consequence of the formation of networked structures. Recently, Dudnik et al. (2016) reported the preparation of poly-[5-(2-hexyldodecyl)-4H-thieno[3,4-c]pyrrole-4,6(5H)-dione-1,3-yl-alt-4,400 -dodecyl-2,20 :50 ,200 terthiophene-5,500 -diyl] (P44; Scheme 2.18) via direct arylation polycondensation between N-alkylated TPD (52) and dibromo-4,400 -dodecyl-2,20 :50 ,200 -terthiophene (53). The protocol showed the added value of being performed in 2-methyltetrahydrofuran (2-MeTHF), which is a green solvent derived from renewable sources (Anderson, 2012; Henderson et al., 2011; ECHA, n.d.). By employing Pd2dba3/P (2-MeOPh)3 as the catalyst pair (1:4), excess Cs2CO3 as the base, and 25 mol% PivOH, the target polymer P44 was obtained with Mn of 19.2 kDa (PDI ¼ 2). The same

Scheme 2.17 Synthesis of P3HT by Pd/C-catalyzed direct arylation polycondensation.

Key trends in sustainable approaches to the synthesis of semiconducting polymers

71

Scheme 2.18 Synthesis of TPD-based copolymers P44–P46 by direct arylation polymerization in a 2-Me-THF green solvent.

authors reported a series of high-performance D-A alternating copolymers (e.g., P45 and P46; Scheme 2.18) synthesized through a similar synthetic scheme, except for the use of a PivOH additive, which was replaced with 2,2-diethylhexanoic acid (DEHA). Indeed, the bulkier nature of DEHA compared to pivalic acid played a crucial role in enhancing the C-H selectivity without negatively affecting the C-H reactivity. All these defect-free polymers showed structural and optical qualities equal to (or even better than) the analogs synthesized via Stille polymerization. Additionally, the photovoltaic performance of P44 and P45 were tested in BHJ solar cell devices with the inverted architecture: ITO/ZnO/copolymer:PC71BM/MoO3/Ag, whereas a conventional cell architecture ITO/poly(3,4-ethylenedioxythiophene): poly(styrenesulfonate) (PEDOT:PSS)/copolymer:PC61BM/LiF/Al was fabricated for P46. Copolymers P44 and P45 synthesized by direct arylation exhibited a PCE of 6.86% (Voc ¼ 0.81 V, Jsc ¼ 12.9 mA/cm2, FF ¼ 65.8%) and 8.19 (Voc ¼ 8.19 V,

72

Handbook of Organic Materials for Electronic and Photonic Devices

Jsc ¼ 15.5 mA/cm2, FF ¼ 68.3%), respectively, which resulted only slightly lower PCEs (7.2% and 8.24%, respectively) than for their counterparts through Stille polymerization. Notably, P45 exhibits the highest efficiency reported to date for a copolymer through direct arylation. Finally, P46 was found to outperform the photovoltaic performance of its Stillederived counterpart (PCE ¼ 5.7% vs 5.1%) (Table 2.1).

2.2.1.3 Click polymerization Click chemistry shares with green chemistry some of the most fundamental principles. The concept of “click chemistry” was first introduced by the Nobel Prize B. Sharpless (Kolb et al., 2001). A set of stringent criteria must be met by a process in order for it to be considered a “click-type” reaction, including (1) generating inoffensive by-products removable by nonchromatographic methods, (2) being carried out under simple reaction conditions, (3) being “spring-loaded” for a single trajectory (i.e., characterized by a high thermodynamic driving force that drives it quickly and irreversibly to high yield of a single reaction product, with high reaction specificity), and (4) involving readily available starting materials and reagents without the use of or only using benign solvents. The copper(I)-catalyzed cycloaddition of azide and alkyne to selectively give 1,2,3-triazoles (CuAAC) reported independently by the research groups of Sharpless (Rostovtsev et al., 2002) and Meldal (Tornøe et al., 2002) is an archetypal example of a click reaction (Fig. 2.5). Since then, click chemistry has found widespread applications in various research areas (Gehringer and Laufer, 2017; Singh et al., 2016). The utility of click reactions in polymer chemistry is also gaining more and more interest. Although most studies reported the postfunctionalization of preformed polymers, recent efforts employed this reaction as an efficient polymerization technique (i.e., click polymerization) (D€ ohler et al., 2017; Barner-Kowollik, 2011). The azidealkyne click reaction in particular holds the promise to become a powerful synthetic tool to develop unprecedented functional materials that are expected to open new opportunities for organic electronics applications. Although several studies indicate that the conjugation in triazole-derived structures is not extended through the triazole moiety ( Jarowski et al., 2008), in the last decade, many interesting examples appeared that suggested changes in the direction of properly designed materials with extended π-electron systems that incorporated 1,2,3triazole moiety into the conjugation path. Triazole-containing molecular systems may function as electron or even ambipolar transporting materials in the fabrication of organic light-emitting diodes (Cheng et al., 2011). Moreover, Mirkin et al. (2011) and Chen et al. (2009) recently demonstrated that the triazole ring-maintained conjugation required for electronic transport in molecular transport junctions to bridge nanogaps, thereby enabling the creation of nanoelectronic devices with diverse functions and applications. Finally, De Miguel et al. (2011) clearly demonstrated that aromatic 1,2,3-triazoles may represent excellent conjugative π-linkers for rapid and efficient photoinduced electron transfer between remote electron donor and electron acceptor moieties—namely, zinc porphyrin and C60, respectively. This is an important

Polymer

Synthetic method

P5

Suzuki

P6

Suzuki

P12

P30 P32 P33 P34 P35b

Suzuki, heterogeneous catalyst Suzuki, homogeneous catalyst Direct arylation Direct arylation Direct arylation, thermal Direct arylation, MW assisted Suzuki Direct arylation Direct arylation Direct arylation Direct arylation

P36 P37 P38

Direct arylation Direct arylation Direct arylation

P24 P25 P28 P29

Voc (V)

Jsc (mA/cm2)

FF (%)

PCE (%)

Mobility (cm2/ Vs) 4  105(μh), 2  102(μe) 7  105(μh), 8  104(μe) 4  104(μh)

Ion/Ioff

References

1  104 1  102

Lee et al. (2011) Lee et al. (2011) Liu et al. (2012)

1.4  105 (μh)

Liu et al. (2012)

1 1 0.78

3 1.5 7.6

30 30 42

0.89 0.39 2.55

8.5 104(μh)

9.2 103

Kuwabara et al. (2013) Kuwabara et al. (2013) Kuwabara et al. (2014)

0.83

9.4

52

4.08

1.1 103(μh)

1.6 104

Kuwabara et al. (2014)

0.60 0.90 0.90 0.88 0.76

2.6 1.03 10.5 10.74 7.8

60 42 60 62 35

0.48 1.90 6.0 5.80 2.0

3.1 105(μh)

4.3 102

Kuwabara et al. (2014) Jo et al. (2012) Jo et al. (2012) Jo et al. (2012) Mercier et al. (2013) and Beaupre et al. (2012) Grenier et al. (2013) Grenier et al. (2013) Pouliot et al. (2013) Continued

73

3.0 104(μe) 3.5 103(μe) 3.0 103 (μe)

Key trends in sustainable approaches to the synthesis of semiconducting polymers

Table 2.1 Summary of BHJ photovoltaic cells and OFETs performance characteristics of poly(arylene)s in this chapter and related references

74

Polymer

Synthetic method

P39

Direct arylation

P40

Direct arylation Suzuki Direct arylation Direct arylation Direct arylation

P44 P45 P46

Voc (V)

Jsc (mA/cm2)

FF (%)

PCE (%)

Mobility (cm2/ Vs) 3

3.0 10 (μe) 0.627 0.640 0.81 0.77 0.99

13.92 12.71 12.9 15.5 10.0

45.5 46.0 65.8 68.3 57.8

3.98 3.74 6.86 8.19 5.71

Ion/Ioff

References Pouliot et al. (2013) Chang et al. (2012) Horie et al. (2010) Dudnik et al. (2016) Dudnik et al. (2016) Dudnik et al. (2016)

Handbook of Organic Materials for Electronic and Photonic Devices

Table 2.1 Continued

Key trends in sustainable approaches to the synthesis of semiconducting polymers

Click

75

Fig. 2.5 Synthesis of 1,2,3 triazoles via 1,3-dipolar cycloaddition of azides and terminal alkynes.

+ H

R

+

R⬘

N

N N

Cu+

N N R⬘ N R 1,2,3-triazoles

factor in applications of organic materials in solar energy storage and photovoltaic devices (Fukuzumi et al., 2014). Similar behaviors have been observed independently by several other groups (Ladomenou et al., 2016) when exploring the use of 1,2,3triazoles to bridge donor-acceptor groups featuring different substitution patterns. In this section, we survey seminal papers exploring the utility of the CuAAC reaction to afford π-conjugated, heteropolymeric structures for organic electronics. The key physical and morphological properties of the resulting materials have been discussed whenever available. Several interesting studies (Kumari et al., 2016; Lang et al., 2013) have been carried out on the use of azide-alkyne click reaction as an effective strategy for tuning the associated self-assembly properties in the postfunctionalization of preformed conjugated polymers. At any rate, a survey of these methods is beyond the scope of this chapter. For a comprehensive picture of the click chemistry approach to diverse π-conjugated polymers (e.g., not the polyarylene-type), including reactions other than CuAAC featuring the essential “click” attributes, the reader is referred to a recent pertinent review article (Marrocchi et al., 2016b). The first report on the synthesis of conjugated polymers by the Cu(I)-catalyzed 1,3dipolar “click” reaction was published in 2005 by van Steenis et al. (2005), who had the intuition of preparing linear poly(triazole)s P47–P48 via polymerization of 2,7diazidofluorene (54) and aromatic diynes 55–57 (Scheme 2.19). The reaction between 54 and 2,5-diethynylpyridine (55) was performed in THF/ CH3CN (9:1 v/v), by employing a 1:1 comonomer molar ratio (about 0.02 M concentration), Cu (about 14 mol%)/Cu(OAc)2 (about 2 mol%) as the catalyst, and tris(benzyltriazolylmethyl)amine (about 0.8 mol%) as the ligand (25°C, 170 h). It is worth mentioning that copper was removed from the reaction mixture via an extensive extraction procedure using N,N-diethylcarbamodithioic acid as a copper salt scavenger. The copolymer P47 was achieved with an Mn up to about 25 kDa (PDI ¼ 1.9). Under the same conditions, copolymers P48 and P49 were obtained with Mn up to 327 kDa (PDI ¼ 1.21) and Mn ¼ 8 kDa (PDI ¼ 1.61), respectively. Interestingly, when the reaction between 54 and 55 was carried out at 10°C for 65 h, the molecular weight of P47 was higher (Mn ¼ 20.6 kDa, PDI ¼ 2.86) than that of the polymer obtained at 25°C for 170 h, thereby suggesting an exothermic nature of the click polymerization reaction. Unfortunately, no reaction yields were reported. The fluorenecontaining copolymer absorption spectra were found to be the superposition of those of the monomers.

76

Handbook of Organic Materials for Electronic and Photonic Devices

Scheme 2.19 Synthesis of click polymers P47–P48.

Furthermore, the authors observed that P47–P48 were highly emissive in THF, with photoluminescence peaks located at 360–380 nm (i.e., the emission was dominated by the fluorene units). The highest quantum yield (ΦF ¼ 55%) was achieved for P49. These findings indicate a poor electronic communication between the different polymeric moieties. However, cyclic voltammetry measurements on a model compound consisting of two piridyl-triazole units bridged by a fluorene unit (i.e., the product from the reaction between 54 and 55) showed one two-electron reduction (1.86 V) but no oxidation, suggesting that the present materials might be useful as conductive materials. Shortly thereafter, Bakbak et al. (2006) prepared the conjugated poly(triazole)s P50–P51 (Scheme 2.20) by reacting 2,7-diazidofluorenes (54), and 4,40 - diazido3,3-dimethoxy-biphenyl (59) with 2,5-dialkyl-1,4-diethynyl benzenes (58) in THF at room temperature, in the presence of CuSO4 (5 mol%)/sodium ascorbate (SA, 5 mol%). These copolymers were obtained in high yields (80%–92%), but for reactions lasting ten days, and their molecular weights were only moderate, with a maximum Mn value of 8.7 kDa achieved for P50-type copolymer (PDI ¼ 5.8). According to 1H- and 13 C-NMR spectroscopy, all poly(triazole)s P50–P51 exhibited regioregular 1,4substitution of the triazole unit and showed blue fluorescence in solution, with quantum yields ΦF up to about 40%. On the other hand, no fluorescence was observed in the solid state, which is likely due to aggregation phenomena (Chen et al., 2003). The optical properties for all the copolymers were independent of their molecular weights, suggesting that the building blocks led to species with localized HOMO/LUMO orbitals. Interestingly, fluorescence measurements and theoretical calculations suggested that protonation at the 3-position of the triazole group in the P51-type

Key trends in sustainable approaches to the synthesis of semiconducting polymers

77

Scheme 2.20 Synthesis of the click polymers P50–P51.

polymer resulted in the decrease of the HOMO-LUMO gap and more delocalized frontier molecular orbitals. On these bases, it can be concluded that the use of alkyne/azide precursors substituted with electron-donating and electron-withdrawing groups, respectively, are expected to lower the band-gap of P50–P51 further, which may become of interest in semiconductor applications. Finally, the authors succeeded in writing crisp nanoscale features into P50-type thin films by using a tip of an atomic force microscopy (AFM) cantilever heated at about 225°C. The absence of tackiness and ripping led the authors to the conclusion that these materials were attractive as novel semiconductors that could have been easily thermally structured. Karim et al. (2008) reported the preparation of π-conjugated soluble poly(triazole)s P52–P53 (Scheme 2.21) by click polymerization (about 90% yield) of 2,7diazido-9,9-dioctylfluorene (50) with 2,7-diethynyl-9,9-dioctylfluorene (57), 4,7 diethynylbenzothiadiazole (60), and 2,7 diethynylcarbazole (61), respectively, employing a 1:1 comonomer molar ratio, CuSO4 5H2O (5 mol%)/sodium ascorbate (10 mol%)/triethylamine (0.2–0.3 mL) as the catalytic system in THF (30–35°C, 48 h). Triethylamine was required to make the polymerization reaction proceed. A maximum Mn value of about 8.3 kDa was obtained for P52 (PDI ¼ 1.92). All the copolymers were purified through multiple Soxhlet extractions with methanol, hexane, and chloroform. P52–P54 are stable up to 300°C (via TGA) and exhibit

78

Handbook of Organic Materials for Electronic and Photonic Devices

Scheme 2.21 Synthesis of the click polymers P52–P54.

optical absorption maxima in the range of 328–350 nm, both in solution and in the solid state, with P52 being the most red-shifted. The well-structured PL spectra of P52–P54 in solution revealed a blue light emission between 370 and 406 nm, whereas their emission spectra in the solid state are slightly red-shifted (15–40 nm). The HOMO and LUMO energy levels of P52 were found to be 5.23 and 3.25 eV, respectively, based on cyclic voltammetry and optical absorption data. Hwang et al. (2011) reported a stepwise methodology based on surface-initiated (Alonzi et al., 2013) Cu (I)-catalyzed click polymerization to synthesize the “brush” polymer P55 (Scheme 2.22). Surface-grafted brush polymers are an interesting alternative for high-performance, long-term-operation organic electronics, possibly helping charge injection and charge transport processes, which are crucial for many devices. Thus, the authors functionalized a quartz substrate with a monolayer of trimethylsilyl-acetylene initiator; next, the activated surface was immersed in a dimethylsulfoxide (DMSO) solution (10 mM) of the bis-azide monomer 62 (Scheme 2.22) in the presence of 10 mol% CuI at 40°C, followed by rinsing in CHCl3. Finally, the substrate was placed into a DMSO solution (10 mM) of the bisacetylene monomer 63 and 10 mol% CuI (40°C), followed by rinsing CHCl3. This sequence was repeated 34 times. Note here that the initially prepared monomer solutions could last for the entire duration of polymerization, thereby avoiding wasting monomers, although some decomposition of such solutions was observed. Moreover, no oligomer/polymer formation was detected in these solutions, which indicated that a controlled stepwise polymerization occurred only at the substrate surface. Atomic force microscopy (AFM) and polarization-dependent ultraviolet photoemission spectroscopy (UPS) data revealed that P55 films exhibited a thickness of

Key trends in sustainable approaches to the synthesis of semiconducting polymers

79

Scheme 2.22 Synthesis of the click polymer P55.

about 28 nm and a surface morphology featuring uniform cylindrical domains of about 90 nm in diameter, with tight packing density, and oriented normally to the substrate. Furthermore, the authors claimed that each domain likely featured a uniform, wellordered packing of the macromolecules. The UV-vis thin-film absorption spectrum exhibited a maximum of about 350 nm, with a broadband spanning to almost 700 nm, as well as an optical band gap of 2.52 eV. The HOMO and LUMO energy levels for P55 were found to be 5.28 eV and 2.76 eV, respectively, based on cyclic voltammetry experiments and optical absorption data.

2.2.1.4 Metal catalyst-free oxidative coupling polymerizations While the advantages offered by metal catalysts are unquestionable, including high activity, selectivity, and functional group tolerance, their replacement with organic substitutes is an attractive prospect for large-scale polymer preparation. This is readily rationalized by the relatively high costs of the transition metals, their contingent toxicity, and their demanding removal from final products (as discussed previously), especially in the case of polymers with complex side chains (Pei and Yang, 1996). In addition, the minimization of metal leaching into the final problem will be beneficial for the quality and performance of the target device, as discussed previously. Consequently, the development of efficient metal-free protocols for materials synthesis represents an important field of research that has seen remarkable growth over the last few years. Examples of these approaches include controlled radical

80

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 2.6 General scheme of CRP methods. Radical trapping by (A) activation/deactivation process; (B) “reversible transfer”, degenerative exchange process.

polymerization (CRP) (Murarka et al., 2012); i.e., atom transfer radical polymerization (ATRP), nitroxide-mediated radical polymerization (NMRP), ring opening polymerization (ROP). All the CRP methods rely on forming equilibria between active species (radicals) and various types of so-called dormant species. Radicals may be reversibly trapped in a deactivation/activation process according to Fig. 2.6A, or they can be involved in a “reversible transfer” degenerative exchange process, according to Fig. 2.6B.

2.3

Nitroxide-mediated radical polymerizations

With respect to π-conjugated polymers, the best results were obtained by using the NMRP-mediated approach. It is based on the reversible capture of the propagating species by nitroxides and formation of alkoxyamines, which are the dormant species in this process (Fig. 2.7). The nitroxide radical is designed to act as a persistent radical, so that instead of reacting with another nitroxide radical, it cross-couples (reversibly) with the growing species and generates the dormant species (Fig. 2.7). Maji et al. (2010) reported the synthesis of polyfluorene P56 via the nitroxidemediated oxidative homopolymerization (Scheme 2.23). Optimal conditions involved the use of the 7,70 -dimagnesated fluorene monomer 65 generated in situ from 7,70 dibromo-9,9,90 ,90 -tetraoctyl[2,20 ]bifluorenyl 64 and butyllithium (Li-Br exchange reaction), followed by an MgBr2 Et2O addition (Li-Mg transmetallation). Further, 2,2,6,6-tetramethyl-piperidine-1-oxyl radical (TEMPO, Fig. 2.7) (2.4 eq.) was employed as an oxidant, and ethyl ether as the reaction medium (reflux temperature, 5 h). The polymer P56 was obtained in a high yield (94%) and with a reasonable molecular weight (Mn ¼ 9.7 kDa) and PDI ¼ 2.11. Notably, the hydroxylamine obtained as the TEMPO reduction product could be readily reoxidized to TEMPO by treatment with air (Maji et al., 2008), thereby making this method economically attractive.

Key trends in sustainable approaches to the synthesis of semiconducting polymers

81

Fig. 2.7 The general mechanism of NMRP.

Scheme 2.23 Synthesis of the polyfluorene P56 by NMRP.

Optical absorption and emission data were congruent with those previously reported in the literature for the regioregular polyfluorene, thereby demonstrating the site selectivity in the formation of P56. Interestingly, metal trace analysis carried out for Ni, Mn, Cu, Pd, and Co revealed that polymer P56, prepared by the TEMPO method, contained transition metal impurities in the range of 0.02–16 ppm, owing to the Mg used to prepare the Grignard monomer (65). On the other hand, polymer P56, prepared via a standard Ni-catalyzed GRIM procedure, was found to contain transition metals in larger amounts (0.03–1200 ppm). In a series of papers, Thelakkat and coworkers (Lang et al., 2010, 2013; Muth et al., 2011) investigated the preparation of semiconducting perylene-bisimide (PB) polymers containing acrylates by direct NMRP as well as by a combination of this latter with click polymerization (as previously discussed) (Bakbak et al., 2006). For instance, the authors recently reported on the synthesis of two novel liquid-crystalline semiconductor polymers—that is, poly(perylene diester benzimidazole acrylate) P57 and poly(perylene diester benzimidazole acrylate) P58 (Scheme 2.24), via the nitroxide-mediated radical homopolymerization. These reactions were carried out in the chlorinated (toxic) solvent 1,2,4-trichlorobenzene (about 120°C, 50 h) with bis(2-ethylhexyl)perylene-3,4-(4-undecyloxyacrylic acid-1,2-benzimidazole)9,10-dicarboxylate (66; 0.8-M concentration) or N-(dodecylacrylic acid)bis(2ethylexyl)perylene-3,4,9,10-tetracarboxyl-monoimide (67; 1.8-M concentration), respectively, as the monomer, and employing 2,2,5-trimethyl-3-(1-phenylethoxy)-

82

Handbook of Organic Materials for Electronic and Photonic Devices

Scheme 2.24 Synthesis of the perylene diester polymers P57 and P58 by NMRP.

4-phenyl-3-azahexane as the unimolecular initiator. Note also that these reactions were carried out in the presence of 0.1 eq. of additional free N-tert-butyl-αisopropyl-α-phenylnitroxide (TIPNO; 69), to shift the equilibrium to the dormant species (68) ([monomer]:[68] ¼ 100:1) and enhance control over the polymerization. These conditions led to polymers P57 and P58 in moderate yields (about 35%), with Mn ¼ 9.4 kDa (PDI ¼ 1.4) and 20.4 kDa (PDI ¼ 1.7), respectively. The authors indicated that because a controlled radical polymerization method was used, the molecular weight distribution was rather broad, indicating that termination reactions occurred. The higher molecular weight achieved for P58 compared to P57 was attributed to a better solubility of the former in 1,2,4-trichlorobenzene. Differential scanning calorimetry, polarized optical microscopy, and X-ray diffraction measurements of P57 and P58 demonstrated a liquid-crystalline order over a broad range of temperatures. The UV-vis measurements in chlorobenzene solution revealed that the presence of the benzimidazole unit in P57 led to a significantly extended absorption in the visible range up to 670 nm compared to P58 (580 nm). Optical band gaps of 1.86 eV for P57 and 2.16 eV for P58 were determined from the absorption edges of these absorption spectra in chlorobenzene solution. A spectral broadening for both polymers in thin-films compared to the solution spectra was observed, which indicated a change in the aggregation in the solid state. Next, HOMO/LUMO energy levels for P57 were 5.68 and 3.52 eV, respectively, whereas a HOMO of 5.39 eV and a LUMO of 3.53 eV were reported for P58, based

Key trends in sustainable approaches to the synthesis of semiconducting polymers

83

on cyclic voltammetry and optical absorption data. Finally, space-charge-limited current (SCLC) measurements revealed that P57 is a n-type semiconductor with an electron mobility as high as 3.2  104 cm2/Vs, whereas p-type behavior was observed for P58, which exhibited a hole mobility of 1.5  104 cm2/Vs.

2.4

Perspective

From this overview, it is clear that over the last 5 years, impressive advances have been made in the development of methodologies toward the poly(arylene) green synthesis. Several insights have been gained into the understanding of both the problematic aspects associated with conventional synthetic methodologies and the implications in the performance of electronic devices when employing alternative reaction conditions, solvents, reagents, and catalysts. More specifically, the direct arylation polymerization scheme has proved to enable the preparation of structurally diverse poly(arylene)s exhibiting physical properties and performances that are comparable to or better than those for the analogs achieved through conventional routes. For this approach, further research is needed to face the selectivity and regioregularity issues when using monomers lacking directing groups. In this regard, focusing on the design of proper catalytic systems seems to be a promising strategy. For all the surveyed methods, future directions should focus on further expanding the substrate scope, thereby enabling access to the most efficient polymers for organic electronics, optimizing the reaction conditions to gain a precise control over the polymerization reactions, and, obviously, simplifying the reaction protocols in terms of eliminating expensive or toxic reagents, as well as minimizing waste. Particular efforts should be directed toward the replacement of traditional solvents that are not compliant with relevant legislation, ranking high on the list of harmful chemicals. A focus on evaluating the costs of sustainable synthetic processes in comparison with their conventional counterparts in terms of waste disposal, labor costs, and energy consumption, as well as on the development of new synthetic methodologies, is also essential.

References Abdo, N.I., Ku, J., El-Shehawy, A.A., Shim, H.-S., Min, J.-K., El-Barbary, A.A., Jang, Y.H., Lee, J.-S., 2013. Synthesis and characterization of low bandgap π-conjugated copolymers incorporating 4,7-bis(3,30 /4,40 -hexylthiophene-2-yl)benzo[c][2,1,3]thiadiazole units for photovoltaic application. J. Mater. Chem. A 1, 10306–10317. Ackermann, L., Lygin, A.V., 2011. Ruthenium-catalyzed direct C–H bond arylations of heteroarenes. Org. Lett. 13, 3332–3335. Alonzi, M., Lanari, D., Marrocchi, A., Petrucci, C., Vaccaro, L., 2013. Synthesis of polymeric semiconductors by a surface-initiated approach. RSC Adv. 3, 23909–23923. Anastas, P.T., Warner, J.C., 1998. Green Chemistry: Theory and Practice. Oxford University Press, New York. Anderson, N.G., 2012. Practical Process Research and Development. A Guide for Organic Chemists. Academic Press, Waltham, MA.

84

Handbook of Organic Materials for Electronic and Photonic Devices

Bakbak, S., Leech, P.J., Carson, B.E., Saxena, S., King, W.P., Bunz, U.H.F., 2006. 1,3-Dipolar cycloaddition for the generation of nanostructured semiconductors by heated probe tips. Macromolecules 39, 6793–6795. Barner-Kowollik, C., Du Prez, F.E., Espeel, P., Hawker, C.W., Junkers, T., Schlaad, H., Van Camp, W., 2011. “Clicking” Polymers or just efficient linking: what is the difference? Angew. Chem. Int. Ed. 50, 60–62. Bartollini, E., Seri, M., Tortorella, S., Facchetti, A., Marks, T.J., Marrocchi, A., Vaccaro, L., 2013. Sustainable synthetic approach to π-conjugated arylacetylenic semiconductors for bulk heterojunction solar cells. RSC Adv. 3, 9288–9295. Beaupre, S., Pron, A., Drouin, S.H., Najari, A., Mercier, L.G., Robitaille, A., Leclerc, M., 2012. Thieno-, furo-, and selenopheno[3,4-c]pyrrole-4,6-dione copolymers: effect of the heteroatom on the electrooptical properties. Macromolecules 45, 6906–6914. Berrouard, P., Najari, A., Pron, A., Gendron, D., Morin, P.-O., Pouliot, J.-R., Veilleux, J., Leclerc, M., 2012. Synthesis of 5-alkyl[3,4-c]thienopyrrole-4,6-dione-based polymers by direct heteroarylation. Angew. Chem. Int. Ed. 51, 2068–2071. Bohra, H., Wang, M., 2017. Direct C–H arylation: a “Greener” approach towards facile synthesis of organic semiconducting molecules and polymers. J. Mater. Chem. A 5, 11550–11571. Brouwer, F., Alma, J., Valkenier, H., Voortman, T.P., Hillebrand, J.T., Chiechi, R.C., Hummelen, J.C., 2011. Using bis(pinacolato)diboron to improve the quality of regioregular conjugated co-polymers. J. Mater. Chem. 21, 1582–1592. Burke, D.J., Lipomi, D.J., 2013. Green chemistry for organic solar cells. Energy Environ. Sci. 6, 2053–2066. Carsten, B., He, F., Son, H.-J., Xu, T., Yu, L., 2011. Stille polycondensation for synthesis of functional materials. Chem. Rev. 111, 1493–1528. Chang, S.-W., Waters, H., Kettle, J., Kuo, Z.-R., Li, C.-H., Yu, C.-Y., Horie, M., 2012. Pdcatalysed direct arylation polymerisation for synthesis of low-bandgap conjugated polymers and photovoltaic performance. Macromol. Rapid Commun. 33, 1927–1932. Chen, J., Law, C.C.W., Lam, J.W.Y., Dong, Y., Lo, S.M.F., Williams, I.D., Zhu, D., Tang, B.Z., 2003. Synthesis, light emission, nanoaggregation, and restricted intramolecular rotation of 1,1-substituted 2,3,4,5-tetraphenylsiloles. Chem. Mater. 15, 1535–1546. Chen, X., Braunschweig, A.B., Wiester, M.J., Yeganeh, S., Ratner, M.A., Mirkin, C.A., 2009. Spectroscopic tracking of molecular transport junctions generated by using click chemistry. Angew. Chem. Int. Ed. 48, 5178–5181. Cheng, C.-H., Wu, C.-A., Wu, F.-I.Y., Shih, C.-H., 2011. m-Terphenyl Compound Derivatives and Application for Organic Light Emitting Diode. US8475939B2. De Miguel, G., Wielopolski, M., Schuster, D.I., Fazio, M.A., Lee, O.P., Haley, C.K., Ortiz, A.L., Echegoyen, L., Clark, T., Guldi, D.M., 2011. Triazole bridges as versatile linkers in electron donor–acceptor conjugates. J. Am. Chem. Soc. 133, 13036–13054. D€ ohler, D., Michael, P., Binder, W.H., 2017. CuAAC-based click chemistry in self-healing polymers. Acc. Chem. Res. 50, 2610–2620. Dou, L., Liu, Y., Hong, Z., Li, G., Yang, Y., 2015. Low-bandgap near-IR conjugated polymers/ molecules for organic electronics. Chem. Rev. 115, 12633–12665. Dudnik, A.S., Aldrich, T.J., Eastham, N.D., Chang, R.P., Facchetti, A., Marks, T.J., 2016. Tinfree direct C–H arylation polymerization for high photovoltaic efficiency conjugated copolymers. J. Am. Chem. Soc. 138, 15699–15709. ECHA, EU REACH Regulation (EC) No 1907/2006. https://echa.europa.eu/regulations/reach/ understanding-reach. Accessed February 2018. EPA, https://www.epa.gov/greenchemistry. Accessed February 2018.

Key trends in sustainable approaches to the synthesis of semiconducting polymers

85

Facchetti, A., Vaccaro, L., Marrocchi, A., 2012. Semiconducting polymers prepared by direct arylation polycondensation. Angew. Chem. Int. Ed. 51, 3520–3523. Fukuzumi, S., Ohkubo, K., Suenobu, T., 2014. Long-lived charge separation and applications in artificial photosynthesis. Acc. Chem. Res. 47, 1455–1464. Gehringer, M., Laufer, S.A., 2017. Click chemistry: novel applications in cell biology and drug discovery. Angew. Chem. Int. Ed. 49, 15504–15505. Gozzi, C., Lavenot, L., Ilg, K., Penalva, V., Lemaire, M., 1997. Direct thiophene arylation catalysed by palladium. Tetrahedron Lett. 38, 8867–8870. Grenier, F., Berrouard, P., Pouliot, J.-R., Tseng, H.-R., Heeger, A.J., Leclerc, M., 2013. Synthesis of new n-type isoindigo copolymers. Polym. Chem. 4, 1836–1841. Grenier, F., Goudreau, K., Leclerc, M., 2017. Robust direct (hetero)arylation polymerization in biphasic conditions. J. Am. Chem. Soc. 139, 2816–2824. G€ unes, S., Neugebauer, H.S., Sariciftci, N.S., 2007. Conjugated polymer-based organic solar cells. Chem. Rev. 107, 1324–1338. Guo, X., Ortiz, R.P., Zheng, Y., Kim, M.G., Zhang, S., Hu, Y., Lu, G., Facchetti, A., Marks, T.J., 2011. Thieno[3,4-c]pyrrole-4,6-dione-based polymer semiconductors: toward highperformance, air-stable organic thin-film transistors. J. Am. Chem. Soc. 133, 13685. Guo, X., Baumgarten, M., Muellen, K., 2013. Designing π-conjugated polymers for organic electronics. Prog. Polym. Sci. 38, 1832–1908. Hayashi, S., Kojima, Y., Koizumi, T., 2015. Highly regioselective Pd/C-catalyzed direct arylation toward thiophene-based π-conjugated polymers. Polym. Chem. 6, 881–885. Hayashi, S., Takigami, A., Koizumi, T., 2016. Palladium immobilized on thiol-modified silica gel for effective direct C H arylation. ChemPlusChem 81, 930–934. He, C.-Y., Fan, S., Zhang, X., 2010. Pd-catalyzed oxidative cross-coupling of perfluoroarenes with aromatic heterocycles. J. Am. Chem. Soc. 132, 12850–12852. Henderson, R.H., Jimenez-Gonzalez, C., Constable, D.J.C., Alston, S.R., Inglis, G.G.A., Fisher, G., Sherwood, J., Binks, S.P., Curzons, A.D., 2011. Expanding GSK’s solvent selection guide–embedding sustainability into solvent selection starting at medicinal chemistry. Green Chem. 13, 854–862. Horie, M., Majewski, L.A., Fearn, M.J., Yu, C.Y., Luo, Y., Song, A.M., Saunders, B.R., Turner, M.L., 2010. Cyclopentadithiophene based polymers—a comparison of optical, electrochemical and organic field-effect transistor characteristics. J. Mater. Chem. 20, 4347–4355. Huang, H., Yang, L., Facchetti, A., Marks, T.J., 2017. Organic and polymeric semiconductors enhanced by noncovalent conformational locks. Chem. Rev. 117, 10291–10318. Hwang, E., Lusker, K.L., Garno, J.C., Losovyj, Y., Nesterov, E.E., 2011. Semiconducting polymer thin films by surface-confined stepwise click polymerization. Chem. Commun. 11990–11992. Jarowski, P.D., Wu, Y.-L., Schweizer, B., Diederich, F., 2008. 1,2,3-Triazoles as conjugative π-linkers in push pull chromophores: importance of substituent positioning on intramolecular charge-transfer. Org. Lett. 10, 3347–3350. Jo, J., Pron, A., Berrouard, P., Leong, W.L., Yuen, J.D., Moon, J.S., Leclerc, M., Heeger, A.J., 2012. Enhanced power conversion efficiency of low band-gap polymer solar cells by insertion of optimized binary processing additives. Adv. Energy Mater. 2, 1397–1403. Kameshima, H., Nemoto, N., Endo, T., 2001. Synthesis and properties of fluorene-based fluorinated polymers. J. Polym. Sci. A 39, 3143–3150. Karim, M.A., Cho, Y.-R., Park, J.S., Kim, S.C., Lee, J.W., Gal, Y.-S., Jin, S.-H., 2008. Novel fluorene-based functional ‘click polymers’ for quasi-solid-state dye-sensitized solar cells. Chem. Commun. 1929–1931.

86

Handbook of Organic Materials for Electronic and Photonic Devices

Kolb, H.C., Finn, M.G., Sharpless, K.B., 2001. Click chemistry: diverse chemical function from a few good reactions. Angew. Chem. Int. Ed. 40, 2004–2021. Koppe, M., Scharber, M., Brabec, C., Duffy, W., Heeney, M., McCulloch, I., 2007. Polyterthiophenes as donors for polymer solar cells. Adv. Funct. Mater. 26, 1371–1376. Kowalski, S., Allard, S., Scherf, U., 2012. Synthesis of poly(4,4-dialkyl-cyclopenta[2,1-b:3,4b0 ]dithiophene-alt-2,1,3-benzothiadiazole) (PCPDTBT) in a direct arylation scheme. ACS Macro Lett. 1, 465–468. Kumari, P., Khawas, K., Hazra, S., Kuila, B.K., 2016. Poly(3-hexyl thiophene)-b-poly(Nisopropyl acrylamide): synthesis and its composition dependent structural, solubility, thermoresponsive, electrochemical, and electronic properties. J. Polym. Sci. A 54, 1785–1794. Kuwabara, J., Nohara, Y., Choi, S.J., Fujinami, Y., Lu, W., Yoshimura, K., Oguma, J., Suenobu, K., Kanbara, T., 2013. Direct arylation polycondensation for the synthesis of bithiophene-based alternating copolymers. Polym. Chem. 4, 947–953. Kuwabara, J., Yasuda, T., Choi, S.J., Lu, W., Yamazaki, K., Kagaya, S., Han, L., Kanbara, T., 2014. Direct arylation polycondensation: a promising method for the synthesis of highly pure, high-molecular-weight conjugated polymers needed for improving the performance of organic photovoltaics. Adv. Funct. Mater. 34, 3226–3233. Ladomenou, K., Nikolaou, V., Charalambidis, G., Coutsolelos, A.G., 2016. “Click”-reaction: an alternative tool for new architectures of porphyrin based derivatives. Coord. Chem. Rev. 306, 1–42, and references therein. Lang, A.S., Muth, M.A., Heinrich, C.D., Carrasco-Orozco, M., Thelakkat, M., 2013. Pendant perylene polymers with high electron mobility. J. Pol. Sci. B 51, 1480–1486. Lee, J.K., Gwinner, M.C., Berger, R., Newby, C., Zentel, R., Friend, R.H., Sirringhaus, H., Ober, C.K., 2011. High-performance electron-transporting polymers derived from a heteroaryl bis(trifluoroborate). J. Am. Chem. Soc. 133, 9949–9951. Lee, E.K., Lee, M.Y., Park, C.H., Lee, H.R., Oh, J.H., 2017. Toward environmentally robust organic electronics: approaches and applications. Adv. Mater. 29, 1703638. Li, C., Liu, M., Pschirer, N.G., Baumgarten, M., M€ullen, K., 2010. Polyphenylene-based materials for organic photovoltaics. Chem. Rev. 110, 6817–6855. Li, W., Hendriks, K.H., Wienk, M.M., Janssen, R.A.J., 2016. Diketopyrrolopyrrole polymers for organic solar cells. Acc. Chem. Res. 49, 78–85. Liu, S.-Y., Li, H.-Y., Shi, M.-M., Jiang, H., Hu, X.-L., Li, W.-Q., Fu, L., Chen, H.-Z., 2012. Pd/C as a clean and effective heterogeneous catalyst for C–C couplings toward highly pure semiconducting polymers. Macromolecules 45, 9004–9009. Liu, W., Tkachov, R., Komber, H., Senkovskyy, V., Schubert, M., Wei, Z., Facchetti, A., Neher, D., Kiriy, A., 2014. Chain-growth polycondensation of perylene diimide-based copolymers: a new route to regio-regular perylene diimide-based acceptors for all-polymer solar cells and n-type transistors. Polym. Chem. 5, 3404–3411. Lu, W., Kuwabara, J., Kanbara, T., 2011. Polycondensation of dibromofluorene analogues with tetrafluorobenzene via direct arylation. Macromolecules 44, 1252–1255. Lu, W., Kuwabara, J., Iijima, T., Higashimura, H., Hayashi, H., Kanbara, T., 2012. Synthesis of π-conjugated polymers containing fluorinated arylene units via direct arylation: efficient synthetic method of materials for OLEDs. Macromolecules 45, 4128–4133. Lu, W., Kuwabara, J., Kanbara, T., 2013. Synthesis of π-conjugated polymer consisting of pyrrole and fluorene units by Ru-catalyzed site-selective direct arylation polycondensation. Macromol. Rapid Commun. 34, 1151–1156. Maji, M.S., Pfeifer, T., Studer, A., 2008. Oxidative homocoupling of aryl, alkenyl, and alkynyl Grignard reagents with TEMPO and dioxygen. Angew. Chem. Int. Ed. 47, 9547–9550.

Key trends in sustainable approaches to the synthesis of semiconducting polymers

87

Maji, M.S., Pfeifer, T.T., Studer, A., 2010. Transition-metal-free synthesis of conjugated polymers from bis-Grignard reagents by using TEMPO as oxidant. Chem. Eur. J. 16, 5872–5875. Marrocchi, A., Lanari, D., Facchetti, A., Vaccaro, L., 2012. Poly(3-hexylthiophene): synthetic methodologies and properties in bulk heterojunction solar cells. Energy Environ. Sci. 5, 8457–8474. Marrocchi, A., Facchetti, A., Lanari, D., Petrucci, C., Vaccaro, L., 2016a. Current methodologies for a sustainable approach to π-conjugated organic semiconductors. Energy Environ. Sci. 9, 763–786. Marrocchi, A., Facchetti, A., Lanari, D., Santoro, A., Vaccaro, L., 2016b. Click-chemistry approaches to π-conjugated polymers for organic electronics applications. Chem. Sci. 7, 6298–6308. Mercier, L.G., Leclerc, M., 2013. Direct (hetero)arylation: a new tool for polymer chemists. Acc. Chem. Res. 46, 1597–1605. Mercier, L.G., Aı¨ch, B.R., Najari, A., Beaupre, S., Berrouard, P., Pron, A., Robitaille, A., Tao, Y., Leclerc, M., 2013. Direct heteroarylation of β-protected dithienosilole and dithienogermole monomers with thieno[3,4-c]pyrrole-4,6-dione and furo[3,4-c]pyrrole-4,6-dione. Polym. Chem. 4, 5252–5260. Mirkin, C.A., Chen, X., Braunschweig, A.B., Wiester, M.J., Xu, X., Daniel, W.L., 2011. Click Chemistry, Molecular Transport Junction and Colorimetric Detection of Copper. US2011/0033940A1. Mohebbi, A.R., Yuen, J., Fan, J., Munoz, C., Wang, M.F., Shirazi, R.S., Seifter, J., Wudl, F., 2011. Emeraldicene as an acceptor moiety: balanced-mobility, ambipolar, organic thinfilm transistors. Adv. Mater. 23, 4644–4648. Molander, G.A., Ellis, N., 2007. Organotrifluoroborates: protected boronic acids that expand the versatility of the suzuki coupling reaction. Acc. Chem. Res. 40, 275–286. Murarka, S., Wertz, S., Studer, A., 2012. Transition-metal-free oxidative coupling reactions for the formation of C–C and C–N bonds mediated by TEMPO and its derivatives. Chimia 66, 413–417. Nehls, B.S., Fuldner, S., Preis, E., Farrell, T., Sherf, U., 2005. Microwave-assisted synthesis of 1,5- and 2,6-linked naphthylene-based ladder polymers. Macromolecules 38, 687–694. Nehls, B.J., Galbrecht, F., Brauer, D.J., Lehmann, C.W., Scherf, U., Farrell, T., 2006. Synthesis and characterization of a helical step-ladder polyarylene. Polym. Chem. 44, 5533–5545. Osedach, T.P., Andrew, T.L., Bulovic, V., 2013. Effect of synthetic accessibility on the commercial viability of organic photovoltaics. Energy Environ. Sci. 6, 711–718. Pei, Q., Yang, Y., 1996. Efficient photoluminescence and electroluminescence from a soluble polyfluorene. J. Am. Chem. Soc. 118, 7416–7417. Pouliot, J.R., Mercier, L.G., Caron, S., Leclerc, M., 2013. Accessing new DPP-based copolymers by direct heteroarylation polymerization. Macromol. Chem. Phys. 214, 453–457. Pouliot, J.-R., Grenier, F., Blaskovitz, J.T., Beaupre, S., Leclerc, M., 2016. Direct (hetero) arylation polymerization: simplicity for conjugated polymer synthesis. Chem. Rev. 116, 14225–14274. Rostovtsev, V.V., Green, L.G., Fokin, V.V., Sharpless, K.B., 2002. A stepwise Huisgen cycloaddition process: copper(I)-catalyzed regioselective “ligation” of azides and terminal alkynes. Angew. Chem. Int. Ed. 41, 2596–2599. Rudenko, A.E., Wiley, C.A., Stone, S.M., Tannaci, J.F., Thompson, B.C., 2012. Semi-random P3HT analogs via direct arylation polymerization. J. Polym. Sci. A 50, 3691–3697. Saudari, S.R., Lin, Y.J., Lai, Y., Kagan, C.R., 2010. Device configurations for ambipolar transport in flexible, pentacene transistors. Adv. Mater. 22, 5063–5068.

88

Handbook of Organic Materials for Electronic and Photonic Devices

Schmidt, G.C., Hoft, D., Haase, K., Hubler, A.C., Karpov, E., Tkakov, R., Stamm, M., Kiriy, A., Haidu, F., Zahn, D.R.T., Yan, H., Facchetti, A., 2014. Naphtalenediimide-based donor– acceptor copolymer prepared by chain-growth catalyst-transfer polycondensation: evaluation of electron-transporting properties and application in printed polymer transistors. J. Mater. Chem. C 2, 5149–5154. Senkovskyy, V., Tchachov, R., Konmber, H., John, A., Sommer, J.-U., Kiriy, A., 2012. Mechanistic insight into catalyst-transfer polymerization of unusual anion-radical naphthalene diimide monomers: an observation of Ni(0) intermediates. Macromolecules 45, 7770–7777. Sevignon, M., Papillon, J., Schulz, E., Lemaire, M., 1999. New synthetic method for the polymerization of alkylthiophenes. Tetrahedron Lett. 40, 5873–5876. Singh, M.S., Chowdhury, S., Koley, S., 2016. Advances of azide-alkyne cycloaddition-click chemistry over the recent decade. Tetrahedron 72, 5257–5283. Strappaveccia, G., Ismalaj, E., Petrucci, C., Lanari, D., Marrocchi, A., Drees, M., Facchetti, A., Vaccaro, L., 2015a. A biomass-derived safe medium to replace toxic dipolar solvents and access cleaner Heck coupling reactions. Green Chem. 17, 365–372. Strappaveccia, G., Luciani, L., Bartollini, E., Marrocchi, A., Pizzo, F., Vaccaro, L., 2015b. γ-Valerolactone as an alternative biomass-derived medium for the Sonogashira reaction. Green Chem. 17, 1071–1076. Suzuki, A., 1999. Recent advances in the cross-coupling reactions of organoboron derivatives with organic electrophiles, 1995–1998. J. Organomet. Chem. 576, 147–168. Thirunavukkarasu, V.S., Parthasarathy, K., Cheng, C.-H., 2010. One-pot synthesis of diarylmethylidenefluorenes and phenanthrenes by palladium-catalyzed multiple C-H bond functionalization. Chem. Eur. J. 16, 1436–1440. Tkachov, R., Karpov, Y., Senkovskyy, V., Raguzin, I., Zessin, J., Lederer, A., Stamm, M., Voit, B., Beryozkina, T., Bakulev, V., Zhao, W., Facchetti, A., Kiriy, A., 2014. Efficient tin-free route to a donor–acceptor semiconducting copolymer with variable molecular weights. Macromolecules 47, 3845–3851. Tornøe, C.W., Christensen, C., Meldal, M., 2002. Peptidotriazoles on solid phase: [1,2,3]triazoles by regiospecific copper(I)-catalyzed 1,3-dipolar cycloadditions of terminal alkynes to azides. J. Organomet. Chem. 67, 3057–3064. Tsami, A., Yang, X.-H., Farrell, T., Neher, D., Holder, H., 2008. Alternating fluorene-di(thiophene)quinoxaline copolymers via microwave-supported Suzuki cross-coupling reactions. Polym. Chem. 46, 7794–7808. van Steenis, D.J.V.C., David, O.R.P., van Strikdonck, G.P.F., van Maarseveen, J.H., Reek, J.N.H., 2005. Click-chemistry as an efficient synthetic tool for the preparation of novel conjugated polymers. Chem. Commun. 0, 4333–4335. Wakioka, M., Kitano, Y., Ozawa, F., 2013. A highly efficient catalytic system for polycondensation of 2,7-dibromo-9,9-dioctylfluorene and 1,2,4,5-tetrafluorobenzene via direct arylation. Macromolecules 46, 370–374. Wakioka, M., Ichihara, N., Kitano, Y., Ozawa, F., 2014. A highly efficient catalyst for the synthesis of alternating copolymers with thieno[3,4-c]pyrrole-4,6-dione units via direct arylation polymerization. Macromolecules 47, 626–631. Wakioka, M., Nakamura, Y., Montgomery, M., Ozawa, F., 2015. Remarkable ligand effect of P (2-MeOC6H4)3 on palladium-catalyzed direct arylation. Organometallics 34, 198–205. Wang, Q., Takita, R., Kikuzaki, Y., Ozawa, F., 2010. Palladium-catalyzed dehydrohalogenative polycondensation of 2-bromo-3-hexylthiophene: an efficient approach to head-to-tail poly (3-hexylthiophene). J. Am. Chem. Soc. 132, 11420–11421.

Key trends in sustainable approaches to the synthesis of semiconducting polymers

89

Wang, D.H., Pron, A., Leclerc, M., Heeger, A.J., 2013. Additive-free bulk-heterojuction solar cells with enhanced power conversion efficiency, comprising a newly designed selenophene-thienopyrrolodione copolymer. Adv. Funct. Mater. 23, 1297–1304. Yokozawa, T., Nanashima, Y., Ohta, Y., 2012. Precision synthesis of n-Type π-conjugated polymers in catalyst-transfer condensation polymerization. ACS Macro Lett. 1, 862–866. Yu, G., Gao, J., Hummelen, J.C., Wudl, F., Heeger, A.J., 1995. Polymer photovoltaic cells: enhanced efficiencies via a network of internal donor-acceptor heterojunctions. Science 270, 1789–1791. Yuan, M.-C., Chiu, M.-Y., Liu, S.P., Chen, C.-M., Wei, K.-H., 2010. A thieno[3,4-c]pyrrole-4,6-dione-based donor  acceptor polymer exhibiting high crystallinity for photovoltaic applications. Macromolecules 43, 6936–6938. Yuen, A.K.L., Hutton, C.A., 2005. Deprotection of pinacolyl boronate esters via hydrolysis of intermediate potassium trifluoroborates. Tetrahedron Lett. 46, 7899–7903. Zhang, W., Wang, Z., Ma, Y., 2013. Microwave-assisted Suzuki coupling reaction for rapid synthesis of conjugated polymers: poly(9,9-dihexylfluorene)s as an example. J. Polym. Sci. A 51, 1950–1955. Zhu, Z., Waller, D., Gaudiana, R., Morana, M., M€uhlbacher, D., Scharber, M., Brabec, C., 2007. Panchromatic conjugated polymers containing alternating donor/acceptor units for photovoltaic applications. Macromolecules 40, 1981–1986.

Further reading Lang, A.S., Neubig, A., Sommer, M., Thelakkat, M., 2010. NMRP versus “Click” chemistry for the synthesis of semiconductor polymers carrying pendant perylene bisimides. Macromolecules 43, 7001–7010. Muth, M.-A., Carrasco-Orozco, M., Thelakkat, M., 2011. Liquid-crystalline perylene diester polymers with tunable charge-carrier mobility. Adv. Funct. Mater. 21, 4510–4518.

This page intentionally left blank

Functional blends of organic materials for optoelectronic applications

3

Valerii Sharapov Department of Chemistry, The University of Chicago, Chicago, IL, United States

3.1

Introduction

Every year, hundreds of new materials are discovered by researchers worldwide. The purpose of these discoveries is identifying materials that possess new exciting properties or that have known properties, but those properties have been greatly improved. These properties may be mechanical, electrical, optical, magnetic, or a combination of those. Optoelectronic materials, as follows from the name, are employed in devices where light is used to affect electronic processes in materials (or vice versa) and thus provide a useful functionality (Ostroverkhova, 2016). Devices that belong to this category include solar cells, light-emitting diodes (LED), light-emitting transistors (LETs), and photodetectors. Often, however, not only one material is used in these devices, but quite a few. They may form sandwichlike structures, be incorporated in a blend with other materials, or a combination of both. When multiple materials are combined into a single device, many potential sources of failure may arise from the mutual incompatibility of device components or instability under device fabrication/operation conditions. Therefore, proper identification of potential materials for high-performing devices has great importance. A mere trial-and-error approach will be highly inefficient, and thus knowledge of fundamental principles governing the interaction of components in a blend is crucial to making a rational choice of materials. This chapter describes blends of organic materials that are used in modern photovoltaic (PV) devices. These blends may consist of organic polymers, small molecules, or a combination of both. First, we consider fundamental thermodynamic and kinetic effects that govern film formation and how these principles may be used to form blends with the required specifications. Next, we give a brief overview of organic PV devices and major processes that facilitate photon-to-electron conversion. Finally, we summarize some of the best-performing PV blends, with a particular emphasis on structure-property relationships that may help to deduce design rules for future generations of high-performing PV materials.

Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00003-6 © 2019 Elsevier Ltd. All rights reserved.

92

3.2

Handbook of Organic Materials for Electronic and Photonic Devices

Thermodynamics and kinetics of polymer mixing

Many devices that have optoelectronic applications contain several components blended into a single film. These components may include electron donor and electron acceptor systems, charge-conducting materials, morphology modifiers (plasticizers), inert polymers serving as a matrix with active material being dispersed in it, and so on. Despite the various requirements for each of these blends, the principles governing film formation are the same. In the end, what determines a final state of a blend is the interplay between the thermodynamic parameters that characterize the blend components and the kinetic details of film formation. From the perspective of thermodynamics, the formation of any blend will be governed by its free energy, which in turn has energetic ΔUmix and entropic ΔSmix terms: ΔFmix ¼ ΔUmix  TΔSmix Because ΔFmix is a thermodynamic potential, a negative value for it indicates favored mixing, which in turn depends on signs of energetic and entropic terms. The entropic component of mixing is always positive, thus promoting mixing; however, its exact value depends on the nature of the mixing species (namely, whether these components are small molecules or polymers). When two small molecules mix (e.g., a solvent and a small-molecule solute), the number of their possible permutations in space is very large and depends on the volume available for mixing. At the same time, when small molecules are connected in a polymer chain, this imposes geometric restrictions on their movement and the number of possible permutations decreases, thus decreasing an entropic component of mixing, but still keeping it positive. Therefore, an energetic term of free energy becomes a parameter that determines whether mixing is going to occur. For multicomponent polymer systems, an energetic term is often conveniently described by Flory-Huggins theory (Huggins, 1941; Flory, 1941), which defines a polymer as a system of cells that occupy a cell space together with a solvent, and interactions within this system are described using the statistical approach. Even though it was developed in order to explain polymer solutions, the theory also may be used to describe polymer blends. According to Flory-Huggins, the total change in free energy ΔFmix upon the mixing of two components A and B may be expressed as follows: ΔFmix ¼ ϕA lnϕA + ϕB ln ϕB + χϕA ϕB kB T where ϕ is the volume fraction of component A or B and χ is the Flory-Huggins interaction parameter, which in turn depends on temperature T as χ ¼ C + D/T, where C and D are constants determined for a specific combination of blend components. The first two terms in this equation originate from entropic mixing and are always negative, thus promoting mixing. The third one describes the energy of interactions upon mixing and may be positive (i.e., opposing mixing), negative (i.e., promoting mixing)

Functional blends of organic materials for optoelectronic applications

93

or zero. Fig. 3.1A (Cheng, 2008) shows an example of the free energy versus composition curve for a symmetrical system (molecular size of components A and B being equal) for several temperatures (T5 to T1). It may be noticed that at the highest temperature T5, the curve has a parabolic shape with an energy minimum at ϕ ¼ 0.5, indicating that components A and B form a stable, homogeneous mixture for all component ratios under these conditions. When the temperature is decreased and reaches a critical level TC, a curve reaches a point at which its curvature changes from concave downward to concave upward, and TC is an inflec2 tion point at which ∂∂ϕΔF 2 ¼ zero. This point indicates that the miscibility of the blend component has reached a limit and the mixture phase separates at temperatures lower than TC. The appearance of two minima on the energy curve indicates phase separation, meaning that two stable phases can be formed; and the composition of these phases

Fig. 3.1 (A) The free energy versus composition curve for varying temperatures; (B) simulated morphology of a blend after spinodal decomposition; (C) simulated morphology of a blend after phase separation through nucleation and growth mechanisms. (A) Modified from Cheng, S.Z.D., 2008. Phase Transitions in Polymers: The Role of Metastable States, first ed. Elsevier Science S.A; (B,C) Reproduced with permission from Garcia-Ojalvo, J., Lacasta, A.M., Sancho, J.M., Toral, R. 1998. Phase separation driven by external fluctuations. Europhys. Lett. 42, 125, Copyright 1998 EDP Sciences.

94

Handbook of Organic Materials for Electronic and Photonic Devices

is determined by a concentration corresponding to an energy minimum. Each of the curves corresponding to temperatures T1 to T4 has two important points. The first one is the point at which ∂ΔF ∂ϕ ¼ 0 (marked with the symbol  in Fig. 3.1A), showing an energy minimum. The second one is the point at which

∂2 ΔF ¼ 0, ∂ϕ2

(marked with the

symbol  in Fig. 3.1A), showing an inflection point at which an energy maximum of the curve (i.e., an unstable region) transitions to an energy minimum (i.e., a stable region). If all minima points are connected by a line, they will form a binodal line, whereas connected inflection points form a spinodal line. The area between the spinodal and binodal lines is called the metastable phase. If the system is unstable (located in the areas of the curve between two spinodal lines), then spontaneous phase separation happens upon the slightest composition fluctuation. This spontaneous phase separation is called spinodal decomposition, and a simulated example of it is shown in Fig. 3.1B (Garcia-Ojalvo et al., 1998). At the same time, if the system is in a metastable state, it is more resistant to fluctuations. In this case, phase separation may occur through a mechanism of nucleation and growth if composition fluctuations are large enough to form stable nuclei of a phase larger than a critical size (Fig. 3.1C). Even though the example given here describes a binary system where solvent and solute molecules have equal size, similar conclusions can be made for more complicated systems with multiple different components, such as polymer blends. Most of the blends used for optoelectronic applications must meet specific requirements for phase separation of components to have good performance. For example, neither finely intermixed nor largely phase-separated blends are used for organic solar cells. The bulk heterojunction solar cells with the highest power-conversion efficiency tend to have donor and acceptor components phase-separated to form domains of around 10 nm in size for optimum charge generation and transport (Dou et al., 2013; Lu and Yu, 2014). Domains much larger or smaller than this, or with low phase purity, typically result in increased charge recombination, and thus deteriorated performance (Lyons et al., 2012; Lakhwani et al., 2014). Therefore, a rational choice of components and their amounts is critical to achieve future progress in the field. It was proposed that studying the eutectic behavior of the mixtures may be a useful tool to use for determining the optimum donor-acceptor ratio in a blend (M€uller et al., 2008). When a series of blends with varying ratios of P3HT and PC61BM was made and their DSC thermograms were measured, it was observed that mixtures with 65 wt% of P3HT showed eutectic behavior (Fig. 3.2A). At the same time, device performance was measured for these blends, and the highest current density was found for blends with hypoeutectic composition, corresponding to a slight excess of PC61BM (see the shadowed area in Fig. 3.2). Similar results were observed for other poly-3-alkylthiophenes and PC71BM; however, eutectic composition for these systems was slightly shifted. Authors argue that blends near the eutectic point have optimized phase separation, which results in well-balanced formation of domains of the optimum size for efficient charge percolation and interfaces with a surface area large enough for efficient exciton dissociation. At the same time, slight excess of PCBM is necessary for higher electron mobility. Similar studies have been conducted for P3HT and small molecule

Functional blends of organic materials for optoelectronic applications

95

Fig. 3.2 (A) Differential scanning calorimetry (DSC) heating thermograms (left) and corresponding temperature/composition diagram (right) for a P3HT:PC61BM system. Crosses represent the onset of crystallization. (B) Performance of solar cell devices for systems with an increasing ratio of poly(3-hexylthiophene) (P3HT). Filled circles represent devices thermally annealed at 140°C, open triangles represent devices melt-quenched from 290°C, and open circles represent devices after further annealing at 140°C. Reproduced with permission from M€uller, C., Ferenczi, T.A.M., Campoy-Quiles, M., Frost, J.M., Bradley, D.D.C., Smith, P., Stingelin-Stutzmann, N., Nelson, J., 2008. Binary organic photovoltaic blends: a simple rationale for optimum compositions. Adv. Mater. 2008, 20, 3510–3515. Copyright 2008 John Wiley and Sons.

nonfullerene acceptors. Authors observed that for blends with acceptors YF25 and K12, devices with the maximum performance are also hypoeutectic, similarly to P3HT:PCBM case (Wolfer et al., 2013). Moreover, highly crystalline and highly amorphous nonfullerene acceptors that did not form eutectic mixtures with P3HT had much inferior performance (Stoltzfus et al., 2016). In recent years, a tremendous number of new PV materials have been developed, making the number of their potential combinations overwhelmingly large. Therefore, a problem of rational choice of potential candidates for further optimization and detailed studies within this huge pool of materials is currently of great importance. Recently, Harald Ade and co-workers proposed an approach aimed at resolving this exact issue (Ye et al., 2018). The authors used a modified Flory-Huggins interaction parameter [χ aa(T), a temperature-dependent amorphous-amorphous interaction parameter] to identify the best combinations of donors and acceptors and establish quantitative correlations between material properties, purity of domains, and device performance. They found that pairs of materials for which the values of χ aa(T) were

96

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 3.3 Dependence of FF on interaction parameter χaa(Tm), estimated from DSC for 15 pairs of electron donors and acceptors. Reproduced with permission from Ye, L., Hu, H., Ghasemi, M., Wang, T., Collins, B.A., Kim, J.H., Jiang, K., Carpenter, J.H., Li, H., Li, Z., et al., 2018. Quantitative relations between interaction parameter, miscibility and function in organic solar cells. Nat. Mater. 17, 253–260, Copyright 2018 Springer Nature.

above a threshold value of about 0.72 formed blends with higher phase purity, and thus higher fill factor (FF), in the solar cell device. Moreover, it was shown that a similar interaction parameter measured from DSC χ aa(Tm) may be an excellent tool to rationalize performance in 15 different donor-acceptor systems (Fig. 3.3). Even though thermodynamic properties play a crucial role in predicting compatibility and providing useful tools for making a rational choice of PV materials, the final outcome largely depends on the kinetics of film formation. For example, when PV film is processed from a fast-drying solvent, it is often observed that very large (>100 nm) domains of polymer and 6,6-phenyl-C61-butyric acid methyl ester (PCBM) are formed. However, the addition of small quantities of high-boilingpoint solvents, such as DIO or o-dichlorobenzene, often results in much finer phase separation and superior device performance. Studies on real time film-drying dynamics helped to support this result (van Franeker et al., 2015). Thus, for a low-boiling-point solvent (chloroform), it was observed that fast spinodal decomposition preceding polymer aggregation occurred during film drying. This resulted in the formation of large aggregates of polymer and PCBM phases that deteriorated device performance. However when a high-boiling-point solvent additive (orthodichlorobenzene or DIO) was added, the mechanism was drastically different. Polymer preaggregation induced by a solvent additive suppressed fast liquid-liquid demixing, which in turn prevented the formation of large PCBM aggregates (Fig. 3.4).

Functional blends of organic materials for optoelectronic applications

97

Fig. 3.4 Schematic diagram showing mechanisms of film formation with (right side) and without (left side) the presence of high boiling point solvent additive. Reproduced with permission from van Franeker, J.J., Turbiez, M., Li, W., Wienk, M.M., Janssen, R.A.J., 2015. A real-time study of the benefits of co-solvents in polymer solar cell processing. Nat. Commun. 6, 6229. Copyright 2015 Springer Nature.

3.3

Blends for organic solar cells

3.3.1 General overview of BHJ solar cells PV devices are designed to convert optical power (i.e., sunlight) into electrical power using the PV effect in semiconductors. The efficiency of this conversion is described by a parameter called power conversion efficiency (PCE), which in turn depends on open circuit voltage (VOC), short circuit current density (JSC), and FF. Historically, since the invention of the solar cells, the first material employed was silicon, which is still widely used in monocrystalline and polycrystalline form in modern commercially available devices. Later, other inorganic semiconductors came into play, such as GaAs, PbS, and CdSe. Such a wide variety of available semiconductors is especially important for photodetector devices (PDs), where both narrow and wide coverage in different parts of the spectrum are important (Garcı´a de Arquer et al., 2017; Jansen-van Vuuren et al., 2016; Hussain and Hussain, 2016) Inorganic materials offer huge advantages for PV devices and PDs, such as broad wavelength coverage, chemical stability, and higher crystallinity. At the same time, they have multiple disadvantages, such as expensive purification and doping methods, usage of toxic and/or rare elements (Cd, In, Te), high density and heavier weight.

98

Handbook of Organic Materials for Electronic and Photonic Devices

Therefore, organic materials may offer a viable alternative to inorganic semiconductors, bringing a rich arsenal of organic chemistry to synthesizing and optimizing these compounds, the potential of using green technologies (Chen et al., 2014; Zhao et al., 2016) to manufacture these devices, and the potential to manufacture lightweight and flexible devices through a roll-to-roll process (Diao et al., 2014; Richter et al., 2017). A typical architecture of these devices includes an active layer, where photon-tocharge conversion occurs, sandwiched between a transparent electrode, usually a transparent conductive oxide, such as indium tin oxide (ITO) or fluorine-doped tin oxide (FTO), and a metal electrode (Ag, Al, Au, etc.). Additional electron- or hole-transporting layers are often inserted to improve interfacial charge transfer (Chueh et al., 2015). An active layer consists of two kinds of material—an electron donor and an electron acceptor, which typically absorb in complementary parts of the spectrum, thus covering a wider range of wavelengths, which maximizes device performance. A photon that is absorbed by one of the components in the active layer transitions the molecule to its excited state, thus creating a Frenkel exciton (a localized electron-hole pair that is tightly bound due to a lower dielectric constant of organic materials) (Torabi et al., 2015; Cho et al., 2014; Armin et al., 2017) The exciton diffuses to the donor/acceptor interface and dissociates, taking advantage of the driving force provided by a favorable downward energy cascade between donor and acceptor (Clarke and Durrant, 2010). The exciton dissociation generates free charge carriers that are collected at the device terminals; electrons are collected at the cathode of the device and holes are collected at the anode. Because generation of free charge carriers takes place at the interface between electron donor and acceptor, a concept of bulk heterojunction (BHJ) was introduced, where donors and acceptors are mixed into a single blend, forming a myriad of interfacial connections and thus significantly increasing efficiency of the process compared to planar heterojunction devices, in which only one heterojunction exists (Fig. 3.5; Yu et al., 1995). When such a complex structure as a BHJ blend is formed, a countless number of parameters need to be optimized to maximize the performance of the device. For example, opto-electronic parameters may include materials’ lightabsorption properties, intramolecular polarization, and intramolecular charge carrier mobility. These properties may be controlled through rigorous design of active materials. For instance, the introduction of molecular units with a quinoidal character results in the formation of flat, rigid polymers with narrow bandgap (Liang et al., 2008; Liang and Yu, 2010; Lu et al., 2015a). This in turn forms polymers with a wider light absorption. In addition, when these units are combined to form an intramolecular donor acceptor pattern in a polymer, this provides an additional driving force for charge separation upon excitation (Collado-Fregoso et al., 2015; Rolczynski et al., 2014). Nowadays, the theoretical computer simulation analysis provides useful tools to predict optoelectronic properties of the material and thus use rational optimization techniques in materials design (Liang et al., 2014; Zhugayevych and Tretiak, 2014). The other group of parameters includes those related to the morphology of PV films. Because the scale of interactions includes millions of molecules forming

Functional blends of organic materials for optoelectronic applications

99

e–

e– h+ e–

e– h+

h+

h+

6 5 4

A

A

3 2 1

(A)

(B)

Fig. 3.5 (A) BHJ solar cell device. Layers in the device are marked as follows: a 1—transparent substrate (glass, polymer), 2—transparent conductive layer (ITO, FTO), 3—hole transporting layer, 4—bulk heterojunction active layer consisting of electron donor and acceptor, 5— electron transporting layer, a 6—metal electrode; (B) planar heterojunction device.

domains and interfaces, it is much more challenging to provide rational predictions for these properties. Therefore, a trial-and-error approach is often used. General guidelines include requirements for phase separation, molecular orientation, and phase purity. Precise phase separation is of major importance for these devices because a very fragile balance must be maintained between exciton dissociation and charge transport (Treat and Chabinyc, 2014; Bartelt et al., 2014; Collins et al., 2013; Huang et al., 2014). When excitons are generated in the bulk of the domain, they diffuse to a donor-acceptor interface and dissociate, thus contributing charge carriers to an electrical circuit. However, for a typical organic material, the exciton diffusion distance before recombination is 9–13 nm, depending on the material; therefore, only excitons generated within this depth from the interface effectively contribute to electrical current (Lin et al., 2014). If domains are too big, then a very small fraction of excitons contribute to current, which decreases the efficiency of a device (Hedley et al., 2013). At the same time, charge transport takes place through a percolating network of domains, so when domains are too small, a formed structure has very finely intermixed donor and acceptor contacts, thus providing multiple points for charge recombination during its transport and again decreasing device efficiency (Huang et al., 2014). Therefore, a very fine balance between charge generation and transport is required.

100

Handbook of Organic Materials for Electronic and Photonic Devices

3.3.2 Ternary blend organic solar cells One of the recently emerging fields in organic photovoltaics (OPVs) is ternary blend solar cells (de Zerio and M€ uller, 2018; Li et al., 2017; An et al., 2016; Savoie et al., 2015; Lu et al., 2015b). A key feature of these systems is that in addition to the main blend component (typically a polymer donor and a small molecule acceptor), another photoactive component is added, which often results in enhanced performance of OPV devices. Some of the most common high-efficiency donors and third components are summarized in Figs. 3.6 and 3.7, respectively, and Table 3.1. In order to be a good third component, a material must meet a range of criteria, including complementary light absorption with the blend main components, a cascade of energy levels facilitating charge transport between the phases, and compatible morphology. Other parameters that are especially important for polymer third components are molecular weight and polydispersity (Zhou et al., 2016). Because of large batchto-batch variations in molecular weight and polydispersity, these blends may suffer from irreproducibility of device performance. Therefore, small-molecule third components were developed as a viable alternative. Typical mechanisms involved in PCE enhancement include harvesting additional photons due to complementary absorption of the third component and relay effects due to energy level cascades, which allow additional harvesting of charge carriers and morphology effects. Additional light harvesting is possible when a third component has a light absorption spectrum that is complementary to a spectrum of the blend’s main components and with a high absorption coefficient. Additional light harvesting is usually required in the areas where main components lack absorption—typically an ultraviolet (UV) region and near-infrared (IR). Organic dyes, such as based on porphyrin and squaraine moieties, showed the most promising results recently in near-IR sensitization. The porphyrin-zinc molecules DPPEZnP-O and its modified version, DPPEZnP-TEH, were successfully employed as third components in PTB7:PC71BM-based devices (Xiao et al., 2016; Nian et al., 2016). The role of these additives was twofold—expanding photon harvesting by an additional absorption in the 800–900-nm region and improving charge transport in devices by taking advantage of cascading energy levels. Researchers observed an increase in PCE from 7.47% to 8.39% for devices with DPPEZnP-O as a third component. Even better results were achieved for DPPEZnP-TEH, where a thiophene unit was added to the side chains, thus expanding conjugation—PCE increased from 7.85% to 9.52% after incorporation of the third component and improved even further to 11.03% after a cathode interfacial layer with higher electrical conductivity was inserted. Another example of near-IR sensitization is the incorporation of squaraine dyes. Small quantities of these compounds were used in quaternary blend solar cells in PTB7/PC71BM and PTB7-Th/PC71BM systems to achieve high PCEs. Two sensitizers have been used in PTB polymers:PC71BM blends—ASSQ, absorbing near 550 nm, and DPSQ, absorbing near 720 nm, therefore covering regions where absorption of the main components is weaker (Goh et al., 2016). When blend components formed a quaternary blend with only 2% of each sensitizer present, the authors achieved an increase in PCE from 8.7% to 10.3% for PTB7 systems, originating

Functional blends of organic materials for optoelectronic applications 101

Fig. 3.6 Molecular structures of donor polymers for high-efficiency organic solar cell devices.

102

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 3.7 Molecular structures of third components used for high-efficiency ternary blend organic solar cells.

mostly from an increase in JSC and FF, and an increase from 9.6% to 10.7% for PTB7Th. They observed a cascade of electron and energy transfer processes—photon energy harvested by ASSQ was passed via a F€ orster resonance energy transfer (FRET) mechanism to DPSQ and PTB7, and charge transfer was observed from PTB7-Th to DPSQ, and eventually to PCBM.

Table 3.1 Summary of device parameters for high-performing ternary blend OPVs and their corresponding binary systems JSC (mA cm22)

FF (%)

PCEmax (%)

Refs.

PTB7:PC71BM PTB7:PC71BM:DPPEZnPTEH PTB7-Th:PC71BM PTB7-Th:PC71BM:ASSQ:DPSQ PTB7-Th:PC71BM (thickness 250 nm) PTB7-Th:PC71BM:BTR (thickness 250 nm) PTB7-Th:PC71BM PTB7-Th:PC71BM:DR3TSBDT PTB7-Th:PC71BM PTB7-Th:PC71BM:p-DTS(FBTTH2)2 PTB7-Th:PC71BM PTB7-Th:PC71BM:ICBA PffBT4T-2OD:PC71BM PffBT4T-2OD:PC71BM:BTR PDBT-T1:PC71BM PDBT-T1:PC71BM:ITIC-Th PBDB-T:PC71BM PBDB-T:PC71BM:ITIC PBDB-T:IDT-2O PBDB-T:IDT-2O: PC71BM PBDB-T:IT-M PBDB-T:IT-M:bis[70]PCBM PBDB-T:IT-M PBDB-T:IT-M:N2200 PDOT: PC71BM PDOT: PC71BM:ITIC PTFB-O:ITIC-Th PTFB-O:ITIC-Th:EIEC-Th PTB7-Th:IDTBR:IDFBR PSTZ:IDIC PSTZ:IDIC:ITIC PBDB-T:IT-M PBDB-T:IT-M:ITCN J52:IT-M J52:IT-M:IEICO

0.747 0.769 0.794 0.789 0.771 0.751 0.785 0.765 0.797 0.737 0.799 0.826 0.74 0.77 0.915 0.934 0.92 0.91 0.86 0.87 0.937 0.952 0.930 0.934 0.94 0.96 0.920 0.948 1.03 0.928 0.953 0.937 0.954 0.843 0.847

16.96 18.68 17.05 17.82 17.8 21.4 19.43 22.63 20.70 21.67 16.70 17.90 17.43 18.28 13.24 15.54 15.9 17.7 15.70 16.80 16.70 17.39 16.64 17.17 13.92 17.49 16.8 16.4 17.2 14.7 17.4 17.05 17.67 17.1 19.7

70.04 74.89 68.60 71.7 50.6 70.0 64.90 68.51 55.50 67.50 67.0 68.8 73.93 74.02 76.2 70.5 65.6 68.3 71.6 72.11 69.0 73.7 71.7 75.5 72.9 66.8 66 72 60 59.1 66.9 68.1 71.2 65.1 66.8

9.09 11.03 9.60 10.70 7.07 11.4 10.10 12.10 9.16 10.78 9.28 10.5 9.76 10.59 9.29 10.48 9.6 11.0 9.65 10.67 10.80 12.20 11.10 12.10 9.54 11.21 10.2 11.2 11.0 8.06 11.1 10.81 12.16 9.4 11.1

Nian et al. (2016) Nian et al. (2016) Goh et al. (2016) Goh et al. (2016) Zhang et al. (2017a) Zhang et al. (2017a) Kumari et al. (2017) Kumari et al. (2017) Zhang et al. (2017b) Zhang et al. (2017b) Cheng et al. (2016) Cheng et al. (2016) Xiao et al. (2018) Xiao et al. (2018) Liu et al. (2017) Liu et al. (2017) Bi et al. (2018) Bi et al. (2018) Zhang et al. (2018a) Zhang et al. (2018a) Zhao et al. (2017) Zhao et al. (2017) An et al. (2018) An et al. (2018) Zhang et al. (2018b) Zhang et al. (2018b) Jiang et al. (2018a) Jiang et al. (2018a) Baran et al. (2017) Su et al. (2017) Su et al. (2017) Jiang et al. (2018b) Jiang et al. (2018b) Yu et al. (2017) Yu et al. (2017)

103

VOC (V)

Functional blends of organic materials for optoelectronic applications

System

104

Handbook of Organic Materials for Electronic and Photonic Devices

The other enhancement mechanisms originate from morphology effects, which may be divided into several categories—crystallinity, control of domain size, and control of domain purity. The addition of crystalline small molecules as third components was recently demonstrated as a viable strategy to achieve higher PCEs. For example, two similar molecules (DR3TSBDT and BTR) were used to increase the crystallinity of the PTB7-Th:PC71BM blend. Despite their structural similarities, incorporation of these molecules had drastically different effects on morphology (Zhang et al., 2017a; Kumari et al., 2017). When highly ordered liquid-crystalline small-molecule benzodithiophene terthiophene rhodamine (BTR) was used as a third component, this allowed the fabrication of thicker devices in which PCE increased from 7.07% to 11.40% for 250-nm film. A significant increase in hole mobility was observed (more than fivefold), which was mostly attributed to a shorter π-π stacking distance, increased crystallinity, and increased domain purity (0.53 for binary versus 0.75 for ternary blend) (Zhang et al., 2017a). Similar results were observed for another system, with PffBT4T-2OD:PC71BM as the main components and BTR as the third component, where the maximum achieved PCE was 10.59% for 280-nm film (Xiao et al., 2018). At the same time, for a modified version of this small-molecule DR3TSBDT, where thiophene units were replaced with alkylthio chains and shorter alkyl chains on the acceptor units, drastically different morphology effects were observed. First, a gradual change in device VOC was observed upon the addition of DR3TSBDT, which typically indicates an alloy formation with the main components, while the opposite trend was observed when BTR-VOC stayed fixed at the PTB7-Th/PC71BM level. In addition, researchers observed the appearance of mixed face-on edge-on phase for polymer blends and an increase in correlation length in qxy direction for ternary blends. Typically, a face-on orientation of π-π stacks is beneficial for charge transport; however, despite the randomization in stack orientation in this case, researchers observed a slight increase in hole and electron mobilities and an overall PCE increase from 10.10% to 12.10% (certified 11.76%) for binary and ternary devices, respectively (Kumari et al., 2017). Another small-molecule third component that demonstrated promising results is the p-DTS(FBTTH2)2 in PTB7-Th/PC71BM system. When a small amount of this small molecule (15%) was added to a polymer blend and film was processed from a chlorobenzene:DIO solution, a study observed PCEs of 9.20% and 10.50% for binary and ternary devices, respectively (Zhang et al., 2015). An addition of the small molecule did not affect phase separation of the components, and it formed an alloy with the blend’s main components, resulting in an increased ratio of face-on- to edge-on-oriented polymer π-π stacks and increased correlation lengths in the film. Very similar results were observed for films coated from a greener, nonhalogenated solvent ortho-xylene, which also allowed thicker films (about 270 nm) and a maximum PCE of 10.78% for optimized ternary blend devices (Zhang et al., 2017b). There are also cases when multiple enhancement mechanisms contribute simultaneously. One class of material in which multiple mechanisms were successfully balanced is a family of fused-ring electron acceptor (FREA) materials. These molecules have a very flat, fused ring donor unit (indacenodithiophene or dithienoindacenodithiophene) in the middle, and two acceptor units at the periphery. The

Functional blends of organic materials for optoelectronic applications

105

donor part is shielded with nonconjugated groups; therefore, its participation in electron-transfer processes in the device is minimized. The molecule becomes endcapped with the electron acceptor moieties. In addition, this family of molecules has unique optical properties, absorbing light in the visible and near-IR parts of the spectrum, which usually lack coverage in other types of materials. Several different types of ternary blends demonstrated superior performance with FREA materials. We will categorize them into three groups: systems with FREA- and fullerene-based acceptors, systems with FREA and another nonfullerene non-FREA acceptor, and systems with two FREA-type acceptors. For example, ITIC-Th was studied as a third component in PDBT-T1:PC71BM systems (Liu et al., 2017). The researchers observed PCE increase from 9.29% for binary to 10.48% for ternary devices (conventional architecture) with 50% content of ITIC-Th; this result was attributed mostly to an increase in VOC and JSC. Enhanced performance originated mostly from an increase in hole and electron mobilities and additional light harvesting in near-IR regions. No major changes took place in molecular packing, and three separate phases were observed—polymer, ITIC-Th, and PC71BM—indicating that both polymer:ITIC-Th and polymer:PC71BM systems form independent percolation networks and thus work in a parallel-like manner. ITIC was also used as a third component in a similar polymer system, PBDB-T: PC71BM (Bi et al., 2018). The optimum composition for this system was PBDB-T: ITIC:PC71BM (1:0.7:0.3), which is much higher than previously seen, and here, PC71BM effectively is the third component. When 30% PC71BM was added to a PBDB-T:ITIC blend, the efficiency increased from 9.6% to 11.0%, mostly due to an increase in JSC and FF. The increase in FF was attributed to an increase in hole and electron mobilities, while the increase in JSC was attributed to enhanced light harvesting and improved morphology (increased donor and acceptor correlation lengths). Moreover, the authors observed that the addition of PC71BM enhances the favorable vertical distribution of components (for inverted architecture); acceptors concentrate at the bottom of the device, and polymer forms a crystalline phase at the interface near hole-collecting electrodes, which improves the efficiency of charge transport. When ITIC was replaced with a wider band-gap acceptor IDT-2O in the same system, a surprisingly high PCE was observed as well, despite the narrower absorption coverage (Zhang et al., 2018a). Ternary blends showed increased PCE (inverted configuration) of 10.67% compared to the binary blends PBDB-T:PC71BM (7.47% PCE) and PBDB-T:IDT-2O (9.65% PCE). The increase originated from enhanced light absorption, increased and more balanced hole and electron mobilities, and improved morphology. The addition of PC71BM to a blend decreased excessive aggregation of IDT-2O, resulting in more balanced phase separation. Interestingly, a less crystalline analog of IDT-2O, where alkyl chains were replaced with bulkier alkylphenyl chains IDT-2B, showed the opposite trend: ternary blend performance decreased upon the addition of PC71BM, which indicated a significant disruption in electron conducting. Very impressive results were achieved for the same polymer PBDB-T and methylated ITIC IT-M when bis[70]PCBM was used as the third component (Zhao et al., 2017). Upon the incorporation of 20% bis[70]PCBM, the efficiency of devices

106

Handbook of Organic Materials for Electronic and Photonic Devices

increased from 10.80% to 12.20% PCE, which is currently one of the highest results for a single-junction ternary blend OPV. The enhancement was attributed to extended absorption in a short-wavelength region, increased hole and electron mobilities, and preferential concentration of bis[70]PCBM on the top surface of the film at the interface between the active layer and a cathode, which is favorable for charge transport in devices with a conventional architecture. Incorporating FREA into polymer:PCBM system also may result in thicker film devices. Thus, the incorporation of 20% ITIC into a PDOT:PC71BM system increased PCE from 9.54% for binary devices to 11.21% for ternary devices in 230-nm film (Zhang et al., 2018b). The device performance was highly dependent on the amount of the solvent additive DIO that was present. Crystallinity of the ITIC phase in the ternary blend increased with increasing content of DIO from 0% to 1%; however, at higher concentrations, large aggregates of ITIC were formed that deteriorated device performance. The optimum PCE was observed for 0.5% DIO. Devices with FREA acceptors and polymer nonfullerene acceptors were used as a viable strategy for high-performing devices. Thus, the naphthalene diimide (NDI)based polymer acceptor N2200 was used as a third component in PBDB-T:ITIC, PBDB-T:ITIC-Th, and PBDB:IT-M blends (An et al., 2018). Impressive results were achieved when 10% of N2200 was incorporated. For blends with PBDB-T and ITIC, PCE increased from 10.03% for binary to 11.41% for ternary blends. Even better results were achieved for ternary devices with ITIC-Th and IT-M—11.40% and 12.10%, respectively. The enhancement originated from more balanced hole and electron mobility, extended light absorption, and a more pronounced face-on orientation in PBDB-T and ITIC upon incorporation of N2200. In addition, these devices showed excellent air stability—80% of initial PCE was retained after storage in air for 1000 h. There are multiple reports on ternary blends where two FREA-type acceptors are present in film with one polymer donor and demonstrate excellent power conversion efficiencies of 11% and higher ( Jiang et al., 2018a,b; Baran et al., 2017; Su et al., 2017; Yu et al., 2017; Zhang et al., 2018c). The common feature of these systems is the presence of highly crystalline, wide band-gap donor polymer and two highly compatible acceptors forming an alloylike acceptor phase. It has been suggested that the ability to form an alloy is highly dependent on surface tension of the materials, and even structurally dissimilar systems having close surface tension may form an alloy ( Jiang et al., 2018a). As a result of alloy formation, VOC has a linear dependence on the device composition, thus avoiding pinning. Also, addition of a third component may fine-tune the crystallinity of the acceptor alloy phase, typically causing crystalline face-on structures to form.

3.4

Conclusion

In this chapter, we have provided an overview of blends of organic materials for PV applications. We have covered fundamental thermodynamic principles that govern the formation of polymer blends and provided specific examples of how these principles

Functional blends of organic materials for optoelectronic applications

107

may be used to predict and study their behavior. Additional examples were given on how processing conditions may affect the kinetics of film formation, and thus largely determine the final outcome and eventual performance of these devices. We have used this material to summarize some of the best-performing blends in modern organic PVs and deduce common characteristics of these blends that resulted in high performance. Special attention has been paid to ternary blend solar cells—one of the emerging fields in OPV that has seen significant progress recently due to tremendous success in developing high-performance, nonfullerene electron acceptors. Due to increasing energy needs, the importance of alternative energy sources such as PVs continue to grow in the future. The search for new PV materials will be accompanied by the simultaneous development of theoretical and experimental tools capable of predicting material properties and avoid the trial-and-error approach in the next generation of high-efficiency PV materials.

References An, Q., Zhang, F., Zhang, J., Tang, W., Deng, Z., Hu, B., 2016. Versatile ternary organic solar cells: a critical review. Energy Environ. Sci. 9, 281–322. An, Q., Zhang, F., Gao, W., Sun, Q., Zhang, M., Yang, C., Zhang, J., 2018. High-efficiency and air stable fullerene-free ternary organic solar cells. Nano Energy 45, 177–183. Armin, A., Stoltzfus, D.M., Donaghey, J.E., Clulow, A.J., Nagiri, R.C.R., Burn, P.L., Gentle, I.R., Meredith, P., 2017. Engineering dielectric constants in organic semiconductors. J. Mater. Chem. C 5, 3736–3747. Baran, D., Ashraf, R.S., Hanifi, D.A., Abdelsamie, M., Gasparini, N., R€ ohr, J.A., Holliday, S., Wadsworth, A., Lockett, S., Neophytou, M., et al., 2017. Reducing the efficiency-stabilitycost gap of organic photovoltaics with highly efficient and stable small molecule acceptor ternary solar cells. Nat. Mater. 16, 363–369. Bartelt, J.a., Douglas, J.D., Mateker, W.R., Labban, A.E., Tassone, C.J., Toney, M.F., Frechet, J.M.J., Beaujuge, P.M., McGehee, M.D., 2014. Controlling solution-phase polymer aggregation with molecular weight and solvent additives to optimize polymerfullerene bulk heterojunction solar cells. Adv. Energy Mater. 4, 1–11. Bi, P., Xiao, T., Yang, X., Niu, M., Wen, Z., Zhang, K., Qin, W., So, S.K., Lu, G., Hao, X., et al., 2018. Regulating the vertical phase distribution by fullerene-derivative in high performance ternary organic solar cells. Nano Energy 46, 81–90. Chen, X., Liu, X., Burgers, M.A., Huang, Y., Bazan, G.C., 2014. Green-solvent-processed molecular solar cells. Angew. Chem. Int. Ed. Eng. 14378–14381. Cheng, S.Z.D., 2008. Phase Transitions in Polymers: The Role of Metastable States, first ed. Elsevier Science S.A., Amsterdam/Boston. Cheng, P., Yan, C., Wu, Y., Wang, J., Qin, M., An, Q., Cao, J., Huo, L., Zhang, F., Ding, L., et al., 2016. Alloy acceptor: superior alternative to pcbm toward efficient and stable organic solar cells. Adv. Mater. 28, 8021–8028. Cho, N., Schlenker, C.W., Knesting, K.M., Koelsch, P., Yip, H.L., Ginger, D.S., Jen, A.K.Y., 2014. High-dielectric constant side-chain polymers show reduced non-geminate recombination in heterojunction solar cells. Adv. Energy Mater. 4, 1301857. Chueh, C.-C., Li, C.-Z., Jen, A.K.-Y., 2015. Recent progress and perspective in solutionprocessed interfacial materials for efficient and stable polymer and organometal perovskite solar cells. Energy Environ. Sci. 8, 1160–1189.

108

Handbook of Organic Materials for Electronic and Photonic Devices

Clarke, T.M., Durrant, J.R., 2010. Charge photogeneration in organic solar cells. Chem. Rev. 110, 6736–6767. Collado-Fregoso, E., Boufflet, P., Fei, Z., Gann, E., Ashraf, S., Li, Z., McNeill, C.R., Durrant, J.R., Heeney, M., 2015. Increased exciton dipole moment translates into charge-transfer excitons in thiophene-fluorinated low-bandgap polymers for organic photovoltaic applications. Chem. Mater. 27, 7934–7944. Collins, B.A., Li, Z., Tumbleston, J.R., Gann, E., Mcneill, C.R., Ade, H., 2013. Absolute measurement of domain composition and nanoscale size distribution explains performance in PTB7:PC71bm solar cells. Adv. Energy Mater. 3, 65–74. de Zerio, A.D., M€uller, C., 2018. Glass forming acceptor alloys for highly efficient and thermally stable ternary organic solar cells. Adv. Energy Mater. 1702741, 1–18. Diao, Y., Shaw, L., Bao, Z., Mannsfeld, S.C.B., 2014. Morphology control strategies for solution-processed organic semiconductor thin films. Energy Environ. Sci. 7, 2145–2159. Dou, L., You, J., Hong, Z., Xu, Z., Li, G., Street, R.A., Yang, Y., 2013. 25th Anniversary article: a decade of organic/polymeric photovoltaic research. Adv. Mater. 25, 6642–6671. Flory, P.J., 1941. Thermodynamics of high polymer solutions. J. Chem. Phys. 9, 660. Garcı´a de Arquer, F.P., Armin, A., Meredith, P., Sargent, E.H., 2017. Solution-processed semiconductors for next-generation photodetectors. Nat. Rev. Mater. 2, 16100. Garcia-Ojalvo, J., Lacasta, A.M., Sancho, J.M., Toral, R., 1998. Phase separation driven by external fluctuations. Europhys. Lett. 42, 125. Goh, T., Huang, J.S., Yager, K.G., Sfeir, M.Y., Nam, C.Y., Tong, X., Guard, L.M., Melvin, P.R., Antonio, F., Bartolome, B.G., et al., 2016. Quaternary organic solar cells enhanced by cocrystalline squaraines with power conversion efficiencies >10%. Adv. Energy Mater. 6, 1600660. Hedley, G.J., Ward, A.J., Alekseev, A., Howells, C.T., Martins, E.R., Serrano, L.A., Cooke, G., Ruseckas, A., Samuel, I.D.W., 2013. Determining the optimum morphology in highperformance polymer-fullerene organic photovoltaic cells. Nat. Commun. 4, 2867. Huang, Y., Kramer, E.J., Heeger, A.J., Bazan, G.C., 2014. Bulk heterojunction solar cells: morphology and performance relationships. Chem. Rev. 114, 7006–8043. Huggins, M.L., 1941. Solutions of long chain compounds. J. Chem. Phys. 9, 440. Hussain, A.M., Hussain, M.M., 2016. CMOS-technology-enabled flexible and stretchable electronics for internet of everything applications. Adv. Mater. 28, 4219–4249. Jansen-van Vuuren, R.D., Armin, A., Pandey, A.K., Burn, P.L., Meredith, P., 2016. Organic photodiodes: the future of full color detection and image sensing. Adv. Mater. 28, 4766–4802. Jiang, K., Zhang, G., Yang, G., Zhang, J., Li, Z., Ma, T., Hu, H., Ma, W., Ade, H., Yan, H., 2018a. Multiple cases of efficient nonfullerene ternary organic solar cells enabled by an effective morphology control method. Adv. Energy Mater. 8, 1701370. Jiang, W., Yu, R., Liu, Z., Peng, R., Mi, D., Hong, L., Wei, Q., Hou, J., Kuang, Y., Ge, Z., 2018b. Ternary nonfullerene polymer solar cells with 12.16% efficiency by introducing one acceptor with cascading energy level and complementary absorption. Adv. Mater. 30, 1–7. Kumari, T., Lee, S.M., Kang, S.-H., Chen, S., Yang, C., 2017. Ternary solar cells with a mixed face-on and edge-on orientation enable an unprecedented efficiency of 12.1%. Energy Environ. Sci. 10, 258–265. Lakhwani, G., Rao, A., Friend, R.H., 2014. Bimolecular recombination in organic photovoltaics. Annu. Rev. Phys. Chem. 65, 557–581. Li, H., Lu, K., Wei, Z., 2017. Polymer/small molecule/fullerene based ternary solar cells. Adv. Energy Mater. 7, 1602540.

Functional blends of organic materials for optoelectronic applications

109

Liang, Y., Yu, L., 2010. A new class of semiconducting polymers for bulk heterojunction solar cells with exceptionally high performance. Acc. Chem. Res. 43, 1227–1236. Liang, Y., Xiao, S., Feng, D., Yu, L., 2008. Control in energy levels of conjugated polymers for photovoltaic application. J. Phys. Chem. C 112, 7866–7871. Liang, C., Wang, Y., Li, D., Ji, X., Zhang, F., He, Z., 2014. Modeling and simulation of bulk heterojunction polymer solar cells. Sol. Energy Mater. Sol. Cells 127, 67–86. Lin, J.D.A., Mikhnenko, O.V., Chen, J., Masri, Z., Ruseckas, A., Mikhailovsky, A., Raab, R.P., Liu, J., Blom, P.W.M., Loi, M.A., et al., 2014. Systematic study of exciton diffusion length in organic semiconductors by six experimental methods. Mater. Horiz. 1, 280–285. Liu, T., Xue, X., Huo, L., Sun, X., An, Q., Zhang, F., Russell, T.P., Liu, F., Sun, Y., 2017. Highly efficient parallel-like ternary organic solar cells. Chem. Mater. 29, 2914–2920. Lu, L., Yu, L., 2014. Understanding low bandgap polymer PTB7 and optimizing polymer solar cells based on it. Adv. Mater. 26, 4413–4430. Lu, L., Zheng, T., Wu, Q., Schneider, A.M., Zhao, D., Yu, L., 2015a. Recent advances in bulk heterojunction polymer solar cells. Chem. Rev. 115, 12666–12731. Lu, L., Kelly, M.A., You, W., Yu, L., 2015b. Status and prospects for ternary organic photovoltaics. Nat. Photonics 9, 491–500. Lyons, B.P., Clarke, N., Groves, C., 2012. The relative importance of domain size, domain purity and domain interfaces to the performance of bulk-heterojunction organic photovoltaics. Energy Environ. Sci. 5, 7657. M€ uller, C., Ferenczi, T.A.M., Campoy-Quiles, M., Frost, J.M., Bradley, D.D.C., Smith, P., Stingelin-Stutzmann, N., Nelson, J., 2008. Binary organic photovoltaic blends: a simple rationale for optimum compositions. Adv. Mater. 20, 3510–3515. Nian, L., Gao, K., Liu, F., Kan, Y., Jiang, X., Liu, L., Xie, Z., Peng, X., Russell, T.P., Ma, Y., 2016. 11% efficient ternary organic solar cells with high composition tolerance via integrated near-IR sensitization and interface engineering. Adv. Mater. 28, 8184–8190. Ostroverkhova, O., 2016. Organic optoelectronic materials: mechanisms and applications. Chem. Rev. 116, 13279–13412. Richter, L.J., DeLongchamp, D.M., Amassian, A., 2017. Morphology development in solutionprocessed functional organic blend films: an in situ viewpoint. Chem. Rev. 117, 6332–6366. Rolczynski, B.S., Szarko, J.M., Son, H.J., Yu, L., Chen, L.X., 2014. Effects of exciton polarity in charge-transfer polymer/pcbm bulk heterojunction films. J. Phys. Chem. Lett. 5, 1856–1863. Savoie, B.M., Dunaisky, S., Marks, T.J., Ratner, M.A., 2015. The scope and limitations of ternary blend organic photovoltaics. Adv. Energy Mater. 5, 1400891. Stoltzfus, D.M., Clulow, A.J., Jin, H., Burn, P.L., Gentle, I.R., 2016. Impact of dimerization on phase separation and crystallinity in bulk heterojunction films containing non-fullerene acceptors. Macromolecules 49, 4404–4415. Su, W., Fan, Q., Guo, X., Meng, X., Bi, Z., Ma, W., Zhang, M., Li, Y., 2017. Two compatible nonfullerene acceptors with similar structures as alloy for efficient ternary polymer solar cells. Nano Energy 38, 510–517. Torabi, S., Jahani, F., Van Severen, I., Kanimozhi, C., Patil, S., Havenith, R.W.A., Chiechi, R.C., Lutsen, L., Vanderzande, D.J.M., Cleij, T.J., et al., 2015. Strategy for enhancing the dielectric constant of organic semiconductors without sacrificing charge carrier mobility and solubility. Adv. Funct. Mater. 25, 150–157. Treat, N.D., Chabinyc, M.L., 2014. Phase separation in bulk heterojunctions of semiconducting polymers and fullerenes for photovoltaics. Annu. Rev. Phys. Chem. 65, 59–81.

110

Handbook of Organic Materials for Electronic and Photonic Devices

van Franeker, J.J., Turbiez, M., Li, W., Wienk, M.M., Janssen, R.A.J., 2015. A real-time study of the benefits of co-solvents in polymer solar cell processing. Nat. Commun. 6, 6229. Wolfer, P., Schwenn, P.E., Pandey, A.K., Fang, Y., Stingelin, N., Burn, P.L., Meredith, P., 2013. Identifying the optimum composition in organic solar cells comprising non-fullerene electron acceptors. J. Mater. Chem. A 1, 5989–5995. Xiao, L., Gao, K., Zhang, Y., Chen, X., Hou, L., CAO, Y., Peng, X., 2016. A complementary absorption small molecule for efficient ternary organic solar cells. J. Mater. Chem. A 4, 5288–5293. Xiao, M., Zhang, K., Jin, Y., Yin, Q., Zhong, W., Huang, F., Cao, Y., 2018. Low temperature processed high-performance thick film ternary polymer solar cell with enhanced stability. Nano Energy 48, 53–62. Ye, L., Hu, H., Ghasemi, M., Wang, T., Collins, B.A., Kim, J.H., Jiang, K., Carpenter, J.H., Li, H., Li, Z., et al., 2018. Quantitative relations between interaction parameter, miscibility and function in organic solar cells. Nat. Mater. 17, 253–260. Yu, G., Gao, J., Hummelen, J.C., Wudl, F., Heeger, A.J., 1995. Polymer photovoltaic cells: enhanced efficiencies via a network of internal heterojunctions. Science 270, 1789–1791. Yu, R., Zhang, S., Yao, H., Guo, B., Li, S., Zhang, H., Zhang, M., Hou, J., 2017. Two wellmiscible acceptors work as one for efficient fullerene-free organic solar cells. Adv. Mater. 29, 1700437. Zhang, J., Zhang, Y., Fang, J., Lu, K., Wang, Z., Ma, W., Wei, Z., 2015. Conjugated polymersmall molecule alloy leads to high efficient ternary organic solar cells. J. Am. Chem. Soc. 137, 8176–8183. Zhang, G., Zhang, K., Yin, Q., Jiang, X., Zhang, G., Zhang, K., Yin, Q., Jiang, X., Wang, Z., Xin, J., et al., 2017a. High-performance ternary organic solar cell enabled by a thick active layer containing a liquid crystalline small molecule donor. J. Am. Chem. Soc. 139, 2387–2395. Zhang, J., Zhao, Y., Fang, J., Yuan, L., Xia, B., Wang, G., Wang, Z., Zhang, Y., Ma, W., Yan, W., et al., 2017b. Enhancing performance of large-area organic solar cells with thick film via ternary strategy. Small 13, 1–8. Zhang, C., Feng, S., Liu, Y., Ming, S., Lu, H., Ma, D., Xu, X., Wu, Y.-Z., Bo, Z., 2018a. High efficiency ternary polymer solar cells based on a fused pentacyclic electron acceptor. J. Mater. Chem. A 6, 6854–6859. Zhang, T., Zhao, X., Yang, D., Tian, Y., Yang, X., 2018b. Ternary organic solar cells with >11% efficiency incorporating thick photoactive layer and nonfullerene small molecule acceptor. Adv. Energy Mater. 8, 1–9. Zhang, M., Gao, W., Zhang, F., Mi, Y., Wang, W., An, Q., Wang, J., Ma, X., Miao, J., Hu, Z., et al., 2018c. Efficient ternary non-fullerene polymer solar cells with PCE of 11.92% and FF of 76.5%. Energy Environ. Sci. 11, 841–849. Zhao, J., Li, Y., Yang, G., Jiang, K., Lin, H., Ade, H., Ma, W., Yan, H., 2016. Efficient organic solar cells processed from hydrocarbon solvents. Nat. Energy 1, 15027. Zhao, W., Li, S., Zhang, S., Liu, X., Hou, J., 2017. Ternary polymer solar cells based on two acceptors and one donor for achieving 12.2% efficiency. Adv. Mater. 29, 1604059. Zhou, N., Dudnik, A.S., Li, T.I.N.G., Manley, E.F., Aldrich, T.J., Guo, P., Liao, H.-C., Chen, Z., Chen, L.X., Chang, R.P.H., et al., 2016. All-polymer solar cell performance optimized via systematic molecular weight tuning of both donor and acceptor polymers. J. Am. Chem. Soc. 138, 1240–1251. Zhugayevych, A., Tretiak, S., 2014. Theoretical description of structural and electronic properties of organic photovoltaic materials. Annu. Rev. Phys. Chem. 66, 305–330.

Organic photonic nanostructures

4

Deirdre M. O’Carroll*,†,‡ *Rutgers University, Department of Materials Science and Engineering, Piscataway, NJ, United States, †Rutgers University, Department of Chemistry and Chemical Biology, Piscataway, NJ, United States, ‡Trinity College Dublin, School of Physics, Dublin 2, Ireland

Organic conjugated (i.e., semiconducting) materials are at the forefront of many nextgeneration photonic and optoelectronic technologies, including organic light-emitting diodes and chemical and biological fluorescent sensors. The photophysical properties of these materials can be controlled and optimized through the formation of nanoscale-confined geometries such as nanoparticles, aggregates, nanofibers, nanowires, and ultrathin films. The development of controlled nanofabrication methods for organic nano-objects has enabled the demonstration of numerous photonic and optoelectronic functions at the nanoscale, such as lasing, photodetection, fluorescence-based chemical sensing and biosensing, and live-cell imaging (Kim et al., 2007; O’Brien et al., 2006; O’Carroll et al., 2008, 2007a; Zhang et al., 2006, 2012; Hlaing et al., 2011; Aryal et al., 2009; Noy et al., 2002; Harfenist et al., 2004; Huang et al., 2003; Li et al., 2004, 2003, 2014, 2018, 2017; Steinhart, 2002; Chuangchote et al., 2007; Valentini et al., 2007; Abidian et al., 2006; Huynh, 2002; Herland et al., 2007; Zussman et al., 2003; Liu et al., 2003, 2013; Cho et al., 2005; Kuo et al., 2007; Cacialli et al., 2004; Steinhart et al., 2004; Martin, 1994; Hulteen and Martin, 1997; Shin, 2004; O’Carroll, 2007; Hwang et al., 2016; Vannahme et al., 2010; Arregui et al., 2010; Avinash and Govindaraju, 2018; Ishihara et al., 2014; Massey et al., 2015; Weber et al., 2016; Fasano et al., 2014; Nguyen and Tran, 2014; Sun et al., 2011; Dai et al., 2015; Fischer et al., 2015; Hu et al., 2017; Qian et al., 2015; Senthilkumar et al., 2016; Shen et al., 2011; Wang et al., 2014). In this chapter, we discuss the photonic characteristics of organic nanostructures and devices, with a focus on how excitons and photons can be manipulated and managed though nanoscale confinement of organic conjugated molecules. We also include case studies from the literature on how internal molecular morphology can be controlled in organic nanoparticles, nanowires, and nanofibers, and in turn, how internal morphology affects the photonic properties and functions of these structures.

4.1

Types of organic photonic nanostructures: Particles, wires, fibers

Nanoscale confinement of organic conjugated materials is of interest for photonic applications, including live-cell imaging, fluorescence sensing, and integrated photonics. In contrast to quantum-confined inorganic materials, most organic nanowires Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00004-8 © 2019 Elsevier Ltd. All rights reserved.

112

Handbook of Organic Materials for Electronic and Photonic Devices

and nanoparticles, which typically contain many tens or hundreds of organic conjugated molecules, do not exhibit predictable size-dependent changes in their photonic properties. This is primarily due to the strongly bound excitons that occur in organic conjugated molecular materials arising from the low dielectric constant and, hence, low electron-hole Coulomb screening (Scholes and Rumbles, 2006). Therefore, organic conjugated molecules can be thought of as intrinsically quantum-confined materials, in that their excitons are typically confined to a single molecular unit or segment. This is particularly true when intermolecular interactions are minimized (e.g., by the addition of long nonconjugated side chains to the conjugated molecule). Improved control of excited-state interactions in organic conjugated materials can be achieved by nanoscale confinement in two or three dimensions (Fig. 4.1), leading to new or more distinct optical behavior compared to thin films or the bulk material (Nguyen et al., 2000; Kim and Swager, 2001). For example, studies of organic conjugated polymers incorporated into nonemissive nanopores have investigated the control of energy transfer along polymer chains by ordering and isolating them from each other (Nguyen, 2000). Two-dimensionally confined organic nanowires and nanofibers have been investigated in order to explore new or enhanced properties of potential interest in miniaturized electronic and photonic devices (Fig. 4.1C and D) (O’Brien et al., 2006; O’Carroll et al., 2007a; O’Carroll, 2007). Such nanowire- and nanofiber-based nanostructures contain tens or hundreds of organic molecules, and their photonic and electronic properties are often determined by internal molecular organization. Likewise, three-dimensionally confined organic nanoparticles have emerged as relevant candidates for bioimaging and chemical sensing due to their increased surface-to-volume ratio and the greater chemical and surface sensitivity of their luminescence compared to inorganic nanomaterials (Fig. 4.1A and B) (Qian et al., 2015; Kuehne et al., 2012; AL-Amar and Burns, 2011; Fu and Yao, 2001; Pecher and Mecking, 2010; Tuncel and Demir, 2010). In the solid state, intramolecular and intermolecular interactions and ordering can alter the size and degree of confinement of excitons in organic thin films and nanostructures; they can become delocalized over more than one molecular segment or molecule and can more readily diffuse or dissociate into mobile charged states such as polarons (AL-Amar and Burns, 2011; Fu and Yao, 2001; Arias et al., 2013; Montilla et al., 2013; Huo et al., 2015; Brown et al., 2003; Cook et al., 2008; Herrmann et al., 2011; Jiang et al., 2002; Mikhnenko et al., 2015). In particular, side-chain-free organic conjugated monomers exhibit high-degrees of crystallinity due to their ability to close-pack in the solid state. This propensity for close-packing can promote exciton quenching due to the formation of nonemissive bimolecular electronic states (K€ohler et al., 2002; Hu et al., 2015). However, superradiance and aggregation-induced emission (AIE) also have been reported for organic molecular aggregates, indicating that certain types of intermolecular ordering arrangements can promote highly emissive electronic states (Muller et al., 2013; Meinardi et al., 2003; Chernyak et al., 1999; Spano and Mukamel, 1989; Balzer and Rubahn, 2005; Ceballos et al., 2017). Therefore, crystalline organic nanostructures can exhibit significantly altered photonic properties compared to those of a single constituent molecule due to intermolecular interactions (Eder et al., 2017; Spano and Silva, 2014; Martin et al., 2013; Sherwood et al., 2009; Spano, 2006).

Fig. 4.1 Top: Schematics of 2D and 3D confined organic nanostructure types; nanoparticle (left), nanowire/nanofiber (right). (A) Scanning electron micrograph image of 420-nm poly(dioctylfluorene-alt-benzothiadiazole), F8BT, nanoparticles; scale bar is 2 μm (inset shows the molecular structure of F8BT) (Kuehne et al., 2012). (B) Photograph of fluorescent dispersions of conjugated polymer particles in water (Kuehne et al., 2012). (C) Scanning electron microscope (SEM) image (40-degrees tilted view) showing a vertically aligned 1,5diaminoanthraquinone (DAAQ) nanowire array grown at a temperature of 160°C, for 5 min by a physical vapor transport method on a Si substrate; scale bar is 1 μm (inset shows the molecular structure of DAAQ) (Zhao et al., 2009). (D) Fluorescence microscopy image (1

End

Do any of the ligands exhibit a distinctly more extended π-conjugation than the others (consider resonance structures that produce the more extended π-conjugation)? c

1

No

Yes

b

Is M in one of it highest oxidation g states?

Select this ligand

f

No Do any ligands have a significantly more polarized substituent at one terminus of a π-conjugation path between this terminus and the metal? e

Yes

Run subroutine to identify the charge-transfer type

Does only one ligand have aromatic d rings?

Yes

Yes

No Is M coordinatively unsaturated? h

Yes

No Does the ligand responsible for the D-π-A ICT coordinate to M via high-energy lone pairs? i

Yes

No MLCT

LMCT

M=D Identify the strongest A in the ligand that can connect to D via a j π-conjugated path

M=A Identify the strongest D in the ligand that can connect to A via a π-conjugated path k

End

End

Fig. 5.5 Workflow for the classification of the origins of dipolar SHG activity in organometallic complexes.

156 Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 5.6 Four Ru(II)(NH)X-4,40 -bipyridine complexes with their respective β0- and βInt-values (Coe et al., 1997).

Molecular engineering of organic and organometallic second-order nonlinear optical materials

157

by workflow decision (a) as the SHG-relevant D-π-A ICT motif because it is the only organic ligand present. It then falls to a subroutine, beginning in workflow decision (b), to determine if the directionality of ICT is metal-to-ligand or ligand-to-metal in nature. Given that ruthenium is in its low-spin 2+ oxidation state, it is coordinatively saturated in this complex, and the D-π-A motif does not involve metal-ligand coordination via high-energy lone pairs, the workflow decision-making steps (g)– (j) determine that the metal must act as the electron donor of the framework, with the terminal N-methyl (14) and N-phenyl (15) performing the role of electron acceptor (i.e., the D-π-A motif employs MLCT). Having used this workflow to identify the D-π-A fragment of the complex that is responsible for its SHG activity and the directionality of the ICT therein, the secondary CT effects of its surrounding ligands are considered. Using the auxiliary chart in Fig. 5.7, the NH3 ligands in 14 and 15 are classified as donors, and push electrons to ruthenium, thus augmenting the donating capabilities of the metal. This will naturally improve the inherent strength of ICT across the D-π-A motif and enhance the SHG response of the complex. The SHG response for 14 and 15 will be similarly affected by the surrounding ligands because they are the same in both complexes. Indeed, 14 and 15 differ only by the nature of their organic ligand, which is naturally part of the D-π-A motif. So the differences in SHG response between 14 and 15 are a result of the different nature of the acceptor group in the organic ligand: the N atom of the pyridyl ring possesses a positive charge, so it will be stabilized to a greater extent by a phenyl ring (15) than by a methyl group (14); consequently,

Fig. 5.7 Assessing competing constituent factors that contribute to the ICT in the D-π-A motif that is responsible for the second-order NLO response.

158

Handbook of Organic Materials for Electronic and Photonic Devices

the push-pull ICT effects will be more pronounced in 15 than in 14, resulting in 15 exhibiting a larger SHG response, as shown by the increase in both β0 and βint. Turning to complexes 16 and 17, they once again consist of a Ru(II) metal center, but unlike 14 and 15, the metal coordinates to only four NH3 ligands, while the other two ligands comprise different π-conjugated organic moieties, 1-methylimidazole (16–17), and 10 -methyl-2-20 -bipyridinium (16), 10 -phenyl-2-20 -bipyridinium (17). As with 14 and 15, the ligand that represents the SHG-relevant D-π-A motif for these complexes is the 2–20 -bipyridinium moiety. The workflow requires a few more steps to make this decision compared to 14 and 15, as shown in the decision trees of Fig. 5.6. This is because 16 and 17 have two organic ligands to choose from, in terms of identifying which one participates in the SHG-relevant D-π-A ICT motif. In this case, the greater extent of π-conjugation in the 2–20 -bipyridinium moiety dictates its selection in workflow decision (c) as being the organic constituent of this D-π-A motif. For similar reasons as those discussed for 14 and 15, its ICT directionality is determined to be MLCT in nature. Regarding their D-π-A ICT strength, 16 and 17 possess one less NH3 ligand than 14 and 15 (i.e., 4), and so there is less augmentation of the electron-donating capabilities of the Ru(II) metal in 16 and 17 from these types of ligands. Regarding the organic ligand that is not involved in the D-π-A motif, the process shown in Fig. 5.7 classifies the 1-methylimidazole ligand as an acceptor. Thus, it will have a deleterious effect upon the SHG response in 16 and 17, especially because it is diametrically opposed to the D-π-A unit geometrically and, as such, the two organic ligands will generate an electronic pull directly against each other; this direct competition will naturally diminish the overall push-pull-induced ICT, and SHG response. This SHG diminution is clearly observed by comparing the βInt of 14 and 16 to that of 15 and 17. Considering the relative SHG-responses of 16 and 17, the same argument presented for comparing 14 and 15 prevails—that is, the βInt value of 17 is naturally larger than that of 16 owing to the greater acceptor strengths of the methyl (16) and phenyl groups (17). In principle, this classification workflow can be applied to all known SHG-active organometallic complexes. Moreover, its decision-tree logic makes it attractive for being encoded into an algorithm by which D-π-A ICT motifs in SHG-active dipolar organometallic complexes could be identified automatically. This would stand to benefit the emerging materials-by-design strategies to predict new SHG materials; see Section 5.5.3 for more detail. This workflow is nonetheless specific to dipolar SHG effects in organometallic complexes; it does not translate to octupolar SHGactive media, which will be discussed in the next section.

5.4

Octupolar SHG chromophores

5.4.1 From dipolar to octupolar materials Until the early 1990s, almost all work on the molecular design of organic and organometallic materials for SHG applications was conducted on dipolar SHG effects in materials. Despite the fact that the nonlinear responses of octupolar structures have revealed that they have good potential, octupolar compounds remain underrepresented

Molecular engineering of organic and organometallic second-order nonlinear optical materials

159

in the literature. The initial work of Zyss and coworkers subdivided octupolar materials into two classes: (1) materials that exhibit a dipolar and an octupolar contribution and (2) materials that exhibit only an octupolar contribution. Zyss and Ledoux (1994) showed that the third-rank tensor of a material, which describes the first molecular hyperpolarizability, could have only two components of order 1 and 3, such that the irreducible components can be written as β ¼ βJ¼1 βJ¼3

(5.18)

Consequently, even in the absence of a dipolar contribution (βJ¼1), a material may not have an overall β value of zero if its molecular symmetry permits an octupolar contribution (βJ¼3). In principle, structures that have both dipolar and octupolar components may exhibit an augmented nonlinear response. Yet, in practice, the dipolar or octupolar SHG effects of a material have contrasting molecular symmetry requirements, which generally results in these two forms of SHG contributions compromising each other’s prospective performance. The noncentrosymmetry requirements of the dipolar component of SHG effects have already been discussed (cf. Section 5.1.2.2). Given the threefold symmetry of an octupolar moment, octupolar SHG activity is naturally restricted to materials whose molecular structures feature threefold molecular symmetry, as classified by the point groups illustrated in Fig. 5.8. Such structures can be formed from a cubic structure (the “Octupole” shown in Fig. 5.8) via two different methods: (1) projection of the cubic structure along a C3-axis, which gives rise to D3h and thence D3-symmetry; or (2) fusion of the charge in the center of the cubic structure, giving rise to Td- or D2d-symmetry. Octupolar SHG-active molecules, therefore, can be either two-dimensional (2D) or three-dimensional (3D) in their structural form. To date, 2D octupolar SHG-active molecules have been the most studied, especially in the field of organic compounds. To this end, substantial efforts have been devoted to removing the typically compromising dipolar component of β for prospective second-order NLO materials, thus inducing a purely octupolar SHG response. Efficient SHG from a material with a vanishing dipole moment was first shown by Zyss et al. (1981). While the lack of a dipole in a molecular structure is not essential to achieve large first molecular hyperpolarizability, it is considered preferable. We now highlight a range of octupolar compounds, the vast majority of which are 2D and organic in form, as per the current literature.

Fig. 5.8 A depiction of the symmetry options for purely octupolar structures.

160

Handbook of Organic Materials for Electronic and Photonic Devices

5.4.2 Examples of octupolar molecules 5.4.2.1 Historical introduction One of the earliest demonstrations of an octupolar molecule exhibiting SHG effects was performed on triaminotrinitrobenzene (TATB), labeled as 18 in Fig. 5.9 (Ledoux et al., 1990) using powder SHG experiments. This study was shortly followed by a complementary computational study of TATB by Bredas et al. (1992).

Fig. 5.9 (Top) A comparison of the NLO response of the organic dipolar industry standard pNA with that of the organic octupolar TATB. These β-values were determined by ab initio DFT calculations at the CHPF/3-21G level of theory (Bredas et al., 1992). (Bottom) A TAITB compound investigated using HRS techniques (Verbiest et al., 1994).

Molecular engineering of organic and organometallic second-order nonlinear optical materials

161

TATB consists of an arene ring that is replaced with NH2 groups in the 1-, 3-, and 5-positions and NO2 groups in the 2-, 4-, and 6-positions (Fig. 5.9). This alternating donor-acceptor substitution pattern around a π-conjugated (planar) phenyl ring that exhibits sixfold symmetry presents a molecule with D3h point group symmetry. Even though TATB does not exhibit a dipole moment owing to the canceling effects of the three opposing NH2-πring-NO2 dipoles on symmetry grounds, it still exhibits a significant β value on account of an octupolar contribution. Indeed, the computationally derived β-value of TATB is by a factor of 1.60 greater than that of the industrystandard dipolar SHG material, pNA (approximately a 60% increase). Due to the volatility of TATB and its solubility in organic solvents, HRS measurements of its hyperpolarizability have not been pursued. However, Verbiest et al. (1994) investigated a slightly modified form of TATB, known as triisopropylaminotrinitrobenze ne (TIATB), labeled 19 in Fig. 5.9. TIATB has D3h octupolar symmetry and has four nonzero β tensorial components, all equal to the βZZZ-component. Verbiest et al. undertook HRS measurements to investigate its octupolar behavior and noted that making a direct comparison of the HRS-derived βZZZ values of TIATB and pNA is difficult, given the large solvent dependence associated with the pNA measurements. However, as the number of dipole interactions between pNA and the solvent is lowest for p-dioxane, a comparison of this value with TIATB was made. Here, Verbiest et al. show that the hyperpolarizability value of TIATB is almost equal to that of pNA, which, given that pNA has only one significant tensorial β component, means that the total β-value of TIATB is approximately twice the size of pNA. A year later, Stadler et al. (1995) reported a study on three similar octupolar (20, 22, 24) structures and their cognate dipolar (21, 23, 25) methyl-terminated analogs (see Fig. 5.10). They determined the dynamic (i.e., frequency dependent) hyperpolarizability βHRS of these compounds at λ ¼ 1064 nm, and they were able to compare the dipolar and octupolar SHG responses. β1064 nm(HRS) for these three octupolar molecules was found to be 2.5 to 4 times higher than those of their dipolar counterparts. They also discovered that the replacement of cyano substituents for aldehyde groups afforded similar HRS results. These seminal studies stimulated substantial interest in octupolar structures, and a variety of studies into the molecular design of 2D octupolar structures with 1, 3, 5-phenyl backbone architectures with either D3- or D3h-symmetry soon followed.

5.4.2.2 Other D3h octupolar molecules Three of the most extensively investigated D3h octupolar molecules are shown in Fig. 5.11. Therein, β0 for 27 was found to be substantially larger than that of 26, whereupon the R1 group in 26 changes from an electron-withdrawing (A ¼ NO2) to an electron-donating group (D ¼ NEt2) in 27, while X simultaneously changes from an electron-donating (D ¼ OMe) to an electron-withdrawing group (A ¼ NO2). This pairwise switch in D, A substitution patterns suggests that X and R1 need to be of a type A and D, respectively, to afford a good SHG response. This indication is corroborated by the finding that changing X from NO2 to CN groups (27 and 28) affords

162

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 5.10 A comparison of the molecular hyperpolarizability of a range of cyano, pyrimidine-5carbonitrile, and aldehyde-substituted octupolar D3h structures investigated by Stadler et al. (1995).

comparable β0 values (i.e., changing X for another electron withdrawing group will have little effect on the molecular hyperpolarizability). That said, the UV-vis absorption band of 28 exhibits a significant bathochromic shift (Δλ ¼ 64 nm) relative to that of 27, indicating that the energy of the electronic transition associated with the SHG response is affected. The molecular design of D3h octupolar molecules also can employ heteroatoms to induce the required point-group symmetry. One example is the series of derivatives shown in Fig. 5.12 (29–31), which consist of a boroxine ring, in which the oxygen

Molecular engineering of organic and organometallic second-order nonlinear optical materials

163

Fig. 5.11 Three 1,3,5-tris-X-2,4,6-tris-Z-type octupolar structures with methoxy, nitro, and cyano substituents at the 1-, 3-, and 5-positions (Cho et al., 2001, 2002).

atoms alternate with all boron atoms that are substituted with para-substituted phenyl rings, leading to D3h symmetry. 29–31 all exhibit good β0-values, whose magnitude follows the chemical substitution pattern, R1 ¼ NMe2 > OMe > SMe2, which tracks their decreasing electron-donating strength, as one would expect.

5.4.2.3 A D3 octupolar molecule: Case study A study by Zyss et al. (1993) on the chiral ion, ruthenium(II)-tris[2,20 -bipyridyl] (RuTB) (32; see Fig. 5.13), is a classic example of an octupolar molecule with D3 symmetry. The complexation of three 2,20 -bipyridyl ligands to a central Ru(II) ion results in the formation of a chiral D3 propeller-type geometry (Fig. 5.13). The molecular hyperpolarizability of this compound emanates from MLCT from the donating metal center to the electron-accepting tris[2,20 -bipyridyl] ligands. This study provided the first evidence of substantial β-values in trigonal organometallic chiral cations. There was nonetheless disagreement about the value of β0 for RuTB. While Zyss et al. reported β0 ¼ (210  60)  1030 esu, Morrison et al. (1996) demonstrated that its β0-value was in fact (25  5)  1030 esu once the HRS signal had been corrected for multiphoton fluorescence contributions. Multiphoton fluorescence deconvolution in HRS experiments has since been reported (Olbrechts et al., 1998; Wostyn et al., 2001). Despite this initial overestimation of β0, the corrected value for RuTB continues to point toward the beneficial contribution that octupolar molecules can play in producing large hyperpolarizability responses.

5.4.2.4 A D2d octupolar molecule: Case study A metalorganic octupolar molecule with tetrahedral D2d-symmetry was reported by S
en
echal et al. (2002) and consists of the complexation of a Zn(II) central metal ion coordinated to two bipyridyl ligands, wherein the para-position of each pyridyl

164

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 5.12 Molecular structures of three boroxine derivatives, revealing the effects of different para-substituents on the molecular hyperpolarizability.

Fig. 5.13 Molecular structure of RuTB (A) and its octupolar framework (B).

ring is replaced by a π-conjugated chain that is terminated by an electron-donating dNBu2 group (see 33 in Fig. 5.14). The very large reported molecular hyperpolarizability of 33 (β0 ¼ 157  1030 esu) was justified by considering that the harmonic wavelength (955 nm) in the HRS measurement was sufficiently far from the optical absorption cutoff of the complex that any multiphoton fluorescence contributions could be regarded as negligible. The

Molecular engineering of organic and organometallic second-order nonlinear optical materials

165

Fig. 5.14 Molecular structure of Zn(II) D2d of S
en
echal et al. (A) and its octupolar framework (B).

result was compared against an analogous Cu(I) structure (Renouard et al., 1999), whose molecular hyperpolarizability was reported as being 50% smaller (β0 ¼ 78  1030 esu). The superior molecular hyperpolarizability of the subject complex was attributed to the Zn(II) being a better electron acceptor that Cu(I).

5.4.2.5 A Td octupolar molecule: Case study The Td-symmetry of octupolar structures is the least studied to date. Such structures possess 3D CT effects, and thus a 3D transition moment. Such structures are also nonpolar, and hence are primed for the development of noncentrosymmetric SHG materials. A series of organotin complexes with Td symmetry were reported by Lequan et al. (1994); the one with the largest molecular hyperpolarizability value (β0 ¼ 42  1030 esu) is shown in Fig. 5.15 (34). This complex consists of a tin metal bonded to four (E)-N,N-dibutyl-4-(p-tolyldiazenyl)aniline ligands in a tetrahedral arrangement. Each ligand possesses an extended level of π-conjugation that includes two arene rings.

166

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 5.15 Molecular structure of the Sn Td structure of Lequan et al. (A) and its octupolar framework (B).

5.5

Future prospects for the molecular engineering of NLO chromophores

This chapter has shown that molecular design strategies to engineer new dipolar organic SHG materials are quite advanced and well employed, while those of dipolar organometallic complexes are much less developed. It is hoped that the recently introduced algorithm, by which the SHG-relevant D-π-A ICT motif in an organometallic complex can be identified and its strength assessed cf. Section 5.3.3.1, will help move this field toward a more rational approach to materials discovery of dipolar organometallic SHG chromophores. The potential for octupolar materials, organic or organometallic, has yet to be realized, and molecular design rules are still being developed. The discovery of more octupolar materials is needed before sufficient data are available for establishing generic relationships between structure and SHG properties. Nonetheless, the field of octupolar SHG materials heralds exciting prospects for new types of NLO applications that employ SHG in a more 3D fashion, given the threefold symmetry of an octupole. This chapter has presented examples of discrete octupolar molecules, as this reflects the natural progress of the field. Nonetheless, the last decade has seen an emerging trend toward the development of 3D framework structures that embed threefold symmetry to generate octupolar SHG activity. Such developments are highlighted next, together with analogous research findings on 3D materials that issue dipolar SHG responses.

Molecular engineering of organic and organometallic second-order nonlinear optical materials

167

5.5.1 SHG in 3D-framework structures Coordination polymers, especially their subclass of metal–organic frameworks (MOFs), are receiving quite a bit of interest in the context of SHG applications owing to their potential for rational design. This is because their construction is inherently based on molecular building blocks that may install the SHG activity in the material, or SHG may be created by virtue of the symmetry arrangement of the 3D framework constructed. In addition, MOFs can afford good thermal and mechanical stability, and they have the ability to form optically transparent media, which may overcome the nonlinearity transparency trade-off problem (Cheng et al., 1991a). The majority of MOFs that have been developed for NLO applications to date contain Zn2+ or Cd2+ metal ions (Evans and Lin, 2002; Mingabudinova et al., 2016; Wang et al., 2012). This choice of metal avoids any undesirable d-d transitions in the visible region of the electromagnetic spectrum, thereby producing materials with good optical transparency. Such metals typically adopt either fourfold or sixfold coordination geometries, which largely governs the type of MOF structures that form by virtue of the metal residing at the nodes of a MOF. The most prevalent structural categories for SHG-active MOFs are one-dimensional (1D) chains and helices, 2D sheets, and 3D diamondoids, whose SHG activity stretches up to about 46, 100, and 40 times that of KDP, respectively (Wang et al., 2012); see Fig. 5.16 for the corresponding structures. These are all noncentrosymmetric structures, in order to satisfy the macroscopic symmetry requirements for SHG. While this requirement could be relaxed in the molecular-engineering efforts of other material types featured in earlier sections of this chapter, those studies were able to design and characterize SHG on the molecular scale and to use external media to address noncentrosymmetry if needed. MOFs do not have this luxury because they innately provide their macroscopic product. Moreover, the SHG properties of MOFs can be characterized only at the macroscopic level. This is because MOFs are insoluble in common solvents owing to their polymeric nature, and so their molecular hyperpolarizability cannot be determined; rather, the Kurtz powder test for bulk SHG activity is the only experimental recourse for SHG property evaluation. Nonetheless, noncentrosymmetry can be ensured in MOFs by employing chiral ligands as their molecular building blocks. Owing to the use of the Kurtz method for SHG property evaluation, it is not possible to distinguish octupolar and dipolar contributions in MOFs, although the space groups of their structures help to determine if any threefold symmetry is present. It transpires that the majority of MOFs developed so far exhibit dipolar SHG effects (Wang et al., 2012). Thus, SHG applications based on dipolar SHG responses can capitalize on 1D, 2D, or 3D MOFs without considering any additional symmetry aspects of these materials. The case for octupolar MOFs is more complicated. The first report of an SHG-active octupolar 2D framework structure concerned a MOF that features a cadmium-based organometallic building block (Fig. 5.17) which, generates threefold symmetry within 2D molecular sheets—that is, it fulfills the D3 octupolar symmetry requirement of an octupolar SHG material (Lin et al., 1999). The cadmium coordination is chiral, rendering the MOF noncentrosymmetric and thus issuing bulk SHG activity, which was found to be 10 times greater than that of KDP.

168

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 5.16 Molecular building blocks of MOF structures that exhibit macroscopic SHG activity of 46, 100, and 40 times that of KDP, respectively.

It was not until 8 years later that the first SHG-active octupolar 3D MOF was reported, wherein Liu et al. (2007) demonstrated SHG from the 3D cadmium-based organometallic complex shown in Fig. 5.17. Its SHG response is about 15 times higher than that of KDP (i.e., 1.5 times that of the 2D MOF portrayed in Fig. 5.17). Moreover, its SHG response was found to be tuneable by varying the cation in the cages of this MOF, which are present by virtue of its 3D nature; to this end, the H2NMe+2 cation of this MOF could be wholly replaced by NH+4 , Na+, or K+ ions to yield derivatives whose SHG performance was about 15.5, 9, or 11 times that of KDP, respectively. This 3D MOF also represents a rare example of an octupolar material that exhibits tetrahedral, Td-symmetry. Accordingly, 3D CT effects manifest (Lequan et al., 1994), which is particularly exciting because the 3D nature of MOFs could be used to capitalize on this type of CT to achieve optimal SHG activity.

Molecular engineering of organic and organometallic second-order nonlinear optical materials

169

Fig. 5.17 Molecular building blocks of the first 2D and 3D MOFs with octupolar point-group symmetry, which are reported to show SHG activity.

Since these first demonstrations, a few more cadmium-based octupolar MOFs have come forward, including several that are part of a wider series of metal-organoboron frameworks (Liu et al., 2008, 2009). These MOFs exhibit D3-symmetry to ensure octupolar SHG activity which presents at the level of three to four times that of KDP. This field is still in its infancy, given that reported examples of octupolar MOFs remain rare; yet, it has exciting prospects, especially as 3D MOFs displaying Tdsymmetry have the capacity to harness the multidimensional CT attributes of Td octupolar SHG phenomena in an optimal fashion.

170

Handbook of Organic Materials for Electronic and Photonic Devices 90

180

135

0

270 50 nm

0

315

225 270 100 nm

45

135

45

180

315

225

90

90 45

135

0

180

225

315 270 150 nm

Fig. 5.18 Angle-dependent HRS measurements on gold metallic nanostructures showing the increasingly multipolar nature of NLO output as the nanostructure size increases from 50 nm (left—dipolar NLO), to 100 nm (middle—quadrupolar NLO character), to 150 nm (right—highly quadrupolar NLO signature). Modified from J. Nappa, I. Russier-Antoine, E. Benichou, C. Jonin, P.F. Brevet, Second harmonic generation from small gold metallic particles: from the dipolar to the quadrupolar response, J. Chem. Phys. 125 (2006) 184712.

5.5.2 Nanotechnology The future prospects for another type of 3D SHG design, concerning nanomaterials, are also worth highlighting. The 3D aspect of nanomaterials is more complicated than that of MOFs. On the one hand, the shape and size of nanoparticles highly affect the SHG response, owing to quantum confinement effects. On the other hand, massive enhancements in SHG responses have been observed from surface plasmon resonance (SPR) effects on nanomaterials, whereby SPRs generate very large local electromagnetic fields (Barnes et al., 2003). To this end, studies on roughened versus unroughened nanosurfaces have shown up to 104 enhancement in the SHG signal (Ray, 2010). Therefore, nanostructures can be considered to have their own light property characteristics, which are distinct from any other type of material. Two possible types of SHG contribution have been classified where SPR effects are in action, the relative extents of which depend on particle size. Below a certain size threshold, dipolar SHG effects dominate; when the particle size becomes sufficiently large that it is no longer comparable to the wavelength of incident light, SHG arises primarily from multipolar contributions owing to quadrupolar electronic effects (Kelly et al., 2003). Polarization-resolved HRS experiments are typically used to observe such contributions, as is illustrated for the case of gold nanoparticles of different sizes in Fig. 5.18 (Nappa et al., 2006). This original work on metal nanoparticles is now being extended to a wide range of inorganic materials and inorganic/organic hybrid materials. For example, multilayers of gold/thiol nanoparticle arrays (Fig. 5.19A) have demonstrated SHG signals, wherein 13 nanoparticle layers produced the maximum enhancement of SHG: 50-fold enhancement relative to a monolayer (Addison et al., 2009). In addition, a gold/organic nanoparticle structure, Au38(SCH2CH2(Phenyl))24, has been shown to be SHG active (Knoppe et al., 2015).

Molecular engineering of organic and organometallic second-order nonlinear optical materials

171

(A)

(B) Fig. 5.19 (A) Gold (AU)/thiol nanostructures that exhibit maximum SHG enhancement from surface plasmon resonance effects at 13 nanoparticle layers; (B) SHG-active nanoarchitectures built from organic push-pull molecules bound to a cyclotetrasiloxane base.

Meanwhile, others have used chemistry to build nanoarchitectures from the ground up, using purely covalent origins. For example, Ronchi et al. (2009) produced SHGactive materials that comprise a cyclotetrasiloxane base appended with one or four organic push-pull tails (Fig. 5.19B). The SHG responses of these monomers and tetramers were compared, revealing that β-values were fairly constant, although there was a substantial increase in the dipole moment, μ, as a consequence of the organic aggregation. These aggregation effects and ground-up nanoarchitecture ideas present intriguing concepts for the future molecular engineering of SHG materials, particularly those driven by rational design.

5.5.3 SHG materials by design The symmetry requirements of SHG-active materials lend them a structural order that makes them attractive for rational molecular design. The manner by which organic and organometallic molecules can be built up from their donor, acceptor, and π-media constituents to create SHG function, in accordance with the structure-property

172

Handbook of Organic Materials for Electronic and Photonic Devices

relationships illustrated in this chapter, is an additional merit of this field of research. Such relationships were encoded into algorithms in an initial attempt at a systematic materials prediction of new SHG-active organic molecules by Cole and Weng (2010). This work was successful in its own right, but the opportunities for much larger-scale, data-driven SHG materials prediction are now apparent, given recent major advances in big data, artificial intelligence, and high-performance computing (HPC). Access to massive data sets has been stimulated by government-regulated openaccess requirements for data and literature, while machine-learning tools have become mainstream and exascale HPC capabilities are now within our grasp. These timely advances present an ideal opportunity to develop molecular-engineering strategies to systematically design and predict new chemicals for a given device application. Therefore, we are ready for the data-driven discovery of new materials with tailored SHG activity.

5.6

Conclusions

This chapter has surveyed the current state of the art in the field of NLO of organic and organometallic materials, especially with regard to the structure-property relationships that govern SHG applications. We have shown how dipolar organic SHG chromophores can be constructed from their molecular building blocks, which feature an electron donor, electron acceptor, and π-medium. The gradual emergence of dipolar organometallic compounds as advanced SHG materials over the last two decades is also described, as well as the manner by which their D-π-A motifs can be identified and characterized, so that more general structure-property relationships can be developed from them. The essential elements for the molecular design of octupolar structures for SHG applications have been examined for both organic and organometallic complexes. Looking ahead, the emerging developments of dipolar and octupolar SHG materials that adopt 3D MOFs or nanostructured architectures are highlighted. The symmetry requirements that condition SHG-active materials to have ordered structures afford ideal opportunities for the rational molecular design of new SHG materials that can be tailored for a given NLO device application. We close on this topic by predicting an exciting future for data-driven molecular engineering of organic and organometallic materials for SHG applications. Finally, it is important to remember that this chapter has necessarily focused on SHG, given that its associated structure-property relationships are the most well developed of all NLO applications for organic media. Nonetheless, it is worth remembering that distinct structure-property relationships are associated with other NLO phenomena, such as THG (Tykwinski et al., 1998) which, have been and continue to be developed for related NLO applications.

References Addison, C.J., Konorov, S.O., Brolo, A.G., Blades, M.W., Turner, R.F.B., 2009. Tuning gold nanoparticle self-assembly for optimum coherent anti-stokes Raman scattering and second harmonic generation response. J. Phys. Chem. C 113, 3586–3592.

Molecular engineering of organic and organometallic second-order nonlinear optical materials

173

Barnes, W.L., Dereux, A., Ebbesen, T.W., 2003. Surface plasmon subwavelength optics. Nature 424, 824. Bredas, J.L., Meyers, F., Pierce, B.M., Zyss, J., 1992. On the second-order polarizability of conjugated π-electron molecules with octupolar symmetry: the case of triaminotrinitrobenzene. J. Am. Chem. Soc. 114, 4928–4929. Champagne, B., Kirtman, B., 2006. Evaluation of alternative sum-over-states expressions for the first hyperpolarizability of push-pull π-conjugated systems. J. Chem. Phys. 125. 024101. Cheng, L.T., Tam, W., Stevenson, S.H., Meredith, G.R., Rikken, G., Marder, S.R., 1991a. Experimental investigations of organic molecular nonlinear optical polarizabilities. 1. Methods and results on benzene and stilbene derivatives. J. Phys. Chem. 95, 10631–10643. Cheng, L.T., Tam, W., Marder, S.R., Stiegman, A.E., Rikken, G., Spangler, C.W., 1991b. Experimental investigations of organic molecular nonlinear optical polarizabilities. 2. A study of conjugation dependences. J. Phys. Chem. 95, 10643–10652. Cho, B.R., Chajara, K., Oh, H.J., Son, K.H., Jeon, S.-J., 2002. Synthesis and nonlinear optical properties of 1,3,5-methoxy-2,4,6-tris(styryl)benzene derivatives. Org. Lett. 4, 1703–1706. Cho, B.R., Lee, S.J., Lee, S.H., Son, K.H., Kim, Y.H., Doo, J.-Y., Lee, G.J., Kang, T.I., Lee, Y.K., Cho, M., Jeon, S.-J., 2001. Octupolar crystals for nonlinear optics: 1,3,5-trinitro-2,4,6-tris(styryl)benzene derivatives. Chem. Mater. 13, 1438–1440. Clays, K., Persoons, A., 1991. Hyper-Rayleigh scattering in solution. Phys. Rev. Lett. 66, 2980–2983. Coe, B.J., Chamberlain, M.C., Essex-Lopresti, J.P., Gaines, S., Jeffery, J.C., Houbrechts, S., Persoons, A., 1997. Large molecular quadratic hyperpolarizabilities in donor/acceptorsubstituted trans-tetraammineruthenium(II) complexes. Inorg. Chem. 36, 3284–3292. Cole, J.M., 2003. Organic materials for second-harmonic generation: advances in relating structure to function. Philos. Trans. A Math. Phys. Eng. Sci. 361, 2751–2770. Cole, J.M., Ashcroft, C.M., 2018. Molecular design rules for second-order non-linear optical organometallic complexes: a generic classification. (to be published). Cole, J.M., Hickstein, D.D., 2013. Molecular origins of nonlinear optical activity in zinc tris(thiourea)sulfate revealed by high-resolution x-ray diffraction data and ab initio calculations. Phys. Rev. B. 88. 184105. Cole, J.M., Lin, T.-C., Ashcroft, C.M., Perez-Moreno, J., Tan, Y., Venkatesan, P., Higginbotham, A.P., Pattison, P., Edwards, A.J., Piltz, R.O., Clays, K., Ilangovan, A., 2016. Relating the structure of geminal amido esters to their molecular hyperpolarizability. J. Phys. Chem. C 120, 29439–29448. Cole, J.M., Weng, Z.F., 2010. Discovery of high-performance organic non-linear optical molecules by systematic ‘smart material’ design strategies. In: Multi-Functional Materials and Structures III, Advanced Materials Research. Trans Tech Publications, Switzerland, pp. 959–962. Evans, O.R., Lin, W., 2002. Crystal engineering of NLO materials based on metal organic coordination networks. Acc. Chem. Res. 35, 511–522. Franken, P.A., Hill, A.E., Peters, C.W., Weinreich, G., 1961. Generation of optical harmonics. Phys. Rev. Lett. 7, 118–119. Hickstein, D.D., Cole, J.M., Turner, M.J., Jayatilaka, D., 2013. Modeling electron density distributions from X-ray diffraction to derive optical properties: constrained wavefunction versus multipole refinement. J. Chem. Phys. 139. 064108. Higginbotham, A.P., Cole, J.M., Blood-Forsythe, M.A., Hickstein, D.D., 2012. Identifying and evaluating organic nonlinear optical materials via molecular moments. J. Appl. Phys. 111. 033512.

174

Handbook of Organic Materials for Electronic and Photonic Devices

Jayatilaka, D., Grimwood, D.J., 2001. Wavefunctions derived from experiment. I. Motivation and theory. Acta Crystallogr. A 57, 76–86. Kaatz, P., Shelton, D.P., 1996. Polarized hyper-Rayleigh light scattering measurements of nonlinear optical chromophores. J. Chem. Phys. 105, 3918–3929. Kelly, K.L., Coronado, E., Zhao, L.L., Schatz, G.C., 2003. The optical properties of metal nanoparticles: the influence of size, shape, and dielectric environment. J. Phys. Chem. B 107, 668–677. Knoppe, S., Vanbel, M., van Cleuvenbergen, S., Vanpraet, L., B€ urgi, T., Verbiest, T., 2015. Nonlinear optical properties of thiolate-protected gold clusters. J. Phys. Chem. C 119, 6221–6226. Kodaira, T., Watanabe, A., Ito, O., Matsuda, M., Clays, K., Persoons, A., 1996. Evaluation of hyperpolarizability of nonlinear optical organic molecules by hyper-Rayleigh scattering. Jpn. J. Appl. Phys. 35, 6074. Kohn, W., Sham, L.J., 1965. Self-consistent equations including exchange and correlation effects. Phys. Rev. 140, A1133–A1138. Kuzyk, M.G., 2005. Compact sum-over-states expression without dipolar terms for calculating nonlinear susceptibilities. Phys. Rev. A. 72. 053819. Kuzyk, M.G., Dirk, C.W. (Eds.), 1998. Characterization Techniques and Tabulations for Organic Nonlinear Optical Materials, Optical Engineering. Marcel Dekker, New York. Lacroix, P.G., Malfant, I., Lepetit, C., 2016. Second-order nonlinear optics in coordination chemistry: an open door towards multi-functional materials and molecular switches. Coord. Chem. Rev. 308, 381–394. Ledoux, I., Zyss, J., Siegel, J.S., Brienne, J., Lehn, J.-M., 1990. Second-harmonic generation from non-dipolar non-centrosymmetric aromatic charge-transfer molecules. Chem. Phys. Lett. 172, 440–444. Lequan, M., Branger, C., Simon, J., Thami, T., Chauchard, E., Persoons, A., 1994. First hyperpolarizability of organotin compounds with Td symmetry. Adv. Mater. 6, 851–853. Levine, B.F., 1976. Donor-acceptor charge transfer contributions to the second order hyperpolarizability. Chem. Phys. Lett. 37, 516–520. Levine, B.F., Bethea, C.G., 1975. Second and third order hyperpolarizabilities of organic molecules. J. Chem. Phys. 63, 2666–2682. Lin, T.-C., Cole, J.M., Higginbotham, A.P., Edwards, A.J., Piltz, R.O., P
erez-Moreno, J., Seo, J.-Y., Lee, S.-C., Clays, K., Kwon, O.-P., 2013. Molecular origins of the highperformance nonlinear optical susceptibility in a phenolic polyene chromophore: electron density distributions, hydrogen bonding, and ab initio calculations. J. Phys. Chem. C 117, 9416–9430. Lin, W., Wang, Z., Ma, L., 1999. A novel octupolar metal- organic NLO material based on a chiral 2D coordination network. J. Am. Chem. Soc. 121, 11249–11250. Liu, Y., Li, G., Li, X., Cui, Y., 2007. Cation-dependent nonlinear optical behavior in an octupolar 3D anionic metal–organic open framework. Angew. Chem. 119, 6417–6420. Liu, Y., Xu, X., Zheng, F., Cui, Y., 2008. Chiral octupolar metal–organoboron NLO frameworks with (14,3) topology. Angew. Chem. 120, 4614–4617. Liu, Y., Xuan, W., Zhang, H., Cui, Y., 2009. Chirality- and threefold-symmetry-directed assembly of homochiral octupolar metal organoboron frameworks. Inorg. Chem. 48, 10018–10023. Long, N.J., 1995. Organometallic compounds for nonlinear optics—the search for En-lightenment!. Angew. Chem. Int. Ed. Eng. 34, 21–38. Marder, S.R., Gorman, C.B., Tiemann, B.G., Perry, J.W., Bourhill, G., Mansour, K., 1993. Relation between bond-length alternation and second electronic hyperpolarizability of conjugated organic molecules. Science 261, 186–189.

Molecular engineering of organic and organometallic second-order nonlinear optical materials

175

Marder, S.R., Perry, J.W., 1993. Molecular materials for second-order nonlinear optical applications. Adv. Mater. 5, 804–815. Mingabudinova, L.R., Vinogradov, V.V., Milichko, V.A., Hey-Hawkins, E., Vinogradov, A.V., 2016. Metal–organic frameworks as competitive materials for non-linear optics. Chem. Soc. Rev. 45, 5408–5431. Morley, J.O., Docherty, V.J., Pugh, D., 1987. Non-linear optical properties of organic molecules. Part 2. Effect of conjugation length and molecular volume on the calculated hyperpolarisabilities of polyphenyls and polyenes. J. Chem. Soc., Perkin Trans. 2, 1351–1355. Morrison, I.D., Denning, R.G., Laidlaw, W.M., Stammers, M.A., 1996. Measurement of first hyperpolarizabilities by hyper-Rayleigh scattering. Rev. Sci. Instrum. 67, 1445–1453. Nalwa, H.S., 1991. Organometallic materials for nonlinear optics. Appl. Organomet. Chem. 5, 349–377. Nappa, J., Russier-Antoine, I., Benichou, E., Jonin, C., Brevet, P.F., 2006. Second harmonic generation from small gold metallic particles: from the dipolar to the quadrupolar response. J. Chem. Phys. 125. 184712. Olbrechts, G., Strobbe, R., Clays, K., Persoons, A., 1998. High-frequency demodulation of multi-photon fluorescence in hyper-Rayleigh scattering. Rev. Sci. Instrum. 69, 2233–2241. Orr, B.J., Ward, J.F., 1971. Perturbation theory of the non-linear optical polarization of an isolated system. Mol. Phys. 20, 513–526. Oudar, J.L., 1977. Optical nonlinearities of conjugated molecules. Stilbene derivatives and highly polar aromatic compounds. J. Chem. Phys. 67, 446–457. Oudar, J.L., Chemla, D.S., 1977. Hyperpolarizabilities of the nitroanilines and their relations to the excited state dipole moment. J. Chem. Phys. 66, 2664–2668. Oudar, J.L., Zyss, J., 1982. Structural dependence of nonlinear-optical properties of methyl(2,4-dinitrophenyl)-aminopropanoate crystals. Phys. Rev. A 26, 2016–2027. Ray, P.C., 2010. Size and shape dependent second order nonlinear optical properties of nanomaterials and their application in biological and chemical sensing. Chem. Rev. 110, 5332–5365. Renouard, T., Bozec, H.L., Brasselet, S., Ledoux, I., Zyss, J., 1999. Tetrahedral bipyridyl copper(I) complexes: a new class of non-dipolar chromophore for nonlinear optics. Chem. Commun. 10, 871–872. Robinson, F.N.H., 1967. Nonlinear optical coefficients. Bell Syst. Tech. J. 46, 913–956. Ronchi, M., Pizzotti, M., Orbelli Biroli, A., Righetto, S., Ugo, R., Mussini, P., Cavazzini, M., Lucenti, E., Salsa, M., Fantucci, P., 2009. Second-order nonlinear optical (NLO) properties of a multichromophoric system based on an ensemble of four organic NLO chromophores nanoorganized on a cyclotetrasiloxane architecture. J. Phys. Chem. C 113, 2745–2760. S
en
echal, K., Maury, O., Le Bozec, H., Ledoux, I., Zyss, J., 2002. Zinc(II) as a versatile template for the design of dipolar and octupolar NLO-phores. J. Am. Chem. Soc. 124, 4560–4561. Stadler, S., Feiner, F., Br€auchle, C., Brandl, S., Gompper, R., 1995. Determination of the first hyperpolarizability of four octupolar molecules and their dipolar subunits via hyperRayleigh scattering in solution. Chem. Phys. Lett. 245, 292–296. Tykwinski, R.R., Gubler, U., Martin, R.E., Diederich, F., Bosshard, C., G€ unter, P., 1998. Structure  property relationships in third-order nonlinear optical chromophores. J. Phys. Chem. B 102, 4451–4465. Verbiest, T., Clays, K., Samyn, C., Wolff, J., Reinhoudt, D., Persoons, A., 1994. Investigations of the hyperpolarizability in organic molecules from dipolar to octopolar systems. J. Am. Chem. Soc. 116, 9320–9323.

176

Handbook of Organic Materials for Electronic and Photonic Devices

Wang, C., Zhang, T., Lin, W., 2012. Rational synthesis of noncentrosymmetric metal–organic frameworks for second-order nonlinear optics. Chem. Rev. 112, 1084–1104. Woi
nska, M., Jayatilaka, D., Dittrich, B., Flaig, R., Luger, P., Woz´niak, K., Dominiak, P.M., Grabowsky, S., 2017. Validation of X-ray wavefunction refinement. ChemPhysChem 18 (23), 3334–3351. Wostyn, K., Binnemans, K., Clays, K., Persoons, A., 2001. Hyper-Rayleigh scattering in the Fourier domain for higher precision: correcting for multiphoton fluorescence with demodulation and phase data. Rev. Sci. Instrum. 72, 3215–3220. Zhou, J., Kuzyk, M.G., 2008. Intrinsic hyperpolarizabilities as a figure of merit for electro-optic molecules. J. Phys. Chem. C 112, 7978–7982. Zyss, J., Berthier, G., 1982. Nonlinear optical properties of organic crystals with hydrogenbonded molecular units: the case of urea. J. Chem. Phys. 77, 3635–3653. Zyss, J., Chemla, D.S., Nicoud, J.F., 1981. Demonstration of efficient nonlinear optical crystals with vanishing molecular dipole moment: second-harmonic generation in 3-methyl-4-nitropyridine-1-oxide. J. Chem. Phys. 74, 4800–4811. Zyss, J., Dhenaut, C., Chauvan, T., Ledoux, I., 1993. Quadratic nonlinear susceptibility of octupolar chiral ions. Chem. Phys. Lett. 206, 409–414. Zyss, J., Ledoux, I., 1994. Nonlinear optics in multipolar media: theory and experiments. Chem. Rev. 94, 77–105.

Further reading Hu, S., Zou, H.-H., Zeng, M.-H., Wang, Q.-X., Liang, H., 2008. Molecular packing variation of crimpled 2D layers and 3D uncommon 65.8 topology: effect of ligand on the construction of metal  quinoline-6-carboxylate polymers. Cryst. Growth Des. 8, 2346–2351. Keller, B.A., Kn€opfle, G., Pr^etre, P., Bosshard, C., G€unter, P., 1993. Second-order nonlinear optical properties of dimethylaminobenzylidene-1,3-indandione (DABI) measured by EFISH in dioxane. Mol. Eng. 2, 325–338. Moylan, C.R., McNelis, B.J., Nathan, L.C., Marques, M.A., Hermstad, E.L., Brichler, B.A., 2004. Challenging the auxiliary donor effect on molecular hyperpolarizability in thiophene-containing nonlinear chromophores: X-ray crystallographic and optical measurements on two new isomeric chromophores. J. Organomet. Chem. 69, 8239–8243. Paley, M.S., Harris, J.M., Looser, H., Baumert, J.C., Bjorklund, G.C., Jundt, D., Twieg, R.J., 1989. A solvatochromic method for determining second-order polarizabilities of organic molecules. J. Organomet. Chem. 54, 3774–3778. Szablewski, M., Thomas, P.R., Thornton, A., Bloor, D., Cross, G.H., Cole, J.M., Howard, J.A.K., Malagoli, M., Meyers, F., Br
edas, J.-L., Wenseleers, W., Goovaerts, E., 1997. Highly dipolar, optically nonlinear adducts of tetracyano-p-quinodimethane: synthesis, physical characterization, and theoretical aspects. J. Am. Chem. Soc. 119, 3144–3154.

Molecular crystals and thin films for photonics

6

Mojca Jazbinsek*, Peter Gu€nter† *Zurich University of Applied Sciences (ZHAW), Zurich, Switzerland, †ETH Zurich and Rainbow Photonics AG, Zurich, Switzerland

6.1

Introduction

Organic materials with second-order nonlinear optical (NLO) functionality are promising for various photonic applications, including light-frequency conversion, terahertzwave generation, electric-field detection, and electro-optic (EO) modulation. These materials are based on NLO molecules (chromophores) with a high molecular nonlinearity, which are most often dipolar. To achieve a macroscopic second-order NLO response, such molecules need to be arranged in a noncentrosymmetric way in a material. This can be done by incorporating the chromophores in a polymer matrix and subsequent electric-field poling, molecular self-assembly into amorphous acentric structures, or self-assembly into single-crystalline acentric structures. This chapter is devoted to single-crystalline organic second-order NLO materials, which are attractive because of their high thermal and photochemical stability due to their highly stable chromophore packing, but they also involve challenging crystal growth in either bulk or thinfilm growth. We discuss state-of-the-art materials, molecular and crystal engineering approaches, as well as processing in bulk and thin-film single-crystalline forms. We also present the most promising photonic applications of single-crystalline organic NLO materials, including integrated EO devices, terahertz-wave generation, and terahertzwave detection.

6.2

Second-order NLO organic crystals

6.2.1 Microscopic and macroscopic NLO effects in organic materials The conventional design of organic NLO molecules is for crystals similar to poled polymers and is commonly based on π-conjugated polar chromophores with an asymmetric response to an external electric field (Dalton et al., 2010, 2015; Zyss, 1994; Bosshard et al., 1995, 2000; Bosshard and Gunter, 1997; Kuzyk, 2000; Clays and Coe, 2003). π-conjugated structures are regions of delocalized electronic charge distribution resulting from the overlap of π-orbitals. The electron distribution can be distorted by substituents (electron donor and acceptor groups) at both sides of the π-conjugated system as illustrated in Fig. 6.1A. Upon applying an external electric Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00006-1 © 2019 Elsevier Ltd. All rights reserved.

178

Handbook of Organic Materials for Electronic and Photonic Devices

1

Induced dipole

1

2= (2)

fs pulse

Polar axis k D

kz

kz

A A

3

(

Main chargetransfer axis z

3

3

2

(0, ,

)

(

(

1+ 1,

(

Second-harmonic generation (SHG)

2

( = 0)

, + , 3=

Sum-frequency generation (SFG)

1 1, 3)

50

274

mm2

500

d33 (1064 nm) ¼ 234 d32 (1064 nm) ¼ 15.6

110

mm2

600

d33 (1907 nm) ¼ 120 r33 (633 nm) ¼ 109 r33 (1319 nm) ¼ 52 r23 (1319 nm) ¼ 30 r13 (1319 nm) ¼ 6.8

212

PF 6

DAPSH(trans-4-dimethylaminoN-phenyl-stilbazolium hexafluorophosphate) (Coe et al., 2002, 2003; Figi et al., 2008b) O

CH3

HO CH 3

H3C

N+

SO3

H3C

CH3

HMQ-TMS(2-(4-hydroxy-3methoxystyryl)-1methylquinolinium 2,4,6trimethylbenzenesulfonate) ( Jeong et al., 2013; Kang et al., 2018) H3C CH2 NH

NO2

BNA(N-benzyl-2-methyl-4nitroaniline) (Fujiwara et al., 2007, 2006; Hashimoto et al., 1997) NC

HO

CN

CH3 CH3

OH1(2-(3-(4-hydroxystyryl)-5,5dimethylcyclohex-2-enylidene) malononitrile) (Kwon et al., 2008; Hunziker et al., 2008b) Notes: λc is the cut-off wavelength in the bulk, d is the NLO coefficient, r is the EO coefficient, and Tm is the melting point.

184

Handbook of Organic Materials for Electronic and Photonic Devices

6.2.3 Examples of organic NLO ionic crystals DAST (4-N,N-dimethylamino-40 -N0 -methyl-stilbazolium tosylate) is the most well known and widely investigated organic EO crystal. DAST was first reported in 1989 by Marder et al. (1989) and Nakanishi et al. (1989) and is still recognized as the state-of-the-art NLO crystal. High optical quality and large DAST crystals can be grown from methanol solution by the slow cooling method (Pan et al., 1996b; Ruiz et al., 2008) and were used for accurate determination of its dielectric, linear, and NLO properties. The reasons for the continuing interest in obtaining high-quality DAST crystals are the high second-order NLO and the EO coefficients, being, respectively, 10 times and twice as large as those of the inorganic standard LiNbO3, in combination with a low dielectric constant, which allows high-speed EO applications and broadband terahertz-wave generation. DAST is an organic salt that consists of a positively charged stilbazolium cation and a negatively charged tosylate anion as shown in Fig. 6.2A. The stilbazolium cation is one of the most efficient NLO-active chromophores that pack in an acentric structure, whereas the counterion tosylate is used to promote noncentrosymmetric crystallization (Marder et al., 1989, 1994). The structure of DAST is shown in Fig. 6.2B and C. The chromophores are packed with their main charge-transfer axis oriented at about θ ¼ 20 degrees with respect to the polar axis a, resulting in a high-order parameter of cos 3 θ ¼ 0:83, which is close to the optimum for the EO and terahertz-wave applications. As a result of the highly ordered packing of highly polarizable constituting molecules, DAST crystals are strongly anisotropic, with a refractive index difference n1  n2 > 0.5 in the visible and infrared (IR) wavelength range; see Fig. 6.3A (Pan et al., 1996a,b). DAST crystals are well suited for applications in telecommunications, with a material absorption that is smaller than 1 cm1 at 1.3 and 1.55 μm wavelengths. The dielectric constants of DAST in the low-frequency range, below acoustic and optical lattice vibrations, were determined as ET1 ¼ 5:2  0:4, ET2 ¼ 4:1  0:4, and

c b H 3C

N

CH3

N H 3C H 3C

SO3

c

(A)

(B) b

a

a

(C)

Fig. 6.2 (A) Molecular units of the ionic DAST crystal. The positively charged, NLO-active chromophore methyl-stilbazolium and the negatively charged tosylate. (B and C) X-ray structure of the ionic DAST crystal with the point group symmetry m showing molecules from one unit cell, projected along the crystallographic axes b and c; hydrogen atoms have been omitted for clarity.

Refractive index n

2.6 2.4

n1

2.2 2.0 1.8 1.6

(A)

n2 n3

600 800 1000 1200 1400 1600 1800 2000

Wavelength (nm)

185

EO coefficient rijkT(pm/V)

Molecular crystals and thin films for photonics 150 125 100

(B)

75

r111

50

r221

25 0 700

r113 900

1100

1300

1500

Wavelength (nm)

Fig. 6.3 (A) Dispersion of refractive indices along the three dielectric axes x1, x2, and x3, where x2 is parallel to the crystallographic axis b and x1 is close to the crystallographic axis a ( Jazbinsek et al., 2008). (B) Dispersion of the largest EO tensor coefficients of DAST: r111, r221, and r113 (Pan et al., 1996a).

ET3 ¼ 3:0  0:3 (Pan et al., 1996a), and are considerably lower than inorganic EO materials, which is very important for high-speed EOs and phase-matched terahertz-wave generation and detection ( Jazbinsek et al., 2008). T The low-frequency (unclamped) EO coefficients rijk of DAST are shown in Fig. 6.3B. The experimentally measured dispersion was modeled by the two-level model according to Eq. (6.6) and is presented by solid curves in Fig. 6.3B (Pan et al., 1996a); the deviation at shorter wavelengths stems from resonance effects when approaching the absorption edge. DAST exhibits large EO coefficients and refractive indices, resulting in a high-EO figure of merit n31 r111 ¼ 455  80 pm/V at wavelength λ ¼ 1535 nm, and therefore, the reduced half-wave voltage vπ ¼ λ/(n3r) compares favorably with inorganic single crystals and poled EO polymers. Theoretical evaluations show that the upper limits of second-order optical nonlinearities of organic crystals have not been reached yet (Cole et al., 2015; Bosshard et al., 2000); therefore, the research and development of novel NLO organic crystals is continuing. Another reason for the new development is a desire for optical properties that would allow applications in various wavelength ranges of the available laser sources, such as optimal phase-matching ranges for terahertz-wave generation, as discussed later in this chapter. Several organic crystalline salts have been identified with similar or even superior NLO properties compared to DAST (Marder et al., 1994; Glavcheva et al., 2005; Yang et al., 2007a; Coe et al., 2002; Figi et al., 2008b; Kim et al., 2011; Kwon et al., 2012; Yin et al., 2012; Chen et al., 2015), also combined with better crystal growth possibilities (Yang et al., 2007b; Ogawa et al., 2008) or better crystal processing possibilities, such as the possibility of cleaving as-grown crystals to produce samples with a desired thickness and optical-quality surfaces without the need of expensive and time-consuming cutting and polishing techniques ( Jeong et al., 2013; Lee et al., 2017). It also has been shown that various crystal engineering approaches may lead to crystals with a desired final morphology (Lee et al., 2016b,c; Kim et al., 2016; Choi et al., 2017). For recent reviews on the development of organic salt materials, see Lee et al. (2016a) and Liu et al. (2016b).

186

Handbook of Organic Materials for Electronic and Photonic Devices

From the variety of new organic NLO salt crystals developed during the last decade, we here mention in particular two of them with very favorable crystal and optical properties, so they already have been used in several photonic application demonstrations: DSTMS (4-N,N-dimethylamino-40 -N0 -methyl-stilbazolium 2,4,6trimethylbenzenesulfonate) and HMQ-TMS (2-(4-hydroxy-3-methoxystyryl)1-methylquinolinium 2,4,6-trimethylbenzenesulfonate). DSTMS first was reported in 2007 (Yang et al., 2007b) and is a DAST derivative: it has the same NLO cation chromophore as DAST (see Fig. 6.2A); only the counteranion, which has negligible NLO properties compared to the cation, is slightly modified (from tosylate or 4-methylbenzenesulfonate to 2,4,6-trimethylbenzenesulfonate). The additional methyl groups on the anion of DSTMS have shown not to dramatically affect both the crystal structure and the NLO properties (Mutter et al., 2007a), but to have an important advantage considering the crystal growth possibilities; the solubility of DSTMS in methanol is about twice as that of DAST (Yang et al., 2007b). It has also been shown that the main terahertz phonon absorption peak in DSTMS is only about half as that of DAST (Stillhart et al., 2008). Due to these improvements, DSTMS already has been used by several research groups, especially related to terahertz photonic applications (Stillhart et al., 2008; Liu et al., 2014, 2017; Vicario et al., 2015a; Monoszlai et al., 2013; Somma et al., 2015; Finneran et al., 2016; Zhang et al., 2016b). The second ionic crystal, HMQ-TMS, has a different, quinolinium-based ionic core structure, which recently has been shown as very attractive, leading to acentric crystal structures with a relatively very high probability (Lee et al., 2016a). HMQ-TMS was first reported in 2013 ( Jeong et al., 2013) and has been used in various terahertz photonics schemes (Brunner et al., 2014; Lu et al., 2015; Vicario et al., 2015b; Lee et al., 2016c; Rovere et al., 2018; Kang et al., 2018). The NLO properties of HMQ-TMS may be comparable to or slightly lower than DAST and DSTMS (Lee et al., 2016a; Kang et al., 2018), but a big advantage of HMQ-TMS is a much simpler optical-quality sample processing possibility based on cleaving. HMQ-TMS has different optical and terahertz properties than DAST and DSTMS, so it may be preferred for particular applications, as discussed later in this chapter.

6.2.4 Examples of organic NLO nonionic crystals The chromophores for highly NLO crystals in general exhibit limited temperature stability. In the case of DAST (and similarly for other highly nonlinear organic salts), the chromophores start to decompose at about the melting temperature, which is for DAST at 256°C. Therefore, the processing possibilities of stilbazolium salts in most practical situations are limited to solution-based techniques. On the other hand, melt growth is very attractive for several reasons (e.g., higher growth rates, higher purity, and very attractive waveguide processing possibilities, as will be discussed in the following section). Until recently, short π-conjugated chromophores with relatively low melting temperatures (Tm < 150°C), but also relatively low first hyperpolarizabilities, such as COANP (2-cyclooctylamino-5-nitropyridine), were the only organic NLO crystals obtained by melt-growth techniques. Unfortunately, the EO figure of merit n3r of these crystals may be one order of magnitude smaller than for the best

Molecular crystals and thin films for photonics

187

stilbazolium salts. Therefore, to design organic EO materials with a broad spectrum of processing possibilities, the challenge is to simultaneously achieve high thermal stability and nonlinearity in one compound. To evade the nonlinearity-thermal stability trade-off of organic crystalline materials, different series of configurationally locked polyene chromophores have been developed (Kwon et al., 2006, 2008). Several of these chromophores crystallize in a noncentrosymmetric structure with a high powder test efficiency of the same order of magnitude as that of DAST. Configurationally locked polyene crystals DAT2 (2-(3-(2-(4-dimethylaminophenyl) vinyl)-5,5-dimethylcyclohex-2-enylidene)malononitrile) and OH1 (2-(3-(4hydroxystyryl)-5,5-dimethylcyclohex-2-enylidene)malononitrile) have been demonstrated as being especially promising for integrated EO applications (Figi et al., 2008a, 2010; Hunziker et al., 2008a; Jazbinsek et al., 2010). The big advantage of DAT2 is its excellent range of possibilities for thin-film processing from all solutions, vapor, and melt, although its EO properties are not as high as for DAST due to a less optimal crystalline packing (Kwon et al., 2006; Figi et al., 2008a). On the other hand, OH1 crystals show the EO effect as high as DAST crystals (Hunziker et al., 2008b). OH1 exhibits very good crystal processing possibilities from solution, for both highoptical-quality bulk crystals (Kwon et al., 2010), as well as large-area, single-crystalline thin films on substrates (Kwon et al., 2009). OH1 has been found to be particularly efficient for terahertz-wave generation (Brunner et al., 2008; Ruchert et al., 2012; Uchida et al., 2013, 2018; Stepanov et al., 2014; Majkic et al., 2014; Li et al., 2014; Vicario et al., 2015c; Shalaby and Christoph, 2015; Brenier, 2015; Ovchinnikov et al., 2016; Liu et al., 2016a). Another very promising, recently identified group of nonionic crystals are isoxazolone-based derivatives (Zhang et al., 2015, 2016a), which also have high thermal stability and show up to three times higher powder second-harmonicgeneration efficiency compared to OH1 and also have already been used for broadband terahertz-wave generation based on difference-frequency generation (Zhang et al., 2016a). Another nonionic crystal often used in applications is BNA (N-benzyl2-methyl-4-nitroaniline), first reported in 1997 (Hashimoto et al., 1997). This nitroaniline derivative is based on the so-called “yellow” nitroaniline crystals MNA (2-methyl-4-nitroaniline) developed in the early days of organic NLO research (Chemla and Zyss, 1987). Because of the yellow color of these crystals, one expects considerably lower NLO coefficients compared to “red” materials such as DAST, DSTMS, and OH1 due to the common nonlinearity-transparency trade-off (see Eq. 6.8). The measured NLO coefficients of BNA are surprisingly high (see Table 6.1), with the measured diagonal NLO coefficient of d333 ¼ 234 pm/V at the fundamental wavelength of 1064 nm (Fujiwara et al., 2007), which is resonantly enhanced. From the measured dispersion of the refractive indices and using a twolevel dispersion model (see Eq. 6.7), we estimate d333  92 pm/V and d333  72 pm/V for BNA at 1500- and 1900-nm fundamental wavelength, respectively. Considering the two-level dispersion model for both the NLO and the EO coefficient, we can estimate the highest EO coefficient of BNA is r33 ¼ 24 pm/V at 1538 nm (Korn et al., 2014). Although some experimental results suggest different

188

Handbook of Organic Materials for Electronic and Photonic Devices

values of the corresponding NLO and EO coefficients (Bernerd et al., 2018; Korn et al., 2014; Wang et al., 2011), BNA crystal has been very successfully used for terahertz-wave generation applications, particularly in the near-infrared pump wavelength regime that is less optimal for high-nonlinearity “red” materials (Kuroyanagi et al., 2006; Miyamoto et al., 2008, 2009; Iwaszczuk et al., 2009; Notake et al., 2012a,b; Shalaby et al., 2016; Thirupugalmani et al., 2017).

6.2.5 Single-crystal growth: Solution, vapor, and melt-based techniques Crystallization of organic materials is based on solution growth, melt growth, or vapor growth. To produce either bulk, thin-film, or wire crystals, different (and in some cases rather complex) growth techniques are required. The choice of an appropriate technique depends on material properties, as well as the desired crystalline form. For many of the state of the art, highly nonlinear organic materials, such as DAST, DSTMS, DAPSH, HMQ-TMS, and OH1, best-quality bulk crystals are grown from solution. Most commonly, slow temperature lowering techniques or slow isothermal evaporation techniques are used, which can be combined with temperature gradients at the growth position. An example of the solution-growth process optimization for obtaining high-optical quality bulk crystals of DAST can be found in Ruiz et al. (2008). Compared to DAST, growth of DSTMS crystals and OH1 crystals from solution may be faster and easier, which is due to the favorable thermodynamic properties of these materials (Yang et al., 2007b; Kwon et al., 2010). Some materials (e.g., HMQTMS) also can be cleaved along a certain crystallographic plane after growth ( Jeong et al., 2013), which importantly reduces the postprocessing requirements for photonic applications. High-quality, single-crystalline thin films of highly NLO materials are essential for the fabrication of integrated photonic devices. If one should start from bulk crystals, then complicated, expensive, and time-consuming cutting, polishing, and structuring procedures are required to fabricate waveguiding devices. Obviously, thin films may be much more compatible with simpler and cheaper waveguiding structures for applications such as EO modulators. Various approaches have been investigated for the fabrication of single-crystalline films, using either solution-, melt-, or vapor-growth techniques; an overview of different approaches for thin-film fabrication is reviewed in more detail in Dalton et al. (2015) and Jazbinsek and Gunter (2011). One of the very attractive solutions for the integrated photonic devices presents a direct growth of the desired microstructures and nanostructures at a desired position. This can be done by first structuring standard inorganic templates, such as glass, silicon, electrodes, and other materials, with void structures at positions where active organic crystalline materials are desired. This method was demonstrated by using the melt-processable material DAT2 (Figi et al., 2008a), the small chromophore COANP (Figi et al., 2009), and BNA (Figi et al., 2011; Korn et al., 2014); these materials were chosen because of their favorable growth characteristics from melt as well

Molecular crystals and thin films for photonics

5 mm

(A)

189

5 mm

(B)

(C)

Fig. 6.4 Examples of the grown crystals: (A) bulk OH1 single crystal that is 4 mm thick (Kwon et al., 2010), (B) thin-film DSTMS single crystal that is 20 μm thick (Yang et al., 2007b), and (C) wire DAT2 single crystal that is 25 nm thick (Figi et al., 2008a).

as their tendency for thin-film formation. By this method, several-millimeter-long, single-crystalline wires with a thickness up to several micrometers down to below 30 nm have been obtained (Figi et al., 2008a). Fig. 6.4 shows some examples of single-crystalline organic bulk, thin-film, and wire crystals.

6.3

Integrated EO applications

The EO effect describes the change of the refractive index upon application of a static (or quasistatic) field and is defined by Eq. (6.3), which gives, assuming Δn ≪ n, the following change of the refractive index Δn for light polarization parallel to the direction of the applied field E: 1 Δn ¼  n3 rE, 2

(6.9)

where r ¼ r333 for the electric field and polarization along the x3 material direction. This gives the following phase change Δϕ for light traveling a distance L in the material upon application of the voltage V over the (electrode) distance D: Δϕ ¼

2π πn3 r L ΔnL ¼  V: λ λ D

(6.10)

For the EO applications, an important parameter is the so-called half-wave voltage Vπ , which is the voltage required to change the phase of the optical field by π: Vπ ¼

λ D , n3 r L

(6.11)

and critically depends on the material EO figure of merit FMEO ¼ n3r, as well as the configuration of the EO material by the geometrical factor D/L. Waveguide

190

Handbook of Organic Materials for Electronic and Photonic Devices

configurations allow small electrode distances D and relatively large propagation distances L. This can decrease the half-wave voltages by more than three orders of magnitude compared to bulk materials, from kilovolts to less than 1 V for best EO materials. Therefore, integrated optics is the best solution to achieve the light modulation and switching needed in telecommunications. For high data transmission rates or fast modulation and switching, the applied field or voltage V is modulated at high frequencies, up to gigahertz (even 100 GHz). Therefore, for larger propagation distances L, the applied field already changes during light propagation, which considerably reduces the final phase change Δϕ. The solution is in the so-called traveling-wave modulators, in which the electric field travels with the optical field. In organic materials, the electric wave travels at about the same speed as the optical wave due to the low dielectric constant in the low-frequency regime E  n2, which is not the case for most inorganic EO materials, where E ≫ n2. This kind of phase matching is important when building high-frequency EO modulators. The low dielectric constant of organic materials also will decrease the power requirement of the EO modulators. Another advantage of organic over inorganic materials is the almost constant EO coefficient over an extremely wide frequency range. This property is essential when building broadband EO modulators and field detectors. The interest in organic crystals stems from these advantages compared to inorganic materials, as well as the fact that the long-term orientational stability and photochemical stability, as well as the optical quality of molecular crystals, may be significantly superior to those of polymers (Rezzonico et al., 2008b). However, compared to polymers, processing of organic EO crystals in thin films and waveguides needed for integrated optics is much more challenging. In the following section, we summarize several structuring techniques that have been used to build waveguides and optical modulators based on organic EO crystals.

6.3.1 Structuring possibilities for organic crystals As for general optical waveguides, organic crystals have to be structured with a submicron precision, so that a suitable refractive index contrast for optical waveguiding is achieved. Although for optical waveguiding even very small refractive index changes of materials on the order of Δn  103 may be sufficient, for small waveguides needed for large-scale integration, as well as for reducing the half-wave voltage of EO modulators, a larger index contrast is desired. For example, to fabricate microring resonators (Little et al., 1997; Gheorma and Osgood, 2002; Rabiei et al., 2002; Rezzonico et al., 2008a) with a small radius below 10 μm, the refractive index of the guiding medium should be at least about Δn  0.5 larger than the surrounding medium to avoid high losses. Organic EO crystals have a relatively high refractive index compared to poled polymers, which can reach the values of inorganic ferroelectric crystals such as LiNbO3 (n  2.2). This basically allows very efficient high-index contrast waveguiding with respect to substrate materials such as silicon dioxide (SiO2). Organic EO crystals are also strongly anisotropic, with birefringence as high as Δn > 0.5 at nonresonant wavelengths, which should be taken into account when designing the waveguides. The

Molecular crystals and thin films for photonics

191

particular best orientation of the waveguides with respect to the propagation direction and optical and electric field orientation depends on the particular tensor properties of the material. Organic crystals are also suitable as active cladding materials to high-index silicon photonic passive waveguides (nSi  3.5), which can result in very compact EO modulators with high figures of merit, if organic crystals can be oriented in a suitable way ( Jazbinsek et al., 2010; Figi et al., 2011; Korn et al., 2014). The main challenges to build integrated EO modulators based on organic crystals are related to their processing possibilities: the organic crystal should be deposited on appropriate substrate materials and in a desired orientation to achieve planar light confinement, and then structured with an appropriate technique to achieve horizontal light confinement. Several approaches and techniques have been developed to fabricate optical waveguides in organic EO crystals. We can distinguish photolithography (Tsuda et al., 1992; Takayama et al., 2001; Kaino et al., 2002; Jazbinsek et al., 2008; Hunziker et al., 2008a), photobleaching (Cai et al., 2003; Mutter et al., 2003; Kaino et al., 2002), femtosecond laser ablation (Dittrich et al., 2003), ion implantation (Mutter et al., 2007b, 2008), electron-beam irradiation (Mutter et al., 2007c), and direct deposition into prestructured inorganic templates (Geis et al., 2004; Figi et al., 2008a, 2010, 2011; Jazbinsek et al., 2010; Korn et al., 2014). A more detailed overview of organic, single-crystalline waveguides and modulators with examples can be found in Jazbinsek and Gunter (2011) and Dalton et al. (2015). Currently, the most promising organic crystals, several of which have been already successfully used for building prototype integrated EO modulators, are listed in Table 6.1. Some of the above-mentioned crystal-processing techniques allow even more complex waveguiding structures than simple, straight waveguides. For example, by melt capillary growth, single-crystalline organic EO microring-resonator filters and modulators were demonstrated (Figi et al., 2009). In this case, the organic material COANP (2-cyclooctylamino-5-nitropyridine) with very good melt-crystallization properties and a moderate EO coefficient r33 ¼ 15  2 pm/V at 633 nm was employed. A top-view transmission microscope image of a COANP crystal grown in a microring resonator channel waveguide is depicted in Fig. 6.5A. Very high single-crystalline quality of these waveguides was confirmed by optical waveguiding characterization. Typical devices fabricated showed almost perfectly symmetric high extinction ratio resonance peaks of about 10 dB, ring losses α ¼ 12  0.3 dB/cm, and a finesse F ¼ 6.2  0.2. The measured TE spectrum of the racetrack resonator shown in Fig. 6.5A showed a Δλ ¼ 110 pm shift in response to an applied voltage of 100 V, corresponding to a frequency tunability of 0.11 GHz/V, which is comparable to what has been reported for ion-sliced LiNbO3 microring resonators (Guarino et al., 2007). A great improvement in performance is expected if materials with state-of-the-art EO figures of merit (n3r of DAST or OH1 are more than one order of magnitude higher than for COANP) and higher index contrast (Δn with respect to borosilicate is at 1.55 μm about 0.15 for COANP and almost 0.7 for OH1 and DAST) can be used for melt growth. Using melt capillary growth of organic NLO crystals, it also has been possible to demonstrate first silicon-organic hybrid (SOH) slot waveguide devices, where a small slot (50–200 nm) in a silicon photonic waveguide is filled with an organic NLO

192

Handbook of Organic Materials for Electronic and Photonic Devices

~140 nm thick COANP nanosheet

0

Floating electrode

Electrode pads

(A)

Dl TE ~ 110 pm

Split-ring electrode

Intensity (dB)

COANP waveguide and microring

–2 –4 0V 100 V 200 V

–6 200 µm

(B)

1573.2

1574.0

1574.8

Wavelength (nm)

Fig. 6.5 (A) Transmission microscope image between crossed polarizers of a COANP waveguide with a racetrack microring resonator grown by the melt capillary method in prefabricated channels. (B) Resonance curve of a TE mode at a wavelength around 1.574 μm (solid line); the dashed and dotted lines are the corresponding electro-optically shifted curves by applying 100 and 200 V to the device electrodes (Figi et al., 2009).

Fig. 6.6 SOH Mach-Zehnder integrated optical device employing an organic-crystal cladding. (A) Cross-section of strip-loaded silicon slot waveguide. Cross-sections of (B) the dominant optical E-field component in a quasi-TE mode and (C) the dominant electric RF field component in the strip-loaded slot waveguide. (D) Mach-Zehnder interferometer modulator with strip-loaded optical slot waveguides, electrical connections, and ground-signal-ground electrodes (GSGs) of coplanar transmission line (Korn et al., 2014).

crystal, providing an EO functionality to an otherwise passive silicon-on-insulator (SOI) waveguide (Figi et al., 2011; Korn et al., 2014). With a silicon slot waveguide (as shown in Fig. 6.6), filled with an organic crystal BNA, it was possible to demonstrate high-speed modulation at 12.5 Gbit/s at a half-wave voltage-length product of VπL ¼ 12 Vmm (Korn et al., 2014). Also, for these devices, a great improvement is expected if crystalline materials with state-of-the-art EO figures of merit can be employed. Although the present performance does not yet reach the best figures achieved with much more widely employed EO polymers (Heni et al., 2017), the approach remains attractive because of the high long-term thermal and photochemical stability of organic crystalline compared to polymeric materials.

Molecular crystals and thin films for photonics

6.4

193

Terahertz-wave generation and detection with organic crystals

The interest in generating terahertz waves stems from the unique interactions of these rays with matter, which can be exploited in a number of applications. For example, optical phonon resonances of crystalline materials and part of vibrational and rotational excitations of molecules are in the terahertz range, which makes terahertz radiation very interesting for spectroscopy and material identification. Other applications include nondestructive material testing and imaging, various research material investigations such as carrier dynamics in semiconductors with subpicosecond time resolution, medical diagnostics, and pharmaceutical characterization (Dhillon et al., 2017; Hwang et al., 2015).

6.4.1 Terahertz-wave generation by difference-frequency generation Generating terahertz waves by optical difference-frequency generation (DFG) requires a pump source consisting of two frequencies ω1 and ω2 that are very close to each other, so that their difference frequency lies in the terahertz range: ωTHz ¼ ω1  ω2. The phase-matching condition should be satisfied as well: Δk ¼ kTHz  ðk1  k2 Þ:

(6.12)

For collinear difference-frequency generation, and assuming that the optical frequencies are close together, so that in first approximation the dispersion in the optical range can be considered as n2 ¼ n1 + (∂n/∂λ)Δλ, Δλ ¼ λ2  λ1, this leads to Δk ¼

ωTHz ðnTHz  ng Þ c

(6.13)

and the following coherence length for terahertz generation: lc ¼

λTHz , 2ðnTHz  ng Þ

(6.14)

where ng ¼ n  (∂n/∂λ)λ is the group index of the optical wave. Eq. (6.14) is valid for a relatively small dispersion in the optical range (i.e., up to several terahertz if we use infrared pump light). For larger terahertz frequencies, it should be calculated as lc ¼

  1 nTHz n1 n2 1  + : 2 λTHz λ1 λ2

(6.15)

For the efficient generation of terahertz waves by difference-frequency generation, besides a high second-order NLO susceptibility, the most important parameter is the low refractive index mismatch Δn ¼ nTHz  ng between the generated terahertz

194

Handbook of Organic Materials for Electronic and Photonic Devices

and the pump optical waves. This is where organic materials are of a big advantage compared to standard inorganic materials such as LiNbO3. Because of the relatively low contribution of the lattice phonon vibrations to the dielectric constant, the dispersion of the refractive index between the optical and the terahertz frequency range is low, and therefore, the phase-matching condition is almost naturally satisfied, while for inorganic materials such as LiNbO3, special phase-matching configurations are needed. For difference-frequency generation in the case of phase-matching and neglecting the pump-light absorption, the visible-to-terahertz conversion efficiency is given by (Sutherland, 2003) ηTHz ¼

2 ω2THz dTHz L2 I0 sinh 2 ðαTHz L=4Þ exp ðα L=2Þ , THz 2E0 c3 n20 nTHz ðαTHz L=4Þ2

(6.16)

where 1 dTHz ¼ n4o r 4

(6.17)

is the NLO susceptibility for terahertz-wave generation, ωTHz the angular frequency of the generated terahertz wave, L the length of the terahertz-generation materials, I0 the pump intensity, αTHz the absorption constant at the terahertz frequency, r the corresponding EO coefficient, and n0 and nTHz the refractive indices at the pump optical and the generated terahertz frequencies, respectively. Besides phase-matching and minimal terahertz absorption, the main material figure of merit for terahertz generation (FMTHz) according to Eq. (6.16) is FMTHz ¼

2 dTHz n6 r 2 ¼ 0 : 2 n0 nTHz 16nTHz

(6.18)

Table 6.2 shows most of these parameters for a series of inorganic and organic crystals, as well as for an EO polymer. As can be seen in this table, the organic crystals OH1, DSTMS, and OH1 show the largest figure of merit and can be phase-matched using pump lasers at telecommunication wavelengths 1.3–1.55 μm. OH1 also shows a very small absorption constant at terahertz frequencies below 3 THz, thus allowing one to use large interaction lengths (Brunner et al., 2008). The optical damage threshold of these organic crystals mainly depends on the optical quality, both of the bulk crystal quality and the quality of surface polishing. Very slow cooling growth with high temperature stability of 0.002°C has to be used for high-damage-threshold materials reaching Idamage > 150 GW/cm2 for 150-fs pulses at 1550 nm (www. rainbowphotonics.com). By using the process of difference-frequency generation, narrowband, frequencytunable terahertz wave generation in DAST crystals under or close to the phase-matched conditions has been demonstrated (Kawase et al., 1999, 2000, 2003; Taniuchi et al., 2000, 2004a,b,c, 2005; Takahashi et al., 2006; Satoh et al.,

DAST DSTMS OH1 LAPCe GaAs ZnTe InP GaP ZnS CdTe LiNbO3

no

ng

nTHz

r (pm/V)

dTHza (pm/V)

FMTHz (pm/V)2

2.13 2.13 2.16 1.6 3.37 2.83 3.2 3.12 2.3 2.82 2.2

2.3c 2.3 2.33 1.8 3.61 2.18 3.16

2.26 2.26 2.28d 1.7 3.63 3.16 3.54f 3.34 2.88 3.24 4.96

47 49 52 52 1.6 4 1.45 1 1.5 6.8 28

240 250 280 85 52 64 38 24 10 110 160

5600 6100 7400 1700 66 160 40 17 7 470 1100

2.18

b

vphonon (THz)

αTHz (cm21)

λ (nm)

22 22 8 >17 7.6 5.3 10 10.8 9.8

20 15 2f 15 0.5 1.3

1500 1500 1350 1500 1560 840

0.2

1000

Molecular crystals and thin films for photonics

Table 6.2 Organic and inorganic NLO materials that have been investigated for optical-to-terahertz frequency conversion and their most relevant parameters

4.8 17

Notes: Where possible, the parameters close to the velocity-matched optical wavelengths and terahertz frequencies are given. Refractive index n0 at the pump optical wavelength λ; group index ng at λ; refractive index nTHz in the terahertz frequency range; the electro-optic coefficient r; the susceptibility dTHz for terahertz-wave generation; figure of merit FMTHz for terahertz generation by OR; optical phonon frequency of the material νphonon in the terahertz range; the absorption αTHz in the terahertz frequency range. a dTHz ¼ 14 n40 r. d2

n6 r 2

FMTHz ¼ n2 nTHz ¼ 16n0 THz . THz

b

0

v > 1.5 THz. v < 1.9 THz. LAPC guest-host polymer (Zheng et al., 2007). f v  1 THz. c

d e

195

196

Handbook of Organic Materials for Electronic and Photonic Devices

THz-wave peak power (W)

2010; Tang et al., 2011; Koichi et al., 2011; Liu and Merkt, 2008; Liu et al., 2009; Notake et al., 2012b; Dolasinski et al., 2015; Zhong et al., 2017; He et al., 2018). The accessible terahertz wavelength range is intrinsically limited by the absorption of the NLO material employed. While with inorganic materials it is difficult to generate waves above 5 THz efficiently due to strong optical phonon resonances, terahertz waves reaching frequencies of above 30 THz with a high peak power can be generated in DAST, as shown in Fig. 6.7 (Takahashi et al., 2006), as well as in other organic NLO crystals, such as OH1 (Uchida et al., 2011, 2013; Majkic et al., 2014; Liu et al., 2016a) and BNA (Miyamoto et al., 2009; Notake et al., 2012b). For the terahertz range larger than several terahertz and for the far-IR range, a slightly noncollinear difference-frequency generation geometry may be used to achieve exact phasematching conditions and improve the efficiency in a desired terahertz range (Satoh et al., 2010). Another possibility is to combine the DFG output generated in various organic crystals to obtain a more homogeneous terahertz spectrum, as demonstrated for DAST and BNA (Notake et al., 2012b). It is straightforward, therefore, to produce narrowband terahertz waves by difference-frequency generation in organic NLO crystals, which is phase-matched without a need for special crystal cuts, at least for some pump-wavelength ranges and thinner crystals (Kawase et al., 1999, 2000, 2003; Taniuchi et al., 2000, 2004a,b,c, 2005; Takahashi et al., 2006; Satoh et al., 2010; Tang et al., 2011; Koichi et al., 2011), but also specially cut to optimize the phase-matching, and therefore the efficiency (Liu and Merkt, 2008; Liu et al., 2009). It may be more challenging to develop an appropriate pump-beam pair with a desired separation in the terahertz frequency range. This is most often done by designing complex optical-parametric-oscillator (OPO) systems, possibly

10 1 0.1 0.01

0

5

10

15 20 25 Frequency (THz)

30

35

Fig. 6.7 Terahertz-wave peak power spectrum generated in a 1-mm-thick DAST crystal using difference frequency generation. A coherent, widely tunable terahertz wave in the broad range of 2–31.5 THz with a high peak power has been generated (Takahashi et al., 2006). (From Takahashi, Y., Adachi, H., Taniuchi, T., Takagi, M., Hosokawa, Y., Onzuka, S., Brahadeeswaran, S., Yoshimura, M., Mori, Y., Masuhara, H., Sasaki, T., Nakanishi, H., 2006. Organic nonlinear optical DAST crystals for electro-optic measurement and terahertz wave generation. J. Photochem. Photobiol. A Chem. 183 (3), 247–252, reprinted with permission from Elsevier).

Molecular crystals and thin films for photonics

197

with a tuning option to tune the generated terahertz frequency. These systems are still very bulky and sensitive today. Much more compact generation systems can be achieved by using OR, as described in the following section.

6.4.2 Terahertz-wave generation by OR By using the process of OR, broadband terahertz radiation can be efficiently generated in noncentrosymmetric NLO crystals pumped by femtosecond pulses. An ultrashort laser pulse (10–200 fs) induces a quasi-static polarization in such materials through OR, which follows the amplitude of the pump pulse and thus acts as a source for the terahertz pulse. In other words, a short laser pulse has an intrinsically broad bandwidth (i.e., a laser beam with a pulse length of 10 or 200 fs has a bandwidth of roughly 40 or 2 THz, respectively, depending on the pulse shape). Different frequency components in such a pulse can mix with each other in a nonlinear crystal by differencefrequency generation, producing a broadband terahertz wave. In the simplest approximation (nondepleted pump approximation, negligible terahertz absorption, planewave approximation, and at phase-matching conditions), the intensity of the generated terahertz intensity is given by (Schneider et al., 2006a) 1 IðωTHz , LÞ ¼ E0 cnTHz jEðωTHz ,LÞj2 2 

(6.19)

2 dTHz I02 ω2THz L2 2E0 c3 n20 nTHz

(6.20)

and the conversion efficiency by η¼

2 IðωTHz , LÞ dTHz I0 ω2THz L2 1 ¼ ¼ FMTHz ω2THz I0 L2 , I0 2E0 c3 n20 nTHz 2E0 c3

(6.21)

where the same figure of merit FMTHz as the one listed in Table 6.2 for differencefrequency generation is relevant. An analytical solution also has been derived for a more general case, considering both optical and terahertz-wave absorption and a general phase-mismatching term, but still considering a nondepleted pump approximation (Schneider et al., 2006a; Schneider, 2010). In the plane-wave approximation, the generated terahertz electric field at the angular terahertz frequency ωTHz ¼ ω pumped by the optical pulse with the central wavelength λ is then given by ω

ω

eið c nTHz + iαTHz =2ÞL  eið c ng + iαÞL  ETHz ðω, λÞ ¼ ω ω , ðnTHz  ng Þ + iðαTHz =2  αÞ ðnTHz + ng Þ + iðαTHz =2 + αÞ nE0 c3 c c (6.22) 2dTHz ω2 IðωÞ

where I(ω) is the Fourier transform of the pump optical pulse and L the material thickness. nTHz is the refractive index and αTHz is the absorption coefficient at the terahertz

198

Handbook of Organic Materials for Electronic and Photonic Devices

3.2

3.2

3.0

3.0

Optical group index

THz refractive index

angular frequency ω. n is the refractive index, ng is the optical group index, and α is the absorption coefficient at the optical wavelength λ. The second term in Eq. (6.22) has a unit of length, and its maximum is reached at exact phase-matching (i.e., velocity matching) nTHz ¼ ng in cases of negligible absorption. Its absolute value is refereed to as the effective generation length and is a useful figure to analyze the thickness dependence of the generated terahertz spectrum, which strongly depends on the material dispersion (Schneider et al., 2006a). A simpler figure of merit resulting from the expression (6.22), which neglects the optical and terahertz absorption effects, is again the coherence length lc, which is given by the same Eq. (6.14) as for DFG. Fig. 6.8 illustrates how velocity matching can be achieved in DAST. Due to the small absorption and the corresponding dispersion near 1.1 THz phase matching can be realized either for pump wavelengths between 830 and 1300 nm, yielding terahertz waves between 0.2 and 1 THz, or for pump waves with wavelengths between 1300 and 1700 nm for the generation of terahertz waves above 1.8 THz (Schneider et al., 2006a,b). The best conditions for velocity matching in DAST are also illustrated in Fig. 6.9A, which shows the coherence length lc according to Eq. (6.14) as a function of the pump optical wavelength λ and the generated terahertz frequency f ¼ ωTHz/(2π). We use the optical and terahertz properties of DAST, as shown in Fig. 6.8, to calculate lc as a function of λ and f. The generation of terahertz waves with frequencies around  1.1 THz is in DAST limited due to a transverse optical phonon (Walther et al., 2000). Other organic materials have different velocity-matching conditions and can be used for

2.8 2.6 2.4 2.2 2.0 1.8 1.6 0 1 2 3 4 5 6 7 8 9 10 11 12 Frequency (THz)

2.8 2.6 2.4 2.2 2.0 1.8 1.6

800

1000 1200 1400 Wavelength (nm)

1600

Fig. 6.8 Refractive index dispersion in the optical and the group index dispersion in the terahertz frequency range for DAST, indicating the optimal ranges for velocity-matching (shaded areas). The optical group index is calculated from the dispersion of n1 reported in Fig. 6.3A. Terahertz refractive indices nTHz of DAST along x1 have been reported by various groups using different approaches (see e.g., Cunningham et al., 2010; Ohno et al., 2010), but unfortunately the reported values nTHz may differ by more than 0.1, which is already an error too high to predict the best ranges for velocity matching. The curve for nTHz shown here is based on the measurements up to 2 THz from Schneider et al. (2006a) and above 2 THz, we use our own measurements obtained using a commercial terahertz spectrometer from Rainbow Photonics AG.

Molecular crystals and thin films for photonics

8

0.4

6 0.2

4 2

(A)

0.0

800

1000

1200

1400

1600

lc (mm)

OH1

0.6

10 8

0.4

6 4

0.2

2 0

(C)

800

1000

1200

1400

1600

Pump wavelength (nm)

lc (mm)

DSTMS

0.6

10 8

0.4

6 4

0.2

2 0

(B)

Pump wavelength (nm)

12

12

Generated frequency (THz)

0.6

10

0

Generated frequency (THz)

lc (mm)

DAST

800

1000

1200

1400

1600

0.0

Pump wavelength (nm) 12

Generated frequency (THz)

Generated frequency (THz)

12

199

lc (mm)

HMQ-TMS

0.6

10 8

0.4

6 4

0

0.0

(D)

0.2

2 800

1000

1200

1400

1600

0.0

Pump wavelength (nm)

Fig. 6.9 Coherence length for velocity matching lc for terahertz generation by using OR in (A) DAST, (B) DSTMS, (C) OH1, and (D) HMQ-TMS as a function of the pump optical wavelength λ and the generated terahertz frequency f. The white area represents the range with best velocity matching, leading to coherence lengths larger than 0.6 mm.

an available pump wavelength and the desired terahertz frequency range (Brunner et al., 2009; Miyamoto et al., 2009; Kim et al., 2011; Kwon et al., 2012; Jeong et al., 2013; Zhang et al., 2016a; Lee et al., 2017, 2018). Fig. 6.9B–D shows similar plots for selected newer materials, DSTMS, OH1, and HMQ-TMS. For DSTMS, we use the optical refractive indices from Mutter et al. (2007a) and terahertz refractive indices reported in Montemezzani et al. (2015) for the calculation of lc in Fig. 6.9B. In DSTMS, the optical phonon absorption near 1.1 THz is suppressed by a heavier counteranion and therefore, this material is superior to DAST within this frequency range (Stillhart et al., 2008). In a broader terahertz frequency range up to 12 THz (for which the terahertz refractive index data is available), DSTMS shows similar or slightly improved terahertz velocity-matching properties compared to DAST. For OH1, we use the optical refractive indices from Hunziker et al. (2008b) and terahertz refractive indices reported in Brunner et al. (2008) and Majkic et al. (2014) for the calculation of lc in Fig. 6.9C. OH1 crystal is a nonionic crystal based on hydrogen bonds and has even higher figure of merit for terahertzwave generation (see Table 6.2) compared to DAST and DSTMS and optimum

200

Handbook of Organic Materials for Electronic and Photonic Devices

velocity-matching between 1200 and 1460 nm for 0.3–2.5 THz, has no absorption at about 1 THz but at about 3 THz (Brunner et al., 2008). Above 3 THz, the optimal phase-matching conditions for OH1 are shifted to longer wavelengths, which is less optimal considering the available pump laser sources. We also include a more recently developed quinolinium-salt derivative HMQ-TMS (see Fig. 6.9D), for which the optical refractive indices are reported in Brunner et al. (2014) and terahertz refractive indices up to 1.5 THz in Brunner et al. (2008) and above 1.5 THz in Vicario et al. (2015b). We can see that the best phase-matching range is shifted toward shorter wavelengths compared to DAST and DSTMS, which can be an advantage for using certain pumpwavelength sources (Vicario et al., 2015b; Rovere et al., 2018). The velocity-matching plots based on the coherence length as presented in Fig. 6.9 do not take into account the influence of the optical and the terahertz-wave absorption. Based on the expression for the terahertz electric field (Eq. 6.22) one can derive an equivalent parameter lopt, the optimum crystal length (Schneider et al., 2006a), which considers both optical/terahertz absorption and optical/terahertz group/refractive indices. The corresponding plots of lopt as a function of f and λ for DAST, DSTMS, OH1, and HMQ-TMS can be found in Lee et al. (2016a). The main conclusions about the best pump wavelength for a certain organic crystal and a desired terahertz frequency range, however, remain similar to those resulting from Fig. 6.9. Organic crystals can be also used for coherent detection of terahertz waves. The most common technique for detecting terahertz waves based on NLO crystals is the so-called EO sampling, based on the change of the polarization state of the probe light in the detecting EO material due to the terahertz field, which is limited to materials with a relatively small or zero birefringence, such as ZnTe and GaP (Wu and Zhang, 1996). For highly birefringent materials, several alternative techniques have been proposed (Han et al., 2000; Schneider et al., 2004; Martin et al., 2010; Ilyakov et al., 2014). A relatively simple approach is the so-called terahertz-induced lensing technique (Schneider et al., 2004), in which the refractive index changes induced by a focused terahertz electric field due to the EO effect are exploited. The measured response depends on the spatial profile of the refractive index change, induced by the spatial profile of the terahertz wave given by 1 Δnðx, y,tÞ ¼  n3 rETHz ðx, y,tÞ: 2

(6.23)

This time-varying refractive index profile acts as a time-varying optical lens and leads to focusing/defocusing of the probe beam, which is can be detected by common optical detection techniques (Schneider et al., 2004; Schneider and Gunter, 2007). Using organic crystals as both terahertz generation and detection materials enables achieving an ultrabroad terahertz spectral bandwidth beyond 15 THz, even when using very compact and affordable femtosecond lasers, as shown in Fig. 6.10. Terahertz generators based on organic crystals are not only highly interesting because of the ultrahigh bandwidth possible, but also because of their high efficiency, leading to terahertz electric fields exceeding 150 MV/m (Vicario et al., 2015c) or even beyond 8 GV/m using special focusing optimization techniques (Shalaby and Hauri,

(A)

201

100

2 Power (normalized)

Signal amplitude (a.u.)

Molecular crystals and thin films for photonics

1 0

0

1

2

3

4 5 6 7 Time (ps)

8

9 10

10 10 10 10 10 10 10

(B)

0

5

10 15 20 Frequency (THz)

25

Fig. 6.10 (A) Terahertz time-domain signal and (B) the corresponding power spectrum achieved in a compact terahertz setup based on DSTMS as terahertz generator, another DSTMS as terahertz detector, and a pump laser with pulse length of 40 fs, 190 mW average power, 100 MHz repetition rate, and 1.55 μm central wavelength (Puc et al., 2018).

2015). These high-intensity terahertz sources offer a range of new possibilities for nonlinear terahertz photonics in various scientific and industrial applications.

6.5

Conclusions and outlook

In this chapter, we have presented the basics, present status, and potential of organic NLO crystalline materials for applications, particularly for integrated EO modulation and terahertz-wave generation. These materials are composed of highly polar chromophores with a highly asymmetric, ultrafast electronic response to external fields. They may exhibit high EO figures of merit of more than n3r ¼ 450 pm/V in the 1.55-μm telecommunication window, a low dielectric dispersion, and a high thermal stability of polar order, and therefore are promising for high-speed and highly efficient EO modulation, as well as phase-matched terahertz generation with high conversion efficiency. In the last few decades, several promising organic NLO crystalline materials have been developed for various applications. The basic molecular design is challenging, mainly because the crystalline packing of highly NLO molecules is not yet possible to predict in order to achieve highly favorable noncentrosymmetric packing potentially useful for second-order NLO applications. Additionally, the crystal growth and processing of these materials is often very challenging. For the materials already identified in the past, the main progress achieved in the last few years has been in the growth of high-optical-quality bulk crystals for terahertz-wave generation and structuring of optical waveguides for EO applications. Several novel materials optimized for these applications also have been developed and characterized in the last few years. In particular, organic NLO crystals have established in the field of terahertz applications due to their figures of merit considerably superior to inorganic materials, which has enabled these materials to enter the commercial market, are used in many laboratories as the source of terahertz waves, and have been also implemented in commercial terahertz spectrometers and terahertz imaging devices.

202

Handbook of Organic Materials for Electronic and Photonic Devices

Besides further development and implementation of organic NLO crystals in highly integrated photonic devices and new terahertz generation/detection schemes, the development of new materials is expected to continue to further extend the fundamental material knowledge on molecular and crystal engineering, as well as to identify optimized materials for novel applications and extend the possible application ranges, considering various pump laser sources, optical pump wavelengths, terahertz frequency ranges, and large-scale integrated photonic structures.

References Bernerd, C., Segonds, P., Debray, J., Notake, T., Koyama, M., Minamide, H., Ito, H., Boulanger, B., 2018. Quadratic nonlinear optical properties of the organic n-benzyl-2-methyl-4-nitroaniline (BNA) biaxial crystal. Opt. Lett. 43 (8), 1818–1821. Bosshard, C., Gunter, P., 1997. Electro-optic effects in organic molecules and polymers. In: Nalwa, H.S., Miyata, S. (Eds.), Nonlinear Optics of Organic Molecules and Polymers. CRC Press, Boca Raton, FL, p. 391. Bosshard, C., Sutter, K., Pr^etre, P., Hulliger, J., Fl€orsheimer, M., Kaatz, P., G€ unter, P., 1995. Organic Nonlinear Optical Materials. Gordon and Breach Science Publishers, Amsterdam. Bosshard, C., B€osch, M., Liakatas, I., J€ager, M., G€unter, P., 2000. Second-order nonlinear optical organic materials: recent developments. In: G€unter, P. (Ed.), Nonlinear Optical Effects and Materials. vol. 72. Brenier, A., 2015. Two-frequency pulsed YLIF4:ND lasing out of the principal axes and THz generation. Opt. Lett. 40 (19), 4496–4499. Brunner, F.D.J., Kwon, O.P., Kwon, S.J., Jazbinsek, M., Schneider, A., Gunter, P., 2008. A hydrogen-bonded organic nonlinear optical crystal for high-efficiency terahertz generation and detection. Opt. Express 16 (21), 16496–16508. Brunner, F.D.J., Schneider, A., Guenter, P., 2009. Velocity-matched terahertz generation by optical rectification in an organic nonlinear optical crystal using a Ti:sapphire laser. Appl. Phys. Lett. 94 (6), 061119. Brunner, F.D.J., Lee, S.H., Kwon, O.P., Feurer, T., 2014. THz generation by optical rectification of near-infrared laser pulses in the organic nonlinear optical crystal HMQ-TMS. Opt. Mater. Express 4 (8), 1586–1592. Cai, B., Komatsu, K., Kaino, T., 2003. Refractive index control and waveguide fabrication of DAST crystals by photobleaching technique. Opt. Mater. 21 (1–3), 525–529. Chemla, D.S., Zyss, J., 1987. Organic Nonlinear Optics: Molecules, Polymers and Crystals. Academic Press, Orlando, FL. Chen, H., Ma, O., Zhou, Y., Yang, Z., Jazbinsek, M., Bian, Y., Ye, N., Wang, D., Cao, H., He, W., 2015. Engineering of organic chromophores with large second-order optical nonlinearity and superior crystal growth ability. Cryst. Growth Des. 15 (11), 5560–5567. Choi, J.Y., Lee, S.J., Lee, S.C., Jeong, C.U., Jazbinsek, M., Yun, H., Kang, B.J., Rotermundd, F., Kwon, O.P., 2017. Quinolinium single crystals with a high optical nonlinearity and unusual out-of-plane polar axis. J. Mater. Chem. C 5 (47), 12602–12609. Clays, K., Coe, B.J., 2003. Design strategies versus limiting theory for engineering large secondorder nonlinear optical polarizabilities in charged organic molecules. Chem. Mater. 15, 642. Coe, B.J., Harris, J.A., Asselberghs, I., Clays, K., Olbrechts, G., Persoons, A., Hupp, J.T., Johnson, R.C., Coles, S.J., Hursthouse, M.B., Nakatani, K., 2002. Quadratic nonlinear optical properties of n-aryl stilbazolium dyes. Adv. Funct. Mater. 12, 110.

Molecular crystals and thin films for photonics

203

Coe, B.J., Harris, J.A., Asselberghs, I., Wostyn, K., Clays, K., Persoons, A., Brunschwig, B.S., Coles, S.J., Gelbrich, T., Light, M.E., Hursthouse, M.B., Nakatani, K., 2003. Quadratic optical nonlinearities of n-methyl and n-aryl pyridinium salts. Adv. Funct. Mater. 13 (5), 347–357. Cole, J.M., Lin, T.-C., Edwards, A.J., Piltz, R.O., Griet, D., Clays, K., Lee, S.C., Kwon, O.P., 2015. Concerted mitigation of O⋯H and C(π)⋯H interactions prospects sixfold gain in optical non linearity of ionic stilbazolium derivatives. ACS Appl. Mater. Interfaces 7 (8), 4693–4698. Cunningham, P.D., Hayden, L.M., 2010. Optical properties of DAST in the THz range. Opt. Express 18 (23), 23620–23625. Dalton, L.R., Sullivan, P.A., Bale, D.H., 2010. Electric field poled organic electro-optic materials: state of the art and future prospects. Chem. Rev. 110 (1), 25–55. Dalton, L., Gunter, P., Jazbinsek, M., Kwon, O.P., Sullivan, P., 2015. Nonlinear Optical Properties of Organic Molecules and Crystals. Cambridge University Press, Cambridge. Dhillon, S.S., Vitiello, M.S., Linfield, E.H., Davies, A.G., Matthias, C.H., Booske, J., Claudio Paoloni, M.G., Weightman, P., Williams, G.P., Castro-Camus, E., Cumming, D.R.S., Simoens, F., Escorcia-Carranza, I., Grant, J., Lucyszyn, S., Kuwata-Gonokami, M., Konishi, K., Koch, M., Charles, A.S., Tyler, L.C., Rupert Huber, A.G., Markelz, Z.D., Vincent, P.W.T., Axel Zeitler, J., Sibik, J., Korter, T.M., Ellison, B., Rea, S., Goldsmith, P., Ken, B.C., Appleby, R., Pardo, D., Huggard, P.G., Krozer, V., Shams, H., Fice, M., Renaud, C., Seeds, A., Stoehr, A., Naftaly, M., Ridler, N., Clarke, R., John, E.C., Michael, B.J., 2017. The 2017 terahertz science and technology roadmap. J. Phys. D. Appl. Phys. 50 (4), 043001. Dittrich, P., Bartlome, R., Montemezzani, G., Gunter, P., 2003. Femtosecond laser ablation of DAST. Appl. Surf. Sci. 220 (1–4), 88–95. Dolasinski, B., Peter, E.P., Joseph, W.H., Cooney, A., 2015. Tunable narrow band difference frequency THz wave generation in DAST via dual seed PPLN OPG. Opt. Express 23 (3), 3669–3680. Figi, H., Jazbinsek, M., Hunziker, C., Koechlin, M., Gunter, P., 2008a. Electro-optic singlecrystalline organic waveguides and nanowires grown from the melt. Opt. Express 16 (15), 11310–11327. Figi, H., Mutter, L., Hunziker, C., Jazbinsek, M., Gunter, P., Coe, B.J., 2008b. Extremely large nonresonant second-order nonlinear optical response in crystals of the stilbazolium salt DAPSH. J. Opt. Soc. Am. B 25 (11), 1786–1793. Figi, H., Jazbinsek, M., Hunziker, C., Koechlin, M., Gunter, P., 2009. Electro-optic tuning and modulation of single-crystalline organic microring resonators. J. Opt. Soc. Am. B 26 (5), 1103–1110. Figi, H., Jazbinsek, M., Hunziker, C., Koechlin, M., Gunter, P., 2010. Integrated electro-optic devices of melt-processable single-crystalline organic films. Proc. SPIE 7599, 1–6. Figi, H., Bale, D.H., Szep, A., Dalton, L.R., Chen, A.T., 2011. Electro-optic modulation in horizontally slotted silicon/organic crystal hybrid devices. J. Opt. Soc. Am. B 28 (9), 2291–2300. Finneran, I.A., Welsch, R., Allodi, M.A., Miller III, T.F., Blake, G.A., 2016. Coherent twodimensional terahertz-terahertz-Raman spectroscopy. Proc. Natl. Acad. Sci. USA 113 (25), 6857–6861. Fujiwara, M., Yanagi, K., Maruyama, M., Sugisaki, M., Kuroyanagi, K., Takahashi, H., Aoshima, S., Tsuchiya, Y., Gall, A., Hashimoto, H., 2006. Second order nonlinear optical properties of the single crystal of n-benzyl 2-methyl-4-nitroaniline: anomalous enhancement of the D(333) component and its possible origin. Jpn J. Appl. Phys. Part I 45 (11), 8676–8685.

204

Handbook of Organic Materials for Electronic and Photonic Devices

Fujiwara, M., Maruyama, M., Sugisaki, M., Takahashi, H., Aoshima, S.I., Cogdell, R.J., Hashimoto, H., 2007. Determination of the D-tensor components of a single crystal of n-benzyl-2-methyl-4-nitroaniline. Jpn J. Appl. Phys. Part I 46 (4A), 1528–1530. Geis, W., Sinta, R., Mowers, W., Deneault, S.J., Marchant, M.F., Krohn, K.E., Spector, S.J., Calawa, D.R., Lyszczarz, T.M., 2004. Fabrication of crystalline organic waveguides with an exceptionally large electro-optic coefficient. Appl. Phys. Lett. 84 (19), 3729–3731. Gheorma, I.L., Osgood, R.M., 2002. Fundamental limitations of optical resonator based highspeed EO modulators. IEEE Photonics Technol. Lett. 14 (6), 795–797. Glavcheva, Z., Umezawa, H., Mineno, Y., Odani, T., Okada, S., Ikeda, S., Taniuchi, T., Nakanishi, H., 2005. Synthesis and properties of i-methyl-4-2-[4-(dimethylamino)phenyl]ethenylpyridinium p-toluenesulfonate derivatives with isomorphous crystal structure. Jpn J. Appl. Phys. Part I 44 (7A), 5231–5235. Guarino, A., Poberaj, G., Rezzonico, D., Degl’Innocenti, R., Gunter, P., 2007. Electro-optically tunable microring resonators in lithium niobate. Nat. Photonics 1 (7), 407–410. Han, P.Y., Tani, M., Pan, F., Zhang, X.C., 2000. Use of the organic crystal DAST for terahertz beam applications. Opt. Lett. 25 (9), 675–677. Hashimoto, H., Okada, Y., Fujimura, H., Morioka, M., Sugihara, O., Okamoto, N., Matsushima, R., 1997. Second-harmonic generation from single crystals of n-substituted 4-nitroanilines. Jpn J. Appl. Phys. Part I 36 (11), 6754–6760. He, Y., Wang, Y., Xu, D., Nie, M., Yan, C., Tang, L., Shi, J., Feng, J., Yan, D., Liu, H., Teng, B., Feng, H., Yao, J., 2018. High-energy and ultra-wideband tunable terahertz source with DAST crystal via difference frequency generation. Appl. Phys. B Lasers Opt. 124 (1), 16. Heni, W., Kutuvantavida, Y., Haffner, C., Zwickel, H., Kieninger, C., Wolf, S., Lauermann, M., Fedoryshyn, Y., Andreas, F.T., Lewis, E.J., Delwin, L.E., Bruce, H.R., Freude, W., Koos, C., Leuthold, J., Larry, R.D., 2017. Silicon-organic and plasmonic-organic hybrid photonics. ACS Photonics 4 (7), 1576–1590. Hunziker, C., Kwon, S.J., Figi, H., Jazbinsek, M., Gunter, P., 2008. Fabrication and phase modulation in organic single-crystalline configurationally locked, phenolic polyene OH1 waveguides. Opt. Express 16 (20), 15903–15914. Hunziker, C., Kwon, S.J., Figi, H., Juvalta, F., Kwon, O.P., Jazbinsek, M., Gunter, P., 2008. Configurationally locked, phenolic polyene organic crystal 2-3-(4-hydroxystyryl)-5, 5-dimethylcyclohex-2-enylidenemalononitrile: linear and nonlinear optical properties. J. Opt. Soc. Am. B 25 (10), 1678–1683. Hwang, H.Y., Fleischer, S., Brandt, N.C., Perkins, B.G., Liu, M.K., Fan, K.B., Sternbach, A., Zhang, X., Averitt, R.D., Nelson, K.A., 2015. A review of non-linear terahertz spectroscopy with ultrashort tabletop-laser pulses. J. Mod. Opt. 62 (18), 1447–1479. Ilyakov, I.E., Kitaeva, G.K., Shishkin, B.V., Akhmedzhanov, R.A., 2014. Laser pulse amplitude changes induced by terahertz waves under linear electro-optic effect. Appl. Phys. Lett. 104 (15), 151107. Iwaszczuk, K., David, G.C., Fujiwara, M., Hashimoto, H., Jepsen, P.U., 2009. Simultaneous reference and differential waveform acquisition in time-resolved terahertz spectroscopy. Opt. Express 17 (24), 21969–21976. Jazbinsek, M., Gunter, P., 2011. Organic electro-optic crystal modulators. In: Chen, A., Murphy, E. (Eds.), Broadband Optical Modulators: Science, Technology, and Applications. CRC Press, London. Jazbinsek, M., Mutter, L., Guenter, P., 2008. Photonic applications with the organic nonlinear optical crystal DAST. IEEE J. Sel. Top. Quantum Electron. 14 (5), 1298–1311. Jazbinsek, M., Hunziker, C., Kwon, S.J., Figi, H., Kwon, O.P., Gunter, P., 2010. Hybrid organic crystal/silicon-on-insulator integrated electro-optic modulators. Proc. SPIE 7599, 1–14.

Molecular crystals and thin films for photonics

205

Jeong, J.H., Kang, B.J., Kim, J.S., Jazbinsek, M., Lee, S.H., Lee, S.C., Baek, I.H., Yun, H., Kim, J., Lee, Y.S., Lee, J.H., Kim, J.H., Rotermund, F., Kwon, O.P., 2013. High-power broadband organic THz generator. Sci. Rep. 3, 3200. Kaino, T., Cai, B., Takayama, K., 2002. Fabrication of DAST channel optical waveguides. Adv. Funct. Mater. 12 (9), 599–603. Kang, B.J., Lee, S.H., Kim, W.T., Lee, S.C., Lee, K., Benacchio, G., Montemezzani, G., Jazbinsek, M., Kwon, O.P., Rotermund, F., 2018. New class of efficient terahertz generators: effective terahertz spectral filling by complementary tandem configuration of nonlinear organic crystals. Adv. Funct. Mater. 28 (15), 1707195. Kawase, K., Mizuno, M., Sohma, S., Takahashi, H., Taniuchi, T., Urata, Y., Wada, S., Tashiro, H., Ito, H., 1999. Difference-frequency terahertz-wave generation from 4-dimethylamino-n-methyl-4-stilbazolium-tosylate by use of an electronically tuned Ti: sapphire laser. Opt. Lett. 24 (15), 1065–1067. Kawase, K., Hatanaka, T., Takahashi, H., Nakamura, K., Taniuchi, T., Ito, H., 2000. Tunable terahertz-wave generation from DAST crystal by dual signal-wave parametric oscillation of periodically poled lithium niobate. Opt. Lett. 25 (23), 1714–1716. Kawase, K., Shikata, J., Ito, H., 2003. Narrow-linewidth tunable terahertz-wave sources using nonlinear optics. Solid State Mid-Infrared Laser Sources 89, 397–423. Kim, P.J., Jeong, J.H., Jazbinsek, M., Kwon, S.J., Yun, H., Kim, J.T., Lee, Y.S., Baek, I.H., Rotermund, F., Gunter, P., Kwon, O.P., 2011. Acentric nonlinear optical n-benzyl stilbazolium crystals with high environmental stability and enhanced molecular nonlinearity in solid state. CrystEngComm 13 (2), 444–451. Kim, J.Y., Lee, S.H., Choi, I.Y., Choi, J.Y., Lee, S.C., Jazbinsek, M., Kim, W.S., Kwak, S.K., Huh, Y.S., Kang, J.W., Kwon, O.P., 2016. Stereoselective inhibitors based on nonpolar hydrocarbons for polar organic crystals. Cryst. Growth Des. 16 (11), 6514–6521. Kn€opfle, G., Schlesser, R., Ducret, R., G€unter, P., 1995. Optical and nonlinear optical properties of 40 -dimethylamino-n-methyl-stilbazolium tosylate (DAST) crystals. Nonlinear Opt. 9, 143–149. Koichi, M., Miyamoto, K., Ujita, S., Saito, T., Ito, H., Omatsu, T., 2011. Dual-frequency picosecond optical parametric generator pumped by a Nd-doped vanadate bounce laser. Opt. Express 19 (19), 18523–18528. Korn, D., Jazbinsek, M., Palmer, R., Baier, M., Alloatti, L., Yu, H., Bogaerts, W., Lepage, G., Verheyen, P., Absil, P., Guenter, P., Koos, C., Freude, W., Leuthold, J., 2014. Electro-optic organic crystal silicon high-speed modulator. IEEE Photonics J. 6 (2), 2700109. Kuroyanagi, K., Fujiwara, M., Hashimoto, H., Takahashi, H., Aoshima, S.I., Tsuchiya, Y., 2006. All organic terahertz electromagnetic wave emission and detection using highly purified n-benzyl-2-methyl-4-nitroaniline crystals. Jpn J. Appl. Phys. Part I Reg. Pap. Brief Commun. Rev. Pap. 45 (5A), 4068–4073. Kuzyk, M.G., 2000. Physical limits on electronic nonlinear molecular susceptibilities. Phys. Rev. Lett. 85, 1218. Kwon, O.P., Ruiz, B., Choubey, A., Mutter, L., Schneider, A., Jazbinsek, M., Gramlich, V., Gunter, P., 2006. Organic nonlinear optical crystals based on configurationally locked polyene for melt growth. Chem. Mater. 18 (17), 4049–4054. Kwon, O.P., Kwon, S.J., Jazbinsek, M., Brunner, F.D.J., Seo, J.I., Hunziker, C., Schneider, A., Yun, H., Lee, Y.S., Gunter, P., 2008. Organic phenolic configurationally locked polyene single crystals for electro-optic and terahertz wave applications. Adv. Funct. Mater. 18 (20), 3242–3250. Kwon, S.J., Hunziker, C., Kwon, O.P., Jazbinsek, M., Gunter, P., 2009. Large-area organic electro-optic single crystalline thin films grown by evaporation-induced local supersaturation with surface interactions. Cryst. Growth Des. 9 (5), 2512–2516.

206

Handbook of Organic Materials for Electronic and Photonic Devices

Kwon, S.J., Jazbinsek, M., Kwon, O.P., Guenter, P., 2010. Crystal growth and morphology control of OH1 organic electro-optic crystals. Cryst. Growth Des. 10 (4), 1552–1558. Kwon, O.P., Kwon, S.J., Jazbinsek, M., Choubey, A., Gramlich, V., Gunter, P.J., Kim, P., Jeong, J.H., Jazbinsek, M., Choi, S.B., Baek, I.H., Kim, J.T., Rotermund, F., Yun, H., Lee, J.S., Gunter, P., Kwon, O.P., 2012. Highly efficient organic THz generator pumped at near-infrared: quinolinium single crystals. Adv. Funct. Mater. 22, 1750–1756. Lee, S.H., Jazbinsek, M., Hauri, C.P., Kwo, O.P., 2016. Recent progress in acentric core structures for highly efficient nonlinear optical crystals and their supramolecular interactions and terahertz applications. CrystEngComm 18 (38), 7180–7203. Lee, S.H., Lee, G.H., Lee, K.H., Jazbinsek, M., Kang, B.J., Rotermund, F., Kwon, O.P., 2016. In situ tailor-made additives for molecular crystals: a simple route to morphological crystal engineering. Cryst. Growth Des. 16 (7), 3555–3561. Lee, S.H., Lee, S.J., Jazbinsek, M., Kang, B.J., Rotermund, F., Kwon, O.P., 2016. Electro-optic crystals grown in confined geometry with optimal crystal characteristics for THz photonic applications. CrystEngComm 18 (38), 7311–7318. Lee, S.H., Lu, J., Lee, S.J., Han, J.H., Jeong, C.U., Lee, S.C., Li, X., Jazbinsek, M., Yoon, W., Yun, H., Kang, B.J., Rotermund, F., Kwon, O.P., 2017. Benzothiazolium single crystals: a new class of nonlinear optical crystals with efficient THz wave generation. Adv. Mater. 29 (30). https://doi.org/10.1002/adma.201701748. Lee, S.J., Kang, B.J., Shin, M.H., Lee, S.C., Lee, S.H., Jazbinsek, M., Yun, H., Kim, D., Rotermund, F., Kwon, O.P., 2018. Efficient optical-to-THz conversion organic crystals with simultaneous electron withdrawing and donating halogen substituents. Adv. Opt. Mater. 6 (2), 1700930. Li, Y., Wu, Z., Zhang, X., Wang, L., Zhang, J., Wu, Y., 2014. Crystal growth and terahertz wave generation of organic NLO crystals: OH1. J. Cryst. Growth 402, 53–59. Little, B.E., Chu, S.T., Haus, H.A., Foresi, J., Laine, J.P., 1997. Microring resonator channel dropping filters. J. Lightwave Technol. 15 (6), 998–1005. Liu, J., Merkt, F., 2008. Generation of tunable Fourier-transform-limited terahertz pulses in 4-n, n-dimethylamino-40 -n0 -methyl stilbazolium tosylate crystals. Appl. Phys. Lett. 93 (13), 131105. Liu, J., Schmutz, H., Merkt, F., 2009. Generation of widely tunable Fourier-transform-limited terahertz pulses using narrowband near-infrared laser radiation. J. Mol. Spectrosc. 256 (1), 111–118. Liu, P., Xu, D., Li, Y., Zhang, X., Wang, Y., Yao, J., Wu, Y., 2014. Widely tunable and monochromatic terahertz difference frequency generation with organic crystal DSTMS. EPL 106 (6), 60001. Liu, P., Zhang, X., Yan, C., Xu, D., Li, Y., Shi, W., Zhang, G., Zhang, X., Yao, J., Wu, Y., 2016. Widely tunable and monochromatic terahertz difference frequency generation with organic crystal 2-(3-(4-hydroxystyryl)-5, 5-dime-thylcyclohex-2-enylidene) malononitrile. Appl. Phys. Lett. 108 (1), 011104. Liu, X., Yang, Z., Wang, D., Cao, H., 2016. Molecular structures and second-order nonlinear optical properties of ionic organic crystal materials. Crystals 6 (12), 158. Liu, B., Bromberger, H., Cartella, A., Gebert, T., Foerst, M., Cavalleri, A., 2017. Generation of narrowband, high-intensity, carrier-envelope phase-stable pulses tunable between 4 and 18 THz. Opt. Lett. 42 (1), 129–131. Lu, J., Hwang, H.Y., Li, X., Lee, S.H., Kwon, O.P., Nelson, K.A., 2015. Tunable multi-cycle THz generation in organic crystal HMQ-TMS. Opt. Express 23 (17), 22723–22729. Majkic, A., Zgonik, M., Petelin, A., Jazbinsek, M., Ruiz, B., Medrano, C., Guenter, P., 2014. Terahertz source at 9.4 THz based on a dual-wavelength infrared laser and quasi-phase matching in organic crystals OH1. Appl. Phys. Lett. 105 (14), 141115.

Molecular crystals and thin films for photonics

207

Marder, S.R., Perry, J.W., Schaefer, W.P., 1989. Synthesis of organic salts with large 2nd-order optical nonlinearities. Science 245 (4918), 626–628. Marder, S.R., Perry, J.W., Yakymyshyn, C.P., 1994. Organic salts with large 2nd-order optical nonlinearities. Chem. Mater. 6 (8), 1137–1147. Martin, M., Mangeney, J., Crozat, P., Mounaix, P., 2010. Optical phase detection in a 4-n,ndimethylamino-40 -n0 -methyl-stilbazolium tosylate crystal for terahertz time domain spectroscopy system at 1.55 μm wavelength. Appl. Phys. Lett. 97 (11), 111112. Meier, U., Bosch, M., Bosshard, C., Pan, F., Gunter, P., 1998. Parametric interactions in the organic salt 4-n, n-dimethylamino-40 -n0 -methyl-stilbazolium tosylate at telecommunication wavelengths. J. Appl. Phys. 83, 3486. Miyamoto, K., Minamide, H., Fujiwara, M., Hashimoto, H., Ito, H., 2008. Widely tunable terahertz-wave generation using an n-benzyl-2-methyl-4-nitro aniline crystal. Opt. Lett. 33 (3), 252–254. Miyamoto, K., Ohno, S., Fujiwara, M., Minamide, H., Hashimoto, H., Ito, H., 2009. Optimized terahertz-wave generation using BNA-DFG. Opt. Express 17 (17), 14832–14838. Monoszlai, B., Vicario, C., Jazbinsek, M., Hauri, C.P., 2013. High-energy terahertz pulses from organic crystals: DAST and DSTMS pumped at Ti:sapphire wavelength. Opt. Lett. 38 (23), 5106–5109. Montemezzani, G., Alonzo, M., Coda, V., Jazbinsek, M., Guenter, P., 2015. Running electric field gratings for detection of coherent radiation. J. Opt. Soc. Am. B 32 (6), 1078–1083. Mutter, L., Jazbinsek, M., Zgonik, M., Meier, U., Bosshard, C., Gunter, P., 2003. Photobleaching and optical properties of organic crystal 4-n, n-dirnethylamino-40 -n0 -methyl stilbazolium tosylate. J. Appl. Phys. 94 (3), 1356–1361. Mutter, L., Fabian, D.J.B., Yang, Z., Jazbinsek, M., Guenter, P., 2007. Linear and nonlinear optical properties of the organic crystal DSTMS. J. Opt. Soc. Am. B 24 (9), 2556–2561. Mutter, L., Guarino, A., Jazbinsek, M., Zgonik, M., Gunter, P., Dobeli, M., 2007. Ion implanted optical waveguides in nonlinear optical organic crystal. Opt. Express 15 (2), 629–638. Mutter, L., Koechlin, M., Jazbinsek, M., Gunter, P., 2007. Direct electron beam writing of channel waveguides in nonlinear optical organic crystals. Opt. Express 15 (25), 16828–16838. Mutter, L., Jazbinsek, M., Herzog, C., Gunter, P., 2008. Electro-optic and nonlinear optical properties of ion implanted waveguides in organic crystals. Opt. Express 16, 731–739. Nakanishi, H., Matsuda, H., Okada, S., Kato, M., 1989. Organic and polymeric ion-complexes for nonlinear optics. In: Doyama, M., Somiya, S., Chang, R.P.H. (Eds.), Materials Research Society International Meeting on Advanced Materials. Materials Research Society, Pittsburgh, PA, pp. 97–104. Notake, T., Nawata, K., Kawamata, H., Matsukawa, T., Minamide, H., 2012. Solution growth of high-quality organic n-benzyl-2-methyl-4-nitroaniline crystal for ultra-wideband tunable DFG-THz source. Opt. Mater. Express 2 (2), 119–125. Notake, T., Nawata, K., Kawamata, H., Matsukawa, T., Qi, F., Minamide, H., 2012. Development of an ultra-widely tunable DFG-THz source with switching between organic nonlinear crystals pumped with a dual-wavelength BBO optical parametric oscillator. Opt. Express 20 (23), 25850–25857. Ogawa, J., Okada, S., Glavcheva, Z., Nakanishi, H., 2008. Preparation, properties and structures of 1-methyl-4-2-[4-(dimethylamino)phenyl]ethenylpyridinium crystals with various counter anions. J. Cryst. Growth 310 (4), 836–842. Ohno, S., Miyamoto, K., Minamide, H., Ito, H., 2010. New method to determine the refractive index and the absorption coefficient of organic nonlinear crystals in the ultra-wideband THz region. Opt. Express 18 (16), 17306–17312. Oudar, J.L., Chemla, D.S., 1977. Hyperpolarizabilities of nitroanilines and their relations to excited-state dipole-moment. J. Chem. Phys. 66, 2664.

208

Handbook of Organic Materials for Electronic and Photonic Devices

Ovchinnikov, A.V., Chefonov, O.V., Molchanov, V.Y., Yushkov, K.B., Vicario, C., Hauri, C., 2016. Generation of frequency-tunable pulsed terahertz radiation by a Cr:forsterite laser system with an acoustooptical control of the pulse temporal profile. Quantum Electron. 46 (12), 1149–1153. Pan, F., Knopfle, G., Bosshard, C., Follonier, S., Spreiter, R., Wong, M.S., Gunter, P., 1996. Electro-optic properties of the organic salt 4-n,n-dimethylamino-40 -n0 -methyl-stilbazolium tosylate. Appl. Phys. Lett. 69 (1), 13–15. Pan, F., Wong, M.S., Bosshard, C., Gunter, P., 1996. Crystal growth and characterization of the organic salt 4-n,n-dimethylamino-40 -n0 -methyl-stilbazolium tosylate (DAST). Adv. Mater. 8 (7), 592. Puc, U., Bach, T., Krajewski, M., Medrano, C., Jazbinsek, M., 2018. Ultrabroadband terahertz time domain spectroscopy based on organic crystals. In: Book of Abstracts, 8th International Workshop on Terahertz Technology and Applications, Kaiserslautern, Germany. Rabiei, P., Steier, W.H., Zhang, C., Dalton, L.R., 2002. Polymer micro-ring filters and modulators. J. Lightwave Technol. 20, 1968–1975. Rezzonico, D., Jazbinsek, M., Guarino, A., Kwon, O.P., Gunter, P., 2008. Electro-optic Charon polymeric microring modulators. Opt. Express 16 (2), 613–627. Rezzonico, D., Kwon, S.J., Figi, H., Kwon, O.P., Jazbinsek, M., Gunter, P., 2008. Photochemical stability of nonlinear optical chromophores in polymeric and crystalline materials. J. Chem. Phys. 128, 124713. Rovere, A., Jeong, Y.G., Piccoli, R., Lee, S.H., Lee, S.C., Kwon, O.P., Jazbinsek, M., Morandotti, R., Razzari, L., 2018. Generation of high-field terahertz pulses in an HMQTMS organic crystal pumped by an ytterbium laser at 1030 nm. Opt. Express 26 (3), 2509–2516. Ruchert, C., Vicario, C., Hauri, C.P., 2012. Scaling submillimeter single-cycle transients toward megavolts per centimeter field strength via optical rectification in the organic crystal OH1. Opt. Lett. 37 (5), 899–901. Ruiz, B., Jazbinsek, M., Gunter, P., 2008. Crystal growth of DAST. Cryst. Growth Des. 8 (11), 4173–4184. Satoh, T., Toya, Y., Yamamoto, S., Shimura, T., Kuroda, K., Takahashi, Y., Yoshimura, M., Mori, Y., Sasaki, T., Ashihara, S., 2010. Generation of mid- to far-infrared ultrashort pulses in 4-dimethylamino-n-methyl-4-stilbazolium tosylate crystal. J. Opt. Soc. Am. B 27 (12), 2507–2511. Schneider, A., 2010. Theory of terahertz pulse generation through optical rectification in a nonlinear optical material with a finite size. Phys. Rev. A 82 (3), 033825. Schneider, A., Gunter, P., 2007. Coherent detection of terahertz pulses based on two-photon absorption in a photodiode. Appl. Phys. Lett. 90 (12), 121125. Schneider, A., Biaggio, I., Gunter, P., 2004. Terahertz-induced lensing and its use for the detection of terahertz pulses in a birefringent crystal. Appl. Phys. Lett. 84 (13), 2229–2231. Schneider, A., Neis, M., Stillhart, M., Ruiz, B., Khan, R.U.A., Gunter, P., 2006. Generation of terahertz pulses through optical rectification in organic DAST crystals: theory and experiment. J. Opt. Soc. Am. B 23 (9), 1822–1835. Schneider, A., Stillhart, M., Gunter, P., 2006. High efficiency generation and detection of terahertz pulses using laser pulses at telecommunication wavelengths. Opt. Express 14 (12), 5376–5384. Shalaby, M., Hauri, C.P., 2015. Demonstration of a low-frequency three-dimensional terahertz bullet with extreme brightness. Nat. Commun. 6, 5976. Shalaby, M., Vicario, C., Thirupugalmani, K., Brahadeeswaran, S., Hauri, C.P., 2016. Intense THz source based on BNA organic crystal pumped at Ti:sapphire wavelength. Opt. Lett. 41 (8), 1777–1780.

Molecular crystals and thin films for photonics

209

Somma, C., Folpini, G., Gupta, J., Reimann, K., Woerner, M., Elsaesser, T., 2015. Ultrabroadband terahertz pulses generated in the organic crystal DSTMS. Opt. Lett. 40 (14), 3404–3407. Stepanov, A.G., Ruchert, C., Levallois, J., Erny, C., Hauri, C.P., 2014. Generation of broadband THz pulses in organic crystal OH1 at room temperature and 10 K. Opt. Mater. Express. 4 (4), 870–875. Stillhart, M., Schneider, A., Guenter, P., 2008. Optical properties of 4-n,n-dimethylamino-40 -n0 methyl-stilbazolium 2,4,6-trimethylbenzenesulfonate crystals at terahertz frequencies. J. Opt. Soc. Am. B 25 (11), 1914–1919. Sutherland, R.L., 2003. Handbook of Nonlinear Optics. Dekker, New York. Takahashi, Y., Adachi, H., Taniuchi, T., Takagi, M., Hosokawa, Y., Onzuka, S., Brahadeeswaran, S., Yoshimura, M., Mori, Y., Masuhara, H., Sasaki, T., Nakanishi, H., 2006. Organic nonlinear optical DAST crystals for electro-optic measurement and terahertz wave generation. J. Photochem. Photobiol. A Chem. 183 (3), 247–252. Takayama, K., Komatsu, K., Kaino, T., 2001. Serially grafted waveguide fabrication of organic crystal and transparent polymer. Jpn J. Appl. Phys. Part I 40 (8), 5149–5150. Tang, M., Minamide, H., Wang, Y., Notake, T., Ohno, S., Ito, H., 2011. Tunable terahertz-wave generation from DAST crystal pumped by a monolithic dual-wavelength fiber laser. Opt. Express 19 (2), 779–786. Taniuchi, T., Shikata, J., Ito, H., 2000. Tunable terahertz-wave generation in DAST crystal with dual-wavelength KTP optical parametric oscillator. Electron. Lett. 36 (16), 1414–1416. Taniuchi, I., Adachi, H., Okada, S., Sasaki, T., Nakanishi, H., 2004a. Continuously tunable THz and far-infrared wave generation from DAST crystal. Electron. Lett. 40 (9), 549–551. Taniuchi, T., Okada, S., Nakanishi, H., 2004b. Widely tunable terahertz-wave generation in an organic crystal and its spectroscopic application. J. Appl. Phys. 95 (11), 5984–5988. Taniuchi, T., Okada, S., Nakanishi, H., 2004c. Widely-tunable THz-wave generation in 2–20 THz range from DAST crystal by nonlinear difference frequency mixing. Electron. Lett. 40 (1), 60–62. Taniuchi, T., Ikeda, S., Mineno, Y., Okada, S., Nakanishi, H., 2005. Terahertz properties of a new organic crystal, 40 -dimethylamino-n-methyl-4-stilbazolium p-chlorobenzenesulfonate. Jpn J. Appl. Phys. Part II 44 (28–32), L932–L934. Thirupugalmani, K., Venkatesh, M., Karthick, S., Maurya, K.K., Vijayan, N., Chaudhary, A.K., Brahadeeswaran, S., 2017. Influence of polar solvents on growth of potentially NLO active organic single crystals of n-benzyl-2-methyl-4-nitroaniline and their efficiency in terahertz generation. CrystEngComm 19 (19), 2623–2631. Tsuda, K., Kondo, T., Saito, F., Kudo, T., Ito, R., 1992. New fabrication method of channel optical wave-guides of organic-crystal using an inorganic photoresist. Jpn J. Appl. Phys. Part II 31 (2A), L134–L135. Uchida, H., Sugiyama, T., Suizu, K., Osumi, T., Kawase, K., 2011. Generation of widely tunable terahertz waves by difference-frequency generation using a configurationally locked polyene 2-[3-(4-hydroxystyryl)-5, 5-dimethylcyclohex-2-enylidene] malononitrile crystal. Terahertz Sci. Technol. 4, 132–136. Uchida, H., Saroj, R.T., Suizu, K., Shibuya, T., Osumi, T., Kawase, K., 2013. Widely tunable broadband terahertz radiation generation using a configurationally locked polyene 2-[3-(4-hydroxystyryl)-5,5-dimethylcyclohex-2-enylidene] malononitrile crystal via difference frequency generation. Appl. Phys. B Lasers Opt. 111 (3), 489–493. Uchida, H., Oota, K., Okimura, K., Kawase, K., Takeya, K., 2018. Single-cycle terahertz pulse generation from oh1 crystal via Cherenkov phase matching. J. Infrared Millim. Terahertz Waves 39 (6), 509–513.

210

Handbook of Organic Materials for Electronic and Photonic Devices

Vicario, C., Jazbinsek, M., Ovchinnikov, A.V., Chefonov, O.V., Ashitkov, S.I., Agranat, M.B., Hauri, C.P., 2015. High efficiency THz generation in DSTMS, DAST and OH1 pumped by Cr:forsterite laser. Opt. Express 23 (4), 4573–4580. Vicario, C., Monoszlai, B., Jazbinsek, M., Lee, S.H., Kwon, O.P., Hauri, C.P., 2015. Intense, carrier frequency and bandwidth tunable quasi single-cycle pulses from an organic emitter covering the terahertz frequency gap. Sci. Rep. 5, 14394. Vicario, C., Ruchert, C., Hauri, C.P., 2015. High field broadband THz generation in organic materials. J. Mod. Opt. 62 (18), 1480–1485. Walther, M., Jensby, K., Keiding, S.R., Takahashi, H., Ito, H., 2000. Far-infrared properties of DAST. Opt. Lett. 25 (12), 911–913. Wang, Z., Sun, W., Chen, A., Kosilkin, I., Bale, D., Larry, R.D., 2011. Organic electro-optic thin films by simultaneous vacuum deposition and laser-assisted poling. Opt. Lett. 36 (15), 2853–2855. Wu, Q., Zhang, X.C., 1996. Ultrafast electro-optic field sensors. Appl. Phys. Lett. 68 (12), 1604–1606. Yang, Z., Jazbinsek, M., Ruiz, B., Aravazhi, S., Gramlich, V., Gunter, P., 2007. Molecular engineering of stilbazolium derivatives for second-order nonlinear optics. Chem. Mater. 19 (14), 3512–3518. Yang, Z., Mutter, L., Stillhart, M., Ruiz, B., Aravazhi, S., Jazbinsek, M., Schneider, A., Gramlich, V., Guenter, P., 2007. Large-size bulk and thin-film stilbazolium-salt single crystals for nonlinear optics and THz generation. Adv. Funct. Mater. 17 (13), 2018–2023. Yin, J., Li, L., Yang, Z., Jazbinsek, M., Tao, X., Guenter, P., Yang, H., 2012. A new stilbazolium salt with perfectly aligned chromophores for second-order nonlinear optics: 4-n,ndimethylamino-40 -n0 -methyl-stilbazolium 3-carboxy-4-hydroxybenzenesulfonate. Dyes Pigments 94 (1), 120–126. Zhang, X., Jiang, X., Li, Y., Lin, Z., Zhang, G., Wu, Y., 2015. Isoxazolone-based single crystals with large second harmonic generation effect. CrystEngComm 17 (38), 7316–7322. Zhang, X., Jiang, X., Liu, P., Li, Y., Tu, H., Lin, Z., Xu, D., Zhang, G., Wu, Y., Yao, J., 2016. Molecular design on isoxazolone-based derivatives with large second-order harmonic generation effect and terahertz wave generation. CrystEngComm 18 (20), 3667–3673. Zhang, Y., Zhang, X., Li, S., Gu, J., Li, Y., Tian, Z., Ouyang, C., He, M., Han, J., Zhang, W., 2016. A broadband THz-TDS system based on DSTMS emitter and LTG InGaAs-InAlAs photoconductive antenna detector. Sci. Rep. 6, 26949. Zheng, X.M., McLaughlin, C.V., Cunningham, P., Hayden, L.M., 2007. Organic broadband terahertz sources and sensors. J. Nanoelectron. Optoelectron. 2 (1), 58–76. Zhong, K., Mei, J., Wang, M., Liu, P., Xu, D., Wang, Y., Shi, W., Yao, J., Teng, B., Xiao, Y., 2017. Compact high-repetition-rate monochromatic terahertz source based on difference frequency generation from a dual-wavelength ND:YAG laser and DAST crystal. J. Infrared Millim. Terahertz Waves 38 (1), 87–95. Zyss, J. (Ed.), 1994. In: Molecular Nonlinear Optics: Materials, Physics, Devices. Academic Press, Boston, MA. Zyss, J., Oudar, J.L., 1982. Relations between microscopic and macroscopic lowest-order optical nonlinearities of molecular-crystals with one-dimensional or two-dimensional units. Phys. Rev. A 26, 2028.

Hybrid perovskites for device applications

7

Kyle Frohna, Samuel D. Stranks Cavendish Laboratory, University of Cambridge, Cambridge, United Kingdom

7.1

Introduction

The perovskite structure has shown enormous utility in many facets of science in the past 50 years. From applications in high-temperature superconductors to piezoelectric sensors (Shinobu and Yuji, 1987; He et al., 2001; Park and Snyder, 1995; Cox et al., 2001; Du et al., 1998), the perovskite structure has been used in a wide array of cutting-edge areas of solid-state physics and materials science. The perovskites commonly employed for these applications are inorganic, metal-oxide-based materials such as SrTiO3 and LaAlO3. During this time, another branch of perovskites involving a metal-halide framework combined with either a small organic cation such as methylammonium (CH3NH+3 , MA) or a small inorganic ion such as cesium began to be studied (Mitzi, 1999, 2001; Mitzi et al., 1994; Onoda-Yamamuro et al., 1992). Variants of these hybrid-organic inorganic perovskites (HOIPs) were studied through the 1990s and early 2000s for use in organic electronic devices such as transistors (Mitzi et al., 2001, 2002) and light-emitting diodes (LEDs) (Chondroudis and Mitzi, 1999). However, it was not until these metal-halide perovskites were explored as a potential solar absorber in dye-sensitized solar cells (DSSCs) that their most remarkable properties were revealed. The archetypal methylammonium lead iodide (MAPbI3) was the first such material to be used in a solar cell configuration in 2009, showing a power conversion efficiency of 3.8% (Kojima et al., 2009). Published efficiencies for MAPbI3-based solar cells increased from 3.8% to 15% in just 4 years (Lee et al., 2012; Liu et al., 2013; Burschka et al., 2013). Since then, improvements to the carrier selective contacts, light management, and compositional engineering of perovskites have enabled the fabrication of solar cells with certified efficiencies above 22% (Fig. 7.1A) (Yang et al., 2017a; Saliba et al., 2016a; Jiang et al., 2017). Metal-halide perovskites also have been subsequently used to great initial success in LEDs, with these devices now showing remarkably high peak external quantum efficiencies (Tan et al., 2014; Xiao et al., 2017; Li et al., 2016a; Zhang et al., 2017), and recently have been studied in more niche applications such as X-ray/gamma ray (Wei et al., 2017a, b; Shrestha et al., 2017) and visible photodetectors (Garcı´a de Arquer et al., 2017; Ahmadi et al., 2017). The paradigm shift associated with HOIPs has largely been driven by their remarkable properties: they exhibit strong absorption and long diffusion lengths associated Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00007-3 © 2019 Elsevier Ltd. All rights reserved.

212

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 7.1 (A) The record power conversion efficiency over the years of perovskite solar cells and other commercial photovoltaic (PV) technologies. (B) The perovskite crystal structure. (C) The lattice parameters of the three phases of MAPbI3 as measured by X-ray diffraction (XRD) as a function of temperature. (D) Absorption spectrum and images of thin films in the MAPb(IxBr1-x)3 phase space as x is tuned from 1 to 0. (E) The band-gap ranges of various 3D HOIP perovskites. The range for each band gap is attained by varying the halide composition from I to Br to Cl. (C) Adapted with permission from Nature Publishing Group. Whitfield, P.S., et al., 2016. Structures, phase transitions and tricritical behavior of the hybrid perovskite methyl ammonium lead iodide. Sci. Rep. 6, 35685. (D) Adapted with permission from the Royal Society of Chemistry. Hoke, E.T., et al., 2015. Reversible photo-induced trap formation in mixed-halide hybrid perovskites for photovoltaics. Chem. Sci. 6 (1), 613–617. (E) Reprinted from Anaya, M., et al., 2017. ABX3 perovskites for tandem solar cells. Joule 1 (4), 769–793 with permission from Elsevier.

Hybrid perovskites for device applications

213

with traditional crystalline semiconductors (De Wolf et al., 2014; Stranks et al., 2013; Brenes et al., 2017), while simultaneously showing a remarkable tolerance to defects and the ability to be solution-processed in a similar way as organic-based devices. However, the added benefits of organic-like properties and their ionic nature mean that they face the same device stability challenges (Leijtens et al., 2015; Asghar et al., 2017). In this chapter, we will discuss in detail the structural properties of HOIPs and band-gap tuning via structural engineering; methods for deposition and fabrication of perovskite films and devices; the material and optoelectronics properties of HOIPs, including carrier dynamics; and, finally, their diverse range of applications.

7.2

Structural properties

Perovskites take the general chemical formula ABX3, with the compositions of interest in this chapter comprised of A as a monovalent cation, B as a divalent metal ion, and X as a halide anion. The common A-site cations are short organic cations MA, NH2CHNH+2 (formamidinium, FA) and/or the inorganic Cs+ (Saliba et al., 2016a; Yi et al., 2016; Eperon et al., 2014). The B-site is typically occupied by either lead (Pb), tin (Sn), or an alloy of the two, with some recent work exploring germanium (Ge), as well as mixed-metal double perovskites (Ogomi et al., 2014; Eperon et al., 2016; Stoumpos et al., 2015; Volonakis et al., 2017). The X-site is almost exclusively occupied by iodine (I) or bromine (Br), although chlorine (Cl) has also been used (Saliba et al., 2016a; Rehman et al., 2017; Akkerman et al., 2015). The basic metal-halide perovskite structure is a three-dimensional (3D) metal-halide framework consisting of corner-sharing octahedra. The metal ion resides at the center of each octahedron and is connected to six adjacent halide atoms. The unit cell of the simplest structure is cubic, with the B-site metal at the center of the unit cell, the X-site halides at the center of each cube face, and the A-site cations at the corner of the cubes (Fig. 7.1B). The metal-halide framework is strongly bound, while the A-site cations are predicted to be free to rotate at room temperature (Wasylishen et al., 1985). To form a stable perovskite structure, the relative sizes of the ions must be suitable. This is captured by the Goldschmidt tolerance factor t: rA + rX t ¼ pffiffiffi , 2ð r B + r X Þ

(7.1)

where rA, rB, and rX are the ionic radii of the A-, B-, and X-site ions, respectively. When the tolerance factor is close to 1, a regular cubic perovskite structure is predicted to form. If the tolerance factor is too far above or below 1, then a nonperovskite phase will form preferentially. In practice, HOIPs have been successfully found in the range of tolerance factors from 0.8 to 1.0 (Kieslich et al., 2014). This limits the compositions of stable perovskite structures that can be formed using the ionic components listed here. Distorted perovskite structures that deviate from the regular arrangement of octahedra also can be stabilized. One distortion that has been well documented is the tilting of adjacent octahedra with respect to one another (Glazer, 1972). Above about 330 K, the MAPbI3

214

Handbook of Organic Materials for Electronic and Photonic Devices

structure is cubic with a space group of Pm3m (Kawamura et al., 2002; Weller et al., 2015). Between about 160 K and 330 K, the most relevant regime to device structures operating at room temperature, the structure becomes tetragonal with a lengthened c-axis, a space group of I4/mcm, and octahedra tilting around the c-axis with adjacent planes of octahedra rotated out of phase with one another. Below about 160 K, the structure is further distorted to orthorhombic with a space group of Pnma (Fig. 7.1C). These structural transitions from orthorhombic to tetragonal to cubic are general to many perovskites, although the transition temperatures vary as a function of composition (Cox et al., 2001; Takahashi et al., 2011; Wang et al., 2017a; Ishidate et al., 1997). The ionic radius of MA results in a tolerance factor for MAPbI3 close to 0.9, whereas the FA-containing perovskite has tolerance factors of approximately 1 (Kieslich et al., 2014). Stabilizing thin films of the cubic FAPbI3 (black phase) is more difficult and requires higher temperatures (Eperon et al., 2014), suggesting that although the tolerance factor is close to 1, the inherent uncertainties in measuring the ionic radius of an organic ion may be underpredicting the tolerance factor in these materials. Cs+ has a smaller ionic radius than either MA or FA, and CsPbI3 has a correspondingly smaller tolerance factor. CsPbI3 also does not form a stable perovskite phase at room temperature (Beal et al., 2016), although the perovskite phase can be synthesized to be metastable (Eperon et al., 2015; Hutter et al., 2017). It is becoming increasingly common for an alloy of FA, MA, and/or cesium (Cs) to be used at the A-site, which allows the formation of a stable perovskite structure as a thin film (Yi et al., 2016; Eperon et al., 2016). This alloying approach is applied to all three distinct sites for multiple reasons, including that the band gap of the perovskite can be tuned by varying the composition at each site, and optimally stable crystal structures can be achieved. In fact, a workhorse composition in the HOIP solar cell community is FAxCs1-xPb(IyBr1-y)3 with x and y commonly varying between 0.6 and 0.85 and 0.6–1.0, respectively (Yi et al., 2016; Bush et al., 2017, 2018), sometimes with additional small fractions of MA (Saliba et al., 2016b). This highlights the considerable phase space available for structure-function optimization.

7.2.1 Band-gap tuning The tuning of absorption and emission wavelengths is one of the most attractive properties of HOIPs, with band gaps from about 1.2–3.0 eV readily achievable through compositional tuning (Fig. 7.1E) (Eperon et al., 2014, 2016; Wang et al., 2016). HOIPs are tolerant of a wide phase space of substituents, which allows the fine tuning of the optical properties of the material. The first methods to controllably tune the band gap were achieved by replacing some iodide with bromide to form the compound MAPb(IxBr1-x)3, where x varies from 0 to 1 (Fig. 7.1D) (Hoke et al., 2015; Kulkarni et al., 2014; Noh et al., 2013). Replacing iodide with bromide in the lattice results in an increase in the band gap of the perovskite. This is rationalized by the character of the states at the valence band maximum being an antibonding hybrid of halide p orbitals (5p for I, 4p for Br) and metal s orbitals (6s for Pb, 5s for Sn), while the conduction band is an antibonding/nonbonding hybrid of primarily metal p orbitals with some halide p character (Grote and Berger, 2015; Brandt et al., 2015). The antibonding

Hybrid perovskites for device applications

215

nature of the valence and conduction band extrema also means that localized states formed due to the bond breaking (i.e., intrinsic defects) will not reside within the gap. This is one proposed explanation for the apparent defect tolerance of HOIPs (the high performance of HOIP devices in the presence of large concentrations of defects) (Brandt et al., 2015). Bromide has a larger ionization potential than iodide due to its greater electronegativity, which decreases the valence-band maximum energy and increases the band gap of Br-based perovskites, while leaving the conduction-band minimum at the same energy (Schulz et al., 2014). The trend also holds for the substitution of Br with Cl, although Cl-based perovskites do not form as readily, in spite of their favorable tolerance factor (Nagabhushana et al., 2016). The B-site also can be used to tune the band gap. By alloying Pb with Sn in a system such as APb1-xSnxI3, the band gap is found to show an anomalous bowing to lower values of about 1.2 eV for x in the range of about 0.5–0.8, in spite of the fact that the pure Pb and pure Sn perovskites have band gaps in the range of 1.6 eV and 1.4 eV, respectively (Fig. 7.2A) (Ogomi et al., 2014; Eperon et al., 2016). Although the A-site cation does not contribute to the density of states at the band edge, replacing the A-site cation can influence the structure of the metal-halide framework, which in turn can influence the band gap (Prasanna et al., 2017). Replacing MA with FA results in a decrease in the band gap of about 100 meV of Pb-based HOIPs, while replacing MA with Cs can result in an increase of the band gap of about 150 meV (Eperon et al., 2014, 2015). Given that the valence band is an antibonding hybrid of metal and halide states, if the overlap between these orbitals increases, the antibonding states will be further destabilized and the band gap decreases; similarly, reducing their overlap will increase the band gap. The two main mechanisms for this are lattice contraction and octahedral tilting (Prasanna et al., 2017). Assuming a cubic starting structure, the substitution of a smaller A-site cation may cause the lattice to contract isotropically, while retaining its cubic structure. In this case, the metal-halide orbital overlap will increase and the band gap will decrease. This is also the case if some Cs is substituted into FASnI3 (Prasanna et al., 2017). Alternatively, the substitution of a smaller cation may cause the structure to distort and the octahedra to tilt relative to one another. In this case, given the directionality of the overlap with the halide p orbital, reducing the X-B-X bond angle below 180 degrees reduces the orbital overlap and increases the band gap. This is true in the equivalent lead system, i.e., when substituting Cs+ into FAPbI3. The wide bandgap range that can be obtained with HOIPs makes them candidates for a range of applications, including both top and bottom cells in tandem solar cells (Section 7.7.1) (Eperon et al., 2016; Bush et al., 2017), as well as colored light emitters (Section 7.2) and photodetectors (Section 7.7.3).

7.2.2 Dimensional tuning 3D perovskites are commonly formed using small cations at the A-site; however, by increasing the size of the cations and tuning stoichiometry, HOIPs also can be fabricated to have a range of dimensionalities (Mitzi, 1999). Two-dimensional (2D) Ruddlesden-Popper HOIPs can be fabricated by including longer organic spacer cations into the structures. These cations are too large to fit into a conventional 3D

216

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 7.2 (A) Plot of the band gap of the FAxCs1-xSnyPb1-yI3 phase space as x varies from 0.7–1 and y varies from 0 to 1 showing anomalous band-gap bowing at intermediate y-values. (B) Structure of the quasi-2D perovskites for different values of n with long CH3(CH2)3NH+3 spacer ions. (C) The antisolvent drip method for the fabrication of HOIP thin films via spin coating, with addition of an antisolvent while spin coating the precursor solution. (D) Crystal structure of CsPbBr3 nanocrystals, and transmission electron microscope (TEM) image of a single nanocrystal and a densely packed nanocrystal film. (E) Photographs of MAPbX3 (X ¼ Cl, Br, I) by introducing a seed crystal into a hot precursor solution multiple times. (A) Adapted with permission from Prasanna, R., et al., 2017. Band gap tuning via lattice contraction and octahedral tilting in perovskite materials for photovoltaics. J. Am. Chem. Soc. 139 (32), 11117–11124. Copyright 2017 American Chemical Society. (B) Reproduced with permission from Stoumpos, C.C., et al., 2016. Ruddlesden–popper hybrid lead iodide perovskite 2D homologous semiconductors. Chem. Mater. 28 (8), 2852–2867. Copyright 2016 American Chemical Society. (C) Adapted with permission from Springer Nature: Nature Materials, Jeon, N.J., et al., 2014. Solvent engineering for high-performance inorganic–organic hybrid perovskite solar cells. Nat. Mater. 13, 897; Copyright 2014. (Continued)

Hybrid perovskites for device applications

217

perovskite structure but can arrange themselves at the surface of quasi-2D perovskite sheets (Chen et al., 2018). By tailoring the ratio of the spacer cations to the rest of the perovskite precursors, one can tune the thickness of the quasi-2D sheet (Fig. 7.2B). The general formula for Ruddlesden-Popper perovskites is S2An-1BnX3n+1, where S is a large organic spacer cation such as n-butylammonium (BA) or phenylethylammonium (PEA) and n is the number of planes of BX6 octahedra making up the 2D sheets. These 2D sheets also strongly absorb light and the organic spacer layers act as a barrier to penetration of water molecules, thereby showing improved environmental stability, albeit somewhat inhibited charge transport over their 3D equivalents (Smith et al., 2014; Tsai et al., 2016). Mixing large and small cations allows one to combine the favorable charge transport properties of 3D HOIPs— low-exciton binding energy and isotropic mobilities—with the intrinsic stability of the 2D analogs (Wang et al., 2017b; Ran et al., 2018; Grancini et al., 2017). Confinement effects in lower-dimensional perovskite structures also lead to an increase in the band gap, which is particularly promising for LEDs emitting in the green and blue (Jia et al., 2018; Kovalenko et al., 2017; Tsai et al., 2018). However, the quantum and dielectric confinement of the carriers in these sheets also causes an increase in the exciton binding energy by an order of magnitude over their 3D counterparts, and these values scale inversely with the thickness of the quasi-2D layer (Blancon et al., 2017; Galkowski et al., 2016). A large exciton-binding energy in the 2D structures could be problematic for designing appropriate PV devices in which extraction of free charges is required, although recent evidence suggests that efficient free-charge generation is still possible, with excitons diffusing to edge states that facilitate rapid dissociation of the excitons into free carriers (Blancon et al., 2017). The 2D sheets also have lower intersheet mobilities, as charges must hop between sheets to be collected at electrodes and resulting devices, although more stable, are currently less efficient than their 3D counterparts (Grancini et al., 2017). Recently, layered nanoplatelets have been fabricated that show highly tunable emission properties and confinement effects, and they can be deposited from colloidal solutions (Weidman et al., 2016; Li et al., 2017). One-dimensional (1D) perovskites (nanowires) also can be produced by tuning the stoichiometry; the increased confinement effects result in further blue-shifted emission (Yuan et al., 2017). Zero-dimensional (0D) perovskite nanocrystals (NCs) can also be synthesized via colloidal synthesis and other means (Fig. 7.2D) (Protesescu et al., 2015). The crystal structure of the NCs is that of the 3D perovskites, but stabilized by long ligands at the surface. Their band gaps can be tuned by a combination of changes to their size and composition. NCs have been reported to exhibit very high

Fig. 7.2, Cont’d (D) Reproduced with permission from Protesescu, L., et al., 2015. Nanocrystals of cesium lead halide perovskites (CsPbX3, X ¼ Cl, Br, and I): novel optoelectronic materials showing bright emission with wide color gamut. Nano Lett. 15 (6), 3692–3696. Copyright 2015 American Chemical Society. (E) Adapted with permission from the Wiley Publishing Group. Liu, Y., et al., 2015. Two-inch-sized perovskite CH3NH3PbX3 (X ¼ cl, Br, I) crystals: growth and characterization. Adv. Mater. 27 (35), 5176–5183.

218

Handbook of Organic Materials for Electronic and Photonic Devices

photoluminescence quantum yields (PLQYs) approaching 100%, even without additional passivation considerations, in stark contrast to chalcogenide-based, core-shell NCs, which require careful passivation (Protesescu et al., 2015; Chen et al., 2013). It is likely that the ligands act as additional and very effective passivating agents to manage surface defects (Kovalenko et al., 2017). Nanocrystals also have been fabricated without ligands by depositing the perovskite precursor into periodic, nanostructured scaffolds (Anaya et al., 2017b).

7.3

Deposition methods

7.3.1 Solution-processing 7.3.1.1 Spin coating Polycrystalline HOIP films are most commonly deposited in the laboratory by spin coating precursor inks onto a substrate. The precursor inks commonly consist of a metal halide salt and an organic/alkali halide salt dissolved in suitable solvents. The substrate is then rotated at several thousand revolutions per minute for a period of seconds to minutes, followed by an annealing of the substrate, which leads to formation of the crystalline perovskite thin film. The high-speed spinning process typically produces a very thin, uniform perovskite precursor phase film, which is then converted to the perovskite phase with the application of heat (Lee et al., 2012). The prevalence of spin coating comes from its speed and ability to screen different processing parameters rapidly, whereas other processes such as vapor deposition can be more arduous for these purposes. This ease of processing likely contributed to the rapid expansion of the HOIP field, as those working in the OPV and DSSC fields could immediately use existing processing methods on these new materials. However, given the scale limitations of spin coating, most devices are made on substrates of approximately 2  2 cm2 or less and the active areas of most reported devices are typically much smaller than 1 cm2. The spin-coating step is often tuned in order to achieve the desired film thickness and morphology, with typical thicknesses for solar cells of about 500 nm. Early work employed the solvent N,N-dimethylformamide (DMF) to dissolve precursors such as PbI2 and MAI (Lee et al., 2012). However, other solvents such as dimethylsulfoxide (DMSO) have since been added to the mixture, where the DMSO acts as a Lewis base and complexes to the PbI2 in the precursor solution (Jeon et al., 2014; Ahn et al., 2015). This complexing leads to a more favorable film morphology than the neat DMF solutions. DMSO is often coupled with a means of rapidly removing the excess solvent through the use of an antisolvent drip method. A volatile organic solvent such as toluene or chlorobenzene, in which the perovskite precursor is not soluble, is expelled onto the substrate during the spin-coating process (Fig. 7.2C) (Bush et al., 2017; Jeon et al., 2014). Once the precursor is annealed, the volatile antisolvent and the DMSO evaporate quickly, resulting in a compact polycrystalline perovskite film in general devoid of pinholes.

Hybrid perovskites for device applications

219

An alternative approach is a two-step sequential deposition process wherein one of the precursors, typically the metal halide, is deposited in the first step. Then the organic halide is deposited on top of the metal halide and the film is converted to the perovskite. Reports of improved film uniformity and device performance have been reported with these approaches through either solution or vapor processing of the organic halide in the second step (Burschka et al., 2013; Liang et al., 1998; Yang et al., 2015a). More recently, there have been efforts to replace DMF and DMSO with less toxic (green) solvents more suitable for large-scale manufacturing. Acetonitrile is a very promising candidate, with a recent demonstration of large-area, smooth perovskite films fabricated via spin coating from this solvent system (Noel et al., 2017a). Spin coating is widely used due to its compatibility with different substrate types and the rate at which films can be produced in a laboratory. However, large-scale photovoltaics (PVs) cannot be manufactured in this way, as the method is not scalable. In order to push HOIPs toward commercialization, more scalable deposition methods are required.

7.3.1.2 Blade coating, slot-die coating, and inkjet printing Blade coating is a scalable deposition method from solution in which the perovskite precursor inks are spread across the substrate by a metal blade. The ink is deposited in front of the blade, and either the blade is spread across the substrate or, more commonly for roll-to-roll processing, the substrate is pulled underneath the blade. This is closely related to slot-die coating, in which there is a slot at the tip of the blade through which the ink is deposited rather than being put in front of the blade. Inkjet printing is another potentially scalable deposition method in which the ink is finely deposited onto the substrate. Perovskite films have been successfully deposited by blade coating, and optimization has shown respectable efficiencies on large areas approaching minimodule size (about 10 cm2) (Yang et al., 2017b, 2015b; Kim et al., 2015; Deng et al., 2015). There has been similar progress with slot-die coating, including fully slot-die-coated devices (Hwang et al., 2015; Vak et al., 2015; Qin et al., 2017). Many of these reports demonstrate the fabrication of fully printed, large-area modules on flexible substrates without the need for vacuum or an inert atmosphere. This suggests that large-scale roll-to-toll manufacturing may be possible with these methods. Preliminary reports using inkjet printing have also demonstrated moderate success (Li et al., 2015; Mathies et al., 2016; Hashmi et al., 2017). Processing under atmospheric conditions removes the need for costly high-vacuum systems, as well as the time required to pump down to a suitable vacuum. The efficiencies of cells fabricated in this way are still lower than their spin-coated counterparts, and it remains to be seen whether these techniques can scale up to full module size.

7.3.2 Vapor deposition Vapor deposition or evaporation involves the heating of precursor powders under a high vacuum until the vapor pressure is sufficient to coat a substrate suspended above the evaporation sources. Early reports document the use of the simultaneous evaporation

220

Handbook of Organic Materials for Electronic and Photonic Devices

(coevaporation) of both the organic halide compound and the lead halide precursors to form perovskite films without a solvent (Liu et al., 2013). Since then, considerable efforts have led to the development of sophisticated, vapor-based deposition methods, including solar cells that are fully processed under vacuum (Longo et al., 2017; Sessolo et al., 2015; Ono et al., 2016). Apart from the obvious advantage of being able to scale this deposition method easily, vapor deposition allows a much greater control of film thickness and more uniform film morphology. There are also inherently fewer impurities associated with evaporating materials in a high or moderate vacuum without solvents. A hybrid between vapor and solution processing known as vapor-assisted solution processing (VASP) has been reported, in which the inorganic metal halide is deposited via solution (or vapor) before the film is exposed to the organic halide vapor (Chen et al., 2014; Zhou et al., 2015). In addition, alternative deposition methods that do not require such low vacuum levels such as chemical vapor deposition (CVD) have been reported, which could lower the inherent cost barrier to vacuum-based processing (Leyden et al., 2016; Luo et al., 2015; Shen et al., 2016).

7.3.3 Single-crystal growth Although the easily fabricated, defect-rich, polycrystalline HOIP films have been the most technologically relevant for solar and LED applications to date, certain applications and fundamental science research benefit from single crystals of HOIPs. Consequently, a variety of methods have been employed to grow HOIP single crystals. Saidaminov et al. report the use of an inverse solubility method, which exploits the fact that HOIPs become less soluble in certain solvents at higher temperatures (Saidaminov et al., 2015). By slowly ramping up the temperature, large HOIP crystals can be crystallized out of solution. Other methods involve introducing a smaller perovskite seed crystal to a hot solution of perovskite precursor (Fig. 7.2E) (Liu et al., 2015), or using antisolvent vapor to assist in the crystallization of the HOIP from solution (Shi et al., 2015). Recently, the inverse solubility method was used to obtain thin-film single crystals on solar cell contact materials for device applications (Chen et al., 2017a). Also, 2D HOIP single crystals have been fabricated by heating solutions of 2D perovskite precursors before slowly cooling to room temperature (Stoumpos et al., 2016; Peng et al., 2017).

7.4

Optoelectronic properties of HOIPs

7.4.1 Absorption coefficient One key property of HOIPs relevant to their performance in PV applications is their large absorption coefficients. The archetypal MAPbI3 has a remarkably sharp absorption onset (De Wolf et al., 2014), with low Urbach energies (about 13meV) that are comparable to other high-quality semiconductors [7.5 meV for gallium arsenide (GaAs) (Johnson and Tiedje, 1995), 9.6 meV for c-Si (Cody, 1992)]. This suggests that they are relatively clean semiconductors despite a moderately large defect density.

Hybrid perovskites for device applications

221

The absorption coefficient of MAPbI3 is also larger than both GaAs and CdTe, two of the most successful inorganic, thin-film solar technologies to date, across a large portion of the solar spectrum (Fig. 7.3A). With absorption coefficients above 105 cm1 across much of the solar spectrum, HOIP-based thin-film devices need only have an absorbing layer thickness of a few hundred nanometers to absorb a large proportion of all incident light. This, in tandem with favorable charge carrier transport properties (which will be discussed in Section 5.2), means that the majority of incident photons will generate photoexcited species that will diffuse to the respective selective contacts in solar-cell or photodetector device structures.

7.4.2 Exciton binding energy When light is absorbed in a material, an electron is promoted to an excited state in the conduction band, leaving behind a positively charged hole in the valence band. The electron and hole can be bound electrostatically as a quasiparticle called an exciton. Whether the exciton dissociates to generate free charges or remains bound depends on the exciton binding energy relative to the thermal energy (about 25 meV at room temperature), and the binding energy in turn depends on the inverse of the material’s dielectric constant. Traditional organic materials such as polymers used in organic solar cells have low relative dielectric constants of about 2, resulting in the formation of stable, tightly bound excitons. HOIPs, by contrast, possess dielectric constants much larger with values between about 5 and 25 reported for low- and high-frequency dielectric constant, respectively (Yang et al., 2017c; Samiee et al., 2014). However, there has been ongoing debate about the frequency (high or low) to consider in order to extract the relevant dielectric constant for perovskites. This has resulted in a wide spread of reported exciton binding energies in the range of about 2–85 meV for MAPbI3 (Yang et al., 2017c; D’Innocenzo et al., 2014; Ishihara, 1994; Zheng et al., 2015a; Men
endez-Proupin et al., 2014; Lin et al., 2014). However, subsequent direct magneto-optic measurements without assumptions on the dielectric constant have revealed a value of 16  2 meV for the exciton binding energy at low temperature (Miyata et al., 2015), and these measurements suggest a lower energy of the order ≪10 meV for the room temperature phase, as has been estimated elsewhere (Fig. 7.3B) (Lin et al., 2014; Yamada et al., 2015). This means that the dominant photoexcited species in 3D perovskites such as MAPbI3 are free carriers, allowing the efficient collection of free charges at selective contacts in solar-cell structures. This is a marked advantage over traditional OPV devices, which require exciton dissociation interfaces and therefore voltage losses introduced by energetically staggered heterojunctions. However, the exciton binding energy has been reported to increase marginally with the band gap of the perovskite (Fig. 7.3B), suggesting that 3D perovskites may make more efficient LEDs in the green and blue ends of the spectrum. As discussed in Section 7.2.2, reducing the dimensionality of the perovskites has the effect of increasing the exciton binding energy because the electrostatic interaction between the electron and hole is no longer screened in all three dimensions.

222

Handbook of Organic Materials for Electronic and Photonic Devices

Hybrid perovskites for device applications

Fig. 7.3 (A) The absorption coefficient as a function of wavelength for two HOIPs and several other semiconductors relevant to optoelectronic applications. (B) The exciton reduced mass μ and binding energy R* as a function of band gap for several studied HOIPs. (C) Exciton binding energy as a function of n in the quasi-2D-layered HOIP system BA2MAn-1PbnI3n+1. (D) Experimentally measured mobility as a function of temperature of MAPbI3 (black circles) and a T-3/2 fit (red dashed line) suggesting that phonons are limiting mobility. (A) Adapted with permission from Springer Nature: Nature Photonics, Green, M.A., Ho-Baillie, A., Snaith, H.J., 2004. The emergence of perovskite solar cells. Nat. Photonics 8, 506. Copyright 2014. (B) Adapted from Galkowski, K., et al., 2016. Determination of the exciton binding energy and effective masses for methylammonium and formamidinium lead tri-halide perovskite semiconductors. Energy Environ. Sci. 9 (3), 962–970 with permission of the Royal Society of Chemistry. (C) Data adapted from Blancon, J.C., et al., 2017. Extremely efficient internal exciton dissociation through edge states in layered 2D perovskites. Science 355 (6331), 1288. (D) Adapted with permission from the Wiley Publishing Group. Milot, R.L., et al., 2015. Temperature-dependent charge-carrier dynamics in CH3NH3PbI3 perovskite thin films. Adv. Funct. Mater. 25 (39), 6218–6227.

223

224

Handbook of Organic Materials for Electronic and Photonic Devices

This leads to an increase in the electrostatic potential experienced by the electron and hole, and a corresponding increase in the binding energy. In the 2D perovskite (BA)2MAn-1PbnI3n+1, the value of n varies between 1 and 5 and the corresponding exciton binding energies varies between about 400 meV for n ¼ 1 to about 200 meV for n ¼ 5 (Fig. 7.3C) (Blancon et al., 2017). In principle, this would lead to lower measured EQE values when compared to the 3D equivalents, as excitons must be dissociated in some way before the carriers can be collected. However, it has been proposed that protected edge states in the 2D materials result in highly efficient exciton dissociation and resulting device performance in spite of the larger binding energy than kT (25 meV at room temperature) (Blancon et al., 2017). The majority of the losses originate from the insulating nature of the inorganic spacing layers. One could envisage tuning the binding energy properties of the material based on the function of the device: a low binding energy would be useful for solar cell absorbers to allow the efficient extraction of carriers under operation, whereas a large binding energy with strong confinement effects would promote efficient radiative recombination in the active layer of an LED.

7.5

Charge carrier properties

HOIPs exhibit a remarkable set of charge carrier properties and dynamics, particularly in spite of the fact that typical perovskite films are solution-processed and polycrystalline and contain relatively high concentrations of trap states (Samiee et al., 2014; Stranks et al., 2014; Leijtens et al., 2016; Wu et al., 2015; Tahara et al., 2016; Wright et al., 2017). Such high densities of trap states would dramatically reduce the efficiency of any conventional inorganic semiconductor-based device, again consistent with at least some degree of defect tolerance (Stranks, 2017). In this section, we discuss some of the key properties of the charge carriers in these HOIPs.

7.5.1 Diffusion lengths The diffusion length of a carrier type in a material can be defined as the average distance that an excited carrier will travel before recombining. The diffusion length can be defined as follows: pffiffiffiffiffiffi LD ¼ Dτ, where D is the diffusion coefficient and τ is the lifetime of the excited carrier. The diffusion coefficient can be written according to the Einstein relation as follows: D¼

μkB T , q

where μ is the charge carrier mobility, kB is the Boltzmann constant, T is the temperature, and q is the charge of an electron. Early work on MAPbI3 prepared using

Hybrid perovskites for device applications

225

Cl-based precursors utilized photoluminescence-quenching measurements to reveal charge carrier lifetimes of hundreds of nanoseconds, diffusion coefficients on the order of 1–5  102 cm2/s and diffusion lengths over a micrometer (Stranks et al., 2013; Xing et al., 2013). Since those first reports, film fabrication techniques have improved considerably, as have associated defect passivation techniques, resulting in reported lifetimes of tens of microseconds and corresponding diffusion lengths of tens of micrometers (Brenes et al., 2017; deQuilettes et al., 2016a; de Quilettes et al., 2015; Zheng et al., 2017; Noel et al., 2014a). The diffusion lengths in other analogous HOIPs, such as the mixed-cation, mixedhalide Cs0.17FA0.83Pb(IxBr1-x)3 systems, have also been reported to be many microns. This means that, in practice, practically all photogenerated carriers will diffuse to the contacts as the diffusion length is many multiples of the absorption depth of visible light in HOIP films (about 100 nm). With clean interfaces and suitable band alignment between the HOIP and the contacts, it is possible to collect close to all photogenerated carriers in a solar-cell device. Sn-based HOIPs such as MASnI3 were predicted to have diffusion lengths of the order of 1 μm if the background, trap-induced carrier density could be reduced below 1015 cm3 (Noel et al., 2014b), and this prediction was recently confirmed, with measured diffusion lengths in single crystals of CsSnI3 approaching this value (Wu et al., 2017). Single crystals of FAPbBr3 and FAPbI3 both have been reported to have diffusion lengths of 19 μm and 6 μm, respectively (Zhumekenov et al., 2016).

7.5.2 Mobility The low mobilities of organic semiconductors for device applications compared to conventional semiconductors have been a challenge for device design. Organic crystals using highly conjugated aromatic systems such as rubrene have hole mobilities of up to 20 cm2/Vs, while more disordered, solution-processed organic films used in organic photovoltaic (OPV) commonly exhibit carrier mobilities in the range of about 105–103 cm2/Vs (Shuttle et al., 2010; Foster et al., 2014; Torabi et al., 2015). These values are in stark contrast with more conventional semiconductors such as GaAs, which commonly exhibit mobilities on the order of 103 cm2/Vs (Sze and Ng, 2006). One might expect that given the long diffusion lengths in HOIPs, the carrier mobilities would be correspondingly large. However, the mobilities of HOIPs are actually modest and take intermediate values between organic and inorganic semiconductors (Brenner et al., 2015). Electron and hole mobilities in single crystals of MAPbI3 and MAPbBr3 have been reported between about 20–120 cm2/Vs (Saidaminov et al., 2015; Shi et al., 2015; Dong et al., 2015). Recent work has revealed mobilities in polycrystalline MAPbI3 perovskite films that are comparable to single crystals, with the total mobility (sum of electron and hole mobilities) being measured by photoconductance to be about 90 cm2/Vs (Brenes et al., 2017). The bromide-rich analogs generally exhibit lower mobilities than the iodidebased HOIPs (Herz, 2017). Interestingly, thin-film Sn-based analogs, MASnI3, exhibit electron and hole mobilities of about 103 cm2/Vs, which are directly

226

Handbook of Organic Materials for Electronic and Photonic Devices

comparable to inorganic, single-crystalline semiconductors (Stoumpos et al., 2013). It is likely that these measurements are also affected by a self-doping (Sn2+ to Sn4+) effect in these materials (Takahashi et al., 2011; Noel et al., 2014b). In fact, care should be exercised when quoting the mobility of a HOIP sample, as there is a very wide spread of measured mobility values for otherwise identical samples, even in the case of single crystals measured with the same technique (Herz, 2017). The experimental uncertainties in these measurements are typically large. Nevertheless, the consensus is that values up to about 100 cm2/Vs for MAPbI3 and related compositions are realistic. The charge carrier mobility in perovskites is clearly limited by some intrinsic processes. Assuming a semiclassical model of electron conductivity such as the Drude model, the mobility can be written as μ ¼ mqτ∗ , where q is the electron charge, τ is the carrier momentum relaxation time (or the average time between scattering events), and m* is the charge carrier effective mass. The carrier effective masses measured experimentally and for the room-temperature phase show exciton reduced masses between 0.1 and 0.2me (where me is the electron mass), and the effective masses of carriers have been calculated to be in the same range for both carrier types (Galkowski et al., 2016; Men
endez-Proupin et al., 2014; Miyata et al., 2015). This strongly suggests that the limiting factor is a short momentum relaxation time that could be dominated either by defect scattering or by coupling of the electrons to lattice vibrations. Experiments have shown a characteristic T-3/2 dependence of the mobility with temperature, strongly suggesting that the mobility at room temperature is limited by electron–phonon interactions (Fig. 7.3D) (Milot et al., 2015; Savenije et al., 2014). Given the ionic and therefore polar nature of the metal-halide framework in HOIPs, Fr€ ohlich scattering, the scattering of carriers from the electric fields produced by longitudinal optical (LO) phonon modes (Fr€ohlich, 1954), may be the dominant interaction in these systems, as is common in other polar semiconductors (Zhou and Bernardi, 2016). Experimental analyses of the PL line broadening of FAPbI3 over a range of temperatures has demonstrated that Fr€ohlich interaction dominates at room temperature (Wright et al., 2016). This also has been confirmed by further experiments and theoretical calculations (Sendner et al., 2016; Filippetti et al., 2016). The Fr€ ohlich interaction is closely related to the concept that polarons are mobile charges that generate ionic displacements and polarizations in the vicinity of the mobile charge from which they subsequently scatter (cf. Section 7.6.3). Effectively, the charge carriers are dragging this ionic displacement with them, which slows them (Fr€ ohlich, 1954; Feynman, 1955). In other words, the Fr€ohlich interaction is a large polaron. Recent calculations of the mobility of a range of perovskites using the Feynman polaron model revealed trends in temperature-dependent mobilities consistent with experiments (Frost, 2017). The calculated mobilities are higher than those mentioned previously; however, the temperature-dependent behavior matches that of experimental results reasonably closely, and the calculated mobilities agree within a factor of 2. This further strengthens the viewpoint that mobility in perovskites is limited by these optical phonon interactions.

Hybrid perovskites for device applications

227

7.5.3 Effective masses Ambipolar transport, where both electrons and holes exhibit similar mobilities, is a desirable property for device applications. Indeed, a marked advantage of Si-based transistors over GaAs is the ability to make Si transistors p- or n-type. This is because the carrier mobilities are comparable in magnitude, μe  1400 cm2/Vs and μh  500 cm2/Vs, and so one does not incur a significant penalty relative to the n-channel if choosing to fabricate a p-channel transistor (Sze and Ng, 2006). This is in stark contrast to GaAs, which, due to its asymmetry in carrier-effective masses caused by the very heavy holes, has a large-magnitude difference in its carrier mobilities, with μe  8000 cm2/Vs and μh  400 cm2/Vs (Sze and Ng, 2006). Such asymmetry is also observed in organic mobilities: for example, often the hole mobilities are reasonable, μh  10 cm2/Vs, and can be treated with a bandlike approximation, whereas the electrons are highly polaronic and show orders of magnitude lower mobility, as is the case in naphthalene (Lee et al., 2017a). These asymmetric constraints limit device design principles such as transistor conduction channels and carrier transport layers, to name just a few. Early reports indicated that HOIPs have close to symmetric carrier lifetimes, which is indicative of balanced carrier transport (Stranks et al., 2013; Xing et al., 2013). This is surprising when compared with their inorganic counterparts such as GaAs, given that the heavy hole contribution to the hole-effective mass tends to reduce hole mobility considerably relative to electron mobility (Sze and Ng, 2006; Nakwaski, 1995). Further theoretical and experimental work since has confirmed these experimental observations, showing very small, balanced effective masses with no perceivable heavy hole contribution in either the room-temperature tetragonal phase of MAPbI3 or in related Sn-based materials (Galkowski et al., 2016; Men
endez-Proupin et al., 2014; Miyata et al., 2015; Giorgi et al., 2013; Umari et al., 2014). The hole effective masses only appear to be 10%–20% heavier than electron-effective masses, in contrast with the near order-of-magnitude difference in GaAs between the effective masses of the electrons and the heavy holes (Sze and Ng, 2006). This is promising for device considerations in which ambipolar transport is important, for example allowing efficient extraction of both generated carriers to the respective electrodes in simple planar heterojunction solar-cell architectures.

7.6

Charge carrier recombination

7.6.1 Recombination regimes Recombination in perovskites can be summarized by the general rate equation: dn ¼ G  k1 n  k2 n2  k3 n3 , dt

(7.2)

where n is the carrier concentration, G is the generation term for photoexcited electrons and holes, k1 is the monomolecular rate constant, k2 is the bimolecular (band-to-

228

Handbook of Organic Materials for Electronic and Photonic Devices

band) recombination constant and k3 is the Auger recombination constant. The dominant recombination process (k1, k2, or k3) depends on the carrier density regime. In most materials, k1 tends to be nonradiative and trap-assisted and of a Shockley-ReadHall type, depending only on the concentration of one type of carrier. This also could include photodoping effects, where one photoexcited carrier is trapped and the large concentration of the other carriers is long-lived, as has been observed in HOIPs (Stranks et al., 2014; Leijtens et al., 2016). At low excitation fluences in which carrier densities are below the density of trap states, this type of recombination dominates; this is observed in HOIPs at excitation densities comparable with solar irradiation (1013–1015 cm3) (Richter et al., 2016). Once the carrier density exceeds the trap density and subgap states are filled, k2 begins to dominate; this occurs at an excitation density of about 1016 cm3 in MAPbI3. Recombination in this regime is primarily band-to-band, radiative recombination between free electrons and holes, and the photoluminescence quantum efficiency (PLQE) increases as a result (Fig. 7.4A) (Stranks, 2017; Richter et al., 2016; Davies et al., 2018). At very high carrier densities, above 1018 cm3 for MAPbI3, k3 in the form of the nonradiative Auger recombination begins to dominate and PLQE decreases again (Fig. 7.4A). This process involves three carriers, where one excited carrier transfers its excess energy to another excited carrier in order to conserve energy, while recombining nonradiatively with a third carrier. Auger recombination limits the efficiency of silicon solar cells due to the long radiative lifetimes associated with an indirect band gap (Kerr et al., 2003). Given that HOIP solar cells typically operate in the monomolecular, trap-dominated regime, it is necessary to reduce the concentration of trap states or suitably passivate them to reach internal PLQE values close to 100% at solar fluences, as has been demonstrated with various methods (Brenes et al., 2017; deQuilettes et al., 2016a; Abdi-Jalebi et al., 2018). An important point here is that given the dependence of the recombination rate on the excitation density due to the different recombination regimes at play, measured carrier lifetimes will vary considerably with excitation density (Stranks and Petrozza, 2016). The lifetime is not a material or sample-based parameter; as a result, it is therefore essential to report the excitation density used when quoting measured lifetimes. Furthermore, given the strong absorption near the band edge of HOIPs and their relatively high dielectric constants, light emitted due to radiative recombination can be totally internally reflected and reabsorbed within the HOIP film in a process known as photon recycling (Richter et al., 2016). The measured or external PLQE of a sample will be lower than the internal PLQE, as not all emitted photons will escape the film.

7.6.2 Nonradiative losses The open-circuit voltage (Voc) of a solar cell depends on the quasi-Fermi level in the active material and on the thermodynamic potential difference between the ground and excited states (Stranks, 2017; Rau, 2007). Both quantities are maximized when all recombination is radiative, and the light is emitted back at the Sun through the same light cone in which the light was incident; any nonradiative recombination leads to

Hybrid perovskites for device applications

229

Fig. 7.4 (A) The external PLQE values measured for MAPbI3 films as a function of carrier density for both continuous-wave (black squares) and pulsed (red triangles) excitation. (B) TRPL measurements of MAPbI3 films at different carrier densities and fits of the data to extract a trap density of about 1016 cm3. (C) Internal and external PLQE values measured for triple-cation (MA,FA,CS) perovskites as a function of potassium passivation fraction, x. TRPL, time-resolved photoluminescence. (A) Adapted with permission from Springer Nature: Nature Communications, Richter, J.M., et al., 2016. Enhancing photoluminescence yields in lead halide perovskites by photon recycling and light outcoupling. Nat. Commun. 7, 13941. Copyright 2016. (B) Reprinted with permission from Stranks, S.D., et al., 2014. Recombination kinetics in organic-inorganic perovskites: excitons, free charge, and subgap states. Phys. Rev. Appl. 2 (3), 034007. Copyright 2014 by the American Physical Society. (C) Adapted with permission from Springer Nature: Nature, Abdi-Jalebi, M., et al., 2018. Maximising and stabilising luminescence in metal halide perovskite device structures. Nature 555 (7697), 497–501, Copyright 2018.

open-circuit voltage losses in a solar cell. Under carrier densities generated under solar illumination (up to about 1015 cm3), recombination is dominated by monomolecular recombination processes in which charges interact with defect sites (cf. ShockleyRead-Hall recombination). HOIPs are commonly deposited via solution or vapor to

230

Handbook of Organic Materials for Electronic and Photonic Devices

produce polycrystalline films. As a result, the most common perovskite samples have very high concentrations of extrinsic defects in addition to any intrinsic defects. While conflicting computational reports suggest a range of formation energies and depths for intrinsic point defects (Yin et al., 2014; Buin et al., 2014), the rapid ionic movement under light and electric bias suggests that formation and migration energies of defects are low, and thus corresponding intrinsic defect densities are also likely to be high. Experimental reports of high concentrations of states within the gap confirm this (Samiee et al., 2014; Leijtens et al., 2016; Adinolfi et al., 2016). The literature suggests that these trap states in MAPbI3 primarily trap electrons rather than holes (Leijtens et al., 2016), although hole traps may still be present too. Kinetic models incorporating subgap trap states have been successfully applied to explain a wide range of fluence-dependent spectroscopic data (Fig. 7.4B) (Stranks et al., 2014; Hutter et al., 2015; Yamada et al., 2014). Although HOIPs are remarkably tolerant to defects and associated trap states, recombination is still influenced by defects, as exemplified by measured internal PLQE values in films falling considerably short of 100% in most cases (Stranks, 2017). This can be visualized on the microscale using confocal PL mapping, where there is considerable heterogeneity between the local luminescence efficiency of different grains in a film (Brenes et al., 2017; de Quilettes et al., 2015; deQuilettes et al., 2016b). The luminescence efficiency typically decreases further when contacts are added because the new interfaces add further nonradiative recombination avenues due to interfacial trap states (Abdi-Jalebi et al., 2018). This is manifested in a VOC loss relative to the detailed balance radiative limit. To reduce the prevalence of nonradiative recombination, several avenues have been explored. Chemical posttreatments of the perovskite surface have shown considerable success, with a wide variety of Lewis bases, including pyridine and tri-noctylphosphine oxide (TOPO), leading to marked improvements in PLQE and lifetimes (deQuilettes et al., 2016a; de Quilettes et al., 2015; Noel et al., 2014a). Atmospheric treatments using a combination of oxygen and water result in similar improvement, with films treated in this way showing internal PLQE values as high as about 90% (Brenes et al., 2017). The inclusion of additives in the perovskite precursor solution can also influence nonradiative recombination, with the lowest VOC losses in solar cells occurring either by controlling the pH of the precursor solution (Noel et al., 2017b) or by introducing alkali halides such as RbI (Saliba et al., 2016a) or KI (Abdi-Jalebi et al., 2018) in order to manage halide vacancies (Fig. 7.4C). Nevertheless, the highest-performing perovskite solar cells to date still suffer from fractions of nonradiative decay, and it will take careful material and interfacial passivation to eliminate it entirely.

7.6.3 Unique recombination pathways One of the early indications of the high quality of HOIP semiconductors was the realization that carrier lifetimes in perovskites are very long (about 1 μs) at low charge densities, even in polycrystalline films (Stranks et al., 2013). With appropriate passivation, the measured lifetimes in these films has increased further still (Brenes et al., 2017; deQuilettes et al., 2016a). This result is particularly surprising given the fact that

Hybrid perovskites for device applications

231

lifetimes in conventional inorganic semiconductors would be extremely short in the presence of the relatively high concentration of defects and grain boundaries present in these films. While initial reports applying Langevin recombination theory to HOIPs suggested that band-to-band recombination was far slower than expected (Wehrenfennig et al., 2014), it is now widely believed that radiative recombination is an inverse absorption process and that calculated k2 values match those measured experimentally (Davies et al., 2018). Nevertheless, at low charge densities, HOIPs exhibit longer lifetimes than expected given the abundance of trap states. Several unique recombination mechanisms have been suggested that further impact recombination beyond the processes described in Eq. (7.2) and could in part explain the apparent defect tolerance, including long-lived traps, photon recycling, large polarons, and the formation of an indirect band gap (deQuilettes et al., 2018). One proposal is that the defects form traps that are shallow and trapped electrons can thermally activate (detrap) back to the conduction band, meaning that the traps do not act as Shockley-Read-Hall nonradiative recombination sites. These trapped charges may be screened from recombination with the opposite carrier. If this detrapping process is slow, then the observed lifetime of the carriers will be long. Modeling of experimental transient data has suggested that the rate of trapping is much higher than the rate of detrapping (Stranks et al., 2014). This is combined with long decays observed in transient photocurrent (Leijtens et al., 2016) and time-resolved microwave conductivity measurements (Bi et al., 2016a), which suggests that these trapping dynamics indeed may be extending the life of carriers. Alternatively, given the soft, ionic nature of HOIPs, and that Fr€ ohlich interactions limit mobility (Wright et al., 2016), large polarons may form in these materials. A polaron is the quasiparticle consisting of a charge carrier and the lattice distortion that it causes. Large polarons involve lattice distortions over many unit cells and are somewhat delocalized. The effective masses of large polarons are larger than free charges, although scattering is reduced, so they retain modest mobilities (Emin, 2012). Optical Kerr effect measurements and first-principles calculations of MAPbI3 and MAPbBr3 show the deformation of the lead halide cage with carrier injection, approximating the size of the polaron to be many unit cells in diameter, consistent with a large polaron model (Miyata et al., 2017). This report suggests that opposite structural distortions of electrons and holes lead to a barrier in charge-carrier recombination, as could the screening of the carrier’s charge by lattice distortions. A further explanation is that photon recycling in highly emissive samples could contribute to an effective, long carrier lifetime (Brenes et al., 2017; Richter et al., 2016). Finally, the Rashba effect has been proposed to impact recombination. In the presence of spin-orbit coupling (SOC) and inversion symmetry breaking of the crystal structure, carriers feel a relativistic magnetic field that splits otherwise spindegenerate band extrema in k-space, introducing an indirect band gap (Bychkov and Rashba, 1984). SOC is present in HOIPs, particularly due to the presence of lead (Even et al., 2013). Bulk inversion symmetry breaking is unlikely (Frohna et al., 2018), although dynamic or surface effects may play a role (Motta et al., 2015; Mosconi et al., 2017). If present, the Rashba effect has been predicted to slow recombination due to the formation of an indirect band gap (Azarhoosh et al., 2016), spin

232

Handbook of Organic Materials for Electronic and Photonic Devices

selection rules, or both (Zheng et al., 2015b). Indirect experimental evidence includes the measurement of an energy barrier to recombination in MAPbI3 consistent with calculated bandstructures (Hutter et al., 2016), as well as the narrowing of the band gap with pressure (Wang et al., 2017c). Angle-resolved photoemission spectroscopy of MAPbBr3 shows a valence band consistent with Rashba splitting, although this could be exaggerated by surface effects (Frohna et al., 2018; Niesner et al., 2016). The Rashba effect may be increasingly important in nanostructured architectures with large surface-to-bulk ratios due to centrosymmetry being inherently broken at the surface. These devices could have profound implications for fields such as spintronics (Kepenekian and Even, 2017).

7.6.4 Ion migration

FB-SC SC-FB

10

Scan rate: 0.15 V/s

6

0 –10 –20 0.0

(A)

Direction Jsc η (%) Voc FF (mA/cm2) (V) FB-SC

22.9

15.5 1.05 0.66

SC-FB

22.8

10.2 0.97 0.46

0.2

0.4 0.6 0.8 Applied bias (V)

Fit EA = 0.27 ± 0.06 ev

10

PL growth rate (%/second)

20

PL (arb. units)

Current density (mAcm–2)

Early in the development of HOIP-based solar cells, researchers were observing anomalous current-voltage transients and hysteresis in the current-voltage curves depending on the direction of the voltage sweep (Fig. 7.5A) (Unger et al., 2014; Snaith et al., 2014). It was these observations that first led researchers to the conclusion that the ionic conductivity of HOIPs was influencing the electrical conductivity measurements. Computational work has suggested that several defects, in particular iodide vacancies, have low formation energies and low migration activation energies. The accumulation of charged ions at interfaces affects charge extraction and the resulting current-voltage measurements (Azpiroz et al., 2015; Haruyama et al., 2015). Direct observations of ion migration under applied bias have also been observed (Yuan et al., 2015; Lee et al., 2017b).

4

1

0.1

3.5

4.0 4.5 1000/T (K–1)

5.0

2 0

1.0

1.2

(B)

1.6

1.7

1.8 1.9 Energy (eV)

2.0

Fig. 7.5 (A) Current voltage scan of a MAPbI3-based solar cell (fabricated using chloride precursors), showing considerable hysteresis depending on the direction of voltage sweep. (B) PL spectra from an MAPb(I0.6Br0.4)3 film over time (45 s in 5-s increments) under illumination (457-nm light, about 15 mW/cm2). (A) Adapted with permission from Snaith, H.J., et al., 2014. Anomalous hysteresis in perovskite solar cells. J. Phys. Chem. Lett. 5 (9), 1511–1515. Copyright 2014 American Chemical Society. (B) Reproduced with permission from the Royal Society of Chemistry. Hoke, E.T., et al., 2015. Reversible photo-induced trap formation in mixed-halide hybrid perovskites for photovoltaics. Chem. Sci. 6 (1), 613–617.

Hybrid perovskites for device applications

233

Additionally, ion migration has been observed under illumination. This was initially discovered in mixed-halide perovskites: when x in the system MAPb(I1-xBrx)3 is increased above 0.2, an anomalous red-shifting of the PL peak in combination with a splitting of the X-ray diffraction (XRD) peaks was observed (Fig. 7.5B) (Hoke et al., 2015). Under illumination, the halides segregate into higher band-gap, bromine-rich regions and lower band-gap, iodine-rich regions, in a process referred to as the Hoke effect. The photoexcited carriers are funneled into the low band-gap regions on which they recombine, giving rise to the red-shifted emission. Interestingly, the process is reversible upon leaving the sample in the dark. Curiously, the photoinduced halide migration in the presence of atmospheric molecules can lead to luminescence increases (Brenes et al., 2017, 2018; deQuilettes et al., 2016b), which highlights the complicated nature of the effect and the intimate relationship between defects (sites of nonradiative decay) and ion migration. The combination of ionic and electronic conductivity could have interesting applications in devices moving forward, although the effect also may be problematic for longer-term stability of some devices. The driving force for light-induced ion migration is still under debate, but it appears to be driven by photogenerated free carriers. There are suggestions that free holes interact with iodine ions to cause changes in their valence occupation (Kim et al., 2018), while others suggest that a charge-carrier gradient is necessary to drive this effect (Barker et al., 2017).

7.6.5 Thermal properties Heat conductivity is another intriguing property of HOIPs and once again appears to fuse selected properties of organic molecular/polymeric-type materials with those of inorganic solids. The thermal conductivity as a function of temperature for HOIP films, and single crystals has been measured by several groups, with reported values in the range of about 0.3–1 W/mK for varying HOIP compositions and phases (Pisoni et al., 2014; Heiderhoff et al., 2017). This is described as ultralow thermal conductivity and is much lower than metals such as silver (about 430 W/mK) (Gero Bernhard Martin et al., 1999), as well as semiconductors such as GaAs (about 55 W/ mK at 300 K) (Carlson et al., 1965). The thermal conductivity of HOIPs is much closer to amorphous insulators and polymers than to other crystalline solids (Choy, 1977). Molecular dynamics and lattice dynamics simulations on MAPbI3 reproduce the thermal conductivities measured experimentally (Qian et al., 2016; Whalley et al., 2016). These calculations have found that the thermal conductivity in MAPbI3 is limited by strongly anharmonic phonon modes, the slow group velocities of acoustic phonons and very short phonon lifetimes. The combination of very low thermal conductivity with moderate carrier conductivity suggests that these materials could make promising candidates for thermoelectric materials. Indeed, some groups have investigated HOIPs for these applications and have predicted moderate thermoelectric figures of merit (He and Galli, 2014; Mettan et al., 2015; Zhao et al., 2017).

234

7.7

Handbook of Organic Materials for Electronic and Photonic Devices

Device applications and challenges

7.7.1 Solar cells The first use of HOIPs in solar cells was as a replacement for more traditional dyes/ sensitizers in mesoporous, TiO2-based, DSSCs, where the perovskite infiltrates the mesoporous TiO2 and, following photoexcitation, electrons inject rapidly into the metal oxide (Kojima et al., 2009). However, HOIPs quickly outgrew this role as a dye replacement, as it was discovered that they could function well in simpler planar perovskite heterojunction architectures (Fig. 7.6A), in which the perovskites could transport electrons and holes with reasonable mobility and those charges exhibit long carrier lifetimes (Lee et al., 2012; Ball et al., 2013; Liu and Kelly, 2013) HOIP solar cells, in contrast to typical p-n junction silicon solar cells, are heterojunction solar cells. This means that the energy band variation in the device, which allows the efficient and selective extraction of carriers, is formed by contacting the perovskite with suitable wide band-gap materials with the correct work functions, rather than by graded doping of the perovskite itself. The standard device architecture for perovskite-based solar cells is that of an n-i-p device (Lee et al., 2012; Edri et al., 2013, 2014). First, an n-type contact is deposited onto a transparent conductive electrode (TCO), the intrinsic (i) perovskite is then deposited on top of this n-type contact, the p-type contact on top of the perovskite, and finally a metallic back contact to allow Ohmic contact with the cell (Fig. 7.6C). An alternative to this method is to use an inverted p-i-n device architecture, in which the device is instead fabricated on the p-type contact. This architecture could be beneficial for perovskite-silicon tandems because silicon homojunction solar cells are most commonly fabricated as a p-type wafer with an n+-doped emitter on their top surface (about 95% market share) (ITRPV, 2017). However, n-type silicon solar cells are predicted to gain considerable market share in the next five years due to their favorable carrier lifetimes, tolerance of prevalent defects, and higher efficiency (ITRPV, 2017; Bandana and Chetan, 2017; NREL, 2018). Thus, both n-i-p and p-in perovskite cells may be important for this application in the next decade. For n-i-p-based devices, both mesoporous and planar architectures are commonly used. The highest-efficiency devices utilize thin mesoporous TiO2 as the electron contact (Yang et al., 2017a; Saliba et al., 2016a; Bi et al., 2016b), and either 2,20 ,7,70 Tetrakis[N,N-di(4-methoxyphenyl)amino]-9,90 -spirobifluorene (spiroOMeTAD) (Fig. 7.6C) (Saliba et al., 2016a; Bi et al., 2016b), or poly(bis(4-phenyl)(2,4,6trimethylphenyl)amine) (PTAA) as the hole transport layer (Fig. 7.6C) (Yang et al., 2017a). The highest-efficiency devices in this architecture in the literature is 22.1%, reported by Yang et al. (2017a) (Fig. 7.6B), although to date a certified higher efficiency of 23.3% has been reported, but not yet published (NREL, 2018). Planar architectures also have been considered due to their lower fabrication complexity and the possibility for alternative, lower-temperature processing methods than the high-temperature sintering required for TiO2. Compact TiO2 contacts have been employed with reasonable success (Liu et al., 2013; Yang et al., 2016), as well as TiO2/C60 bilayers (P
erez-del-Rey et al., 2018),

Fig. 7.6 (A) Schematics (top to bottom) of the dye-sensitized, the mesoporous, and the planar perovskite solar cells. (B) JV curves of high-performing perovskite solar cells. (C) Schematic and cross-sectional scanning electron microscope (SEM) image of a high-performing n-i-p solar cell. (D) Energy-level diagram showing many common HOIPs, p-type, and n-type contacts used in the fabrication of high efficiency HOIP solar cells. (A) Adapted with permission from the Nature Publishing Group. Gr€atzel M., 2014. The light and shade of perovskite solar cells. Nat. Mater. 13, 838. (B) Data adapted from Yang, W.S., et al., 2017a. Iodide management in formamidinium-lead-halide–based perovskite layers for efficient solar cells. Science 356 (6345), 1376; Saliba, M., et al., 2016a. Incorporation of rubidium cations into perovskite solar cells improves photovoltaic performance. Science, doi:10.1126/science.aah5557. (C) Adapted with permission from Wiley Publishing Group. Aitola, K., et al., 2017. High temperature-stable perovskite solar cell based on low-cost carbon nanotube hole contact. Adv. Mater. 29 (17), 1606398. (D) Reproduced from Chueh, C.-C., Li, C.-Z., Jen, A.K.Y., 2015. Recent progress and perspective in solution-processed interfacial materials for efficient and stable polymer and organometal perovskite solar cells. Energy Environ. Sci. 8 (4), 1160–1189. with permission of the Royal Society of Chemistry.

236

Handbook of Organic Materials for Electronic and Photonic Devices

and efficiencies above 20% have been reported on small areas (Tan et al., 2017). Other planar electron contacts include C60 (Yoon et al., 2016), SnO2/C60 bilayers (Wang et al., 2017b), and SnOx, which has produced the highest fully planar n-i-p conversion efficiency to date, at 20.7% (Anaraki et al., 2016). Alternative hole contacts for the ni-p architecture include N4,N4,N400 ,N400 -tetra([1,10 -biphenyl]-4-yl)-[1,10 :40 ,100 -terphenyl]-4,400 -diamine (TaTm) (P
erez-del-Rey et al., 2018), and 20 ,70 -bis(bis(4methoxyphenyl)amino)spiro[cyclopenta[2,1-b:3,4-b0 ]dithiophene-4,90 -fluorene] (FDT) (Saliba et al., 2016c). Devices with the p-i-n configuration have been primarily investigated in the planar device architecture. The most widely reported contacts have been the organic PTAA and the inorganic NiOx. Devices with PTAA were reported with high efficiency, and due to the hydrophobicity of the PTAA, they typically yield larger perovskite grains than other commonly used hole transport layers (Bi et al., 2015). PTAA devices are often paired with the combination of C60 and bathocuproine (BCP) as electron-collecting layers. Refinements to the interface between the perovskite and the C60, by either cross-linking and doping the fullerene layer (Bai et al., 2016) or passivating surface trap states with quaternary ammonium-halide ions (Zheng et al., 2017), have enabled efficiencies above 20%. NiOx has been investigated due to its potential scalable deposition methods and stability; however, initial efforts have produced relatively low efficiencies, particularly low open-circuit voltages (Wang et al., 2014; Jeng et al., 2014). The problems with VOC have been attributed to shunt pathways introduced by poor wetting of the perovskite film to the NiOx as well as trap states at the NiOx/perovskite interface (Manders et al., 2013). Doping the NiOx with Mg/Li, Cu, and Cs (Chen et al., 2015a, 2017b; Jung et al., 2015) has led to an efficiency as high as 19% in the Cs case, although their performance is still considerably lower than the alternatives. A plethora of organic molecules and polymers also have been used as the p-type contact in p-i-n solar cells, including poly(N,N0 -bis(4-butylphenyl)-N,N0 -bis(phenyl)benzidine) (poly-TPD) (Cheng et al., 2017), and poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) (Chiang et al., 2017). A summary of the energy levels for many common HOIPs and p- and n-type contacts is shown in Fig. 7.6D. One of the most attractive applications for perovskites are tandem solar cells, in which two or more solar cells are stacked on top of one another monolithically or mechanically. Tandem solar cells are an exciting avenue to increase the efficiency of solar cells beyond the single-junction Shockley Queisser limit. By splitting the solar spectrum between two or more subcells, one can effectively absorb the same number of photons, but at a higher potential because each photoexcited carrier experiences less thermalization, resulting in boosts in efficiency beyond that of single-junction solar cells. Simulations under real-world conditions reveal that perovskite-silicon and perovskite-perovskite tandems could both reach practical efficiencies of over 30% (Fig. 7.7A) (Horantner and Snaith, 2017; H€ orantner et al., 2017). Mechanically stacked or four-terminal tandems (4T) involve the combination of two separate solar cells stacked on top of one another. The top, wide-band-gap cell contains two transparent electrodes and is not connected in series with the bottom, narrow band-gap cell.

Hybrid perovskites for device applications

237

(Continued) Fig. 7.7 See legend on next page.

238

Fig. 7.7, Cont’d (A) Contour plot of the maximum practical efficiency of a perovskite-perovskite tandem solar cell as a function of top and bottom cell band gaps with existing architectures. (B) EQE plot and (C) cross-section schematic of the record 23.6% efficiency monolithic perovskite-silicon tandem. (D) Current-voltage plot of the perovskite-perovskite tandem solar cell, as reported by Eperon et al. (E) Maximum power point tracking and environmental test conditions of SnO2 and TiO2-based HOIP solar cells for 1000 h. (A) Adapted with permission from H€orantner, M.T., et al., 2017. The potential of multijunction perovskite solar cells. ACS Energy Lett. 2 (10), 2506–2513. Copyright 2017 American Chemical Society. (C) Adapted with permission from Springer Nature: Nature Energy, Bush, K.A., et al., 2017. 23.6%-efficient monolithic perovskite/silicon tandem solar cells with improved stability. Nat. Energy 2, 17009, Copyright 2017. (D) Data adapted from Eperon, G.E., et al., 2016. Perovskite-perovskite tandem photovoltaics with optimized band gaps. Science 354 (6314), 861. (E) Adapted with permission from Springer Nature: Nature Energy, Christians, J.A., et al., 2018. Tailored interfaces of unencapsulated perovskite solar cells for >1,000 hour operational stability. Nat. Energy 3 (1), 68–74. Copyright 2018.

Handbook of Organic Materials for Electronic and Photonic Devices

Hybrid perovskites for device applications

239

In contrast, monolithic tandems (2T) involve the fabrication of the top cell directly on top of the bottom cell with an electrical interconnect—either a tunnel junction or recombination layer—to connect the cells in series. Given that the cells are in series, the current generated in each subcell must match, as the overall current will be limited by the lowest current subcell, while the voltages are additive. Perovskite-silicon tandems are likely to be the first HOIP devices available on the market, as they are not competing directly with silicon, but rather boosting the existing technology. Tandems require transparent contacts in order for the top cell to allow light to reach the bottom cell. In the perovskite-silicon tandem architecture, a variety of perovskite compositions have been combined with numerous transparent contact materials, including silver nanowires (Bailie et al., 2015; Mailoa et al., 2015), ZnO nanoparticles + ITO (Bush et al., 2016), SnO2 + ITO (Bush et al., 2017), and MoOx + indium zinc oxide (IZO) (Duong et al., 2017; Shen et al., 2018). The highest-reported perovskite-silicon tandem efficiencies to date stand at 23.6% for 2 T (Fig. 7.7B and C) (Bush et al., 2017), and 26.4% for 4 T (Duong et al., 2017). The replacement of Pb with Sn to achieve low band gaps (about 1.2 eV) also has allowed the fabrication of perovskite-perovskite tandems. This concept was first demonstrated by Eperon et al., showing 2 T efficiencies of 17.0% and 4 T efficiencies above 20% (Fig. 7.7D) (Eperon et al., 2016). Further works have seen the efficiencies of the perovskite-perovskite 2T tandems increase above 18% (Rajagopal et al., 2017; Forga´cs et al., 2017). Although there have been concerns about the environmental stability of Sn-based HOIPs (Leijtens et al., 2017), these solar cells may well be the ultimate goal for perovskite solar cells—in principle, the entire cell can be deposited using scalable vapor/solvent deposition methods on roll-to-roll flexible substrates, but they are also capable of achieving efficiencies above the Shockley-Queisser limit for a single junction cell. Alternatively, HOIPs may be paired with organic bulk heterojunction solar cells, although the efficiency gains reported thus far have been more modest (Chen et al., 2015b; Liu et al., 2016a). Finally, we discuss the stability of HOIP solar cells. Although the initial devices reported were very unstable (Kojima et al., 2009), optimization of contacts, interfaces, and the perovskite itself have led to considerable enhancements in stability. Initial reports of intrinsic thermal instability of MAPbI3 above 85°C and degradation in response to applied fields and humidity suggested that the MA cation would not withstand conventional accelerated stability tests (Leijtens et al., 2015; Conings et al., 2015) and led to compositional engineering of HOIPs in pursuit of increased stability and efficiency. FA- and Cs-based perovskites have greater thermal stability, but poorer phase stability (Eperon et al., 2014; Beal et al., 2016). This was overcome by alloying FA and Cs at the A-site (Li et al., 2016b; McMeekin et al., 2016). Efficiency was increased further by reintroducing small fractions of MA to the A-site (Saliba et al., 2016b). This system, along with the archetypal MAPbI3, are current workhorses in the field (Correa-Baena et al., 2017). Enhanced thermal stability, reduced ion migration, enhanced charge carrier mobilities and increased efficiencies have been observed with the addition of small amounts of Rb and K (Saliba et al., 2016a; Abdi-Jalebi et al., 2018; Hu et al., 2018). Encapsulated perovskite solar cells are now passing

240

Handbook of Organic Materials for Electronic and Photonic Devices

International Electrotechnical Commission (IEC) tests such as the 85°C, 85% relative humidity, 1000-h damp heat test (Bush et al., 2017), as well as the 40°C to 85°C, 200-cycle test (Cheacharoen et al., 2018). Unencapsulated solar cells have been shown to retain >90% of their efficiency with continuous maximum-power point tracking over >500 hours, with much of the degradation reversible by leaving the cell in the dark (Fig. 7.7E) (Bush et al., 2017; Jodlowski et al., 2017; Christians et al., 2018). Demonstrations of stability in 100-cm2 modules retaining at least the initial device performance while held at an open circuit for over a year have been shown (Grancini et al., 2017). However, the combination of the highest efficiencies and stabilities, particularly in large areas, have yet to be realized, and this is arguably the main challenge for commercializing HOIP solar cells. It is likely that the choice of contact will be essential for both stability and performance.

7.7.2 Light-emitting diodes (LEDs) The 3D perovskite structure has a band gap that in principle can be tuned across the entire visible spectrum by tweaking the composition (Sutherland and Sargent, 2016); the nanostructured analogs such as nanocrystals and nanoplatelets allow further control of the band gap (Stoumpos et al., 2016; Protesescu et al., 2015). The combination of facile band-gap tuning, high luminescence yields, and low-cost deposition methods make HOIPs very promising candidates for light-emission applications. Furthermore, the significant photon-recycling effects in perovskite structures due to strong overlap between absorption and emission spectra means that there is an opportunity to increase the light outcoupled significantly in the forward direction, which otherwise would be laterally waveguided out in an OLED. Perovskites also have the advantage of a peakemission-wavelength full width at half maximum (FWHM) that can be tuned to be narrow or very broad, depending on the desired application (Sadhanala et al., 2015; Zhang et al., 2016a; Wu et al., 2018). The first LEDs were demonstrated in the near-infrared (NIR) region using a thinemissive layer of MAPb(IxCl1-x)3, TiO2 as an electron-injection layer and poly(9,90 dioctylfluorene) (F8) as the hole injection layer, demonstrating a maximum EQE of a modest 0.76% (Tan et al., 2014). Efficient LEDs in the NIR have since been demonstrated by Yuan et al. using quasi-2D perovskites of the form PEA2MAn-1PbnI3n+1 where n has an average value of 5, and peak EQE values of above 8% and radiances up to 80 W sr1 m2 were produced (Yuan et al., 2016). The most efficient NIR HOIP LEDs also have been fabricated using a mixture of quasi-2D perovskite layers arranged in self-assembled, multiple quantum well structures with peak EQEs at time of writing above 11% and a power-to-light conversion efficiency of 5.5% at a current density of 100 mA/cm2. Monochromatic LEDs have been fabricated with emission wavelengths across the visible spectrum. Red LEDs using either nanocrystals of CsPb(Br/I)3 emitting at 688 nm (Zhang et al., 2016b) or multiple quantum wells with perovskites of the form (NMA)2Csn-1PbnI3n+1, where NMA is 1-naphthylmethylammonium emitting at 694 nm, both show peak EQE values above 7% (Chang et al., 2018). 2D sheets of the form PEA2FAn-1PbnBr3n+1 were used to create green LEDs emitting at 532 nm,

Hybrid perovskites for device applications

241

with peak EQE values above 14% (Yang et al., 2018), the highest for any HOIP LED so far (Fig. 7.8A and B). Blue LEDs also have been reported using nanocrystals of the form CsPb(BrxCl1-x)3 emitting at 490 nm with peak EQE values of 1.9% (Pan et al., 2016), though we have yet to see truer blue LEDs. There have also been reports of white HOIP LEDs using polymer-HOIP composites, although at comparatively low efficiencies (Yao et al., 2017). LEDs face similar stability challenges to solar cells, although perhaps more extreme conditions, given the high biases that must be applied across the diodes to produce light. The current state of the art for stability in perovskite LEDs shows lifetimes of the order of 10 h, not years (Tsai et al., 2018).

7.7.3 Photodetectors HOIPs have long carrier lifetimes, moderate carrier mobilities and low intrinsic carrier densities (Shi et al., 2015), which make them promising candidates for photodetection applications (Ahmadi et al., 2017). It took several years after the initial solar cell reports for photodetectors to become feasible due to the incompatibility of photodetection with mesoporous scaffolds (Garcı´a de Arquer et al., 2017). The first reports used device architectures analogous to solar cells, and showed tremendous promise with very large detectivities across the visible spectrum (closely related to the signal to noise ratio) of about 1012 Jones and linear dynamic ranges of well over 100 dB (Lin et al., 2015; Sutherland et al., 2015). Trap passivation allowed for the lowering of the dark current and increasing the signal to noise ratio further, resulting in light intensities of the order about pW cm2 able to be readily detected in the linear regime (Fig. 7.8C) (Fang and Huang, 2015). Alternative device architectures such as thin film transistors (Wang et al., 2015), and bilayers (Ma et al., 2016) have been reported with similar detectivities, highlighting the flexibility of HOIPs for detection applications. There have also been demonstrations of HOIPs in X-ray detection. Producing sensitive X-ray detectors is important because if one can detect a lower intensity of X-rays, one can expose the patient or scientific sample to a lower dose of potentially harmful radiation (Brenner et al., 2001). The heavy atoms such as lead and iodine make lead-based HOIPs interesting candidates for this application because X-ray attenuation depends heavily on the atomic number (Hubbell and Seltzer, 1995). The attenuation coefficients for X-rays tend to be much smaller than for visible photons, which means that X-ray detectors must in general be thick to produce appreciable signal. In order for the carriers to be extracted, the mobility-lifetime product (μτ) must be large. Early reports using a device based on a sintered MAPbI3 pellet demonstrated sensitivities, measured in terms of the charge extracted at constant bias from different doses of X-ray radiation, of about 2500 μC GyAir cm2, which is very close to commercialized CdTe detectors (Shrestha et al., 2017). These measurements were performed at high electric fields of 0.2 V μm1. Devices using single crystals of MAPbBr3 have also been developed and exhibit a considerably lower sensitivity of about 80 μC GyAir cm2 but at a far lower applied field of 1  104 V μm1, which is considerably more sensitive than common commercialized technologies at such a low bias (Wei et al., 2016). Both studies indicate that the large μτ value for HOIPs plays a crucial role in their performance. Similar devices have also been fabricated

242 Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 7.8 (A) An SEM cross section and (B) EQE and luminance at 532 nm versus injected current density for a PEA2FAn-1PbnBr3n+1 (n ¼ 3)-based HOIP LED with the highest report peak EQE to date. (C) Specific detectivity as a function of wavelength for an MAPbI3-based photodetector capable of resolving incident power densities as low as about 1 pW/cm2. (D) Schematic and (E) laser emission spectra as a function of wavelength and pump fluence of a HOIP microsphere-based laser exhibiting the highest reported Q-factors to date. (B) Adapted with permission from Springer Nature: Nature Communications, Yang, X., et al., 2018. Efficient green light-emitting diodes based on quasi-two-dimensional composition and phase engineered perovskite with surface passivation. Nat. Commun. 9 (1), 570. Copyright 2018. (C) Adapted with permission from Wiley Publishing Group. Fang, Y., Huang, J., 2015. Resolving weak light of sub-picowatt per square centimeter by hybrid perovskite photodetectors enabled by noise reduction. Adv. Mater. 27 (17), 2804–2810. (E) Adapted with permission from Tang, B., et al., 2017. Single-mode lasers based on cesium lead halide perovskite submicron spheres. ACS Nano 11 (11), 10681–10688. Copyright 2017 American Chemical Society.

Hybrid perovskites for device applications

243

to detect gamma rays (Wei et al., 2017b). The low cost of HOIP fabrication makes them ideal candidates to replace conventional semiconductors in this space.

7.7.4 Lasing materials The long carrier lifetimes, tuneable band gaps and relatively high internal PLQE values, particularly at high fluences, make HOIPs promising candidates for lasing applications (Deschler et al., 2014). Carriers can be readily pumped into the excited state, remain there for sufficient time as to create a carrier population inversion, and will recombine primarily radiatively, resulting in high efficiency, coherent, stimulated emission. The first reports of a HOIP-based laser used MAPb(IxCl1-x)3 as a gain medium in a cavity that was optically pumped with an excitation of 2.33 eV and produced high intensity, narrow bandwidth emission at about 1.6 eV (Deschler et al., 2014). Since this first investigation, the HOIP lasing community has turned to nanostructured or lower-dimensional perovskite materials in a similar manner to the LED community due to their intrinsically high PLQE values. Reports using colloidal CsPbX3 (X ¼ I, Br, Cl) nanocrystals have shown lasing across the visible range in thin films using silica microspheres as cavities that trap light via total internal reflection with low pump thresholds (PTh), the energy the sample must be pumped with to show lasing, of about 5 μJ cm2 (Yakunin et al., 2015). CsPbX3 microspheres have also been used as their own cavity and gain medium simultaneously in single-mode lasers, showing extremely low PTh of about 0.4 μJ cm2 and the highest lasing quality factors (Q about 6000) ever reported for this type of laser (Fig. 7.8D and E) (Tang et al., 2017). MAPbX3 nanowires have also been used as simultaneous cavity and gain media as the nanowire acts as a waveguide with reflecting material at each end (Zhu et al., 2015; Liu et al., 2017), with low PTh values of 0.6 μJ cm2 and Q factors as high as 3600 reported by Zhu et al. Replacing the MA cation in nanowires with FA or Cs has produced similarly high-quality lasing (Q factor for FA nanowires of 2300 and PTh of about 6 μJ cm2) but increased stability of up to 109 cycles (Fu et al., 2016; Eaton et al., 2016), a marked improvement over the 107 reported for MA counterparts. However, the repetition rate of the pump lasers mean that 109 cycles only corresponds to a lifetime of about 60 min. These nanowires are actually 3D HOIPs either selectively grown in one dimension through the slow release of precursors or the confinement of precursor solutions. HOIP nanoplatelets grown on lithographically patterned hexagonal boron nitride act as an ordered array of micro-lasers (Liu et al., 2016b). The hexagonal platelets act as their own cavity and gain medium and exhibit decent Q factors of about 1200 and moderate PTh of about 10 μJ cm2 depending on the platelet size. 3D HOIPs deposited into regular arrays such as photonic crystals, gratings and patterned polymeric frameworks also exhibit high-quality lasing (Brenner et al., 2016; Sch€ unemann et al., 2017; Saliba et al., 2016d).

7.8

Conclusions

The hybrid nature of HOIPs in both their structure and their properties has allowed for an explosion in research interest in a wide range of optoelectronic applications. In many cases, the best properties from both inorganic and organic materials are

244

Handbook of Organic Materials for Electronic and Photonic Devices

inherited by these materials. The rapid, simple and cheap fabrication methods have enabled this boon and have resulted in the fastest rise in efficiency of any solar cell technology in history (cf. Fig. 7.1). It seems that given the rapid progress of the field and increasing public interest, commercialization is imminent. However, the challenges that face HOIPs are great and overcoming these challenges will determine whether HOIPs will remain a niche material or a center-stage semiconductor. Stability is a major concern and the wider field must now focus on making high efficiency, large-area devices and modules stable over long time periods in a range of atmospheric conditions; recent developments are encouraging on this front. While the cost of perovskite fabrication is low, the cost of silicon solar cell fabrication is lowering by the day, with costs plummeting well below a dollar per watt (ISE, 2017). It remains to be seen whether HOIPs will be able to compete with the moving price target of the incumbent silicon, though tandem structures with high performance show tremendous promise to do so. In any case, HOIPs are a fascinating system of fundamental study and their unique properties and applications will be a topic of interest for many years to come.

References Abdi-Jalebi, M., et al., 2018. Maximising and stabilising luminescence in metal halide perovskite device structures. Nature 555 (7697), 497–501. Adinolfi, V., et al., 2016. The in-gap electronic state spectrum of methylammonium lead iodide single-crystal perovskites. Adv. Mater. 28 (17), 3406–3410. Ahmadi, M., Wu, T., Hu, B., 2017. A review on organic–inorganic halide perovskite photodetectors: device engineering and fundamental physics. Adv. Mater. 29 (41), 1605242. Ahn, N., et al., 2015. Highly reproducible perovskite solar cells with average efficiency of 18.3% and best efficiency of 19.7% fabricated via Lewis base adduct of lead(II) iodide. J. Am. Chem. Soc. 137 (27), 8696–8699. Akkerman, Q.A., et al., 2015. Tuning the optical properties of cesium lead halide perovskite nanocrystals by anion exchange reactions. J. Am. Chem. Soc. 137 (32), 10276–10281. Anaraki, E.H., et al., 2016. Highly efficient and stable planar perovskite solar cells by solutionprocessed tin oxide. Energy Environ. Sci. 9 (10), 3128–3134. Anaya, M., et al., 2017b. Strong quantum confinement and fast photoemission activation in CH3NH3PbI3 perovskite nanocrystals grown within periodically mesostructured films. Adv. Opt. Mater. 5 (8), 1601087. Asghar, M.I., et al., 2017. Device stability of perovskite solar cells—a review. Renew. Sust. Energ. Rev. 77 (Suppl C), 131–146. Azarhoosh, P., et al., 2016. Research update: relativistic origin of slow electron-hole recombination in hybrid halide perovskite solar cells. APL Mater. 4 (9), 091501. Azpiroz, J.M., et al., 2015. Defect migration in methylammonium lead iodide and its role in perovskite solar cell operation. Energy Environ. Sci. 8 (7), 2118–2127. Bai, Y., et al., 2016. Enhancing stability and efficiency of perovskite solar cells with crosslinkable silane-functionalized and doped fullerene. Nat. Commun. 7, 12806. Bailie, C.D., et al., 2015. Semi-transparent perovskite solar cells for tandems with silicon and CIGS. Energy Environ. Sci. 8 (3), 956–963. Ball, J.M., et al., 2013. Low-temperature processed meso-superstructured to thin-film perovskite solar cells. Energy Environ. Sci. 6 (6), 1739–1743.

Hybrid perovskites for device applications

245

Bandana, S., Chetan, S.S., 2017. N-type solar cells: advantages, issues, and current scenarios. Mater. Res. Express 4 (7), 072001. Barker, A.J., et al., 2017. Defect-assisted photoinduced halide segregation in mixed-halide perovskite thin films. ACS Energy Lett. 2 (6), 1416–1424. Beal, R.E., et al., 2016. Cesium lead halide perovskites with improved stability for tandem solar cells. J. Phys. Chem. Lett. 7 (5), 746–751. Bi, C., et al., 2015. Non-wetting surface-driven high-aspect-ratio crystalline grain growth for efficient hybrid perovskite solar cells. Nat. Commun. 6, 7747. Bi, Y., et al., 2016a. Charge carrier lifetimes exceeding 15 μs in methylammonium lead iodide single crystals. J. Phys. Chem. Lett. 7 (5), 923–928. Bi, D., et al., 2016b. Efficient luminescent solar cells based on tailored mixed-cation perovskites. Sci. Adv. 2 (1), e1501170. https://doi.org/10.1126/sciadv.1501170. Blancon, J.C., et al., 2017. Extremely efficient internal exciton dissociation through edge states in layered 2D perovskites. Science 355 (6331), 1288. Brandt, R.E., et al., 2015. Identifying defect-tolerant semiconductors with high minority-carrier lifetimes: beyond hybrid lead halide perovskites. MRS Commun. 5 (2), 265–275. Brenes, R., et al., 2017. Metal halide perovskite polycrystalline films exhibiting properties of single crystals. Joule 1 (1), 155–167. Brenes, R., et al., 2018. The impact of atmosphere on the local luminescence properties of metal halide perovskite grains. Adv. Mater. 30, 1706208. Brenner, D.J., et al., 2001. Estimated risks of radiation-induced fatal cancer from pediatric CT. Am. J. Roentgenol. 176 (2), 289–296. Brenner, T.M., et al., 2015. Are mobilities in hybrid organic–inorganic halide perovskites actually “high”? J. Phys. Chem. Lett. 6 (23), 4754–4757. Brenner, P., et al., 2016. Highly stable solution processed metal-halide perovskite lasers on nanoimprinted distributed feedback structures. Appl. Phys. Lett. 109 (14), 141106. Buin, A., et al., 2014. Materials processing routes to trap-free halide perovskites. Nano Lett. 14 (11), 6281–6286. Burschka, J., et al., 2013. Sequential deposition as a route to high-performance perovskitesensitized solar cells. Nature 499, 316. Bush, K.A., et al., 2016. Thermal and environmental stability of semi-transparent perovskite solar cells for tandems enabled by a solution-processed nanoparticle buffer layer and sputtered ITO electrode. Adv. Mater. 28 (20), 3937–3943. Bush, K.A., et al., 2017. 23.6%-efficient monolithic perovskite/silicon tandem solar cells with improved stability. Nat. Energy 2, 17009. Bush, K.A., et al., 2018. Compositional engineering for efficient wide band gap perovskites with improved stability to photoinduced phase segregation. ACS Energy Lett. 3 (2), 428–435. Bychkov, Y.A., Rashba, E.I., 1984. Properties of a 2D electron gas with lifted spectral degeneracy. J. Exp. Theor. Phys. Lett. 39 (2), 78–81. Carlson, R.O., Slack, G.A., Silverman, S.J., 1965. Thermal conductivity of GaAs and GaAs1  xPx laser semiconductors. J. Appl. Phys. 36 (2), 505–507. Chang, J., et al., 2018. Enhanced performance of red perovskite light-emitting diodes through the dimensional tailoring of perovskite multiple quantum wells. J. Phys. Chem. Lett. 9 (4), 881–886. Cheacharoen, R., et al., 2018. Design and understanding of encapsulated perovskite solar cells to withstand temperature cycling. Energy Environ. Sci. 11 (1), 144–150. Chen, O., et al., 2013. Compact high-quality CdSe–CdS core–shell nanocrystals with narrow emission linewidths and suppressed blinking. Nat. Mater. 12, 445. Chen, Q., et al., 2014. Planar heterojunction perovskite solar cells via vapor-assisted solution process. J. Am. Chem. Soc. 136 (2), 622–625.

246

Handbook of Organic Materials for Electronic and Photonic Devices

Chen, W., et al., 2015a. Efficient and stable large-area perovskite solar cells with inorganic charge extraction layers. Science 350 (6263), 944. Chen, C.-C., et al., 2015b. Perovskite/polymer monolithic hybrid tandem solar cells utilizing a low-temperature, full solution process. Mater. Horiz. 2 (2), 203–211. Chen, Z., et al., 2017a. Thin single crystal perovskite solar cells to harvest below-bandgap light absorption. Nat. Commun. 8 (1), 1890. Chen, W., et al., 2017b. Cesium doped NiOx as an efficient hole extraction layer for inverted planar perovskite solar cells. Adv. Energy Mater. 7 (19), 1700722. Chen, Y., et al., 2018. 2D Ruddlesden–popper perovskites for optoelectronics. Adv. Mater. 30 (2), 1703487. Cheng, Y., et al., 2017. 18% high-efficiency air-processed perovskite solar cells made in a humid atmosphere of 70% RH. Solar RRL 1 (9), 1700097. Chiang, C.-H., et al., 2017. The synergistic effect of H2O and DMF towards stable and 20% efficiency inverted perovskite solar cells. Energy Environ. Sci. 10 (3), 808–817. Chondroudis, K., Mitzi, D.B., 1999. Electroluminescence from an organic inorganic perovskite incorporating a quaterthiophene dye within lead halide perovskite layers. Chem. Mater. 11 (11), 3028–3030. Choy, C.L., 1977. Thermal conductivity of polymers. Polymer 18 (10), 984–1004. Christians, J.A., et al., 2018. Tailored interfaces of unencapsulated perovskite solar cells for >1,000 hour operational stability. Nat. Energy 3 (1), 68–74. Cody, G.D., 1992. Urbach edge of crystalline and amorphous silicon: a personal review. J. NonCryst. Solids 141, 3–15. Conings, B., et al., 2015. Intrinsic thermal instability of methylammonium lead trihalide perovskite. Adv. Energy Mater. 5 (15), 1500477. Correa-Baena, J.-P., et al., 2017. Promises and challenges of perovskite solar cells. Science 358 (6364), 739. Cox, D.E., et al., 2001. Universal phase diagram for high-piezoelectric perovskite systems. Appl. Phys. Lett. 79 (3), 400–402. D’Innocenzo, V., et al., 2014. Excitons versus free charges in organo-lead tri-halide perovskites. Nat. Commun. 5, 3586. Davies, C.L., et al., 2018. Bimolecular recombination in methylammonium lead triiodide perovskite is an inverse absorption process. Nat. Commun. 9 (1), 293. de Quilettes, D.W., et al., 2015. Impact of microstructure on local carrier lifetime in perovskite solar cells. Science 348 (6235), 683. De Wolf, S., et al., 2014. Organometallic halide perovskites: sharp optical absorption edge and its relation to photovoltaic performance. J. Phys. Chem. Lett. 5 (6), 1035–1039. Deng, Y., et al., 2015. Scalable fabrication of efficient organolead trihalide perovskite solar cells with doctor-bladed active layers. Energy Environ. Sci. 8 (5), 1544–1550. deQuilettes, D.W., et al., 2016a. Photoluminescence lifetimes exceeding 8 μs and quantum yields exceeding 30% in hybrid perovskite thin films by ligand passivation. ACS Energy Lett. 1 (2), 438–444. deQuilettes, D.W., et al., 2016b. Photo-induced halide redistribution in organic–inorganic perovskite films. Nat. Commun. 7, 11683. deQuilettes, D.W., et al., 2018. Unique Recombination Pathways in Halide Perovskites. Submitted. Deschler, F., et al., 2014. High photoluminescence efficiency and optically pumped lasing in solution-processed mixed halide perovskite semiconductors. J. Phys. Chem. Lett. 5 (8), 1421–1426.

Hybrid perovskites for device applications

247

Dong, Q., et al., 2015. Electron-hole diffusion lengths >175 μm in solution-grown CH3NH3PbI3 single crystals. Science 347 (6225), 967. Du, X.-h., et al., 1998. Crystal orientation dependence of piezoelectric properties of lead zirconate titanate near the morphotropic phase boundary. Appl. Phys. Lett. 72 (19), 2421–2423. Duong, T., et al., 2017. Rubidium multication perovskite with optimized bandgap for perovskite-silicon tandem with over 26% efficiency. Adv. Energy Mater. 7 (14), 1700228. Eaton, S.W., et al., 2016. Lasing in robust cesium lead halide perovskite nanowires. Proc. Natl. Acad. Sci. 113 (8), 1993. Edri, E., et al., 2013. High open-circuit voltage solar cells based on organic–inorganic lead bromide perovskite. J. Phys. Chem. Lett. 4 (6), 897–902. Edri, E., et al., 2014. Elucidating the charge carrier separation and working mechanism of CH3NH3PbI3 xClx perovskite solar cells. Nat. Commun. 5, 3461. Emin, D., 2012. Polarons. Cambridge University Press, Cambridge. Eperon, G.E., et al., 2014. Formamidinium lead trihalide: a broadly tunable perovskite for efficient planar heterojunction solar cells. Energy Environ. Sci. 7 (3), 982–988. Eperon, G.E., et al., 2015. Inorganic caesium lead iodide perovskite solar cells. J. Mater. Chem. A 3 (39), 19688–19695. Eperon, G.E., et al., 2016. Perovskite-perovskite tandem photovoltaics with optimized band gaps. Science 354 (6314), 861. Even, J., et al., 2013. Importance of spin–orbit coupling in hybrid organic/inorganic perovskites for photovoltaic applications. J. Phys. Chem. Lett. 4 (17), 2999–3005. Fang, Y., Huang, J., 2015. Resolving weak light of sub-picowatt per square centimeter by hybrid perovskite photodetectors enabled by noise reduction. Adv. Mater. 27 (17), 2804–2810. Feynman, R.P., 1955. Slow electrons in a polar crystal. Phys. Rev. 97 (3), 660–665. Filippetti, A., et al., 2016. Low electron-polar optical phonon scattering as a fundamental aspect of carrier mobility in methylammonium lead halide CH3NH3PbI3 perovskites. Phys. Chem. Chem. Phys. 18 (22), 15352–15362. Forga´cs, D., et al., 2017. Efficient monolithic perovskite/perovskite tandem solar cells. Adv. Energy Mater. 7 (8), 1602121. Foster, S., et al., 2014. Electron collection as a limit to polymer:PCBM solar cell efficiency: effect of blend microstructure on carrier mobility and device performance in PTB7:PCBM. Adv. Energy Mater. 4 (14), 1400311. Fr€ ohlich, H., 1954. Electrons in lattice fields. Adv. Phys. 3 (11), 325–361. Frohna, K., et al., 2018. Inversion symmetry and bulk Rashba effect in methylammonium lead iodide perovskite single crystals. Nat. Commun. 9 (1), 1829. Frost, J.M., 2017. Calculating polaron mobility in halide perovskites. Phys. Rev. B 96 (19), 195202. Fu, Y., et al., 2016. Nanowire lasers of formamidinium lead halide perovskites and their stabilized alloys with improved stability. Nano Lett. 16 (2), 1000–1008. Galkowski, K., et al., 2016. Determination of the exciton binding energy and effective masses for methylammonium and formamidinium lead tri-halide perovskite semiconductors. Energy Environ. Sci. 9 (3), 962–970. Garcı´a de Arquer, F.P., et al., 2017. Solution-processed semiconductors for next-generation photodetectors. Nat. Rev. Mater. 2, 16100. Gero Bernhard Martin, F., et al., 1999. Quantitative thermal conductivity measurements with nanometre resolution. J. Phys. D. Appl. Phys. 32 (5), L13. Giorgi, G., et al., 2013. Small photocarrier effective masses featuring ambipolar transport in methylammonium lead iodide perovskite: a density functional analysis. J. Phys. Chem. Lett. 4 (24), 4213–4216.

248

Handbook of Organic Materials for Electronic and Photonic Devices

Glazer, A., 1972. The classification of tilted octahedra in perovskites. Acta Crystallogr. B 28 (11), 3384–3392. Grancini, G., et al., 2017. One-year stable perovskite solar cells by 2D/3D interface engineering. Nat. Commun. 8, 15684. Grote, C., Berger, R.F., 2015. Strain tuning of tin–halide and lead–halide perovskites: a firstprinciples atomic and electronic structure study. J. Phys. Chem. C 119 (40), 22832–22837. Haruyama, J., et al., 2015. First-principles study of ion diffusion in perovskite solar cell sensitizers. J. Am. Chem. Soc. 137 (32), 10048–10051. Hashmi, S.G., et al., 2017. Air processed inkjet infiltrated carbon based printed perovskite solar cells with high stability and reproducibility. Adv. Mater. Technol. 2 (1), 1600183. He, Y., Galli, G., 2014. Perovskites for solar thermoelectric applications: a first principle study of CH3NH3AI3 (A ¼ Pb and Sn). Chem. Mater. 26 (18), 5394–5400. He, T., et al., 2001. Superconductivity in the non-oxide perovskite MgCNi3. Nature 411, 54. Heiderhoff, R., et al., 2017. Thermal conductivity of methylammonium lead halide perovskite single crystals and thin films: a comparative study. J. Phys. Chem. C 121 (51), 28306–28311. Herz, L.M., 2017. Charge-carrier mobilities in metal halide perovskites: fundamental mechanisms and limits. ACS Energy Lett. 2 (7), 1539–1548. Hoke, E.T., et al., 2015. Reversible photo-induced trap formation in mixed-halide hybrid perovskites for photovoltaics. Chem. Sci. 6 (1), 613–617. Horantner, M.T., Snaith, H.J., 2017. Predicting and optimising the energy yield of perovskiteon-silicon tandem solar cells under real world conditions. Energy Environ. Sci. 10 (9), 1983–1993. H€ orantner, M.T., et al., 2017. The potential of multijunction perovskite solar cells. ACS Energy Lett. 2 (10), 2506–2513. Hu, Y., et al., 2018. Understanding the role of cesium and rubidium additives in perovskite solar cells: trap states, charge transport, and recombination. Adv. Energy Mater. 1703057. Hubbell, J.H., Seltzer, S.M., 1995. Tables of x-ray mass attenuation coefficients and mass energy-absorption coefficients 1 keV to 20 meV for elements z ¼ 1 to 92 and 48 additional substances of dosimetric interest. NIST, Gaithersburg, MD, USA. Hutter, E.M., et al., 2015. Charge carriers in planar and meso-structured organic–inorganic perovskites: mobilities, lifetimes, and concentrations of trap states. J. Phys. Chem. Lett. 6 (15), 3082–3090. Hutter, E.M., et al., 2016. Direct–indirect character of the bandgap in methylammonium lead iodide perovskite. Nat. Mater. 16, 115. Hutter, E.M., et al., 2017. Vapour-deposited cesium lead iodide perovskites: microsecond charge carrier lifetimes and enhanced photovoltaic performance. ACS Energy Lett. 2 (8), 1901–1908. Hwang, K., et al., 2015. Toward large scale roll-to-roll production of fully printed perovskite solar cells. Adv. Mater. 27 (7), 1241–1247. ISE, F., 2017. Photovoltaics report. . Ishidate, T., et al., 1997. Phase diagram of BaTiO3. Phys. Rev. Lett. 78 (12), 2397–2400. Ishihara, T., 1994. Optical properties of PbI-based perovskite structures. J. Lumin. 60–61, 269–274. ITRPV, 2017. International Technology Roadmap for Photovoltaic Report. . Jeng, J.-Y., et al., 2014. Nickel oxide electrode interlayer in CH3NH3PbI3 perovskite/PCBM planar-heterojunction hybrid solar cells. Adv. Mater. 26 (24), 4107–4113. Jeon, N.J., et al., 2014. Solvent engineering for high-performance inorganic–organic hybrid perovskite solar cells. Nat. Mater. 13, 897.

Hybrid perovskites for device applications

249

Jia, G., et al., 2018. Super air stable quasi-2D organic-inorganic hybrid perovskites for visible light-emitting diodes. Opt. Express 26 (2), A66–A74. Jiang, Q., et al., 2017. Planar-structure perovskite solar cells with efficiency beyond 21%. Adv. Mater. 29 (46), 1703852. Jodlowski, A.D., et al., 2017. Large guanidinium cation mixed with methylammonium in lead iodide perovskites for 19% efficient solar cells. Nat. Energy 2 (12), 972–979. Johnson, S.R., Tiedje, T., 1995. Temperature dependence of the Urbach edge in GaAs. J. Appl. Phys. 78 (9), 5609–5613. Jung, J.W., Chueh, C.-C., Jen, A.K.Y., 2015. A low-temperature, solution-processable, Cu-doped nickel oxide hole-transporting layer via the combustion method for highperformance thin-film perovskite solar cells. Adv. Mater. 27 (47), 7874–7880. Kawamura, Y., Mashiyama, H., Hasebe, K., 2002. Structural study on cubic–tetragonal transition of CH3NH3PbI3. J. Phys. Soc. Jpn. 71 (7), 1694–1697. Kepenekian, M., Even, J., 2017. Rashba and Dresselhaus couplings in halide perovskites: accomplishments and opportunities for spintronics and spin–orbitronics. J. Phys. Chem. Lett. 8 (14), 3362–3370. Kerr, M.J., Cuevas, A., Campbell, P., 2003. Limiting efficiency of crystalline silicon solar cells due to coulomb-enhanced Auger recombination. Prog. Photovolt. Res. Appl. 11 (2), 97–104. Kieslich, G., Sun, S., Cheetham, A.K., 2014. Solid-state principles applied to organic-inorganic perovskites: new tricks for an old dog. Chem. Sci. 5 (12), 4712–4715. Kim, J.H., et al., 2015. Enhanced environmental stability of planar heterojunction perovskite solar cells based on blade-coating. Adv. Energy Mater. 5 (4), 1401229. Kim, G.Y., et al., 2018. Large tunable photoeffect on ion conduction in halide perovskites and implications for photodecomposition. Nat. Mater. 17, 445–449. Kojima, A., et al., 2009. Organometal halide perovskites as visible-light sensitizers for photovoltaic cells. J. Am. Chem. Soc. 131 (17), 6050–6051. Kovalenko, M.V., Protesescu, L., Bodnarchuk, M.I., 2017. Properties and potential optoelectronic applications of lead halide perovskite nanocrystals. Science 358 (6364), 745. Kulkarni, S.A., et al., 2014. Band-gap tuning of lead halide perovskites using a sequential deposition process. J. Mater. Chem. A 2 (24), 9221–9225. Lee, M.M., et al., 2012. Efficient hybrid solar cells based on meso-superstructured organometal halide perovskites. Science 338 (6107), 643. Lee, N.E., et al., 2017a. Charge transport in organic molecular semiconductors from first principles: the band-like hole mobility in naphthalene crystal. Phys. Rev. B 97, 115203 arXiv:1712.00490. Lee, H., et al., 2017b. Direct experimental evidence of halide ionic migration under bias in CH3NH3PbI3-xClx-based perovskite solar cells using GD-OES analysis. ACS Energy Lett. 2 (4), 943–949. Leijtens, T., et al., 2015. Stability of metal halide perovskite solar cells. Adv. Energy Mater. 5 (20), 1500963. Leijtens, T., et al., 2016. Carrier trapping and recombination: the role of defect physics in enhancing the open circuit voltage of metal halide perovskite solar cells. Energy Environ. Sci. 9 (11), 3472–3481. Leijtens, T., et al., 2017. Mechanism of tin oxidation and stabilization by lead substitution in tin halide perovskites. ACS Energy Lett. 2 (9), 2159–2165. Leyden, M.R., Jiang, Y., Qi, Y., 2016. Chemical vapor deposition grown formamidinium perovskite solar modules with high steady state power and thermal stability. J. Mater. Chem. A 4 (34), 13125–13132.

250

Handbook of Organic Materials for Electronic and Photonic Devices

Li, S.-G., et al., 2015. Inkjet printing of CH3NH3PbI3 on a mesoscopic TiO2 film for highly efficient perovskite solar cells. J. Mater. Chem. A 3 (17), 9092–9097. Li, G., et al., 2016a. Highly efficient perovskite nanocrystal light-emitting diodes enabled by a universal crosslinking method. Adv. Mater. 28 (18), 3528–3534. Li, Z., et al., 2016b. Stabilizing perovskite structures by tuning tolerance factor: formation of formamidinium and cesium lead iodide solid-state alloys. Chem. Mater. 28 (1), 284–292. Li, J., et al., 2017. 2D behaviors of excitons in cesium lead halide perovskite nanoplatelets. J. Phys. Chem. Lett. 8 (6), 1161–1168. Liang, K., Mitzi, D.B., Prikas, M.T., 1998. Synthesis and characterization of organic  inorganic perovskite thin films prepared using a versatile two-step dipping technique. Chem. Mater. 10 (1), 403–411. Lin, Q., et al., 2014. Electro-optics of perovskite solar cells. Nat. Photonics 9, 106. Lin, Q., et al., 2015. Low noise, IR-blind organohalide perovskite photodiodes for visible light detection and imaging. Adv. Mater. 27 (12), 2060–2064. Liu, D., Kelly, T.L., 2013. Perovskite solar cells with a planar heterojunction structure prepared using room-temperature solution processing techniques. Nat. Photonics 8, 133. Liu, M., Johnston, M.B., Snaith, H.J., 2013. Efficient planar heterojunction perovskite solar cells by vapour deposition. Nature 501, 395. Liu, Y., et al., 2015. Two-inch-sized perovskite CH3NH3PbX3 (X ¼ cl, Br, I) crystals: growth and characterization. Adv. Mater. 27 (35), 5176–5183. Liu, Y., et al., 2016a. High efficiency tandem thin-perovskite/polymer solar cells with a graded recombination layer. ACS Appl. Mater. Interfaces 8 (11), 7070–7076. Liu, X., et al., 2016b. Periodic organic–inorganic halide perovskite microplatelet arrays on silicon substrates for room-temperature lasing. Adv. Sci. 3 (11), 1600137. Liu, P., et al., 2017. Organic–inorganic hybrid perovskite nanowire laser arrays. ACS Nano 11 (6), 5766–5773. Longo, G., et al., 2017. Fully vacuum-processed wide band gap mixed-halide perovskite solar cells. ACS Energy Lett. 3 (1), 214–219. Luo, P., et al., 2015. Uniform, stable, and efficient planar-heterojunction perovskite solar cells by facile low-pressure chemical vapor deposition under fully open-air conditions. ACS Appl. Mater. Interfaces 7 (4), 2708–2714. Ma, C., et al., 2016. Heterostructured WS2/CH3NH3PbI3 photoconductors with suppressed dark current and enhanced photodetectivity. Adv. Mater. 28 (19), 3683–3689. Mailoa, J.P., et al., 2015. A 2-terminal perovskite/silicon multijunction solar cell enabled by a silicon tunnel junction. Appl. Phys. Lett. 106 (12), 121105. Manders, J.R., et al., 2013. Solution-processed nickel oxide hole transport layers in high efficiency polymer photovoltaic cells. Adv. Funct. Mater. 23 (23), 2993–3001. Mathies, F., et al., 2016. Multipass inkjet printed planar methylammonium lead iodide perovskite solar cells. J. Mater. Chem. A 4 (48), 19207–19213. McMeekin, D.P., et al., 2016. A mixed-cation lead mixed-halide perovskite absorber for tandem solar cells. Science 351 (6269), 151. Men
endez-Proupin, E., et al., 2014. Self-consistent relativistic band structure of the CH3NH3PbI3 perovskite. Phys. Rev. B 90 (4), 045207. Mettan, X., et al., 2015. Tuning of the thermoelectric figure of merit of CH3NH3MI3 (M]Pb, Sn) photovoltaic perovskites. J. Phys. Chem. C 119 (21), 11506–11510. Milot, R.L., et al., 2015. Temperature-dependent charge-carrier dynamics in CH3NH3PbI3 perovskite thin films. Adv. Funct. Mater. 25 (39), 6218–6227. Mitzi, D.B., 1999. Synthesis, structure, and properties of organic-inorganic perovskites and related materials. In: Progress in Inorganic Chemistry. John Wiley & Sons, Inc, New York, pp. 1–121.

Hybrid perovskites for device applications

251

Mitzi, D.B., 2001. Templating and structural engineering in organic-inorganic perovskites. J. Chem. Soc. Dalton Trans. 1, 1–12. Mitzi, D.B., et al., 1994. Conducting tin halides with a layered organic-based perovskite structure. Nature 369, 467. Mitzi, D.B., Dimitrakopoulos, C.D., Kosbar, L.L., 2001. Structurally tailored organic inorganic perovskites: optical properties and solution-processed channel materials for thin-film transistors. Chem. Mater. 13 (10), 3728–3740. Mitzi, D.B., et al., 2002. Hybrid field-effect transistor based on a low-temperature meltprocessed channel layer. Adv. Mater. 14 (23), 1772–1776. Miyata, A., et al., 2015. Direct measurement of the exciton binding energy and effective masses for charge carriers in organic–inorganic tri-halide perovskites. Nat. Phys. 11, 582. Miyata, K., et al., 2017. Large polarons in lead halide perovskites. Sci. Adv. 3 (8), e1701217. https://doi.org/10.1126/sciadv.1701217. Mosconi, E., Etienne, T., De Angelis, F., 2017. Rashba band splitting in organohalide lead perovskites: bulk and surface effects. J. Phys. Chem. Lett. 8 (10), 2247–2252. Motta, C., et al., 2015. Revealing the role of organic cations in hybrid halide perovskite CH3NH3PbI3. Nat. Commun. 6, 7026. Nagabhushana, G.P., Shivaramaiah, R., Navrotsky, A., 2016. Direct calorimetric verification of thermodynamic instability of lead halide hybrid perovskites. Proc. Natl. Acad. Sci. 113 (28), 7717. Nakwaski, W., 1995. Effective masses of electrons and heavy holes in GaAs, InAs, A1As and their ternary compounds. Phys. B Condens. Matter 210 (1), 1–25. Niesner, D., et al., 2016. Giant Rashba splitting in CH3NH3PbBr3 organic-inorganic perovskite. Phys. Rev. Lett. 117 (12), 126401. Noel, N.K., et al., 2014a. Enhanced photoluminescence and solar cell performance via Lewis base passivation of organic–inorganic lead halide perovskites. ACS Nano 8 (10), 9815–9821. Noel, N.K., et al., 2014b. Lead-free organic-inorganic tin halide perovskites for photovoltaic applications. Energy Environ. Sci. 7 (9), 3061–3068. Noel, N.K., et al., 2017a. A low viscosity, low boiling point, clean solvent system for the rapid crystallisation of highly specular perovskite films. Energy Environ. Sci. 10 (1), 145–152. Noel, N.K., et al., 2017b. Unveiling the influence of pH on the crystallization of hybrid perovskites, delivering low voltage loss photovoltaics. Joule 1 (2), 328–343. Noh, J.H., et al., 2013. Chemical management for colorful, efficient, and stable inorganic– organic hybrid nanostructured solar cells. Nano Lett. 13 (4), 1764–1769. NREL, 2018. PV Efficiency Chart. . Ogomi, Y., et al., 2014. CH3NH3SnxPb(1-x)I3 perovskite solar cells covering up to 1060 nm. J. Phys. Chem. Lett. 5 (6), 1004–1011. Ono, L.K., et al., 2016. Organometal halide perovskite thin films and solar cells by vapor deposition. J. Mater. Chem. A 4 (18), 6693–6713. Onoda-Yamamuro, N., Matsuo, T., Suga, H., 1992. Dielectric study of CH3NH3PbX3 (X ¼ Cl, Br, I). J. Phys. Chem. Solids 53 (7), 935–939. Pan, J., et al., 2016. Highly efficient perovskite-quantum-dot light-emitting diodes by surface engineering. Adv. Mater. 28 (39), 8718–8725. Park, C., Snyder, R.L., 1995. Structures of high-temperature cuprate superconductors. J. Am. Ceram. Soc. 78 (12), 3171–3194. Peng, W., et al., 2017. Ultralow self-doping in two-dimensional hybrid perovskite single crystals. Nano Lett. 17 (8), 4759–4767. P
erez-del-Rey, D., et al., 2018. Interfacial modification for high-efficiency vapor-phasedeposited perovskite solar cells based on a metal oxide buffer layer. J. Phys. Chem. Lett. 9 (5), 1041–1046.

252

Handbook of Organic Materials for Electronic and Photonic Devices

Pisoni, A., et al., 2014. Ultra-low thermal conductivity in organic–inorganic hybrid perovskite CH3NH3PbI3. J. Phys. Chem. Lett. 5 (14), 2488–2492. Prasanna, R., et al., 2017. Band gap tuning via lattice contraction and octahedral tilting in perovskite materials for photovoltaics. J. Am. Chem. Soc. 139 (32), 11117–11124. Protesescu, L., et al., 2015. Nanocrystals of cesium lead halide perovskites (CsPbX3, X ¼ Cl, Br, and I): novel optoelectronic materials showing bright emission with wide color gamut. Nano Lett. 15 (6), 3692–3696. Qian, X., Gu, X., Yang, R., 2016. Lattice thermal conductivity of organic-inorganic hybrid perovskite CH3NH3PbI3. Appl. Phys. Lett. 108 (6), 063902. Qin, T., et al., 2017. Amorphous hole-transporting layer in slot-die coated perovskite solar cells. Nano Energy 31, 210–217. Rajagopal, A., et al., 2017. Highly efficient perovskite–perovskite tandem solar cells reaching 80% of the theoretical limit in photovoltage. Adv. Mater. 29 (34), 1702140. Ran, C., et al., 2018. Bilateral interface engineering toward efficient 2D–3D bulk heterojunction tin halide lead-free perovskite solar cells. ACS Energy Lett. 3 (3), 713–721. Rau, U., 2007. Reciprocity relation between photovoltaic quantum efficiency and electroluminescent emission of solar cells. Phys. Rev. B 76 (8), 085303. Rehman, W., et al., 2017. Photovoltaic mixed-cation lead mixed-halide perovskites: links between crystallinity, photo-stability and electronic properties. Energy Environ. Sci. 10 (1), 361–369. Richter, J.M., et al., 2016. Enhancing photoluminescence yields in lead halide perovskites by photon recycling and light out-coupling. Nat. Commun. 7, 13941. Sadhanala, A., et al., 2015. Blue-green color tunable solution processable organolead chloride– bromide mixed halide perovskites for optoelectronic applications. Nano Lett. 15 (9), 6095–6101. Saidaminov, M.I., et al., 2015. High-quality bulk hybrid perovskite single crystals within minutes by inverse temperature crystallization. Nat. Commun. 6, 7586. Saliba, M., et al., 2016a. Incorporation of rubidium cations into perovskite solar cells improves photovoltaic performance. Science. https://doi.org/10.1126/science.aah5557. Saliba, M., et al., 2016b. Cesium-containing triple cation perovskite solar cells: improved stability, reproducibility and high efficiency. Energy Environ. Sci. 9 (6), 1989–1997. Saliba, M., et al., 2016c. A molecularly engineered hole-transporting material for efficient perovskite solar cells. Nat. Energy 1, 15017. Saliba, M., et al., 2016d. Structured organic–inorganic perovskite toward a distributed feedback laser. Adv. Mater. 28 (5), 923–929. Samiee, M., et al., 2014. Defect density and dielectric constant in perovskite solar cells. Appl. Phys. Lett. 105 (15), 153502. Savenije, T.J., et al., 2014. Thermally activated exciton dissociation and recombination control the carrier dynamics in organometal halide perovskite. J. Phys. Chem. Lett. 5 (13), 2189–2194. Schulz, P., et al., 2014. Interface energetics in organo-metal halide perovskite-based photovoltaic cells. Energy Environ. Sci. 7 (4), 1377–1381. Sch€ unemann, S., et al., 2017. Halide perovskite 3D photonic crystals for distributed feedback lasers. ACS Photonics 4 (10), 2522–2528. Sendner, M., et al., 2016. Optical phonons in methylammonium lead halide perovskites and implications for charge transport. Mater. Horiz. 3 (6), 613–620. Sessolo, M., et al., 2015. Photovoltaic devices employing vacuum-deposited perovskite layers. MRS Bull. 40 (8), 660–666. Shen, P.S., et al., 2016. Low-pressure hybrid chemical vapor growth for efficient perovskite solar cells and large-area module. Adv. Mater. Interfaces 3 (8), 1500849.

Hybrid perovskites for device applications

253

Shen, H., et al., 2018. Mechanically-stacked perovskite/CIGS tandem solar cells with efficiency of 23.9% and reduced oxygen sensitivity. Energy Environ. Sci. 11 (2), 394–406. Shi, D., et al., 2015. Low trap-state density and long carrier diffusion in organolead trihalide perovskite single crystals. Science 347 (6221), 519. Shinobu, H., Yuji, M., 1987. High Tc superconductors of the perovskite structure oxides. Jpn. J. Appl. Phys. 26 (S3–2), 1027. Shrestha, S., et al., 2017. High-performance direct conversion X-ray detectors based on sintered hybrid lead triiodide perovskite wafers. Nat. Photonics 11, 436. Shuttle, C.G., et al., 2010. Measurement of charge-density dependence of carrier mobility in an organic semiconductor blend. Adv. Funct. Mater. 20 (5), 698–702. Smith, I.C., et al., 2014. A layered hybrid perovskite solar-cell absorber with enhanced moisture stability. Angew. Chem. Int. Ed. 53 (42), 11232–11235. Snaith, H.J., et al., 2014. Anomalous hysteresis in perovskite solar cells. J. Phys. Chem. Lett. 5 (9), 1511–1515. Stoumpos, C.C., Malliakas, C.D., Kanatzidis, M.G., 2013. Semiconducting tin and lead iodide perovskites with organic cations: phase transitions, high mobilities, and near-infrared photoluminescent properties. Inorg. Chem. 52 (15), 9019–9038. Stoumpos, C.C., et al., 2015. Hybrid germanium iodide perovskite semiconductors: active lone pairs, structural distortions, direct and indirect energy gaps, and strong nonlinear optical properties. J. Am. Chem. Soc. 137 (21), 6804–6819. Stoumpos, C.C., et al., 2016. Ruddlesden–popper hybrid lead iodide perovskite 2D homologous semiconductors. Chem. Mater. 28 (8), 2852–2867. Stranks, S.D., 2017. Nonradiative losses in metal halide perovskites. ACS Energy Lett. 2 (7), 1515–1525. Stranks, S.D., Petrozza, A., 2016. Revisiting photocarrier lifetimes in photovoltaics. Nat. Photonics 10, 562. Stranks, S.D., et al., 2013. Electron-hole diffusion lengths exceeding 1 micrometer in an organometal trihalide perovskite absorber. Science 342 (6156), 341. Stranks, S.D., et al., 2014. Recombination kinetics in organic-inorganic perovskites: excitons, free charge, and subgap states. Phys. Rev. Appl. 2 (3), 034007. Sutherland, B.R., Sargent, E.H., 2016. Perovskite photonic sources. Nat. Photonics 10, 295. Sutherland, B.R., et al., 2015. Sensitive, fast, and stable perovskite photodetectors exploiting interface engineering. ACS Photonics 2 (8), 1117–1123. Sze, S.M., Ng, K.K., 2006. Appendix G properties of Si and GaAs. In: Physics of Semiconductor Devices. John Wiley & Sons, Inc., Hoboken, NJ, p. 790. Tahara, H., et al., 2016. Experimental evidence of localized shallow states in orthorhombic phase of CH3NH3PbI3 perovskite thin films revealed by photocurrent beat spectroscopy. J. Phys. Chem. C 120 (10), 5347–5352. Takahashi, Y., et al., 2011. Charge-transport in tin-iodide perovskite CH3NH3SnI3: origin of high conductivity. Dalton Trans. 40 (20), 5563–5568. Tan, Z.-K., et al., 2014. Bright light-emitting diodes based on organometal halide perovskite. Nat. Nanotechnol. 9, 687. Tan, H., et al., 2017. Efficient and stable solution-processed planar perovskite solar cells via contact passivation. Science 355 (6326), 722. Tang, B., et al., 2017. Single-mode lasers based on cesium lead halide perovskite submicron spheres. ACS Nano 11 (11), 10681–10688. Torabi, S., et al., 2015. Strategy for enhancing the dielectric constant of organic semiconductors without sacrificing charge carrier mobility and solubility. Adv. Funct. Mater. 25 (1), 150–157.

254

Handbook of Organic Materials for Electronic and Photonic Devices

Tsai, H., et al., 2016. High-efficiency two-dimensional Ruddlesden–Popper perovskite solar cells. Nature 536, 312. Tsai, H., et al., 2018. Stable light-emitting diodes using phase-pure Ruddlesden–Popper layered perovskites. Adv. Mater. 30 (6), 1704217. Umari, P., Mosconi, E., De Angelis, F., 2014. Relativistic GW calculations on CH3NH3PbI3 and CH3NH3SnI3 perovskites for solar cell applications. Sci. Rep. 4, 4467. Unger, E.L., et al., 2014. Hysteresis and transient behavior in current-voltage measurements of hybrid-perovskite absorber solar cells. Energy Environ. Sci. 7 (11), 3690–3698. Vak, D., et al., 2015. 3D printer based slot-die coater as a lab-to-fab translation tool for solutionprocessed solar cells. Adv. Energy Mater. 5 (4), 1401539. Volonakis, G., et al., 2017. Cs2InAgCl6: a new lead-free halide double perovskite with direct band gap. J. Phys. Chem. Lett. 8 (4), 772–778. Wang, K.-C., et al., 2014. Low-temperature sputtered nickel oxide compact thin film as effective electron blocking layer for mesoscopic NiO/CH3NH3PbI3 perovskite heterojunction solar cells. ACS Appl. Mater. Interfaces 6 (15), 11851–11858. Wang, Y., et al., 2015. Hybrid graphene–perovskite phototransistors with ultrahigh responsivity and gain. Adv. Opt. Mater. 3 (10), 1389–1396. Wang, L., et al., 2016. Pressure-induced structural evolution and band gap shifts of organometal halide perovskite-based methylammonium lead chloride. J. Phys. Chem. Lett. 7 (24), 5273–5279. Wang, K.-H., et al., 2017a. Structural and photophysical properties of methylammonium lead tribromide (MAPbBr3) single crystals. Sci. Rep. 7 (1), 13643. Wang, Z., et al., 2017b. Efficient ambient-air-stable solar cells with 2D–3D heterostructured butylammonium-caesium-formamidinium lead halide perovskites. Nat. Energy 2, 17135. Wang, T., et al., 2017c. Indirect to direct bandgap transition in methylammonium lead halide perovskite. Energy Environ. Sci. 10 (2), 509–515. Wasylishen, R.E., Knop, O., Macdonald, J.B., 1985. Cation rotation in methylammonium lead halides. Solid State Commun. 56 (7), 581–582. Wehrenfennig, C., et al., 2014. High charge carrier mobilities and lifetimes in organolead trihalide perovskites. Adv. Mater. 26 (10), 1584–1589. Wei, H., et al., 2016. Sensitive X-ray detectors made of methylammonium lead tribromide perovskite single crystals. Nat. Photonics 10, 333. Wei, W., et al., 2017a. Monolithic integration of hybrid perovskite single crystals with heterogenous substrate for highly sensitive X-ray imaging. Nat. Photonics 11, 315. Wei, H., et al., 2017b. Dopant compensation in alloyed CH3NH3PbBr3-xClx perovskite single crystals for gamma-ray spectroscopy. Nat. Mater. 16, 826. Weidman, M.C., et al., 2016. Highly tunable colloidal perovskite nanoplatelets through variable cation, metal, and halide composition. ACS Nano 10 (8), 7830–7839. Weller, M.T., et al., 2015. Complete structure and cation orientation in the perovskite photovoltaic methylammonium lead iodide between 100 and 352 K. Chem. Commun. 51 (20), 4180–4183. Whalley, L.D., et al., 2016. Phonon anharmonicity, lifetimes, and thermal transport in CH3NH3PbI3 from many-body perturbation theory. Phys. Rev. B 94 (22), 220301. Wright, A.D., et al., 2016. Electron–phonon coupling in hybrid lead halide perovskites. Nat. Commun. 7, 11755. Wright, A.D., et al., 2017. Band-tail recombination in hybrid lead iodide perovskite. Adv. Funct. Mater. 27 (29), 1700860. Wu, X., et al., 2015. Trap states in lead iodide perovskites. J. Am. Chem. Soc. 137 (5), 2089–2096.

Hybrid perovskites for device applications

255

Wu, B., et al., 2017. Long minority-carrier diffusion length and low surface-recombination velocity in inorganic lead-free CsSnI3 perovskite crystal for solar cells. Adv. Funct. Mater. 27 (7), 1604818. Wu, Z., et al., 2018. Broadband white-light emission with a high color rendering index in a twodimensional organic-inorganic hybrid perovskite. J. Mater. Chem. C 6 (5), 1171–1175. Xiao, Z., et al., 2017. Efficient perovskite light-emitting diodes featuring nanometre-sized crystallites. Nat. Photonics 11, 108. Xing, G., et al., 2013. Long-range balanced electron- and hole-transport lengths in organicinorganic CH3NH3PbI3. Science 342 (6156), 344. Yakunin, S., et al., 2015. Low-threshold amplified spontaneous emission and lasing from colloidal nanocrystals of caesium lead halide perovskites. Nat. Commun. 6, 8056. Yamada, Y., et al., 2014. Photocarrier recombination dynamics in perovskite CH3NH3PbI3 for solar cell applications. J. Am. Chem. Soc. 136 (33), 11610–11613. Yamada, Y., et al., 2015. Photoelectronic responses in solution-processed perovskite CH3NH3PbI3 solar cells studied by photoluminescence and photoabsorption spectroscopy. IEEE J. Photovoltaics 5 (1), 401–405. Yang, W.S., et al., 2015a. High-performance photovoltaic perovskite layers fabricated through intramolecular exchange. Science 348 (6240), 1234. Yang, Z., et al., 2015b. High-performance fully printable perovskite solar cells via bladecoating technique under the ambient condition. Adv. Energy Mater. 5 (13), 1500328. Yang, D., et al., 2016. Surface optimization to eliminate hysteresis for record efficiency planar perovskite solar cells. Energy Environ. Sci. 9 (10), 3071–3078. Yang, W.S., et al., 2017a. Iodide management in formamidinium-lead-halide–based perovskite layers for efficient solar cells. Science 356 (6345), 1376. Yang, M., et al., 2017b. Perovskite ink with wide processing window for scalable highefficiency solar cells. Nat. Energy 2, 17038. Yang, Z., et al., 2017c. Unraveling the exciton binding energy and the dielectric constant in single-crystal methylammonium lead triiodide perovskite. J. Phys. Chem. Lett. 8 (8), 1851–1855. Yang, X., et al., 2018. Efficient green light-emitting diodes based on quasi-two-dimensional composition and phase engineered perovskite with surface passivation. Nat. Commun. 9 (1), 570. Yao, E.-P., et al., 2017. High-brightness blue and white LEDs based on inorganic perovskite nanocrystals and their composites. Adv. Mater. 29 (23), 1606859. Yi, C., et al., 2016. Entropic stabilization of mixed A-cation ABX3 metal halide perovskites for high performance perovskite solar cells. Energy Environ. Sci. 9 (2), 656–662. Yin, W.-J., Shi, T., Yan, Y., 2014. Unusual defect physics in CH3NH3PbI3 perovskite solar cell absorber. Appl. Phys. Lett. 104 (6), 063903. Yoon, H., et al., 2016. Hysteresis-free low-temperature-processed planar perovskite solar cells with 19.1% efficiency. Energy Environ. Sci. 9 (7), 2262–2266. Yuan, Y., et al., 2015. Photovoltaic switching mechanism in lateral structure hybrid perovskite solar cells. Adv. Energy Mater. 5 (15), 1500615. Yuan, M., et al., 2016. Perovskite energy funnels for efficient light-emitting diodes. Nat. Nanotechnol. 11, 872. Yuan, Z., et al., 2017. One-dimensional organic lead halide perovskites with efficient bluish white-light emission. Nat. Commun. 8, 14051. Zhang, X., et al., 2016a. Enhancing the brightness of cesium lead halide perovskite nanocrystal based green light-emitting devices through the interface engineering with perfluorinated ionomer. Nano Lett. 16 (2), 1415–1420.

256

Handbook of Organic Materials for Electronic and Photonic Devices

Zhang, X., et al., 2016b. Bright perovskite nanocrystal films for efficient light-emitting devices. J. Phys. Chem. Lett. 7 (22), 4602–4610. Zhang, L., et al., 2017. Ultra-bright and highly efficient inorganic based perovskite lightemitting diodes. Nat. Commun. 8, 15640. Zhao, T., Wang, D., Shuai, Z., 2017. Doping optimization of organic-inorganic hybrid perovskite CH3NH3PbI3 for high thermoelectric efficiency. Synth. Met. 225, 108–114. Zheng, K., et al., 2015a. Exciton binding energy and the nature of emissive states in organometal halide perovskites. J. Phys. Chem. Lett. 6 (15), 2969–2975. Zheng, F., et al., 2015b. Rashba spin–orbit coupling enhanced carrier lifetime in CH3NH3PbI3. Nano Lett. 15 (12), 7794–7800. Zheng, X., et al., 2017. Defect passivation in hybrid perovskite solar cells using quaternary ammonium halide anions and cations. Nat. Energy 2, 17102. Zhou, J.-J., Bernardi, M., 2016. Ab initio. Phys. Rev. B 94 (20), 201201. Zhou, H., Chen, Q., Yang, Y., 2015. Vapor-assisted solution process for perovskite materials and solar cells. MRS Bull. 40 (8), 667–673. Zhu, H., et al., 2015. Lead halide perovskite nanowire lasers with low lasing thresholds and high quality factors. Nat. Mater. 14, 636. Zhumekenov, A.A., et al., 2016. Formamidinium lead halide perovskite crystals with unprecedented long carrier dynamics and diffusion length. ACS Energy Lett. 1 (1), 32–37.

Further reading Anaya, M., et al., 2017a. ABX3 perovskites for tandem solar cells. Joule 1 (4), 769–793. Chueh, C.-C., Li, C.-Z., Jen, A.K.Y., 2015. Recent progress and perspective in solutionprocessed interfacial materials for efficient and stable polymer and organometal perovskite solar cells. Energy Environ. Sci. 8 (4), 1160–1189. Gr€atzel, M., 2014. The light and shade of perovskite solar cells. Nat. Mater. 13, 838. Green, M.A., Ho-Baillie, A., Snaith, H.J., 2014. The emergence of perovskite solar cells. Nat. Photonics 8, 506. Leijtens, T., Bush, K.A., Prasanna, R., McGehee, M.D., 2018. Opportunities and challenges for tandem solar cells using metal halide perovskite semiconductors. Nat. Energy 3, 828–838. Stranks, S.D., Snaith, H.J., 2015. Metal-halide perovskites for photovoltaic and light-emitting devices. Nat. Nanotechnol. 10, 391–402. Whitfield, P.S., et al., 2016. Structures, phase transitions and tricritical behavior of the hybrid perovskite methyl ammonium lead iodide. Sci. Rep. 6, 35685.

Part Two Mechanisms

This page intentionally left blank

Frenkel exciton dynamics: A theoretical perspective

8

Oliver Ku€hn University of Rostock, Institute of Physics, Rostock, Germany

8.1

Frenkel excitons in organic materials

The term “exciton” denotes a quasiparticle (i.e., a collective electronic excitation of some medium). It was been introduced by Frenkel in 1931 (Frenkel, 1931) as a form of excitation wave in crystals. In fact, Frenkel excitons (FEs) are bound states of electron-hole pairs, both residing on the same lattice cell or molecule. The situation where the electron-hole separation is much larger (
10 nm) corresponds to what is called a Wannier-Mott exciton; it is typical for inorganic bulk semiconductors. The difference in separation comes with different binding energies for the electron-hole pair. FEs in organic crystals have typical binding energies of about 1 eV, whereas the binding energy of Wannier-Mott excitons in inorganic crystals is only about 10 meV. Within the simple H-atom-like picture, this behavior is often rationalized in terms of the dielectric shielding of the electron-hole interaction, which is much larger in inorganic than in organic materials (for an overview on excitons in organic solids, see Agranovich (2009) and Schwoerer and Wolf (2007)). The local electronic excitations of the molecules interact with each other via the Coulomb potential, which leads to a splitting of the electronic transitions. For molecular crystals, this effect is known as Davydov splitting, and it can be found in cases where two interacting molecules form a unit cell (Davydov, 1964). More generally, the interaction between local excitations in molecular assemblies such as aggregates leads to a characteristic shift of the absorption bands. Here, one distinguishes J- and H-aggregates, which show a red and blue shift, respectively, with respect to the monomer absorption (May and K€ uhn, 2011).1 While excitons in molecular crystals (e.g., polyacenes) have been studied for decades, FEs in molecular nanoscale systems have attracted this level of attention only more recently. These include conjugated polymers (CPs), molecular aggregates, and photosynthetic pigment protein complexes (Scholes and Rumbles, 2006). What distinguishes nanoscale systems from crystals is their flexibility, which facilitates the tuning of the optical and transport properties (W€urthner et al., 2011). Inspiration comes from nature, with its highly efficient nanoscale machinery for the conversion of sunlight into energy stored in chemical bonds (Blankenship, 2014). Light 1

Note that the usage of the term “Frenkel exciton” is often not that strict when it comes to distinguishing the local (noninteracting) excitation and the delocalized excitation, which results from the interaction between local excitations.

Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00008-5 © 2019 Elsevier Ltd. All rights reserved.

260

Handbook of Organic Materials for Electronic and Photonic Devices

harvesting and charge separation are realized in pigment-protein complexes, where reaction centers responsible for charge separation are surrounded by antennae capturing photons and transferring the excitation energy by means of FEs. Bulk heterojunctions (BHJs) in organic photovoltaics (OPVs) are mimicking this setup by using polymer-fullerene blends (Scharber and Sariciftci, 2013). A substantially enhanced control of the assembly of chromophores for artificial light harvesting on the nanoscale can be reached with the emerging deoxyribonucleic acid (DNA) origami technology (Woller et al., 2013). During the last decade, evidence has been gathered that nature’s design not only fulfills the purpose of optimizing the rates for incoherent kinetics of energy and charge transfer (CT). In addition, nontrivial quantum effects, such as wave packet coherence, appear to play a prominent role. Although there were indications of exciton delocalization in the strongly coupled antenna system of purple bacteria 20 years ago (Chachisvilis et al., 1997; K€ uhn and Sundstr€ om, 1997), the breakthrough came with the implementation of two-dimensional (2D) electronic spectroscopy (Brixner et al., 2005). The Fenna-Matthews-Olson (FMO) complex, which acts like a wire for electronic excitation energy connecting the peripheral antennae and reaction centers in green sulfur bacteria, has played a prominent role (Engel et al., 2007). Indications of quantum coherence have been reported at 77 K, as well as at room temperature (Panitchayangkoon et al., 2010), although the latter results have been questioned recently (Duan et al., 2017). In any case, these findings have been extremely stimulating, far beyond the particular biological applications. It has been observed that the exploitation of quantum effects in artificial light harvesting should be part of the agenda (Romero et al., 2014; Bredas et al., 2017; Scholes et al., 2017). The FE theory for molecular systems is well established (for a recent review, see Schr€ oter et al., 2015a). Of central importance for the understanding of the dynamics and spectroscopy of condensed phase systems, like crystals, pigment-protein complexes, or polymers, is the incorporation of exciton-vibrational coupling (EVC). For decades, intermolecular and intramolecular vibrations have been considered as forming a heat bath that gives rise to the exciton phase and energy relaxation (May and K€ uhn, 2011). This comforted the status of experiment and theory, the latter being limited by computer power and numerical methods to treat rather simple models. The emergence of ultrafast, nonlinear spectroscopy has changed this picture, and today, vibrational degrees of freedoms (DOFs) are known to play various roles. The development in experimentation has been paralleled by enormous progress made in theoretical methodology and numerical implementation. Still, the complexity of a system like the FMO complex provides a challenge, which is best seen by comparing the simulation results obtained for the exciton transfer (Duan et al., 2017; Ishizaki et al., 2010; Kreisbeck and Kramer, 2012; Lee and Coker, 2016; Schulze and K€ uhn, 2015). This chapter is organized as follows. The next section introduces the FE Hamiltonian model, and Section 8.3 provides a brief account on quantum dynamics methods for open and closed systems. The connection to spectroscopy is outlined in Section 8.4. Three applications (i.e., the determination of spectral densities, the manifestation of EVC in absorption and emission line shapes, and the quantum dynamics

Frenkel exciton dynamics: A theoretical perspective

261

of a model for the FMO exciton transport) are discussed in Section 8.5. Some final remarks are given in Section 8.6.

8.2

Model Hamiltonian

8.2.1 Electronic part In the following discussion, it is assumed that the considered system can be decomposed into Nagg molecular units (monomers), which interact only weakly with each other, such that monomers keep their chemical identity. In particular their absorption and emission spectra are supposed to be identical to those of the isolated monomers. Two types of excitations can be distinguished: local FEs and nonlocal CT transitions. In order to describe FE, as well as CT transitions, an electron-hole representation as described by Merrifield can be introduced (Merrifield, 1961). The zero-order basis consists of states jme, nhi, where me and nh correspond to an electron and a hole at site m and n, respectively. All other monomers k6¼(m, n) are supposed to be in their ground state.2 The general representation of the electronic Hamiltonian thus reads H ðelÞ ðRÞ ¼

XX

Hme nh , me nh ðRÞjme nh ⟩me nh j:

(8.1)

mn mn

Here, we included the parametric dependence on all nuclear DOFs, R. Eq. (8.2) can be separated as follows: H ðelÞ ¼ H ðFEÞ + H ðCTÞ + V ðFECTÞ

(8.2)

The first term describes FEs (ζ mn ¼ 1  δmn): H ðFEÞ ¼

X

½δmn Eme mh + ζ mn Jme mh ,ne nh jme mh ihne nh j:

mn

(8.3)

The second term in Eq. (8.2) gives the contributions due to CT states: H ðCTÞ ¼

XX

h i ζ mn ζ mn δmn δnn Eme nh + Jme nh , me nh jme nh ihme nh j:

(8.4)

mn mn

The coupling between both types of states is written as: V ðFECTÞ ¼

XX

ζ mm ½δmn ζ mn + δmn ζ mn Jme nh , me nh jme nh ihme nh j:

(8.5)

mn mn

2

Note that the superscripts e/h are used for the bookkeeping of energies and states only; for the actual summation indices, m/n are taken.

262

Handbook of Organic Materials for Electronic and Photonic Devices

Here, Eme mh and Eme nh are the energies of the FE and CT states, respectively. The general Coulomb integral is given by: Z Jme nh , me nh ¼

dr1 dr2

φ∗me ðr1 Þφme ðr2 Þφnh ðr1 Þφ∗nh ðr2 Þ jr1  r2 j

:

(8.6)

There is a huge body of literature on this model of coupled exciton and CT states (cf. the recent review in Polkehn et al., 2018). The parameterization based on quantum chemistry calculations is straightforward if a local scheme is used (e.g., Pl€otz et al., 2017). Often, the supermolecule approach is employed. Here, one first obtains the adiabatic states (i.e., the eigenstates of H(el)). These data have to be transformed to the diabatic basis with respect to the monomers.3 Different strategies have been developed in this respect (Polkehn et al., 2018; Arago´ and Troisi, 2015; Liu et al., 2015; Xie et al., 2017).

8.2.2 Exciton-vibrational coupling In the following section, we will focus on the effect of EVC. To ease the notation, CT states will not be considered. However, the presented model can be applied to CT states in a straightforward way (e.g., Polkehn et al., 2018). Denoting the FE basis by jmemhijmi, the FE Hamiltonian becomes (May and K€uhn, 2011) H ðFEÞ ¼

X

hmm jmihnj

mn

(8.7)

with (Em ¼ Eme mh and Jmn ¼ Jme mh ,ne nh ) hmn ¼ δmn Em + Jmn :

(8.8)

The exciton Hamiltonian matrix, hmn, depends on the nuclear coordinates. Hence, intermolecular and intramolecular nuclear motions characterizing the thermal ensemble cause a variation of hmn (i.e., along a sampling trajectory, it becomes time dependent). Introducing averaged energies, hhmmi, and the respective fluctuations, δhmn(t) ¼ hmn(t) hhmmi, Eq. (8.7) becomes H ðFEÞ ¼

X mn

½hhmm i + δhmn ðtÞjmihnj:

(8.9)

The fluctuations can be characterized by means of their correlation functions and the related spectral distributions; that is

3

Note that in terms of the individual monomers, these states are still adiabatic ones in the sense of the BornOppenheimer approach (May and K€uhn, 2011).

Frenkel exciton dynamics: A theoretical perspective

263

Z Ckl,mn ðωÞ ¼

dteiωt hδhkl ðtÞδhmn ð0Þi:

(8.10)

Having at hand an FE parameterization, there is a variety of methods for calculating Ckl,mn(ω) (Lee and Coker, 2016; Pl€ otz et al., 2017; Olbrich et al., 2011; Renger et al., 2012; Fornari et al., 2016). Insight into Ckl,mn(ω) is provided upon the introduction of normal modes of vibration (May and K€ uhn, 2011; Schr€ oter et al., 2015a; K€uhn et al., 1997; Renger et al., 2001). To the lowest order in the (dimensionless) normal mode coordinates, {Qξ}, the fluctuations can be written as δhmn ðtÞ ¼

X

ħωξ gmn ðξÞQξ ðtÞ,

ξ

(8.11)

where gmn(ξ) is a dimensionless coupling matrix. Its diagonal elements are related to the Huang-Rhys factor as follows: Sm, ξ ¼ [gmm(ξ)]2/2. Usually, normal modes can be differentiated into local intramolecular modes, {Qm,ξ}, and intermolecular as well as ðphÞ environmental modes (both sometimes called phonon modes), fQξ g. The harmonic approximation enables one to evaluate Eq. (8.10) analytically yielding (May and K€ uhn, 2011) 2 CHO kl,mn ðωÞ ¼ 2πω ½1 + nðωÞ½J kl,mn ðωÞ  J kl,mn ðωÞ:

(8.12)

Here, n(ω) represents the Bose-Einstein distribution and the spectral density (SD) has been introduced according to J kl,mn ðωÞ ¼

X

gkl ðξÞgmn ðξÞδðω  ωξ Þ:

ξ

(8.13)

The harmonic oscillator bath itself is described by the Hamiltonian ! 1X ∂2 2 HB ¼ ħωξ  2 + Qξ : 2 ξ ∂Qξ

(8.14)

In summary, the effect of EVC due to thermal fluctuations can be described in harmonic approximation by the system-bath Hamiltonian4 H ¼ HS + HB + HSB

4

(8.15)

Notice that there is a formal correspondence of Eq. (8.15) to the Holstein Hamiltonian of electron transfer (Holstein, 1959).

264

Handbook of Organic Materials for Electronic and Photonic Devices

with HS ¼ H(FE) according to Eq. (8.3)5 and HSB ¼

XX mn

ħωξ gmn ðξÞQξ jmihnj:

ξ

(8.16)

With increasing computer power, models describing EVC to a finite number of modes explicitly became popular (Eisfeld et al., 2005; Polyutov et al., 2012; Butkus et al., 2014; Pl€ otz et al., 2016). The most simple example is a molecular dimer with intramolecular vibrational coordinates {Qm,ξ} (m ¼ 1, 2) (Philpott, 1967). In the electronic ground state vibrations are uncoupled and the electron-vibrational state can be described by j0, Mg1 ,Mg2 i, where Mgm denotes the multimode vibrational state at monomer m in its electronic ground state (M ¼ (M1, M2, …), with Mξ being the quantum number for the ξth mode). There are two configurations for electronically excited states (i.e., j1,Me1 , Mg2 i and j2, Mg1 , Me2 i, where Mem refers to the modes in the excited state at monomer m). For the present model j0,Mg1 ,Mg2 i is already an eigenstate. The exciton-vibrational eigenstates can be obtained by expansion into this given basis set (Schr€ oter et al., 2015a).

8.3

Dynamics of excitons

8.3.1 Time-dependent exciton-vibrational wave packets Given the FE Hamiltonian, including a finite number of vibrational modes according to the Huang-Rhys model, textbook knowledge can be used to formulate the timedependent Schr€ odinger equation in terms of coupled, first-order differential equations. But, due to exponential scaling with the number of DOFs, this is a meaningless task for any system of interesting size. In their seminal 1990 paper, Meyer and coworkers presented a more clever approach (i.e., the Multiconfiguration Time-Dependent Hartree [MCTDH]), which uses time-dependent basis functions to create a more compact representation of the moving wave packet (Meyer et al., 1990) (for reviews, see Beck et al., 2000; Meyer et al., 2009; Meyer, 2011). MCTDH has been a success, with applications in many different fields (Meyer et al., 2009). Moving toward really high-dimensional problems, approximating condensed phase systems with continuous spectra became possible with the so-called multilayer extension called ML-MCTDH (Wang and Thoss, 2003; Manthe, 2008; Vendrell and Meyer, 2011). The simple structure of the FE Hamiltonian makes it ideally suited for ML-MCTDH simulations. Surprisingly, there are not many applications yet. Engel and coworkers were the first to apply MCTDH to the calculation of excitonvibrational properties of small linear aggregates (up to eight sites, including one vibrational mode per monomer) (Seibt et al., 2009). Further applications have been reported for linear perylene bisimide (PBI)-type aggregates with up to six sites and five modes 5

Here, we assumed that the matrix elements of the FE Hamiltonian have been obtained for a representative average minimum configuration of the nuclei.

Frenkel exciton dynamics: A theoretical perspective

265

per monomer (Ambrosek et al., 2012), as well as for dimer and trimer aggregates with the focus on nonadiabatic interband transitions (Schr€oter and K€uhn, 2013). Multidimensional exciton-vibrational wave packet dynamics in the FMO complex and the light-harvesting complex 2 of purple bacteria (including hundreds of DOFs) have been discussed in a series of papers (Schulze and K€uhn, 2015; Schulze et al., 2016, 2017; Shibl et al., 2017). Finally, Burghardt and coworkers have used ML-MCTDH to describe interfacial exciton dissociation in BHJs (Tamura and Burghardt, 2013) and singlet fission in molecular crystals (Tamura et al., 2015). In the following discussion, the basic equations of ML-MCTDH will be introduced briefly for an EVC model comprising only intramolecular modes. Let us denote the various electronic configurations (i.e., local FE or CT excitations, by jKi) and assume that they form a complete basis such that the time-dependent state vector can be expanded as follows: jΨ ðQ;tÞi ¼

X

ψ K ðQ;tÞjKi:

(8.17)

K

The nuclear coordinates are in the D ¼ Nagg  Nvib dimensional set Q ¼ fQ1 , …,QNagg g  fQ1 , …, QD g, with Nvib being the number of vibrational modes per monomer. Eq. (8.17) defines state-specific vibrational wave packets, ψ K(Q;t), which will be discussed in the following (dropping the state index). MCTDH uses time-dependent basis functions, so-called single particle functions (SPFs), which evolve with the propagating wave packet (Meyer et al., 1990; Beck et al., 2000). Thus, in MCTDH, the wave packet is represented as follows: ψðQ;tÞ ¼

n1 X j1 ¼1

…

nD X

A1j1 …jD ðtÞ  ϕj11;1 ðQ1 ,tÞ…ϕj1D;D ðQD , tÞ:

(8.18)

jD ¼1

κ Note that in contrast to the standard textbook approach, the SPFs, ϕ1, λ (Qκ , t), are time dependent. Thus, the number of basis functions, nκ, needed per DOF κ is smaller than in the standard approach. This does not change the exponential scaling, but the base of the exponential, nκ, can be reduced. The equations of motion for the MCTDH expansion coefficients, A1j1 …jD ðtÞ and SPFs are obtained by applying the Dirac variational principle (Beck et al., 2000). To this end, the time-dependent SPFs are expanded into ðκÞ a time-independent (primitive) basis χ j ðQκ Þ themselves, that is

ϕλ1;κ ðQκ , tÞ ¼

Nκ X ðκÞ Aλ2;;jκ ðtÞχ j ðQκ Þ:

(8.19)

j¼1

MCTDH can be seen as a two-layer approach. This becomes clear upon the interpretation of Aλ2;;jκ ðtÞ as an additional set of expansion coefficients to represent the timedependent SPFs of the upper (i.e., first) layer in a time-independent primitive basis

266

Handbook of Organic Materials for Electronic and Photonic Devices

in the lower (i.e., second) layer. In Aλ2;;jκ ðtÞ, the superscript 2 and κ correspond to the second layer and the κth mode, respectively, whereas the subscript λ refers to the indices of the respective SPF of the first layer. In the same spirit, the standard method, having only a single set of A coefficients, would correspond to a single-layer approach. ML-MCTDH builds on the idea to combine strongly correlated DOFs into p logical coordinates (combined modes or particles) yielding the set fq11 ,…, q1p g with q1κ ¼ fQ1κ , …,Qdκ g, where the superscript 1 refers to the first layer and dκ is the dimension of the κth particle. The wave packet within the combined modes’ picture for the uppermost (first) layer can be written as Ψ ðq11 ,…, q1p , tÞ ¼

n1 X

…

j1 ¼1

np X jp ¼1

A1j1 …jp ðtÞϕj11;1 ðq11 ,tÞ…ϕj1p;p ðq1p ,tÞ:

(8.20)

In ML-MCTDH, the multidimensional SPFs for the logical coordinates, ϕλ1;κ ðq1κ ,tÞ, are then considered as multidimensional wave packets of the second layer and expanded according to ϕλ1;κ ðq1κ ,tÞ ¼ ϕλ2;κ ðQ12;κ ,…,Qd2;κ κ , tÞ nκ , 1 nκ, dκ X X ¼ … Aλ2;;jκ1 …jdκ ðtÞϕ2j1;κ;1 ðQ12;κ , tÞ…ϕ2jd;κκ;dκ ðQd2;κ κ , tÞ: j1

(8.21)

j dκ

In the third layer, these wave packets are represented in the primitive basis, that is ϕ2λ ;κ;σ ðQσ2;κ , tÞ ¼

Nα X ðαÞ A3λ;;jκ;σ ðtÞχ j ðQσ2;κ Þ,

α¼σ+

j¼1

κ1 X

di :

(8.22)

i

While this illustrative example considered three layers, the general idea can easily applied to cases with more layers, including particles with decreasing dimensionality when stepping toward the bottom layer. Within this scheme, electronic states are incorporated by introducing an electronic coordinate, Qel, in the uppermost layer. By construction, ML-MCTDH allows a numerical solution of the time-dependent Schr€ odinger equation with controlled convergence. Finally, it should be pointed out that although the Huang-Rhys model provides a simple exciton-vibrational coupling scheme, leading to the uncoupled dynamics of displaced harmonic oscillators in the monomer case, the dynamics becomes highly anharmonic and nonadiabatic due to the Coulomb interaction.

8.3.2 Reduced-density matrix dynamics ML-MCTDH builds on the discretization of the SD of vibrational modes. Further, it does not include temperature. Although the latter restriction can be relaxed (see the development of this idea in Manthe and Huarte-Larran˜aga (2001) and Nest and

Frenkel exciton dynamics: A theoretical perspective

267

Kosloff (2007)), the treatment of a continuous, low-frequency bath at a finite temperature is numerically too demanding. At this point, one can resort to the reduced-density matrix theory (May and K€ uhn, 2011). For the separation of the total Hamiltonian according to Eq. (8.15), the equation of motion for the reduced-density operator ρS(t) ¼ trB[ρtot(t)] reads as (assuming ρtot(0) ¼ ρS(0)ρB(0), with ρB(0) being the Boltzmann equilibrium operator for the bath) (May and K€uhn, 2011) ∂ρS i ¼  ½HS , ρS   RρS : ħ ∂t

(8.23)

In deriving this so-called quantum master equation, the second-order perturbation theory with respect to HS-B has been invoked, as well as the Markov approximation. Phase and energy relaxation due to the system-bath interaction is introduced by the relaxation superoperator R. To guarantee correct thermalization, Eq. (8.23) should be formulated in the exciton eigenstate representation that follows diagonalization of HS. This gives the eigenstates jαi ¼

X

cm ðαÞjmi

m

(8.24)

with energies Eα. For HSB, it follows X

HSB ¼

Ku Φu ,

u¼ðα, βÞ

(8.25)

P with Ku ¼ jαihβj and Φu ¼ ξ ħωξ Qξ gαβ ðξÞ being system and bath operators, respecP tively (gαβ ðξÞ ¼ mn gmn ðξÞcm ðαÞcn ðβÞ). Using Eq. (8.25), the relaxation superoperator reads RρS ¼

X ½Ku ,Λu ρS  ρS Λ{u : u

(8.26)

The operator Λu can be expressed via the bath correlation function: Cuv ðtÞ ¼ trB ½Φu ðtÞΦs vð0ÞρB ,

(8.27)

as Λu ¼

XZ v

∞

dtCuv ðtÞKv ðtÞ:

(8.28)

0

In the present Huang-Rhys-type model, the system-bath coupling is linear in the harmonic bath coordinate and the Fourier transform of the bath correlation function takes the form given by Eq. (8.12) with ðkl,mnÞ ! ðαβ, γδÞ.

268

Handbook of Organic Materials for Electronic and Photonic Devices

To illustrate the use of Eq. (8.23), we employ the energy representation in terms of the exciton eigenstates jαi, which yields the so-called Redfield equations: X ∂ραβ ¼ iωαβ ραβ  Rαβ,κλ ρκλ : ∂t κλ

(8.29)

Further, one can invoke the Bloch approximation, which decouples populations and coherence; that is, X κλ

Rαβ,κλ ρκλ  δαβ

X ðkακ ραα  kκα ρκκ Þ + ð1  δαβ ÞΓ αβ ραβ κ

(8.30)

Here, the transition rates between pairs of states are given by the expression kαβ ¼ CHO αβ,γδ ðωαβ Þ,

(8.31)

and the dephasing rates are obtained as Γ αβ ¼

1X ðkακ + kβκ Þ: 2 κ

(8.32)

Eqs. (8.31), (8.32) establish the connection between phase and energy relaxation and the SD function. Quantum master equations like Eq. (8.23) are still the method of choice for the treatment of larger system sizes. For small and moderate system dimensions (about 10 sites), more accurate methods exist that overcome the limitations imposed by the perturbative and Markovian treatment.6 Most prominent is the so-called hierarchical equations of motion (HEOM) method, initially developed by Tanimura and Kubo (1989), and later by Ishizaki and Tanimura (2005); for reviews, see Schr€oter et al. (2015a) and Tanimura (2006). HEOM builds on a particular form of the bath correlation function (i.e., it should be expressible as a sum of exponential functions). This allows one to unravel the non-Markovian dynamics of the reduced-density matrix in terms of a hierarchy of coupled auxiliary density matrices. There are various applications to exciton dynamics in aggregates (Ishizaki and Fleming, 2009; Yan and K€uhn, 2012; Schr€ oter et al., 2015b), photosynthetic antennae (Kreisbeck and Kramer, 2012; Str€ umpfer and Schulten, 2012; Kreisbeck and Aspuru-Guzik, 2016), and polymers (Hughes et al., 2014). These density matrix approaches account for the coherent system evolution, as well as decoherence due to the system-bath interaction. Often, the opposite limit of incoherent rate transfer is applicable. In principle, it can be obtained from Eq. (8.23) by focusing only on the diagonal elements of the density matrix. More common is the

6

Note that these limitations can be partially relaxed by including a few vibrational modes into the system Hamiltonian (see, e.g., K€uhn et al., 1996; van Grondelle and Novoderezhkin, 2006; Liu and K€uhn, 2016).

Frenkel exciton dynamics: A theoretical perspective

269

approach taken by F€ orster theory, which assumes a weak Coulomb coupling, JDA, between donor and acceptor molecules such that the local equilibration at the donor precedes the actual transfer. This yields the Golden Rule-type expression for the donor-to-acceptor transfer rate (May and K€ uhn, 2011): kDA ¼

2π jJDA j2 ħ

Z

∞

dωFD ðωÞAA ðωÞ:

(8.33)

0

Here, FD(ω) and AA(ω) are the area-normalized fluorescence and absorption spectra of the donor and acceptor, respectively. In cases, where the donor and acceptor themselves consist of strongly coupled monomers, F€ orster theory still can be applied, but it has to be modified to account for the weak coupling between the individually delocalized donor and acceptor exciton states (Şener et al., 2011). F€orster theory has found wide application to describe hopping-type dynamics in molecular nanoscale systems (see, e.g., Şener et al., 2011; Beljonne et al., 2009; Scholes, 2003; Wolter et al., 2017).

8.4

Spectroscopy

The properties of excitons in molecular nanoscale systems can be investigated using electronic spectroscopy. Within the dipole approximation, the light-matter interaction Hamiltonian is given by (May and K€ uhn, 2011) Hfield ðtÞ ¼ d  EðtÞ,

(8.34)

with E(t) being the vector of the classical electric laser field. The dipole moment operator can be expressed in terms of the FE states as follows: d¼

X

dm jmih0j + h:c:,

m

(8.35)

with dm being the transition dipole matrix element at the mth monomer. Laser-driven exciton dynamics can be investigated by adding Hfield(t) to the equations of motion. In this case, the analysis of the dynamics usually involves investigating the behavior of populations and coherences pertaining to the exciton states (Schulze et al., 2016). The calculation of observables in nonlinear spectroscopy is more involved (Mukamel, 1995; K€ uhn and Lochbrunner, 2009). Here, one needs to obtain the macroscopic polarization, which is defined as the expectation value of the dipole operator (nmol is the homogeneous volume density of the molecules in the sample): Pðr,tÞ ¼ nmol trS ½dρS ðtÞ:

(8.36)

270

Handbook of Organic Materials for Electronic and Photonic Devices

In the limit of weak fields, one can apply perturbation theory with respect to Hfield(t) to obtain signals in various orders with respect to the multiple incoming fields. In general these signals can be expressed in terms of multitime correlation functions of the dipole operator. The various nonlinear signals can be differentiated by means of phasematching between the wave vectors of the incoming fields and that of the outgoing signal field (for a review, see Abramavicius et al., 2009). Absorption and emission line shapes follow from the linear response function (i.e., a dipole-dipole autocorrelation function). Provided that the eigenstates of the considered system are known (e.g., the exciton-vibrational eigenstates) and assuming homogenous line-broadening this yields the well-known Lorentzian line shape expression (see example in Section 8.5.2).

8.5

Applications

8.5.1 Spectral densities In the following discussion, the determination and analysis of SDs will be examined for the exemplary case of a PTCDI (3,4,9,10-perylene-tetracarboxylic-diimide) bulk crystal. Compared with pigment-protein complexes or polymer systems, molecular crystals have the advantage of reduced structural flexibility. Thus, the harmonic approximation for the nuclear DOFs is well justified, and the focus can be placed on the actual information provided by the SDs. Organic thin films based on perylene derivatives are discussed as candidates for applications organic or hybrid optoelectronics (Ferrere et al., 1997). Excitonic properties and the influence of vibrations have been discussed, for example, in Gisslen and Scholz (2009). Thin films of PTCDI have been studied experimentally, for example, in Topple et al. (2011), and by a combined theoretical and experimental investigation in Megow et al. (2015). Recently, we have applied molecular dynamics simulations combined with TD-DFTB7 electronic structure calculations to the determination of intramolecular and intermolecular SDs for PTCDI bulk crystals (Pl€ otz et al., 2017, 2018). The crystal structure is shown in Fig. 8.1A, where typical structural motifs (head-tail, stacked, and step) are highlighted. Fig. 8.1B–D shows spectral distributions of monomer excitation and Coulomb coupling energies. The intramolecular fluctuations are characterized by Cmm,mm(ω) in Fig. 8.1B. We notice that there are several peaks in the range between 1080 and 1550 cm1, a region typical for Franck-Condon active modes in perylene derivatives. In order to assign such spectra to specific vibrational modes, a generalization of the concept of normal modes at stationary geometries is required. This can be accomplished by using the tensorial definition of the vibrational density of states (Mathias and Baer, 2011). Generalized normal mode analysis provides effective dynamical normal modes, which take into account the thermal motion of the nuclei. 7

TD-DFTB: Linear response time-dependent density functional theory-based tight-binding (Niehaus et al., 2001; Pl€ otz et al., 2014).

Frenkel exciton dynamics: A theoretical perspective

271 14 TBFE TBFEm

Cmn,mn (ps–1)

21 10 8 6 4 2 0

(C)

(A)

1500

Step Head-tail Head-tail: dipole

0.7

500

Cmn,mn (ps–1)

Cmm,mm (ps–1)

1000

Wavenumber (cm–1) 0.8

600

400 300 200

0.6 0.5 0.4 0.3 0.2

100

(B)

500

0.9

700

0

0

0.1 0

500 1000 1500 Wavenumber (cm–1)

0.0

(D)

0

500 1000 1500 Wavenumber (cm–1)

Fig. 8.1 Spectral density of PTCDI within molecular crystal. (A) PTCDI crystal structure with exemplary head-tail (yellow, red), stacked (red, blue), and step (yellow, blue) configurations (inset: monomer structure). (B) Spectral distribution of monomer excitation energy fluctuations (inset: normal mode displacements for the strongest coupled mode). (C) Spectral distribution of Coulomb coupling fluctuations for a stacked configuration (TBFE: fully flexible, TBFEm: averaged transition charges). (D) Same as (C) (TBFE), but for the step and head-tail configurations. Also shown is the result of the dipole approximation to the head-tail case. (Reproduced from Pl€otz, P.A., Megow, J., Niehaus, T., K€ uhn, O., 2017. J. Chem. Phys. 146, 084112, with the permission of AIP Publishing.)

As an example, the inset in Fig. 8.1B shows the displacement vectors of the effective normal mode corresponding to the highest peak around 1550 cm1. The effect of thermal fluctuations on the Coulomb coupling according to Cmn,mn(ω) is shown in Fig. 8.1C and D. Two approaches are compared for the case of the stacked configuration in Fig. 8.1C. TBFE denotes the approach, including flexible transition charges, whereas for TBFEm, the transition charges are fixed to their averaged values. Clearly, the latter case cannot capture the high-frequency intramolecular contributions, which are caused by the polarization of the electron density due to nuclear motions. Comparing Fig. 8.1C and D, we notice a pronounced dependence of the intensities on the actual geometrical motif. Furthermore, Fig. 8.1D contains results obtained for the dipole approximation, which completely fails to give the correct intensities.

272

Handbook of Organic Materials for Electronic and Photonic Devices

8.5.2 Vibronic effects in absorption and emission The coupling between excitonic and vibrational DOFs leads to a complicated-level structure of exciton-vibrational eigenstates, which will be reflected in absorption and emission spectra. This is illustrated in Fig. 8.2 for the case of a PBI (chemical structure of monomer in Fig. 8.2A) dimer model, including two vibrational modes with a different frequency per monomer. This system shows interesting biphasic aggregation behavior (Pl€otz et al., 2016; Fennel et al., 2013). Provided that there is sufficient concentration, J-type, stack-slipped dimers and H-type stacked dimers exist (Fig. 8.2B). In the H-dimer, the space-demanding tert-butylphenoxy groups at the bay positions are oriented away from the contact plane. This prevents a close approach of a third molecule, which thus inhibits further aggregation. In the J-type configuration, the tertbutylphenoxy groups of both molecules are displaced from each other and can stay near the contact plane. This facilitates the association of further molecules, such that long J-aggregates are formed. There is a stronger binding for the H-dimer than for the J-dimer. As a consequence, upon increasing temperature, J-like aggregates dissociate, leaving only H-dimers intact. This causes a drop in fluorescence yield until the temperature is high enough to cause dissociation of the H-dimer as well. Different species, monomers, H- and J-dimers can be identified by their absorption and emission spectra, which have been extracted from the experimental data by fitting to an aggregation model (shown by dashed lines in the top row of Fig. 8.2C) (Pl€otz et al., 2016). Fig. 8.2C (bottom row) shows a systematic investigation of the dependence of these spectra on the strength of the Coulomb coupling. Starting from the monomer case (J ¼ 0), the absorption spectrum shows a red shift for the J-dimer (J < 0) and a blue shift for the H-dimer (J > 0). The actual shape of the spectrum is determined by the EVC. The same holds true for the emission spectrum, but in this case, the overall intensity decreases with increasing positive coupling, and the vibronic-type side band is located in the low-energy part of the main peak. For this example, comparison between calculated and experimental spectra allowed to deduce the strength of the Coulomb coupling for the different cases (Pl€ otz et al., 2016). The spectral line shapes can be understood in terms of the exciton-vibrational eigenstates, shown in Fig. 8.2D and E for absorption and emission, respectively. The left panel of Fig. 8.2D shows the exciton-vibrational eigenstates as a function of the Coulomb coupling. The states are color-coded according to the oscillator strength for transitions from the exciton-vibrational ground state. The right panel shows the same exciton-vibrational eigenstates, but color-coded with the electronic character. First, we notice that most of the transitions are actually dark. For J < 0, the spectrum is simpler, that is, the lowest transition is of electronic character, which is followed by a Franck-Condon like progression of transitions with mixed character. For J > 0, the lowest transition is of electronic character and is also dark. For small coupling, most oscillator strengths are still carried by a transition to a state that has a dominant electronic character. For larger couplings, the dominant transition is to a state with mixed electron-vibrational character.

Frenkel exciton dynamics: A theoretical perspective

273

(A)

I

M

(C)

O

E

T

T

(B)

T

(D)

(E)

Fig. 8.2 Vibronic effects on spectra of PBI dimer. (A) Chemical structure of the considered PBI derivative. (B) Geometries of H- and J-like dimers in the gas phase. (C) Absorption (left) and emission (right) spectra; top row shows selected spectra (solid: theory, dashed: experiment) from the dependence on the Coulomb coupling given in the bottom row (chosen cuts indicated by lines). (D) Coupling strength-dependent energy levels for a dimer, including a low- and a high-frequency vibrational mode. The color code gives the oscillator strength for absorption and the electronic character (scaled such that 0.25 corresponds to 100%) of the transition. (E) Same as (D), but for emission (color code oscillator strength, upper/lower panel: without/with temperature weighting (T ¼ 331 K)). (Transition frequencies and Coulomb couplings are given in units of higher vibrational frequency; transitions are shifted by an electronic gap.) (From Pl€otz, P.A., Polyutov, S.P., Ivanov, S.D., Fennel, F., Wolter, S., Niehaus, T., Xie, Z., Lochbrunner, S., W€urthner, F., K€uhn, O., 2016. PhysChemChemPhys 18, 25110–25119.)

Next, we focus on the analysis of the emission spectrum. Fig. 8.2E shows the exciton-vibrational eigenstates as a function of the Coulomb coupling and is colorcoded according to the transition from the electronically excited to the ground state. The upper panel gives the bare oscillator strength, which results in a spectrum that

274

Handbook of Organic Materials for Electronic and Photonic Devices

doesn’t resemble the one in Fig. 8.2C. This agreement is achieved only after including the Boltzmann weight for the initial states, as is done in the lower panel. In other words, for the present EVC model, it is not the oscillator strength as such, which yields vanishing emission for the H-dimer, but the negligible thermal population of those exciton-vibrational states that carry oscillator strength.

8.5.3 Energy flow in the FMO complex The capabilities of the ML-MCTDH approach are illustrated in Fig. 8.3, which contains results for the dynamics of the FMO complex (Schulze and K€uhn, 2015; Schulze et al., 2016, 2017). Fig. 8.3A shows the complex with the BChl a monomers 1–8. The decomposition of the exciton eigenenergies into local states is given in Fig. 8.3B; the exciton Hamiltonian has been taken from Moix et al. (2011). For this system, an experimental SD is available from low-temperature, site-selected fluorescence (Wendling et al., 2000). It has been discretized into 74 seven-modes per site in the interval [2:300] cm1 using the Huang-Rhys model. In the following discussion only the seven-site simulation will be covered (518 vibrational DOFs), starting from the initial population of site 1, assuming that only this site is connected to the chlorosome antenna. The site population dynamics, Pm(t) ¼ ρmm(t) ¼ hmjΨ (t)ihΨ (t)jni, is shown in the lower right of Fig. 8.3D. First, we notice that only sites 1–3 are noticeably populated, which is already clear from inspecting the decomposition of the exciton energies. Second, there is a pronounced population beating between sites 1 and 2, which is a consequence of the strong Coulomb coupling, as well as the local initial state preparation. The populations of sites 1 and 2 decay along with the rise of the population of site 3; the latter shows almost no oscillation. It is interesting to note that this model nicely reproduces typical time scales for this complex, with no fit parameter. Moreover, it yields coherence dephasing and population relaxation by virtue of the highdimensional vibrational space coupled to the exciton dynamics. While similar results can be obtained, for example, by using the HEOM approach (Kreisbeck and Kramer, 2012), explicit wave packet propagation provides a means to follow the dynamics of the “bath” vibrations explicitly. In Fig. 8.3 we plot local vibronic (i.e., in an electronically excited state; upper row) and vibrational (i.e., in an electronically ground state; lower row) energies (with respect to the zero-point energy), according to the modes taken from the SD (shown at left). Two basic mechanisms can be identified from this plot: (1) vibronic resonance-assisted exciton transfer between sites 2 and 3, giving a pronounced vibronic excitation in the vicinity of the electronic gap (cf. Fig. 8.3B and C); (2) Vibrational excitation in the electronic ground states of sites 1 and 2, which requires that the modes’ frequencies exceed the transition frequency due to the Coulomb coupling between these sites. This leads to the threshold behavior at about 160 cm1. The bare excitonic transition frequencies for the considered population flow are in the range of a peak in the SD, as can be seen in Fig. 8.3C. In view of the differences between experimental and calculated SDs (e.g., Wendling et al., 2000; Lee and Coker, 2016) one might ask whether the actual details really matter with the exciton dynamics. Using ML-MCTDH, it has been found that the efficiency of transfer to site 3 is

Frenkel exciton dynamics: A theoretical perspective

275

Energy (cm–1)

500

1 8

6

300 200 100 0

7

5

400

(B)

2

1 2 3 4 5 6 7 8

3

(A)

0

1

2

0.012 0.010

0.003

0.006 0.008

0.002

0.004

0.001 0

(C) 300

Experiment Discrete modes

0.004

0

0.002

Huang-Rhys factor

4

Spectral density (1/cm–1)

BChl a monomer 0.005

0 100 150 200 250 300

50

Wavenumber (cm–1)

3

4

5

6

250 200

Wavenumber (cm–1)

150 100 50 0

Vibro-exc-site 2

Vibro-exc-site 1

Vibro-exc-site 3

0.1

250

0.8

200

0

50

100

150

0.6

0.2

50

(D)

150

0.4

100

0

1

Vibra-exc-site 1 0 200 400 600

800

Vibra-exc-site 2 0 200 400 600

800

0

200

400

600

800 1000

Time (fs)

Fig. 8.3 Quantum simulation of energy flow in the FMO complex. (A) Monomeric FMO pigment-protein complex with relevant BChl a molecules labeled as 1–8 (Tronrud et al., 2009). (B) Spectrum of exciton energies and their decomposition into local states. (C) Experimental SD (Wendling et al., 2000) and stick spectrum for discretized model (arrows: relevant transitions, see panel (B)). (D) Population dynamics of a seven-site model (lower right) and local nonequilibrium vibrational/vibronic energy of modes (left: SD). (Reproduced from Schulze, J., Shibl, M.F., Al-Marri, M.J., K€ uhn, O., 2016. J. Chem. Phys. 144, 185101, with the permission of AIP Publishing.)

276

Handbook of Organic Materials for Electronic and Photonic Devices

essentially governed by the total Huang-Rhys factor, whereas the details of its spectral distribution are relevant for the nonequilibrium vibronic and vibrational excitation, as well as for the pattern of exciton relaxation (Schulze et al., 2017).

8.6

Final remarks

Simulations of exciton dynamics and spectroscopy have made a major step forward in the recent years. Nowadays, methods of electronic structure theory are capable of making realistic predictions of electronic transition energies and Coulomb couplings, as well as of the influence of the coupling to nuclear DOFs. This gives access to the FE Hamiltonian (supplemented by CT excitations, if necessary), including spectral densities that can be interpreted within the Huang-Rhys type model of shifted harmonic oscillator modes. This setup is ideally suited for quantum dynamics simulations, using either the density matrix or the ML-MCTDH approach. In particular, the latter has the advantage of providing access to all details of the exciton-vibrational wavepacket. This enables one to unravel effects of vibrations on the coherent quantum dynamics that go far beyond their previously assumed role as a heat bath for disposal of excess energy. Exploiting coherence, for example, to enhance the performance of artificial light-harvesting and conversion, will be a future challenge.

Acknowledgments Financial support from the Deutsche Forschungsgemeinschaft (Sfb 652) is gratefully acknowledged. Part of this work was made possible by the NPRP Grant No. NPRP 7-227-1-034 from the Qatar National Research Fund (a member of the Qatar Foundation). I am particularly grateful to Per-Arno Pl€otz and Jan Schulze (University of Rostock) for their enthusiastic work on the PTCDI/PBI and FMO project, respectively.

References Abramavicius, D., Palmieri, B., Voronine, D.V., Sanda, F., Mukamel, S., 2009. Chem. Rev. 109, 2350–2408. Agranovich, V., 2009. Excitations in Organic Solids. Oxford University Press, Oxford. Ambrosek, D., K€ohn, A., Schulze, J., K€uhn, O., 2012. J. Phys. Chem. A 116, 11451–11458. Arago´, J., Troisi, A., 2015. J. Chem. Phys. 142, 164107. Beck, M.H., J€ackle, A., Worth, G.A., Meyer, H.D., 2000. Phys. Rep. 324, 1–105. Beljonne, D., Curutchet, C., Scholes, G.D., Silbey, R.J., 2009. J. Phys. Chem. B 113, 6583–6599. Blankenship, R.E., 2014. Molecular Mechanisms of Photosynthesis. Wiley-Blackwell, Hoboken, NJ. Bredas, J.L., Sargent, E.H., Scholes, G.D., 2017. Nat. Mater. 16, 35–44. Brixner, T., Stenger, J., Vaswani, H.M., Cho, M., Blankenship, R.E., Fleming, G.R., 2005. Nature 434, 625–628. Butkus, V., Valkunas, L., Abramavicius, D., 2014. J. Chem. Phys. 140, 034306.

Frenkel exciton dynamics: A theoretical perspective

277

Chachisvilis, M., K€ uhn, O., Pullerits, T., Sundstr€om, V., 1997. J. Phys. Chem. B 101, 7275–7283. Davydov, A.S., 1964. Sov. Phys. Usp. 7, 145. Duan, H.G., Prokhorenko, V.I., Cogdell, R.J., Ashraf, K., Stevens, A.L., Thorwart, M., Miller, R.J.D., 2017. Proc. Natl. Acad. Sci. USA 114, 8493–8498. Eisfeld, A., Braun, L., Strunz, W.T., Briggs, J.S., Beck, J., Engel, V., 2005. J. Chem. Phys. 122, 134103. Engel, G.S., Calhoun, T.R., Read, E.L., Ahn, T.K., Mancˇal, T., Cheng, Y.C., Blankenship, R.E., Fleming, G.R., 2007. Nature 446, 782–786. Fennel, F., Wolter, S., Xie, Z., Pl€otz, P.A., K€uhn, O., W€ urthner, F., Lochbrunner, S., 2013. J. Am. Chem. Soc. 135, 18722–18725. Ferrere, S., Zaban, A., Gregg, B.A., 1997. J. Phys. Chem. B 101, 4490–4493. Fornari, R.P., Arago´, J., Troisi, A., 2016. J. Phys. Chem. C 120, 7987–7996. Frenkel, J., 1931. Phys. Rev. 37, 17–44. Gisslen, L., Scholz, R., 2009. Phys. Rev. B 80, 115309. Holstein, T., 1959. Ann. Phys. 8, 325–342. Hughes, K.H., Cahier, B., Martinazzo, R., Tamura, H., Burghardt, I., 2014. Chem. Phys. 442, 111–118. Ishizaki, A., Fleming, G.R., 2009. J. Chem. Phys. 130, 234111. Ishizaki, A., Tanimura, Y., 2005. J. Phys. Soc. Jpn 74, 3131–3134. Ishizaki, A., Calhoun, T.R., Schlau-Cohen, G.S., Fleming, G.R., 2010. PhysChemChemPhys 12, 7319–7337. Kreisbeck, C., Aspuru-Guzik, A., 2016. Chem. Sci. 7, 4174–4183. Kreisbeck, C., Kramer, T., 2012. J. Phys. Chem. Lett. 3, 2828–2833. K€ uhn, O., Lochbrunner, S., 2009. In: David, L.A. (Ed.), Encyclopedia of Applied Spectroscopy. Wiley-VCH, Weinheim, p. 769. K€ uhn, O., Sundstr€om, V., 1997. J. Chem. Phys. 107, 4154–4164. K€ uhn, O., Renger, T., May, V., 1996. Chem. Phys. 204, 99–114. K€ uhn, O., Renger, T., May, V., Voigt, J., Pullerits, T., Sundstr€ om, V., 1997. Trends Photochem. Photobiol. 4, 213–256. Lee, M.K., Coker, D.F., 2016. J. Phys. Chem. Lett. 7, 3171–3178. Liu, X., K€uhn, O., 2016. Chem. Phys. 481, 272–280. Liu, W., Lunkenheimer, B., Settels, V., Engels, B., Fink, R.F., K€ ohn, A., 2015. J. Chem. Phys. 143, 084106. Manthe, U., 2008. J. Chem. Phys. 128, 164116. Manthe, U., Huarte-Larran˜aga, F., 2001. Chem. Phys. Lett. 349, 321–328. Mathias, G., Baer, M.D., 2011. J. Chem. Theory Comput. 7, 2028–2039. May, V., K€uhn, O., 2011. Charge and Energy Transfer Dynamics in Molecular Systems. WileyVCH, Weinheim. Megow, J., K€orzd€orfer, T., Renger, T., Sparenberg, M., Blumstengel, S., Henneberger, F., May, V., 2015. J. Phys. Chem. C 119, 5747–5751. Merrifield, R.E., 1961. J. Chem. Phys. 34, 1835–1839. Meyer, H.D., 2011. WIREs Comput. Mol. Sci. 2, 351–374. Meyer, H.D., Manthe, U., Cederbaum, L.S., 1990. Chem. Phys. Lett. 165, 73–78. Meyer, H.D., Gatti, F., Worth, G., 2009. Multidimensional Quantum Dynamics. Wiley-VCH, Weinheim. Moix, J., Wu, J., Huo, P., Coker, D., Cao, J., 2011. J. Phys. Chem. Lett. 2, 3045–3052. Mukamel, S., 1995. Principles of Nonlinear Optical Spectroscopy. Oxford University Press, Oxford.

278

Handbook of Organic Materials for Electronic and Photonic Devices

Nest, M., Kosloff, R., 2007. J. Chem. Phys. 127, 134711. Niehaus, T.A., Suhai, S., Della Sala, F., Lugli, P., Elstner, M., Seifert, G., Frauenheim, T., 2001. Phys. Rev. B 63, 085108. Olbrich, C., Str€umpfer, J., Schulten, K., Kleinekath€ofer, U., 2011. J. Phys. Chem. Lett. 2, 1771–1776. Panitchayangkoon, G., Hayes, D., Fransted, K.A., Caram, J.R., Harel, E., Wen, J., Blankenship, R.E., Engel, G.S., 2010. Proc. Natl. Acad. Sci. USA 107, 12766–12770. Philpott, M.R., 1967. J. Chem. Phys. 47, 4437. Pl€ otz, P.A., Niehaus, T., K€uhn, O., 2014. J. Chem. Phys. 140, 174101. Pl€ otz, P.A., Polyutov, S.P., Ivanov, S.D., Fennel, F., Wolter, S., Niehaus, T., Xie, Z., Lochbrunner, S., W€urthner, F., K€uhn, O., 2016. PhysChemChemPhys 18, 25110–25119. Pl€ otz, P.A., Megow, J., Niehaus, T., K€uhn, O., 2017. J. Chem. Phys. 146, 084112. Pl€ otz, P.-A., Megow, J., Niehaus, T., K€uhn, O., 2018. J. Chem. Theory Comput. 14, 5001–5010. Polkehn, M., Eisenbrandt, P., Tamura, H., Burghardt, I., 2018. Int. J. Quantum Chem. 118, e25502. Polyutov, S., K€uhn, O., Pullerits, T., 2012. Chem. Phys. 394, 21–28. Renger, T., May, V., K€uhn, O., 2001. Phys. Rep. 343, 137–254. Renger, T., Klinger, A., Steinecker, F., Schmidt am Busch, M., Numata, J., M€ uh, F., 2012. J. Phys. Chem. B 116, 14565–14580. Romero, E., Augulis, R., Novoderezhkin, V.I., Ferretti, M., Thieme, J., Zigmantas, D., van Grondelle, R., 2014. Nat. Phys. 10, 676–682. Scharber, M.C., Sariciftci, N.S., 2013. Prog. Polym. Sci. 38, 1929–1940. Scholes, G.D., 2003. Annu. Rev. Phys. Chem. 54, 57–87. Scholes, G.D., Rumbles, G., 2006. Nat. Mat. 5, 683–696. Scholes, G.D., Fleming, G.R., Chen, L.X., Aspuru-Guzik, A., Buchleitner, A., Coker, D.F., Engel, G.S., van Grondelle, R., Ishizaki, A., Jonas, D.M., Lundeen, J.S., McCusker, J.K., Mukamel, S., Ogilvie, J.P., Olaya-Castro, A., Ratner, M.A., Spano, F.C., Whaley, K.B., Zhu, X., 2017. Nature 543, 647. Schr€ oter, M., K€uhn, O., 2013. J. Phys. Chem. A 117, 7580. Schr€ oter, M., Ivanov, S.D., Schulze, J., Polyutov, S.P., Yan, Y., Pullerits, T., K€ uhn, O., 2015. Phys. Rep. 567, 1–78. Schr€ oter, M., Pullerits, T., K€uhn, O., 2015. Ann. Phys. 527, 536–545. Schulze, J., K€uhn, O., 2015. J. Phys. Chem. B 119, 6211–6216. Schulze, J., Shibl, M.F., Al-Marri, M.J., K€uhn, O., 2016. J. Chem. Phys. 144, 185101. Schulze, J., Shibl, M.F., Al-Marri, M.J., K€uhn, O., 2017. Chem. Phys. 497, 10–16. Schwoerer, M., Wolf, H.C., 2007. Organic Molecular Solids. Wiley-VCH, Weinheim. Seibt, J., Winkler, T., Renziehausen, K., Dehm, V., W€urthner, F., Meyer, H.D., Engel, V., 2009. J. Phys. Chem. A 113, 13475–13482. Şener, M., Str€umpfer, J., Hsin, J., Chandler, D., Scheuring, S., Hunter, C.N., Schulten, K., 2011. ChemPhysChem 12, 518–531. Shibl, M.F., Schulze, J., Al-Marri, M.J., K€uhn, O., 2017. J. Phys. B At. Mol. Opt. Phys. 50, 184001. Str€ umpfer, J., Schulten, K., 2012. J. Chem. Theory Comput. 8, 2808–2816. Tamura, H., Burghardt, I., 2013. J. Am. Chem. Soc. 135, 16364–16367. Tamura, H., Huix-Rotllant, M., Burghardt, I., Olivier, Y., Beljonne, D., 2015. Phys. Rev. Lett. 115, 107401. Tanimura, Y., 2006. J. Phys. Soc. Jpn 75, 082001. Tanimura, Y., Kubo, R., 1989. J. Phys. Soc. Jpn 58, 101. Topple, J.M., Burke, S.A., Ji, W., Fostner, S., Tekiel, A., Gr€ utter, P., 2011. J. Phys. Chem. C 115, 217–224.

Frenkel exciton dynamics: A theoretical perspective

279

Tronrud, D.E., Wen, J., Gay, L., Blankenship, R.E., 2009. Photosynth. Res. 100, 79–87. van Grondelle, R., Novoderezhkin, V.I., 2006. PhysChemChemPhys 8, 793–807. Vendrell, O., Meyer, H.D., 2011. J. Chem. Phys. 134, 044135. Wang, H., Thoss, M., 2003. J. Chem. Phys. 119, 1289–1299. Wendling, M., Pullerits, T., Przyjalgowski, M.A., Vulto, S.I.E., Aartsma, T.J., van Grondelle, R., van Amerongen, H., 2000. J. Phys. Chem. B 104, 5825. Woller, J.G., Hannestad, J.K., Albinsson, B., 2013. J. Am. Chem. Soc. 135, 2759–2768. Wolter, S., Westphal, K.M., Hempel, M., W€urthner, F., K€ uhn, O., Lochbrunner, S., 2017. J. Phys. B At. Mol. Opt. Phys. 50, 184005. W€ urthner, F., Kaiser, T.E., Saha-M€oller, C.R., 2011. Angew. Chem. Int. Ed. 50, 3376–3410. Xie, Y., Jiang, S., Zheng, J., Lan, Z., 2017. J. Phys. Chem. A 121, 9567–9578. Yan, Y., K€uhn, O., 2012. New J. Phys. 14, 105004.

This page intentionally left blank

Strong light-matter interactions and exciton-polaritons in organic materials

9

Arko Graf*,†,a, Laura Tropf*,a, Jana Zaumseil†, Malte C. Gather* *Organic Semiconductor Centre, SUPA, School of Physics and Astronomy, University of St Andrews, St Andrews, United Kingdom, †Institute for Physical Chemistry, Universit€at Heidelberg, Heidelberg, Germany

9.1

Introduction

Polariton is the general term for any mixed light-matter excitation. Polaritons form when photons strongly couple for instance to plasmons, phonons, or excitons. This chapter focuses on exciton-polaritons, which are mixed light-matter quasiparticles that form when a cavity photon couples to an exciton (Fig. 9.1A). The properties of the exciton and the photon are significantly altered if (1) photon and exciton are in resonance (i.e., a photon can excite an exciton and a decaying exciton will create a cavity photon); and (2) the rate of energy exchange is faster than the decay rates of both constituents. In this so-called strong coupling regime, the photon and exciton can no longer be distinguished and the excitations of the system are described more accurately by exciton-polaritons. The properties of these hybrid particles are intriguing because they are a mix of those of the photon and the exciton: Like the photon, they have a very light mass and can be probed optically. From the exciton, they inherit large interaction cross sections and spin. This mixture of properties can be accurately controlled through material design and cavity engineering and offers opportunities for numerous potential applications (e.g., quantum simulators, optically controlled spin switches, or tuning energy levels to control chemical reactions or excitonic devices). Moreover, polaritons follow bosonic statistics. This can give rise to a stimulated final state population, and thus to polariton condensation or polariton lasing. Polariton lasers are of interest, as they can have lower lasing thresholds than conventional lasers. Before discussing the properties of these new quasiparticles and the advantages of organic polaritons in detail, let us first take a look at the separate constituents (i.e., the cavity photon and the exciton) to explore how polaritons form. A cavity photon is confined between two mirrors. This leads to a deviation from the dispersion relation of the free space photon, where E ¼ ℏck with photon energy E and photon wave vector

a

These authors contributed equally.

Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00009-7 © 2019 Elsevier Ltd. All rights reserved.

282

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 9.1 (A) Schematic representation of an empty (left) and strongly coupled (center) microcavity with metal mirrors (bottom: opaque, top: semitransparent) and a bare organic film with excitons (right). (B) Corresponding dispersion relations of the weakly or uncoupled exciton and cavity photon (gray dashed lines) and of the strongly coupled system, where the photon and the exciton form upper (UP, blue dash-dotted line) and lower (LP, green line) polariton branches. Also marked are the detuning δ and Rabi-splitting ℏΩ.

k. Due to the confinement, only discrete wavelengths (and thus energies) are allowed in the direction perpendicular to the mirrors. For a cavity of thickness LC, we have λ? C, j ¼

2LC nC hc ) E? j, j ¼ 1, 2,… C, j ¼ 2LC nC j

(9.1)

Here, nC denotes the refractive index in the cavity and j the index number of the mode. When taking into account all angles and considering in-plane wave vectors to be much smaller than those perpendicular to the mirrors (kk ≪ k?), this yields the dispersion relation for the lowest-energy cavity photon (j ¼ 1): hc EC ¼ nC

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  2 ℏ2 kk2 2π hc + kk2  + , LC 2LC nC 2mC

(9.2)

ℏπ with the band mass of the photon mC ¼ cL  105 me . Because kk is closely related to C the emission angle θ, the dispersion relation can be directly observed via angledependent or far-field optical probing. In fact, all fundamental properties of the polaritons inside the cavity are probed by investigating the characteristics of the light leaking out of the cavity. Excitons of a disordered system, as found in most organic systems, do not show any dispersion; i.e., EX(kjj)¼const.1 The dispersion relations of an exciton and a photon in an uncoupled or weakly coupled cavity system are shown as dashed lines in Fig. 9.1B. While the energy of the exciton is fixed and depends solely on the material, the photon energy can be tuned by changing the thickness of the cavity. Their difference at kk ¼ 0 is called the detuning of the cavity: δ ¼ EX  EC.

1

In periodically ordered systems, excitons do acquire a band mass, but as this is a lot higher than mC, it can be approximated to be flat as well.

Strong light-matter interactions and exciton-polaritons in organic materials

283

If the cavity photon is brought close to resonance with the exciton, the system will couple because the excitation of the exciton is concurrent with the absorption of a photon, and in turn, its relaxation results in the emission of a photon. The magnitude of the coupling constant depends on the oscillator strength of the molecule, f, on the number qffiffiffiffiffiffiffi of molecules, N, and on the mode volume V: g∝ f N V . If this coupling constant is large with respect to the losses, the system enters the strong coupling regime and the exciton and photon branch split into the lower polariton (LP) and upper polariton (UP) branches, as illustrated by the solid and dash-dotted lines in Fig. 9.1B. This splitting is called anticrossing of the uncoupled branches and is a key characteristic of the polariton dispersion. In a simplified model, the strongly coupled system can be described by a two-level coupled oscillator. This model approximates the exciton as a single level and does not consider the inhomogeneous broadening caused by disorder, nor necessarily multiple vibronic replica. Nevertheless, this simple model was shown to be sufficient to describe microcavities containing organic crystals, small molecules, and polymers (Kena-Cohen et al., 2008; Holmes and Forrest, 2004; Takada et al., 2003). The two coupled states are the photonic cavity mode (with energy Ec and line width γ c, which originates from the finite lifetime of the cavity photon) and the exciton (with energy Ex and homogeneous line width γ x, which originates from the finite lifetime of the exciton). The diagonalization of the Hamiltonian yields the new eigenstates of the system, UP and LP, with energies EUP and ELP, respectively: EUP=LP ¼

ðEx + Ec Þ  iðγ x + γ c Þ 1  2 2

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 4g2 + ½ðEx  Ec Þ  iðγ x  γ c Þ2 :

(9.3)

Because the photon energy depends on the angle of the light inside the cavity and the cavity thickness, so do the energy and properties of the polaritons. At the point of resonance, Ec  Ex ¼ 0, both polaritons consist of equal parts of exciton and photon, and the splitting of the modes is minimal. The splitting energy at this point is called Rabi splitting: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ℏΩ ¼ 4g2  ðγ x  γ c Þ2 :

(9.4)

It is a measure for the coupling strength g of the system. Its form clearly shows that there is a limit for the strong coupling regime, which depends on the losses of the system: g

jγ x  γ c j : 2

(9.5)

This statement about the threshold for strong coupling is unrelated to the (experimentally accessible) visibility of the splitting. The visibility, also an important quantity in practice, will depend on the sum of the line widths of both constituents; i.e., strong coupling will be clearly resolvable only if

284

Handbook of Organic Materials for Electronic and Photonic Devices

ℏΩ 

γx + γc : 2

(9.6)

Having understood under which circumstances we can observe polaritons, we will now briefly summarize their key characteristics, some of which will be discussed in detail later in this chapter. Depending on the detuning between exciton and photon, the branches contain different fractions of the cavity photon and exciton: With increasing photon energy, the UP will become more photonic and the LP will become more excitonic, and vice versa. As a consequence of the polariton character being tunable from mainly photonic to largely excitonic, we can control the properties of these new quasiparticles, like the responsiveness to magnetic fields, the interaction between polaritons, and the band mass (which can even change signs). The regions of negative band mass give rise to nonlinear effects like bistability (Baas et al., 2004) or solitons (Sich et al., 2011). Because energy minimization is one of the most fundamental physical principles, the ability to tune and manipulate the energy levels of a strongly coupled microcavity could be employed in numerous applications. In the realm of organic materials, for example, one could influence chemical reactions (Hutchison et al., 2012) or enhance the efficiency of photovoltaic devices (PVDs) by tuning the singlet fission rate (Martı´nez-Martı´nez et al., 2018). Another interesting aspect is the delocalization of the polaritonic wave packet, in contrast to the highly localized organic exciton. There have been indications that it may be possible to exploit this delocalization to enhance charge transport behavior (Orgiu et al., 2015); however, the question of whether these findings can be generalized to other materials and systems is still controversial (Hagenm€ uller et al., 2017; Graf et al., 2017). Moreover, light-matter hybridization allows the light in the cavity to be manipulated via its excitonic fraction, such as by electric and magnetic fields to which photons are not usually responsive—a highly promising feature for opto-electronic applications (Sanvitto and Kena-Cohen, 2016). Further intriguing properties arise from the bosonic nature of polaritons, which is a result of both the exciton and the photon being bosons. Like all bosons, polaritons undergo stimulated scattering into their common ground state and form a condensate at sufficiently high densities (Kasprzak et al., 2006). Superfluidity and vortex formation have been observed as well (Amo et al., 2011; Lagoudakis et al., 2008). Polariton condensates are unusual for two reasons: (1) the extensive tunability of the system and (2) condensation can also occur at room temperature due to the low mass of the LP2 (Daskalakis et al., 2013, 2014; Plumhof et al., 2013). Most research on strongly coupled systems to date has been performed on inorganic semiconductors (Kasprzak et al., 2006; Weisbuch et al., 1992; Schneider et al., 2013). These inorganic microcavities are mostly fabricated by epitaxial growth, the observed

2

The limiting factor for the temperature of the condensates is usually the binding energy of the excitons, which renders excitons in most inorganic materials unstable at room temperature. However, organic materials, but also wide-band-gap semiconductors or transition metal dichalcogenides, persist well above room temperature.

Strong light-matter interactions and exciton-polaritons in organic materials

285

Rabi splittings lie around 10 meV and the exciton binding energies typically around 4 meV (Sanvitto and Kena-Cohen, 2016). While these materials were shown to exhibit exciting effects like Bose-Einstein condensation or quantum fluidity, the low exciton-binding energies restrict the phenomena to cryogenic temperatures (Byrnes et al., 2014). Wide-band-gap semiconductors (e.g., GaN or ZnO), which have higher binding energies and larger coupling strengths, in turn present technological challenges (Sanvitto and Kena-Cohen, 2016). Organic semiconductors, in contrast, often exhibit far greater Rabi splitting and exciton-binding energy than inorganic materials. As a consequence, organic polaritons persist up to room temperature, which, along with the relative ease of fabrication, makes them particularly appealing for applications. The challenges that one faces when using organic materials for strong coupling arise from the complex nature of the disordered environment and their limited stability. This chapter reviews the current state of the art of experimental research in organic polaritons. We start by covering the different materials used for strong coupling investigations. Next, we review reports on the strong coupling of more than one excitonic state to the cavity photon, which is of interest (e.g., to mix excitonic states of organic and inorganic materials). The subsequent section then presents studies on polariton condensation, and the last section expands on electrically pumped exciton-polaritons before we summarize and give our perspective of possible future developments.

9.2

Materials for strong light-matter coupling

This section provides an overview of the various organic materials that have been used to study strong coupling in organic microcavities. First, we compare the sample design of inorganic and organic microcavities, focusing on different mirror materials. Then, we present the organic materials used for strong coupling. Finally, unconventional organic materials for strong coupling (i.e., materials at the crossover to inorganic or biological materials) will be reviewed. A summary of the studied materials, cavities, and their strong coupling parameters is given in Table 9.1 (though it is not exhaustive). An overview of the chemical structures of some of the organic compounds reviewed in this chapter appears in Fig. 9.2. Strong coupling phenomena were first discovered in microcavities containing Rydberg atoms and later, semiconductor quantum wells (Weisbuch et al., 1992; Rempe et al., 1987). These cavities used distributed Bragg reflectors (DBRs) as mirrors, resulting in high-quality factors (typically Q > 1000). The first experiments on organic materials adjusted this sample design to the material characteristics of organic semiconductors and used fabrication processes that are compatible with these. For example, instead of confining the excitons in quantum wells as in most inorganic systems, the entire cavity was filled with the active material. The top DBR was omitted because the deposition of DBRs on top of organic semiconductors without damaging the organic material was considered challenging to achieve. Instead, the cavities consisted of a bottom DBR mirror, on top of which the organic film was processed.

Table 9.1 Summary of organic materials used for strong coupling and relevant parameters Absorption width (meV)

ħΩ (meV)

Dorg (nm)

Material class

Material

Oscillator strength/ max. abs. coeff.

Aggregating small molecules

4TBPPZn

2 1015 cm2

90  5

80…160

100

NTCDA

?

150

360

60

TDBC

70

265

6

Cyanine dye

106 cm1 (George et al., 2015) ?

45  5

300

190

Cyanine dye

?

58

180

120

SPI-MC

?

450

700

PBPS

1.3 105 cm1

200

J-aggregates

Polymers

MeLPPP Organic crystals Amorphous small molecules Perovskites

Concentration

Cavity

Varied in polystyrene Neat film

DBRAg DBRAl Ag-Ag

126

Neat film / bilayer 5 1019 mol cm1 in PVA Unknown in PVA 60% in PMMA

430

120

Neat film

60

116

35

Neat film

Anthracene

?

10

216

70

Neat film

TDAF

?

500

980

70

Neat film

(C6H5C2H4– NH3)2PbI4

5 1015 cm2

80

140

50

Neat film

Ag-Ag DBRAg Ag-Ag DBRAl DBRDBR DBRDBR Al-Al DBRAg

First author of refs. Lidzey et al. (1998) Holmes and Forrest (2004) Tischler et al. (2005) Hobson et al. (2002) Lidzey et al. (1999) Schwartz et al. (2011) Takada et al. (2003) Plumhof et al. (2013) Kena-Cohen et al. (2008) Kena-Cohen et al. (2013) Brehier et al. (2006)

Transition metal dichalcogenides Biological materials

Carbon allotropes

MoS2

?

30

46

0.65

Neat film

DBRDBR

Liu et al. (2014)

eGFP

?

150

195

500

Neat film

Chlorosomes

?

105

150

205

20% in PVA

DBRDBR Ag-Ag

(6,5) SWCNTs

240 cm1

21

112

250

Dibenzo[hi,st] ovalene (“nanographene”)

1.8 105 M1 cm1 (Paterno` et al., 2017)

135

126

540

1.7% in polymer 50% in PMMA

Dietrich et al. (2016) Coles et al. (2014b) Graf et al. (2016) Coles et al. (2017a)

Au-Au Al-Al

288

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 9.2 Overview of the chemical structures of the organic materials used for strong coupling.

The top mirror was a thermally evaporated metal film of aluminium or silver (Lidzey et al., 1998). Later experiments showed that cavities with two metal mirrors actually exhibit an enhanced mode splitting due to better mode confinement (Hobson et al., 2002). Only after this was established, this simpler fabrication method, which leads to cavities with even lower quality factors, became widely accepted (Kena-Cohen et al., 2013; Mazzeo et al., 2014; Tischler et al., 2005). Nevertheless, cavities with high Q-factors, which are crucial for obtaining sufficiently long polariton lifetimes (and thus for observing nonlinear effects, for instance) still rely on highly reflective mirrors like DBRs. Deposition of DBRs by sputtering on top of organic materials was demonstrated in 2001 (Anni et al., 2001), and indeed it was first used for strongly coupled microcavities in 2004 (Song et al., 2004). However, DBR-clad cavities with organic materials were only routinely fabricated for strong coupling experiments when lasing from organic polaritons was pursued and demonstrated (Kena-Cohen et al., 2008; Kena-Cohen and Forrest, 2010).

Strong light-matter interactions and exciton-polaritons in organic materials

289

Initial experiments that used organic semiconductors for strong coupling tried to closely imitate the material characteristics of pioneering research on inorganic microcavities (i.e., their spectrally narrow absorption). Thus, the first demonstration of organic exciton polaritons by Lidzey et al. coupled the cavity photon to the narrow Soret band of 4TBPPZn (see Fig. 9.2), which was doped into polystyrene (Lidzey et al., 1998). Due to a large Rabi splitting ( 160 meV) in these structures, the mode anticrossing was clearly observable in reflectance, despite the low Q-factor of the cavity (Q  85) relative to inorganic microcavities. The authors tuned the Rabi splitting by changing the concentration of the absorber in the film, which is arguably more easily achieved with organic materials than with inorganic quantum well structures. Potential applications of organic polaritons, however, rely on the luminescence from strongly coupled cavities (Sanvitto and Kena-Cohen, 2016), which is expected to be very weak in 4TBPPZn due to its low photoluminescence quantum yield (PLQY) (Lidzey et al., 1998). Photoluminescence (PL) from organic polaritons was thus first observed using a different material class: cyanine dyes, which form J-aggregates when processed from solution and show a narrow absorption line width, a small Stokes shift, and furthermore, a high PLQY (Lidzey et al., 1999). The J-aggregate in a polyvinylalcohol matrix was clad between a bottom DBR and a top silver mirror. As in previous experiments with inorganic semiconductors, luminescence was observed only from the LP. This is expected because the system aims to minimize its energy and thus, relaxation pathways from the UP to the LP are efficient. The same cyanine dye was later also used to demonstrate polariton formation in a silver-clad metal-only cavity (Hobson et al., 2002). In 2005, Tischler et al. reported the first electroluminescence from polaritons in organic materials, three years before inorganic polariton LEDs were demonstrated (Tischler et al., 2005). They used metal films both as mirrors and as electrodes on either side of J-aggregated films of the cyanine dye TDBC (see Fig. 9.2). Molecules that can form J-aggregates dominated organic polariton experiments for a long time (Hobson et al., 2002; Tischler et al., 2005; Lidzey et al., 1999; Schouwink et al., 2002; Oda et al., 2009; Virgili et al., 2011; Coles et al., 2014a). The reasons for this were (1) their narrow line width due to the high order in the stacked aggregates and (2) their large oscillator strength, which arises from the delocalization of the exciton over several aggregated molecules. The same properties were sought in other organic materials, and thus the number of material classes showing strong coupling was extended. For instance, it was established that polymers also could support the formation of polaritons (Takada et al., 2003). The first demonstration of strong coupling in an organic polymer used a σ-conjugated polysilane with a particularly rigid silicon backbone, and thus a spectrally narrow absorption band. MeLPPP, the polymer that enabled the first observation of organic polariton condensation in two dimensions (2D) and in zero dimensions (0D), also has a rigid backbone and thus reduced disorder for increased π-conjugation (see Fig. 9.2; Plumhof et al., 2013; Scafirimuto et al., 2017). Similarly, the first report on strong coupling in thermally evaporated organic materials (often referred to as “small molecules”) used the compound 1,4,5,8naphthalenetetracarboxylic dianhydride (NTCDA; see Fig. 9.2), which has a much

290

Handbook of Organic Materials for Electronic and Photonic Devices

narrower absorption line width than other representatives of this class of materials (Holmes and Forrest, 2004). The small extent of inhomogeneous broadening is a result of the polycrystalline morphology (i.e., ordered structure) of the deposited NTCDA film. The reduced disorder also revealed the excitonic substructure in its absorption and emission spectrum, which originates from different vibronic transitions. As a consequence, a middle polariton (MP) was observed in the cavity between the 0–0 and the 0–1 transitions. The strategy of approximating organic systems to their inorganic counterpart was perfected by Kena-Cohen et al., who grew a single crystal of anthracene between two DBRs (Kena-Cohen et al., 2008). The single crystal showed three clearly distinct excitonic modes, which strongly coupled to the photon mode. Furthermore, a clear dependence of the splitting on the crystal axes was demonstrated. This approach proved successful because the first polariton lasing in organic materials was later observed in an anthracene single-crystal microcavity (Kena-Cohen and Forrest, 2010). Other experiments, however, explored organic materials with relatively broad absorption spectra (several hundreds of millielectron volts), both processed from solution and using vacuum deposition techniques (Kena-Cohen et al., 2013; Schwartz et al., 2011). In these materials (namely, a switchable spiropyran/merocyanine system and the small-molecule TDAF, see Fig. 9.2), the observed mode splitting was so large (≳ 25% of the exciton energy) that the cavities entered the so-called ultrastrong coupling regime (Ciuti et al., 2005). In this regime, antiresonant coupling terms in the Hamiltonian become important and result in a strongly asymmetric splitting around the exciton energy (for example). By now, a substantial number of materials have been reported to show ultrastrong coupling. As both the organic and inorganic materials have drawbacks and advantages for the strong coupling regime, materials at the border between the two material classes have received increased attention recently. Carbon nanotubes and 2D materials like monolayered transition metal dichalcogenides and nanographene are solution processable and have large exciton-binding energies, similar to organic materials. However, they share their high structural order and stability with inorganic semiconductors. All these materials have been reported to show strong coupling (Liu et al., 2014; Graf et al., 2016; Coles et al., 2017a) and are potentially interesting for polariton devices because they have high charge carrier mobility in comparison to conventional organic semiconductors. Another inherently hybrid organic-inorganic material class are the perovskites, which also can enter the strong coupling regime, as demonstrated by Brehier et al. (2006). While the coupled exciton stems from the inorganic part of the perovskite-innate quantum well, some of its characteristics like a large excitonbinding energy and a relatively broad absorption (80 meV) are comparable to those of organic materials. A different group of materials studied in the strong coupling regime are biomolecules. Fluorescent proteins were chosen, for example, because their unique architecture leads to high PLQYs and little exciton quenching. The enhanced green fluorescent protein (eGFP), in which the fluorophore is protected by a barrel-like structure, supports very high excitation densities, thus enabling the only demonstration of two nonlinear thresholds in an organic material to date (the first corresponding to polariton lasing, the second to photon lasing) (Dietrich et al., 2016). A further step into biology

Strong light-matter interactions and exciton-polaritons in organic materials

291

was taken by Coles et al., who first reported chlorosomes (optical antennas), followed by bacteria containing the chlorosomes to be strongly coupled to cavity photons (Coles et al., 2014b, 2017b). In this biological system, the modified energy levels could be used to control the energy-transfer pathways within the organism (e.g., photosynthesis processes) noninvasively. From this discussion, it becomes evident that a vast number and variety of organic molecules exhibits strong coupling. The material parameters extracted from a number of reports are summarized in Table 9.1, though the large degree of freedom in preparing the cavities limits the comparability between samples. An important difference between experiments is the number of molecules or oscillators present in the cavity. This can be changed by changing the thickness of the organic film or by changing the concentration in a host. However, by changing the concentration of the active material, not only the number of absorbers is changed, but also the environment of each exciton. The proximity to other excitons affects possible decay channels, and thus the width of the transition and its efficiency. In general, excitons in organic materials have narrower, more intense transitions when being isolated (e.g., by being diluted in a matrix). It remains to be established which materials and sample structure will prove ideal for the practical use of strong coupling. Most likely, this will largely depend on the application: While lasing applications will focus on highly photostable materials in cavities with high Q-factors, applications relying on energy shifts will employ materials and cavity structures leading to the highest Rabi splitting.

9.3

Mixing different excitonic states via strong coupling

So far, we have discussed the coupling of cavity photons to a single excitonic mode. In this section, we consider how the cavity photon can hybridize two different excitonic modes by simultaneously coupling to both of them. The resulting quasiparticles are called hybrid polaritons and are a mixture of the photon and both excitons. They are typically defined by three anticrossing modes: the UP, MP, and LP. In order for the light to be able to hybridize two excitons, they both need to be in resonance with the cavity mode. Thus, the energy difference of the two excitons to be hybridized must be smaller than their combined coupling strengths. Even before organic polaritons had been realized, a theoretical paper by Agranovich et al. suggested using hybrid polaritons in mixed organic-inorganic systems (Agranovich et al., 1997). The authors argued that by coupling the states of weakly bound Wannier-Mott (inorganic material) and strongly bound Frenkel (organic material) excitons via a cavity photon, one could merge their properties in a beneficial way. More specifically, one could enhance the relaxation rates in the region kk  0 (which is crucial for polariton condensation) by efficient scattering between polariton branches. Such enhanced devices also could profit from combining the high oscillator strengths and large exciton-binding energies characteristic of organic materials with the high carrier mobilities and large nonlinearities typical for inorganic semiconductors. The realization of a hybrid Frenkel-Wannier-Mott microcavity proved challenging, however, as both the fabrication processes and the

292

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 9.3 First hybrid organic polaritons (Lidzey et al., 2000). (A) Absorbance of a control film of EX1 j Polystyrenej EX2 on glass, where EX1 and EX2 are two cyanine dyes. (B) Cavity structure with stacked films between a DBR and silver mirror. (C) Anticrossing modes appearing in angle-dependent and spectrally resolved reflectance. Reproduced with permission from Lidzey, D.G., 2000. Photon-mediated hybridization of frenkel excitons in organic semiconductor microcavities. Science 288, 1620–1623.

material properties have to be compatible. In addition, the hybridization of the two intrinsically different excitons requires high Q-cavities—the fabrication of which can potentially damage the organic material—in order to resolve the small Rabi splittings inherent to Wannier-Mott excitons. However, even without introducing Wannier-Mott excitons into the microcavities, there are various ways to explore the hybridization between two matter-states in organic-only systems. The first demonstration of hybrid organic-only polaritons was published not long after the first report on organic polaritons (Lidzey, 2000). Lidzey et al. prepared a microcavity in which two different J-aggregate forming cyanine dyes were separated by a polystyrene film (see Fig. 9.3). This transparent film served to block any direct energy transfer (like F€ orster transfer) that requires close vicinity of the materials. The excitons of the cyanine dyes were close in energy (only 60 meV apart), so that the energy difference was smaller than the Rabi splitting that previously had been observed in these systems. Indeed, the dispersion of the microcavity showed three coupled modes, which avoided a crossing of the uncoupled photon and exciton energies. A three-level, coupled oscillator model described the data well and yielded nearly equal mixing fractions of the photon and both excitons for one point along the MP. Further work on hybrid polaritons in organic-only systems demonstrated that a single material can undergo hybridization between different vibronic replicas of the same electronic state (Holmes and Forrest, 2004). Also, the quantized modes of molecular vibrations of two materials were shown to undergo coupling when combined in a resonant microcavity (Muallem et al., 2016). Although these quasiparticles are not exciton-polaritons, but rather vibron-polaritons emitting in the infrared (IR), they are of interest because materials hybridized in this way might behave differently in chemical reactions. Experimental investigations of hybrid organic-inorganic polaritons started in 2006. Since then, cavity photons have been shown to hybridize excitons in epitaxially grown, inorganic quantum wells with excitons in small molecules and a J-aggregate-forming

Strong light-matter interactions and exciton-polaritons in organic materials

293

material (Holmes et al., 2006; H€ ofner et al., 2015; Paschos et al., 2017). Perovskites also were used as organic counterparts in hybrid organic-inorganic polariton devices (Lanty et al., 2011). Other reports focused on all solution-processable cavity materials and demonstrated photon-mediated hybridization of a perovskite (now providing the inorganic exciton) or of ZnO nanoparticles with small organic molecules (Wenus et al., 2006; Slootsky et al., 2014). This list (which is not comprehensive) illustrates that organic and inorganic excitons with very diverse properties can be coupled, with some even showing persistence of polaritons up to room temperature (Wenus et al., 2006; Slootsky et al., 2014). In most studies, the hybridization of two excitons and a photon is characterized by the three polariton modes that have been described here. From fitting the three-level coupled oscillator model, the mixing fractions of the different branches are determined. These fractions indicate that a large contribution from all three constituents is present only for a narrow energy (and angle) range of the MP. A qualitatively different hybridization occurs when the energy difference of the coupled excitons is much smaller than their Rabi splitting: In this case, only two polariton branches appear, which show an increased total coupling strength (see also Fig. 9.4) (Slootsky et al., 2014). Those are described by a two-level oscillator model, in which the hybridization of the exciton species is now spread more uniformly over both the UP and LP branches. Although the energy-dependent contributions of the various exciton species are harder to disentangle in this scenario, the wider spectral region of their mixing may prove beneficial. While this two-branch hybridization was first observed in an organic-inorganic microcavity, the effect was also confirmed in a blend of two bio-materials (i.e., fluorescent proteins) (Dietrich et al., 2017). Many reports on hybrid polaritons give evidence of the presence of hybridization by providing dispersion relations obtained from reflectance measurements, but do not analyze the phenomenon further. In the following discussion, we will focus on a few selected publications that give further insight into the physics or applications arising from these threefold coupled quasiparticles. An interesting application of hybrid polaritons is the possibility to enhance the transfer of energy between polariton branches. This, however, cannot be investigated by simple reflectance measurements. Instead, the energy transfer is studied by measuring the PL of hybrid polaritons, as discussed in two publications from the Lidzey group (Coles et al., 2014a; Wenus et al., 2006). The first study, by Wenus et al., found that the dispersion of the PL from a microcavity containing a perovskite hybridized with an organic dye did not follow the polariton dispersion obtained from reflectance. The PL only coincided with a polariton branch observed in reflectance if the cavity photon was dominant and the fraction of the exciton with low PLQY (i.e., of the organic material) was low (Wenus et al., 2006). While these results could be interpreted as an energy transfer from the inorganic to the organic exciton via the middle polariton (MP), they were not conclusive. The second, more comprehensive investigation by Coles et al. studied the energy transfer between two J-aggregating materials that were coupled by a cavity photon. In addition to PL measurements, these experiments comprised the analysis of a blend of the

294

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 9.4 Hybrid ZnO-organic polaritons (Slootsky et al., 2014). Angle-resolved reflectance spectra of (A) a control ZnO-only cavity, split into UP and LP (LP is the unlabeled lower yellow line in the graph); (B) a control NTCDA-only cavity, split into UP and MP (LP is cut off as not relevant); and (C) a hybrid ZnO-NTCDA cavity, showing a larger split into UP and MP than either of the microcavities containing only one of the materials. Both UP and MP are assumed to be equally hybridized between ZnO and NTCDA. Symbols represent the minima of the measured reflectance (background), solid lines show the results of the coupled oscillator fit, and dashed lines represent the uncoupled exciton mode and cavity photon. Reproduced with permission from Slootsky, M., Liu, X., Menon, V.M., Forrest, S.R., 2014. Room temperature Frenkel-Wannier-Mott hybridization of degenerate excitons in a strongly coupled microcavity. Phys. Rev. Lett. 112, 076401.

uncoupled materials and the use of PL excitation spectroscopy3 (Coles et al., 2014a). The authors found that while the uncoupled, blended film showed only negligible energy transfer, energy transfer became significant once the different exciton species 3

PL excitation spectroscopy measures the output intensity as a function of excitation parameters (in this case, energy, and angle).

Strong light-matter interactions and exciton-polaritons in organic materials

295

were hybridized by the microcavity photon. Furthermore, they demonstrated that the MP had the highest scattering efficiency in the LP ground state. The scattering efficiency depended strongly on the relative fraction of the two exciton species, thus highlighting the role of the hybridization. While potential applications have always been a motivation for studying hybrid polaritons, the first hybrid polariton device was demonstrated only recently, by Paschos et al. (2017). Here, the microcavity consisted of an inorganic part, which had electrical contacts and was hybridized by the cavity photon to an organic film of a J-aggregate dye. The energy levels were designed such that the inorganic exciton was energetically close to the minimum of the LP, which was a hybrid between photon, Frenkel, and Wannier-Mott excitons. The inorganic excitons thus provided a reservoir to populate the LP branch. It was shown that under optical pumping, the hybrid polaritons undergo polariton condensation into the ground state of the LP branch, and that this effect relies on the hybrid character of the cavity. On the one hand, polariton lasing could not be observed with a purely inorganic polariton at this energy. On the other hand, the lasing depended on the population of the ground state by the inorganic exciton reservoir and inherited its large nonlinearities. Paschos et al. also achieved significant (although not macroscopic) ground-state occupation by electrical pumping. The research outlined in this section has demonstrated the complexity of the field of hybrid polaritons. However, we are just beginning to explore the rich physics and potential advantages arising from the hybridization and it will be exciting to see what further discoveries will follow from this.

9.4

Polariton condensation

As briefly discussed in the first section of this chapter, polaritons can undergo condensation into their ground state, the LP at kk ¼ 0. As bosons, the scattering rate toward a given state, such as the ground state, scales with 1 + N, where N is the average occupation of that state. At sufficient density (N > 1), the scattering rate into a given state thus becomes a stimulated process. In this regime, a macroscopic occupation of the ground state can form, which is often referred to as nonequilibrium Bose-Einstein condensate (BEC) or polariton condensate. This state differs from the conventional BEC because one of the fundamental conditions for a BEC, the thermal equilibrium of the condensing system, is not fulfilled. However, despite the polaritons having a finite lifetime, the condensate exhibits key signatures of a conventional BEC such as full (spatial and temporal) coherence. The coherence of the photons leaking from the condensate is the reason why this phenomenon is also called polariton lasing. Strictly speaking, most polariton lasers are not BECs (only one sample to date has shown a perfect Bose-Einstein distribution (Sun et al., 2017)), and some authors distinguish between a polariton laser and a polariton condensate (Byrnes et al., 2014). Mostly, however, the two terms are used interchangeably, as we will do here. Unlike in a conventional laser (referred to as a photon laser in the following), the onset of polariton lasing does not require a population inversion and hence can be observed at thresholds below those of conventional photon lasing. While polariton

296

Handbook of Organic Materials for Electronic and Photonic Devices

lasing was first observed at low temperatures and in inorganic materials, the stability of Frenkel excitons in organic materials enables the observation of polariton lasing at room temperature in these materials. For organic materials, polariton lasing was first shown in single crystals of anthracene by Kena-Cohen and Forrest (2010). Since then, polariton lasing was observed in several other organic materials—namely, in conjugated polymers (Scafirimuto et al., 2017; Plumhof et al., 2013), oligomers (Daskalakis et al., 2014, 2015; Bobrovska et al., 2017), fluorescent proteins (Dietrich et al., 2016), and dyes embedded into a matrix (Cookson et al., 2017). The main features of polariton lasers are summarized in Fig. 9.5 and will be discussed next.

Fig. 9.5 See figure legend on opposite page. (Continued)

Strong light-matter interactions and exciton-polaritons in organic materials

297

Fig. 9.5, Cont’d Signatures of polariton condensation in organic materials. (A) Angular reflectivity spectra of a microcavity containing TDAF that confirm strong light-matter coupling between the cavity mode (C) and the exciton (X) by the formation of UP and LP with a characteristic anticrossing behavior. (B) Angle and spectrally resolved PL from the LP below (left) and above (right) the polariton-lasing threshold in the same TDAF microcavity. In the polariton-lasing regime, the emission blue shifts compared to the LP, which is attributed to polariton-polariton interaction. However, the emission remains well separated from the cavity mode. (C) Real-space interferogram of the polariton condensate revealing macroscopic coherence. Here, BODIPY-Br was used as active material. (D) Polariton and conventional lasing threshold in a fluorescent protein-filled microcavity versus the energetic position of the LP and the cavity mode, respectively. The observation of a second threshold based on stimulated emission confirms the fundamentally different character of polariton lasing. In addition, photon-lasing thresholds are significantly above those of polariton lasing (here scaled by a factor of 10). (E) Reduction in the excited-state lifetime of polaritons in the LP for increasing polariton density, as shown in a cavity filled with crystalline anthracene. Above threshold, enhanced scattering results in a dramatically reduced polariton lifetime. (F) Superlinear increase of the output power above the polariton-lasing threshold in a microcavity containing MeLPPP. Stimulated scattering toward the polariton ground state enhances the emitted intensity. The blue shift of the condensate peak with increasing pump power is apparent in the inset, which shows luminescence spectra at three excitation densities. Reproduced with permission from (A,B) Daskalakis, K.S., Maier, S.A., Murray, R., KenaCohen, S., 2014. Nonlinear interactions in an organic polariton condensate. Nat. Mater. 13, 271–278. (C) Cookson, T., et al., 2017. A yellow polariton condensate in a dye filled microcavity. Adv. Opt. Mater. 5, 1700203; (D) Dietrich, C.P., et al., 2016. An exciton-polariton laser based on biologically produced fluorescent protein. Sci. Adv. 2, e1600666; (E) KenaCohen, S., Forrest, S.R., 2010. Room-temperature polariton lasing in an organic single-crystal microcavity. Nat. Photonics 4, 371–375; (F) Plumhof, J.D., St€ oferle, T., Mai, L., Scherf, U., Mahrt, R.F., 2013. Room-temperature Bose-Einstein condensation of cavity exciton-polaritons in a polymer. Nat. Mater. 13, 247–252.

Many of the characteristic features of polariton lasing were already observed in a pioneering study of crystalline anthracene (Kena-Cohen and Forrest, 2010). In this research, the organic material was nonresonantly pumped with a fs-laser. First, strong light-matter coupling was demonstrated by angle-resolved spectroscopy below and above the threshold. Second, a significantly enhanced population of the ground state, and thus relaxation toward this state, was found above the polariton-lasing threshold. Third, thresholdlike behavior and a strong nonlinear increase of the output power were observed. Fourth, the onset of polariton condensation was shown to be accompanied by a sharp decrease in lifetime and spectral narrowing. In anthracene, single-scattering events were found to be the dominant relaxation source; polariton-polariton or polariton-exciton interactions were weak. In 2013, the Mahrt group achieved polariton condensation in a conjugated polymer and studied the difference between that and conventional photon lasing in considerable detail (Plumhof et al., 2013). This work employed a cavity partially filled with a thin, spin-coated layer of MeLPPP (see Fig. 9.2). In particular, temperature-dependent measurements revealed a phonon-driven relaxation with thermalization of the

298

Handbook of Organic Materials for Electronic and Photonic Devices

polaritons at room temperature. The authors also observed a significant blue shift of the polariton condensate with respect to the LP, which was attributed to polaritonpolariton interaction. The observed relaxation and repulsive interaction are characteristic features of polaritons that were previously observed in inorganic semiconductors. Furthermore, the lasing threshold was found to increase with decreasing temperature. This was attributed to a lack of efficient relaxation pathways at low temperatures. Most important, however, long-range coherence of the polariton condensate was shown, which revealed a coherent phase of the macroscopic wavefunction that formed across the entire pump spot (100 μm). Very recently, the same group reported on polariton condensation in a 0D Gaussian-shaped defect cavity using the same polymer MeLPPP (Scafirimuto et al., 2017). The in-plane confinement was achieved in an open-cavity configuration with a movable DBR top mirror that was deposited onto a prepatterned substrate with Gaussian-shaped indentations. These results open the door to investigating the behavior of quantum fluids in organic materials in more complex potential landscapes. Nearly simultaneously with the MeLPPP work, nonlinear polariton interaction was demonstrated for polariton condensates formed in cavities filled with the oligomer TDAF (Daskalakis et al., 2014). These interactions again manifested in a powerdependent blue shift of the condensate. The authors present a simple model based on exciton saturation at localized sites to describe the observed effects. In a follow-up study, the same group investigated the emergence of coherence in these inherently disordered systems (Daskalakis et al., 2015). The disorder is highlighted by spot-to-spot fluctuations of the polariton condensate in real space. While a coherent polariton cloud was observed for small pump spots, dislocations and even separated condensates can form for larger pump spots. In the latter case, the mutual coherence is lost. A recent study on polariton condensates in TDAF revealed a dynamic instability of the condensates due to interaction with the exciton reservoir (Bobrovska et al., 2017). In single-shot experiments of the condensate, the authors showed excellent agreement of the experimental results with their simulations. Importantly, the instability did not simply arise from the disorder in organic materials, but it was found that the exciton reservoir played an important role in determining the condensation dynamics and instabilities. In 2016, Dietrich et al. realized polariton lasing in a biologically produced fluorescent protein, namely eGFP (Dietrich et al., 2016). The special molecular structure of eGFP—which has a fluorescent π-electron system at its core that is protected by a mostly inert protein shell—leads to low exciton-exciton annihilation rates. This behavior was found to be advantageous for realizing polariton condensation with nanosecond pump pulses (i.e., with pulses much longer than the polariton’s lifetime). The researchers used a simple lamination method to integrate eGFP into the microcavity, and then coupled a high-order cavity mode to the excited state of the protein. In addition to showing polariton lasing, the resulting devices also showed conventional photon lasing when excited at a substantially higher pumping rate—a phenomenon that had not been observed before in organic systems. In addition, the dependence of these thresholds on the detuning of the cavity was investigated.

Strong light-matter interactions and exciton-polaritons in organic materials

299

The polariton-lasing threshold was found to be below that of photon lasing in the entire investigated range. Another difference between polariton and photon lasing was observed in the spectral position; polariton lasing in eGFP was most efficient (i.e., maximum optical gain) at spectral positions that were blue-shifted from the wavelength of the lowest photon-lasing threshold. These results further highlight the distinction between polariton and photon lasing. Recently, Cookson et al. achieved polariton condensation from a fluorescent dye, bromine-substituted boron-dipyrromethene (BODIPY-Br), that was doped into an inert host material (polystyrene) (Cookson et al., 2017). The characteristic behavior of polariton condensates was studied using single-shot experiments. This work will likely open the door to an even wider range of host-guest systems as active materials for polariton lasers. A fundamentally new system was explored by Paschos et al., who investigated polariton condensation in a hybrid organic-inorganic microcavity by stacking an organic J-aggregating film on top of a set of inorganic GaAs quantum wells (Paschos et al., 2017). In addition to showing the common characteristics of polariton condensates, the authors demonstrated the crucial role of the inorganic material as an exciton reservoir for lowering the condensation threshold. These findings pave the way for applications that simultaneously exploit the advantages of organic and inorganic materials. In particular, the control over the lasing threshold via an auxiliary state might prove essential for optimizing future devices. Comparing the materials for which polariton condensation has been demonstrated so far, a few similarities emerge. Most have a relatively narrow line width and a high PLQY, even at high pumping densities. As suggested by theoretical work, in disordered organic materials with low intermolecular interactions, the majority of excitations form incoherent excitons in the reservoir (Agranovich et al., 2003). In this picture, a high PLQY (and thus a long exciton lifetime) might help to harvest excitons. In general, the ideal material for polariton lasing provides a fast and efficient relaxation process for polaritons and can sustain the creation of large ground-state occupancies. In addition to efficient pumping of the LP, long polariton lifetimes are desirable for the accumulation of large numbers of polaritons. Given that the excited-state lifetime of singlets in most organic materials is in the nanosecond range, polariton losses are dominated by the cavity-photon lifetime. Hence, all reported polariton lasers based on organic materials use all-dielectric DBR-based cavities, which typically reach cavity-photon lifetimes of hundreds of femtoseconds due to quality factors of several hundreds. Improving the polariton lifetime by further advancing cavity fabrication, therefore, is expected to lead to reduced condensation thresholds and to enable polariton lasing in a wider range of materials. Furthermore, many of the polariton lasers reported so far have used relatively large negative detuning of the cavity. Finding an optimum light-matter ratio at a given level of detuning might further reduce the threshold. New classes of solution-processable, low-dimensional materials might be helpful to overcome some of these limitations, as they combine some of the properties of organic and inorganic excitonic materials.

300

9.5

Handbook of Organic Materials for Electronic and Photonic Devices

Electrical pumping

As shown in the previous section, the thresholds for polariton lasing are typically below those for conventional photon lasing. Unlike photon lasing, which requires population inversion and sufficient optical gain, polariton lasing depends solely on the polariton density in the ground state of the LP. As a result, polariton lasing ultimately might be achievable in materials that lack high PL quantum efficiency but offer efficient relaxation paths. These characteristics spark hope that exciton-polaritons may enable electrically pumped lasing—a long-standing goal of the organic emitter community. At the same time, electrical injection of polaritons into organic devices represents a practical alternative to study fundamental physical phenomena at room temperature. In 2005, Tischler et al. reported on electrically pumped exciton-polaritons for the first time (Tischler et al., 2005). They employed an organic light-emitting diode (OLED) device architecture with two reflective silver (Ag) electrodes that simultaneously acted as the cavity mirrors. The matter component was introduced by a 6-nm-thin film, consisting of four bilayers of polyelectrolyte/TDBC J-aggregate with a sharp excitonic transition at an energy of 2.08 eV (wavelength, 595 nm). These bilayers were incorporated into a multilayer OLED stack, in which charge-transport layers of varying thickness were used to tune the cavity resonance. The oscillator strength of these layers led to a Rabi splitting of 265 meV in this strongly coupled microcavity. The metal-clad cavities had a quality factor of 10 and a current density of up to 0.1 A cm2 was achieved without affecting the Rabi splitting. Angle-resolved EL spectra revealed that the majority of the emission followed the LP. Additionally, a small amount of uncoupled excitonic emission and emission from the UP were observed. The uncoupled EL was attributed to disorder in the J-aggregate, leading to uncoupled excitons. In contrast to optical excitation, electrical pumping excites all emitters equally, including uncoupled emitters with transition dipole moments perpendicular to the surface. The UP emission was ascribed to emissions from a charge transport layer used in the device that was spectrally filtered through the UP branch. In the context of developing an electrically pumped polariton laser, the following challenges became apparent from this first work on electrical pumping: (1) polariton relaxation must be investigated and optimized; (2) the current density, and thus the polariton pump rate, must be increased; and (3) the polariton lifetime, and therefore the quality factor of the cavity, must be increased. In general, one might assume that electrical excitation is advantageous for the population and relaxation of polariton states due to polariton-electron scattering. Lodden et al. studied the polariton population mechanism under electrical excitation using a metal-clad microcavity-OLED structure similar to the structure described previously, but in this case, it had a 70-nm-thick layer of tetraphenylporphyrin (TPP) as the emissive material (Lodden and Holmes, 2011). These cavities showed strong coupling and polariton emission upon electrical excitation. In comparison to optical pumping, the authors observed no change of the population of the polariton branches or of the relaxation within the LP under electrical excitation.

Strong light-matter interactions and exciton-polaritons in organic materials

301

The Lidzey group later reported electrical pumping of exciton-polaritons in an OLED-type structure with an enhanced cavity quality factor of about 60 (Christogiannis et al., 2013). Rather than using a metal bottom mirror, they employed a DBR in combination with a transparent indium tin oxide (ITO) anode. The matter component of the polaritons was introduced by a thin layer of TDBC dye J-aggregates. In angle-resolved EL spectra, a significant polariton relaxation bottleneck was observed. The inefficient relaxation led to a strongly reduced external quantum efficiency (EQE) of the polariton device relative to a noncoupled control OLED, as a result of insufficient scattering to the polariton branches. Again, the population of the polariton branches was found to be identical for optical and electrical excitation of the exciton reservoir. More recent work focused on electrical pumping of ultrastrongly coupled excitonpolaritons in metal-clad cavity-OLEDs. The Kena-Cohen group achieved electrical pumping of polaritons in the ultrastrong coupling regime using TDAF (Fig. 9.6A and B; Gubbin et al., 2014). In these devices, an EQE of 0.1% was achieved. While this was drastically lower than the EQE achieved in current state-of-the-art OLEDs (>20%), it was a high value for electrically pumped organic polariton devices at the time. Using a squaraine dye in a similar cavity-OLED, Mazzeo et al. achieved electrical pumping of polaritons with efficient relaxation toward the ground state (Mazzeo et al., 2014). In later work, the same group reported optimized devices that showed EL in the ultrastrong coupling regime and achieved even higher coupling strengths (Gambino et al., 2014). This led to a nearly dispersionless emission (i.e., very little shift in wavelength with angle, with a narrow emission line width). Conventional cavity-OLEDs operating in the weak coupling regime show a very substantial and generally undesired color shift with angle. The dispersionless emission of strongly coupled devices thus may prove useful for applications, such as in information displays where emissions with both narrow line width and small wavelength shift are required to achieve high color purity and good viewing angle stability. A limiting factor in the efficiency of the devices reviewed here is the low charge carrier mobility in organic semiconductors. Already in 1997, it was proposed to overcome this challenge by pumping inorganic polaritons electrically, which then couple via the cavity photon to the organic excitons placed in the same cavity (Agranovich et al., 1997). This was only realized in 2017 by Paschos et al. in a hybrid LED in which the inorganic part was contacted below a bottom (semiconducting) DBR and directly above GaAs quantum wells, but below a J-aggregate film and a top (dielectric) DBR (Fig. 9.6C and D). This device proved to support optically pumped lasing and currents up to 5 mA, though electroluminescence from the bottom DBR reduced the number of polaritons forming in the quantum wells. By comparing the blue shift of the electrically pumped device to that observed under optical pumping, the authors deduced that the polariton density in their device was fourfold lower than what would be required to reach the lasing threshold. Most recently, we reported on electrically pumped polaritons in cavity-integrated light-emitting field-effect transistors (cavity-LEFETs) using monochiral singlewalled carbon nanotubes (SWCNTs) (Graf et al., 2017) and a low band-gap diketopyrrolopyrole copolymer (DPPT-BT) (Held et al., 2017) as emissive materials. Unlike

Fig. 9.6 Device architectures for electrical pumping and polariton electroluminescence: (A) Microcavity-OLED structure with active material TDAF. The mirrors of the cavity also serve as electrical contacts. (B) Ultrastrong coupling and electrically pumped emission from exciton-polaritons in the structure shown in (A). Symbols represent minima in reflectance spectra, lines are obtained from fits. (C) Hybrid LED structure in which only the inorganic part is contacted electrically, but the DBR-clad cavity contains inorganic GaAs quantum wells and a film of a J-aggregate. (D) Electroluminescence from the hybrid LED at 1.6 mA. (E) LEFET structure using SWCNTs as active material. The current flows between source and drain, while photons/polaritons propagate perpendicular to the current. (F) Electrical pumping of excitonpolaritons in a cavity-LEFET as shown in (E) at low (left) and high (right) pumping rates. Reproduced with permission from (A,B) Gubbin, C.R., Maier, S.A., Kena-Cohen, S., 2014. Low-voltage polariton electroluminescence from an ultrastrongly coupled organic light-emitting diode. Appl. Phys. Lett. 104, 233302; (C,D) Paschos, G.G., et al., 2017. Hybrid organic-inorganic polariton laser. Sci. Rep. 7, 11377; (E,F) Graf, A., et al., 2017. Electrical pumping and tuning of exciton-polaritons in carbon nanotube microcavities. Nat. Mater. 16, 911–917.

Strong light-matter interactions and exciton-polaritons in organic materials

303

the OLED structure, cavity-LEFETs rely on an in-plane current flow (Fig. 9.6E). Thus, the optical cavity can be adjusted largely independently from modifications regarding the electrical performance of the device. Upon electrical excitation, complete thermalization was observed for emission from the LP branch. Owing to the high charge carrier mobility of SWCNTs, extremely high pumping rates were achieved in these devices (current densities up to 18,600 A cm2 in the case of SWCNTs, Fig. 9.6F), but this did not result in polariton lasing to date. In the presented device geometry, the exciton-pumping rate was calculated to be about 1027 cm3 s1, which we estimate corresponds to an occupation of the ground state of 0.004. To increase the polariton density further, several points should be addressed. First, the polariton lifetime could be increased by using highly reflective, all-dielectric mirrors. Second, the overall EQE of the devices has to be enhanced. Finally, the scattering mechanism from the reservoir into the LP must be understood and optimized to increase the number of excitations that actually form polaritons. In order to systematically improve the performance of electrically pumped polariton devices, it is essential to be able to compare fundamental parameters of their polariton characteristics and the electrical properties of these devices. Due to the vast variety of materials and device geometries used for organic emitters, we want to emphasize the most important characteristics that should be included in reports on electrical pumping of polaritons. In addition to characterizing the polariton dispersions (UP and LP branches), the strong coupling under electrical pumping should be monitored. This is particularly important for high pumping rates, which might reduce the oscillator strength of the coupled excitonic transition. In terms of electrical performance, at least the current-voltage characteristics and the EQE versus current density should be given. Improved charge transport and emission properties, in combination with the efficient population of the LP branch in cavity-integrated electroluminescent organic devices, may finally lead to electrically pumped polariton lasers at room temperature, complementary to the currently existing inorganic polariton laser diodes (Schneider et al., 2013).

9.6

Summary and outlook

This chapter reviewed organic microcavity-polaritonics with regard to different materials and the use of the strong coupling regime to hybridize two distinct matter-states. Moreover, the condensation and electrical excitation of polaritons were discussed. In particular, the vast variety and diversity of organic materials that have been used for strong coupling experiments were presented. These range from classic organic materials like polymers or J-aggregates to less conventional types like biologically produced materials and extend to materials at the border between the organic/ inorganic classification, such as carbon nanotubes or 2D materials. Moreover, reports on the possibility of coupling different excitonic states via the cavity photon were presented.

304

Handbook of Organic Materials for Electronic and Photonic Devices

The versatility of this method was demonstrated by the diversity of different excitons that can be coupled in this manner. Polariton condensation in organic materials was reviewed. This included general key features of condensates, such as the interaction-induced blue shift and spatial coherence as well as polariton-specific peculiarities like the two-threshold behavior and the dynamical instabilities. Finally, we presented the state of the art in pumping polaritons electrically, which will likely be necessary for many practical applications of polaritonic devices. As mentioned previously, the use of organic materials in the strong coupling regime is particularly appealing for applications at and above room temperature. For instance, the color stability of displays at large angles and their color purity could be improved by using ultrastrongly coupled microcavities. Similarly, a new method for tuning the color of emissive devices could consist of changing their coupling strength (e.g., by changing the applied electric field). Another field where strong light-matter coupling could be useful is chemical landscape engineering. By shifting the energetic levels in molecules, chemical reactions can be influenced, as shown by a decreased reaction rate in a strongly coupled microcavity with respect to a weakly coupled control sample. Perhaps the most discussed candidate for applications, however, is the polariton laser. Its reduced threshold compared to a conventional photon laser raises particularly high hopes for demonstrating an electrically pumped, organic laser—a goal that has so far remained elusive. The realization of these potential devices and even the improvement of proof-ofconcept demonstrations require a deeper understanding of this relatively young field of research. Organic thin films are complex systems, including disorder and multiple radiative and nonradiative relaxation channels, and hence their theoretical description is inherently complicated. Although there has been significant progress on both theoretical and experimental organic polariton research over the last few years, many questions remain, such as to which extent the delocalization of the polaritonic wave packet can change and improve charge transport in organic films with intrinsically low charge carrier mobilities. Also, the necessary requirements for achieving polariton lasing in organic microcavities are not clear to date. The role of the width of the absorption band and of the Stokes shift is still debated, and the best way of maximizing the relaxation efficiency toward the bottom of the LP is unclear. Potential strategies for addressing this involve changing the light-matter ratio, changing the coupling strength, and maximizing the PLQY by doping to suppress competing processes. From these open questions, it is evident that a better understanding of the fundamental processes is necessary to allow us to select the best materials and to optimize devices for a given application. This shows that while the potential prospects of organic polaritonics are very exciting, there is still a great amount of challenging research ahead.

References Agranovich, V., Benisty, H., Weisbuch, C., 1997. Organic and inorganic quantum wells in a microcavity: Frenkel-Wannier-Mott excitons hybridization and energy transformation. Solid State Commun. 102, 631–636.

Strong light-matter interactions and exciton-polaritons in organic materials

305

Agranovich, V., Litinskaia, M., Lidzey, D., 2003. Cavity polaritons in microcavities containing disordered organic semiconductors. Phys. Rev. B 67, 085311. Amo, A., et al., 2011. Polariton superfluids reveal quantum hydrodynamic solitons. Science 332, 1167–1170. Anni, M., et al., 2001. Organic μ cavities based on thermally evaporated TeOx-LiF distributed Bragg reflectors. Appl. Phys. Lett. 79, 1381–1383. Baas, A., Karr, J., Eleuch, H., Giacobino, E., 2004. Optical bistability in semiconductor microcavities. Phys. Rev. A 69, 023809. Bobrovska, N., Matuszewski, M., Daskalakis, K.S., Maier, S.A., Kena-Cohen, S., 2017. Dynamical instability of a nonequilibrium exciton-polariton condensate. ACS Photonics 5, 111–118. Brehier, A., Parashkov, R., Lauret, J.S., Deleporte, E., 2006. Strong exciton-photon coupling in a microcavity containing layered perovskite semiconductors. Appl. Phys. Lett. 89, 171110. Byrnes, T., Kim, N.Y., Yamamoto, Y., 2014. Exciton–polariton condensates. Nat. Phys. 10, 803–813. Christogiannis, N., et al., 2013. Characterizing the electroluminescence emission from a strongly coupled organic semiconductor microcavity LED. Adv. Opt. Mater. 1, 503–509. Ciuti, C., Bastard, G., Carusotto, I., 2005. Quantum vacuum properties of the intersubband cavity polariton field. Phys. Rev. B 72, 115303. Coles, D.M., et al., 2014a. Polariton-mediated energy transfer between organic dyes in a strongly coupled optical microcavity. Nat. Mater. 13, 712–719. Coles, D.M., et al., 2014b. Strong coupling between chlorosomes of photosynthetic bacteria and a confined optical cavity mode. Nat. Commun. 5, 5561. Coles, D.M., et al., 2017a. Strong exciton–photon coupling in a nanographene filled microcavity. Nano Lett. 17, 5521–5525. Coles, D., et al., 2017b. A nanophotonic structure containing living photosynthetic bacteria. Small 13, 1701777. Cookson, T., et al., 2017. A yellow polariton condensate in a dye filled microcavity. Adv. Opt. Mater. 5, 1700203. Daskalakis, K.S., et al., 2013. All-dielectric GaN microcavity: strong coupling and lasing at room temperature. Appl. Phys. Lett. 102, 101113. Daskalakis, K.S., Maier, S.A., Murray, R., Kena-Cohen, S., 2014. Nonlinear interactions in an organic polariton condensate. Nat. Mater. 13, 271–278. Daskalakis, K.S.S., Maier, S.A.A., Kena-Cohen, S., 2015. Spatial coherence and stability in a disordered organic polariton condensate. Phys. Rev. Lett. 115, 035301. Dietrich, C.P., et al., 2016. An exciton-polariton laser based on biologically produced fluorescent protein. Sci. Adv. 2, e1600666. Dietrich, C.P., et al., 2017. Strong coupling in fully tunable microcavities filled with biologically produced fluorescent proteins. Adv. Opt. Mater. 5, 1600659. Gambino, S., et al., 2014. Exploring light-matter interaction phenomena under ultrastrong coupling regime. ACS Photonics 1, 1042–1048. George, J., et al., 2015. Ultra-strong coupling of molecular materials: spectroscopy and dynamics. Faraday Discuss. 178, 281–294. Graf, A., Tropf, L., Zakharko, Y., Zaumseil, J., Gather, M.C., 2016. Near-infrared excitonpolaritons in strongly coupled single-walled carbon nanotube microcavities. Nat. Commun. 7, 13078. Graf, A., et al., 2017. Electrical pumping and tuning of exciton-polaritons in carbon nanotube microcavities. Nat. Mater. 16, 911–917.

306

Handbook of Organic Materials for Electronic and Photonic Devices

Gubbin, C.R., Maier, S.A., Kena-Cohen, S., 2014. Low-voltage polariton electroluminescence from an ultrastrongly coupled organic light-emitting diode. Appl. Phys. Lett. 104, 233302. Hagenm€uller, D., Schachenmayer, J., Sch€utz, S., Genes, C., Pupillo, G., 2017. Cavity-enhanced transport of charge. Phys. Rev. Lett. 119, 223601. Held, M., et al., 2017. Ultrastrong coupling of electrically pumped near-infrared excitonpolaritons in high mobility polymers. Adv. Opt. Mater. 6, 1700962. Hobson, P.A., et al., 2002. Strong exciton–photon coupling in a low-Q all-metal mirror microcavity. Appl. Phys. Lett. 81, 3519. H€ ofner, M., Sadofev, S., Kobin, B., Hecht, S., Henneberger, F., 2015. Hybrid polaritons in a resonant inorganic/organic semiconductor microcavity. Appl. Phys. Lett. 107, 181109. Holmes, R.J., Forrest, S.R., 2004. Strong exciton-photon coupling and exciton hybridization in a thermally evaporated polycrystalline film of an organic small molecule. Phys. Rev. Lett. 93, 186404. Holmes, R.J., Kena-Cohen, S., Menon, V.M., Forrest, S.R., 2006. Strong coupling and hybridization of Frenkel and Wannier-Mott excitons in an organic-inorganic optical microcavity. Phys. Rev. B 74, 186404. Hutchison, J.A., Schwartz, T., Genet, C., Devaux, E., Ebbesen, T.W., 2012. Modifying chemical landscapes by coupling to vacuum fields. Angew. Chem. 124, 1624–1628. Kasprzak, J., et al., 2006. Bose–Einstein condensation of exciton polaritons. Nature 443, 409–414. Kena-Cohen, S., Forrest, S.R., 2010. Room-temperature polariton lasing in an organic singlecrystal microcavity. Nat. Photonics 4, 371–375. Kena-Cohen, S., Davanc¸o, M., Forrest, S., 2008. Strong exciton-photon coupling in an organic single crystal microcavity. Phys. Rev. Lett. 101, 116401. Kena-Cohen, S., Maier, S.A., Bradley, D.D.C., 2013. Ultrastrongly coupled exciton-polaritons in metal-clad organic semiconductor microcavities. Adv. Opt. Mater. 1, 827–833. Lagoudakis, K.G., et al., 2008. Quantized vortices in an exciton–polariton condensate. Nat. Phys. 4, 706–710. Lanty, G., et al., 2011. Hybrid cavity polaritons in a ZnO-perovskite microcavity. Phys. Rev. B 84, 195449. Lidzey, D.G., 2000. Photon-mediated hybridization of frenkel excitons in organic semiconductor microcavities. Science 288, 1620–1623. Lidzey, D.G., et al., 1998. Strong exciton-photon coupling in an organic semiconductor microcavity. Nature 395, 53–55. Lidzey, D.G., et al., 1999. Room temperature polariton emission from strongly coupled organic semiconductor microcavities. Phys. Rev. Lett. 82, 3316–3319. Liu, X., et al., 2014. Strong light–matter coupling in two-dimensional atomic crystals. Nat. Photonics 9, 30–34. Lodden, G.H., Holmes, R.J., 2011. Thermally activated population of microcavity polariton states under optical and electrical excitation. Phys. Rev. B 83, 075301. Martı´nez-Martı´nez, L.A., Du, M., Ribeiro, F., Kena-Cohen, S., Yuen-Zhou, J., 2018. Polaritonassisted singlet fission in acene aggregates. J. Phys. Chem. Lett. 9, 1951–1957. Mazzeo, M., et al., 2014. Ultrastrong light-matter coupling in electrically doped microcavity organic light emitting diodes. Appl. Phys. Lett. 104, 233303. Muallem, M., Palatnik, A., Nessim, G.D., Tischler, Y.R., 2016. Strong light-matter coupling and hybridization of molecular vibrations in a low-loss infrared microcavity. J. Phys. Chem. Lett. 7, 2002–2008. Oda, M., et al., 2009. Strong exciton-photon coupling and its polarization dependence in a metal-mirror microcavity with oriented PIC J-aggregates. Phys. Status Solidi C 6, 288–291.

Strong light-matter interactions and exciton-polaritons in organic materials

307

Orgiu, E., et al., 2015. Conductivity in organic semiconductors hybridized with the vacuum field. Nat. Mater. 14, 1123–1129. Paschos, G.G., et al., 2017. Hybrid organic-inorganic polariton laser. Sci. Rep. 7, 11377. Paterno`, G.M., et al., 2017. Synthesis of Dibenzo [hi,st] ovalene and its amplified spontaneous emission in a polystyrene matrix. Angew. Chem. Int. Ed. 56, 6753–6757. Plumhof, J.D., St€oferle, T., Mai, L., Scherf, U., Mahrt, R.F., 2013. Room-temperature BoseEinstein condensation of cavity exciton-polaritons in a polymer. Nat. Mater. 13, 247–252. Rempe, G., Walther, H., Klein, N., 1987. Observation of quantum collapse and revival in a oneatom maser. Phys. Rev. Lett. 58, 353–356. Sanvitto, D., Kena-Cohen, S., 2016. The road towards polaritonic devices. Nat. Mater. 15, 1061–1073. Scafirimuto, F., Urbonas, D., Scherf, U., Mahrt, R.F., St€ oferle, T., 2017. Room-temperature exciton-polariton condensation in a tunable zero-dimensional microcavity. ACS Photonics 5, 85–89. Schneider, C., et al., 2013. An electrically pumped polariton laser. Nature 497, 348–352. Schouwink, P., Berlepsch, H., D€ahne, L., Mahrt, R., 2002. Dependence of Rabi-splitting on the spatial position of the optically active layer in organic microcavities in the strong coupling regime. Chem. Phys. 285, 113–120. Schwartz, T., Hutchison, J.A., Genet, C., Ebbesen, T.W., 2011. Reversible switching of ultrastrong light-molecule coupling. Phys. Rev. Lett. 106, 196405. Sich, M., et al., 2011. Observation of bright polariton solitons in a semiconductor microcavity. Nat. Photonics 6, 50–55. Slootsky, M., Liu, X., Menon, V.M., Forrest, S.R., 2014. Room temperature Frenkel-WannierMott hybridization of degenerate excitons in a strongly coupled microcavity. Phys. Rev. Lett. 112, 076401. Song, J.-H., He, Y., Nurmikko, A.V., Tischler, J., Bulovic, V., 2004. Exciton-polariton dynamics in a transparent organic semiconductor microcavity. Phys. Rev. B 69, 235330. Sun, Y., et al., 2017. Bose-Einstein condensation of long-lifetime polaritons in thermal equilibrium. Phys. Rev. Lett. 118, 016602. Takada, N., Kamata, T., Bradley, D.D.C., 2003. Polariton emission from polysilane-based organic microcavities. Appl. Phys. Lett. 82, 1812. Tischler, J., Bradley, M., Bulovic, V., Song, J., Nurmikko, A., 2005. Strong coupling in a microcavity LED. Phys. Rev. Lett. 95, 036401. Virgili, T., et al., 2011. Ultrafast polariton relaxation dynamics in an organic semiconductor microcavity. Phys. Rev. B 83, 245309. Weisbuch, C., Nishioka, M., Ishikawa, A., Arakawa, Y., 1992. Observation of the coupled exciton-photon mode splitting in a semiconductor quantum microcavity. Phys. Rev. Lett. 69, 3314–3317. Wenus, J., et al., 2006. Hybrid organic-inorganic exciton-polaritons in a strongly coupled microcavity. Phys. Rev. B 74, 235212.

This page intentionally left blank

Advances in modeling the physics of disordered organic electronic devices

10

Michael C. Heiber*,a, Alexander Wagenpfahl†,a, Carsten Deibel† *Center for Hierarchical Materials Design (CHiMaD), Northwestern University, Evanston, IL, United States, †Institut f€ur Physik, Technische Universit€at Chemnitz, Chemnitz, Germany

10.1

Introduction

Organic electronic devices have attracted significant attention over the last several decades as a potential low-cost, lightweight, flexible, semitransparent, and customizable solution for a wide variety of applications not well suited to traditional inorganic technologies. The most researched device applications have been organic photovoltaics (OPVs), organic light-emitting diodes (OLEDs), and organic field-effect transistors (OFETs). Creating high-performance organic electronic devices has required a detailed understanding of the mechanisms and processes that make them work, as well as those that lead to performance losses, so that materials and device architectures can be appropriately tuned. Due to the complexity of the materials’ structures and the resulting charge-carrier and exciton generation, transport, and recombination mechanisms, elucidating the physics of these devices to create predictive physical models has been a challenging task. Modeling techniques can span a large window of length and timescales, from very detailed simulations of individual molecules all the way to simpler analytical equations for device performance metrics. When building a model for a complex system, it is always important to determine how much detail is needed to capture the dominant physical phenomena that dictate the performance variable of primary interest. This becomes especially important for computational models, where calculation time constraints are the primary limitation on model complexity. Among the variety of modeling techniques that have been used, one could broadly classify them into two categories, primarily based on their level of detail: (1) microscopic techniques and (2) macroscopic techniques. In microscopic techniques, the goal is to build a mechanistic model that simulates the individual optoelectronic mechanisms that together determine the overall device response and performance. In macroscopic techniques, analytical effective medium expressions typically are used to capture overall behavior at the device level.

a

Both authors contributed equally to this work.

Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00010-3 © 2019 Elsevier Ltd. All rights reserved.

310

Handbook of Organic Materials for Electronic and Photonic Devices

In this chapter, after a short discussion of the microscopic techniques, we will highlight the most popular one, kinetic Monte Carlo (KMC), describe its unique strengths, and explain how KMC simulations are constructed. We then briefly overview how KMC has been used to simulate the unique physics of charge carrier and exciton transport in organic semiconductors (OSCs), but we place the primary focus on a more detailed review that highlights how KMC simulations have been used to develop physical models for OPVs and OLEDs. Moving on to macroscopic simulations, we introduce the popular drift-diffusion technique and provide a brief tutorial on how the method is derived. We then highlight examples where drift-diffusion methods have been used to understand and optimize OPVs and OLEDs. In the end, we provide an outlook on the organic electronic device simulation and modeling field with recommendations for directions of future research.

10.2

Microscopic simulation and modeling

There are several microscopic modeling techniques where the goal is to simulate each of the individual optoelectronic mechanisms occurring in a device. These techniques are primarily used to understand the emergent device behavior and how experimentally observed phenomena can be traced back to the complex combination of fundamental mechanisms and materials structural features. Such techniques include primarily master equation and KMC methods, each of which can be performed at different length scales and with different levels of detail. In both cases, the individual mechanisms can be defined at a coarse-grained level, where the details of the individual molecules are disregarded; or at an ab initio level, where the positions of the molecules, their orientation, and the resulting electronic states and transitions are all derived from theory. While the use of multiscale ab initio KMC simulations is increasing, the more coarse-grained-lattice KMC implementation has been the most common technique for investigating many of the physical questions relevant for disordered organic electronic devices due to extreme computational cost of ab inito methods. In some cases, even lattice KMC is too computationally expensive, and researchers have preferred to use the much faster master equation methods. However, master equation methods are often inappropriate due to the challenges of correctly including particle-particle interactions that are particularly important for simulating electronic devices that contain charge carriers (Houili et al., 2006; Casalegno et al., 2013), and master equation methods can overestimate diffusion coefficients in disordered materials (Stehr et al., 2011). Here, we will focus our discussion mainly on lattice-KMC simulations, looking at (1) how KMC simulations are constructed, (2) how they are applied to simulate the unique transport physics in disordered OSCs, and (3) how they have been used in conjunction with experiments to elucidate the physics of OPVs and OLEDs.

Advances in modeling the physics of disordered organic electronic devices

311

10.2.1 Introduction to KMC simulations Monte Carlo methods are a general class of algorithms that use random numbers to solve a wide range of problems. The first and most popular Monte Carlo algorithms were developed to determine equilibrium properties of systems. However, in the 1960s, researchers developed a new algorithm to describe the dynamic behavior of nonequilibrium systems (Voter, 2007), now most commonly called KMC. KMC simulation models have been applied to many different types of problems, but in the late 1970s and early 1980s, the method was adapted to describe the dispersive charge and exciton transport observed in disordered OSCs, through which B€assler and colleagues developed the well-known Gaussian disorder model (GDM) (B€assler, 1993). Expanding on this early work, numerous groups around the world have used KMC methods to simulate a wide variety of mechanisms and physical phenomena present in OSC devices. The KMC algorithm is used to simulate a physical system as it transitions from one microstate to the next over time. The method assumes that the dynamic evolution of the system has the properties of a Markov chain, whereby each transition is a stochastic process in which only knowledge of the present state is needed to determine the transition to the next state, and that the transition time is much shorter than the time between transitions (Voter, 2007). For organic electronic devices, the system’s state is defined by the positions of all charge carriers and excitons within the active semiconductor layer(s). Transitions occur whenever an additional exciton or charge is created or destroyed, or when any of them move. To then construct a KMC simulation, the transition rate constant must be known for all relevant optoelectronic mechanisms in the material of interest, and missing mechanisms always will be a potential source of systematic error. For each mechanism, one must explicitly define an analytical rate equation derived from theory or empirical relationships. The majority of KMC simulation studies have used a cubic, 3D lattice to represent a portion of the OSC materials in the device, whereby all charge carriers or excitons reside on discrete lattice sites. Without access to detailed knowledge about the positions of individual molecules or without the need for such detail, many important physical phenomena that result from the unique molecular nature of OSCs can be cast onto a cubic lattice with a lattice size resolution of around 1 nm, which is roughly the size of a single molecule or a short segment of a polymer. While using a lattice is not required, implementing the KMC algorithm with a lattice is significantly easier from a software design perspective, and the regularity of the lattice reduces computational complexity. Nevertheless, when one would like to capture the detailed impact of molecular structure and packing, precise positions of each molecule and its neighbors are needed, which requires a more detailed, off-lattice implementation. While lattice models have been used to capture positional disorder in molecularly doped materials, off-lattice models can simulate positional disorder effects more accurately (Oelerich et al., 2017). In most cases, there will be a number of possible transitions (events) at any given time. To determine which event will occur and when, most studies use the Gillespie

312

Handbook of Organic Materials for Electronic and Photonic Devices

first reaction method because it is very computationally efficient for systems with a large number of independent events (Gillespie, 1976). For each possible event, the event wait time is calculated, t¼

 ln X , k

(10.1)

where X is a uniform random number between 0 and 1 and k is the rate constant of the particular event. The wait time signifies the amount of time that must elapse before the particular event occurs. During the simulation, the wait time for all independent events is calculated and stored in a queue. Then the event with the shortest wait time is executed, and the wait time of the executed event becomes the time step for the iteration. Following the execution of each event, the queue is updated by adding or removing any newly enabled or disabled events, respectively, and then the wait time of the remaining events is decremented by the time step. Despite over three decades of using KMC simulations in the OSC field, there are still no widely accepted standard KMC software tools. However, in recent years, various software packages have become available namely, VOTCA-CTP for multiscale charge-transport simulations (www.votca.org) by the Andrienko group, MorphCT for multiscale charge-transport simulations (bitbucket.org/cmelab/morphct) by the Jankowski group, Bumblebee for OLED simulations (simbeyond.com/bumblebee) by Simbeyond, hophop for off-lattice charge-transport simulations (github.com/ janoliver/hophop) by Oelerich, and Excimontec for a variety of organic electronic devices (github.com/MikeHeiber/Excimontec) by Heiber. Of these, VOTCA-CTP, MorphCT, hophop, and Excimontec are freely available, open-source software tools under active development that may be of interest to readers looking to get started in this field. Nevertheless, most of the existing literature has been produced by research groups that maintain private codebases of varying complexity, efficiency, and reliability.

10.2.2 Fundamental transport modeling in OSCs 10.2.2.1 Charge transport In contrast to highly ordered inorganic semiconductors, charge carriers in disordered OSCs are relatively localized and form small polarons. As a result, charge-carrier motion proceeds via a thermally activated hopping mechanism. In addition, disordered semiconductors often exhibit dispersive charge transport, whereby some charge carriers move through the material much slower than others and the rate of transport becomes time dependent (Scher and Montroll, 1975). In most cases, this dispersion has been shown to be caused by a broadened distribution of hopping site energies (diagonal disorder) due to the fluctuations in the molecular orientations, conformation, and the surrounding dipolar environment, but positional disorder (off-diagonal disorder) due to fluctuations in the distance between molecules also can provide another type of disorder that affects charge transport (B€assler, 1993).

Advances in modeling the physics of disordered organic electronic devices

313

To implement the broadened distribution of hopping site energies, the lattice sites can be randomly assigned values from the appropriate density of states (DOS) probability distribution. Most commonly, a Gaussian distribution is used to describe the energetic disorder, but in some cases, an exponential distribution or even more complex forms have been used (MacKenzie et al., 2012; Oelerich et al., 2012; Lange et al., 2013). In addition to the shape of the distribution, a number of studies have identified the presence of correlated energetic disorder in OSCs arising from statistical fluctuations in the molecular dipole orientation or molecular orientation correlations. As a result, there can be local clusters of high-energy or low-energy sites, and the presence and length scale of these site energy correlations can have a significant effect on charge transport (Novikov, 2003). Regardless of the DOS, two predominant models are used to describe the charge hopping transport mechanism: the Miller-Abrahams (Miller and Abrahams, 1960) and Marcus models (Marcus, 1993). In many cases, both models produce similar behavior. However, the details of these models, the various DOS models, and how they both affect the electric field, temperature, charge-carrier density, and time dependence of charge transport have been a heavily investigated and sometimes highly debated topic of fundamental solid-state physics research over the years, and several extensive review papers have been written on these subjects that are outside the scope of this chapter (Coehoorn and Bobbert, 2012; Baranovskii, 2014).

10.2.2.2 Exciton transport Due to the low dielectric constant of OSCs, excitons have a relatively large binding energy, are relatively localized, and are typically modeled as neutral quasiparticles. Two varieties of exciton states also can exist, as defined by their spin state, singlet or triplet. Singlet-exciton states are the primary product of optical excitation and typically emit a photon when relaxing back to the ground state. Conversely, when spinuncorrelated electrons and holes meet in an OSC, triplet-exciton formation usually prevails over singlet formation, and direct relaxation of triplet excitons to the ground state is a spin-forbidden transition that often results in significantly longer lifetimes and very low radiative efficiencies. In addition, exciton annihilation processes can occur whereby an exciton is quenched when encountering another exciton or a charge carrier. All together, there is a complex series of possible excitonic mechanisms, and for microscopic device simulations, it is important to make sure that all the relevant mechanisms are included for the device type and operating conditions of interest. While the details of exciton transport models have been somewhat less studied than charge-transport models, there are several good reviews on the physics of exciton transport and diffusion and the use of KMC simulation techniques in this area (Menke and Holmes, 2014; Bjorgaard and K€ ose, 2015; Mikhnenko et al., 2015). Like for charge carriers, KMC simulations of both singlet- and triplet-exciton transport have been shown to produce similar dispersive characteristics and can be similarly modeled using a broadened distribution of hopping site energies (Sch€onherr et al., 1980; Richert and B€assler, 1985). In this case, a Gaussian DOS is also most commonly used. As a result, following photoexcitation, the rate of exciton diffusion slows over

314

Handbook of Organic Materials for Electronic and Photonic Devices

time as the excitons hop to lower-energy sites in the tail of the DOS. Singlet-exciton transport is normally simulated using the F€ orster resonant energy transfer (FRET) model (F€ orster, 1948), and triplet-exciton hopping is usually simulated using the Dexter electron exchange model (Dexter, 1953). In general, singlet excitons can diffuse 5–20 nm from their origin before relaxing back to the ground state, but triplet excitons often can diffuse significantly farther due to their extended lifetimes (Mikhnenko et al., 2015).

10.2.3 Modeling organic photovoltaics In an OPV device, a complex series of optoelectronic mechanisms converts photon energy into electrical energy and dictates the final efficiency of the conversion process. A schematic illustration of these processes is presented in Fig. 10.1. First, light is absorbed by an active OSC layer to form singlet excitons. To overcome the excitonbinding energy and efficiently generate free charge carriers, a blend of OSCs that phase-separate to form a bulk heterojunction (BHJ) structure is used. In most cases, an electron-donating material (donor) and an electron-accepting material (acceptor) are utilized to create an energetic driving force for exciton dissociation into a

Fig. 10.1 Primary OPV current generation and loss mechanisms. (1) Photocurrent generation process showing an exciton (green) diffusing to the DA interface and dissociating into charge carriers that separate and are extracted at the proper electrodes, (2) exciton recombination, (3) geminate charge recombination where CT states are unable to separate, and (4) bimolecular charge recombination where free charge carriers meet and recombine.

Advances in modeling the physics of disordered organic electronic devices

315

charge-transfer (CT) state at a donor-acceptor (DA) interface. As a result, when singlet excitons are created, they must diffuse to the interface to create a photocurrent. In high-performance devices, the CT state then efficiently separates to form free charge carriers. The free charge carriers then must travel though the complex BHJ morphology network, with electrons restricted to acceptor domains and holes to donor domains, before each carrier is extracted at its respective electrode. In opposition, there are a number of loss processes that reduce the photocurrent. Any excitons that are unable to reach the interface during their lifetime will relax back to the ground state. Additionally, if CT states are unable to form free charge carriers, the charges can recombine through a geminate recombination process. However, even if free charge carriers are generated, electrons and holes originating from separate excitons can meet and recombine through a bimolecular charge recombination process. For a more complete description of the physics of OPV devices, the reader is referred to several more detailed review papers that summarize the development of this field over the years (Clarke and Durrant, 2010; Deibel and Dyakonov, 2010; Proctor et al., 2013; Heeger, 2014). Groves (2013a) also has provided a prior overview of how KMC simulations in particular have contributed to advances in our understanding of OPV device physics. Here, we provide an updated overview and alternate perspective on this field. In addition, while a number of KMC studies have developed methods for simulating full current-voltage curves and have used them to make predictions about optimal devices, here, we primarily focus on work that elucidates the physics of the more detailed fundamental processes.

10.2.3.1 Bulk heterojunction morphology models The nanoscale morphology present in BHJ OPVs has been repeatedly shown to have a significant impact on device performance. As a result, retaining nanoscale detail in KMC simulations is particularly critical for developing structure-propertyperformance relationships. To model the BHJ morphology, Peumans et al. introduced an Ising-based method (Peumans et al., 2003), and this concept was later simplified and applied to KMC simulations by Watkins et al. (2005) to create the first BHJ OPV KMC simulations. Since these pioneering studies, Ising-based morphologies have been widely used in KMC simulations to elucidate OPV device physics. While this model may not capture all morphological features in BHJ blends, qualitatively, it produces a nanoscale bicontinuous morphology that is a reasonable and computationally accessible starting model. The ability to efficiently create Ising-based BHJ morphologies for KMC simulations and perform detailed 3D structural analysis is available through the open-source Ising_OPV software package developed by Heiber and Dhinojwala (2014) and Heiber (2016). Cahn-Hilliard model morphologies often have been used by the Groves group as a more complex alternative (Lyons et al., 2011, 2012). Most studies have used these BHJ morphology models as simplified models for a bicontinuous microstructure with pure donor and pure acceptor phases, and then they determined how simple structural features such as the interfacial-area-to-volume ratio and domain size affect a variety of different device behaviors. However, some

316

Handbook of Organic Materials for Electronic and Photonic Devices

efforts have been made to understand the impact of more detailed morphological features, such as mixed domains (McNeill et al., 2007; Lyons et al., 2012; Gagorik et al., 2013; Heiber and Dhinojwala, 2013; Jones et al., 2014) and anisotropic morphologies (Heiber et al., 2017), but these more complex morphological features that have been observed experimentally have not yet been rigorously probed using microscopic simulations.

10.2.3.2 Exciton diffusion and dissociation dynamics To produce an electrical current, photogenerated excitons must reach a DA interface before returning to the ground state, and as discussed in Section 10.2.2, most OSCs have a short exciton diffusion length, which places limitations on the maximum domain size of the BHJ structure. If domains are too large, too many excitons created in the interior of the domains are unable to reach an interface for dissociation. A number of KMC simulations have been used to probe this phenomenon and provide predictions about the optimal domain size when including exciton diffusion limitations with other domain size-dependent processes (Watkins et al., 2005; Yang and Forrest, 2008; Lyons et al., 2012). However, a number of studies have indicated that the domain-size restrictions due to the short exciton diffusion length may not always be so strict (Banerji, 2013; Caruso and Troisi, 2012). If excitons are delocalized over a larger volume, charge transfer can occur even when the exciton is centered on a molecule that is not at the interface. Another side effect of delocalization is a significant change in the exciton dissociation dynamics. When excitons are localized, the majority of the excitons must take time to diffuse to an interface for dissociation. However, when delocalized, a significant fraction of the excitons can dissociate immediately after creation. Guo et al. first experimentally quantified this behavior in P3HT:PCBM blends and determined that up to 50% of the excitons undergo immediate dissociation, and the rest require diffusion prior to dissociation (Guo et al., 2010); based on this finding, they concluded that P3HT has an exciton delocalization radius of about 4–7 nm. Taking this idea even further, Kaake et al. (2012) proposed that ultrafast dissociation of highly delocalized (>30 nm) excitons is a dominant mechanism occurring in many BHJ blends. In support of this conclusion, Banerji has highlighted a number of other experimental studies that support similar conclusions (Banerji, 2013). Probing this phenomenon further, Heiber and Dhinojwala used KMC simulations to model Guo’s experimental data and demonstrated the interplay among domain size, domain purity, and prompt dissociation. They estimated that excitons are likely to be significantly less delocalized than previously concluded from simpler analysis techniques (Heiber and Dhinojwala, 2013). They found that a significant amount of prompt exciton dissociation can be expected if domains are impure or if there is significant interfacial mixing and that the observation of prompt dissociation cannot be safely attributed to exciton delocalization without taking into account the details of the BHJ morphology. Overall, it remains unclear how much exciton delocalization increases the maximum domainsize limitation in state-of-the-art OPVs.

Advances in modeling the physics of disordered organic electronic devices

317

10.2.3.3 Charge separation The efficient splitting of CT states into free charge carriers is a critical step in maximizing photocurrent generation. While a number of DA blends have demonstrated highly efficient charge separation, many others have not. Explaining how the donor and acceptor chemical structure, the resulting morphology at the interface and within the domains, and the energy levels of the relevant states each contribute to the resulting charge separation has been a hotly debated, fundamental physical question of major importance for the development of OPV materials. As a result, a number of recent reviews have been written that describe in greater detail the experimental and theoretical developments related to the charge-separation process (Dimitrov and Durrant, 2014; Gao and Ingan€as, 2014; Few et al., 2015). KMC simulations have played a key role in the development of new models and the understanding of the charge-separation process due to the ability to implement and then test new conceptual models. Immediately after exciton dissociation, traditionally the separation between the two charges in the CT state is assumed to be small enough that there should be an appreciable CT-state Coulomb binding energy, as depicted in Fig. 10.2A. In the traditional Onsager-Braun bound polaron-pair model, a thermally and/or field-activated separation process is required to overcome this binding energy and create free charges, and the overall separation yield depends on the CT state’s lifetime (Braun, 1984). When extending beyond the analytical Onsager-Braun model to investigate bound CT states at DA heterojunctions, KMC studies have shown that the overall chargeseparation yield and the electric-field dependence are both significantly affected by the presence of the heterojunction itself (Peumans and Forrest, 2004; Groves et al., 2008; Wojcik et al., 2010), and also the temperature (Offermans et al., 2005), electric field (Peumans and Forrest, 2004; Groves et al., 2008, 2010), mobility (Groves et al., 2008; Wojcik et al., 2010; Heiber and Dhinojwala, 2012), BHJ domain size (Marsh et al., 2007; Groves et al., 2008), and the energetic disorder of the materials (Offermans et al., 2005; Groves et al., 2010; Heiber and Dhinojwala, 2012). In addition, the separation yield remains critically dependent on the CT state’s lifetime (Offermans et al., 2005; Wojcik et al., 2010; Deibel et al., 2009a; Heiber and Dhinojwala, 2012). KMC simulations built on the bound-CT-state model largely predict relatively low charge-separation yields at device-relevant electric fields unless a very long CT-state lifetime is assumed (Offermans et al., 2005; Marsh et al., 2007; Heiber and Dhinojwala, 2012). In order to explain the highly efficient charge separation that occurs in many optimized BHJ blends, a number of theories have been proposed. Peumans and Forrest (2004) proposed and implemented a hot-CT-state dissociation model into KMC simulations, in which the initial charge-separation distance (thermalization radius) of the relaxed CT state simply correlates with the excess energy of exciton dissociation, without simulating the thermalization process itself. Building on this early work, the hot-CT-state model has evolved to its current form, where it is assumed that hot, delocalized CT states with a lower binding energy are formed, which makes it easier for charge carriers to separate, as depicted in

Handbook of Organic Materials for Electronic and Photonic Devices

Potential energy

318

Bound CT state model

(A)

Potential energy

Hot delocalized states

Hot CT state model

Potential energy

(B)

Charge delocalization model

Potential energy

(C)

(D)

Interfacial disordered region

Interfacial energy cascade model

Charge separation distance

Fig. 10.2 Leading charge-separation models. Schematic diagram of the energetic landscape in leading charge-separation models: (A) Bound-CT-state model, (B) hot-CT-state model, (C) charge-delocalization model, (D) interfacial energy cascade model.

Advances in modeling the physics of disordered organic electronic devices

319

Fig. 10.2B. Hot-CT states contain an electron and hole that are able to move much faster and diffuse away from each other. However, the benefit of the excess energy lasts only as long as the charge carriers retain the excess energy while occupying the higher vibrational and/or electronic states, typically lasting only around 0.5 ps (Jailaubekov et al., 2013). Heiber and Dhinojwala (2012) expanded the Peumans model by also simulating the thermalization process and including a parameterized excess energy thermalization rate and an enhanced mobility for hot charge carriers. However, they showed that a major mobility enhancement, an extremely slow excess energy relaxation rate, or both were needed to increase the charge-separation yield significantly. More recently, Jones et al. (2014) probed this problem further in model BHJ blends without simulating the thermalization process, and similarly concluded that an unexpectedly large thermalization radius is needed to reach high-separation yields. Nevertheless, the hot-CT-state model has been supported by a number of studies that find a correlation between the DA offset energy and the free charge-carrier generation yield (Dimitrov and Durrant, 2014) or show that charge separation through hot states is more efficient (Bakulin et al., 2012; Grancini et al., 2013; Dimitrov and Durrant, 2014). At the same time, in at least some blends, efficient charge separation has been observed from relaxed CT states (Vandewal et al., 2014) and in blends with a negligible energetic offset driving force (Menke et al., 2017). In such cases, even if excess energy relaxation is not needed to produce a high-charge separation yield, another related explanation for this phenomena proposed by many studies is that relaxation of the charge carriers into the DOS tail provides an additional longer-time-scale (100 ps) energetic relaxation process that promotes charge separation (Offermans et al., 2003, 2005; Groves et al., 2010; van Eersel et al., 2012; Howard et al., 2014). In this model, the charge-carrier mobility at short times is significantly enhanced until the charge carriers relax into the DOS tail. As a result, the separation yield depends not only on the width of the DOS, but also on the initial energetic position of the electrons and holes within their respective DOS distributions (Groves et al., 2010). Alternatively, Deibel et al. (2009a) proposed that the delocalization of relaxed hole polarons along a polymer chain could be a primary cause of efficient charge separation in polymer:fullerene blends and used KMC simulations to demonstrate a large increase in charge-separation yield when implementing delocalization. This idea has evolved to become the charge delocalization model, whereby following charge transfer, the electron and/or hole spreads out along a polymer chain or between adjacent molecules, causing the effective Coulomb attraction between the electron and hole to be greatly reduced, as depicted in Fig. 10.2C. The delocalization is assumed to be present even in the relaxed polaron states, and not a transient property associated with excess energy. Deibel et al. implemented hole delocalization by placing partial charges along a polymer-chain segment (Deibel et al., 2009a), but delocalization also can be implemented by treating the charge carriers as Gaussian spheres (Gagorik et al., 2015). In addition to charge delocalization, exciton delocalization has been proposed to have a significant impact. Guo et al. (2010) proposed that the magnitude of exciton delocalization controls the initial electron-hole separation distance following exciton

320

Handbook of Organic Materials for Electronic and Photonic Devices

dissociation. KMC simulations including exciton delocalization by Heiber and Dhinojwala (2012) have shown that the magnitude of the exciton delocalization radius can have a strong impact on the separation yield. However, like in the case of the hotCT-state model, a very large initial separation distance resulting from highly delocalized excitons is needed to reach the high-separation yields observed experimentally, and as discussed in the “Exciton diffusion and dissociation dynamics” subsection, there are still questions as to whether such highly delocalized excitons are present. Another prominent explanation is the interfacial energy cascade model. Experimental studies have shown that disorder in many OSCs increases the band gap due to a shift in the energy levels. Due to greater disorder near the DA heterojunction, McMahon et al. (2011) proposed that the energy of the transport states near the interface is higher than the domain interior, as depicted in Fig. 10.2D. In addition to disorder, complex electrostatic and induction effects can alter the energetic landscape near the interface and provide an energy gradient that favors charge separation (D’Avino et al., 2016). As a result, there can be an energetic driving force that promotes charge separation. Groves (2013b) has shown using KMC simulations with a simple interfacial energy cascade model that there is an overall increase in the charge-separation yield and a reduced electric-field dependence that depends on the magnitude of the cascade energy shift. However, he also demonstrated that if the interfacial region is too thick, the separation yield decreases again because the charge carriers are unable to reach the lower-energy regions before geminate recombination occurs. In addition, detailed morphological measurements have shown that the more disordered interfacial regions form diffuse/mixed interfaces (Collins et al., 2011). While the disordered mixed phase is still proposed to act as an energy-cascade interlayer so long as pure, more ordered domains are still present ( Jamieson et al., 2012), the complex interface can complicate the resulting behavior. Using KMC simulations, Lyons et al. (2012) demonstrated that charge-carrier percolation limitations in the mixed regions can increase the geminate recombination losses, and Burke and McGehee (2014) also showed a decreased separation yield when the interfacial region is mixed. In addition, increased disorder near the interface would likely be expected to reduce charge delocalization, thereby increasing the CT-state binding energy and reducing the charge-carrier mobility. Even with the interfacial energy cascade, Burke and McGehee (2014) showed that the separation yield is still strongly dependent on the mobility and the CT-state lifetime. In recent years an entropic driving force has been gaining traction to explain the charge-separation process. Early on, Clarke and Durrant (2010) proposed that entropic effects should be included in the conceptual explanation for the charge-separation process, and that a significant entropic driving force for charge separation exists at a DA interface. Instead of being simply dictated by the electrostatic potential as a function of the charge-carrier separation distance, they argued that the charge-separation process should be thought of in terms of the overall free energy of the system as the charges separate, which includes the gain in configurational entropy due to the large number of charge-separated states. Gregg provided further analysis of this concept by quantifying the entropy gain and total free-energy change as charges separate,

Advances in modeling the physics of disordered organic electronic devices

321

depending on the dimensionality of the system (Gregg, 2011), and later, KMC simulations confirmed the impact of dimensionality (Giazitzidis et al., 2014; Nan et al., 2016). However, further theoretical work also has shown how energetic disorder limits the entropy gain due to fewer accessible states (Savoie et al., 2014; Hood and Kassal, 2016), and the entropic driving force can be lowered by reducing the BHJ domain size (Vithanage et al., 2013) and by having mixed interfaces (Lyons et al., 2012). In addition, the majority of studies have used equilibrium arguments to describe a definitively nonequilibrium process, and Giazitzidis et al. (2014) have calculated the free-energy curve under nonequilibrium conditions and found significant increases in the free energy of charge-separated states. Overall, while the entropic driving force certainly can play a role, it should not be argued to be the dominant reason for the unexpectedly high-charge separation yield. As Baranovskii et al. (2012) point out, Onsager theory already inherently includes entropy effects. In addition, we emphasize here that the numerous early KMC studies showing high-geminate recombination when implementing the bound-polaron-pair model already include entropy effects.

10.2.3.4 Charge transport in bulk heterojunction films While many studies have investigated the details of charge transport in neat OSC films, less is known about how the complex BHJ microstructure affects charge transport in OPVs. However, KMC simulation studies using model BHJ morphologies have revealed significantly lower mobilities than in a neat material due to the morphology (Frost et al., 2006; Groves et al., 2009), and showed that the presence of bottlenecks and dead ends can lead to charge buildup and increased recombination (Donets et al., 2013). On the other hand, morphologies containing domains with enhanced vertical alignment and direct charge-transport pathways give better charge-transport efficiency and less recombination (Donets et al., 2015). Using a master equation approach, Koster (2010) also probed charge transport in model BHJ blends and showed that the electric-field dependence of the mobility changes from positive to negative when comparing neat and BHJ films at low charge-carrier densities. Finally, Koster (2010) explained how the opposing domains in a BHJ morphology represent barriers to transport and that, when a barrier forces charge carriers to hop perpendicular to, or even against the electric field to bypass the barrier, these hops are driven by diffusion, and an increased electric field does not assist in their motion. In fact, an increased electric field can even hinder long-range transport if hopping against the field direction is needed to bypass a barrier. Expanding on this work further, Heiber et al. (2017) investigated how the quality of the charge-transport pathways in various BHJ morphologies, quantified by the average tortuosity, affects the transport behavior relative to a neat material system as a function of energetic disorder and temperature. KMC time-of-flight transport simulations showed that as tortuosity increases, the overall magnitude of the mobility is greatly reduced, and there is a dramatic decrease in electric-field dependence, as shown in Fig. 10.3. Interestingly, negative field dependence was shown not to be an inherent property of BHJ blends, but instead, the field dependence was found to depend on a combination of the tortuosity, energetic disorder, and temperature.

Handbook of Organic Materials for Electronic and Photonic Devices

Normalized mobility, mm/m0

322

10–2

10–3

10–4

10–5

(A)

σ/kBT = 3 Neat, t = 1 BHJ (d = 8 nm), t = 1.05 BHJ (d = 8 nm), t = 1.10 BHJ (d = 8 nm), t = 1.15 BHJ (d = 8 nm), t = 1.24

0.4

0.6

(B)

0.8

1.0

1.2

(F/F0)1/2

Fig. 10.3 Charge transport through a tortuous BHJ morphology. (A) Schematic showing how charge carriers must navigate through a complex, tortuous pathways. (Reproduced with permission from Heiber, M.C., Kister, K., Baumann, A., Dyakonov, V., Deibel, C., Nguyen, T.Q., 2017. Impact of tortuosity on charge-carrier transport in organic bulk heterojunction blends. Phys. Rev. Appl. 8, 054043.) (B) The impact of tortuosity (τ) on the electric-field dependence of the charge-carrier mobility. (Data from Heiber, M.C., Kister, K., Baumann, A., Dyakonov, V., Deibel, C., Nguyen, T.Q., 2017. Impact of tortuosity on charge-carrier transport in organic bulk heterojunction blends. Phys. Rev. Appl. 8, 054043.)

In addition to morphology effects, KMC modeling of experimental data has demonstrated that the charge-carrier mobility decreases over time as charge carriers relax into the DOS tail, and in some blends, this process is slow enough that quasiequilibrium mobilities are poor descriptors of the charge extraction process in at least some OPVs (Melianas et al., 2014, 2015). Not only does this phenomenon affect the chargeseparation yield, as we discussed in the “Charge separation” subsection, but it also potentially complicates the operation of BHJ OPVs and experimental device characterization. While it has not been shown how detailed morphological features combine with the DOS to affect mobility-relaxation behavior, morphological traps have been proposed to slow mobility relaxation further (Melianas et al., 2014).

10.2.3.5 Bimolecular charge recombination In most well-performing OPVs with efficient charge separation, bimolecular charge recombination is the dominant loss pathway and must be minimized in order to realize more efficient devices. We refer the reader to more comprehensive review papers discussing bimolecular charge recombination in OPVs for further details (Proctor et al., 2013; Lakhwani et al., 2014; G€ ohler et al., 2018). Bimolecular recombination is defined as a second-order process, Rrec ¼ 

dn ¼ kbr np, dt

(10.2)

Advances in modeling the physics of disordered organic electronic devices

323

where kbr is the bimolecular recombination rate coefficient, n is the electron density, and p is the hole density. The starting model for bimolecular recombination in OPVs has been the Langevin model (Langevin, 1903), which assumes an encounter-limited reaction where the time it takes for an electron and hole to meet in the film is rate limiting. The result is that the recombination coefficient is primarily dictated by the mobility of the charge carriers. The Langevin model also assumes a spatially and energetically homogeneous and isotropic system, which is not strictly valid in BHJ OPVs. Typically, electrons are restricted to the acceptor phase and holes to the donor phase, and recombination can occur only at a DA interface. These spatial limitations on charge-carrier motion and recombination location are expected to alter the bimolecular recombination rate and cause deviations from the Langevin model. Perhaps not surprisingly, BHJ OPVs have frequently exhibited two major deviations. As recently summarized by Heiber et al. (2016), a literature survey shows first, that super-second-order recombination kinetics has been measured in many blend systems, and second, that the measured recombination rate sometimes is found to be up to several orders of magnitude lower than predicted by the Langevin model. As a result, assuming equal electron and hole densities, the recombination rate can be defined in greater detail as Rrec ¼ kbr ðnÞn2 ! Rrec ∝ n2 + δ ,

(10.3)

where the apparent higher recombination order is typically attributed to a carrier density-dependent recombination coefficient ðkbr ðnÞ∝nδ Þ. To quantify the so-called reduced recombination phenomenon, the magnitude of kbr is compared to the Langevin model, resulting in the characteristic reduction factor ðζ ¼ kbr =kL Þ. When highly reduced recombination was first observed, Pivrikas et al. (2005) proposed that a reduced recombination rate is an inherent property of BHJ blends due to the spatial segregation of the charge carriers, and a large number of subsequent studies have adopted this argument. However, in a number of other BHJ blends, recombination rates much closer to the Langevin model have been observed. Clarke et al. (2005) showed that even blends with very similar morphologies can have dramatically different bimolecular-recombination rates. While the measured effective reduction factor can be significantly smaller than the true value due to charge-carrier concentration gradients within the active layer (Deibel et al., 2009b), reduced recombination still has been observed in cases where there are not significant gradient effects. Understanding the various factors that ultimately determine kbr has been a critical fundamental endeavor. KMC simulations have shown that the domain size of model BHJ structures do not alter the recombination order, and when using a Gaussian DOS, second-order recombination kinetics are observed (Heiber et al., 2015; Coropceanu et al., 2017). However when using an exponential DOS, the recombination order is correlated with the characteristic energy of the distribution (Nelson, 2003; Coropceanu et al., 2017; G€ohler et al., 2018). In such a case, an exponential DOS leads to carrier density-dependent mobilities that then cause the carrier density dependence of the recombination

324

Handbook of Organic Materials for Electronic and Photonic Devices

coefficient. In addition, questions remain how to explain observations of blends where transport behavior appears to indicate a Gaussian DOS and recombination behavior indicates an exponential DOS (Gorenflot et al., 2014). As recently summarized by Street (2015), one leading explanation is that a small density of exponential trap states forms at the DA interface and recombination through these trap states accounts for the higher-order kinetics. In such a case, less traps may be present in the interior of the domains, and charge transport may be dictated by hopping within the primarily Gaussian DOS present in the interior of the domains (Gorenflot et al., 2014). Another idea is that regardless of the DOS shape, the charge-carrier mobilities can relax over time as the charge carriers relax into the DOS tail, as discussed in the OPV charge-transport subsection. Under these nonquasiequilibrium conditions, the charge-carrier mobility is primarily time-dependent. In such cases, the recombination coefficient can be timedependent as well, which causes super-second-order recombination kinetics that does not directly correlate with the DOS shape (Mozer et al., 2005; Clarke et al., 2012; Kurpiers and Neher, 2016). An important first step toward a more complete understanding of reduced recombination has been the testing of the fundamental effect of phase separation on kbr under encounter-limited conditions. In early theoretical work, Koster et al. (2006) developed the minimum mobility model, arguing that the bimolecular-recombination rate in a BHJ blend should be limited by the mobility of the slowest carrier. Groves and Greenham (2008) then used KMC simulations to show that the recombination rate lies somewhere between the Langevin model and the minimum mobility model with a weak dependence on domain size. However, other KMC simulations (Hamilton et al., 2010) have indicated that the domain size could have a much larger impact. Expanding on this work, Heiber et al. (2015) investigated the combined impact of the domain size and the mobilities on the recombination rate coefficient and the reduction factor, as shown in Fig. 10.4. Heiber et al. (2016) then further investigated the charge-carrier density dependence to develop the power mean model. The power mean model predicts a continuous deviation from the Langevin model as the domain-size increases that begins to approach the minimum mobility model when domains are very large. Most importantly, this work, as well as previous results by Groves and Greenham (2008) show that phase separation itself reduces the recombination coefficient by only about a factor of 2 when using domain sizes that are typical of optimized OPV devices and equal mobilities, and that highly reduced recombination is not an inherent property of BHJ blends (Heiber et al., 2015, 2016). Instead, the leading explanation for highly reduced recombination is that the recombination rate is not encounter-limited; rather, it is highly dependent on the lifetime of the CT state and the CT redissociation rate (Hilczer and Tachiya, 2010; Ferguson et al., 2011; Burke et al., 2015; Heiber et al., 2016; Coropceanu et al., 2017). Given the importance of the CT state’s lifetime and dissociation rate in the initial charge-separation process, correlations between efficient charge-separation behavior and reduced bimolecular recombination would seem likely, but the connections between the two phenomena have not been deeply investigated. Future work making connections between the two may help refine the range of competing models for each process and arrive at a more broadly applicable unifying model.

(A)

10 –7

10 –8

d = 25 nm d = 35 nm d = 45 nm d = 55 nm

10 –9

10 –10

0.1

me/m h = 1 me/m h = 10 me/m h = 100

kL (d = 5) kmin (d = 55) 10 –11

325

1 d = 5 nm d = 10 nm d = 15 nm d = 20 nm

Reduction factor (ζ)

Recombination coefficient, kbr (cm3s–1)

Advances in modeling the physics of disordered organic electronic devices

10–3 1/2

(m em h)

10–2 2

–1 –1

(cm V s )

0.01

(B)

0

10

20

30

40

50

60

Domain size (nm)

Fig. 10.4 Encounter-limited bimolecular recombination in a BHJ. The impact of domain size and charge-carrier mobilities on the bimolecular recombination coefficient (A) and reduction factor (B). (Data from Heiber, M.C., Baumbach, C., Dyakonov, V., Deibel, C., 2015. Encounter-limited charge-carrier recombination in phase-separated organic semiconductor blends. Phys. Rev. Lett. 114, 136602.)

10.2.4 Modeling organic light-emitting diodes In a simple sense, OLEDs operate opposite to OPVs. In an OLED, charge carriers are injected at the opposing electrodes and ideally meet within the active emissive layer to form an exciton that then decays radiatively. As a result, understanding how charge carriers are injected into the OSC layers, transported through the material, and then eventually meet is important. And understanding exciton formation, diffusion, and radiative properties is needed as well. In addition, any nonradiative recombination mechanisms will lower the quantum efficiency of the device and must be minimized. All together, controlling each of these processes is critical for designing highefficiency OLEDs. We refer readers to more comprehensive reviews for details about the operation and design of OLEDs (Murawski et al., 2013; Reineke et al., 2013), and Kordt et al. (2015) have published an informative review on OLED modeling. Here, we highlight specific examples where KMC simulations have played a key role in understanding the fundamental device physics and processes in OLEDs. A variety of early KMC studies were done to model the injection of charges from a metal into a semiconductor layer using a hopping model, showing changes to the electric-field and temperature dependence of current injection due to charge hopping from the Fermi level of the metal into the tail of the DOS (Gartstein et al., 1996; Wolf et al., 1999). As a result of this phenomenon, a number of KMC and master equation studies have predicted the formation of current-density filaments in neat, disordered films (Yu et al., 2001; Tutisˇ et al., 2004; van der Holst et al., 2011). In a particularly thorough study, van der Holst et al. (2011) used KMC simulations of single-carrier diodes to show how the filaments change due to the injection barrier height (Δ)

326

Handbook of Organic Materials for Electronic and Photonic Devices Uncorrelated disorder s/kBT = 3, D = 0 eV

(A)

(C)

Correlated disorder s/kBT = 3, D = 0 eV

(B) Correlated disorder s/kBT = 6, D = 1 eV

Uncorrelated disorder s/kBT = 6, D = 1 eV

(D)

Fig. 10.5 Filamentary charge transport in single-carrier diodes. Visualization of current-density hot spots from KMC simulations of charge transport in single-carrier diodes comparing cases of low energetic disorder and no injection barrier with an uncorrelated (A) and a correlated (B) Gaussian DOS and a high-injection barrier and high-energetic disorder with an uncorrelated (C) and a correlated (D) Gaussian DOS. (Adapted with permission from van der Holst, J.J.M., van Oost, F.W.A., Coehoorn, R., Bobbert, P.A., 2011. Monte Carlo study of charge transport in organic sandwich-type singlecarrier devices: effects of Coulomb interactions. Phys. Rev. B 83, 085206.)

and the energetic disorder (σ) with both an uncorrelated and correlated Gaussian DOS, as shown in Fig. 10.5. One of the primary drivers of this phenomenon is the chargeinjection hot spots due to fluctuations in the injection barrier resulting from the energetic disorder (Tutisˇ et al., 2004; van der Holst et al., 2011). As shown in Fig. 10.5A and B, with no injection barrier and low energetic disorder, injection hot spots are weak and the filaments largely disperse as they move away from the bottom injecting surface (van der Holst et al., 2011). However, the filaments can be enhanced further by the presence of correlated energetic disorder, as depicted in Fig. 10.5C and D (Yu et al., 2001; van der Holst et al., 2011). In general, the effect of the currentdensity filaments is particularly enhanced when the films are thin. Given the possible presence of significant current-density hot spots in the active layer, the primary question for OLEDs is how this affects the exciton-formation process. Most likely, injection hot spots for electrons and holes will not be correlated, so it may be more difficult for electrons and holes to encounter each other; and major questions include how large this effect is and how it can be minimized to create high-radiative-efficiency devices. In simple cases, recombination is expected to be described by the Langevin model, but KMC simulation studies have identified conditions where deviations from the Langevin model occur, due to transport anisotropy (Ries and B€assler, 1984; Gartstein et al., 1996; Groves and Greenham, 2008), energetic disorder (Richert et al., 1989; Gartstein et al., 1996), and the electric field (Albrecht and B€assler, 1995; Gartstein et al., 1996). While van der Holst et al. concluded that the Langevin model still works well in an isotropic system with the an uncorrelated Gaussian DOS at low electric fields so long as accurate mobility values

Advances in modeling the physics of disordered organic electronic devices

327

are used (van der Holst et al., 2009), in cases where significant current-density filaments are expected, exciton formation and the resulting emission profiles can be highly inhomogeneous (Mesta et al., 2013; Shen and Giebink, 2015). Mesta et al. (2013) have shown that capturing these details is particularly important for accurately modeling and designing multilayer white OLEDs. In many of the more efficient phosphorescent OLEDs, in which photon emission comes from triplet excitons, so-called efficiency roll-off occurs where the quantum efficiency and luminous efficacy of the device decreases at higher current densities (Murawski et al., 2013). Due to the generally longer lifetime of triplet excitons, the nonradiative recombination mechanisms of triplet-triplet annihilation and tripletpolaron quenching can become significant due to the higher steady-state densities of triplet excitons and charge carriers in the device when operating at higher current densities or in current-density filament regions. To reduce this loss process and improve OLED performance, understanding the origins of efficiency roll-off is critical. A number of KMC studies have studied these phenomena to construct predictive models that relate electronic properties to expected losses so that device design can be optimized. In a particularly illuminating study, van Eersel et al. (2014) implemented a multilayer KMC model to fit experimental electrical and optical measurements and study the competing loss mechanisms and demonstrated the dominance of the triplet-polaron-quenching mechanism in two common host-guest OLED blends. Coehoorn et al. (2015) then showed that the triplet-polaron-quenching mechanism also can be linked to emitter molecule degradation. In these cases, triplet-polaronquenching causes degradation of OLED devices, and there is a direct relationship between roll-off behavior and device stability. Shen and Giebink (2015) further showed the molecular-degradation hot spots that can form due to the filamentary charge transport and the locally enhanced triplet-polaron quenching that occurs there. As new approaches are implemented to mitigate the undesirable roll-off and degradation processes, KMC simulation and modeling are sure to continue to be vital to understanding the complex kinetics and the impact of materials’ structural features.

10.3

Macroscopic simulation

When wanting to simulate full device behavior, especially at high charge-carrier densities, KMC simulation techniques can become prohibitively expensive, computationally speaking. Thus, alternative techniques are required. A major increase in calculation speed is obtained if the system is no longer calculated stepwise for each exciton or charge carrier, but rather by an analytic system of equations that solves the problem for all of them simultaneously. The drift-diffusion system of equations used for macroscopic simulations also offers the unique possibility to apply analytic theories without the requirement of numerical assumptions, such as the cubic lattice often used in KMC simulations. Here, we present a brief overview of how state-of-theart drift-diffusion simulations for application in OSC devices are built, and then highlight the important developments in the fields of OPV and OLED research using drift-diffusion simulations over the last decade.

328

Handbook of Organic Materials for Electronic and Photonic Devices

10.3.1 Introduction to drift-diffusion simulations The drift-diffusion system of equations is based on the Boltzmann-transport equation, which generally describes particle transport due to particle-density gradients and applied driving forces. For charge transport in semiconductor devices, this equation is often reduced, but not limited to, drift and diffusion in three dimensions, as follows: Jn ðxÞ ¼ qnðxÞμn FðxÞ qDn rnðxÞ |fflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflffl} + |fflfflfflfflfflfflffl{zfflfflfflfflfflfflffl} , drift

(10.4)

diffusion

which is often denoted as the drift-diffusion equation. In this example for electrons, the electron current density (Jn) is determined by the elementary charge (q), the local electron density (n), the electron mobility (μn), the local electric field (F), the electron diffusion coefficient (Dn), and the local electron density gradient (rn). Most of these (bold) parameters can have a dependency on the spatial coordinate x. In the 1960s, efficient methods to discretize the drift-diffusion equation into a series of finite elements were developed. One of the main issues was that small changes in the electric field can lead to an exponential change of the charge-carrier density. This implies that the charge-carrier density does not scale linearly between two elements and that intermediate results might exceed the numerical range, which leads to severe numerical errors during the calculation. Scharfetter and Gummel developed a mathematical discretization scheme still used today for most drift-diffusion simulations. In its original and commonly used form, it requires the validity of the Einstein relation and ignores spatial variations in the charge-carrier mobility (Selberherr, 1984). The Einstein relation was demonstrated to be valid in OSCs, at least with nondispersive transport, where charge carriers are thermalized in the DOS (Wetzelaer et al., 2011). The charge-carrier mobilities can strongly depend on the charge-carrier density due to the energetic and spatial disorder that causes the broadened density of localized states. This implies that the current flow due to positional variations of the charge-carrier mobility must be smaller than drift or diffusive charge transport to produce valid simulation results. To simulate a real device, an analogous equation to Eq. (10.4) must be defined for holes as well as for all other particles of interest, such as excitons and CT states. Then, the temporal dependence of each particle type is determined using the continuity equation, X X ∂n 1 ¼ rJn ðxÞ  Ri ðxÞ + Gj ðx,tÞ, ∂t q i j

(10.5)

shown here for electrons. The continuity equation defines the current transport to and from each discretized spatial element, as well as the generation rate and recombination rate within each element. For each particle type, the generation rate is a sum of all possible generation mechanisms, and the recombination rate is a sum of all possible depletion mechanisms into other particles or the ground state. These rates interconnect

Advances in modeling the physics of disordered organic electronic devices

329

all particle densities and result in the final population ratio between all considered particle densities by following detailed balance principles. Next, the electric field generated by the spatial distribution of charge carriers is calculated using the Poisson equation, rFðxÞ ¼ 

q ðn  p + CÞ, E0 Er

(10.6)

where Er is the relative dielectric constant, E0 is the vacuum permittivity constant, and C is the doping concentration. Spatial variations of Er are easily included in order to implement a stack of semiconducting layers. Finally, the system of equations created by Eqs. (10.4)–(10.6) is solved numerically. A variety of efficient approaches have been developed during the last decades. Which system is used mainly depends on the desired applied-voltage regime, the convergence accuracy of the selected algorithm, and the convergence speed. They all include a transformation of the physical variables into a numeric number space, a simultaneous or subsequent solution of all considered equations in an iterative loop, and then a retransformation into physical parameters. The discretization of the spatial dimensions is often disregarded in the implementation of drift-diffusion simulations. Mathematical requirements exist for the maximum distance in space (and in time) between two calculation steps to achieve valid simulation results (Selberherr, 1984; Mock, 1983). As an alternative to implementing and solving the drift-diffusion system of equations oneself, many open-source and commercial software packages are available to facilitate a first contact with drift-diffusion simulations. The most prominent examples are PC1D (sourceforge.net/projects/pc1d/), ASA (www.tudelft.nl), Setfos (www.fluxim.com/setfos/), Sentaurus (www.synopsys.com), and gpvdm (www. gpvdm.com).

10.3.2 Parameterization of disorder effects Drift-diffusion simulations require an analytic model for describing the drift and diffusion transport processes of each particle. This is particularly critical for charge-carrier transport, where a microscopic hopping model must be converted to a macroscopic analytic form. For disordered OSCs, the two most common models are the multiple-trapping-and-release (MTR) model and various empirical forms of the GDM.

10.3.2.1 Multiple-trapping-and-release model The MTR model considers the transport of an ensemble of localized charge carriers in a disorder broadened DOS (Shklovskii and Efros, 1984). The distribution is mostly assumed to be Gaussian or exponential, but one is not limited to one of these two examples. For hopping transport, charge carriers can either hop to an energetically lower state or they can be thermally excited to a state with higher energy.

330

Handbook of Organic Materials for Electronic and Photonic Devices

The restrictions are the limited thermal energy and availability of vacant states within the range of the hopping mechanism. The MTR model assumes that charge carriers at or above the so-called transport energy (Etr) can be interpreted as mobile, comparable to band transport, but charge carriers below Etr are interpreted as energetically trapped until they are again released to the transport energy. There are several ways of calculating the transport energy based on different hopping models (Oelerich et al., 2014), but the MTR model effectively discretizes the total population of charge carriers into two distinct groups. As a result, charge-carrier mobility always has to be considered as either the mobility of solely conductive charges (nc), or more commonly, as the lower average mobility of all charge carriers (n ¼ nc + nt). In the latter case, the effective mobility (μ) can be parameterized by the fraction (θ) of free charge carriers to the total charge-carrier density, and the corresponding mobility of the free charge carriers (μc), μ ¼ μc  θ ¼ μc

nc : nt + nc

(10.7)

The density of each subpopulation is calculated by integrating over all occupied states relative to the transport energy, nt ¼ nc ¼

ð Etr

DOSðEÞ  fFD ðE,EF , TÞdE

(10.8)

DOSðEÞ  fFD ðE, EF ,TÞdE

(10.9)

∞

ð∞ Etr

with the Fermi-Dirac function fFD(E, EF, T), where E is the state energy, EF is the Fermi energy, and T is the temperature. Whether the DOS is better described as exponential distribution, Gaussian distribution, or even a more complex function is still under consideration (Baranovskii, 2014). A Gaussian DOS distribution shows a fairly constant ratio of nc/nt for most relevant charge-carrier density regimes for OPV and OLED applications. An exponential DOS results in an exponentially decreasing share of conductive charges with decreasing total charge-carrier density (Mehraeen et al., 2013). The different behavior originates from the amount of available states above the Fermi energy, as shown in Fig. 10.6. A larger amount of available hopping sites allows for higher charge-carrier mobilities in Gaussian DOS distributions. Chargecarrier density-dependent mobilities with any distribution of the DOS can be directly calculated with drift-diffusion simulations (Baranovskii, 2014).

10.3.2.2 Gaussian disorder models Another approach to modeling the charge-carrier mobility is to fit results of KMC transport simulations using empirical analytic expressions. Such an approach was successfully introduced by B€assler (1993) to parameterize the mobility in a Gaussian DOS, known as the GDM. The model captured how the mobility depends on the

Advances in modeling the physics of disordered organic electronic devices

e

Transport energy

e

DOS

331

Transport energy

DOS

s2

kT

eF

eF Carrier density Carrier density

Fig. 10.6 Influence of the DOS on mobility. Quasiequilibrium density of occupied states and transport energies of an exponentially distributed density of states (left) and Gaussiandistributed density of states (right). (Reproduced with permission from Baranovskii, S.D., 2014. Theoretical description of charge transport in disordered organic semiconductors. Phys. Status Solidi B 251, 487–525.)

energetic disorder (σ) and temperature (T) in the low charge-carrier density regime, and includes a Poole-Frenkel dependence on the electric field (F), 

2σ μ∝ μ0 exp  3kB T

2 !

pffiffiffi  exp Aðσ, TÞ F :

(10.10)

Additional fit parameters are summarized by A. Due to shortcomings with the original GDM, several extensions have been derived to more accurately capture additional physical phenomena. The correlated disorder model (CDM) introduces spatial correlations between site energies due to randomly oriented permanent dipoles, which extends the Poole-Frenkel electric-field dependence of the mobility to low field strengths (Novikov et al., 1998). Later, the extended GDM was developed to capture the charge-carrier density dependence that appears at higher carrier densities, but again without considering energetic correlation. Eventually, both extensions were merged to form the extended correlated disorder model (ECDM) (Bouhassoune et al., 2009). However, these models were all derived from lattice KMC simulations, and recent work has shown how the electric-field dependence of the mobility changes when including positional disorder in off-lattice KMC simulations and has argued for the use of an effective temperature model in drift-diffusion simulations (Oelerich et al., 2017). The GDM-based models limit simulations to scenarios with a Gaussian DOS, but effective temperature models also exist for an exponential DOS. Regardless, any of these transport models can be inserted directly into the drift-diffusion system of equations, so long as the current created by the spatially dependent charge-carrier mobilities remains small enough.

332

Handbook of Organic Materials for Electronic and Photonic Devices

10.3.3 Modeling organic photovoltaics 10.3.3.1 BHJ morphology effects Simulating OPVs is uniquely challenging due to the complex, 3D BHJ morphology that is required to produce efficient devices. In many cases, the effective medium approach is used in macroscopic simulations, where electrons are transported on the lowest unoccupied molecular orbital (LUMO) of the acceptor phase and holes are transported in the highest occupied molecular orbital (HOMO) of the donor phase. While both levels are separated in energy, no spatial constraints for electrons and holes exist in an effective medium model. Morphology effects on recombination and charge transport, among others, have to be included parametrically. In addition, band-gap differences between the donor and acceptor materials are not treated explicitly, and an effective band gap is used instead. As discussed in the “Microscopic simulation and modeling” section, the details of the BHJ morphologies are often important, so this is one of the main reasons to apply 2D or 3D drift-diffusion models instead of effective medium models. Among the first, Buxton and Clarke (2006) investigated the concentration profiles of excitons, electrons, and holes in 2D morphologies formed by diblock copolymers. Other work considered the optimum ratio between well-mixed BHJs and regions with neat donor and acceptor phases. It turned out that the average domain size in the well-mixed regions is important to the device efficiency. If one compares the influence of well-blended BHJ morphologies containing neat domains of donor and acceptor molecules in a simulated 3D device, the question of whether a complex 3D morphology can be calculated by a 1D drift-diffusion simulation can be answered. It was found that the total current from a solar cell can be approximated by the sum of two 1D simulations: one accounting for currents originating from the mixed phase, and one for charge carriers from an interface region between the mixed-phase and neat (acceptor) domains. Because the current generated from the blended phase always exceeds the contribution of the neat phases, 1D simulations are often sufficient to predict the current-voltage characteristics (Bartesaghi and Koster, 2015). In summary, this means that in contrast to KMC simulations, effects of percolation are often not explicitly addressed in macroscopic simulations. However, effects of a stack of electrically active materials typical for devices can be addressed only in macroscopic simulations, which show variations in the spatial charge-carrier density distributions (to cite one example). This effect makes recombination spatially dependent—a fact neglected by simple recombination models. This is one reason for observed lower-than-expected recombination rates in OSCs by averaging charge-extraction experiments.

10.3.3.2 Charge recombination There are a variety of charge-recombination mechanisms and models, including the Langevin, Onsager-Braun, Shockley-Read-Hall (SRH), Auger, and other models (Wagenpfahl, 2017). Any of these can be directly implemented into a drift-diffusion simulation to probe their influence on the device performance.

Advances in modeling the physics of disordered organic electronic devices

333

In traditional solar cells, the dominant charge-carrier recombination mechanism can be estimated from current-voltage characteristics. The J-V curve is often treated as an ideal diode described by the Shockley equation. For a perfect pn-junction, the recombination mechanism can be approximated by the ideality factor,  nid ¼

kB T ∂ ln J q ∂V

1

:

(10.11)

Values of unity indicate direct Langevin recombination, whereas values of 2 and 2/3 indicate SRH and Auger recombination, respectively (G€ohler et al., 2018). Most experimental measurements on OPVs do not show distinct values, but rather more complex dependencies, such as charge-carrier density (illumination) or temperature. Drift-diffusion simulations can help to understand this by modeling a more realistic scenario than captured by the Shockley equation. Under the assumption of exponential tail states, ideality factors between unity and 1.5 can be obtained by additional recombination pathways through tail-state traps. Also, the spatial distribution of charge carriers in the device can decrease the ideality factor (Kirchartz and Nelson, 2012). However, ideality factors above 1 are not necessarily a sign of trap-assisted recombination. In contrast to the recombination of geminate pairs such as excitons or CT states, charge-carrier losses mediated by tail states seem to be most relevant to reproduce current-voltage curves with ideality factors not equal to 1 (Soldera et al., 2012). Due to the weak charge-carrier density dependence of mobilities in a system with a Gaussian DOS, exponential tail states or deep traps might be necessary to model current-voltage characteristics with higher ideality factors. Also, numerical fits to current-voltage characteristics point to the relevance of trap-assisted recombination (Liu and Li, 2011). However, the dominant mechanism can depend on illumination conditions. At low illumination intensities, recombination tends to be dominated by trap-assisted recombination, whereas direct Langevin recombination becomes more prominent with higher light intensities around one sun (Tress et al., 2013). With the help of macroscopic simulations, it was shown that current-voltage characteristics of OPVs often cannot be described by the Shockley equation. Instead, the low charge-carrier mobilities typical for OSCs lead to a current-transport-limited current-voltage curve, in which charge-carrier accumulations hinder charge extraction. It is, therefore, much more precise to determine ideality factors from Voc-Jsc pairs with changing temperature or illumination instead of from dark current-voltage curves (Tvingstedt and Deibel, 2016). The ideality factor is not only sensitive to recombination in the bulk of the film but also to surface recombination. In cases where electrons and holes are mutually transferred into a metal electrode, ideality factors less than or equal to 1 are reported. If one charge-carrier species is not efficiently transported across an interface, S-shaped current-voltage curves are found, which no longer follow the Shockley equation and show a plateau of the current increase around the open-circuit voltage. Besides surface recombination, charge-carrier mobility mismatch between two adjacent semiconductor layers, or blocking layers, also have been found to create such S-shaped

334

Handbook of Organic Materials for Electronic and Photonic Devices

current-voltage curves. The common origin in all cases is the formation of spacecharge regions under certain voltage regimes. Thus, S-shaped curves indicate a current transport limitation during extraction or an injection barrier into the adjacent layer. Corresponding simulations have helped to explain such a charge extraction limitation, thereby enabling researchers to pinpoint the back contact as the bottleneck in one case (H€ ubler et al., 2011).

10.3.3.3 Optimizing material properties and device design A key strength of macroscopic simulations is the ability to simulate current-voltage characteristics and predict how the device performance depends on variations in material properties, such as charge-carrier mobility. Parameters typically used to quantify a solar cell are the open-circuit voltage (Voc), short-circuit current (Jsc), fill factor (FF), and power conversion efficiency (PCE). All of these values are a function of the parameters used to simulate a solar cell. Since the early applications of macroscopic simulations in OPVs, it has become clear that the PCE does not automatically increase with a higher charge-carrier mobility. The reason for this is that, while high mobility leads to a high Jsc and low remaining charge-carrier densities in the device, this is accompanied by a larger Langevin recombination rate and a lower Voc. For low charge-carrier mobilities, Voc is high with a lower recombination rate, but slow charge exaction limits Jsc, which thereby reduces the PCE. It was also demonstrated that balanced (equal) electron and hole mobilities are required to gain a high PCE, as shown in Fig. 10.7 (Wagenpfahl et al., 2010). Rules for designing efficient OPVs have been

Fig. 10.7 Power conversion efficiency as function of balanced mobilities. Driftdiffusion simulations predict optimal mobility to maximize PCE in OPVs for a standard device (solid black line), a device with surface passivation (dashed blue line), and for limited transfer velocities of electrons and holes over both metalsemiconductor interfaces (dotted red line). (Data from Wagenpfahl, A., Deibel, C., Dyakonov, V., 2010. Organic solar cell efficiencies under the aspect of reduced surface recombination velocities. IEEE J. Sel. Top. Quantum Electron. 16 (6), 1759–1763.)

Advances in modeling the physics of disordered organic electronic devices

335

derived from these findings. Also, the energetic disorder has to be recognized as a relevant factor for the photocurrent. In addition, the question of how to design a device with optimum PCE using a provided set of materials can be answered. The first question may be whether a planar bilayer heterojunction or a BHJ architecture is better. As it turns out, both device structures differ in the charge-carrier distribution as a function of the metal electrode work function and the specific properties of the semiconductor layers, such as the exciton diffusion length (Foertig et al., 2012). Regardless of the heterojunction architecture used, interference effects in thin films must be considered to simulate the exciton generation process accurately. Optical interference has been shown to influence the photocurrent of organic solar cells significantly, as shown in Fig. 10.8. Capturing this phenomenon requires knowledge about the complex refractive index and the thicknesses of all layers to generate accurate optical absorption profiles that determine the overall exciton generation rate and the resulting photocurrent. Especially in complex structures such as tandem solar cells, optical simulations are required to tune the thickness of the layers optimally so that each subcell contributes the same current density and the recombination is minimized. All together, drift-diffusion simulations can combine all optical and electronic effects to predict the optimal heterojunction architecture and layer thicknesses needed to reach the highest PCE (H€ausermann et al., 2009). –10

–10

–9

–9

–8

–8

Jsc (mA/cm2)

–7

–7 reff

–6 reff = 1 %

–5

–4

reff = 10 %

–3 –2

reff = 25 %

–3

reff = 50 %

–2

reff = 100 %

–1 0

–5

reff = 5 %

–4

–1

Optical simulation

0

50

–6

100 150 200 P3HT:PCBM thickness (nm)

250

300

Fig. 10.8 Optical interference effects in P3HT:PCBM solar cells. Transfer-matrix simulations are used to predict optical interference effects in P3HT:PCBM BHJ OPVs, leading to several peaks in the photocurrent as a function of the layer thickness that are reduced by recombination losses. (Adapted with permission from H€ausermann, R., Knapp, E., Moos, M., Reinke, N.A., Flatz, T., Ruhstaller, B., 2009. Coupled optoelectronic simulation of organic bulk-heterojunction solar cells: parameter extraction and sensitivity analysis. J. Appl. Phys. 106 (10), 104507.)

336

Handbook of Organic Materials for Electronic and Photonic Devices

10.3.3.4 Transient simulations A number of important experiments in OPV research are based on the transient response of devices, which reveals the charge-carrier generation and recombination kinetics. With these techniques, a short illumination pulse generally is used to create charge carriers, and then one measures the transient current response of the device under an applied bias. Transient drift-diffusion simulations allow one to model such experiments, but it is challenging due to issues with the stability of the differential equation solver (Mock, 1983), as well as the limitation of most macroscopic simulations to the transport of thermalized charge carriers. Recent developments show that the limitations of dispersive transport can be overcome (Felekidis et al., 2016). Using transient macroscopic simulations, transient photocurrent curves can be reproduced, and one can conclude that the initial increase of the photocurrent is a function of the free charge-carrier formation, charge-transport, and charge-carrier recombination. Fits to photocurrent transients have shown the influence of charge-carrier trapping and detrapping, as well as the importance of trap-assisted recombination. In some cases, an exponential distribution of tail-state traps seems to be relevant for modeling the experimental data (MacKenzie et al., 2012). Different slopes in the transient photocurrent decay were attributed to changing from a nondispersive regime to a dispersive charge-transport regime within an exponential DOS (Christ et al., 2013). Another typical application of transient experiments is the estimation of chargecarrier mobilities. Depending on the average extraction time after the excitation pulse and the device thickness, the charge-carrier mobility can be calculated. The photo-charge extraction by linear increasing voltage (photo-CELIV) technique applies a linearly increasing voltage ramp to extract the charge carriers and can be used to estimate the charge-carrier mobilities. Because analytic approximations of the photo-CELIV response are often less accurate, transient simulations can provide more accurate parameter extraction (Neukom et al., 2011). These benefits also apply to other experiments, including the standard transient photocurrent measurements.

10.3.3.5 Extracting accurate fit parameters A frequently raised question is whether macroscopic simulations can be used to extract fit parameters from a set of measurements and use them to predict the device behavior accurately. Fits to steady-state current-voltage characteristics are generally possible, but not unambiguous. Parametrical fits often are used to reproduce certain measurements, but it was recently demonstrated that most fit parameters show a correlation to each other, and therefore, they cannot be easily separated. An accurate and unique parameter set can be extracted only from a set of steady-state and transient measurements (Neukom et al., 2012).

Advances in modeling the physics of disordered organic electronic devices

337

10.3.4 Modeling organic light-emitting diodes Applications of the drift-diffusion model for OLEDs are less numerous in the recent research literature. Due to the progress of the OLED technology, it can be expected that significant work has been conducted in companies without revealing the results to the general public. There should be no doubt that drift-diffusion simulations play a very important role in the design of state-of-the-art OLED devices and displays. However, current transport and calculating the origin of the generated photons is just one step in the development of OLED devices. The propagation and out coupling of the generated light is comparably important. In terms of charge-carrier transport, the current-voltage characteristics of OLEDs can be fit by macroscopic simulations. Tunneling currents from metals into OSCs seem to be occasionally necessary for that purpose (Martin et al., 2005; Altazin et al., 2016). Analogous to OPVs, the influence of charge-carrier injection layers, layer thicknesses, or charge-carrier mobilities can be systematically studied to maximize the radiative recombination of charge carriers. The recombination zone can be engineered by choosing semiconductor layers (or Dopants) that accumulate electrons and holes at a mutual position (Erickson and Holmes, 2013). Charge-carrier gradients and charge-carrier mobilities limit the overall recombination efficiency (Kasparek et al., 2018). Charge-carrier mobilities and activation energies often are measured by charge extraction and impedance spectroscopy experiments. However, it was shown that Arrhenius plots, which are used to examine the temperature dependence of charge-carrier mobilities, are of limited use in certain OLEDs because chargecarrier gradients and nonlinear electric-field distributions inside the device structure lead to systematic errors (Z€ ufle et al., 2017). The integration of single OLEDs to more complex devices such as displays can be seen as the next step. For these applications, the switching (on/off/pulse) behavior of OLEDs and the corresponding light emission must be optimized (Pflumm et al., 2008). Aging effects need to be understood such as the appearance of additional hole traps in aged devices made of the often-used semiconductor known as “super yellow” PPV (Niu et al., 2016). With the knowledge of the radiative recombination efficiency, complex multilayer OLEDs with several recombination centers with different wavelengths can be simulated. In combination with light-propagation calculations, the angular-dependent emission of white-light LEDs can be simulated (Perucco et al., 2012). Even the most complex systems, such as AMOLED displays, can be simulated at large scales, even though additional assumptions and calculations might be required (Diethelm et al., 2018). Overall, drift-diffusion simulations are a powerful tool to model and predict the electrical behavior of OPV and OLED devices in steady-state and transient conditions. Avoiding prohibitive costs of precise microscopic transport models tracking each individual particle, drift-diffusion simulations investigate the influence of materials’ properties and device architectures using a faster analytic framework. Drift-diffusion simulations represent a link between theoretical models and real-life applications, where other models can be applied only in certain regions of a more complex structure. Macroscopic simulations allow this kind of analysis, and therefore, they are an important tool to understand the physics of OPVs, OLEDs, and other devices.

338

10.4

Handbook of Organic Materials for Electronic and Photonic Devices

Conclusions and outlook

As with all computational methods, there are pros and cons to using KMC, driftdiffusion, or other methods. There are physical phenomena uniquely captured by KMC simulations and cases where drift-diffusion modeling is more appropriate. To choose the most appropriate technique, researchers should carefully consider their specific scientific question and the resources that are available. While coarse-grained KMC simulations are often much less computationally intensive than ab initio methods, most device-simulation problems require the use of high-performance computing clusters. Alternatively, less computationally expensive techniques, like driftdiffusion and master equation methods, can be reasonably executed on a personal computer. Despite these significant costs, for problems where one is interested in the impact of materials’ microstructure on device performance or on nonequilibrium phenomena, KMC is usually the most appropriate tool. However, due to the major computational cost of explicit electrostatic interaction calculations and simulating charge injection with Ohmic contacts, fitting KMC simulations to experimental device data is often prohibitively expensive. When wanting to perform extensive experimental device data-fitting, drift-diffusion techniques are much more appropriate. Looking to the future, there are many avenues that will lead to further progress. From an application standpoint, it is still not possible to predict a priori the performance of a particular material or blend in a specific device application. This limitation stems largely from shortcomings in our physical models on many length scales. At the molecular level, there has recently been an increasing effort to simulate the arrangement of the OSC molecules and calculate the resulting electronic properties of the material. However, validation of these predictions is often difficult due to measurement limitations, large computational cost, and difficulty verifying with experiments. Nevertheless, as molecular-simulation methods and computational power continue to improve, it will be important to utilize this information in device simulations to help us understand which molecular details dictate the device physics. Even on a larger scale, materials’ microstructure at the nanoscale to microscale due to crystalline-amorphous domains and DA domains can have a major impact on device performance. Microscopic modeling techniques could be greatly improved by integrating more information from state-of-the-art morphology characterization techniques to construct more accurate structure-property relationships. Further, for predicting device performance using macroscopic device modeling, the accuracy could be improved significantly if analytical expressions were explicitly derived from microscopic simulations. Fitting drift-diffusion models to device data also could be improved significantly if fit uniqueness was enhanced by fitting to higher-dimensionality datasets, including a series of measurements with various temperatures, illumination intensities, layer thicknesses, and other data, or to multiple types of measurement techniques. To make these broad multiscale-modeling efforts possible, there is a distinct need in the organic electronics community to create high-quality, open, and interoperable simulation, modeling, and analysis software tools.

Advances in modeling the physics of disordered organic electronic devices

339

Acknowledgments Michael C. Heiber is supported by financial assistance award No. 70NANB14H012 from the US Department of Commerce, National Institute of Standards and Technology, as part of the Center for Hierarchical Materials Design (CHiMaD).

References Albrecht, U., B€assler, H., 1995. Langevin-type charge carrier recombination in a disordered hopping system. Phys. Status Solidi B 191, 455–459. Altazin, S., Z€ufle, S., Knapp, E., Kirsch, C., Schmidt, T., J€ager, L., Noguchi, Y., Br€ utting, W., Ruhstaller, B., 2016. Simulation of OLEDs with a polar electron transport layer. Org. Electron. 39, 244–249. Bakulin, A.A., Rao, A., Pavelyev, V.G., van Loosdrecht, P.H.M., Pshenichnikov, M.S., Niedzialek, D., Cornil, J., Beljonne, D., Friend, R., 2012. The role of driving energy and delocalized states for charge separation in organic semiconductors. Science 335, 1340–1344. Banerji, N., 2013. Sub-picosecond delocalization in the excited state of conjugated homopolymers and donor-acceptor copolymers. J. Mater. Chem. C 1, 3052–3066. Baranovskii, S.D., 2014. Theoretical description of charge transport in disordered organic semiconductors. Phys. Status Solidi B 251, 487–525. Baranovskii, S.D., Wiemer, M., Nenashev, A.V., Jansson, F., Gebhard, F., 2012. Calculating the efficiency of exciton dissociation at the interface between a conjugated polymer and an electron acceptor. J. Phys. Chem. Lett. 3, 1214–1221. Bartesaghi, D., Koster, L.J.A., 2015. The effect of large compositional inhomogeneities on the performance of organic solar cells: a numerical study. Adv. Funct. Mater. 25 (13), 2013–2023. B€assler, H., 1993. Charge transport in disordered organic photoconductors. Phys. Status Solidi B 175, 15–56. Bjorgaard, J.A., K€ose, M.E., 2015. Simulations of singlet exciton diffusion in organic semiconductors: a review. RSC Adv. 5, 8432–8445. Bouhassoune, M., van Mensfoort, S., Bobbert, P., Coehoorn, R., 2009. Carrier-density and fielddependent charge-carrier mobility in organic semiconductors with correlated Gaussian disorder. Org. Electron. 10 (3), 437–445. Braun, C.L., 1984. Electric field assisted dissociation of charge transfer states as a mechanism of photocarrier production. J. Chem. Phys. 80, 4157–4161. Burke, T.M., McGehee, M.D., 2014. How high local charge carrier mobility and an energy cascade in a three-phase bulk heterojunction enable >90% quantum efficiency. Adv. Mater. 26, 1923–1928. Burke, T.M., Sweetnam, S., Vandewal, K., McGehee, M.D., 2015. Beyond Langevin recombination: how equilibrium between free carriers and charge transfer states determines the open-circuit voltage of organic solar cells. Adv. Energy Mater. 5, 1500123. Buxton, G.A., Clarke, N., 2006. Predicting structure and property relations in polymeric photovoltaic devices. Phys. Rev. B 74 (8), 085207. Caruso, D., Troisi, A., 2012. Long-range exciton dissociation in organic solar cells. Proc. Natl. Acad. Sci. 109, 13498–13502.

340

Handbook of Organic Materials for Electronic and Photonic Devices

Casalegno, M., Bernardi, A., Raos, G., 2013. Numerical simulation of photocurrent generation in bilayer organic solar cells: comparison of master equation and kinetic Monte Carlo approaches. J. Chem. Phys. 139, 024706. Christ, N., Kettlitz, S., Mescher, J., Valouch, S., Lemmer, U., 2013. Dispersive transport in the temperature dependent transient photoresponse of organic photodiodes and solar cells. J. Appl. Phys. 113 (23), 234503. Clarke, T.M., Durrant, J.R., 2010. Charge photogeneration in organic solar cells. Chem. Rev. 110, 6736–6767. Clarke, T.M., Rodovsky, D.B., Herzing, A.A., Peet, J., Dennler, G., DeLongchamp, D., Lungenschmied, C., Mozer, A.J., 2005. Significantly reduced bimolecular recombination in a novel silole-based polymer:fullerene blend. Adv. Energy Mater. 1, 1062–1067. Clarke, T.M., Peet, J., Denk, P., Dennler, G., Lungenschmied, C., Mozer, A.J., 2012. NonLangevin bimolecular recombination in a silole-based polymer: PCBM solar cell measured by time-resolved charge extraction and resistance-dependent time-of-flight techniques. Energy Environ. Sci. 5, 5241–5245. Coehoorn, R., Bobbert, P.A., 2012. Effects of Gaussian disorder on charge carrier transport and recombination in organic semiconductors. Phys. Status Solidi A 209, 2354–2377. Coehoorn, R., van Eersel, H., Bobbert, P.A., Janssen, R.A.J., 2015. Kinetic Monte Carlo study of the sensitivity of OLED efficiency and lifetime to materials parameters. Adv. Funct. Mater. 25, 2024–2037. Collins, B.A., Tumbleston, J.R., Ade, H., 2011. Miscibility, crystallinity, and phase development in P3HT/PCBM solar cells: toward an enlightened understanding of device morphology and stability. J. Phys. Chem. Lett. 2, 3135–3145. Coropceanu, V., Bredas, J.L., Mehraeen, S., 2017. Impact of active layer morphology on bimolecular recombination dynamics in organic solar cells. J. Phys. Chem. C 121, 24954–24961. D’Avino, G., Muccioli, L., Castet, F., Poelking, C., Andrienko, D., Soos, Z.G., Cornil, J., Beljonne, D., 2016. Electrostatic phenomena in organic semiconductors: fundamentals and implications for photovoltaics. J. Phys. Condens. Matter. 28, 433002. Deibel, C., Dyakonov, V., 2010. Polymer-fullerene bulk heterojunction solar cells. Rep. Prog. Phys. 73, 096401. Deibel, C., Strobel, T., Dyakonov, V., 2009. Origin of the efficient polaron-pair dissociation in polymer-fullerene blends. Phys. Rev. Lett. 103, 036402. Deibel, C., Wagenpfahl, A., Dyakonov, V., 2009. Origin of reduced polaron recombination in organic semiconductor devices. Phys. Rev. B 80, 075203. Dexter, D.L., 1953. A theory of sensitized luminescence in solids. J. Chem. Phys. 21, 836–859. Diethelm, M., Penninck, L., Altazin, S., Hiestand, R., Kirsch, C., Ruhstaller, B., 2018. Quantitative analysis of pixel crosstalk in AMOLED displays. J. Inf. Disp. 316, 1–9. Dimitrov, S.D., Durrant, J.R., 2014. Materials design considerations for charge generation in organic solar cells. Chem. Mater. 26, 616–630. Donets, S., Pershin, A., Christlmaier, M.J.A., Baeurle, S.A., 2013. A multiscale modeling study of loss processes in block-copolymer-based solar cell nanodevices. J. Chem. Phys. 138, 094901. Donets, S., Pershin, A., Baeurle, S.A., 2015. Optimizing the fabrication process and interplay of device components of polymer solar cells using a field-based multiscale solar-cell algorithm. J. Chem. Phys. 142, 184902. Erickson, N.C., Holmes, R.J., 2013. Investigating the role of emissive layer architecture on the exciton recombination zone in organic light-emitting devices. Adv. Funct. Mater. 23 (41), 5190–5198.

Advances in modeling the physics of disordered organic electronic devices

341

Felekidis, N., Melianas, A., Kemerink, M., 2016. Nonequilibrium drift-diffusion model for organic semiconductor devices. Phys. Rev. B 94 (3), 1–7. Ferguson, A.J., Kopidakis, N., Shaheen, S.E., Rumbles, G., 2011. Dark carriers, trapping, and activation control of carrier recombination in neat P3HT and P3HT:PCBM blends. J. Phys. Chem. C 115, 23134–23148. Few, S., Frost, J.M., Nelson, J., 2015. Models of charge pair generation in organic solar cells. Phys. Chem. Chem. Phys. 17, 2311–2325. Foertig, A., Wagenpfahl, A., Gerbich, T., Cheyns, D., Dyakonov, V., Deibel, C., 2012. Nongeminate recombination in planar and bulk heterojunction organic solar cells. Adv. Energy Mater. 2 (12), 1483–1489. F€ orster, T., 1948. Zwischenmolekulare energiewanderung und fluoreszenz. Ann. Phys. 437, 55–75. Frost, J.M., Cheynis, F., Tuladhar, S.M., Nelson, J., 2006. Influence of polymer-blend morphology on charge transport and photocurrent generation in donor-acceptor polymer blends. Nano Lett. 6, 1674–1681. Gagorik, A.G., Mohin, J.W., Kowalewski, T., Hutchison, G.R., 2013. Monte Carlo simulations of charge transport in 2D organic photovoltaics. J. Phys. Chem. Lett. 4, 36–42. Gagorik, A.G., Mohin, J.W., Kowalewski, T., Hutchison, G.R., 2015. Effects of delocalized charge carriers in organic solar cells: predicting nanoscale device performance from morphology. Adv. Funct. Mater. 25, 1996–2003. Gao, F., Ingan€as, O., 2014. Charge generation in polymer-fullerene bulk-heterojunction solar cells. Phys. Chem. Chem. Phys. 16, 20291. Gartstein, Y.N., Conwell, E.M., Rice, M.J., 1996. Electron-hole collision cross section in discrete hopping systems. Chem. Phys. Lett. 249, 451–458. Giazitzidis, P., Argyrakis, P., Bisquert, J., Vikhrenko, V.S., 2014. Charge separation in organic photovoltaic cells. Org. Electron. 15, 1043–1049. Gillespie, D.T., 1976. A general method for numerically simulating the stochastic time evolution of coupled chemical reactions. J. Comp. Phys. 22, 403–434. G€ ohler, C., Wagenpfahl, A., Deibel, C., 2018. Nongeminate recombination in organic solar cells. Adv. Energy Mater. 1 (1), 1–2. Gorenflot, J., Heiber, M.C., Baumann, A., Lorrmann, J., Gunz, M., K€ampgen, A., Dyakonov, V., Deibel, C., 2014. Nongeminate recombination in neat P3HT and P3HT: PCBM blend films. J. Appl. Phys. 115, 144502. Grancini, G., Maiuri, M., Fazzi, D., Petrozza, A., Egelhaaf, H.J., Brida, D., Cerullo, G., Lanzani, G., 2013. Hot exciton dissociation in polymer solar cells. Nat. Mater. 12, 29–33. Gregg, B.A., 2011. Entropy of charge separation in organic photovoltaic cells: the benefit of higher dimensionality. J. Phys. Chem. Lett. 2, 3013–3015. Groves, C., 2013. Developing understanding of organic photovoltaic devices: Monte Carlo models of geminate and non-geminate recombination, charge transport, and charge extraction. Energy Environ. Sci. 6, 3202–3217. Groves, C., 2013. Suppression of geminate charge recombination in organic photovoltaic devices with a cascaded energy heterojunction. Energy Environ. Sci. 6, 1546–1551. Groves, C., Greenham, N.C., 2008. Bimolecular recombination in polymer electronic devices. Phys. Rev. B 78, 155205. Groves, C., Marsh, R.A., Greenham, N.C., 2008. Monte Carlo modeling of geminate recombination in polymer-polymer photovoltaic devices. J. Chem. Phys. 129, 114903. Groves, C., Koster, L.J.A., Greenham, N.C., 2009. The effect of morphology upon mobility: implications for bulk heterojunction solar cells with nonuniform blend morphology. J. Appl. Phys. 105, 094510.

342

Handbook of Organic Materials for Electronic and Photonic Devices

Groves, C., Kimber, R.G.E., Walker, A.B., 2010. Simulation of loss mechanisms in organic solar cells: a description of the mesoscopic Monte Carlo technique and an evaluation of the first reaction method. J. Chem. Phys. 133, 144110. Guo, J., Ohkita, H., Benten, H., Ito, S., 2010. Charge generation and recombination dynamics in poly(3-hexylthiophene)/fullerene blend films with different regioregularities and morphologies. J. Am. Chem. Soc. 132, 6154–6164. Hamilton, R., Shuttle, C.G., O’Regan, B., Hammant, T.C., Nelson, J., Durrant, J.R., 2010. Recombination in annealed and nonannealed polythiophene/fullerene solar cells: transient photovoltage studies versus numerical modeling. J. Phys. Chem. Lett. 1, 1432–1436. H€ausermann, R., Knapp, E., Moos, M., Reinke, N.A., Flatz, T., Ruhstaller, B., 2009. Coupled optoelectronic simulation of organic bulk-heterojunction solar cells: parameter extraction and sensitivity analysis. J. Appl. Phys. 106 (10), 104507. Heeger, A.J., 2014. 25th anniversary article: bulk heterojunction solar cells: understanding the mechanism of operation. Adv. Mater. 26, 10–28. Heiber, M.C., 2016. Ising_OPV v3.0. https://doi.org/10.5281/zenodo.60505. Heiber, M.C., Dhinojwala, A., 2012. Dynamic Monte Carlo modeling of exciton dissociation in organic donor-acceptor solar cells. J. Chem. Phys. 137, 014903. Heiber, M.C., Dhinojwala, A., 2013. Estimating the magnitude of exciton delocalization in regioregular P3HT. J. Phys. Chem. C 117, 21627–21634. Heiber, M.C., Dhinojwala, A., 2014. Efficient generation of model bulk heterojunction morphologies for organic photovoltaic device modeling. Phys. Rev. Appl. 2, 014008. Heiber, M.C., Baumbach, C., Dyakonov, V., Deibel, C., 2015. Encounter-limited charge-carrier recombination in phase-separated organic semiconductor blends. Phys. Rev. Lett. 114, 136602. Heiber, M.C., Nguyen, T.Q., Deibel, C., 2016. Charge carrier concentration dependence of encounter-limited bimolecular recombination in phase-separated organic semiconductor blends. Phys. Rev. B 93, 205204. Heiber, M.C., Kister, K., Baumann, A., Dyakonov, V., Deibel, C., Nguyen, T.Q., 2017. Impact of tortuosity on charge-carrier transport in organic bulk heterojunction blends. Phys. Rev. Appl. 8, 054043. Hilczer, M., Tachiya, M., 2010. Unified theory of geminate and bulk electron-hole recombination in organic solar cells. J. Phys. Chem. C 114, 6808–6813. Hood, S.N., Kassal, I., 2016. Entropy and disorder enable charge separation in organic solar cells. J. Phys. Chem. Lett. 7, 4495–4500. Houili, H., Tutisˇ, E., Batistic, I., Zuppiroli, L., 2006. Investigation of the charge transport through disordered organic molecular heterojunctions. J. Appl. Phys. 100, 033702. Howard, I.A., Etzold, F., Laquai, F., Kemerink, M., 2014. Nonequilibrium charge dynamics in organic solar cells. Adv. Energy Mater. 4, 1301743. H€ ubler, A., Trnovec, B., Zillger, T., Ali, M., Wetzold, N., Mingebach, M., Wagenpfahl, A., Deibel, C., Dyakonov, V., 2011. Printed paper photovoltaic cells. Adv. Energy Mater. 1 (6), 1018–1022. Jailaubekov, A.E., Willard, A.P., Tritsch, J.R., Chan, W.L., Sai, N., Gearba, R., Kaake, L.G., Williams, K.J., Leung, K., Rossky, P.J., Zhu, X.Y., 2013. Hot charge-transfer excitons set the time limit for charge separation at donor/acceptor interfaces in organic photovoltaics. Nat. Mater. 12, 66–73. Jamieson, F.C., Domingo, E.B., McCarthy-Ward, T., Heeney, M., Stingelin, N., Durrant, J.R., 2012. Fullerene crystallisation as a key driver of charge separation in polymer/fullerene bulk heterojunction solar cells. Chem. Sci. 3, 485–492.

Advances in modeling the physics of disordered organic electronic devices

343

Jones, M.L., Dyer, R., Clarke, N., Groves, C., 2014. Are hot charge transfer states the primary cause of efficient free-charge generation in polymer:fullerene organic photovoltaic devices? A kinetic Monte Carlo study. Phys. Chem. Chem. Phys. 16, 20310–20320. Kaake, L.G., Jasieniak, J.J., Ronald, C., Bakus, I., Welch, G.C., Moses, D., Bazan, G.C., Heeger, A.J., 2012. Photoinduced charge generation in a molecular bulk heterojunction material. J. Am. Chem. Soc. 134, 19828–19838. Kasparek, C., R€orich, I., Blom, P.W., Wetzelaer, G.J.A., 2018. Efficiency of solution-processed multilayer polymer light-emitting diodes using charge blocking layers. J. Appl. Phys. 123 (2), 024504. Kirchartz, T., Nelson, J., 2012. Meaning of reaction orders in polymer:fullerene solar cells. Phys. Rev. B 86, 165201. Kordt, P., van der Holst, J.J.M., Helwi, M.A., Kowalsky, W., May, F., Badinksi, A., Lennartz, C., Andrienko, D., 2015. Modeling of organic light emitting diodes: from molecular to device properties. Adv. Funct. Mater. 5, 1955–1971. Koster, L.J.A., 2010. Charge carrier mobility in disordered organic blends for photovoltaics. Phys. Rev. B 81, 205318. Koster, L.J.A., Mihailetchi, V.D., Blom, P.W.M., 2006. Bimolecular recombination in polymer/ fullerene bulk heterojunction solar cells. Appl. Phys. Lett. 88, 052104. Kurpiers, J., Neher, D., 2016. Dispersive non-geminate recombination in an amorphous polymer:fullerene blend. Sci. Rep. 6, 26832. Lakhwani, G., Rao, A., Friend, R.H., 2014. Bimolecular recombination in organic photovoltaics. Annu. Rev. Phys. Chem. 65, 557–581. Lange, I., Kniepert, J., Pingel, P., Dumsch, I., Allard, S., Janietz, S., Scherf, U., Neher, D., 2013. Correlation between the open circuit voltage and the energetics of organic bulk heterojunction solar cells. J. Phys. Chem. Lett. 4, 3865–3871. Langevin, P., 1903. Recombinaison et mobilites des ions dans les gaz. Ann. Chim. Phys. 28, 433. Liu, L., Li, G., 2011. Investigation of recombination loss in organic solar cells by simulating intensity-dependent current-voltage measurements. Sol. Energy Mat. Sol. C 95 (9), 2557–2563. Lyons, B.P., Clarke, N., Groves, C., 2011. The quantitative effect of surface wetting layers on the performance of organic bulk heterojunction photovoltaic devices. J. Phys. Chem. C 115, 22572–22577. Lyons, B.P., Clarke, N., Groves, C., 2012. The relative importance of domain size, domain purity and domain interfaces to the performance of bulk-heterojunction organic photovoltaics. Energy Environ. Sci. 5, 7657–7663. MacKenzie, R.C.I., Shuttle, C.G., Chabinyc, M.L., Nelson, J., 2012. Extracting microscopic device parameters from transient photocurrent measurements of P3HT:PCBM solar cells. Adv. Energy Mater. 2, 662–669. Marcus, R.A., 1993. Electron transfer reactions in chemistry. Theory and experiment. Rev. Mod. Phys. 65, 599–610. Marsh, R.A., Groves, C., Greenham, N.C., 2007. A microscopic model for the behavior of nanostructured organic photovoltaic devices. J. Appl. Phys. 101, 083509. Martin, S.J., Walker, A.B., Campbell, A.J., Bradley, D.D.C., 2005. Electrical transport characteristics of single-layer organic devices from theory and experiment. J. Appl. Phys. 98 (6), 063709. McMahon, D.P., Cheung, D.L., Troisi, A., 2011. Why holes and electrons separate so well in polymer/fullerene photovoltaic cells. J. Phys. Chem. Lett. 2, 2737–2741.

344

Handbook of Organic Materials for Electronic and Photonic Devices

McNeill, C.R., Westenhoff, S., Groves, C., Friend, R.H., Greenham, N.C., 2007. Influence of nanoscale phase separation on the charge generation dynamics and photovoltaic performance of conjugated polymer blends: balancing charge generation and separation. J. Phys. Chem. C 111, 19153–19160. Mehraeen, S., Coropceanu, V., Bredas, J.L., 2013. Role of band states and trap states in the electrical properties of organic semiconductors: hopping versus mobility edge model. Phys. Rev. B 87 (19), 195209. Melianas, A., Pranculis, V., Devizˇis, A., Glubinas, V., Ingan€as, O., Kemerink, M., 2014. Dispersion-dominated photocurrent in polymer:fullerene solar cells. Adv. Funct. Mater. 24, 4507–4514. Melianas, A., Etzold, F., Savenije, T.J., Laquai, F., Ingan€as, O., Kemerink, M., 2015. Photogenerated carriers lose energy during extraction from polymer-fullerene solar cells. Nat. Commun. 6, 8778. Menke, S.M., Holmes, R.J., 2014. Exciton diffusion in organic photovoltaic cells. Energy Environ. Sci. 7, 499–512. Menke, S.M., Ran, N.A., Bazan, G.C., Friend, R.H., 2017. Understanding energy loss in organic solar cells: toward a new efficiency regime. Joule 2, 1–11. Mesta, M., Carvelli, M., de Vries, R.J., van Eersel, H., van der Holst, J.J.M., Schober, M., Furno, M., L€ussem, B., Leo, K., Loebl, P., Coehoorn, R., Bobbert, P.A., 2013. Molecular-scale simulation of electroluminescence in a multilayer white organic light-emitting diode. Nat. Mater. 12, 652–658. Mikhnenko, O.V., Blom, P.W.M., Nguyen, T.Q., 2015. Exciton diffusion in organic semiconductors. Energy Environ. Sci. 8, 1867–1888. Miller, A., Abrahams, E., 1960. Impurity conduction at low concentrations. Phys. Rev. 120, 745–755. Mock, M.S., 1983. Analysis of Mathematical Models of Semiconductor Devices. Boole Press Limited, Dublin. € Mozer, A.J., Dennler, G., Sariciftci, N.S., Westerling, M., Pivrikas, A., Osterbacka, R., Jusˇka, G., 2005. Time-dependent mobility and recombination of the photoinduced charge carriers in conjugated polymer/fullerene bulk heterojunction solar cells. Phys. Rev. B 72, 035217. Murawski, C., Leo, K., Gather, M.C., 2013. Efficiency roll-off in organic light emitting-diodes. Adv. Mater. 25, 6801–6827. Nan, G., Zhang, X., Lu, G., 2016. The lowest-energy charge-transfer state and its role in charge separation in organic photovoltaics. Phys. Chem. Chem. Phys. 18, 17546–17556. Nelson, J., 2003. Diffusion-limited recombination in polymer-fullerene blends and its influence on photocurrent collection. Phys. Rev. B 67, 155209. Neukom, M., Reinke, N., Ruhstaller, B., 2011. Charge extraction with linearly increasing voltage: a numerical model for parameter extraction. Sol. Energy 85 (6), 1250–1256. Neukom, M., Z€ufle, S., Ruhstaller, B., 2012. Reliable extraction of organic solar cell parameters by combining steady-state and transient techniques. Org. Electron. 13 (12), 2910–2916. Niu, Q., Wetzelaer, G.J.A.H., Blom, P.W.M., Cr€aciun, N.I., 2016. Modeling of electrical characteristics of degraded polymer light-emitting diodes. Adv. Electron. Mat. 2 (8), 1600103. Novikov, S.V., 2003. Charge-carrier transport in disordered polymers. J. Polym. Sci. Pol. Phys. 41, 2584–2594. Novikov, S.V., Dunlap, D.H., Kenkre, V.M., Parris, P.E., Vannikov, A.V., 1998. Essential role of correlations in governing charge transport in disordered organic materials. Phys. Rev. Lett. 81, 4472–4475.

Advances in modeling the physics of disordered organic electronic devices

345

Oelerich, J.O., Huemmer, D., Baranovskii, S.D., 2012. How to find out the density of states in disordered organic semiconductors. Phys. Rev. Lett. 108, 226403. Oelerich, J.O., Jansson, F., Nenashev, A.V., Gebhard, F., Baranovskii, S.D., 2014. Energy position of the transport path in disordered organic semiconductors. J. Phys. Condens. Matter. 26 (25), 255801. Oelerich, J.O., Nenashev, A.V., Dvurechenskii, A.V., Gebhard, F., Baranovskii, S.D., 2017. Field dependence of hopping mobility: lattice models against spatial disorder. Phys. Rev. B 96, 195208. Offermans, T., Meskers, S.C.J., Janssen, R.A.J., 2003. Charge recombination in a poly (para-phenylene vinyl)-fullerene derivative composite film studied by transient, nonresonant, hole-burning spectroscopy. J. Chem. Phys. 119, 10924–10929. Offermans, T., Meskers, S.C.J., Janssen, R.A.J., 2005. Monte-Carlo simulations of geminate electron-hole pair dissociation in a molecular heterojunction: a two-step dissociation mechanism. Chem. Phys. 308, 125–133. Perucco, B., Reinke, N.A., Rezzonico, D., Knapp, E., Harkema, S., Ruhstaller, B., 2012. On the exciton profile in OLEDs-seamless optical and electrical modeling. Org. Electron. 13 (10), 1827–1835. Peumans, P., Forrest, S.R., 2004. Separation of geminate charge-pairs at donor-acceptor interfaces in disordered solids. Chem. Phys. Lett. 398, 27–31. Peumans, P., Uchida, S., Forrest, S.R., 2003. Efficient bulk heterojunction photovoltaic cells using small-molecular-weight organic thin films. Nature 425, 158–162. Pflumm, C., Gartner, C., Lemmer, U., 2008. A numerical scheme to model current and voltage excitation of organic light-emitting diodes. IEEE J. Quantum Electron. 44 (7–8), 790–798. Pivrikas, A., Jusˇka, G., Mozer, A.J., Scharber, M., Arlauskas, K., Sariciftci, N.S., Stubb, H., € Osterbacka, R., 2005. Bimolecular recombination coefficient as a sensitive testing parameter for low-mobility solar-cell materials. Phys. Rev. Lett. 94, 176806. Proctor, C.M., Kuik, M., Nguyen, T.Q., 2013. Charge carrier recombination in organic solar cells. Prog. Polym. Sci. 38, 1941–1960. Reineke, S., Thomschke, M., L€ussem, B., Leo, K., 2013. White organic light-emitting diodes: status and perspective. Rev. Mod. Phys. 85, 1245–1293. Richert, R., B€assler, H., 1985. Energetic relaxation of triplet excitations in vitreous benzophenone. Chem. Phys. Lett. 118, 235–239. Richert, R., Pautmeier, L., B€assler, H., 1989. Diffusion and drift of charge carriers in a random potential: deviation from Einstein’s law. Phys. Rev. Lett. 63, 547–550. Ries, B., B€assler, H., 1984. Monte Carlo study of bimolecular charge-carrier recombination in anisotropic lattices. Chem. Phys. Lett. 108, 71–75. Savoie, B.M., Rao, A., Bakulin, A.A., Gelinas, S., Movaghar, B., Friend, R.H., Marks, T.J., Ratner, M.A., 2014. Unequal partnership: asymmetric roles of polymeric donor and fullerene acceptor in generating free charge. J. Am. Chem. Soc. 136, 2876–2884. Scher, H., Montroll, E.W., 1975. Anomalous transit-time dispersion in amorphous solids. Phys. Rev. B 12, 2455–2477. Sch€onherr, G., Eiermann, R., B€assler, H., Silver, M., 1980. Dispersive exciton transport in a hopping system with Gaussian energy distribution. Chem. Phys. 52, 287–298. Selberherr, S., 1984. Analysis and Simulation of Semiconductor Devices. Springer-Verlag, Wien, New York. Shen, Y., Giebink, N.C., 2015. Monte Carlo simulations of nanoscale electrical inhomogeneity in organic light-emitting diodes and its impact on their efficiency and lifetime. Phys. Rev. Appl. 4, 054017.

346

Handbook of Organic Materials for Electronic and Photonic Devices

Shklovskii, B.I., Efros, A.L., 1984. Electronic Properties of Doped Semiconductors. Springer Series in Solid-State Sciences, vol. 45. Springer, Berlin, Heidelberg, p. 388. Soldera, M., Taretto, K., Kirchartz, T., 2012. Comparison of device models for organic solar cells: band-to-band vs. tail states recombination. Phys. Status Solidi A 209 (1), 207–215. Stehr, V., Pfister, J., Fink, R.F., Engels, B., Deibel, C., 2011. First-principles calculations of anisotropic charge-carrier mobilities in organic semiconductor crystals. Phys. Rev. B 83, 155208. Street, R.A., 2015. Electronic structure and properties of organic bulk-heterojunction interfaces. Adv. Mater. 28, 3814–3830. Tress, W., Leo, K., Riede, M., 2013. Dominating recombination mechanisms in organic solar cells based on ZnPc and C60. Appl. Phys. Lett. 102 (16), 163901. Tutisˇ, E., Batistic, I., Berner, D., 2004. Injection and strong current channeling in organic disordered media. Phys. Rev. B 70, 161202. Tvingstedt, K., Deibel, C., 2016. Temperature dependence of ideality factors in organic solar cells and the relation to radiative efficiency. Adv. Energy Mater. 6 (9), 1502230. van der Holst, J.J.M., van Oost, F.W.A., Coehoorn, R., Bobbert, P.A., 2009. Electron-hole recombination in disordered semiconductors: validity of the Langevin formula. Phys. Rev. B 80, 235202. van der Holst, J.J.M., van Oost, F.W.A., Coehoorn, R., Bobbert, P.A., 2011. Monte Carlo study of charge transport in organic sandwich-type single-carrier devices: effects of Coulomb interactions. Phys. Rev. B 83, 085206. van Eersel, H., Janssen, R.A.J., Kemerink, M., 2012. Mechanism for efficient photoinduced charge separation at disordered organic heterojunctions. Adv. Funct. Mater. 22, 2700–2708. van Eersel, H., Bobbert, P.A., Janssen, R.A.J., Coehoorn, R., 2014. Monte Carlo study of efficiency roll-off of phosphorescent organic light-emitting diodes: evidence for dominant role of triplet-polaron quenching. Appl. Phys. Lett. 105, 143303. Vandewal, K., Albrecht, S., Hoke, E.T., Graham, K.R., Widmer, J., Douglas, J.D., Schubert, M., Mateker, W.R., Bloking, J.T., Burkhard, G.F., Sellinger, A., Frechet, J.M.J., Amassian, A., Riede, M.K., McGehee, M.D., Neher, D., Salleo, A., 2014. Efficient charge generation by relaxed charge-transfer states at organic interfaces. Nat. Mater. 13, 63–68. Vithanage, D.A., Devizˇis, A., Abramavicˇius, V., Infahsaeng, Y., Abramavicˇius, D., MacKenzie, R.C.I., Keivanidis, P.E., Yartsev, A., Hertel, D., Nelson, J., Sundstr€ om, V., Gulbinas, V., 2013. Visualizing charge separation in bulk heterojunction organic solar cells. Nat. Commun. 4, 2334. Voter, A.F., 2007. Introduction to the kinetic Monte Carlo method. In: Sickafus, K.E., Kotomin, E.A., Uberuaga, B.P. (Eds.), Radiation Effects in Solids. In: vol. 1. Springer, Dordrecht, pp. 1–23. Wagenpfahl, A., 2017. Mobility dependent recombination models for organic solar cells. J. Phys. Condens. Matter. 29 (37), 373001. Wagenpfahl, A., Deibel, C., Dyakonov, V., 2010. Organic solar cell efficiencies under the aspect of reduced surface recombination velocities. IEEE J. Sel. Top. Quantum Electron. 16 (6), 1759–1763. Watkins, P.K., Walker, A.B., Verschoor, G.L.B., 2005. Dynamical Monte Carlo modelling of organic solar cells: the dependence of internal quantum efficiency on morphology. Nano Lett. 5, 1814–1818. Wetzelaer, G.A.H., Koster, L.J.A., Blom, P.W.M., 2011. Validity of the Einstein relation in disordered organic semiconductors. Phys. Rev. Lett. 107, 066605.

Advances in modeling the physics of disordered organic electronic devices

347

Wojcik, M., Michalak, P., Tachiya, M., 2010. Geminate electron-hole recombination in organic solids in presence of a donor-acceptor heterojunction. Appl. Phys. Lett. 96, 162102. Wolf, U., Arkhipov, I., B€assler, H., 1999. Current injection from a metal to a disordered hopping system. I. Monte Carlo simulation. Phys. Rev. B 59, 7507–7513. Yang, F., Forrest, S.R., 2008. Photocurrent generation in nanostructured organic solar cells. ACS Nano 2, 1022–1032. Yu, Z.G., Smith, D.L., Saxena, A., Martin, R.L., Bishop, A.R., 2001. Molecular geometry fluctuations and field-dependent mobility in conjugated polymers. Phys. Rev. B 63, 085202. Z€ ufle, S., Altazin, S., Hofmann, A., J€ager, L., Neukom, M.T., Br€ utting, W., Ruhstaller, B., 2017. Determination of charge transport activation energy and injection barrier in organic semiconductor devices. J. Appl. Phys. 122 (11), 115502.

This page intentionally left blank

Doping in organic semiconductors

11

Hannes Hase, Ingo Salzmann Concordia University, Montr
eal, QC, Canada

11.1

Introduction

As late as 1931, Wolfgang Pauli, one of the brightest minds of his time, commented in a letter to Rudolf Peierls, another pioneer of modern physics, “One shouldn’t work on semiconductors, that’s a real mess [orig., Schweinerei], who knows whether semiconductors exist at all” (translated from German by the authors; see also von Meyenn, 1985). Finally, recognizing the relevance and high potential of semiconducting materials certainly represents one of the most important paradigm shifts in the physics of the 20th century. The pioneering work of Sir Alan Herries Wilson shortly thereafter (Wilson, 1931a,b) established that the fundamental physics of inorganic semiconductors was formulated for the first time by providing the basis of electron band theory, a view that has dominated solid-state physics ever since. In particular, by highlighting the role of impurities acting as donors and acceptors, as well as that of electron and hole conduction, and by distinguishing between extrinsic and intrinsic semiconductors, this fundamentally new perception represents the foundation of inorganic semiconductor doping. This view and its technological exploitation constitute the basis for the multitude of electronic devices used in today’s information society. Highly recommendable historic reviews cover this pioneering time of semiconductor physics in great detail (Busch, 1989; Madelung, 1999).

11.1.1 Inorganic semiconductors To highlight the marked contrast to the mechanisms of doping organic semiconductors—the ongoing research that is the focus of this chapter—we first summarize some aspects of the well-established mechanisms for doping inorganic semiconductors. Doping is done by introducing in a controlled manner individual impurity atoms into an otherwise highly pure semiconductor (e.g., boron into silicon for p-type doping) at typical dopant concentrations as low as 10
6 to 10
3 (Fig. 11.1). Through covalent interaction with the host atoms forming the crystalline matrix, electronic defect states are generated within the fundamental energy gap of the formally intrinsic (i.e., undoped) semiconductor. With p-type doping silicon with boron, for instance, empty states with energies slightly above the valence-band edge are created [44 meV for B in Si (Sze, 1981)]. As all available electronic states are occupied following Fermi-Dirac statistics, this altered density of states (DOS) leads to a shift of the Fermi level (EF)—that is, the electron chemical potential—from midgap down toward Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00011-5 © 2019 Elsevier Ltd. All rights reserved.

350

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 11.1 Introducing covalently bonded boron atoms as p-dopants into a crystalline silicon host matrix generates mobile holes in the inorganic semiconductor (left). Schematic energylevel diagram for p-type doping as a function of dopant concentration (right). VB denotes the valence band of the inorganic semiconductor, CB its conduction band, EF the Fermi energy, and f(E) the Fermi function. From Salzmann, I., Heimel, G., 2015. Toward a comprehensive understanding of molecular doping organic semiconductors (review). J. Electron Spectrosc. Relat. Phenom. 204, 208–222. doi:10.1016/j.elspec.2015.05.001. Copyright 2015 Elsevier.

the valence-band edge. As the acceptor states are shallow by design, essentially all of them become occupied with electrons at room temperature. As a result, mobile holes are generated in the valence band. From here, n-type doping proceeds in full analogy, now with deliberately introduced shallow donor states (e.g., phosphorous for silicon) that are designed to lie closely below the conduction-band edge [46 meV for P in Si (Sze, 1981)]. Again, complete ionization of these states at room temperature, now accompanied by an upward shift of EF, generates mobile charge carriers (now electrons) in the n-doped inorganic semiconductor. Importantly, for higher doping concentrations, but typically still at levels in the order of %, EF eventually crosses the valence-band/conduction-band edge, thus entering the regime of degenerate p/n-doping. There, the inorganic semiconductor turns essentially metallic. Doping inorganic semiconductors allows for changing the alignment of their electronic bands relative to those of electrodes and to differently doped semiconductors. Such interfacial phenomena are employed in p-n junctions or field-effect transistors (for example) and allow a multitude of device functionalities to be exploited in silicon-based electronics. The p/ n-doping efficiency, defined as holes/electrons generated per added dopant, of typically  1 leads to a dramatic increase in semiconductor conductivity even at doping ratios well below 10
5 (Sze, 1981), which allows for retaining high charge carrier mobility, as the crystalline order of the highly pure host remains largely unperturbed and only a few scattering centers are introduced.

11.1.2 Organic semiconductors Despite the high level of control that doping provides over electrical materials parameters, tuning the band gap of inorganic semiconductors for opto-electronic applications such as displays or solid-state lighting remains a technological challenge.

Doping in organic semiconductors

351

Moreover, the energy-intensive production necessary for inorganic semiconductors makes them less appealing for low-cost, large-area applications. It is precisely this field in which the high potential of electronics based on organic semiconductors (OSCs) lies; that is, conjugated molecules (COMs) and conjugated polymers (CPs). While the phenomena of electroluminescence and photovoltaic (PV) response of OSCs have been known for around half a century (Chamberlain, 1983; Pope et al., 1963; Tang and Albrecht, 1975), it was not until the 1980s that reports on promising organic light-emitting diodes (OLEDs, cf. Chapter 21), organic photovoltaic cells (OPVCs, cf. Chapter 20), and organic thin-film transistors (OTFTs, cf. Chapter 23) (Horowitz, 1998; Horowitz et al., 1989; Tang, 1986; Tang and VanSlyke, 1987) finally triggered the intense research efforts on practical applications of OSCs that still prevails today. The current success in the application of OSCs, particularly that in the field of OLED-based display technology, is facilitated by the versatility of this materials class, as opto-electronic properties can be tailored to application-specific demands that employ strategies of synthetic organic chemistry. Equally important, OSCs have high absorption coefficients and are conveniently processable on flexible substrates through vacuum sublimation (COMs), versatile solution-processing methods (for both CPs and soluble COMs), or both, including spin coating, blade coating, slot-die coating, inkjet printing, or roll-to-roll processing (Richter et al., 2017). For a detailed discussion of the fundamental properties of OSCs, their chemical design strategies, and hybrid structures with inorganic semiconductors, see Part I of this book; Part III focuses on details of OSC-processing techniques, while Part IV highlights in detail the many applications that can be realized using OSCs.

11.2

Doping organic semiconductors

11.2.1 Initial doping approaches In the 1970s, the entire field of organic semiconductor research originated in the successful doping of both CPs (Basescu et al., 1987; Chiang et al., 1977; Shirakawa et al., 1977) and COMs (Yamamoto et al., 1979) through the admixture of halides and alkali metals. Using this approach, the conductivity of OSCs could be increased by multiple orders of magnitude. The results of early studies on the electronic structure of these doped OSCs employing ultraviolet photoelectron spectroscopy (UPS) were fully in line with the expectations from the physics of doped inorganic semiconductors (L€ogdlund et al., 1989). For example, upon p-doping poly(3-hexylthiophene) (P3HT) with the inorganic dopant NOPF6, its electronic bands shifted toward EF, which is indicative of charge transfer between the OSC and the dopant. The pioneering studies of this period triggered an immense interest in OSCs, which finally culminated in the award of the 2000 Nobel Prize in Chemistry for the discovery of conductive polymers to Alan J. Heeger, Alan G. MacDiarmid, and Hideki Shirakawa (Heeger, 2001). For a summary of the state of knowledge—both theoretical and experimental—as it was established already in the late 1980s, see the comprehensive reviews by Heeger et al. (1988) and Br
edas and Street (1985). Given its key role in

352

Handbook of Organic Materials for Electronic and Photonic Devices

inorganic semiconductor technology, one might assume that doping had played a key role in the early days of organic electronics as well. However, the resounding success of employing doped OSCs in practical applications was inhibited by issues inherent to this approach. Small dopants exhibit a strong tendency to diffuse, which renders doped injection layers for ohmic contacts or p-n junctions unstable under operating conditions (Parthasarathy et al., 2001).

11.2.2 Breakthrough of doping by using molecular dopants To resolve these issues, molecular dopants have been introduced, which finally led to the breakthrough of doping OSCs for use in practical applications. Today, all commercially available OLED displays—representing a multibillion-dollar market— comprise molecularly doped OSCs, as do white OLEDs, which hold great potential for future large-area lighting (Reineke et al., 2009). In all practical applications, molecular electron donors are employed as n-dopants and, conversely, molecular electron acceptors like tetracyanoquinodimethane (TCNQ; Melby et al., 1962) and its derivatives as p-dopants for both CPs and COMs; the chemical structures of typical materials used in the field are depicted in Fig. 11.2A. Indeed, following the approach of molecular electrical doping, the conductivity of CPs (Yim et al., 2008) and COMs (Harada et al., 2010) can be tuned over multiple orders of magnitude, as illustrated for the chemically similar CP P3HT and the related COM quarterthiophene (4T) in Fig. 11.2B. Doping OSCs allows concurrently reducing both ohmic losses in transport layers and injection barriers at interfaces to electrodes, which enables establishing low-contact resistances independent of the initial electrode work function. In p-i-n multilayer structures where p-doped hole- and n-doped electron-transport layers sandwich a thin, intrinsic emission layer, both effects are exploited for OLEDs and OPVCs. In the latter, electron- and hole-transport layers are realized with molecularly doped wide-gap OSCs, which sandwich the photovoltaically active donor/ acceptor heterojunction. This prevents parasitic exciton recombination at the metallic electrodes and such layers act as conducting spacers for optical device optimization (L€ ussem et al., 2013a; Maennig et al., 2004). There, the spatial region of constructive optical field interference upon light reflection at the metallic back-electrode can be tailored to lie within the photovoltaically active region by adjusting the transport-layer thickness. In tandem solar cells, doped interlayers also are employed in the recombination zone (Ameri et al., 2009; L€ ussem et al., 2013a); and doping the active layer in CP-based bulk heterojunction solar cells was shown to lead to considerable improvements in OPVC performance (Zhang et al., 2013). Likewise, in the highly promising field of perovskite photovoltaics, using molecularly doped hole-transport layers has been demonstrated to improve the overall device performance (Zhang et al., 2016), again, by reducing ohmic losses. For OLEDs, the p-i-n concept also has been successfully employed (He et al., 2004; Walzer et al., 2007), where the doped transport layers now sandwich intrinsic emission layers, and electron/hole blocking layers in between confine the charges for efficient emission and reduce exciton quenching at the electrodes.

Doping in organic semiconductors

353

Fig. 11.2 See legend on next page.

354

Fig. 11.2 (A) Top: the chemical structures of the OSCs (left) and p-dopants (right) discussed in this chapter. Bottom: the chemical structure of the dimeric n-dopant [RuCp ∗Mes]2 that separates into two cationic monomers along with the two suggested underlying processes. Process 1: electron transfer from the dimer to the host’s excited state, followed by cleavage of the cation dimer. Process 2: dimer-to-host electron transfer through an intermolecular charge-transfer absorption, followed by cleavage of the cation dimer. (B) Thin-film conductivity of F4TCNQ-doped 4T (top) and P3HT (bottom). (A) Bottom image modified from Lin, X., Wegner, B., Lee, K.M., Fusella, M.A., Zhang, F., Moudgil, K., Rand, B.P., Barlow, S., Marder, S.R., Koch, N., Kahn, A., 2017. Beating the thermodynamic limit with photo-activation of n-doping in organic semiconductors. Nat. Mater. 16, 1209–1215. doi:10.1038/nmat5027. Copyright 2017 Macmillan Publishers Ltd; (B) From Salzmann, I., Heimel, G., Oehzelt, M., Winkler, S., Koch, N., 2016. Molecular electrical doping of organic semiconductors: fundamental mechanisms and emerging dopant design rules. Acc. Chem. Res. 49, 370–378. doi:10.1021/acs.accounts.5b00438. Copyright 2016 American Chemical Society.

Handbook of Organic Materials for Electronic and Photonic Devices

Doping in organic semiconductors

355

In contrast to OPVCs, however, interfacial energy offsets between the electrode Fermi level and the frontier molecular orbital levels of the OSC layers matter for OLEDs, which here represent electron- and hole-injection barriers. Dopants dispersed in the transport layers effectively reduce these barriers upon contact with the electrodes, as they can undergo charge transfer. This leads to an increase/decrease of the respective electrode work function for p/n-dopants and, therefore, to a reduction of the respective injection barrier (Br€ oker et al., 2008; Koch et al., 2005). For OTFTs, molecular n/p-doping has been demonstrated to reduce the bias stress effect and therefore significantly improve device stability for applications in logic circuits (Hein et al., 2014), and increase the field-effect mobility (Ma et al., 2008). For the first fully operative (top-contact) inversion transistor reported by L€ussem et al., an n-doped OSC layer at the gate dielectric and the same p/n-doped OSC at the interface to the top electrodes was employed, allowing efficient hole/electron injection (L€ussem et al., 2013b); details on the device physics of doped OTFTs can be found in a recent comprehensive review by the same author (L€ ussem et al., 2016). Clearly, OSCs are particularly appealing when they can play to their strengths in application fields that are hardly accessible for silicon-based electronics. Recently, Wang et al. (2017) reported stable OTFT performance of devices established on paper with performance benchmarks comparable to those of control devices on solid substrates. This was promoted by interface engineering, again through decorating the silver source and drain electrodes with a molecular p-dopant (Choi et al., 2016). However, using molecular dopants entails issues that severely limit this approach for practical applications. For CPs, which cannot be processed via vacuum deposition as they are thermally fragile due to their high molecular weight, the most obvious problem is limited solubility of dopants in a common solvent with the OSC. This has been comprehensively discussed for F4TCNQ and its derivatives by Li et al. (2015). There, the authors synthesized derivatives where one or two cyanogroups are ester-substituted (see Fig. 11.2A) for increased solubility in organic solvents like chloroform. While solubility can be increased by orders of magnitude by this approach (see Table 11.1), this comes at the price of a reduced electron affinity (EA), which makes these dopants less efficient for ionizing CPs like P3HT. To overcome solubility issues in common solvents, Jacobs et al. (2016) suggested sequential doping of OSC and dopant using orthogonal solvents (i.e., solvents that preferentially solve only one species); a similar study was recently done by Chew et al. (2017). This approach relies on the diffusion of the dopant into the OSC film and leads to higher conductivity as the microstructure is less adversely affected. CPs like P3HT typically form semicrystalline films where amorphous regions coexist with crystalline film portions. By sequential deposition, the dopants tend to penetrate both the amorphous part and the alkyl chain region of the crystalline portion; however, that does not affect the packing of the polymer backbone, which leads to higher conductivities than for films prepared from a common solution. A similar approach was suggested by Hynynen et al. (2017) and Lim et al. (2018), who performed vapor phase infiltration of P3HT films. There, exposing preprepared P3HT films to F4TCNQ vapor leads again to superior structural properties of the doped films and enhanced conductivity. Also, by exposure to F4TCNQ vapor, Jacobs et al. (2015) reported a new role for dopants in facilitating lateral patterning,

356

Handbook of Organic Materials for Electronic and Photonic Devices

Table 11.1 LUMO positions ELUMO (electron affinities) from cyclovoltammetry (CV) and solubilities in chloroform of F4TCNQ derivatives (chemical structures in Fig. 11.2A), as introduced by Li et al. (2015) Compound

ELUMO (eV)

Solubility (mmol/L)

F4TCNQ F4MCTCNQ F4OCTCNQ F4DMCDCNQ F4DMCDCNQ


5.23
5.14
5.12
4.97
4.93

2.0 113.8 204.1 67.9 88.2

where their vapor deposition through a shadow mask generates spatially well defined doped and undoped areas. The authors then exploited the fact that P3HT:F4TCNQ integer charge transfer complexes are insoluble in common organic solvents and, therefore, undoped sites could simply be washed off with a substance like ο-dichlorobenzene. They further showed that upon exposure to 405-nm wavelength light, P3HT:F4TCNQ breaks up and the remaining P3HT can be removed by tetrahydrofuran (THF), which enables subtractive patterning. See a recent review by Jacobs and Moul
e (2017) that covers these advances in detail. Another general issue intrinsic to OSCs is environmental stability, which typically necessitates encapsulation in practical applications. By nature, this is even more critical for highly reactive electron donors/acceptors employed as n/p-dopants. In particular, the low ionization energy (IE) of n-dopants renders them susceptible to oxidation. A valuable approach to overcoming this problem is using air-stable dopant precursors which, upon incorporation into the host material, liberate a more reductive species (leaving residuals in the structure), or whose charge transfer is activated through bond formation/cleavage. Recently, Lin et al. (2017) presented an efficient approach without such residues, where they doped POPy2 with the n-dopant dimer [RuCp∗ Mes]2. Upon irradiation, cleavage of the dimer occurs following the two possible reaction routes depicted at the bottom of Fig. 11.2A, which finally induces electron transfer to the host OSC. While this process was observed for deep red light as well, it turned out to be most efficient in the ultraviolet. For the latter, conductivities increased by approximately six orders of magnitude and showed excellent stability over several hundreds of hours. While, on one hand, dopants may be subject to degradation, on the other hand, they can also help to improve device stability. Nikolka et al. (2017) found that nanometersized voids in CP films, prone to be filled by water molecules, are a key factor in charge trapping and device degradation. These voids, however, also can be filled by dopants. For their investigations, they chose to add F4TCNQ, TCNQ, and ABN to the CP IDTBT in an OFET device and observed improved performance stability. The latter two dopants have dimensions similar to that of F4TCNQ, but they do not significantly affect the overall electronic landscape, as their EAs are sufficiently lower than the IE of the polymer. This clearly demonstrates that the stabilizing function of a dopant presence augments the regular doping effects found for stronger dopants.

Doping in organic semiconductors

357

11.2.3 Contrast to inorganic semiconductors The probably most apparent difference between an inorganic semiconductor like silicon and an OSC are the natures of the elemental entities forming the semiconductor (i.e., atoms for the inorganic semiconductor and COMs/CPs for the organic semiconductor). By nature, for the latter, orientation does matter for essentially all (and therefore typically anisotropic) physical properties, as COMs/CPs are extended, nonrotationally invariant objects. This explains the strong tendency to polymorphism of organic molecular solids in general (Bernstein, 2002)—a key subject in the pharmaceutical sciences (Datta and Grant, 2004; Rodrı´guez-Spong, 2004)—and that of OSC thin films adsorbed on solid surfaces in particular ( Jones et al., 2016). In this context, it has been shown for OFETs that different polymorphs of the same channel material can affect the transistor’s characteristics significantly due to their various crystalline packings ( Jurchescu et al., 2009). Furthermore, these fundamentally different elemental entities bond fundamentally differently, either covalently (inorganic semiconductors) or via Van der Waals forces and electrostatic interactions (OSCs). This not only translates into a significantly lower bond strength of OSCs (responsible for their solubility and low melting points), but also entails marked differences in their electronic structure. These differences readily explain the highly dissimilar phenomenology of charge transport. In particular, the width of the energy bands formed by the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) states in an OSC is typically at least one order of magnitude lower—even in single crystals (Machida et al., 2010)—than that of crystalline silicon. Therefore, bandlike transport is more the exception than the rule, and apart from very pure OSC single crystals and polymers along the chain (Fratini et al., 2016), charges move by thermally assisted hopping from site to site due to the disorder—in particular, in thin OSC films relevant for devices in organic electronics. Another key difference with inorganic semiconductors is the significantly lower dielectric constants of OSCs (typically around 3, in contrast to >10 for inorganic semiconductors), which confines optical absorption and emission processes either to the orbitals of one single molecule or to immediately adjacent molecules. Thus, formed electron-hole pairs remain strongly Coulombically bound in the form of Frenkel excitons (B€assler and K€ohler, 2011). Fig. 11.3 (left) summarizes some key differences between silicon and the prototypical OSC pentacene and shows a juxtaposition between the DOS of single-crystalline silicon and amorphous silicon to highlight the role of structural order for DOS (taken from Horowitz, 2015). In an ideal inorganic semiconductor, the DOS is parabolic at the band edges where the mobile charge carriers are in terms of energy (electrons at the conduction band edge, holes at the valence band edge). Therefore, the DOS falls to zero in the semiconductor gap. Disorder introduces localized states into the fundamental gap region, as illustrated for amorphous silicon in Fig. 11.3 (right), and the region at the band edge can be modeled with an exponential DOS (Horowitz, 2015). In the same vein, the DOS of OSCs is governed by disorder, and a Gaussian DOS is commonly used there to model the electronic state distribution (B€assler, 1993; Oehzelt et al., 2014). The fact that OSCs are highly disordered systems, in contrast to inorganic

358

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 11.3 Some aspects of differences between silicon and the OSC pentacene, along with a juxtaposition of the DOS of single-crystalline and amorphous silicon. Image from Horowitz, G., 2015. Validity of the concept of band edge in organic semiconductors. J. Appl. Phys. 118, 115502. doi:10.1063/1.4931061. Copyright 2015 AIP Publishing.

semiconductors, has recently been shown to be ultimately responsible for the separation of electron and holes after initial doping-related charge transfer—that is, the formation of an ion pair (IPA) (Tietze et al., 2018). Tietze et al. showed that in fact, it is due to energetic disorder in OSCs that efficient carrier release can occur after initial IPA formation upon doping, as the dissociation has an activation energy of only a few tens of megaelectron volts (and therefore is well in the kBT range at room temperature), despite Coulomb binding energies of several 100 meV. They found that amorphous systems like MeO-TPD:F6TCNNQ and Ir(piq)3:F6TCNNQ yielded an activation energy of 9.1 and 9.5 meV, respectively, while for ZnPc:F6TCNNQ, a system of higher order, it more than doubled, to 20.7 meV (Tietze et al., 2018).

11.3

Electronic structure upon molecular electrical doping

Numerous studies have been dedicated to elucidating the electronic structure of OSCs upon doping, revealing a phenomenology that largely resembles that known from inorganic semiconductors. However, some subtle but central questions remained open and have long been under debate. Early studies by UPS, an experimental technique that allows for rather directly assessing the occupied DOS of matter (H€ufner, 2003), showed the expected shift of the CP valence band EF for P3HT doped with NOPF6 (L€ ogdlund et al., 1989), as illustrated in Fig. 11.4A. For pristine P3HT, the onset of the valence-band-related photoemission (determined by a tangent through the inflection point of the data) is observed about 1 eV below the Fermi energy (bottom curve), pointing to a midgap position of EF. Upon saturation doping, a finite DOS is observed at EF, which is in line with what one would expect for a degenerate inorganic semiconductor.

Doping in organic semiconductors

359

Fig. 11.4 (A) UPS data of NOPF6-doped P3HT; the finite DOS around EF is highlighted. (B) Photoemission investigation (UPS and XPS on the O1s core level) of thin films of MeOTPD doped with F4TCNQ up to high mole ratios. (C) Obsolete perception on the energetics of the singly occupied (polaronic) state upon p-doping. (D) Same as (C), but for n-doping. Modified from (A) L€ogdlund, M., Lazzaroni, R., Stafstr€om, S., Salaneck, W.R., Br
edas, J.-L., 1989. Direct observation of charge-induced π-electronic structural changes in a conjugated polymer. Phys. Rev. Lett. 63, 1841–1844. doi:10.1103/PhysRevLett.63.1841. Copyright 1989 American Physical Society; (B) Olthof, S., Tress, W., Meerheim, R., L€ ussem, B., Leo, K., 2009b. Photoelectron spectroscopy study of systematically varied doping concentrations in an organic semiconductor layer using a molecular p-dopant. J. Appl. Phys. 106, 103711. doi:10.1063/1.3259436. Copyright 2009 AIP Publishing; (C) Gao, W., Kahn, A., 2003a. Controlled p doping of the hole-transport molecular material N,N0 -diphenyl-N,N0 -bis (1-naphthyl)-1,10 -biphenyl-4,40 -diamine with tetrafluorotetracyanoquinodimethane. J. Appl. Phys. 94, 359–366. doi:10.1063/1.1577400. Copyright 2003 AIP Publishing; (D) Steinm€ uller, D., Ramsey, M.G., Netzer, F.P., 1993. Polaron and bipolaronlike states in n-doped bithiophene. Phys. Rev. B 47, 13323–13329. doi:10.1103/PhysRevB.47.13323. Copyright 1993 American Physical Society.

However, this observation remains isolated in the research literature, and therefore, it is fair to challenge this interpretation. Although the most obvious implication of doping-induced mutual ionization of OSC and dopant would be the presence of singly occupied states—which, until recently, had been thought to lie in the fundamental gap of the pristine OSC (Br
edas and Street, 1985; Bubnova et al., 2014; Heeger et al.,

360

Handbook of Organic Materials for Electronic and Photonic Devices

€ 1988; L€ ogdlund et al., 1989; Osterbacka et al., 2000)—such states never have been observed close to or at EF in other studies. In comprehensive UPS experiments, again using F4TCNQ, Olthof et al. (2009b) observed for various COMs a shift of the HOMO emission toward EF upon increasing the p-dopant concentration (Fig. 11.4B). This shift turned out to be rigid for valence and core levels assessed by UPS and X-ray photoelectron spectroscopy (XPS), respectively, retaining a largely constant ionization energy of the doped films. In numerous studies (Gao and Kahn, 2001, 2003a,b,c; Olthof et al., 2009a,b), such shifts are found to saturate at high dopant loading, and no evidence for the OSC turning degenerate (in the sense of an inorganic semiconductor) has been observed whatsoever. In this context, Olthof et al. state that “it is surprising that it is not possible to move the HOMO any closer to the Fermi energy than these 0.35 eV” (Olthof et al., 2009b), which it is, indeed, on the basis of the still-prevalent perception of the energetics of polaronic states in OSCs (Br
edas and Street, 1985; Bubnova et al., 2014; € Heeger et al., 1988; L€ogdlund et al., 1989; Osterbacka et al., 2000). This by-now obsolete view is illustrated for p-doping in Fig. 11.4C (Gao and Kahn, 2003a) and for n-doping in Fig. 11.4D (Steinm€ uller et al., 1993). We shared this disturbance in a related UPS study, where we could not observe such states at EF whatsoever in 10:1 and 1:1 mixed films of pentacene and F4TCNQ (Salzmann et al., 2012a).

11.3.1 The energetics of charges in organic semiconductors A model to explain the absence of photoemission intensity at EF upon doping was only recently suggested by Winkler et al. (2015), providing a revised view of the polaron energetics in OSCs. Fig. 11.5 depicts a detailed schematic energy-level diagram of the traditional view of positive polaron energetics, where a molecular cation is embedded into an environment of neutral molecules. Upon removal of an electron from the HOMO of a neutral molecule, geometrical relaxation occurs and the thusly singlyoccupied molecular orbital (SOMO) level is thought to shift into the fundamental energy gap of the OSC by the reorganization energy λ. This state, therefore, should be readily observable in UPS, while the relaxed LUMO should be observable by inverse photoelectron spectroscopy (IPES) probing the unoccupied DOS (Br
edas and Street, 1985; cf. Fig. 11.4C). To challenge this view experimentally, the authors realized a situation of largely unperturbed molecular cations by depositing a thin layer of fullerene (C60) on MoO3-coated Au(111) in the ultrahigh vacuum (UHV), where the MoO3 dielectric serves both for inhibiting orbital hybridization with the metal substrate (Heimel et al., 2013) and increasing the substrate work function beyond the IE of C60. To establish electronic equilibrium, electrons are transferred from C60 across the dielectric into Au(111), generating the desired cations for investigation via UPS and IPES. Similar to the studies on doped OSCs, these experiments showed no UPS intensity at EF. Interestingly, however, a new photoemission feature at a binding energy that is higher than that of the neutral molecule occurred while evidence for the presence of a relaxed unoccupied state in the fundamental gap came from IPES, which is in line with the traditional view (cf. Fig. 11.5A).

Doping in organic semiconductors

361

Fig. 11.5 (A) Traditional view of single-particle energy levels for neutral molecules (green) surrounding a cation (red); ϕel,∞ is the common vacuum level. (B) Left: revised energy levels in molecular semiconductors comprising polarons. The reorganization energy λ corresponds to the difference between IE0 and EA+, the IE of the neutral molecule, and the EA of the positive cation. The on-site Coulomb-interaction U causes a split into two HOMO-derived sublevels of energies EA+ and IE+ below the vacuum level. Right: Intersite Coulomb-interaction then causes a distance dependent shift (see ϕel,local) of the neutral molecular levels close to the cation by up to a value of V. Modified from Winkler, S., Amsalem, P., Frisch, J., Oehzelt, M., Heimel, G., Koch, N., 2015. Probing the energy levels in hole-doped molecular semiconductors. Mater. Horiz. 2, 427–433. doi:10.1039/C5MH00023H, published 2015 by The Royal Society of Chemistry under CC BY 3.0 (https://creativecommons.org/licenses/by/3.0/).

To explain this observation, the authors provided a revised picture of the electronic structure of a positive polaron in a neutral OSC matrix, as illustrated in Fig. 11.5B. There, IE0, the energy necessary to ionize a neutral molecule in the solid, equals the gain in energy upon returning it to the relaxed cation (electron affinity EA+) and the reorganization energy λ. Moreover, in this view, the cation ionization energy IE+ > IE0, which is due to the on-site Coulomb interaction between electrons in the HOMO (“Hubbard U”; Hubbard, 1963), leads to a split into an upper, unoccupied sublevel (shifted by λ into the gap of the OSC) and a lower occupied sublevel that lies at U-λ outside the OSC gap. From their experimental data, the authors deduced for a valence hole in solid C60 a value for U of about 1.4 eV (Winkler et al., 2015). Due to intersite Coulomb interaction (supported by the low OSC dielectric constant), for neutral molecules close to the cation—that is, within the Coulomb potential well (ϕel,local), a shift of their energy levels by up to V is observed. A revised picture of the negative polaron (i.e., a molecular anion) embedded in a matrix of neutral OSC molecules proceeds in the same vein (Winkler et al., 2015). We note that a similar view has been put forward more recently by Png et al. (2016), who investigated the p-doping of triarylamine-fluorene copolymers. Their findings also provide solid evidence for the splitting of the SOMO levels through Hubbard interaction and energy-level shifts induced by Coulomb interaction. Through these effects, the SOMO level is pushed below the valence-band edge of the neutral polymer for p-doping (cf. Fig. 11.5B). As a result, UPS experiments cannot provide direct evidence for charge transfer upon molecularly p-doping OSCs, as the SOMO photoemission is masked by that of the neutral OSC matrix.

362

Handbook of Organic Materials for Electronic and Photonic Devices

11.3.2 Indications for doping-related ion-pair (IPA) formation However, evidence for charge transfer upon doping can be deduced from various other spectroscopic techniques. In ultraviolet-visible/near-infrared (UV-vis/NIR) absorption spectroscopy, transitions characteristic of both the ionized dopant and the ionized OSC are observed, which gives strong evidence for the formation of IPAs upon doping. For example, upon p-doping regioregular (rr) P3HT by F4TCNQ, in addition to the fundamental absorption of neutral P3HT observed slightly above 2 eV (its optical gap), two sharp absorptions at 1.43 and 1.62 eV are observed (labeled as ICT) and assigned to optical transitions of the ionized dopant (M
endez et al., 2013; Pingel and Neher, 2013). These transitions are superimposed on broad absorptions due to the P3HT-positive polarons P1 and P2. However, on the basis of the revised picture of the polaron energetics in OSCs (cf. Fig. 11.5B), the traditional assignment of these € transitions (Osterbacka et al., 2000), as illustrated in Fig. 11.6B (left), cannot be left unchallenged. Augmenting the work in Winkler et al. (2015), Heimel (2016) recently provided a revised scheme for the indexation of these transitions in CPs in a theoretical study, taking into account mean-field Coulomb interaction, adiabatic lattice relaxation, and the delocalization of excess charges. On this basis, calculations using density functional theory (DFT) extend the view presented in Fig. 11.5B to the polymer case, again yielding IE+ > IE0 and reproducing the experimentally observed optical transitions P1 and P2, as schematically illustrated in Fig. 11.6B (right). While in principle, the traditional view suggests three optical subgap transitions, in the revised view, P1 is a transition from the valence band into the upper, unoccupied intragap state, and P2 from the valence-band edge (and lower-lying occupied levels) in the upper, empty intragap state (and higher-lying unoccupied levels) (Heimel, 2016). Upon increasing the degree of fluorination from TCNQ over FTCNQ, F2TCNQ, to F4TCNQ, their EA can be increased from about 4.2 eV (TCNQ) to 5.1 eV (F4TCNQ) (Kanai et al., 2009). Employing these differently strong dopants to P3HT in common solution does not change their UV-vis/NIR absorption characteristics in thin, solid films qualitatively (Fig. 11.6A, right); only the relative strength of the IPA related features is attenuated. For TCNQ, no such features are observed whatsoever, as the EA of the dopant is significantly lower than the IE of the OSC. Conversely, chemically modifying the OSC peripheral substitution pattern can significantly alter its electronic structure (Heimel et al., 2009, 2011), especially its IE. Recently, Li et al. (2017) highlighted the role of the chemical nature of the side chains for polythiophene, where adding a sulfur unit between the thiophene backbone and the dodecyl side chains led to the as-yet-highest reported nonionic conductivity among films made from dopantpolymer solution of 350 S/cm (employing NOBF4 as a dopant). This is understood as a direct consequence from the decrease in IE (leading to more efficient IPA formation) and, furthermore, an overall superior microstructure. Likewise, a similar approach was taken by Kroon et al. (2017), who attached oligoethylene glycol side chains to a polythiophene denoted p(g42T-T). Doping with F4TCNQ yielded conductivities up to 100 S/cm and, most important, enhanced thermal stability up to 150°C, a temperature at which P3HT doped with F4TCNQ would have lost a large part of its dopant content already (Kroon et al., 2017). The authors find the dopants to be

Doping in organic semiconductors

363

Fig. 11.6 (A) UV-vis/NIR absorption spectroscopy on F4TCNQ-doped P3HT at increasing dopant concentrations showing subgap absorptions assigned to the dopant anion (labeled as ICT) and the positive polaron in P3HT; reference spectra of pure neutral dopant and anion are provided; doping ratios given as dopants per polymer repeat unit. (B) Left: traditional view of the energy levels in a CP (cf. Fig. 11.5A), comprising positive polarons with the respective suggested optical transitions. Right: revised view of the energy levels for polarons in an OSC with the optical transitions observed in the experiment shown in (A). Dark- and light-grayshaded boxes symbolize the CP valence and conduction bands, respectively; straight, solid, and vertical arrows indicate electrons with according spin. Modified from (A) M
endez, H., Heimel, G., Winkler, S., Frisch, J., Opitz, A., Sauer, K., Wegner, B., Oehzelt, M., R€othel, C., Duhm, S., T€obbens, D., Koch, N., Salzmann, I., 2015. Charge-transfer crystallites as molecular electrical dopants. Nat. Commun. 6. doi:10.1038/ ncomms9560, published 2015 by Macmillan Publishers Ltd. under CC BY 4.0 (https:// creativecommons.org/licenses/by/4.0/); (B) Heimel, G., 2016. The optical signature of charges in conjugated polymers. ACS Cent. Sci. 2, 309–315. doi:10.1021/acscentsci.6b00073. Copyright 2016 American Chemical Society.

preferentially embedded into the oligo ethylene glycol side-chain region, the polar nature of which promotes binding to F4TCNQ. Probably the most direct technique to assess the formation of IPAs upon doping OSCs is electron paramagnetic resonance (EPR) spectroscopy, from which the density of singly occupied states can be derived. In a study on, again, F4TCNQ doped P3HT, Gao et al. (2013a) observed the doping-ratio dependent increase in ionic species; the data is shown in Fig. 11.7A. In this study, the authors juxtaposed the spin densities in F4TCNQ doped regioregular (rr) and regiorandom (ra) P3HT; i.e., the nominally same CP of different regiostereo regularities. Ideal rr-P3HT comprises only head-to-tail

364 Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 11.7 (A) EPR data of (top) rr-P3HT and (bottom) ra-P3HT doped with F4TCNQ at increasing dopant loading. (B) FTIR data in the region of the dopant C^N stretching modes for 4T doped with increasingly fluorinated TCNQ derivatives; the insets besides the data show the offset between the frontier energy levels (LUMO level of the dopant versus HOMO of 4T); the red dashed curve is the same for rr-P3HT. (A) From Gao, J., Niles, E.T., Grey, J.K., 2013a. Aggregates promote efficient charge transfer doping of poly(3-hexylthiophene). J. Phys. Chem. Lett. 4, 2953–2957. doi:10.1021/jz401555x. Copyright 2013 American Chemical Society; (B) Modified from M
endez, H., Heimel, G., Winkler, S., Frisch, J., Opitz, A., Sauer, K., Wegner, B., Oehzelt, M., R€ othel, C., Duhm, S., T€ obbens, D., Koch, N., Salzmann, I., 2015. Charge-transfer crystallites as molecular electrical dopants. Nat. Commun. 6. doi:10.1038/ncomms9560, published 2015 by Macmillan Publishers Ltd. under CC BY 4.0 (https:// creativecommons.org/licenses/by/4.0/).

Doping in organic semiconductors

365

coupling, while ra-P3HT consists of head-to-head, head-to-tail, and tail-to-tail in random order (Chen et al., 1995; Jiang et al., 2001). These differences in the chemical structure cause rr-P3HT to grow in the form of semicrystalline films, while ra-P3HT is amorphous. For comparable doping ratios, the authors found a g-value in the range of what is expected for organic radicals and recorded an integrated EPR signal stronger by almost three orders of magnitude for rr-P3HT upon saturation doping (beyond a dopant to monomer ration of 1:4). The authors concluded that efficient doping occurs only for the rr species due to the formation of well-ordered aggregates with more rigid and planar backbones promoting hole delocalization along the backbone, while the large torsional disorder impedes such intrachain hole delocalization, resulting in tightly bound species (Gao et al., 2013a; Kanemoto et al., 2008). This perception is supported by UV-vis/NIR data, where the IPA-related features (cf. Fig. 11.6A) are negligibly weak for ra-P3HT. Therefore, whether efficient doping via IPA formation can occur appears to be mainly attributed to structural properties (i.e., to the chemical structure of the CP, as well as to the microstructure of the doped films and their ability to form aggregates). IPA formation can be deduced further from Fourier-transform infrared spectroscopy (FTIR), where the degree of charge transfer (δ) can be assessed by characteristic shifts (Δν) of dopant-related vibrational bands. For TCNQ and its fluorinated derivatives, a linear relationship between δ and Δν has been proposed, where Δν varies between 0 for the neutral dopant (δ ¼ 0) and νion for its radical anion (δ ¼ 1). While the modes around 1500 cm
1 have been suggested by Meneghetti and Pecile (1986), as being most suitable for assessing δ with successful studies on mixed single crystals of COMs with TCNQ, F2TCNQ, and F4TCNQ (Salzillo et al., 2016), doped CPs like P3HT show intense spectral features in that region masking these modes (Rodrigues et al., 2013). In such cases, one instead must resort to the region of C^N stretching modes around 2200 cm
1 for assessing δ (Chappell et al., 1981; Kampar and Neilands, 1986; M
endez et al., 2015; Meneghetti and Pecile, 1986; Pingel et al., 2010; Salzmann et al., 2016); however, be aware that these modes of the cyano-wings are highly sensitive to intermolecular interactions, and thus polymorphism (Meneghetti and Pecile, 1986). This was theoretically explored on the DFT level by Haworth et al. (2014), who derived the vibrational bands for Li+ F4TCNQ
complexes. They found that these C ^ N modes are highly sensitive to the relative placement of the Li countercation. Nevertheless, the shift of these modes in P3HT doped with F4TCNQ (dashed red curve in Fig. 11.7B) equals that observed for dopants in salts with alkali metals (Na, K, Cs), where δ ¼ 1. This is, again, strong evidence for IPA formation in molecularly doped P3HT. However, most interestingly, a comparison with the thiophene oligomer 4T reveals significantly smaller shifts upon doping with the whole dopant series TCNQ—F4TCNQ (M
endez et al., 2015). Importantly, this points to a fractional charge transfer of about δ ¼ ¼ for the thiophene oligomer instead of IPA, as found for the polythiophene.

366

Handbook of Organic Materials for Electronic and Photonic Devices

11.3.3 Ground-state charge-transfer complex (CPX) formation Fractional charge transfer (δ < 1) is indicative of the second process that can occur upon molecularly doping OSCs; that is, the formation of an overall charge-neutral ground-state charge transfer complex (CPX) between the dopant and the OSC. Their frontier molecular orbitals hybridize, forming a doubly occupied bonding and an empty antibonding supramolecular orbital, as illustrated in Fig. 11.8A. It could be shown that in a H€ uckel-type model, the occupied (ECPX,H) and unoccupied (ECPX,L) frontier energy levels of such a CPX are given as follows (M
endez et al., 2013, 2015; Salzmann et al., 2016):

Fig. 11.8 (A) Energy-level splitting upon CPX formation, schematically illustrated for 4T/ F4TCNQ together with DFT-calculated isosurface plots of the bonding and antibonding supramolecular hybrid orbitals (center), the HOMO of 4T (left), and the LUMO of F4TCNQ (right); ECPX gap denotes the transport gap of the CPX and Evac the vacuum level. (B) Top: UV-vis/ NIR data on F4TCNQ-doped 4T with increasing dopant loading up to 1:1 blends. Bottom: UV-vis/NIR data of 4T/dopant blends (1:1 ratio) for the TCNQ—F4TCNQ dopant series of increasing dopant strength. (C) UPS/IPES data of 1:1 blends of 4T and F4TCNQ (right) as compared to a pristine 4T reference film (left) demonstrating both an increased IE and a reduced gap upon CPX formation; data are given as binding energy with respect to the vacuum level (Evac). Modified from M
endez, H., Heimel, G., Winkler, S., Frisch, J., Opitz, A., Sauer, K., Wegner, B., Oehzelt, M., R€othel, C., Duhm, S., T€obbens, D., Koch, N., Salzmann, I., 2015. Charge-transfer crystallites as molecular electrical dopants. Nat. Commun. 6. doi:10.1038/ncomms9560, published 2015 by Macmillan Publishers Ltd. under CC BY 4.0 (https://creativecommons.org/ licenses/by/4.0/).

Doping in organic semiconductors

ECPX, H=L ¼

HOSC + LDOP 1  2 2

367

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðHOSC
LDOP Þ2 + 4β2

(11.1)

Therefore, their separation in energy depends on the HOMO level energy of the OSC (HOSC), the LUMO level energy of the p-dopant (LDOP), and a resonance integral β, which takes into account the orbital overlap. Importantly, these levels do not need to be in resonance, and this process can be well operative for HOSC > LDOP; i.e., in a case where the IE of the OSC is larger than the EA of the p-dopant (M
endez et al., 2015). This process contributes to the binding energy of the OSC/dopant pair and therefore represents a driving force toward the formation of such complexes (Salzmann et al., 2012a). Apart from shifts of the diagnostic vibrational bands in FTIR that are lower than in the IPA case, CPX formation can be deduced from UV-vis/NIR spectroscopy, where the characteristic absorptions of the IPA (Fig. 11.6A) are not observed but different subgap absorptions arise. They are interpreted as the optical transitions between the supramolecular hybrid orbitals of the CPX and, as ECPX,H and ECPX,L depend on LDOP via Eq. (11.1), the absorptions shift toward lower energy for stronger dopants (i.e., the split is smaller for dopants of higher EA), as shown in Fig. 11.8B; transition energies and dopant EAs there show a linear dependence (M
endez et al., 2013, 2015). Upon doping, such CPX states then become embedded into the organic semiconductor DOS and act as precursor states for the eventual generation of mobile charge carriers. As apparent from Fig. 11.8A, the empty supramolecular hybrid state ECPX,L lies at lower binding energy than the HOMO of the OSC (HOSC). As all available states are occupied according to Fermi-Dirac statistics, only a fraction of these states are then ionized at room temperature—certainly less than if CPX formation had not occurred (β ¼ 0).

11.3.4 Overall electronic structure of doped OSCs for both scenarios These considerations regarding the modifications to the DOS upon molecularly doping OSC now allow one to put together a comprehensive view of the energy levels, the resulting DOS, and its occupation, both for IPA and CPX formation, and for both pand n-doping (Salzmann et al., 2016). The complete emerging picture is summarized in Fig. 11.9, with Gaussian-distributed occupied states of the pristine OSC given to account for energetic disorder (Tietze et al., 2018), under consideration of the reorganization energy λ, and of U1 and U2, the respective Hubbard U of the p/n-dopant (cf. Fig. 11.5). For the case of p-doping, the IE of the CPX is higher and its EA lower than the IE of the OSC. The CPX, therefore, introduces a doubly occupied level below the transport manifold of the OSC and an empty level above (Fig. 11.9, left). For the generation of mobile charges (holes) in the doped OSC by this very process, the CPX needs to be ionized (Fig. 11.9, top left). Note that as the energy levels of the negatively ionized CPX have not yet been experimentally observed, they are approximated by those of the neutral CPX (marked with #), as deduced from UPS/IPES

368

Handbook of Organic Materials for Electronic and Photonic Devices

Doping in organic semiconductors

Fig. 11.9 A comprehensive picture of energy levels and the corresponding DOSs upon molecular p/n-doping for the two alternative scenarios of IPA formation (top and bottom) and the formation of CPXs (left and right) and their subsequent ionization; λ is the reorganization energy and U1 and U2 denote the Hubbard U of p/n-dopant and OSC, respectively; the levels of the ionized CPX (marked with #) are approximated by those of the neutral species. From Salzmann, I., Heimel, G., Oehzelt, M., Winkler, S., Koch, N., 2016. Molecular electrical doping of organic semiconductors: fundamental mechanisms and emerging dopant design rules. Acc. Chem. Res. 49, 370–378. doi:10.1021/acs.accounts.5b00438. Copyright 2016 American Chemical Society.

369

370

Handbook of Organic Materials for Electronic and Photonic Devices

(Fig. 11.8A). In full analogy, the right-center and bottom sketches in Fig. 11.9 depict the corresponding neutral CPX and the ionized scenario for n-type doping. The center sketches illustrate the DOS for IPA formation, the major route toward forming mobile electrons and holes in doped OSCs, center top for p-doping and bottom for n-doping (Salzmann et al., 2016).

11.4

Key role of the microstructure in doping OSCs

As the unoccupied states of the CPX can lie deep within the fundamental gap upon (p-) n-doping (Fig. 11.9), the process of CPX formation in the doping of OSCs is clearly detrimental to mobile charge generation and might explain to some extent the typically low doping efficiencies for OSCs compared to their inorganic counterparts. Moreover, high dopant loadings necessary for maximum conductivity (cf. Fig. 11.2B) deteriorate the packing of crystalline OSCs. Prototypical materials like pentacene soon turn amorphous upon doping with something like F4TCNQ (Kleemann et al., 2012; Salzmann et al., 2012a), which is detrimental to the charge carrier mobility. For increasing the doping efficiency in such systems, the process of CPX formation must be suppressed, as it turns an initially strong dopant (F4TCNQ) into a weak dopant (the CPX). In this view, the resonance integral β becomes the key parameter that governs the degree of energy-level splitting in the CPX (M
endez et al., 2013, 2015; Salzmann et al., 2012a, 2016) that is responsible for the loss in dopant strength due to CPX formation. As the splitting depends on the electronic coupling between OSC and dopant, the coupling plays the key role, which itself depends on their mutual wavefunction overlap. Clearly, the wavefunction overlap subtly depends on the mutual arrangement of the species; that is, the nanostructure of the doped films. Fig. 11.10 shows some examples of doping-related modifications of the thin-film structure, as assessed by X-ray diffraction (XRD) techniques. For thin-film analysis, out-of-plane XRD, also referred to as specular XRD, or Θ-2Θ scans, is employed to investigate periodicities perpendicular to the substrate. For crystalline OSCs like pentacene or oligothiophenes on inert substrates (e.g., the common gate dielectric SiO2), such experiments typically yield peak series corresponding to lattice planes parallel to the substrate. If only one peak series is observed, such as the (h00) series (with the Miller index h ¼ 1, 2, 3, …), the (100) lattice plane is parallel to the substrate and the film is said to be grown in (100) texture, as shown in Fig. 11.10A. There, pentacene is found to grow in two different polymorphs of different lattice spacing, which is apparent by the occurrence of two similar series in the data. As the SiO2 substrate is amorphous (and therefore anisotropic), individual crystalline pentacene grains of lateral sizes up to the μm range (Zhang et al., 2011) are azimuthally statistically distributed (forming a crystalline, 2D powder) and the film is said to be grown in (100) fiber-texture, as all crystallites share a common crystal axis, known as the fiber-axis. For applying more advanced, mainly synchrotron radiation-based, diffraction techniques like grazing-incidence XRD (GIXRD)—a technique frequently employed for the analysis of thin films due to its high surface sensitivity—fiber-

Doping in organic semiconductors 371

Fig. 11.10 (A) Specular XRD (CuKα radiation) on F4TCNQ- and F6TCNNQ-doped pentacene, which turns amorphous upon dopant admixture via vacuum codeposition. (B) Left (top): Specular XRD on C10-BTBT doped with F4TCNQ of increasing dopant loading, where the peaks marked with a square are due to a 1:1 mixed structure. Left (bottom): C10-BTBT 1:1 mixed with the series of differently strong dopants TCNQ, FTCNQ, F2TCNQ, F4TCNQ, and F6TCNNQ, all showing highly similar mixed crystal structures. Right: Illustration of the 1:1 mixed crystal structure of C10-BTBT and F4TCNQ. The structural motif of cofacially packed dopant/OSC molecules forms a 3D structure of parallel mixed stacks. The corresponding unit cell parameters determined by single-crystal XRD are highly similar. (A) From Kleemann, H., Schuenemann, C., Zakhidov, A.A., Riede, M., L€ ussem, B., Leo, K., 2012. Structural phase transition in pentacene caused by molecular doping and its effect on charge carrier mobility. Org. Electron. 13, 58–65. doi:10.1016/j.orgel.2011.09.027. Copyright 2012 Elsevier; (B) Modified from M
endez, H., Heimel, G., Opitz, A., Sauer, K., Barkowski, P., Oehzelt, M., Soeda, J., Okamoto, T., Takeya, J., Arlin, J.-B., Balandier, J.-Y., Geerts, Y., Koch, N., Salzmann, I., 2013. Doping of organic semiconductors: impact of dopant strength and electronic coupling. Angew. Chem. Int. Ed. 52, 7751–7755. doi:10.1002/anie.201302396. Copyright 2013 John Wiley and Sons.

372

Handbook of Organic Materials for Electronic and Photonic Devices

textured growth is mandatory. GIXRD data is typically depicted in the form of 2D maps, so-called reciprocal space maps, which, in fact, represent cuts through reciprocal space. While a single crystal is represented in reciprocal space by a reciprocal lattice that is characteristic for its crystal structure, in fiber-textured films, the reciprocal lattice points of the small, individual, single-crystalline OSC grains degenerate to rings, and a reciprocal space map therefore yields the whole information on the crystal structure from the intersection with these rings (Salzmann et al., 2012b). From an indexation of the reflections in such data, full structure solutions can be derived (Birkholz et al., 2006; Robinson and Tweet, 1992; Salzmann et al., 2011, 2012b). As seen in Fig. 11.10A for F4TCNQ and F6TCNNQ doped pentacene, both (100) peak series are already strongly attenuated at 4 mol% dopant admixture (note the data representation on a logarithmic scale) and the film soon turns amorphous. However, for cases of similar molecular dimensions of the conjugated cores (Vogel et al., 2010), mixed crystal growth can occur, as shown in Fig. 11.10B for the OSC C10-BTBT doped with the whole TCNQ—F4TCNQ series (and the even stronger dopant F6TCNNQ). For all the different combinations, highly similar structures are found with a structural motif of cofacially stacked OSC/dopant pairs that form a mixed-stack crystal structure (Fig. 11.10B, right). This motif allows efficient wavefunction overlap between the frontier molecular orbitals of OSC and dopant, which translates into a resonance integral β of about 0.5 eV. Note that a highly similar structure is found for the 4T/TCNQ—F4TCNQ series, as deduced from GIXRD (M
endez et al., 2015), and also shows CPX formation (cf. Figs. 11.7B and 11.8) with β of similar magnitude. In these cases, 1:1 mixed CPX crystallites are dispersed in the pristine OSC matrix and, therefore, take over the role of the dopant. It is surprising that for essentially all CPs of sufficiently low IE, IPA formation appears to be common (Gao et al., 2013b; Pingel and Neher, 2013; Salzmann et al., 2016; Sosorev and Paraschuk, 2014), while numerous examples for fractional charge transfer have been documented for COMs (M
endez et al., 2013, 2015; Salzmann et al., 2012a, 2016; Yoshida et al., 2016). As suggested previously, it might be related to the ability to closely pack and optimize wavefunction overlap in crystalline COMs, which induces CPX formation, leading to substantial values for β. In a study (Ghani et al., 2015) on a series of F4TCNQ-doped thiophene-based copolymers, however, both processes were reported to occur simultaneously, wherein CPX formation was suggested to occur through local interaction of the dopants with the quinoxaline units in the backbone of the copolymer. In donor-acceptor copolymers, Di Nuzzo et al. (2015) showed that the localization of the dopant is decisive for IPA formation, and p-dopants located close to the acceptor moiety do not contribute to efficient charge transfer. The fact that there is indeed a subtle difference in the microstructure that can decide between IPA and CPX formation has been recently shown for P3HT and F4TCNQ by Jacobs et al. (2018), as illustrated in Fig. 11.11. By modifying the film preparation conditions, they found that structures showing exclusive IPA or CPX formation can be established. If prepared from a common chlorobenzene (CB) solution via spin coating at 60°C, a semicrystalline film grows that has been reported in numerous studies (Duong et al., 2013; Gao et al., 2013b; M
endez et al., 2015), showing a mixed

Doping in organic semiconductors

373

crystal structure of F4TCNQ intercalated between the P3HT chains (green curves in Fig. 11.11A, GIXRD in Fig. 11.11B(3)). However, if prepared at 80°C from CB, both the optical, vibrational, and structural properties of the films fundamentally change and, instead of IPA formation, strong evidence for the formation of CPXs is found (blue curves in Fig. 11.11A, GIXRD in Fig. 11.11B(1)), where no IPA related absorption features are observed in UV-vis/NIR, and where FTIR data allows deducing δ ¼ 0.6; that is, fractional charge transfer for doped P3HT. The microstructure of the film, as assessed by GIXRD, is changed into a significantly more closely packed mixed crystal structure (see the sharp and strong in-plane diffraction feature at Qxy ¼ 1.8 nm
1 that stems from F4TCNQ π-stacked with the P3HT backbone) with intercalated alkyl chains (deduced from shifts of the strong features around Qxy ¼ 0 at higher Qz). This data strongly suggests that the microstructure and the mutual arrangement of OSC and dopant indeed decide the formation of CPXs and that, as CPX formation is detrimental for the formation of mobile charges upon doping, structurally incompatible dopants might be the better choice for efficiently doping OSCs. To ultimately inhibit CPX formation, it has been suggested to increase the steric demand of chemically modified molecular dopants by attaching bulky and inert side groups to their chemical structures, which emerges as design rule from these observations (Salzmann et al., 2012a, 2016; Salzmann and Heimel, 2015).

11.5

Summary and outlook

Today’s information society substantially depends on our ability to dope inorganic semiconductors, such as silicon, in a controllable manner, thereby tuning their electrical properties to application-specific demands. However, for opto-electronic devices, organic semiconductors have emerged as a superior alternative owing to the ease of tuning their optical gap through chemical variability and their potential for low-cost, large-area processing on flexible substrates. This material class has shown a resounding success in novel display technology, where doped organic semiconductors are ubiquitous in commercial applications. Despite this tremendous success, however, intense research efforts are still dedicated to understanding the fundamental mechanisms at work in the electrical doping of organic semiconductors, as doping efficiencies are still low. In this chapter, we juxtaposed the doping of inorganic semiconductors to that of organics, discussed the pathway leading to its success by employing molecular dopants, and reviewed novel approaches in their processing and chemical design. While doping OSCs is, in fact, the oldest topic in the field, which originated in the successful doping of polyacetylene in the late 1970s, major challenges remain to be resolved. Today, the doping of CPs and molecules represents a highly contemporary and dynamic area of research (M€ uller and Salzmann, 2018), where increasing the fundamental understanding can be directly translated into technological progress. We also reviewed the two major competing processes that have been identified to occur upon dopant admixture to the organic semiconductor host: (1) the direct

374 Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 11.11 (A) Top: UV-vis/NIR data of pristine and F4TCNQ-doped P3HT prepared at different temperatures. If processed at 80°C from CB, no IPArelated subgap absorptions are observed, but instead transitions are assigned to CPX formation (blue curve, cf. Fig. 11.8B). Bottom: In FTIR, a lower shift of the C^N stretching modes translates into δ ¼ 0.6; that is, fractional charge transfer. (B) GIXRD data on (1), the sample showing fractional charge transfer, (2), that sample after conversion to δ ¼ 1 via postpreparation solvent exposure, (3), the δ ¼ 1 case directly prepared at lower temperature (60°C), and (4), a pristine P3HT presented as reference. Modified from Jacobs, I.E., Cendra, C., Harrelson, T.F., Bedolla Valdez, Z.I., Faller, R., Salleo, A., Moul
e, A.J., 2018. Polymorphism controls the degree of charge transfer in a molecularly doped semiconducting polymer. Mater. Horiz. doi:10.1039/C8MH00223A. Copyright 2018 The Royal Society of Chemistry.

Doping in organic semiconductors

375

formation of ion pairs through integer charge transfer between the dopant and the host, and (2), the formation of ground-state charge transfer complexes, which need to be ionized to generate mobile charge carriers. These processes entail a significantly different DOS, which overall leads to a comprehensive picture of the electronic structure upon doping. By taking into account the revised perception of the energetics of excess charges in organic semiconductors, the absence of photoemission intensity at the Fermi energy in such doped systems becomes apparent. Parameters like the doping efficiency, as well as the respective doping process at work, must be deduced from techniques like optical and vibrational spectroscopy instead. Recent studies clearly highlight the key role that the local microstructure plays for the process of charge transfer, for the subsequent carrier release and, in particular, for the doping efficiency. It has been shown that the ability of OSC and dopant species to electronically couple decides over the formation of ground-state charge transfer complexes, a process clearly detrimental to the doping efficiency. As the coupling depends on the overlap in wavefunctions between organic semiconductor and molecular dopant, structures where their conjugated cores cannot come into intimate contact must be regarded as superior for efficiently generating mobile charges. This represents an important guideline for the chemical design of the future—more efficient molecular dopants, where spatially separating the dopants from the semiconductor species could be achieved by adding side groups of high steric demand to the functional dopant cores. Obviously, progress in the field depends to a large extent on progress in the synthesis of novel doping agents and organic semiconductor materials. Employing bulkier dopants not only is beneficial to suppress electronic coupling, but also further inhibits dopant diffusion and reduces their volatility, which are key aspects for establishing stable bipolar devices employing p-i-n and p-n junctions. For deeper insight into the field, we refer the inclined reader to recent general reviews on molecular electrical doping of organic semiconductors (e.g., Jacobs and Moul
e, 2017; Salzmann et al., 2016; L€ ussem et al., 2013a), and to an applicationrelated review focusing on molecular dopants employed in OTFTs (L€ussem et al., 2016) as starting points to delve further into this topic.

Appendix. Compound abbreviations OSC

Full name

4T C10-BTBT C60 ICTBT Ir(piq)3 MeO-TPD P3HT

Quarterthiophene 2,7-didecyl[1]benzothieno[3,2-b][1]benzothiophene C60 fullerene Indacenodithiophene-cobenzothiadiazole Tris[1-phenylisoquinoline-C2,N]iridium(III) N,N,N0 ,N0 -tetrakis(4-methoxyphenyl)benzidine Poly(3-hexylthiophene-2,5-diyl) Continued

376

Handbook of Organic Materials for Electronic and Photonic Devices

Continued OSC

Full name

p(g42T-T) POPy2 ZnPc

n/a (a polythiophene with oligo-ethylene glycol side chains) Phenyldi(pyren-2-yl)phosphine oxide Zinc phthalocyanine

Dopant

Full name

ABN NOBF4 NOPF6 [RuCp∗ Mes]2

4-Aminobenzonitrile Nitrosyl tetrafluoroborate Nitrosyl hexafluorophosphate (Pentamethylcyclopentadienyl)(1,3,5-trimethylbenzene) ruthenium dimer 7,7,8,8-tetracyanoquinodimethane 2-fluoro-7,7,8,8-tetracyanoquinodimethane 2,5-difluoro-7,7,8,8-tetracyanoquinodimethane 2,3,5,6-tetrafluoro-7,7,8,8-tetracyanoquinodimethane Methyl 2-cyano-2-[4-(dicyanomethylene)-2,3,5,6tetrafluorocyclohexa-2,5-dien-1-ylidene]acetate Octyl 2-cyano-2-[4-(dicyanomethylene)-2,3,5,6tetrafluorocyclohexa-2,5-dien-1-ylidene]acetate Dimethyl 2,20 -(2,3,5,6-tetrafluorocyclohexa-2,5-diene-1,4diylidene)-bis(cyanoacetate) Dioctyl 2,20 -(2,3,5,6-tetrafluorocyclohexa-2,5-diene-1,4diylidene)-bis(cyanoacetate) 2,2-(perfluoronaphthalene-2,6-diylidene)dimalononitrile

TCNQ FTCNQ F2TCNQ F4TCNQ F4MCTCNQ F4OCTCNQ F4DMCDCNQ F4DOCDCNQ F6TCNNQ

References Ameri, T., Dennler, G., Lungenschmied, C., Brabec, C.J., 2009. Organic tandem solar cells: a review. Energy Environ. Sci. 2, 347. https://doi.org/10.1039/b817952b. Basescu, N., Liu, Z.-X., Moses, D., Heeger, A.J., Naarmann, H., Theophilou, N., 1987. High electrical conductivity in doped polyacetylene. Nature 327, 403–405. https://doi.org/ 10.1038/327403a0. B€assler, H., 1993. Charge transport in disordered organic photoconductors a monte carlo simulation study. Phys. Status Solidi B 175, 15–56. https://doi.org/10.1002/pssb.2221750102. B€assler, H., K€ohler, A., 2011. Charge transport in organic semiconductors. In: Metzger, R.M. (Ed.), Unimolecular and Supramolecular Electronics. In: I, Springer, Berlin, Heidelberg, pp. 1–65. https://doi.org/10.1007/128_2011_218. Bernstein, J., 2002. Polymorphism in molecular crystals. In: International Union of Crystallography monographs on crystallography. Oxford University Press; Clarendon Press, Oxford; New York. Birkholz, M., Fewster, P.F., Genzel, C., 2006. Thin Film Analysis by X-ray Scattering. WileyVCH, Weinheim. Br
edas, J.L., Street, G.B., 1985. Polarons, bipolarons, and solitons in conducting polymers. Acc. Chem. Res. 18, 309–315. https://doi.org/10.1021/ar00118a005.

Doping in organic semiconductors

377

Br€ oker, B., Blum, R.-P., Frisch, J., Vollmer, A., Hofmann, O.T., Rieger, R., M€ ullen, K., Rabe, J.P., Zojer, E., Koch, N., 2008. Gold work function reduction by 2.2eV with an air-stable molecular donor layer. Appl. Phys. Lett. 93, 243303. https://doi.org/ 10.1063/1.3049616. Bubnova, O., Khan, Z.U., Wang, H., Braun, S., Evans, D.R., Fabretto, M., Hojati-Talemi, P., Dagnelund, D., Arlin, J.-B., Geerts, Y.H., Desbief, S., Breiby, D.W., Andreasen, J.W., Lazzaroni, R., Chen, W.M., Zozoulenko, I., Fahlman, M., Murphy, P.J., Berggren, M., Crispin, X., 2014. Semi-metallic polymers. Nat. Mater. 13, 190–194. https://doi.org/ 10.1038/nmat3824. Busch, G., 1989. Early history of the physics and chemistry of semiconductors-from doubts to fact in a hundred years. Eur. J. Phys. 10, 254–264. https://doi.org/10.1088/0143-0807/10/ 4/002. Chamberlain, G.A., 1983. Organic solar cells: a review. Sol. Cells 8, 47–83. https://doi.org/ 10.1016/0379-6787(83)90039-X. Chappell, J.S., Bloch, A.N., Bryden, W.A., Maxfield, M., Poehler, T.O., Cowan, D.O., 1981. Degree of charge transfer in organic conductors by infrared absorption spectroscopy. J. Am. Chem. Soc. 103, 2442–2443. https://doi.org/10.1021/ja00399a066. Chen, T.-A., Wu, X., Rieke, R.D., 1995. Regiocontrolled synthesis of poly(3-alkylthiophenes) mediated by Rieke zinc: their characterization and solid-state properties. J. Am. Chem. Soc. 117, 233–244. https://doi.org/10.1021/ja00106a027. Chew, A.R., Ghosh, R., Shang, Z., Spano, F.C., Salleo, A., 2017. Sequential doping reveals the importance of amorphous chain rigidity in charge transport of semi-crystalline polymers. J. Phys. Chem. Lett. 8, 4974–4980. https://doi.org/10.1021/acs.jpclett.7b01989. Chiang, C.K., Fincher, C.R., Park, Y.W., Heeger, A.J., Shirakawa, H., Louis, E.J., Gau, S.C., MacDiarmid, A.G., 1977. Electrical conductivity in doped polyacetylene. Phys. Rev. Lett. 39, 1098–1101. https://doi.org/10.1103/PhysRevLett.39.1098. Choi, S., Fuentes-Hernandez, C., Wang, C.-Y., Khan, T.M., Larrain, F.A., Zhang, Y., Barlow, S., Marder, S.R., Kippelen, B., 2016. A study on reducing contact resistance in solution-processed organic field-effect transistors. ACS Appl. Mater. Interfaces 8, 24744–24752. https://doi.org/10.1021/acsami.6b07029. Datta, S., Grant, D.J.W., 2004. Crystal structures of drugs: advances in determination, prediction and engineering. Nat. Rev. Drug Discov. 3, 42–57. https://doi.org/10.1038/ nrd1280. Di Nuzzo, D., Fontanesi, C., Jones, R., Allard, S., Dumsch, I., Scherf, U., von Hauff, E., Schumacher, S., Da Como, E., 2015. How intermolecular geometrical disorder affects the molecular doping of donor–acceptor copolymers. Nat. Commun. 6. https://doi.org/ 10.1038/ncomms7460. Duong, D.T., Wang, C., Antono, E., Toney, M.F., Salleo, A., 2013. The chemical and structural origin of efficient p-type doping in P3HT. Org. Electron. 14, 1330–1336. https://doi.org/ 10.1016/j.orgel.2013.02.028. Fratini, S., Mayou, D., Ciuchi, S., 2016. The transient localization scenario for charge transport in crystalline organic materials. Adv. Funct. Mater. 26, 2292–2315. https://doi.org/ 10.1002/adfm.201502386. Gao, W., Kahn, A., 2001. Controlled p-doping of zinc phthalocyanine by coevaporation with tetrafluorotetracyanoquinodimethane: a direct and inverse photoemission study. Appl. Phys. Lett. 79, 4040–4042. https://doi.org/10.1063/1.1424067. Gao, W., Kahn, A., 2003a. Controlled p doping of the hole-transport molecular material N,N0 diphenyl-N,N0 -bis(1-naphthyl)-1,10 -biphenyl-4,40 -diamine with tetrafluorotetracyanoquinodimethane. J. Appl. Phys. 94, 359–366. https://doi.org/10.1063/1.1577400.

378

Handbook of Organic Materials for Electronic and Photonic Devices

Gao, W., Kahn, A., 2003b. Effect of electrical doping on molecular level alignment at organic– organic heterojunctions. Appl. Phys. Lett. 82, 4815–4817. https://doi.org/10.1063/ 1.1585123. Gao, W., Kahn, A., 2003c. Electrical doping: the impact on interfaces of π-conjugated molecular films. J. Phys. Condens. Matter 15, S2757–S2770. https://doi.org/10.1088/09538984/15/38/014. Gao, J., Niles, E.T., Grey, J.K., 2013a. Aggregates promote efficient charge transfer doping of poly(3-hexylthiophene). J. Phys. Chem. Lett. 4, 2953–2957. https://doi.org/10.1021/ jz401555x. Gao, J., Roehling, J.D., Li, Y., Guo, H., Moul
e, A.J., Grey, J.K., 2013b. The effect of 2,3,5,6tetrafluoro-7,7,8,8-tetracyanoquinodimethane charge transfer dopants on the conformation and aggregation of poly(3-hexylthiophene). J. Mater. Chem. C 1, 5638. https://doi.org/ 10.1039/c3tc31047g. Ghani, F., Opitz, A., Pingel, P., Heimel, G., Salzmann, I., Frisch, J., Neher, D., Tsami, A., Scherf, U., Koch, N., 2015. Charge transfer in and conductivity of molecularly doped thiophene-based copolymers. J. Polym. Sci. Part B Polym. Phys. 53, 58–63. https://doi. org/10.1002/polb.23631. Harada, K., Sumino, M., Adachi, C., Tanaka, S., Miyazaki, K., 2010. Improved thermoelectric performance of organic thin-film elements utilizing a bilayer structure of pentacene and 2,3,5,6-tetrafluoro-7,7,8,8-tetracyanoquinodimethane (F4-TCNQ). Appl. Phys. Lett. 96, 253304. https://doi.org/10.1063/1.3456394. Haworth, N.L., Lu, J., Vo, N., Le, T.H., Thompson, C.D., Bond, A.M., Martin, L.L., 2014. Diagnosis of the redox levels of TCNQF4 compounds using vibrational spectroscopy. ChemPlusChem 79, 962–972. https://doi.org/10.1002/cplu.201402013. He, G., Walzer, K., Pfeiffer, M.P., Leo, K., Pudzich, R., Salbeck, J., 2004. Ultra-high-efficiency electrophosphorescent p-i-n OLEDs with double emission layers. In: Kafafi, Z.H., Lane, P.A. (Eds.), Proc. SPIE 5519, Organic Light-Emitting Materials and Devices VIII. p. 42. https://doi.org/10.1117/12.563310. Heeger, A.J., 2001. Semiconducting and metallic polymers: the fourth generation of polymeric materials (nobel lecture). Angew. Chem. Int. Ed. 40, 2591–2611. https://doi.org/10. 1002/1521-3773(20010716)40:14< 2591::AID-ANIE2591 > 3.0.CO;2-0. Heeger, A.J., Kivelson, S., Schrieffer, J.R., Su, W.-P., 1988. Solitons in conducting polymers. Rev. Mod. Phys. 60, 781–850. https://doi.org/10.1103/RevModPhys.60.781. Heimel, G., 2016. The optical signature of charges in conjugated polymers. ACS Cent. Sci. 2, 309–315. https://doi.org/10.1021/acscentsci.6b00073. Heimel, G., Salzmann, I., Duhm, S., Rabe, J.P., Koch, N., 2009. Intrinsic surface dipoles control the energy levels of conjugated polymers. Adv. Funct. Mater. 19, 3874–3879. https://doi. org/10.1002/adfm.200901025. Heimel, G., Salzmann, I., Duhm, S., Koch, N., 2011. Design of organic semiconductors from molecular electrostatics. Chem. Mater. 23, 359–377. https://doi.org/10.1021/cm1021257. Heimel, G., Duhm, S., Salzmann, I., Gerlach, A., Strozecka, A., Niederhausen, J., B€ urker, C., Hosokai, T., Fernandez-Torrente, I., Schulze, G., Winkler, S., Wilke, A., Schlesinger, R., Frisch, J., Br€oker, B., Vollmer, A., Detlefs, B., Pflaum, J., Kera, S., Franke, K.J., Ueno, N., Pascual, J.I., Schreiber, F., Koch, N., 2013. Charged and metallic molecular monolayers through surface-induced aromatic stabilization. Nat. Chem. 5, 187–194. https://doi.org/ 10.1038/nchem.1572. Hein, M.P., Zakhidov, A.A., L€ussem, B., Jankowski, J., Tietze, M.L., Riede, M.K., Leo, K., 2014. Molecular doping for control of gate bias stress in organic thin film transistors. Appl. Phys. Lett. 104, 013507. https://doi.org/10.1063/1.4861168.

Doping in organic semiconductors

379

Horowitz, G., 1998. Organic field-effect transistors. Adv. Mater. 10, 365–377. https://doi.org/ 10.1002/(SICI)1521-4095(199803)10:5 < 365::AID-ADMA365 >3.0.CO;2-U. Horowitz, G., 2015. Validity of the concept of band edge in organic semiconductors. J. Appl. Phys. 118, 115502. https://doi.org/10.1063/1.4931061. Horowitz, G., Fichou, D., Peng, X., Xu, Z., Garnier, F., 1989. A field-effect transistor based on conjugated alpha-sexithienyl. Solid State Commun. 72, 381–384. https://doi.org/ 10.1016/0038-1098(89)90121-X. Hubbard, J., 1963. Electron correlations in narrow energy bands. Proc. R. Soc. Math. Phys. Eng. Sci. 276, 238–257. https://doi.org/10.1098/rspa.1963.0204. H€ ufner, S., 2003. Photoelectron spectroscopy: principles and applications. In: Advanced Texts in Physics. third ed. Springer, Berlin, New York. Hynynen, J., Kiefer, D., Yu, L., Kroon, R., Munir, R., Amassian, A., Kemerink, M., M€ uller, C., 2017. Enhanced electrical conductivity of molecularly p-doped poly(3-hexylthiophene) through understanding the correlation with solid-state order. Macromolecules 50, 8140–8148. https://doi.org/10.1021/acs.macromol.7b00968. Jacobs, I.E., Moul
e, A.J., 2017. Controlling molecular doping in organic semiconductors. Adv. Mater. 29, 1703063. https://doi.org/10.1002/adma.201703063. Jacobs, I.E., Li, J., Burg, S.L., Bilsky, D.J., Rotondo, B.T., Augustine, M.P., Stroeve, P., Moul
e, A.J., 2015. Reversible optical control of conjugated polymer solubility with sub-micrometer resolution. ACS Nano 9, 1905–1912. https://doi.org/10.1021/nn506820d. Jacobs, I.E., Aasen, E.W., Oliveira, J.L., Fonseca, T.N., Roehling, J.D., Li, J., Zhang, G., Augustine, M.P., Mascal, M., Moul
e, A.J., 2016. Comparison of solution-mixed and sequentially processed P3HT:F4TCNQ films: effect of doping-induced aggregation on film morphology. J. Mater. Chem. C 4, 3454–3466. https://doi.org/10.1039/C5TC04207K. Jacobs, I.E., Cendra, C., Harrelson, T.F., Bedolla Valdez, Z.I., Faller, R., Salleo, A., Moul
e, A.J., 2018. Polymorphism controls the degree of charge transfer in a molecularly doped semiconducting polymer. Mater. Horiz. https://doi.org/10.1039/C8MH00223A. € Jiang, X., An, C., Osterbacka, R., Vardeny, Z., 2001. FTIR studies of charged photoexcitations in regio-regular and regio-random poly(3-alkylthiophene) films. Synth. Met. 116, 203–206. https://doi.org/10.1016/S0379-6779(00)00487-2. Jones, A.O.F., Chattopadhyay, B., Geerts, Y.H., Resel, R., 2016. Substrate-induced and thinfilm phases: polymorphism of organic materials on surfaces. Adv. Funct. Mater. 26, 2233–2255. https://doi.org/10.1002/adfm.201503169. Jurchescu, O.D., Mourey, D.A., Subramanian, S., Parkin, S.R., Vogel, B.M., Anthony, J.E., Jackson, T.N., Gundlach, D.J., 2009. Effects of polymorphism on charge transport in organic semiconductors. Phys. Rev. B. 80https://doi.org/10.1103/PhysRevB.80.085201. Kampar, E., Neilands, O., 1986. Degree of charge transfer in donor–acceptor systems of the π–π type. Russ. Chem. Rev. 55, 334–342. https://doi.org/10.1070/ RC1986v055n04ABEH003193. Kanai, K., Akaike, K., Koyasu, K., Sakai, K., Nishi, T., Kamizuru, Y., Nishi, T., Ouchi, Y., Seki, K., 2009. Determination of electron affinity of electron accepting molecules. Appl. Phys. A 95, 309–313. https://doi.org/10.1007/s00339-008-5021-1. Kanemoto, K., Muramatsu, K., Baba, M., Yamauchi, J., 2008. Polaron dynamics of heavily doped regioregular and regiorandom poly(3-alkylthiophenes) revealed by electron spin resonance spectroscopy. J. Phys. Chem. B 112, 10922–10926. https://doi.org/10.1021/ jp804065k. Kleemann, H., Schuenemann, C., Zakhidov, A.A., Riede, M., L€ ussem, B., Leo, K., 2012. Structural phase transition in pentacene caused by molecular doping and its effect on charge carrier mobility. Org. Electron. 13, 58–65. https://doi.org/10.1016/j.orgel.2011.09.027.

380

Handbook of Organic Materials for Electronic and Photonic Devices

Koch, N., Duhm, S., Rabe, J.P., Vollmer, A., Johnson, R.L., 2005. Optimized hole injection with strong electron acceptors at organic-metal interfaces. Phys. Rev. Lett. 95. https://doi.org/ 10.1103/PhysRevLett.95.237601. Kroon, R., Kiefer, D., Stegerer, D., Yu, L., Sommer, M., M€ uller, C., 2017. Polar side chains enhance processability, electrical conductivity, and thermal stability of a molecularly p-doped polythiophene. Adv. Mater. 29, 1700930. https://doi.org/10.1002/adma. 201700930. Li, J., Zhang, G., Holm, D.M., Jacobs, I.E., Yin, B., Stroeve, P., Mascal, M., Moul
e, A.J., 2015. Introducing solubility control for improved organic P-type dopants. Chem. Mater. 27, 5765–5774. https://doi.org/10.1021/acs.chemmater.5b02340. Li, H., DeCoster, M.E., Ireland, R.M., Song, J., Hopkins, P.E., Katz, H.E., 2017. Modification of the poly(bisdodecylquaterthiophene) structure for high and predominantly nonionic conductivity with matched dopants. J. Am. Chem. Soc. 139, 11149–11157. https://doi.org/ 10.1021/jacs.7b05300. Lim, E., Peterson, K.A., Su, G.M., Chabinyc, M.L., 2018. Thermoelectric properties of poly(3hexylthiophene) (P3HT) doped with 2,3,5,6-tetrafluoro-7,7,8,8-tetracyanoquinodimethane (F4 TCNQ) by vapor-phase infiltration. Chem. Mater. 30, 998–1010. https:// doi.org/10.1021/acs.chemmater.7b04849. Lin, X., Wegner, B., Lee, K.M., Fusella, M.A., Zhang, F., Moudgil, K., Rand, B.P., Barlow, S., Marder, S.R., Koch, N., Kahn, A., 2017. Beating the thermodynamic limit with photoactivation of n-doping in organic semiconductors. Nat. Mater. 16, 1209–1215. https:// doi.org/10.1038/nmat5027. L€ ogdlund, M., Lazzaroni, R., Stafstr€om, S., Salaneck, W.R., Br
edas, J.-L., 1989. Direct observation of charge-induced π-electronic structural changes in a conjugated polymer. Phys. Rev. Lett. 63, 1841–1844. https://doi.org/10.1103/PhysRevLett.63.1841. L€ ussem, B., Riede, M., Leo, K., 2013a. Doping of organic semiconductors. Phys. Status Solidi A 210, 9–43. https://doi.org/10.1002/pssa.201228310. L€ ussem, B., Tietze, M.L., Kleemann, H., Hoßbach, C., Bartha, J.W., Zakhidov, A., Leo, K., 2013b. Doped organic transistors operating in the inversion and depletion regime. Nat. Commun. 4. https://doi.org/10.1038/ncomms3775. L€ ussem, B., Keum, C.-M., Kasemann, D., Naab, B., Bao, Z., Leo, K., 2016. Doped organic transistors. Chem. Rev. 116, 13714–13751. https://doi.org/10.1021/acs.chemrev.6b00329. Ma, L., Lee, W.H., Park, Y.D., Kim, J.S., Lee, H.S., Cho, K., 2008. High performance polythiophene thin-film transistors doped with very small amounts of an electron acceptor. Appl. Phys. Lett. 92, 063310. https://doi.org/10.1063/1.2883927. Machida, S., Nakayama, Y., Duhm, S., Xin, Q., Funakoshi, A., Ogawa, N., Kera, S., Ueno, N., Ishii, H., 2010. Highest-occupied-molecular-orbital band dispersion of rubrene single crystals as observed by angle-resolved ultraviolet photoelectron spectroscopy. Phys. Rev. Lett. 104. https://doi.org/10.1103/PhysRevLett.104.156401. Madelung, O., 1999. Schottky-Spenke-Welker: Erinnerungen an die “Gr€ underjahre” der Halbleiterphysik in Deutschland nach dem Zweiten Weltkrieg. Phys. J. 55, 54–58. https:// doi.org/10.1002/phbl.19990550614. Maennig, B., Drechsel, J., Gebeyehu, D., Simon, P., Kozlowski, F., Werner, A., Li, F., Grundmann, S., Sonntag, S., Koch, M., Leo, K., Pfeiffer, M., Hoppe, H., Meissner, D., Sariciftci, N.S., Riedel, I., Dyakonov, V., Parisi, J., 2004. Organic p-i-n solar cells. Appl. Phys. A 79, 1–14. https://doi.org/10.1007/s00339-003-2494-9. Melby, L.R., Harder, R.J., Hertler, W.R., Mahler, W., Benson, R.E., Mochel, W.E., 1962. Substituted quinodimethans. II. Anion-radical derivatives and complexes of 7,7,8,

Doping in organic semiconductors

381

8-tetracyanoquinodimethan. J. Am. Chem. Soc. 84, 3374–3387. https://doi.org/10.1021/ ja00876a029. M
endez, H., Heimel, G., Opitz, A., Sauer, K., Barkowski, P., Oehzelt, M., Soeda, J., Okamoto, T., Takeya, J., Arlin, J.-B., Balandier, J.-Y., Geerts, Y., Koch, N., Salzmann, I., 2013. Doping of organic semiconductors: impact of dopant strength and electronic coupling. Angew. Chem. Int. Ed. 52, 7751–7755. https://doi.org/10.1002/ anie.201302396. M
endez, H., Heimel, G., Winkler, S., Frisch, J., Opitz, A., Sauer, K., Wegner, B., Oehzelt, M., R€othel, C., Duhm, S., T€obbens, D., Koch, N., Salzmann, I., 2015. Charge-transfer crystallites as molecular electrical dopants. Nat. Commun. 6. https://doi.org/10.1038/ ncomms9560. Meneghetti, M., Pecile, C., 1986. Charge–transfer organic crystals: molecular vibrations and spectroscopic effects of electron–molecular vibration coupling of the strong electron acceptor TCNQF 4. J. Chem. Phys. 84, 4149–4162. https://doi.org/10.1063/1.450086. M€ uller, C., Salzmann, I., 2018. MRS fall 2017 symposium: organic semiconductors—surface, interface, bulk doping, and charge transport. Chem. Mater. 30, 3151–3154. https://doi.org/ 10.1021/acs.chemmater.8b01683. Nikolka, M., Nasrallah, I., Rose, B., Ravva, M.K., Broch, K., Sadhanala, A., Harkin, D., Charmet, J., Hurhangee, M., Brown, A., Illig, S., Too, P., Jongman, J., McCulloch, I., Bredas, J.-L., Sirringhaus, H., 2017. High operational and environmental stability of high-mobility conjugated polymer field-effect transistors through the use of molecular additives. Nat. Mater. 16, 356–362. https://doi.org/10.1038/nmat4785. Oehzelt, M., Koch, N., Heimel, G., 2014. Organic semiconductor density of states controls the energy level alignment at electrode interfaces. Nat. Commun.. 5https://doi.org/10.1038/ ncomms5174. Olthof, S., Meerheim, R., Schober, M., Leo, K., 2009a. Energy level alignment at the interfaces in a multilayer organic light-emitting diode structure. Phys. Rev. B. 79https://doi.org/ 10.1103/PhysRevB.79.245308. Olthof, S., Tress, W., Meerheim, R., L€ussem, B., Leo, K., 2009b. Photoelectron spectroscopy study of systematically varied doping concentrations in an organic semiconductor layer using a molecular p-dopant. J. Appl. Phys. 106, 103711. https://doi.org/ 10.1063/1.3259436. € Osterbacka, R., An, C.P., Jiang, X.M., Vardeny, Z.V., 2000. Two-dimensional electronic excitations in self-assembled conjugated polymer nanocrystals. Science 287, 839. https://doi. org/10.1126/science.287.5454.839. Parthasarathy, G., Shen, C., Kahn, A., Forrest, S.R., 2001. Lithium doping of semiconducting organic charge transport materials. J. Appl. Phys. 89, 4986–4992. https://doi.org/ 10.1063/1.1359161. Pingel, P., Neher, D., 2013. Comprehensive picture of p-type doping of P3HT with the molecular acceptor F4 TCNQ. Phys. Rev. B. 87https://doi.org/10.1103/PhysRevB.87.115209. Pingel, P., Zhu, L., Park, K.S., Vogel, J.-O., Janietz, S., Kim, E.-G., Rabe, J.P., Br
edas, J.-L., Koch, N., 2010. Charge-transfer localization in molecularly doped thiophene-based donor polymers. J. Phys. Chem. Lett. 1, 2037–2041. https://doi.org/10.1021/jz100492c. Png, R.-Q., Ang, M.C.Y., Teo, M.-H., Choo, K.-K., Tang, C.G., Belaineh, D., Chua, L.-L., Ho, P.K.H., 2016. Madelung and Hubbard interactions in polaron band model of doped organic semiconductors. Nat. Commun. 7, 11948. https://doi.org/10.1038/ncomms11948. Pope, M., Kallmann, H.P., Magnante, P., 1963. Electroluminescence in organic crystals. J. Chem. Phys. 38, 2042–2043. https://doi.org/10.1063/1.1733929.

382

Handbook of Organic Materials for Electronic and Photonic Devices

Reineke, S., Lindner, F., Schwartz, G., Seidler, N., Walzer, K., L€ ussem, B., Leo, K., 2009. White organic light-emitting diodes with fluorescent tube efficiency. Nature 459, 234–238. https://doi.org/10.1038/nature08003. Richter, L.J., DeLongchamp, D.M., Amassian, A., 2017. Morphology development in solutionprocessed functional organic blend films: an in situ viewpoint. Chem. Rev. 117, 6332–6366. https://doi.org/10.1021/acs.chemrev.6b00618. Robinson, I.K., Tweet, D.J., 1992. Surface X-ray diffraction. Rep. Prog. Phys. 55, 599–651. https://doi.org/10.1088/0034-4885/55/5/002. Rodrigues, A., Castro, M.C.R., Farinha, A.S.F., Oliveira, M., Tom
e, J.P.C., Machado, A.V., Raposo, M.M.M., Hilliou, L., Bernardo, G., 2013. Thermal stability of P3HT and P3HT:PCBM blends in the molten state. Polym. Test. 32, 1192–1201. https://doi.org/ 10.1016/j.polymertesting.2013.07.008. Rodrı´guez-Spong, B., 2004. General principles of pharmaceutical solid polymorphism A supramolecular perspective. Adv. Drug Deliv. Rev. 56, 241–274. https://doi.org/ 10.1016/j.addr.2003.10.005. Salzillo, T., Masino, M., Kociok-K€ohn, G., Di Nuzzo, D., Venuti, E., Della Valle, R.G., Vanossi, D., Fontanesi, C., Girlando, A., Brillante, A., Da Como, E., 2016. Structure, stoichiometry, and charge transfer in cocrystals of perylene with TCNQ-F x. Cryst. Growth Des. 16, 3028–3036. https://doi.org/10.1021/acs.cgd.5b01663. Salzmann, I., Heimel, G., 2015. Toward a comprehensive understanding of molecular doping organic semiconductors (review). J. Electron Spectrosc. Relat. Phenom. 204, 208–222. https://doi.org/10.1016/j.elspec.2015.05.001. Salzmann, I., Nabok, D., Oehzelt, M., Duhm, S., Moser, A., Heimel, G., Puschnig, P., Ambrosch-Draxl, C., Rabe, J.P., Koch, N., 2011. Structure solution of the 6,13pentacenequinone surface-induced polymorph by combining X-ray diffraction reciprocal-space mapping and theoretical structure modeling. Cryst. Growth Des. 11, 600–606. https://doi.org/10.1021/cg1015143. Salzmann, I., Heimel, G., Duhm, S., Oehzelt, M., Pingel, P., George, B.M., Schnegg, A., Lips, K., Blum, R.-P., Vollmer, A., Koch, N., 2012a. Intermolecular hybridization governs molecular electrical doping. Phys. Rev. Lett. 108. https://doi.org/10.1103/ PhysRevLett.108.035502. Salzmann, I., Moser, A., Oehzelt, M., Breuer, T., Feng, X., Juang, Z.-Y., Nabok, D., Della Valle, R.G., Duhm, S., Heimel, G., Brillante, A., Venuti, E., Bilotti, I., Christodoulou, C., Frisch, J., Puschnig, P., Draxl, C., Witte, G., M€ ullen, K., Koch, N., 2012b. Epitaxial growth of π-stacked perfluoropentacene on graphene-coated quartz. ACS Nano 6, 10874–10883. https://doi.org/10.1021/nn3042607. Salzmann, I., Heimel, G., Oehzelt, M., Winkler, S., Koch, N., 2016. Molecular electrical doping of organic semiconductors: fundamental mechanisms and emerging dopant design rules. Acc. Chem. Res. 49, 370–378. https://doi.org/10.1021/acs.accounts.5b00438. Shirakawa, H., Louis, E.J., MacDiarmid, A.G., Chiang, C.K., Heeger, A.J., 1977. Synthesis of electrically conducting organic polymers: halogen derivatives of polyacetylene, (CH)x. J. Chem. Soc. Chem. Commun. 578. https://doi.org/10.1039/c39770000578. Sosorev, A.Y., Paraschuk, D.Y., 2014. Charge-transfer complexes of conjugated polymers. Isr. J. Chem. 54, 650–673. https://doi.org/10.1002/ijch.201400063. Steinm€uller, D., Ramsey, M.G., Netzer, F.P., 1993. Polaron and bipolaronlike states in n-doped bithiophene. Phys. Rev. B 47, 13323–13329. https://doi.org/10.1103/PhysRevB.47.13323. Sze, S.M., 1981. Physics of Semiconductor Devices, second ed. Wiley, New York. Tang, C.W., 1986. Two-layer organic photovoltaic cell. Appl. Phys. Lett. 48, 183–185. https:// doi.org/10.1063/1.96937.

Doping in organic semiconductors

383

Tang, C.W., Albrecht, A.C., 1975. Photovoltaic effects of metal–chlorophyll-a–metal sandwich cells. J. Chem. Phys. 62, 2139–2149. https://doi.org/10.1063/1.430780. Tang, C.W., VanSlyke, S.A., 1987. Organic electroluminescent diodes. Appl. Phys. Lett. 51, 913–915. https://doi.org/10.1063/1.98799. Tietze, M.L., Benduhn, J., Pahner, P., Nell, B., Schwarze, M., Kleemann, H., Krammer, M., Zojer, K., Vandewal, K., Leo, K., 2018. Elementary steps in electrical doping of organic semiconductors. Nat. Commun. 9. https://doi.org/10.1038/s41467-018-03302-z. Vogel, J.-O., Salzmann, I., Duhm, S., Oehzelt, M., Rabe, J.P., Koch, N., 2010. Phase-separation and mixing in thin films of co-deposited rod-like conjugated molecules. J. Mater. Chem. 20, 4055. https://doi.org/10.1039/b927594k. von Meyenn, K., 1985. 1930-1939, Wolfgang Pauli. Wissenschaftlicher Briefwechsel mit Bohr, Einstein, Heisenberg u. Springer, New York, NY. Walzer, K., Maennig, B., Pfeiffer, M., Leo, K., 2007. Highly efficient organic devices based on electrically doped transport layers. Chem. Rev. 107, 1233–1271. https://doi.org/10.1021/ cr050156n. Wang, C.-Y., Fuentes-Hernandez, C., Chou, W.-F., Kippelen, B., 2017. Top-gate organic fieldeffect transistors fabricated on paper with high operational stability. Org. Electron. 41, 340–344. https://doi.org/10.1016/j.orgel.2016.11.026. Wilson, A.H., 1931a. the theory of electronic semi-conductors. Proc. R. Soc. Math. Phys. Eng. Sci. 133, 458–491. https://doi.org/10.1098/rspa.1931.0162. Wilson, A.H., 1931b. The theory of electronic semi-conductors. II. Proc. R. Soc. Math. Phys. Eng. Sci. 134, 277–287. https://doi.org/10.1098/rspa.1931.0196. Winkler, S., Amsalem, P., Frisch, J., Oehzelt, M., Heimel, G., Koch, N., 2015. Probing the energy levels in hole-doped molecular semiconductors. Mater. Horiz. 2, 427–433. https://doi.org/10.1039/C5MH00023H. Yamamoto, Y., Yoshino, K., Inuishi, Y., 1979. Electrical properties of phthalocyanine-halogen complexes. J. Phys. Soc. Jpn. 47, 1887–1891. https://doi.org/10.1143/JPSJ.47.1887. Yim, K.-H., Whiting, G.L., Murphy, C.E., Halls, J.J.M., Burroughes, J.H., Friend, R.H., Kim, J.-S., 2008. Controlling electrical properties of conjugated polymers via a solution-based p-type doping. Adv. Mater. 20, 3319–3324. https://doi.org/10.1002/ adma.200800735. Yoshida, Y., Isomura, K., Kumagai, Y., Maesato, M., Kishida, H., Mizuno, M., Saito, G., 2016. Coronene-based charge-transfer complexes. J. Phys. Condens. Matter 28, 304001. https:// doi.org/10.1088/0953-8984/28/30/304001. Zhang, J., Wu, Y., Duhm, S., Rabe, J.P., Rudolf, P., Koch, N., 2011. Formation of intra-island grain boundaries in pentacene monolayers. Phys. Chem. Chem. Phys. 13, 21102. https:// doi.org/10.1039/c1cp21506j. Zhang, Y., Zhou, H., Seifter, J., Ying, L., Mikhailovsky, A., Heeger, A.J., Bazan, G.C., Nguyen, T.-Q., 2013. Molecular doping enhances photoconductivity in polymer bulk heterojunction solar cells. Adv. Mater. 25, 7038–7044. https://doi.org/10.1002/ adma.201302159. Zhang, Y., Elawad, M., Yu, Z., Jiang, X., Lai, J., Sun, L., 2016. Enhanced performance of perovskite solar cells with P3HT hole-transporting materials via molecular p-type doping. RSC Adv. 6, 108888–108895. https://doi.org/10.1039/C6RA21775C.

This page intentionally left blank

Spintronics and magnetic field effects in organic semiconductors and devices

12

Tho Duc Nguyen*, Eitan Ehrenfreund†, Zeev Valy Vardeny‡ *Department of Physics and Astronomy, University of Georgia, Athens, GA, United States, † Physics Department, Technion—Israel Institute of Technology, Haifa, Israel, ‡Physics and Astronomy Department, University of Utah, Salt Lake City, UT, United States

12.1

Introduction

Over the past two decades, the electron spin has transformed from an exotic subject in classroom lectures to a degree of freedom that materials scientists and engineers exploit in new electronic devices nowadays. This interest has been motivated from the prospect of using the spin degree of freedom, in addition to the charge, as an information-carrying physical quantity in electronic devices, thus changing the device functionality in an entirely new paradigm, which has been dubbed spintronics (Wolf et al., 2001; Zutic et al., 2004). This interest culminated by awarding the 2007 Nobel Prize in Physics to Drs. Albert Fert and Peter Gru˝nberg for the discovery and application of the giant magneto-resistance (GMR). More recently, the spintronics field (also called spin-based electronics) has focused on hybrids of ferromagnetic (FM) electrodes and semiconductors, in particular spin injection and transport in the classical semiconductor gallium arsenide (Awschalom and Flatte, 2007). However, spin injection into semiconductors has been a challenge because of the impedance matching problem between the FM and semiconductor (Yunus et al., 2008). In any case, this research has yielded much original physical insight, but no successful applications yet. For the last few years, interest also has risen in similar phenomena in the organics. The organic spintronics field was launched via two articles by Dediu et al. (2002) that suggested spin injection in organic semiconductors (OSCs), and Xiong et al. (2004), which demonstrated spin-polarized currents in thin films with a small organic molecule (namely, Alq3). OSCs can absorb and emit light while transporting charge, and this leads to optoelectronic devices such as photovoltaic cells (PVCs), organic light-emitting diodes (OLEDs), and organic field-effect transistors (OFETs). It is thus expected that adding control of the electron spin to the multifunctional characteristics of these versatile materials should yield novel magnetic devices in the near future. However, there are serious challenges to be faced in understanding the properties of spin-polarized current in OSCs and obtain high-quality devices. Most important, the approach in studying spin injection and transport in organic thin films is fundamentally different from that used for their inorganic counterparts. Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00012-7 © 2019 Elsevier Ltd. All rights reserved.

386

Handbook of Organic Materials for Electronic and Photonic Devices

OSCs are composed of light elements that have weak spin-orbit interaction; consequently, they should possess long spin relaxation times (Ruden and Smith, 2004; Pramanik, 2007; McCamey, 2008). Moreover, hyperfine interaction (HFI) has been thought to play an important role in organic magneto-transport (Bobbert et al., 2009). For example, if the HFI constant a determines the spin lattice relaxation time, TSL of the injected carriers, and consequently their spin diffusion length in a device setting, then the device performance may be enhanced simply by manipulating the nuclear spins of the organic spacer atoms. Moreover the HFI may also play an important role in other organic magneto-electronic devices such as two-terminal devices (Pulizzi, 2009). In addition to spin-polarized carrier injection into OSCs, another fascinating effect has been observed when applying a small external magnetic field to OLEDs that was dubbed the magnetic field effect (MFE; Hayashi, 2004), in which both the current and electroluminescence (EL) increase by as much as 40% at room temperature in relatively small magnetic fields of about 100 Gauss. In fact, MFE in organics was discovered about four decades ago (Ern and Merrifield, 1968); however, renewed interest in it has recently risen since it was realized that the MFE in OLEDs has tremendous potential applications. In spite of recent research efforts in organic MFE, the underlying mechanism and basic experimental findings remain hotly debated. Finally, HFI can influence other spin response processes such as optically detected magnetic resonance (ODMR) in OSC films (McCamey, 2008). ODMR is actually a magnetic field effect induced by microwave (MW) absorption that mixes the spin sublevels of polaron pair species. Therefore, gaining an understanding of ODMR in organic films and electronic devices may shed light on the organic MFE phenomenon. Moreover, ODMR spectroscopy is unique in that it readily measures the spin relaxation time in OSCs, and thus information may be obtained about the electronic spin interaction strength. Therefore, we have used this technique to obtain HFI strength in the same materials that are used for spin injection and MFE studies. In this chapter, we review some of the research achieved at the University of Utah in the field of organic spintronics (Wang and Vardeny, 2010; Nguyen et al., 2010a,b; Nguyen et al., 2011a,b); Specifically, we investigate the role of HFI in various organic magneto-electronic devices and films by replacing all strongly coupled hydrogen atoms (1H, nuclear spin I ¼ ½ ) in the organic π-conjugated polymer poly(dioctyloxy) phenyl vinylene (DOO-PPV) spacer (dubbed H-polymer in this discussion), with deuterium atoms (2H, I ¼ 1) [hereafter dubbed D-polymer; see Fig. 12.1] having much smaller a—namely, a(D) ¼ a(H)/6.5 (Carrington and McLachlan, 1967). We studied the influence of this hydrogen isotope exchange on the magnetic response of three spin-dependent processes: l

l

l

Magneto-electroluminescence (MEL) and magneto-conductance (MC) responses in OLEDs, which are presented in Section 12.2. In this section, we also summarize the newly discovered, ultrasmall MFE, such as in MEL and MC at B about 1–2 nm, the HFI with protons is expected to be dominant and actually determines the ISC rate. The MFE in organic devices is analogous to the MFE observed in chemical and biochemical reactions, where it was explained using the radical pair (RP) mechanism (Brocklehurst and McLauchlan, 1996; Timmel, 1998). In this model, the HFI, Zeeman, and exchange interactions are taken into account. It is assumed that RPs are immobile (and hence the radical diffusion is ignored), but the overall decay rate k of the RP is explicitly taken into account. The steady-state singlet fraction of the RP population (singlet yield ФS) is then calculated from the coherent time evolution of RP wave functions subjected to this interaction. It is clear that when ℏk is much larger than the energy of the HFI and exchange interactions, then the MFE is negligibly small because the pairs disappear before any spin exchange between the spin sublevels can occur. In contrast, for a relatively small decay rate ℏk, the MFE becomes substantially larger, and for a negligible exchange interaction, the singlet yield behaves approximately as ФS  B2/(B20 + B2) (the upside-down Lorentzian function of B), with B0  aHF/gμB, where aHF is the HFI constant. Then, the half-width at halfmaximum (HWHM) is B1/2 ¼ B0. In organic devices, the PP species play the role of RPs, and the calculated MC (and MEL) response may be expressed in terms of the singlet ФS and triplet ФT PP yields in an external magnetic field B. It is important to note here that finite MC can occur only when the singlet decay rate is different from that of the triplet (Ehrenfreund and Vardeny, 2012). In this section, we show our recent advances in understanding the MFE in OSC two-terminal devices using both experiment and theoretical methods. Specifically, the role of the HFI is readily demonstrated.

12.2.1 Experimental methods and results The MEL and MC measurements in OLEDs have been generally conducted on devices with a typical area of 5 mm2, where the organic spacers are deposited on a hole transport layer: poly(3,4-ethylenedioxythiophene) [PEDOT]-poly(styrene sulfonate) [PSS]. For bipolar devices, we capped the bilayer structure with a transparent anode: indium tin oxide (ITO), and a cathode: calcium (protected by aluminum film). For MEH-PPV, we also fabricated unipolar devices: the hole-unipolar device was in the form of ITO/PEDOT-PSS/MEH-PPV/Au, whereas the electron-unipolar device was Al/LiF (about 2 nm)/MEH-PPV/Ca/Al; these devices did not show EL. The organic diodes were transferred to a cryostat that was placed between the two poles of an electromagnet producing magnetic fields up to about 300 mT. The devices were driven at constant bias V, using a Keithley 236 apparatus; and the current I was measured while sweeping B.

390

Handbook of Organic Materials for Electronic and Photonic Devices

The MC and MEL responses are defined, respectively, via MCðBÞ ¼ ΔI ðBÞ=I ð0Þ ¼

I ðBÞ  I ðB ¼ 0Þ I ð B ¼ 0Þ

MEL ¼ ΔELðBÞ=ELð0Þ ¼

ELðBÞ  ELðB ¼ 0Þ ELðB ¼ 0Þ

(12.1)

where △ I and △ EL are the field-induced changes in the current and EL intensity, respectively. Fig. 12.3A shows the MEL response of two OLED devices based on H- and D-polymers with the same thickness df, measured at the same bias V; a very similar MC response was measured simultaneously with MEL (Fig. 12.3B). The MEL and MC responses are narrower in the D-polymer device; in fact, the field, B1/2 for the MEL is about twice as large in the H-polymer device than in the D-polymer device. We also found that B1/2 increases with V (Fig. 12.3A, inset) (Wang et al., 2008); actually, B1/2 increases approximately linearly with the device’s electric field, E ¼ (V  Vbi)/df, where Vbi is the built-in potential in the device, which is related to the onset bias voltage where EL and MEL are observed (Bloom et al., 2007, 2009). Importantly, in all cases, we found that B1/2(H) > B1/2(D) for devices having the same value of the electric field, E (Fig. 12.3A, inset).

1.0

4

0

(A)

0

MC (%)

8

1

H-DOO-PPV mT

MEL (%)

2

0.5 D-DOOPPV H-DOOPPV

D-DOO-PPV

E (107 V/m) 0

–40

1

2

–20

0.0

3

0

20

40

(B)

–60

–40

–20

0 20 B (mT)

40

60

Fig. 12.3 Isotope dependence of MEL (A) and MC (B) responses in OLEDs based on DOOPPV polymers. Room-temperature MEL and MC responses of D- and H-polymers (solid and dashed lines, respectively) measured at bias voltage V ¼ 2.5 V. Inset to (A): the field B1/2, at half the MEL maximum for the two polymers, plotted versus the applied bias voltage, V, which is given in terms of the internal electric field in the polymer layer, E ¼ [V  Vbi]/df, where Vbi is the built-in potential in the device and df is the active layer thickness; the lines are linear fits to “guide the eye.” From Nguyen, T.D., Gautam, B.R., Ehrenfreund, E., Vardeny, Z.V., 2010a. Magnetoconductance response in unipolar and bipolar organic diodes at ultrasmall fields. Phys. Rev. Lett., 105, 166804; Nguyen, T.D., Hukic-Markosian, G., Wang, F., Wojcik, L., Li, X.-G., Ehrenfreund, E., Vardeny, Z.V. 2010b. Isotope effect in spin response of [Pi]-conjugated polymer films and devices. Nat. Mater., 9, 345–352, with permission.

Spintronics and magnetic field effects in organic semiconductors and devices

391

MEL and MC in OLEDs may be considered as an example of a much broader research field that deals with MFE in physics (Groff et al., 1974), chemistry, and biology (Timmel, 1998; Hayashi, 2004). For MFE that results from pairs of radical ions, it was empirically realized (Weller et al., 1983) that B1/2 scales with the HFI constant aHF. In fact, a semiempirical law was advanced (Weller et al., 1983), which for opposite-charge radicals with same aHF reads as B1=2  2aHF ½I ðI + 1Þ1=2 =gμB

(12.2)

where I is the nuclear spin quantum number. In fact, the nuclear spin in the analysis that led to Eq. (12.2) was treated classically (Weller et al., 1983); therefore, it should be considered a crude approximation. Nevertheless, using Eq. (12.2), we get B1/2(H)  aHF(H)√3/gμB for the H-polymer with I ¼ ½, and B1/2(D)  2aHF(D)√2/gμB for the D-polymer with I ¼ 1. Taking aHF(H) and aHF(D) from the fits to the ODMR lines that will be obtained in Section 12.4, we find that B1/2(H)  6 mT for the H-polymer, which is reduced to B1/2(D)  1.5 mT for the D-polymer, compared with the experimental B1/2 values of about 6 mT and about 3 mT, respectively, obtained at small E (Fig. 12.4A, inset). Nonetheless, from the obtained reduction in B1/2, we conclude that the MEL response in OLED of π-CPs is mainly due to the HFI (Sheng et al., 2006; Desai et al., 2007a). Surprisingly, Fig. 12.4A and B show that the MEL and MC has yet another component at low B (that is dubbed “ultra-small-field MEL/MC”; namely, USMEL and USMC), which has an opposite sign as that of the positive MEL (MC) at higher fields. A similar low-field component was also observed in some biochemical reactions (Brocklehurst and McLauchlan, 1996) and anthracene crystals (Belaid et al., 2002),

0.6

D-DOOPPV H-DOOPPV

0.5

MC (%)

MEL (%)

1.0

0.0

0.0

(A)

–2

0.3

–1

0

1

2

(B)

–2

–1

0 B (mT)

1

2

Fig. 12.4 Room-temperature MEL (MC) response of D- and H-polymers (solid and dashed lines, respectively) measured at bias voltage V ¼ 2.5 V, plotted for jB j 0 (Bergeson et al., 2008; Sheng et al., 2006); therefore, the negative USMEL component should be due to an ISC rate increase between PP spin sublevels at low B. This might happen, for example, if there is a level crossing (LC) between PPS and PPT spin sublevels (Hayashi, 2004). But because the two PPT spin sublevels with Sz ¼ 1 split linearly with B (Bergeson et al., 2008), an isotope-dependent LC in the PP spin sublevels at very low field cannot be easily accounted for with the four basic spin wave functions of PPS and PPT, which are traditionally considered in simple MEL models (Bergeson et al., 2008; Sheng et al., 2006). Therefore, we were led to conclude that additional spin wave functions are needed to explain the USMEL response. The USMFE response is not limited to bipolar devices. In Fig. 12.5A and B, we show MC(B) responses of hole-only and electron-only MEH-PPV diodes; similar responses were measured for DOO-PPV devices (Nguyen et al., 2011a). The high-field MC in unipolar devices is negative (Fig. 12.5A; Wang et al., 2008), and thus the USMFE response here appears as negative-to-positive sign reversal with a maximum at Bm  0.8 mT for the electron-only device and Bm  0.1 mT for the hole-only device

MC/MCmax

0.2 x10

0.0 x2

–0.2 –0.4

(B)

–2

–1

0

MC (%)

0.0 –0.4

1

2

Hole only (x40) Electron only

MEH-PPV

–0.8 –1.2 –30

(A)

–20

–10

0

10

20

30

B (mT)

Fig. 12.5 Normalized MC(B)/USMC(B) response for (A) j B j ¼ 4Σm jPSmm j2/M + 4Σm6¼n jPSmn j2/M, where the summations are restricted to degenerate levels, for which ωmn(B) ¼ 0. Here, the first term contributes to the MFEM(B) response, whereas the second term contributes to the MFELC(B) response that modulates < ρS(t ¼ ∞) > primarily at B ¼ 0, where the S-T degeneracy is relatively high (see Fig. 12.7). The combination of the monotonous MFEM(B) and MFELC(B) components at B  0 explains, in principle, the USMFE response in organic devices. When allowing for SP spin decay, ρS(t) in Eq. (12.3) should then be revised to reflect the disappearance of SP with time. Furthermore, for MFE to occur, the decay rates of singlet and triplet configurations must be different from each other. Thus, in a decaying system, the population of each of the M levels would decay at a different rate, γ n (n ¼ 1,…,M). Under these conditions, Eq. (12.3) for the singlet fraction is given by (Timmel, 1998)   4 XM  S 2 P  cos ðωmn tÞeγ nm t ρS ðtÞ ¼ Tr ρðtÞPS ¼ m,n¼1 mn M

(12.4)

where γ nm ¼ γ n + γ m. Eq. (12.4) expresses the fact the singlet (or triplet) time evolution contains both a coherent character [through the cos(ωmnt) factor] and an exponential decay factor. Then the measured MFE (i.e., MC and MEL) may be calculated using Eq. (12.4). For instance, if the dissociation yields are kSD and kTD for the singlet and triplet configurations, respectively, then the time-dependent dissociated fraction of either the singlet or triplet is kαDρα(t) (α ¼ S,T), and thus the dissociation yield is (Ehrenfreund and Vardeny, 2012) Z

∞

ΦαD ¼ 0

kαD ρα ðtÞdt ¼

4X α kαD γ nm Pn,m σ m,n ð0Þ 2 M n, m γ nm + ω2nm

(12.5)

The total dissociation yield is ΦD ¼ ΦSD + ΦTD, and the MC(B) response is then given by MCðBÞ ¼

ΦD ðBÞ  ΦD ð0Þ ΦD ð0Þ

(12.6)

For a slow decay, such that k ≪ aHF/ℏ, the abrupt MFELC(B) obtained at B ¼ 0 in the absence of the spin decay is spread over a field range of the order of ℏk/gμB, after which ΦS(B) increases again due to the more dominant MFEM(B) component at large B. For the MEL response, the final expression depends on the radiative recombination path of the SE and the detailed relaxation route from PP to the SE. For instance, in

Spintronics and magnetic field effects in organic semiconductors and devices

397

polymers where the SE-TE gap is relatively large (say, > 10% of the SE energy), a substantial SE-TE intersystem is crossing through the SOC. As a result, PPT (PPS) may transform not only to TE (SE), but also to SE (TE). Let us denote the effective SE (TE) generation rates from the PPα (α ¼ S,T) configuration as kα,SE (kα,TE). Then, similar to MC, we can define the SE generation yield, ΦSE ¼ ΦS, SE + ΦT, SE, where Φα, SE is given by Eq. (12.5), in which kαD is replaced by kα,SE . Because the EL is proportional to the SE density, the MEL response is still given by Eq. (12.6), in which ΦD is replaced by ΦSE. Fig. 12.8 shows the singlet yield and resulting MEL(B) response of the H-polymer OLED. Importantly, the calculated MEL response captures the experimental USMEL response, comprising a negative component with a minimum at Bmin  0.5 mT, that changes sign to positive MEL with an approximate B2/(B20 + B2) shape with B0  4.5 mT. The high field shape, namely B2/(B20 + B2), is the generic feature in this model. For small values of the exchange interaction, B0 is determined mainly by the HFI constant aHF; also, the USMEL response is a strong function of the decay constant k, as shown in Fig. 12.11. The negative component with Bmin appears only for relatively long decay times (e.g., ℏk/aHF 0.1). For Jex/aHF > 1, the characteristic SMEL response is no longer distinguishable. We note that in Fig. 12.8, the MEL HWHM (¼ 4.5 mT) is not exactly equal to aHF/ gμB, presumably because of the contribution of the USMEL component at low B. We also note that high field resolution is needed for observing the USMEL component,

FS

0.7 0.6 0.5

MEL (%)

(A)

–16

–8

0

8

16

–16

–8

0 B (mT)

8

16

0.8 0.4 0.0

(B)

Fig. 12.8 Calculated magnetic field response of the singlet yield (A) and magneto-conductance (B) for a two-proton PP, where g1 ¼ g2 ¼ g 2, a1 ¼ a2 ¼ a, with a/gμB ¼ 3.5 mT, J ¼ 0, δTS ¼ 0.96, and ℏk/a ¼ 2 103. The resulting MEL response HWHM is about 4.5 mT, and Bmin is about 0.5 mT. From Nguyen, T.D., Hukic-Markosian, G., Wang, F., Wojcik, L., Li, X.-G., Ehrenfreund, E., Vardeny, Z.V. 2010b. Isotope effect in spin response of [Pi]-conjugated polymer films and devices. Nat. Mater., 9, 345–352, with permission.

398

Handbook of Organic Materials for Electronic and Photonic Devices

and this might be the main reason why this component has not been observed thus far in organic magneto-transport. The calculated MEL response for various decay rate constant, k is shown in Fig. 12.9, in which Bmin is strongly dependent on k. Moreover, the model calculation obtained using ℏk/aHF  0.002 also nicely reproduces the USMEL effect (Fig. 12.10), where the calculated Bmin occurs at about 0.7 and 0.3 mT in the H- and D-polymer, respectively; in excellent agreement with the experiment (Fig. 12.4A). Fig. 12.11 shows the calculated MC(B) response, together with its energy sublevels for an axially symmetric anisotropic HFI with N1 ¼ N2 ¼ 1 (I ¼ ½; M ¼ 16), where aHF(electron)/gμB ¼ 3aHF(hole)/gμB ¼ 3 mT, these parameters extracted from the unipolar MEH-PPV MC(B) response, J ¼ 0, δTS ¼ 0.96, and an exponential SP decay ħk/ aHF ¼ 0.001. The calculated MC(B) response captures the experimental USMC response comprising a negative component having its minimum at Bm  aHF/6gμB ¼ 0.5 mT, with an approximate positive B2/(B20 + B2) shape at large B, and B0  4.5 mT. The excellent agreement between theory and experiment, including both Bm, and the USMC intricate response and relative amplitude, validates the model used.

12.2.3 Conclusions In the research investigations described in this section, we elucidated the role of HFIs in the MFE response of conductivity and EL in various OLEDs based on DOO-PPV isotopes that have different hyperfine coupling constants. We showed that both MEL

0.2 k/a = 0.2

0.0

–0.2 0.0

0.1 0.04 0.02 0.002

0.4 0.3

gmBBmin/a

MEL (%)

0.4

0.2 0.1

k/a

0.0 0.00

0.0002

0.5

gmBB/a

0.05

1.0

0.10

1.5

Fig. 12.9 Calculated MEL response for the two-polaron, two-proton model, for various decay rate constants k (given here in units of a). The inset shows the calculated dependence of Bmin on k; it approximately follows the functional dependence, Bmin/a  (ℏk/a)0.28. From Nguyen, T.D., Hukic-Markosian, G., Wang, F., Wojcik, L., Li, X.-G., Ehrenfreund, E., Vardeny, Z.V. 2010b. Isotope effect in spin response of [Pi]-conjugated polymer films and devices. Nat. Mater., 9, 345–352, with permission.

Spintronics and magnetic field effects in organic semiconductors and devices

399

2

Model

MEL (%)

0

(A)

–30

–20

2

–10

0

10

20

30

1

2

3

Protons Deutrons

0

(B) –3

–2

–1

0 B (mT)

Fig. 12.10 Simulations of the MEL response in the two polymers, which reproduces the response data in Figs. 12.3 and 12.4 based on the described model (Timmel, 1998) using the calculated spin sublevels given in Fig. 12.7. From Nguyen, T.D., Hukic-Markosian, G., Wang, F., Wojcik, L., Li, X.-G., Ehrenfreund, E., Vardeny, Z.V. 2010b. Isotope effect in spin response of [Pi]-conjugated polymer films and devices. Nat. Mater., 9, 345–352, with permission.

and MC are strongly dependent on the HFI constant of the materials; the stronger the HFI constant, the broader is the MFE response. Our findings may be useful for material scientists for designing the novel materials for MFE applications. Furthermore, we analyzed the novel USMFE response at B and jf > are vibrational states of the system, and Qk are its normal modes. Research from Jensen and coworkers has shown that the interaction between molecules and the metal surface will affect the EM enhancement, suggesting that the traditional view of treating the EM and chemical enhancement effects as separate entities may provide an incomplete view (Payton et al., 2014). They found that accounting for the presence of a molecule near a metal nanoparticle surface actually could decrease the enhancement relative to the EFEM predicted by an electrodynamics simulation, based on the identity of the metal and the orientation of the molecule, as shown in Fig. 16.4. Thus, while EM enhancement is the dominant mechanism for enhanced SERS signals, one must be careful when using calculated fields (in the absence of a molecule) as an absolute predictor of SERS enhancement in real experiments.

534

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 16.3 An energy-level diagram describing all possible transitions in the metal-molecule SERS system. I and K are the ground state and excited state of the molecule, respectively. F is the Fermi level of the metal. Columns A and B show the possibilities of a molecule ➔ metal and metal ➔ molecule CT, respectively. The μ terms are electronic transition moments, and the h terms are vibronic coupling terms, connecting the energy levels as shown. Reproduced with permission from Lombardi, J.R., Birke, R.L., 2008. A unified approach to surface-enhanced Raman spectroscopy. J. Phys. Chem. C 112, 5605–5617, Copyright 2008 American Chemical Society.

Fig. 16.4 SERS EFs of pyridine on (A) silver and (B) gold as a function of the distance between the center of mass of the molecule and the metal surface. Vertical lines indicate the equilibrium bonding distance. The dots indicate atomistic simulation results, which explicitly take the molecule into account, and the dashed line indicates the j E j4 enhancement alone. Note that the j E j4 method may either overestimate or underestimate the EF. Reproduced with permission from Payton, J.L., Morton, S.M., Moore, J.E., Jensen, L., 2014. A hybrid atomistic electrodynamics–quantum mechanical approach for simulating surface-enhanced Raman scattering. Acc. Chem. Res. 47, 88–99, Copyright 2013 American Chemical Society.

SERS as a characterization method for metal-organic interactions

535

16.2.2 SERS examples 16.2.2.1 SERS vs. normal Raman SERS has been carried out on a wide range of substrates, including electrochemically roughened silver electrodes, island films, metal films over nanospheres (FONs) (Dick et al., 2002), synthetically and lithographically prepared nanoparticles, and randomly assembled silver colloid clusters. It is thought that the tight junctions between nanoparticles in a cluster produce extremely high electric fields at hot spots, leading to significant Raman-scattering enhancements. For the example SERS spectrum displayed in Fig. 16.5, silver colloids were aggregated with NaCl and incubated with berberine (Lombardi and Birke, 2008). By comparing the resulting SERS spectrum with the conventional solution-phase Raman spectrum of berberine, also shown in the figure, we can identify several interesting features. First, the SERS spectrum clearly shows a greater signal-to-noise ratio than the conventional Raman spectrum, even though the intensities between the two spectra cannot be directly compared, as the spectra are not normalized to the number of molecules sampled [see Eq. (16.1)]. In addition, the relative intensity of different modes varies widely among the spectral peaks from normal Raman to SERS. For example, the peak at 729 cm
1 becomes dominant in the SERS spectrum, while showing only midlevel intensity relative to the other peaks in the normal Raman data. On the other hand, the peak at 1203 cm
1 is barely observed above the noise in the SERS data, despite having a strong signal in the normal Raman spectrum. These differences occur because Raman modes

Fig. 16.5 Comparison of surface-enhanced and normal Raman spectra for berberine on silver colloid clusters. Note the differences in relative peak heights and peak positions. Reproduced with permission from Lombardi, J.R., Birke, R.L., 2008. A unified approach to surface-enhanced Raman spectroscopy. J. Phys. Chem. C 112, 5605–5617, Copyright 2008 American Chemical Society.

536

Handbook of Organic Materials for Electronic and Photonic Devices

originate from different bonds within the analyte molecule, and those closer to the metal surface are more strongly enhanced (via both EM and CT processes) than those farther away. Also, for some modes, the frequencies are shifted slightly between the normal Raman and SERS spectra. This is explained by slight differences in molecule conformation and bond lengths when the molecule is adsorbed on a surface versus being free in solution.

16.2.2.2 SERRS (resonance Raman) As discussed in Section 16.2.1, one of the contributions to the SERS signal is from electronic resonances of the analyte at the laser wavelength; this is referred to as SERRS. In resonance Raman scattering, rather than being excited to an instantaneous virtual state by the laser (as shown in Fig. 16.1), the molecule is excited to an electronic excited state. As the molecule relaxes back to the ground state, vibrational energy levels in the neighborhood of the excited electronic state are probed, and the emitted photons have corresponding energy shifts. Thus, the enhancement due to resonance Raman is actually not related to the presence of a metal surface at all, even though it is taken into account when calculating the total EF in SERS. The main advantage of resonance Raman is generally that it greatly enhances a small subset of vibrational modes of the molecule, allowing specific structures to be targeted. The combination of resonance Raman and SERS has proven to be a powerful technique. For example, SERRS has been used to study photobleaching of R6G (Maher et al., 2002) and to probe the vibrational modes of carbon nanotubes (CNTs; Kneipp et al., 2001).

16.2.2.3 Single-molecule SERS The enhancement offered by SERS can be significant enough to enable the detection of Raman scattering from even a single molecule. To date, the most popular substrate for single-molecule SERS detection is silver colloid clusters (Nie and Emery, 1997; Michaels et al., 2000), although there are reports of achieving single-molecule sensitivity using gaps between nanoparticles and thin-metal-film substrates (Park and Kim, 2010) and even single-nanoparticle substrates (Zrimsek et al., 2013). When Raman spectra are taken from a molecule (or few molecules) adsorbed to a single cluster, fluctuations are seen in both the total intensity of the scattering and the relative intensities among the various peaks. This “blinking” behavior is very indicative of single or few-molecule activity. Single-molecule SERS can be further proven by employing a bianalyte technique, in which a silver colloid cluster substrate is incubated with a mixture of two analytes at a sufficiently low concentration that only one analyte should adsorb to a single cluster (Le Ru et al., 2006; Van Duyne et al., 2007, 2011; Zrimsek et al., 2016). In the singlemolecule regime, spectral signatures are observed from either one or the other analyte, but rarely both; through statistical analysis, single-molecule behavior can be confirmed. See Fig. 16.6 for an example of such data for benzotriazole (BTZ) and rhodamine dyes (Le Ru et al., 2006).

SERS as a characterization method for metal-organic interactions

537

Fig. 16.6 Bianalyte SERS. (A) Averaged SERS spectra for R6G alone, BTZ alone, and mixtures of the two analytes. (B) Single SERS spectra from a sample with both analytes present, showing the characteristics of BTZ only, R6G only, and both. Reproduced with permission from Le Ru, E.C., Meyer, M., Etchegoin, P.G., 2006. Proof of single-mokeule sensitivity in surface enhanced Raman scattering (SERS) by means of a two-analyte technique. J. Phys. Chem. B 110, 1944–1948, Copyright 2006 American Chemical Society.

16.3

SERS applications

16.3.1 Fundamental studies of metal-organic interfaces As a tool for the fundamental investigation of metal-organic interfaces, SERS can provide a wealth of information. For example, SERS has been used to study the adsorption geometry of 1,4-benzenedithiol (BDT) on the surface of gold and silver colloids ( Joo et al., 2001). While BDT forms two Ag-S bonds on the surface of the silver, resulting in a planar geometry of the molecule with respect to the surface, the molecule can access both a planar and perpendicular orientation on gold surfaces, depending on the concentration of BDT. SERS also has been used to study how surface charge affects the orientation of molecules such as cytochrome-c on a gold surface (Yu and Golden, 2007). Using self-assembled monolayers (SAMs) terminated with different functional groups, the authors tuned the surface charge of the gold and

538

Handbook of Organic Materials for Electronic and Photonic Devices

followed how the orientation of the cytochrome-c was modulated by measuring changes in the SERS intensity ratio between totally symmetric and nontotally symmetric modes of the molecule. SERS has also been used to follow the surface-induced photoreduction of 4-nitrobenzenethiol (NBT) to 4-aminobenzenethiol (ABT) on nitric acid–etched copper films (Shin et al., 2007). While adsorbed NBT has stable SERS peaks on gold and silver films, a signature NBT SERS peak at 1330 cm
1 disappears over time and the final spectrum matches ABT, indicating surface-induced photoconversion of NBT to ABT on the copper surface. In addition to probing adsorbate geometry and photoconversion at interfaces, SERS is useful for following analyte dynamics at the metal-molecule interface. For example, by combining SERS with superresolution imaging, the diffusion behavior of a single dye molecule within the tightly confined volumes of a SERS hot spot can be followed in real time (Willets and Stranahan, 2010; also see Fig. 16.6). These studies revealed analyte mobility on the surface of the metal nanoparticles and showed that the SERS intensity depends strongly on the spatial position of the mobile analyte on the nanoparticle surface (Weber et al., 2012; Titus et al., 2012).

16.3.2 Fundamental studies of plasmon-enhanced photochemistry Plasmon-enhanced photochemical reactions, or plasmonic photocatalysis, constitute an active research area, of which SERS experiments form an important part (Sun and Xu, 2012). Briefly, plasmonic nanoparticles can promote chemical reactions at their surface in the presence of light excitation in several ways, including via local field enhancement, local heating, enhanced scattering, and the addition of hot charge carriers. The coincident presence of SERS provides a useful way to study these processes. A well-known example of a system exhibiting plasmonic photocatalysis is the selective conversion of p-aminothiophenol (PATP) to p,p0 -dimercaptoazobenzene (DMAB). SERS has been used to help unravel the reaction mechanism, such as with respect to the role of oxygen, and to address the question of how the different enhancement mechanisms (local heating, CT, etc.) affect this particular system (Chen et al., 2014). For example, ultrafast SERS studies on this system have shown the importance of hot carriers for driving this reaction on silver nanoparticles (Brandt et al., 2016) as well as the role of oxygen (Zhang and Wang, 2018). Hot carriers also have been implicated in the photoreduction of Fe3+ to Fe2+, which is monitored by SERS by attaching the iron to surface-bound cyano groups and tracking the shift in the CN-stretching frequency as the oxidation state of the iron changes (Kim et al., 2014). While these experiments suggest CT of a plasmonic hot carrier to an adsorbed analyte, other SERS experiments have also suggested a possible direct charge excitation pathway into a coupled molecule-metal system (similar to the CT states associated with chemical enhancement), which may destabilize adsorbed molecules and further promote photocatalytic conversion (Boerigter et al., 2016).

SERS as a characterization method for metal-organic interactions

539

16.3.3 SERS in forensics SERS has been employed in several fields of forensic science, ranging from the identification of drugs to the detection of explosives (Fikiet et al., 2018). For example, Ray et al. reported the detection of the explosive trinitrotoluene (TNT) with picomolar sensitivity via SERS of cysteine-modified gold nanoparticle clusters (Ray et al., 2009). In this method, the presence of TNT causes particle aggregation due to its complexation with cysteine; the aggregation then induces a great SERS enhancement (up to nine orders of magnitude). The fact that the sensor is aggregation-based causes an easily seen color change with TNT detection. The sensor is specific; that is, the high SERS response is seen only for TNT, and not for other explosives or heavy metals. Later, Tsukruk et al. demonstrated an explosives sensor based upon porous alumina membranes coated with cetyltrimethylammonium bromide (CTAB)–capped gold nanoparticle clusters (Tsukruk et al., 2009). This sensor was not aggregation-based, so it lacked the colorimetric aspect of the sensor discussed here, but it was sensitive to as little as 15–30 molecules of TNT or dinitrotoluene (DNT, a related explosive). This great improvement in sensitivity was attributed to the increased gold surface area made available by using the porous membranes with densely deposited nanoparticles, and to the waveguiding effect of the alumina. Paper-based SERS substrates also have been used for explosives identification; they offer another advantage in that they can be integrated with other analytical techniques, such as mass spectrometry, to provide better identification and quantitation performance metrics (Fedick et al., 2017).

16.3.4 SERS of deoxyribonucleic acid (DNA) and proteins Many SERS applications have been developed for deoxyribonucleic acid (DNA) detection and identification. Perhaps the best known of these is the DNA sensor of Mirkin et al. (Cao et al., 2002). This sensor is based on single-stranded DNA sequences A1 and A2, which are complementary to the two ends of target sequence B. A1 is immobilized on a substrate, while A2, modified with a SERS probe molecule, decorates the surface of gold nanoparticles in solution. If B is introduced into the solution, one end of the target hybridizes to the tethered A1 sequence, while the other end hybridizes to the SERS-tagged A2 sequence, effectively tethering the SERS probe and the gold nanoparticles to the surface. Silver nanoparticles are then grown on the surface of the tethered gold nanoparticles to induce SERS from the attached probe. This method has been shown to detect specific DNA sequences at fM concentrations and can be easily multiplexed to sense multiple DNA sequences, simply by introducing different SERS probes. While this strategy is based on using a capture strand to target a specific molecule, more recent work has focused on identification of the bases present in DNA without advance knowledge of the target sequence. For example, Halas and coworkers have shown that using aluminum as a SERS substrate allows individual bases to be identified and even quantified, as seen in Fig. 16.7 (Tian et al., 2017). Aluminum offers a native oxide layer that prevents chemical interactions between the substrate and the DNA, which typically distort the spectral features relative to normal Raman

540

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 16.7 Raman spectra of 9-mers of DNA bases (A) cytosine, (B) adenine, (C) thymine, and (D) guanine as measured via SERS on an aluminum substrate (top), via normal Raman (middle), and via SERS on a gold substrate (bottom). Reproduced with permission from Tian, S., Neumann, O., Mcclain, M.J., Yang, X., Zhou, L., Zhang, C., Nordlander, P., Halas, N.J., 2017. Aluminum nanocrystals: a sustainable substrate for quantitative SERS-based DNA detection. Nano Lett. 17, 5071–5077, Copyright 2017 American Chemical Society.

scattering, while still maintaining sufficiently high EM enhancement to enable the high sensitivity associated with SERS. SERS has also proven effective for protein detection, for example in the work of Porter et al., who have reported an immunosensor for prostate-specific antigen (PSA), sensitive to the pg/mL-level (Grubisha et al., 2003). In this scheme, anti-PSA is immobilized on a gold surface and blood serum containing PSA is flowed over the surface to capture PSA. Then, gold nanoparticles labeled with both a SERS probe dye and a second antibody specific to the target are added, forming SERS-active junctions between the gold nanoparticles and substrate. SERS spectra were then measured via a scanned fiber bundle and the concentration of PSA was determined from the intensity of the SERS peaks of the probe dye. SERS measurements on proteins also have helped to elucidate aspects of the SERS mechanism itself. For example, Hofkens

SERS as a characterization method for metal-organic interactions

541

and co-workers have measured single-molecule SERS of green fluorescent protein (GFP) on silver colloid clusters (Habuchi et al., 2003). They were able to observe single-molecule blinking, as well as transitions between the protonated and deprotonated forms of individual molecules. This experiment also revealed important information about the size of SERS hot spots. Because GFP is about 4  8 nm in size, whereas previous single-molecule SERS studies were carried out with small dyes (400 nF/cm2 could be demonstrated (Lee et al., 2015). An interesting alternative to inorganic high-k dielectrics are innovative polymer dielectrics such as low-temperature, cross-linkable copolymers (Wang et al., 2017) or P(VDF-TrFE)-based dielectrics (Baeg et al., 2012; Khim et al., 2016), including novel relaxor ferroelectric terpolymers (Zhu and Wang, 2012; Tang et al., 2017, 2015; Schmidt et al., 2015). These materials can exhibit extraordinarily high k-values (e.g., k ¼ 50) for P(VDF-TrFE-CTFE) (Zhu and Wang, 2012), and provide high capacitance even if applied in a multilayer stack with low-k polymer dielectrics such as PS (Xu et al., 2017) or Cytop (Tang et al., 2017, 2015; Schmidt et al., 2015), an approach that has become increasingly common in organic thin-film devices. For instance, an area-normalized capacitance >100 nF/cm2 was reported by Tang et al. (2015) for the combination of Cytop and P(VDF-TrFE-CFE). Such multilayer dielectrics are either fabricated by subsequent coating steps, including screen printing and flexography (Schmidt et al., 2015), or via selforganization during the deposition of a blend of low- and high-k polymer (Khim et al., 2016; Xu et al., 2017). In the latter work, the ferroelectric high-k dielectric P (VDF-TrFE) was blended with the amorphous, low-k polymer PMMA and when spin coated, formed double-layer dielectrics as a result of vertical phase separation

570

Handbook of Organic Materials for Electronic and Photonic Devices

between the blend materials. In general, blended dielectric can provide a range of advantageous characteristics, including reduced roughness, low hysteresis, and improved operational stability compared to the ferroelectric polymer itself. Though not yet realized, the fact that typical stratification behavior of two materials when processed by spin coating is often reproduced when applying scalable methods such as blade coating suggests that these polymer-dielectric films possibly can be deposited with scalable processing techniques.

17.6

Summary

In this chapter, we highlighted some of the more recent advances in the solution coating of materials for organic electronics applications. We feel that a lot of progress has been made during the past few years in advancing the understanding of the complexity of the coating process and utilizing this grown understanding for the fabrication of more efficient or higher-performance devices. This has led to a number of innovative coating solutions that sometimes involve exotic implements such as diamond dust lapping paper or Chinese brushes. It is worth pointing out that a large number of processing degrees of freedom inherent to solution-coating methods allows one to attack the problem of improving active device layer properties across various length scales. Many of the increases in device performance highlighted herein have come from improvements in device layer morphology on the mesoscale, but in some cases, the improvement stems from influencing molecular alignment, such as in the case of polymer growth in nanogrooves, or when the coating method kinetically stabilizes metastable high-performance polymorphs of a given material (e.g., TIPS-Pn). In our opinion, achieving greater control over the crystalline/molecular texture of organic thin films remains more challenging than controlling macroscale and mesoscale wetting behaviors. We’d like to close this chapter by identifying two specific challenges that we feel remain to be tackled in the field of solution-processed organic electronics: (1) Modern vertical electronic devices that rely on efficient transport in the out-of-plane direction, such as many photonic devices but especially high-speed transistors, would profit tremendously from the availability of coating methods that can create favorable out-ofplane textures (i.e., aligning molecules or polymer chains such that the most efficient charge transport axis coincides with the substrate normal). (2) The device performance gap between lab-scale solution coating methods and commercially feasible production schemes remains large and may even have widened for some of the newer lab coating methods. It remains a challenge to demonstrate that a method that produces world-record single devices—often with a very small device area and coated at unrealistically low submillimeter-per-second speeds—does not lose its performance edge over conventional techniques when area and throughput are being scaled up. We hope that the next iteration of this series can report on some exciting research accomplishments in these directions.

Advances in solution processing of organic materials for devices

571

References Baeg, K.-J., Khim, D., Kim, J., Han, H., Jung, S.-W., Kim, T.-W., Kang, M., Facchetti, A., Hong, S., Kim, D., Noh, Y., 2012. Controlled charge transport by polymer blend dielectrics in top-gate organic field-effect transistors for low-voltage-operating complementary circuits. ACS Appl. Mater. Interfaces 4, 6176–6184. Barrows, A.T., Pearson, A.J., Kwak, C.K., Dunbar, A.D.F., Buckley, A.R., Lidzey, D.G., 2014. Efficient planar heterojunction mixed-halide perovskite solar cells deposited via spraydeposition. Energy Environ. Sci. 7, 2944–2950. Basescu, N., Liu, Z.-X., Moses, D., Heeger, A.J., Naarmann, H., Theophilou, N., 1987. High electrical conductivity in doped polyacetylene. Nature 327, 403–405. Chen, J.-D., Cui, C., Li, Y.Y.-Q., Zhou, L., Ou, Q.-D., Li, C., Li, Y.Y.-Q., Tang, J.-X., 2015. Single-junction polymer solar cells exceeding 10% power conversion efficiency. Adv. Mater. 27, 1035–1041. Cheng, X., Caironi, M., Noh, Y.-Y., Wang, J., Newman, C., Yan, H., Facchetti, A., Sirringhaus, H., 2010. Air stable cross-linked cytop ultrathin gate dielectric for high yield low-voltage top-gate organic field-effect transistors. Chem. Mater. 22, 1559–1566. Choi, D.Y., Kang, H.W., Sung, H.J., Kim, S.S., 2013. Annealing-free, flexible silver nanowirepolymer composite electrodes via a continuous two-step spray-coating method. Nanoscale 5, 977–983. De, S., Coleman, J.N., 2010. Are there fundamental limitations on the sheet resistance and transmittance of thin graphene films? ACS Nano 4, 2713–2720. del Pozo, F.G., Fabiano, S., Pfattner, R., Georgakopoulos, S., Galindo, S., Liu, X., Braun, S., Fahlman, M., Veciana, J., Rovira, C.C., Crispin, X., Berggren, M., Mas-Torrent, M., 2016. Single crystal-like performance in solution-coated thin-film organic field-effect transistors. Adv. Funct. Mater. 26, 2379–2386. Diao, Y., Tee, B.C.-K.C.K.C.-K., Giri, G., Xu, J., Kim, D.H., Becerril, H.A., Stoltenberg, R.M., Lee, T.H., Xue, G., Mannsfeld, S.C.B.B., Bao, Z., 2013. Solution coating of large-area organic semiconductor thin films with aligned single-crystalline domains. Nat. Mater. 12, 665–671. Diao, Y., Zhou, Y., Kurosawa, T., Shaw, L., Wang, C., Park, S., Guo, Y., Reinspach, J.a., Gu, K., Gu, X., Tee, B.C.K., Pang, C., Yan, H., Zhao, D., Toney, M.F., Mannsfeld, S.C.B., Bao, Z., 2015. Flow-enhanced solution printing of all-polymer solar cells. Nat. Commun. 6, 7955. Epstein, A.J., Macdiarmid, A.G., 1995. Polyanilines – from solitons to polymer-metal, from chemical curiosity to technology. Synth. Met. 69, 179–182. Fan, X., Wang, J., Wang, H., Liu, X., Wang, H., 2015. Bendable ITO-free organic solar cells with highly conductive and flexible PEDOT:PSS electrodes on plastic substrates. ACS Appl. Mater. Interfaces 7, 16287–16295. Ford, M.J., Wang, M., Patel, S.N., Phan, H., Segalman, R.A., Nguyen, T.-Q., Bazan, G.C., 2016. High mobility organic field-effect transistors from majority insulator blends. Chem. Mater. 28, 1256–1260. Galindo, S., Tamayo, A., Leonardi, F., Mas-Torrent, M., 2017. Adv. Control of polymorphism and morphology in solution sheared organic field-effect transistors. Funct. Mater. 27, 1700526. Giri, G., Park, S., Vosgueritchian, M., Shulaker, M.M., Bao, Z., 2014. High-mobility, aligned crystalline domains of tips-pentacene with metastable polymorphs through lateral confinement of crystal growth. Adv. Mater. 26, 487–493.

572

Handbook of Organic Materials for Electronic and Photonic Devices

Giri, G., Verploegen, E., Mannsfeld, S.C.B., Atahan-Evrenk, S., Kim, D.H., Lee, S.Y., Becerril, H.a., Aspuru-Guzik, A., Toney, M.F., Bao, Z., 2011. Tuning charge transport in solution-sheared organic semiconductors using lattice strain. Nature 480, 504–508. Grau, G., Cen, J., Kang, H., Kitsomboonloha, R., Scheideler, W.J., Subramanian, V., 2016. Gravure-printed electronics: recent progress in tooling development, understanding of printing physics, and realization of printed devices. Flex. Print. Electron. 1, 023002. Grau, G., Subramanian, V., 2016. Fully high-speed gravure printed, low-variability, highperformance organic polymer transistors with sub-5 V operation. Adv. Electron. Mater. 2, 1500328. Groenendaal, L., Jonas, F., Freitag, D., Pielartzik, H., Reynolds, J.R., 2000. Poly(3,4-ethylenedioxythiophene) and its derivatives. Past, present, and future. Adv. Mater. 12, 481–494. Gu, X., Shaw, L., Gu, K., Toney, M.F., Bao, Z., 2018. The meniscus-guided deposition of semiconducting polymers. Nat. Commun. 9, 534. Gu, X., Zhou, Y., Gu, K., Kurosawa, T., Guo, Y., Li, Y., Lin, H., Schroeder, B.C., Yan, H., Molina-Lopez, F., Tassone, C.J., Wang, C., Mannsfeld, S.C.B., Yan, H., Zhao, D., Toney, M.F., Bao, Z., 2017. Roll-to-roll printed large-area all-polymer solar cells with 5% efficiency based on a low crystallinity conjugated polymer blend. Adv. Energy Mater. 7, 1602742. Haase, K., Teixeira da Rocha, C., Hauenstein, C., Zheng, Y., Hambsch, M., Mannsfeld, S.C.B., 2018. High-mobility, solution-processed organic field-effect transistors from C8-BTBT: polystyrene blends. Adv. Electron. Mater. 1800076. Hambsch, M., Reuter, K., Stanel, M., Schmidt, G., Kempa, H., F€ ugmann, U., Hahn, U., H€ubler, A.C.C., 2010. Uniformity of fully gravure printed organic field-effect transistors. Mater. Sci. Eng. B 170, 93–98. Han, Y., Barnes, G., Lin, Y.-H., Martin, J., Al-Hashimi, M., AlQaradawi, S.Y., Anthopoulos, T.D., Heeney, M., 2016. Doping of large ionization potential indenopyrazine polymers via Lewis acid complexation with tris(pentafluorophenyl)borane: a simple method for improving the performance of organic thin-film transistors. Chem. Mater. 28, 8016–8024. He, D., Zhang, Y., Wu, Q., Xu, R., Nan, H., Liu, J., Yao, J., Wang, Z., Yuan, S., Li, Y., Shi, Y., Wang, J., Ni, Z., He, L., Miao, F., Song, F., Xu, H., Watanabe, K., Taniguchi, T., Xu, J.-B., Wang, X., 2014. Two-dimensional quasi-freestanding molecular crystals for highperformance organic field-effect transistors. Nat. Commun. 5, 5162. Hecht, D.S., Heintz, A.M., Lee, R., Hu, L., Moore, B., Cucksey, C., Risser, S., 2011a. High conductivity transparent carbon nanotube films deposited from superacid. Nanotechnology. 22075201. Hecht, D.S., Hu, L., Irvin, G., 2011b. Emerging transparent electrodes based on thin films of carbon nanotubes, graphene, and metallic nanostructures. Adv. Mater. 23, 1482–1513. Hu, L., Liu, T., Duan, J., Ma, X., Ge, C., Jiang, Y., Qin, F., Xiong, S., Jiang, F., Hu, B., Gao, X., Yi, Y., Zhou, Y., 2017. An amidine-type n-dopant for solution-processed field-effect transistors and perovskite solar cells. Adv. Funct. Mater. 27, 1703254. Huseynova, G., Xu, Y., Nketia Yawson, B., Shin, E.-Y., Lee, M.J., Noh, Y.-Y., 2016. P-type doped ambipolar polymer transistors by direct charge transfer from a cationic organic dye Pyronin B ferric chloride. Org. Electron. 39, 229–235. Janneck, R., Pilet, N., Bommanaboyena, S.P., Watts, B., Heremans, P., Genoe, J., Rolin, C., 2017. Highly crystalline C8-BTBT thin-film transistors by lateral homo-epitaxial growth on printed templates. Adv. Mater. 29, 1703864. Jo, J.-W., Kim, J., Kim, K.-T., Kang, J.-G., Kim, M.-G., Kim, K.-H., Ko, H., Kim, Y.-H., Park, S.K., 2015. Highly stable and imperceptible electronics utilizing photoactivated

Advances in solution processing of organic materials for devices

573

heterogeneous sol-gel metal-oxide dielectrics and semiconductors. Adv. Mater. 27, 1182–1188. Kang, H., Kitsomboonloha, R., Jang, J., Subramanian, V., 2012. High-performance printed transistors realized using femtoliter gravure-printed sub-10 μm metallic nanoparticle patterns and highly uniform polymer dielectric and semiconductor layers. Adv. Mater. 24, 3065–3069. Karpov, Y., Erdmann, T., Raguzin, I., Al-Hussein, M., Binner, M., Lappan, U., Stamm, M., Gerasimov, K.L., Beryozkina, T., Bakulev, V., Anokhin, D.V., Ivanov, D.A., G€unther, F., Gemming, S., Seifert, G., Voit, B., Di Pietro, R., Kiriy, A., 2016. High conductivity in molecularly p-doped diketopyrrolopyrrole-based polymer: the impact of a high dopant strength and good structural order. Adv. Mater. 28, 6003–6010. Karpov, Y., Erdmann, T., Stamm, M., Lappan, U., Guskova, O., Malanin, M., Raguzin, I., Beryozkina, T., Bakulev, V., G€unther, F., Gemming, S., Seifert, G., Hambsch, M., Mannsfeld, S., Voit, B., Kiriy, A., 2017. Molecular doping of a high mobility diketopyrrolopyrrole–dithienylthieno[3,2-b]thiophene donor–acceptor copolymer with F6TCNNQ. Macromolecules 50, 914–926. Karpov, Y., Kiriy, N., Al-Hussein, M., Hambsch, M., Beryozkina, T., Bakulev, V., Mannsfeld, S.C.B., Voit, B., Kiriy, A., 2018. Hexacyano-[3]-radialene anion-radical salts: a promising family of highly soluble p-dopants. Chem. Commun. 54, 307–310. Kempa, H., Hambsch, M., Reuter, K., Stanel, M., Schmidt, G.C., Meier, B., H€ ubler, A.C., Hubler, A.C., 2011. Complementary ring oscillator exclusively prepared by means of gravure and flexographic printing. IEEE Trans. Electron Devices 58, 2765–2769. Kheradmand-Boroujeni, B., Schmidt, G.C., Hoft, D., Bellmann, M., Haase, K., Ishida, K., Shabanpour, R., Meister, T., Carta, C., Ghesquiere, P., Hubler, A.C., Ellinger, F., 2016. A fully-printed self-biased polymeric audio amplifier for driving fully-printed piezoelectric loudspeakers. IEEE Trans. Circuits Syst. I Regul. Pap. 63, 785–794. Khim, D., Baeg, K.-J., Caironi, M., Liu, C., Xu, Y., Kim, D.-Y., Noh, Y.-Y., 2014. Control of ambipolar and unipolar transport in organic transistors by selective inkjet-printed chemical doping for high performance complementary circuits. Adv. Funct. Mater. 24, 6252–6261. Khim, D., Xu, Y., Baeg, K.-J., Kang, M., Park, W.-T., Lee, S.-H., Kim, I.-B., Kim, J., Kim, D.-Y., Liu, C., Noh, Y.-Y., 2016. Large enhancement of carrier transport in solution-processed field-effect transistors by fluorinated dielectric engineering. Adv. Mater. 28, 518–526. Kim, N., Kee, S., Lee, S.H., Lee, B.H., Kahng, Y.H., Jo, Y.-R., Kim, B.-J., Lee, K., 2014. Highly conductive PEDOT:PSS nanofibrils induced by solution-processed crystallization. Adv. Mater. 26, 2268–2272. Kim, S.J., Jang, M., Yang, H.Y., Cho, J., Lim, H.S., Yang, H., Lim, J.A., 2017. Instantaneous pulsed-light cross-linking of a polymer gate dielectric for flexible organic thin-film transistors. ACS Appl. Mater. Interfaces 9, 11721–11731. Le Berre, M., Chen, Y., Baigl, D., 2009. From convective assembly to LandauLevich deposition of multilayered phospholipid films of controlled thickness. Langmuir 25, 2554–2557. Lee, E.K., Kim, J.Y., Chung, J.W., Lee, B.-L., Kang, Y., 2014. Photo-crosslinkable polymer gate dielectrics for hysteresis-free organic field-effect transistors with high solvent resistance. RSC Adv. 4, 293–300. Lee, W.-J., Park, W.-T., Park, S., Sung, S., Noh, Y.-Y., Yoon, M.-H., 2015. Large-scale precise printing of ultrathin sol-gel oxide dielectrics for directly patterned solution-processed metal oxide transistor arrays. Adv. Mater. 27, 5043–5048.

574

Handbook of Organic Materials for Electronic and Photonic Devices

Li, S., Feng, L., Zhao, J., Guo, X., Zhang, Q., 2015. Low temperature cross-linked, high performance polymer gate dielectrics for solution-processed organic field-effect transistors. Polym. Chem. 6, 5884–5890. Liang, Y., Wang, Y., Mu, C., Wang, S., Wang, X., Xu, D., Sun, L., 2018. Achieving high opencircuit voltages up to 1.57 V in hole-transport-material-free MAPbBr 3 solar cells with carbon electrodes. Adv. Energy Mater. 8, 1701159. Lin, F., Guo, C., Chuang, W., Wang, C.-L., Wang, Q., Liu, H., Hsu, C.-S., Jiang, L., 2017. Directional solution coating by the Chinese brush: a facile approach to improving molecular alignment for high-performance polymer TFTs. Adv. Mater. 29, 1606987. Liu, C., Jang, J., Xu, Y., Kim, H.-J., Khim, D., Park, W.-T., Noh, Y.-Y., Kim, J.-J., 2015. Effect of doping concentration on microstructure of conjugated polymers and characteristics in N-type polymer field-effect transistors. Adv. Funct. Mater. 25, 758–767. Ma, W., Reinspach, J., Zhou, Y., Diao, Y., McAfee, T., Mannsfeld, S.C.B., Bao, Z., Ade, H., 2015. Tuning local molecular orientation-composition correlations in binary organic thin films by solution shearing. Adv. Funct. Mater. 25, 3131–3137. Mei, A., Li, X., Liu, L., Ku, Z., Liu, T., Rong, Y., Xu, M., Hu, M., Chen, J., Yang, Y., Gratzel, M., Han, H., 2014. A hole-conductor-free, fully printable mesoscopic perovskite solar cell with high stability. Science (80) 345, 295–298. Minemawari, H., Yamada, T., Matsui, H., Tsutsumi, J., Haas, S., Chiba, R., Kumai, R., Hasegawa, T., 2011. Inkjet printing of single-crystal films. Nature 475, 364–367. Molina-Lopez, F., Yan, H., Gu, X., Kim, Y., Toney, M.F., Bao, Z., 2017. Electric field tuning molecular packing and electrical properties of solution-shearing coated organic semiconducting thin films. Adv. Funct. Mater. 27, 1605503. Naab, B.D., Gu, X., Kurosawa, T., J. W. F. To, Salleo, A., Bao, Z., 2016. Role of polymer structure on the conductivity of N-doped polymers. Adv. Electron. Mater. 2, 1600004. Naab, B.D., Guo, S., Olthof, S., Evans, E.G.B., Wei, P., Millhauser, G.L., Kahn, A., Barlow, S., Marder, S.R., Bao, Z., 2013b. Mechanistic study on the solution-phase n-doping of 1,3dimethyl-2-aryl-2,3-dihydro-1 H-benzoimidazole derivatives. J. Am. Chem. Soc. 135, 15018–15025. Naab, B.D., Himmelberger, S., Diao, Y., Vandewal, K., Wei, P., Lussem, B., Salleo, A., Bao, Z., 2013a. High mobility N-type transistors based on solution-sheared doped 6,13-bis (triisopropylsilylethynyl)pentacene thin films. Adv. Mater. 25, 4663–4667. Nair, I.I.J., Varma, M.R., Sebastian, M.T., 2016. Low cost room temperature curable alumina ink for printed electronic applications. J. Mater. Sci. Mater. Electron. 27, 9891–9899. Niazi, M.R., Li, R., Abdelsamie, M., Zhao, K., Anjum, D.H., Payne, M.M., Anthony, J., Smilgies, D.-M.M., Amassian, A., 2016. Contact-induced nucleation in high-performance bottom-contact organic thin film transistors manufactured by large-area compatible solution processing. Adv. Funct. Mater. 26, 2371–2378. Niazi, M.R., Li, R., Li, E.Q., Kirmani, A.R., Abdelsamie, M., Wang, Q., Pan, W., Payne, M.M., Anthony, J.E., Smilgies, D.-M., Thoroddsen, S.T., Giannelis, E.P., Amassian, A., Qiang Li, E., Kirmani, A.R., Abdelsamie, M., Wang, Q., Pan, W., Payne, M.M., Anthony, J.E., Smilgies, D.-M., Thoroddsen, S.T., Giannelis, E.P., Amassian, A., 2015. Solution-printed organic semiconductor blends exhibiting transport properties on par with single crystals. Nat. Commun. 6, 8598. Nikolka, M., Nasrallah, I., Rose, B., Ravva, M.K., Broch, K., Sadhanala, A., Harkin, D., Charmet, J., Hurhangee, M., Brown, A., Illig, S., Too, P., Jongman, J., McCulloch, I., Bredas, J.-L., Sirringhaus, H., 2016. High operational and environmental stability of high-mobility conjugated polymer field-effect transistors through the use of molecular additives. Nat. Mater. 16, 356–362.

Advances in solution processing of organic materials for devices

575

Noh, Y.-J.J., Kim, S.-S.S., Kim, T.-W.W., Na, S.-I.I., 2014. Cost-effective ITO-free organic solar cells with silver nanowire–PEDOT:PSS composite electrodes via a one-step spray deposition method. Sol. Energy Mater. Sol. Cells 120, 226–230. Novoselov, K.S., Geim, A.K., Morozov, S.V., Jiang, D., Zhang, Y., Dubonos, S.V., Grigorieva, I.V., Firsov, A.A., 2004. Electric field effect in atomically thin carbon films. Science (80) 306, 666–669. Panidi, J., Paterson, A.F., Khim, D., Fei, Z., Han, Y., Tsetseris, L., Vourlias, G., Patsalas, P.a., Heeney, M., Anthopoulos, T.D., 2018. Remarkable enhancement of the hole mobility in several organic small-molecules, polymers, and small-molecule:polymer blend transistors by simple admixing of the Lewis acid p-dopant B(C6F5)3. Adv. Sci. 5, 1700290. Park, S., Giri, G., Shaw, L., Pitner, G., Ha, J., Koo, J.H., Gu, X., Park, J., Lee, T.H., Nam, J.H., Hong, Y., Bao, Z., 2015. Large-area formation of self-aligned crystalline domains of organic semiconductors on transistor channels using CONNECT. Proc. Natl. Acad. Sci. 112, 5561–5566. Park, S., Kim, C.-H., Lee, W.-J., Sung, S., Yoon, M.-H., 2017. Sol-gel metal oxide dielectrics for all-solution-processed electronics. Mater. Sci. Eng. R Rep. 114, 1–22. Park, Y.M., Daniel, J., Heeney, M., Salleo, A., 2011. Room-temperature fabrication of ultrathin oxide gate dielectrics for low-voltage operation of organic field-effect transistors. Adv. Mater. 23, 971–974. Pingel, P., Arvind, M., K€olln, L., Steyrleuthner, R., Kraffert, F., Behrends, J., Janietz, S., Neher, D., 2016. p-Type doping of poly(3-hexylthiophene) with the strong Lewis acid tris(pentafluorophenyl)borane. Adv. Electron. Mater. 2, 1600204. Ro, H.W., Downing, J.M., Engmann, S., Herzing, A.A., DeLongchamp, D.M., Richter, L.J., Mukherjee, S., Ade, H., Abdelsamie, M., Jagadamma, L.K., Amassian, A., Liu, Y., Yan, H., 2016. Morphology changes upon scaling a high-efficiency, solution-processed solar cell. Energy Environ. Sci. 9, 2835–2846. Rossbauer, S., M€uller, C., Anthopoulos, T.D., 2014. Comparative study of the N-type doping efficiency in solution-processed fullerenes and fullerene derivatives. Adv. Funct. Mater. 24, 7116–7124. Schmidt, G.C., Bellmann, M., Meier, B., Hambsch, M., Reuter, K., Kempa, H., H€ ubler, A.C., 2010. Modified mass printing technique for the realization of source/drain electrodes with high resolution. Org. Electron. 11, 1683–1687. Schmidt, G.C., H€oft, D., Bhuie, M., Haase, K., Bellmann, M., Haidu, F., Lehmann, D., Zahn, D.R.T., H€ubler, A.C., 2013. Modified poly(3,4-ethylenedioxythiophene):poly (styrenesulfonate) source/drain electrodes for fully printed organic field-effect transistors consisting of a semiconductor blend. Appl. Phys. Lett. 103, 113302. Schmidt, G.C., H€oft, D., Haase, K., Bellmann, M., Kheradmand-Boroujeni, B., Hassinen, T., Sandberg, H., Ellinger, F., H€ubler, A.C., 2015. Fully printed flexible audio system on the basis of low-voltage polymeric organic field effect transistors with three layer dielectric. J. Polym. Sci. Part B Polym. Phys. 53, 1409–1415. Schott, S., Gann, E., Thomsen, L., Jung, S.-H.H., Lee, J.K., McNeill, C.R., Sirringhaus, H., 2015. Charge-transport anisotropy in a uniaxially aligned diketopyrrolopyrrole-based copolymer. Adv. Mater. 27, 7356–7364. Sele, C.W., von Werne, T., Friend, R.H., Sirringhaus, H., 2005. Lithography-free, self-aligned inkjet printing with sub-hundred-nanometer resolution. Adv. Mater. 17, 997–1001. Søndergaard, R., H€osel, M., Angmo, D., Larsen-Olsen, T.T., Krebs, F.C., 2012. Roll-to-roll fabrication of polymer solar cells. Mater. Today 15, 36–49. Søndergaard, R.R., H€osel, M., Krebs, F.C., 2013. Roll-to-roll fabrication of large area functional organic materials. J. Polym. Sci. Part B Polym. Phys. 51, 16–34.

576

Handbook of Organic Materials for Electronic and Photonic Devices

Sun, B., Hong, W., Thibau, E.S., Aziz, H., Lu, Z.-H., Li, Y., 2015. Polyethylenimine (PEI) as an effective dopant to conveniently convert ambipolar and p-type polymers into unipolar n-type polymers. ACS Appl. Mater. Interfaces 7, 18662–18671. Tang, W., Li, J., Zhao, J., Zhang, W., Yan, F., Guo, X., 2015. High-performance solutionprocessed low-voltage polymer thin-film transistors with low-k/high-k bilayer gate dielectric. IEEE Electron Device Lett. 36, 950–952. Tang, W., Zhao, J., Huang, Y., Ding, L., Li, Q., Li, J., You, P., Yan, F., Guo, X., 2017. Bias stress stability improvement in solution-processed low-voltage organic field-effect transistors using relaxor ferroelectric polymer gate dielectric. IEEE Electron Device Lett. 38, 748–751. Teixeira da Rocha, C., Haase, K., Zheng, Y., L€offler, M., Hambsch, M., Mannsfeld, S.C.B., 2018. Solution coating of small molecule/polymer blends enabling ultralow voltage and high-mobility organic transistors. Adv. Electron. Mater. 1800141. Tetzner, K., Schroder, K.A., Bock, K., 2014. Photonic curing of sol–gel derived HfO2 dielectrics for organic field-effect transistors. Ceram. Int. 40, 15753–15761. Tseng, H., Ying, L., Hsu, B.B.Y., Perez, L.A., Takacs, C.J., Bazan, G.C., Heeger, A.J., 2012. Mobility field effect transistors based on macroscopically oriented regioregular copolymers. Nano Lett. 12, 6353–6357. Vaklev, N.L., M€uller, R., Muir, B.V.O., James, D.T., Pretot, R., van der Schaaf, P., Genoe, J., Kim, J.-S., Steinke, J.H.G., Campbell, A.J., 2014. High-performance flexible bottom-gate organic field-effect transistors with gravure printed thin organic dielectric. Adv. Mater. Interfaces. 1, 1300123. van de Craats, A.M., Stutzmann, N., Bunk, O., Nielsen, M.M., Watson, M., M€ ullen, K., Chanzy, H.D., Sirringhaus, H., Friend, R.H., 2003. Meso-epitaxial solution-growth of self-organizing discotic liquid-crystalline semiconductors. Adv. Mater. 15, 495–499. Wang, C., Lee, W., Nakajima, R., Mei, J., Kim, D.H., Bao, Z., 2013. Thiol–ene cross-linked polymer gate dielectrics for low-voltage organic thin-film transistors. Chem. Mater. 25, 4806–4812. Wang, Y., Huang, X., Li, T., Wang, Z., Li, L., Guo, X., Jiang, P., 2017. Novel crosslinkable high-k copolymer dielectrics for high-energy-density capacitors and organic field-effect transistor applications. J. Mater. Chem. A. 5, 20737–20746. Wei, P., Oh, J.H., Dong, G., Bao, Z., 2010. Use of a 1 H-Benzoimidazole derivative as an n -type dopant and to enable air-stable solution-processed n-channel organic thin-film transistors. J. Am. Chem. Soc. 132, 8852–8853. Worfolk, B.J., Andrews, S.C., Park, S., Reinspach, J., Liu, N., Toney, M.F., Mannsfeld, S.C.B., Bao, Z., 2015. Ultrahigh electrical conductivity in solution-sheared polymeric transparent films. Proc. Natl. Acad. Sci. 112, 14138–14143. Wu, Z., Chen, Z., Du, X., Logan, J.M., Sippel, J., Nikolou, M., Kamaras, K., Reynolds, J.R., Tanner, D.B., Hebard, A.F., Rinzler, A.G., 2004. Transparent, conductive carbon nanotube films. Science (80) 305, 1273–1276. Xia, Y., Sun, K., Ouyang, J., 2012. Solution-processed metallic conducting polymer films as transparent electrode of optoelectronic devices. Adv. Mater. 24, 2436–2440. Xu, C., He, P., Liu, J., Cui, A., Dong, H., Zhen, Y., Chen, W., Hu, W., 2016. A general method for growing two-dimensional crystals of organic semiconductors by “Solution Epitaxy.” Angew. Chemie 128, 9671–9675. Xu, M., Xiang, L., Xu, T., Wang, W., Xie, W., Zhou, D., 2017. Low-voltage operating flexible ferroelectric organic field-effect transistor nonvolatile memory with a vertical phase separation P(VDF-TrFE-CTFE)/PS dielectric. Appl. Phys. Lett. 111, 183302. Yang, J., Yan, D., 2009. Weak epitaxy growth of organic semiconductor thin films. Chem. Soc. Rev. 38, 2634.

Advances in solution processing of organic materials for devices

577

Ye, X., Lin, H., Yu, X., Han, S., Shang, M., Zhang, L., Jiang, Q., Zhong, J., 2015. High performance low-voltage organic field-effect transistors enabled by solution processed alumina and polymer bilayer dielectrics. Synth. Met. 209, 337–342. Zhang, F., Dai, X., Zhu, W., Chung, H., Diao, Y., 2017b. Large modulation of charge carrier mobility in doped nanoporous organic transistors. Adv. Mater. 29, 1700411. Zhang, F., Qu, G., Mohammadi, E., Mei, J., Diao, Y., 2017c. Solution-processed nanoporous organic semiconductor thin films: toward health and environmental monitoring of volatile markers. Adv. Funct. Mater. 27, 1701117. Zhang, Y., Zhuang, X., Su, Y., Zhang, F., Feng, X., 2014. Polyaniline nanosheet derived B/N co-doped carbon nanosheets as efficient metal-free catalysts for oxygen reduction reaction. J. Mater. Chem. A 2, 7742. Zhang, Z., Peng, B., Ji, X., Pei, K., Chan, P.K.L., 2017a. Marangoni-effect-assisted bar-coating method for high-quality organic crystals with compressive and tensile strains. Adv. Funct. Mater. 27, 1703443. Zhu, L., Wang, Q., 2012. Novel ferroelectric polymers for high energy density and low loss dielectrics. Macromolecules 45, 2937–2954.

This page intentionally left blank

Advances in device fabrication scale-up methods

18

Tatsuo Hasegawa Department of Applied Physics, The University of Tokyo, Tokyo, Japan

18.1

Introduction

18.1.1 Printed electronics toward scale-up production of organic electronic devices Studies of organic electronic and photonic materials have had a new outlook in the last decade because of broad interests to utilize solution- or printing-based technologies for the scale-up production of a variety of electronic circuits and devices with these materials (Ogawa, 2015). This area of research is frequently referred to as printed electronics, which basically relies on the solution-processable nature of organic electronic materials under ambient conditions. It is highly expected that printed electronics technology enables the production of large-area and flexible electronic and photonic devices without the use of huge vacuum facilities and with the adoption of roll-to-roll processes, which should lead to considerably low material and energy consumption during the production. Thus, it is expected to revolutionize the future electronics industry to accelerate prevailing flexible electronics devices such as lightweight and wearable human interface or energy-harvesting devices. The most fundamental and ultimate target in printed electronics technology is to produce active-matrix (AM) backplanes, composed of an array of huge number of printed thin-film transistors (TFTs) (Fukuda et al., 2014; Sakai et al., 2015; Bucella et al., 2015; Tsutsumi et al., 2018). The AM backplane allows versatile uses in large-area electronics products, such as displays, electronic papers, electronic skins, or sensor sheets, with controlling the on/off switching of each pixel (Dimitrakopoulos and Malenfant, 2002; Gelinck et al., 2010; Noda et al., 2011; Steudel et al., 2012; Kaltenbrunner et al., 2013; Hammock et al., 2013). Thus, the scale-up production, in which the pixel number often reaches far more than a million and corresponds to the case of a 14-inch device with 100 pixel-per-inch (ppi) resolution, is required for the printed electronics technologies of manufacturing printed TFTs. Here, we briefly describe a particular example for the scale-up production of printed TFTs based on organic semiconductors (OSCs) and the resultant operation as the AM backplane (Tsutsumi et al., 2018). Fig. 18.1 presents the appearance, composition, and characteristics of a trial product of the A4-size flexible AM backplane, manufactured by a combination of printing technologies without a vacuum, on a polyethylene naphthalate sheet used as the substrate. Each AM pixel is composed of one-transistor-one-capacitor cell, arranged at a Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00018-8 © 2019 Elsevier Ltd. All rights reserved.

580

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 18.1 An A4-size flexible AM backplane and their TFT characteristics. (A) Whole image, (B) optical micrograph, (C) circuit diagram, (D) expanded micrograph, and (E) schematic cross section for each pixel, as well as (F) transfer characteristics and (G) histogram of the mobility. Reproduced from Figs. 1 and 2 in Tsutsumi, J., Matsuoka, S., Kamata, T., Hasegawa, T., 2018. Fast optical inspection of operations of large-area active-matrix backplane by gate modulation imaging. Org. Electron. 55, 187–193.

150-ppi resolution. All the conductive wires and electrodes were fabricated by a reverse offset printing technique, while the semiconductor layers were deposited and patterned by an inkjet-printing technique. The respective TFTs, with a top-gate, bottom-contact configuration, is composed of successively overlaid layers—silver source/drain electrodes, p-type polymer semiconductor layers, polymer gate insulator layers, silver gate electrodes, and passivation layers. A class of donor-acceptor (DA) copolymer OSCs is used to form the channel layer. The average mobility is estimated at 0.42 cm2/Vs with a standard deviation of 0.04 cm2/Vs in the saturation regime, and at 0.29 cm2/Vs with a standard deviation of 0.07 cm2/Vs in the linear regime. The on/off current ratio reaches as high as 106, and the subthreshold slope is about 0.4 V per decade, with a threshold voltage of about 2 V under the saturation regime. An advanced mass-inspection technique is used to investigate the operation for the above product. Fig. 18.2 demonstrates the operation of the A4-size AM backplane, visualized by the gate-modulation imaging technique (Tsutsumi et al., 2018, 2015). The technique is based on a difference image sensing of the observed optical microscopic images between at alternately biased gate-on and gate-off states, which

Advances in device fabrication scale-up methods

581

Fig. 18.2 Gate-modulation imaging experiment. (A) A schematic for the setup of the gatemodulation imaging technique, (B) large-area optical micrograph, and (B) the corresponding gate-modulation image for an A4-size AM backplane, measured under the transistor inspection mode. Reproduced from Figs. 3 and 4 in Tsutsumi, J., Matsuoka, S., Kamata, T., Hasegawa, T., 2018. Fast optical inspection of operations of large-area active-matrix backplane by gate modulation imaging. Org. Electron. 55, 187–193.

eventually allows the fast and highly sensitive detection of accumulated charges in the channels of respective TFTs. Actually, the obtained difference image [i.e., the gatemodulation image shown in Fig. 18.2C] is composed of an array of red dots over the whole microscope view inside the central circle, clearly demonstrating the normal operation of respective organic TFTs (OTFTs). As the area of the circle is 4 cm2, as many as 15,000 pixels can be inspected simultaneously within 3 min. The OTFTs with no gate-modulation signal pixels are also clearly observed from the gatemodulation image, which can be ascribed to the defective devices. The production with elimination or minimization of these defects should be an important challenge for the mass production of scale-up devices. A possible future role of these technologies is to decorate any plastic surfaces electronically by arraying a large number of printed TFTs that function as the AM backplane for displays or sensors in realizing the “ubiquitous electronics” society. Nonetheless, these products may still be far from satisfactory, in terms of both the device performance and the throughput as required for mass production and popularization. In the subsequent sections of this chapter, we discuss the further intriguing scientific and technological subjects and developments to realize innovations with printed electronics technologies.

582

Handbook of Organic Materials for Electronic and Photonic Devices

18.1.2 Difference between document printing and production of electronic devices In this section, we describe general problems encountered in the course of printing processes for manufacturing electronic devices. The printing technology can be generally defined as a process to allocate and deposit microdroplet ink on substrate surfaces, followed by spreading, drying, and solidification. The respective fluidic processes can cause serious problems in order to form and stack high-resolution patterned layers of electronic functional materials that have enough high uniformity at atomic or molecular scales. The situation is in sharp contrast to the conventional purposes of printing technologies because the crystallinity, grain size, or grain boundaries of coloring matter are not a matter of concern after the deposited inks are dried out, so long as the coloring matter is precisely positioned and deposited. First, consideration should be given to the wettability (or surface energy) of the substrate surface. The spreading of fluidic ink is considerably affected by the surface wettability. The manipulation of fluidic ink becomes increasingly difficult on a highly hydrophobic surface. It causes a serious dilemma for the print production of semiconductor layers because the use of highly hydrophobic gate-insulator surface is quite effective to improve the device characteristics of OTFTs, as is associated with the elimination of charge traps (Bl€ ulle et al., 2014). This subject is discussed later in this chapter. On the other hand, it is possible to produce surface-energy (i.e., hydrophilic/hydrophobic) patterned surfaces by a variety of surface chemical modification techniques. One technique is traditionally utilized as lithographic printing to obtain high resolution of printed patterns. It allows to control the ink coverage on the surface, based on a simple principle that the high-surface-energy (or the hydrophilic) area is primarily wetted. Several reports demonstrated that predefined patterned surface-energy modification can be used to attain fine thin-film patterning for manufacturing printed electronics circuits (Noda et al., 2013). Nonetheless, we also have to know that the ink coverage does not necessarily coincide with the predefined surface-energy pattern, especially in the case of ultrafine or rather complicated patterning (Noda et al., 2013). An example is depicted in Fig. 18.3. The origin of discrepancy is ascribed to the competition between the surface tension of microdroplets and surface energy of the patterned substrate surface. An efficient algorithm was recently developed to simulate the equilibrium microdroplet shape on the surface-energy pattered substrate (Matsui et al., 2012). Note that the feature strongly limits the resolution of printed pattern, so long as the physisorption effect of the fluidic ink on the surface is utilized. Another problem in the print production of electronic devices originates from the evaporation of the solvent (or the dispersant) from the solution (or the colloid). According to the fluid science, solvent (or dispersant) evaporation occurs efficiently at around the solid-liquid-air contact line of the sessile droplet (de Gennes and Brochard-Wyart, 2004). The resultant outward capillary flow carries solute (or dispersoid) toward the contact line of droplets, leading to the formation of ring-like deposits, as presented in Fig. 18.4, which is known as the coffee-ring effect (Singh et al., 2010; Lim et al., 2008). So it often becomes extremely difficult to form uniform thin, solid

Advances in device fabrication scale-up methods

583

Fig. 18.3 Experimental wetting and spreading characteristics of a fluid on surface-energypatterned substrate. Thin black solid lines represent the hydrophilic/hydrophobic boundary. Reproduced from Fig. 5 in Noda, Y., Matsui, H., Minemawari, H., Yamada, T., Hasegawa, T., 2013. Observation and simulation of microdroplet shapes on surface-energy-patterned substrates: contact line engineering for printed electronics. J. Appl. Phys. 114, 044905.

Fig. 18.4 Typical observation of the coffee-ring effect from a deposited sessile droplet. Reproduced from Fig. 7 in Noda, Y., Minemawari, H., Matsui, H., Yamada, T., Arai, S., Kajiya, T., Doi, M., Hasegawa, T., 2015. Underlying mechanism of inkjet printing of uniform organic semiconductor films through antisolvent crystallization. Adv. Funct. Mater. 25, 4022–4031.

584

Handbook of Organic Materials for Electronic and Photonic Devices

layers from a tiny volume of low-viscous droplets by precipitation (or sedimentation) within the droplet through solvent (or dispersant) evaporation, though this nonuniformity problem apparently may not be exposed in the conventional printing technologies, because the solvents (or dispersants) are efficiently absorbed into the cellulose fabrics in the case of using papers as the printed medium.

18.2

Printing of semiconductor layers

18.2.1 Materials—Soluble organic semiconductors (OSCs) Soluble OSCs play major roles in printed electronics technologies because relatively high performance TFTs can be obtained by the convenient solution-based thin-film processing for the materials under ambient conditions (Sirringhaus, 2014). These materials are basically composed of π-conjugated systems, such as fused acene or heteroacene. Fig. 18.5 presents well-known examples for the soluble OSCs that afford relatively high performance OTFTs. The materials are classified into small moleculebased OSCs (1–3) (Anthony et al., 2001; Anthony, 2006; Ebata et al., 2007; Takimiya et al., 2011; Iino et al., 2015) and conjugated polymer (CP)-based OSCs (4–6) (Sirringhaus et al., 1999; McCulloch et al., 2006; Zhang et al., 2013; Kronemeijer et al., 2012; Venkateshvaran et al., 2014). As the figure shows, electronically inactive substituents, such as the alkyl or triisopropylsilylethynyl group, are attached to the π-conjugated system. Carrier transport in these OSCs depends sensitively on the molecular structure, intermolecular packing motif, and structural or energetic disorder.

Fig. 18.5 Examples of soluble OSCs. 1: TIPS-pentacene, 2: diC8-BTBT, 3: Ph-BTBT-Cn, 4: P3HT, 5: PBTTT, 6: PDPPBT.

Advances in device fabrication scale-up methods

585

Small molecule-based OSCs usually exhibit higher crystallinity than CP-based ones. Thus, they afford the OTFTs showing higher performance, and their molecular packing motif is investigated and categorized in detail. A current subject is optimizing characteristics such as molecular packing motif, solvent solubility, and thermal stability of these materials by the application of a wide variety of chemical substitutions to the π-conjugated system. Useful CP-based OSCs also have been obtained traditionally with materials showing a semicrystalline nature, as presented in compounds 4 and 5 (Sirringhaus et al., 1999; McCulloch et al., 2006; Zhang et al., 2013). Recently, DA-type CPs, such as compound 6 (Kronemeijer et al., 2012; Venkateshvaran et al., 2014), have achieved a higher performance than the semicrystalline ones, although they are amorphous or have less crystalline order. There should still be a wide range of promising molecular materials to be improved and developed.

18.2.2 Techniques for printing semiconductor layers Several recent reports have demonstrated that some printing- or solution-based thinfilm processing is more suitable to afford OTFTs showing higher performance than vacuum-based thin-film processing. In these studies, several processing conditions are tuned to optimize the production of uniform channel layers with flat semiconductor-insulator interfaces to facilitate efficient two-dimensional (2D) carrier transport along the channel layers. Among them, spin coating is widely accepted as a standard technique to form polycrystalline thin films for the initial check of applicability of the soluble OSCs. The spin-coating process consists of a production of thin solution layers on top of the substrate surface by centrifugal force as a result of spinning of substrates, and of a subsequent formation of thin solid films by uniform solvent evaporation from the entire liquid-air interfaces (Krebs, 2009; Diao et al., 2014). This mechanism may be reasonable for the spontaneous formation of uniform thin films for the materials showing either amorphous or layered crystallinity. Actually, the production of thin solution layers is the common elementary step for the soluble OSCs that show high layered crystallinity, as discussed later in this chapter. Nonetheless, the use of spin coating is limited for the scale-up production because of the large material loss and also the difficulty in controlling film growth. The blade-coating technique, as schematically shown in Fig. 18.6, is convenient and widely used to fabricate uniform OSC films. The technique is a class of meniscus-guided coating that utilizes the evolution of solution meniscus (or the interface between air and the thin solution layer) formed between the substrate and the coating blade (Diao et al., 2014). The solvent evaporates mostly from the meniscus around which the solution is concentrated. When the solution reaches supersaturation, the OSCs begin to crystallize. The crystallization process may allow the alignment of orientation for the obtained layered crystalline films and often affords singlecrystalline films (Uemura et al., 2009; Giri et al., 2011; Diao et al., 2013; Hamai et al., 2017; Yamamura et al., 2018; Arai et al., 2018). The technique in principle is the same as slot-die coating, solution shearing, edge casting, or dip-coating. The push-coating technique, as schematically shown in Fig. 18.7, is also developed as an alternative to spin coating (Ikawa et al., 2012). The technique uses viscoelastic

586

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 18.6 Schematic of the blade-coating technique. Reproduced from Fig. 1 in Hamai, T., Arai, S., Minemawari, H., Inoue, S., Kumai, R., Hasegawa, T., 2017. Tunneling and origin of large access resistance in layered-crystal organic transistors. Phys. Rev. Appl. 8, 054011.

Fig. 18.7 (A) Schematic of the push-coating technique, and (B) thin-film patterning after the coating process. Reproduced from Figs. 1 and 3, in Ikawa, M., Yamada, T., Matsui, H., Minemawari, H., Tsutsumi, J., Horii, Y., Chikamatsu, M., Azumi, R., Kumai, R., Hasegawa, T., 2012. Simple push coating of polymer thin-film transistors. Nat. Commun. 3, 1176.

stamp composed of polydimethylsiloxane (PDMS). A uniform thin solution layer is first formed by spreading a tiny amount of semiconductor solution between the stamp and the substrate. Solvent is then slowly extracted from the thin solution layer and absorbed by the stamp, eventually forming a thin solid film on the substrate surface. The process is useful to fabricate uniform CP-based semiconductor films on highly hydrophobic substrate surfaces. However, it is hardly applicable to the small molecule–based semiconductors because they are also adsorbed into the stamp as

Advances in device fabrication scale-up methods

587

Fig. 18.8 Schematic of the DS-IJP technique. Reproduced from Fig. 1 in Minemawari, H., Yamada, T., Matsui, H., Tsutsumi, J., Haas, S., Chiba, R., Kumai, R., Hasegawa, T., 2011. Inkjet printing of single-crystal films. Nature 475, 364.

solute materials. A thin-film patterning method with use of the inverse printing method was also proposed, as presented in Fig. 18.7. The double-shot inkjet printing (DS-IJP) technique, as schematically shown in Fig. 18.8, is a novel solution process to use two kinds of microdroplets (Minemawari et al., 2011). The technique utilizes two inkjet heads, which allows the incorporation of the concept of antisolvent crystallization into the microdroplet process. In the process, an antisolvent microdroplet is first deposited on substrates, and then the semiconductor solution microdroplet is overdeposited at the same position, to form mixed sessile droplets on the substrate surfaces. The time of occurrence is divided by the technique between molecular crystallization and solvent evaporation. Furthermore, the process allows for obtaining single-crystal films over the whole area of liquid-air interfaces of the mixed droplets, which is in striking contrast to the case of conventional macroscopic antisolvent crystallization. It is demonstrated that the microdroplet phenomenon creates a thin solution layer on the antisolvent droplet surface, as the initial contact dynamics are driven by the difference of surface tension of the liquids (Noda et al., 2015). Then the gradual diffusion of liquid molecules within the thin solution layer affords an ideal field for the layered crystalline thin-film formation. The spray-coating technique, reported recently, should be quite similar to this process (Rigas et al., 2016).

18.2.3 Toward control of molecular self-organization in solution In this section, we first discuss the microscopic crystallization processes in solution and then the molecular designs of OSCs suitable for the crystallization processes, for the goal of further improving and developing the printing technologies of OSC films. It has been demonstrated, so far, that highly uniform polycrystalline or single-crystalline thin films could be obtained for some OSCs by the solution processes (Diao et al., 2014, 2013; Uemura et al., 2009; Giri et al., 2011; Hamai et al., 2017; Yamamura et al., 2018; Arai et al., 2018; Ikawa et al., 2012; Minemawari et al., 2011; Noda et al., 2015; Rigas et al., 2016). In the process, the film growth

588

Handbook of Organic Materials for Electronic and Photonic Devices

proceeds by a more self-organizing manner compared to the vacuum-based thin-film processing. In this respect, inherent characteristics as to the self-organizing nature into the layered crystallinity is crucial to achieve the uniformly extended layeredcrystalline channel layers, and so the solution process should be designed to promote the self-organization process by taking advantage of the inherent materials characteristics. Recent molecular dynamics (MD) simulation study reveals that the molecular aggregation process of layered-crystalline OSCs proceeds in a very different manner within the solution layer from the usual crystallization process (Yoneya et al., 2017). It is reported that the molecules undergo the spontaneous formation of lyotropic liquidcrystalline phase at the air-liquid interface through the diffusion and self-assembly before the solid crystalline layer formation. A typical example for the monomolecular layer formation of the compound 2 is illustrated in Fig. 18.9. An important point is that the air-liquid interface acts as an orientation field for the spontaneous aligned layer formation of molecules with their long axes perpendicular to the interface by forming lyotropic liquid-crystalline layers. It is also found that the lyotropic liquid-crystalline layer serves as a precursor for single-crystalline film growth. Because the air-liquid interface is ubiquitous in all the solution processes, this result provides a clue to understand why the spontaneous orientation and layered film growth are more promoted in the solution processes. The mechanism is also a reason why the uniaxially aligned crystalline film growth is achieved from the solution layer, which is actually much thicker than the obtained solid films (by about 30 times). Note that the crystallization

Fig. 18.9 MD simulation of diC8-BTBT (compound 2) in solution. (A) Initial structure and snapshots after (B) 1 ns, (C) 4 ns, (D) 10 ns, and (E) 50 ns. The monomolecular layer of diC8BTBT is formed due to the self-assemble process at the air-liquid interface, as shown by the dashed square in (E). Reproduced from Fig. 1 in Yoneya, M., Minemawari, H., Yamada, T., Hasegawa, T., 2017. Interface-mediated self-assembly in inkjet printing of single-crystal organic semiconductor films. J. Phys. Chem. C 121, 8796–8803.

Advances in device fabrication scale-up methods

589

Fig. 18.10 An LHB packing motif of Ph-BTBT-Cn (compound 3, n ¼ 10), viewed along the axis perpendicular to the molecular layer. Reproduced from Fig. 1 in Arai, S., Inoue, S., Hamai, T., Kumai, R., Hasegawa, T., 2018. Semiconductive single molecular bilayers realized using geometrical frustration. Adv. Mater. 30, 1707256.

mechanism is very different from the usual one, in which the nucleation and subsequent crystal growth dominates the crystallization process. The appearance of temporary lyotropic liquid-crystal phase is closely correlated with the layered herringbone (LHB) packing of the materials that is shown in Fig. 18.10. Note that LHB packing is composed of two types of intermolecular contacts—the T-shaped (core-to-edge) and slipped parallel (core-to-core) contacts between the neighboring planar π-electron skeleton, which is known to be the most suitable, among a variety of packing motifs, for the production of highperformance OTFTs (Campbell et al., 1961; Sundar et al., 2004; Yamamoto and Takimiya, 2007). To achieve the soluble OSCs, π-conjugated skeletons are usually modified by introducing alkyl chains. Interestingly, the alkylation not only affords high solvent solubility, but also enhances the layered crystallinity of the materials, according to the systematic studies of alkyl chain-length dependence of the packing motif for BTBT derivatives (i.e., compounds 2 and 3) (Inoue et al., 2015; Minemawari et al., 2017). The substitution of the π-conjugated system by relatively long alkyl chains leads to the highly layered-crystalline packing motif, where the π-electron skeleton and the alkyl chains form independent layers, respectively, as presented in Fig. 18.11. Furthermore, the asymmetric substitution of the π-conjugated system with long alkyl chains leads to the formation of bilayer-type LHB packing. The layered crystallinity is more enhanced in the bilayer-type packing motif than in the usual layer-by-layer packing of the symmetric molecules (Arai et al., 2018).

590

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 18.11 A molecular packing motif of Ph-BTBT-Cn (compound 3) single crystals (n ¼ 3, 4, 5, 6, 8, and 10), projected perpendicular to the orientation of molecular long axes. Reproduced from Fig. 3 in Inoue, S., Minemawari, H., Tsutsumi, J., Chikamatsu, M., Yamada, T., Horiuchi, S., Tanaka, M., Kumai, R., Yoneya, M., Hasegawa, T., 2015. Effects of substituted alkyl chain length on solution-processable layered organic semiconductor crystals. Chem. Mater. 27, 3809–3812.

18.3

Printing of metal wiring

18.3.1 Materials and techniques for printing electrodes and wiring Any electronic circuits and devices, including the AM backplanes and OTFTs, require fine metal electrodes and wirings as indispensable building blocks. Much effort has been dedicated, so far, to developing conductive inks and techniques for the print production of fine metal patterns (Kamyshny and Magdassi, 2014; Kamyshny et al., 2011). One of the most promising materials is concentrated metal nanocolloids, or nanometal inks, composed of highly concentrated metal nanoparticles dispersed in water or organic solvents. A great variety of metal nanoparticles and nanocolloids has been investigated and developed, so far, with various metal species, particle sizes and distributions, various encapsulating agents for colloidal dispersion, and dispersant. Two approaches have been adopted to prepare the metal nanoparticles. The topdown approach is to prepare nanoparticles by breaking bulk metal into small particles by using techniques such as mechanical grinding, laser ablation, rapid condensation of metal vapor, or plasma excitation of bulk. The bottom-up approach utilizes so-called synthetic chemistry, where the metal nanoparticles are obtained by reducing metal ions or decomposing precursor metal-complex molecules. The latter approach seems more advantageous, in terms of the cost and mass productions, than the former.

Advances in device fabrication scale-up methods

591

The metal nanoparticles are dispersed in appropriate dispersant to obtain stable metal nanocolloids. An example of the dense silver nanocolloids is presented in Fig. 18.12. The dispersed metal nanoparticles should be kept to exhibit rapid Brownian motion in the stable nanocolloids. Active or unstable bare metal surfaces should be encapsulated with insulating layers, such as polymeric material or various kinds of surfactants, to prevent aggregation or precipitation through the collision between metal nanoparticles. Currently, silver is the most reported material for conductive ink. Although inkjet printing is usually used to obtain metal wiring patterning, the obtained resolution is not enough for producing the AM backplane with resolution higher than 100 ppi or transparent conductive electrodes. The printed deposits also often suffer from nonuniform layer thickness due to the coffee-ring effect. It also should be mentioned that the encapsulating insulating layer, which is necessary for stable colloid dispersion, must be removed after the printing deposition to restore metal conductivity. So a postprinting process is required for the formation of continuous metallic contacts. This is usually conducted for the printed patterns by thermal annealing, light pulse irradiation, microwave (MW) irradiation, plasma sintering, or chemical sintering, while these processes have to be done under relatively mild conditions, in which the properties of the substrate material are not affected. Due to these difficulties, printed patterns (except for those described next) have not yet offered the material quality, pattern resolution, substrate compatibility, substrate adhesion, or throughput for mass production that are required for the industrial application of high-resolution AM backplanes, so long as conventional printing technologies are utilized.

Fig. 18.12 The appearance of dense silver nanoink and schematic of the dispersed silver nanoparticle. This example shows particular promising silver nanoparticles for the printed electronics; they are obtained by thermal decomposition of oxalate-bridging silver alkylamine complexes and are encapsulated by alklamines/alkylacids, and they can be dispersed in octane/ butanol at 40–60 wt%. Reproduced from Fig. 1 in Aoshima, K., Hirakawa, Y., Togashi, T., Kurihara, M., Arai, S., Hasegawa, T., 2018. Unique coexistence of dispersion stability and nanoparticle chemisorption in alkylamine/alkylacid encapsulated silver nanocolloids. Sci. Rep. 8, 6133.

592

Handbook of Organic Materials for Electronic and Photonic Devices

18.3.2 Novel nanoparticle chemisorption printing technique for ultrafine wiring Recently, a groundbreaking printing technique was reported by taking advantage of the unique nature of a specific silver nanocolloids that show unique self-sintering characteristics (Yamada et al., 2016). The technique, called surface photoreactive nanometal printing (SuPR-NaP), is based on unique chemisorption phenomena of silver nanoparticles on activated patterned polymer surfaces, as schematically presented in Fig. 18.13. The technique is composed only of a simple two-step process: the first is the fabrication of a patterned activated surface by masked vacuum ultraviolet (VUV) irradiation of the perfluoro-polymer layer surface, and the second is blade coating to expose the silver nanocolloids on the patterned activated surface for a short period of time (less than 1 s) at room temperature. It allows easy, high-speed, and large-area manufacturing of ultrafine metal wiring with a minimum line width of 0.8 μm, which is conductive without any postheating treatment, and strongly adheres on the substrate surface. These features are in striking contrast to the conventional printing technique which suffers from the limitations of the physisorption phenomena of the fluidic ink. The technique utilizes the specific characteristics of a class of silver nanocolloids, obtained by thermal decomposition of oxalate-bridging silver alkylamine complexes, as presented in Fig. 18.12 (Itoh et al., 2009). The peculiar nature of these silver nanocolloids is that the high dispersion stability is preserved for several months at room temperature, whereas the silver nanoparticles are readily self-fused with each other, eventually exhibiting metallic conductivity at room temperature if the metal nanocolloids are dried out (i.e., the dispersant evaporated). A recent study based on the confocal dynamic light scattering demonstrated that these silver nanocolloids are composed of a unique balance of ligand formulation (see Fig. 18.12) and dispersant composition, and that the unique balance enables the rapid silver nanoparticle chemisorption with maintaining the high dispersion stability of the silver nanocolloids, eventually leading to ultrahigh-resolution patterning (Aoshima et al., 2018). The technique is applicable to the production of flexible and transparent touch-screen sensor sheets composed of printed ultrafine metal mesh, as presented in Fig. 18.14.

Fig. 18.13 Schematic of the SuPR-NaP technique. Reproduced from Fig. 1 in Yamada, T., Fukuhara, K., Matsuoka, K., Minemawari, H., Tsutsumi, J., Fukuda, N., Aoshima, K., Arai, S., Makita, Y., Kubo, H., Togashi, T., Kurihara, M., Hasegawa, T., 2016. Nanoparticle chemisorption printing technique for conductive silver patterning with submicron resolution. Nat. Commun. 7, 11402.

Advances in device fabrication scale-up methods

593

Fig. 18.14 (A) Photograph of an 18-cm-wide, capacitive-type, transparent flexible touchscreen sensor sheet, fabricated by the SuPR-NaP technique with a poly(ethylene terephthalate) (PET) substrate. (B) Expanded microscope image of the sensor sheet, and the comparison with (C) an expanded microscope image of a 10,000-yen bill. The resolution of the SuPR-NaP technique is more than 10 times higher than the traditional high definition printing technique.

The mass production line for the product has just been established. The technique also should be quite useful in the production of OTFTs and highresolution, AM backplanes (Aoshima et al., 2017; Kitahara et al., 2017).

18.4

Outlook

Solution processability of organic materials is expected to provide essential advantages to the production of large-area and flexible electronics devices. To realize the scale-up production of these devices, the use and development of printing technologies are also quite promising. Owing to the intensive studies, some novel printing techniques for the formation of high-mobility OSC layers, as well as of ultrafine metal wiring, have been developed recently. The focus of these printing methods is to take full advantage of the unique characteristics of the electronic materials, such as the high-layered crystallinity of semiconducting organic molecules. Based on these findings and understanding, further design and development will proceed for the materials and processes. Additionally, the next important challenge is to unify the printing technologies for channel semiconductor layers and ultrafine metal wirings.

594

Handbook of Organic Materials for Electronic and Photonic Devices

References Anthony, J.E., 2006. Functionalized acenes and heteroacenes for organic electronics. Chem. Rev. 106, 5028. Anthony, J.E., Brooks, J.S., Eaton, D.L., Parkin, S.R., 2001. Functionalized pentacene: improved electronic properties from control of solid-state order. J. Am. Chem. Soc. 123, 9482. Aoshima, K., Arai, S., Fukuhara, K., Yamada, T., Hasegawa, T., 2017. Surface modification of printed silver electrodes for efficient carrier injection in organic thin-film transistors. Org. Electron. 41, 137. Aoshima, K., Hirakawa, Y., Togashi, T., Kurihara, M., Arai, S., Hasegawa, T., 2018. Unique coexistence of dispersion stability and nanoparticle chemisorption in alkylamine/alkylacid encapsulated silver nanocolloids. Sci. Rep. 8, 6133. Arai, S., Inoue, S., Hamai, T., Kumai, R., Hasegawa, T., 2018. Semiconductive single molecular bilayers realized using geometrical frustration. Adv. Mater. 30, 1707256. Bl€ ulle, B., H€ausermann, R., Batlogg, B., 2014. Approaching the trap-free limit in organic singlecrystal field-effect transistors. Phys. Rev. Appl. 1, 034006. Bucella, S.G., Luzio, A., Gann, E., Thomsen, L., McNeill, C.R., Pace, G., Perinot, A., Chen, Z., Facchetti, A., Caironi, M., 2015. Macroscopic and high-throughput printing of aligned nanostructured polymer semiconductors for MHz large-area electronics. Nat. Commun. 6, 8394. Campbell, R.B., Robertson, J.M., Trotter, J., 1961. The crystal and molecular structure of pentacene. Acta Crystallogr. 14, 705. de Gennes, P.-G., Brochard-Wyart, F., 2004. Capillarity and Wetting Phenomena: Drops, Bubbles, Pearls. Springer-Verlag, New York. Diao, Y., Tee, B.C.K., Giri, G., Xu, J., Kim, D.H., Becerril, H.A., Stoltenberg, R.M., Lee, T.H., Xue, G., Mannsfeld, S.C.B., Bao, Z., 2013. Solution coating of large-area organic semiconductor thin films with aligned single-crystalline domains. Nat. Mater. 12, 665. Diao, Y., Shaw, L., Bao, Z., Mannsfeld, S.C.B., 2014. Morphology control strategies for solution-processed organic semiconductor thin films. Energy Environ. Sci. 7, 2145. Dimitrakopoulos, C.D., Malenfant, P.R.L., 2002. Organic thin film transistors for large area electronics. Adv. Mater. 14, 99. Ebata, H., Izawa, T., Miyazaki, E., Takimiya, K., Ikeda, M., Kuwabara, H., Yui, T., 2007. Highly soluble [1]benzothieno[3,2-b]benzothiophene (BTBT) derivatives for highperformance, solution-processed organic field-effect transistors. J. Am. Chem. Soc. 129, 15732. Fukuda, K., Takeda, Y., Yoshimura, Y., Shiwaku, R., Tran, L.T., Sekine, T., Mizukami, M., Kumaki, D., Tokito, S., 2014. Fully-printed high-performance organic thin-film transistors and circuitry on one-micron-thick polymer films. Nat. Commun. 5, 4147. Gelinck, G., Heremans, P., Nomoto, K., Anthopoulos, T.D., 2010. Organic transistors in optical displays and microelectronic applications. Adv. Mater. 22, 3778. Giri, G., Verploegen, E., Mannsfeld, S.C.B., Atahan-Evrenk, S., Kim, D.H., Lee, S.Y., Becerril, H.A., Aspuru-Guzik, A., Toney, M.F., Bao, Z., 2011. Tuning charge transport in solution-sheared organic semiconductors using lattice strain. Nature 480, 504. Hamai, T., Arai, S., Minemawari, H., Inoue, S., Kumai, R., Hasegawa, T., 2017. Tunneling and origin of large access resistance in layered-crystal organic transistors. Phys. Rev. Appl. 8, 054011.

Advances in device fabrication scale-up methods

595

Hammock, M.L., Chortos, A., Tee, B.C.-K., Tok, J.B.-H., Bao, Z., 2013. The evolution of electronic skin (e-skin): a brief history, design considerations, and recent progress. Adv. Mater. 25, 5997. Iino, H., Usui, T., Hanna, J., 2015. Liquid crystals for organic thin-film transistors. Nat. Commun. 6, 6828. Ikawa, M., Yamada, T., Matsui, H., Minemawari, H., Tsutsumi, J., Horii, Y., Chikamatsu, M., Azumi, R., Kumai, R., Hasegawa, T., 2012. Simple push coating of polymer thin-film transistors. Nat. Commun. 3, 1176. Inoue, S., Minemawari, H., Tsutsumi, J., Chikamatsu, M., Yamada, T., Horiuchi, S., Tanaka, M., Kumai, R., Yoneya, M., Hasegawa, T., 2015. Effects of substituted alkyl chain length on solution-processable layered organic semiconductor crystals. Chem. Mater. 27, 3809. Itoh, M., Kakuta, T., Nagaoka, M., Koyama, Y., Sakamoto, M., Kawasaki, S., Umeda, N., Kurihara, M., 2009. Direct transformation into silver nanoparticles via thermal decomposition of oxalate-bridging silver oleylamine complexes. J. Nanosci. Nanotechnol. 9, 1. Kaltenbrunner, M., Sekitani, T., Reeder, J., Yokota, T., Kuribara, K., Tokuhara, T., Drack, M., Schwodiauer, R., Graz, I., Bauer-Gogonea, S., Bauer, S., Someya, T., 2013. An ultralightweight design for imperceptible plastic electronics. Nature 499, 458. Kamyshny, A., Magdassi, S., 2014. Conductive nanomaterials for printed electronics. Small 10, 3515. Kamyshny, A., Steinke, J., Magdassi, S., 2011. Metal-based inkjet inks for printed electronics. Open Appl. Phys. J. 4, 19. Kitahara, G., Aoshima, K., Tsutsumi, J., Minemawari, H., Arai, S., Hasegawa, T., 2017. Lowvoltage operation of organic thin-film transistors based on ultrafine printed silver electrodes. Org. Electron. 50, 426. Krebs, F.C., 2009. Fabrication and processing of polymer solar cells: a review of printing and coating techniques. Sol. Energy Mater. Sol. Cells 93, 394. Kronemeijer, A.J., Gili, E., Shahid, M., Rivnay, J., Salleo, A., Heeney, M., Sirringhaus, H., 2012. A selenophene-based low-bandgap donor–acceptor polymer leading to fast ambipolar logic. Adv. Mater. 24, 1558. Lim, J.A., Lee, W.H., Lee, H.S., Lee, J.H., Park, Y.D., Cho, K., 2008. Self-organization of inkjet-printed triisopropylsilylethynyl pentacene via evaporation-induced flows in a drying droplet. Adv. Funct. Mater. 18, 229. Matsui, H., Noda, Y., Hasegawa, T., 2012. Hybrid energy-minimization simulation of equilibrium droplet shapes on hydrophilic/hydrophobic patterned surfaces. Langmuir 28, 15450. McCulloch, I., Heeney, M., Bailey, C., Genevicius, K., Macdonald, I., Shkunov, M., Sparrowe, D., Tierney, S., Wagner, R., Zhang, W.M., Chabinyc, M.L., Kline, R.J., McGehee, M.D., Toney, M.F., 2006. Liquid-crystalline semiconducting polymers with high charge-carrier mobility. Nat. Mater. 5, 328. Minemawari, H., Yamada, T., Matsui, H., Tsutsumi, J., Haas, S., Chiba, R., Kumai, R., Hasegawa, T., 2011. Inkjet printing of single-crystal films. Nature 475, 364. Minemawari, H., Tanaka, M., Tsuzuki, S., Inoue, S., Yamada, T., Kumai, R., Shimoi, Y., Hasegawa, T., 2017. Enhanced layered-herringbone packing due to long alkyl chain substitution in solution-processable organic semiconductors. Chem. Mater. 29, 1245. Noda, M., Kobayashi, N., Katsuhara, M., Yumoto, A., Ushikura, S., Yasuda, R., Hirai, N., Yukawa, G., Yagi, I., Nomoto, K., Urabe, T., 2011. An OTFT-driven rollable OLED display. J. Soc. Inf. Disp. 19, 316.

596

Handbook of Organic Materials for Electronic and Photonic Devices

Noda, Y., Matsui, H., Minemawari, H., Yamada, T., Hasegawa, T., 2013. Observation and simulation of microdroplet shapes on surface-energy-patterned substrates: contact line engineering for printed electronics. J. Appl. Phys. 114, 044905. Noda, Y., Minemawari, H., Matsui, H., Yamada, T., Arai, S., Kajiya, T., Doi, M., Hasegawa, T., 2015. Underlying mechanism of inkjet printing of uniform organic semiconductor films through antisolvent crystallization. Adv. Funct. Mater. 25, 4022. Ogawa, S. (Ed.), 2015. Organic Electronics Materials and Devices. Springer, Japan. Rigas, G.-P., Payne, M.M., Anthony, J.E., Horton, P.N., Castro, F.A., Shkunov, M., 2016. Spray printing of organic semiconducting single crystals. Nat. Commun. 7, 13531. Sakai, S., Soeda, J., Haeusermann, R., Matsui, H., Mitsui, C., Okamoto, T., Ito, M., Hirose, K., Sekiguchi, T., Abe, T., Uno, M., Takeya, J., 2015. All solution-processed organic singlecrystal transistors with high mobility and low-voltage operation. Org. Electron. 22, 1. Singh, M., Haverinen, H.M., Dhagat, P., Jabbour, G.E., 2010. Inkjet printing – process and its applications. Adv. Mater. 22, 673. Sirringhaus, H., 2014. Organic field-effect transistors: the path beyond amorphous silicon. Adv. Mater. 26, 1319. Sirringhaus, H., Brown, P.J., Friend, R.H., Nielsen, M.M., Bechgaard, K., LangeveldVoss, B.M.W., Spiering, A.J.H., Janssen, R.A.J., Meijer, E.W., Herwig, P., de Leeuw, D.M., 1999. Two-dimensional charge transport in self-organized, high-mobility conjugated polymers. Nature 401, 685. Steudel, S., Myny, K., Schols, S., Vicca, P., Smout, S., Tripathi, A., van der Putten, B., van der Steen, J.-L., van Neer, M., Schuetze, F., Hild, O.R., van Veenendaal, E., van Lieshout, P., van Mil, M., Genoe, J., Gelinck, G., Heremans, P., 2012. Design and realization of a flexible QQVGA AMOLED display with organic TFTs. Org. Electron. 13, 1729. Sundar, V.C., Zaumseil, J., Podzorov, V., Menard, E., Willett, R.L., Someya, T., Gershenson, M.E., Rogers, J.A., 2004. Elastomeric transistor stamps: reversible probing of charge transport in organic crystals. Science 303, 1644. Takimiya, K., Shinamura, S., Osaka, I., Miyazaki, E., 2011. Thienoacene-based organic semiconductors. Adv. Mater. 23, 4347. Tsutsumi, J., Matsuoka, S., Yamada, T., Hasegawa, T., 2015. Gate-modulation imaging of organic thin-film transistor arrays: visualization of distributed mobility and dead pixels. Org. Electron. 25, 289. Tsutsumi, J., Matsuoka, S., Kamata, T., Hasegawa, T., 2018. Fast optical inspection of operations of large-area active-matrix backplane by gate modulation imaging. Org. Electron. 55, 187. Uemura, T., Hirose, Y., Uno, M., Takimiya, K., Takeya, J., 2009. Very high mobility in solution-processed organic thin-film transistors of highly ordered [1]benzothieno[3,2-b] benzothiophene derivatives. Appl. Phys. Express 2, 111501. Venkateshvaran, D., Nikolka, M., Sadhanala, A., Lemaur, V., Zelazny, M., Kepa, M., Hurhangee, M., Kronemeijer, A.J., Pecunia, V., Nasrallah, I., Romanov, I., Broch, K., McCulloch, I., Emin, D., Olivier, Y., Cornil, J., Beljonne, D., Sirringhaus, H., 2014. Approaching disorder-free transport in high-mobility conjugated polymers. Nature 515, 385. Yamada, T., Fukuhara, K., Matsuoka, K., Minemawari, H., Tsutsumi, J., Fukuda, N., Aoshima, K., Arai, S., Makita, Y., Kubo, H., Togashi, T., Kurihara, M., Hasegawa, T., 2016. Nanoparticle chemisorption printing technique for conductive silver patterning with submicron resolution. Nat. Commun. 7, 11402.

Advances in device fabrication scale-up methods

597

Yamamoto, T., Takimiya, K., 2007. Facile synthesis of highly π-extended heteroarenes, dinaphtho[2,3-b:20 ,30 -f]chalcogenopheno[3,2-b]chalcogenophenes, and their application to field-effect transistors. J. Am. Chem. Soc. 129, 2224. Yamamura, A., Watanabe, S., Uno, M., Mitani, M., Mitsui, C., Tsurumi, J., Isahaya, N., Kanaoka, Y., Okamoto, T., Takeya, J., 2018. Wafer-scale, layer-controlled organic single crystals for high-speed circuit operation. Sci. Adv. 4, 5758. Yoneya, M., Minemawari, H., Yamada, T., Hasegawa, T., 2017. Interface-mediated selfassembly in inkjet printing of single-crystal organic semiconductor films. J. Phys. Chem. C 121, 8796. Zhang, X., Bronstein, H., Kronemeijer, A.J., Smith, J., Kim, Y., Kline, R.J., Richter, L.J., Anthopoulos, T.D., Sirringhaus, H., Song, K., Heeney, M., Zhang, W., McCulloch, I., DeLongchamp, D.M., 2013. Molecular origin of high field-effect mobility in an indacenodithiophene–benzothiadiazole copolymer. Nat. Commun. 4, 2238.

This page intentionally left blank

Device stability in organic optoelectronics

19

Ayse Turak Department of Engineering Physics, McMaster University, Hamilton, ON, Canada

Chapter Points l

l

l

l

l

l

Interfaces play a major role in the stability of devices. After efficiency, lifetime is the second-most-important parameter for organic devices. Delamination, morphological instability, vulnerability to moisture and oxygen, and chemical attack at interfaces all play a major role in device stability. The additional complexity with new layers sometimes can lead to new degradation mechanisms that were not observed without it, pointing to the importance of stability testing as an intrinsic part of device design. There is a need to follow set guidelines (such as the International Summit on Organic Photovoltaic Stability (ISOS) testing protocols) during testing to ensure that rigorous comparisons are made and can be trusted to describe mechanisms. There is a need to adopt a statistical approach to prevent cherry picking of best results to determine the efficacy of changes to devices to improve stability.

19.1

Introduction

Organic optoelectronic devices based on π-conjugated semiconducting organic molecules are a rapidly growing focus of research. Organic light-emitting diodes (OLEDs) are already available in lighting and displays (Tremblay, 2016), and organic photovoltaics (OPVs) have begun to achieve efficiencies that can compete with amorphous silicon (Kang et al., 2016). However, these technologies suffer from sensitivity to atmospheric contaminants and internal instabilities that limit their effective lifetimes. After efficiency, lifetime is the second-most-important parameter for the commercialization of organic devices. Inorganic semiconductors are mostly chemically stable and insensitive to light and the ambient environment; for organic devices, polymer or small-molecule semiconducting layers, organic or inorganic interlayers, inorganic electrodes, and interfaces between them are all potential locations for degradation. Consequently, degradation for organic electronics is highly complex and cannot be described by a single mechanism. Degradation is a continuous process that occurs during both operation and storage. The three forms of degradation are loss of conjugation and irreversible deterioration of the active organic layers, degradation of the interface conductive properties, and mechanical disintegration of a device, which all change the measured electrical properties. Different applications of organic devices have different stability requirements. Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00019-X © 2019 Elsevier Ltd. All rights reserved.

600

Handbook of Organic Materials for Electronic and Photonic Devices

Stability for short-term displays, such as cell phone monitors, requires different criteria than long-term high-performance, solid-state lighting. Solar cells face exposure to environments that are not considered for OLEDs, aggravated by the degradation of many organic molecules upon exposure to light. The desire for flexible substrates for these technologies brings a different set of challenges. All these requirements have stimulated research on device stabilization in the 40 years since the first devices were produced. This chapter highlights the current state of research in device stability in OLEDs and OPVs, with discussion of the critical role that interfaces play in their performance. The chapter begins with a brief overview of degradation definitions and measurement approaches. It then proceeds to examine major mechanisms of device breakdown, including delamination, morphological changes, interfacial chemical reactions, and modification due to environmental and intrinsic factors. As many of the issues related to degradation, particularly at interfaces, are common for both OPV and OLEDS, and for polymer and small-molecule active layers, all types will be discussed within this chapter.

19.2

Degradation measurements and decay curves

Device lifetime is greatly affected by the driving voltage, number of duty cycles, length of rest cycles, initial luminance or power conversion efficiency (PCE), illumination conditions, deposition conditions, and exposed environment. This makes it difficult to compare reported lifetime values between different groups (Krebs et al., 2009; Gevorgyan et al., 2014). Efforts over the last decade have made considerable progress in establishing standardized protocols for OLED and OPV characterization (Castro, 2015; Blakesley et al., 2014; Rozanski et al., 2014; Shrotriya et al., 2006b; Corazza et al., 2014; Roesch et al., 2015). Stability testing protocols were proposed for OPVs by consensus among 21 international research groups in May 2011 to improve the reliability of reported values (Reese et al., 2011), and are in the process of being adopted by other researchers (Gevorgyan et al., 2011b, 2014, 2017; Tanenbaum et al., 2012; Kumar et al., 2016; Corazza et al., 2014; Tan et al., 2013). The ISOS guidelines offer instructions on how to perform a stability test of OPVs and similar technologies under different conditions, such as dark tests (shelf life), outdoor tests, indoor weathering, and thermal cycling tests, in an attempt to bring more consistency to the study of degradation. The use of standardized degradation testing setups, such as those proposed by Gevorgyan et al. (2011a), Reese et al. (2010), and Mogus et al. (2018) also can help reduce the variability of reported results. As degradation is quite complex in organic systems, it is critical to apply multiple protocols when benchmarking new approaches to stabilization, and to report the results of multiple tests accurately to achieve statistical significance. Recent studies have shown that it is necessary to adopt a statistical approach to correlate measured values meaningfully with any changes to the device (Luber and Buriak, 2013; Vasilak et al., 2017; Bovill et al., 2015). For example, though MoOx interlayers have been

Device stability in organic optoelectronics

601

shown to be successful in improving stability compared to PEDOT:PSS in certain devices and under certain conditions (see Section 19.3.3), Bovill et al. (2015) using a multiplexer to collect multiple samples at once and Kumar et al. (2016) under different ISOS testing protocols have shown that the opposite is true if the average device results are compared. Gevorgyan et al. (2016a,b) performed a comprehensive metaanalysis of OPV stability studies, including all reported lifetimes from 2001 to 2015, to establish baseline results for organic solar cell lifetimes, which can be used as a benchmark for new devices. The performance of organic devices can vary in a number of ways over time depending on its own properties and on external circumstances—there can be a smooth linear or exponential decrease from the beginning; there can be a sudden increase for a short period (the forming period) or a sudden decrease for short periods (the burn-in period), followed by a monotonic decay; or there can be some catastrophic loss at any point in time, as shown in Fig. 19.1. In any discussion of device degradation, it is important to clarify the terms of reference. The most common description of lifetime for OLEDs is the illumination halflife (t50): the time that it takes for the luminance to decrease to half of its initial value (Aziz and Popovic, 2004). In OPVs, a similar standard has been used in the past with t50 defined as the time for the PCE to decrease to half of its initial value ( Jørgensen et al., 2008). Currently, it is more common to report the t80, the time when the device has decayed to 80% of its initial performance (Reese et al., 2011). As decay curves can follow a variety of shapes, five time constants have been adopted to fully describe decay curves, as described in Table 19.1. In this chapter, t50 and t80 will be the lifetime values for OLEDs and OPVs, respectively, unless otherwise stated. In addition to times taken from the decay curve, the degradation can be described by rate constants, obtained by fitting a suitable function to the decay curve (for OPVs, see Conings et al., 2010; Schuller et al., 2004; for OLEDs, see F
ery et al., 2005). The most common shape of the decay curves for both OLED and OPVs is a rapid initial drop in performance followed by monotonic decay over longer time periods, as shown in Fig. 19.1. Such a curve is described by two exponentials.

Initial time Burn-in/irreversible damage

1.0

L Jsc or o L o Jsc

80% decay time (OPVs) Onset of second exponential (OPVs) 50% decay time (OLEDs)

0.5 Catastrophic failure

to

t 80

ts

t 50

Time

operation time or shelf time

Fig. 19.1 Representative decay curve, highlighting the various aspects of device degradation for both OLEDs and OPVs. The time constants necessary for reporting device stability are indicated with drop-down lines.

602

Handbook of Organic Materials for Electronic and Photonic Devices

Table 19.1 List of commonly used time constants to define device stability t0 t50 t80 ts

Initial measurement time immediately after device fabrication Time for performance to decay to 50% of initial measured value (half-life) Time for performance to decay to 80% of initial measured value Second testing measurement time defined arbitrarily by user: mainly for OPVs, onset of 2nd linear or exponential decay Time for performance to decay to 80% of measured value at t ¼ ts

ts, 80

L JSC or 0 ¼ aekint t + aekdeg t L0 JSC

(19.1)

where a and b are constants, and kint and kdeg are the degradation rate constants for the initial and monotonic decay terms, respectively. The two different stages of degradation observed in typical decay curves are important to consider independently, as they are believed to arise from different conditions: 1. The initial decay is attributed to interfacial degradation, often driven by extrinsic factors (such as moisture) (Turak, 2013b); and 2. The longer time-scale decay to intrinsic or oxidation-driven degradation of the bulk active layers (Turak, 2013b and references therein, as well as Scholz et al., 2015; So and Kondakov, 2010; Jørgensen et al., 2008).

Although two exponentials are widely used, it is also common for researchers to ignore the initial degradation and fit only the monotonic decay. They usually then relate the intrinsic losses to changes in bulk electrical properties of the active layers (e.g., quenching, trap formation, annihilation centers). F
ery et al. (2005) were able to use a stretched exponential function to describe the whole behavior, using the annihilation of emissive centers as the phenomenological basis. Coehoorn et al. (2015) performed a 3D kinetic Monte Carlo simulation using the same assumption, which reliably modeled the effect of changing the doping concentration in the active layer. Recently, Roesch et al. (2015), developed further by Gevorgyan et al. (2017), proposed the use of a new metric, the “lifetime energy yield” (LEY), as a measure of the reliability of a device over its entire lifetime. LEY0 is calculated as the integral over the fitted double exponential function over the operational lifetime: 0

LEY ¼

ð ts ,80

ηðtÞdt  1 kW=m2

(19.2)

ts

Using the ts as the initial condition that gives the largest LEY by maximizing the product of time and performance along the degradation curve allows a comparable ts that is independent of the actual decay curve shape. Decay curves, as in Fig. 19.1, which give a global overview of the loss of device properties, are not sufficient to fully describe the degradation processes occurring in organic devices. Visual inspection of the device with microscopy and photocurrent,

Device stability in organic optoelectronics

603

electroluminescence (EL), or photoluminescence (PL) mapping have shown that devices operate in nonhomogeneous ways, particularly for interfacial degradation.

19.2.1 Accelerated testing As seen in Eq. (19.1), for OLEDs, the critical parameter is the normalized luminance intensity at a fixed voltage over time. Increases in voltage at constant current are also sometimes tracked, and changes in emission color have been noted. The common benchmark value defined for OLEDs is the lifetime for an initial luminance of 100 Cd/m2. As t50 generally decreases as a function of applied voltage, accelerated lifetime testing can be performed for higher initial luminance, with an acceleration factor n, assuming Ln0 t50 ¼ const

(19.3)

For OPVs, the PCE is tracked, but it is the normalized short-circuit current (JSC) that is used to determine rate constants, as the open-circuit voltage (Voc) typically does not change over time, though sometimes degradation of the fill factor (FF) has been observed (Katz et al., 2006). It has been assumed that OPV degradation follows an Arrhenius process, so activation energies can be calculated (Conings et al., 2010; De Bettignies et al., 2006; Gevorgyan et al., 2008; Haillant, 2011; Schuller et al., 2004). Rather than used directly to describe degradation, these are most often applied to accelerated testing methods (De Bettignies et al., 2006; Schuller et al., 2004), where the acceleration factor K is the ratio between the constants at two temperatures, assuming a common activation energy Ea: K¼

   kdegT2 Ea 1 1 ¼ exp  kdegT1 kB T T1 T2

(19.4)

However, acceleration testing has not as been well studied by the OPV community, and it is controversial, as constant illumination at high temperatures does not adequately represent typical operational conditions (Krebs and Norrman, 2007). Hauch et al. (2008a), for example, observed very little photobleaching of polymer solar cells under outdoor conditions, whereas this was a major degradation feature during indoor accelerated tests (Visoly-Fisher et al., 2015). As solar cells also tend to recover their performance somewhat when kept in dark conditions (Katz et al., 2006; Krebs and Norrman, 2007; Reese et al., 2008), which occurs naturally during normal operation, a high degree of caution should be used for any accelerated results. The selection of aging conditions (e.g., light spectrum, intensity, temperature, atmospheric gases) can all have a major impact on the resulting degradation dynamics (Hermenau et al., 2010), making comparisons challenging even when following the ISOS protocols (Tanenbaum et al., 2012). The nonlinear dependence of acceleration factors on light intensity and time (Madsen et al., 2013) make it impossible to establish a single degradation rate for polymers degraded under concentrated light. Dosecorrected degradation rates have been proposed as a solution by Madsen et al.

604

Handbook of Organic Materials for Electronic and Photonic Devices

(2013), and yet the convolution of light intensity and temperature increases is a fundamental problem in accelerated testing, which complicates the interpretation of the results (Visoly-Fisher et al., 2015). Additionally, acceleration factors cannot be considered general, as they are really applicable only for a given device architecture: changing one constituent or layer can nullify calculated acceleration factors (Gevorgyan et al., 2008). Reese et al. (2011) state that accelerated testing cannot be considered as a standardized procedure for OPV stability characterization. Gevorgyan et al. (2008) cautions that exact experimental details need to be included when reporting acceleration factors for adequate reproducibility and comparison. Despite the limitations, the need to understand degradation processes within a reasonable experimental time frame has led to increased interest in applying accelerated testing (Visoly-Fisher et al., 2015; Hermenau et al., 2010; Madsen et al., 2013; Corazza et al., 2014; Gevorgyan et al., 2017).

19.3

Degradation mechanisms

Device performance can decrease in three ways: inhomogeneous loss of performance (“dark spots”), sudden catastrophic electrical shorting, or gradual loss without change in device appearance. Many mechanisms have been proposed to describe the device degradation. These mechanisms may act simultaneously or at different times in the device life cycle, they may enhance each other or act at cross purposes, and they may occur at a fixed time or apply continuously throughout the lifetime. Known degradation mechanisms include diffusion of oxygen and water, photobleaching, crystallization or oxidation of organic layers, interlayer and electrode diffusion, electrode reaction with the organic materials, electrode oxidation, phase segregation or intermixing, dewetting, delamination, and the formation of particles, bubbles, and cracks. These effects are shown schematically in Fig. 19.2. Interested readers are directed to in-depth topical reviews focused on interfacial degradation (Turak, 2013b), on Electrode

Active layers Hole transport layer (HTL) Electrode Substrate (transparent)

(A)

Electrode

Electrode

Active layers

Active layers

Hole transport layer (HTL) Electrode Substrate (transparent)

(B)

Electrode

Active layers

Hole transport layer (HTL) Electrode Substrate (transparent)

(C)

Hole transport layer (HTL) Electrode Substrate (transparent)

(D)

Fig. 19.2 Schematics of device structures undergoing various forms of degradation. (a) Photobleaching and charge-carrier degradation lead to a gradual loss of performance. (b) Chemical reactions at electrode interfaces can lead to inhomogenous loss of properties and catastrophic device destruction. (c) Morphological instabilities, including dewetting, interfacial intermixing, delamination, and donor-acceptor phase desegregation, can result in complete device breakdown. (d) A key approach in improving device stability lies in the introduction of interface-stabilizing buffer materials.

Device stability in organic optoelectronics

605

polymer photovoltaics (PVs) ( Jørgensen et al., 2008, 2012; Gevorgyan et al., 2016a,b, 2017), and on OLEDs (So and Kondakov, 2010; Scholz et al., 2015).

19.3.1 Operational and environmental delamination of layers One of the key mechanisms for device failure comes from the delamination of the various layers during storage or operation. The most vulnerable interface is typically the top metal contact, as degradation at the top contact proceeds more rapidly than any other failure process (Turak, 2013b). In OLEDs and OPVs, adhesion failure at the top contact is generally a cohesive failure, with the fracture plane roughly 5–10 ∘A within the organic layer (Turak et al., 2002; Vasilak et al., 2017), outside any interfacial interaction zone. With the rapid rise and development of stretchable OPVs (Kaltenbrunner et al., 2012) and OLEDs (White et al., 2013; Liang et al., 2013; Yin et al., 2016), delamination becomes even more significant. Electrode delamination in stretchable electronic devices is liable to occur when stress builds up in the device beyond the interfacial adhesion strength when stretched. This can limit the stretchability of certain device structures, which depends on the interface properties. It is crucial to have a method to assess the integrity and adhesion of this interface rapidly due to its importance to the performance and stability of organic devices. In recent years, several approaches have been used to examine adhesion and reliability in OPVs and OLEDs, including double-cantilever-beam (DCB) (Dupont et al., 2012, 2015; Greenbank et al., 2017), four-point bending (FPB) (Brand et al., 2012; Phatak et al., 2012), atomic force microscopy (AFM) (Tong et al., 2009; Momodu et al., 2014), normal force testing (Vasilak et al., 2017), and peel testing (Kim et al., 2018; Turak et al., 2002, 2009; Vasilak et al., 2017). Vasilak et al. (2017) recently proposed the use of normal force testing as a standard approach for adhesion testing, as a quick and easy technique that applies to a wide variety of samples without special sample preparation (i.e., can be performed in situ), and that uses a statistical approach to assess the integrity of the top contact. Using this approach, they were able to differentiate between different metal-organic interfaces and between different processing conditions. When combined with a Weibull analysis of the breaking strength and expected failure, the normal force method is well suited to examine the influence of changing processing parameters, exposure to electrical stress, or environmental conditions in a device directly rather than in idealized samples that approximate the conditions of the sample. Adopting a statistical treatment of a large volume of tests should be part of the standard protocol for investigating adhesion in order to accommodate the unavoidable variability in morphology and interfacial structure found in most organic devices (Vasilak et al., 2017). Adhesion can be expressed in two forms: fundamental adhesion and practical adhesion. Fundamental adhesion represents the summation of all interfacial intermolecular interactions that must be overcome to separate two materials; practical adhesion, on the other hand, is the value measured by an adhesion test. Practical adhesion, therefore, will depend on the fundamental adhesion, on the influencing effects of manufacturing, on the parameters of the measuring technique and on the

606

Handbook of Organic Materials for Electronic and Photonic Devices

environmental conditions during testing. The consequences of practical adhesion are exemplified in recent studies of adhesion of poly(3-hexylthiophene):(6,6)phenyl-C61-butyric acid methyl ester (P3HT:PCBM)/PEDOT:PSS interfaces in OPV devices. Using DCB, FPB, and AFM adhesion measurement approaches have resulted in adhesion strength values of 0.4 J/m2 for DCB (Dupont et al., 2012), 1.9 J/m2 for FPB (Brand et al., 2012), and 40 J/m2 for AFM (Tong et al., 2009), a variation of more than three orders of magnitude. Yet, the values from all three of these techniques have successfully correlated changes in measured adhesion to changes in processing or environmental conditions (Dupont et al., 2012, 2015; Brand et al., 2012; Phatak et al., 2012; Momodu et al., 2014). To monitor systematic development or intervention efforts, it is not critical that adhesion measurements yield values that agree with theoretical predictions. This is particularly true for organic devices, where theoretical predications are also complicated by the fact that there are significant and unavoidable variations in morphology and interfacial structure, despite best efforts to achieve identical processing conditions (Luber and Buriak, 2013). The spontaneous phase separation of the immiscible donor and acceptor molecules in bulk heterojunction (BHJ) OPVs leads to a random distribution of phase-separated sections of varying shapes and sizes, with a heterogeneous dispersion of components throughout the film (Watts et al., 2009; Ma et al., 2007). Additionally, organic devices always contain a distribution of flaws at the interface, due to the variability in organic film deposition and surface roughness. This results in a random distribution of dark spots (Turak, 2013b; Azrain et al., 2018) during device operation: the weakest spots delaminate first and form darkened regions that decrease the performance, and hence device failure. Removing and redepositing the electrode changes the location of these delaminated spots (Turak, 2013b), suggesting a large variation in surface structures, even with efforts to achieve identical processing conditions. Without a meaningfully reproducible theoretical framework, there is a need for a statistical approach to correlate measured values with changes to the device.

19.3.1.1 Operational delamination Generally, the operation of an organic device leads to significant heating at the electrode interface due to the high electric fields. The relatively low mobility of charge carriers through the active organic layers and poor heat dissipation leads to significant Joule heating of the organic layers as bias is applied (G€arditz et al., 2007; Preezant and Tessler, 2006; Tyagi et al., 2016). Under great electrical stress, dome like protrusions and dynamic delamination of the top contact have been observed by many groups (Turak, 2013b; Azrain et al., 2018), and some examples are highlighted in Fig. 19.3 (Momodu et al., 2014; Czerw et al., 2004; Akande et al., 2010). These so-called electrode bubbles have been observed with diameters up to 800 μm (Shin et al., 2006a), but they typically tend to be around 20 μm. The delamination of the top-side contact can be so severe that roughened buckling protrusions 10 the thickness of the electrode (up to 5 μm have been observed by Cumpston et al. (1997) for 50 nm electrode films), and can even lead to complete delamination of

Device stability in organic optoelectronics

607

Fig. 19.3 (A) Optical microscopy image showing the evolution of blisters/bubbles in a degraded OLED after day 4. (B) An edge-on view of the delaminated “ring” where the separation of the cathode from the emissive polymer is clearly visible. (C) Spiral blister formation in OLEDs. Scanning electron micrograph showing the distribution of blisters on the OLED surface at 1000. (Part A: Reprinted from Momodu, D.Y., Tong, T., Zebaze Kana, M.G., Chioh, A.V., Soboyejo, W.O., 2014. Adhesion and degradation of organic and hybrid organic-inorganic light-emitting devices. J. Appl. Phys. 115 (8), 084504, with the permission of AIP Publishing. Part B: Reprinted from Czerw, R., Carroll, D.L., Woo, H.S., Kim, Y.B., Park, J.W., 2004. Nanoscale observation of failures in organic light-emitting diodes. J. Appl. Phys. 96 (1), 641–644, with the permission of AIP Publishing. Part C: Reprinted from Akande, W.O., Akogwu, O., Tong, T., Soboyejo, W., 2010. Thermally induced surface instabilities in polymer light emitting diodes. J. Appl. Phys. 108 (2), 023510, with the permission of AIP Publishing.)

the entire device from the bottom indium tin oxide (ITO) surface (Turak, 2013b). Momodu et al. (2014) developed a finite element model to describe the blister formation using fracture mechanics. As melting and rupture occurs even under helium (He) and argon (Ar) atmospheres (Savvateev et al., 1997), bubble formation results mainly from Joule heating. Akande et al. (2010) calculated that temperatures as high as 160°C are possible during device operation. Liao et al. (2000), Savvateev et al. (1997), and Ahn et al. (2005) have seen similar bubble formation with thermal annealing around the glass transition temperatures of the active layers. For systems with poorly functioning electrodes, such as Mg/LiF for Alq3- and C60-based OLEDs (Turak, 2013b), significant bubbling and device failure under mild electrical stress (6 V) has been observed, as a result of high Joule heating from the high barrier to charge injection (Shin et al., 2006b; Turak, 2006). The bubbles typically contain O2 and H2 from absorbed moisture and decomposition products of the active layers or ITO (Turak, 2013b). The presence of gases inside the bubbles has led Aziz et al. (1998) and Do et al. (1994, 1996) to suggest a secondary galvanic corrosion mechanism. Due to the thermal mismatch between the organic layers and the top contact, swelling of the active layers with Joule heating also can lead to bubble formation through delamination. As shown schematically in Fig. 19.4, bubble formation and device degradation typically proceed as follows (Turak, 2013b): 1. Inhomogeneous electrical fields cause localized melting in the electrode. Radial temperature distributions from initiating defects result in thermal compressive stresses at the hot spots. 2. Gases are formed by Joule heating-related chemical reactions (e.g., ITO decomposition, active layer degradation, metal electrode oxidation).

608

Handbook of Organic Materials for Electronic and Photonic Devices

Increasing temperature

Loss of adhesion due to the presence of dust particles (exaggerated) or some other physical processes

a1, E1 a2, E2

Thermal stresses due to Joule Heating (compressive)

(A)

Periodic growth

(B) Fig. 19.4 Schematic drawing of the spiral blister growth model (a) radial temperature distribution develops at hot spots within the device due to the presence of chemical, structural, or interfacial defects. (b) Buckling occurs due to thermal compressive stresses at these hot spots. Adjacent and similar temperature distributions lead to the coalescence of the blisters. (Adapted from Akande, W.O., Akogwu, O., Tong, T., Soboyejo, W., 2010. Thermally induced surface instabilities in polymer light emitting diodes. J. Appl. Phys. 108 (2), 023510, with the permission of AIP Publishing.) 3. An overpressure develops at the weakest part of the device namely, the top-electrode/electron-transporting layer (ETL) interface causing the electrode to protrude. 4. Buckling occurs due to thermal compressive stresses at the hot spots. 5. Injected current is significantly higher at the perimeter of the bubble due to amplification of the electric field at the bubble edges, leading to more heating and growth of the bubble diameters. Adjacent and similar temperature distributions lead to the coalescence of the blisters. 6. Local current density becomes catastrophic, leading to the carbonization of active layers. Short-term healing of the current-voltage characteristics is sometimes observed, with the overall current increasing as charge carriers begin to flow through other parts of the device. 7. As more and more bubbles form and coalesce, the device fails catastrophically.

While many groups suggest that overpressure can result from decomposition of the active layers, infrared (IR) spectra after bubbling show no sign of organic breakdown, and removal of the cathode leaves no trace of perturbed areas on the organic surface

Device stability in organic optoelectronics

609

(Aziz et al., 1998; Savvateev et al., 1997). This suggests that the bubble formation is almost completely related to the top-side contact, where Gardonio et al. (2007) observed aluminum (Al) oxidation at the site of ruptured bubbles. Liao et al. (2000) observed that the thickness of electrode layer affects the appearance, density, and size of bubbles, where thinner layers typically formed smaller bubbles, as evolved gases were able to escape through the grain boundaries and pinholes in the electrode, rather than accumulating at the electrode interface.

19.3.1.2 Environmental delamination Phatak et al. (2012) observed a strong correlation between dark-spot formation and interfacial adhesion, as measured by an FPB test, suggesting that the primary mechanism is the delamination of the electrode at regions where adhesion is poor. Defect sites in the electrode where adhesion is the lowest are also often sites of delamination of the top electrode, and the evolution of dark spots due to storage time is more commonly associated only with cathode delamination. The electrode defect acts as a local avenue for ingress of the ambient air and moisture. Possible mechanisms for delamination include a change in volume associated with the formation of oxides at the interface (Schaer et al., 2001), swelling or dissolving of the hygroscopic interlayers at inorganic electrodes, such as poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) or LiF (Turak et al., 2009; Phatak et al., 2012); or swelling of the polymer layers with water or solvent vapors (Emerson et al., 2013). The critical role of moisture in adhesion was confirmed both by Phatak et al. (2012), by FPB, and by Vasilak et al. (2017), using the normal force test for adhesion. Phatak et al. (2012), by exposing OLEDs to 100% relative humidity (RH), observed rapid formation of dark spots; Vasilak et al. (2017), by placing Al/P3HT:PCBM/ITO OPV samples in a vacuum desiccator to remove all moisture before testing, observed no delamination of the top electrode, with sample failure at the tape-glass interface, suggesting that the removal of moisture increased the strength of the Al/P3HT:PCBM interface beyond the strength of the tape used to mount the samples. Hygroscopic films such as PEDOT:PSS and LiF are particularly vulnerable to moisture attack. When humidity is increased to 55% RH, Lang et al. (2009) saw a reduction in the mechanical strength of PEDOT:PSS films by a factor of 3. Higher RH values lead to water uptake by the hydrophilic and hygroscopic PSS rich shell, and thus to a swelling of the material and larger distances between grains. Larger distances lead to reduced cohesion due to weakened hydrogen bonds, and thus reduced mechanical strength. This often leads to sample failure at the PEDOT:PSS interfaces (Tong et al., 2009; Dupont et al., 2015). Interfacial adhesion at the top contact is similarly affected by the presence of hygroscopic interlayers. Turak et al. (2009) observed that removal of the cathode by an adhesive-tape-peel test was significantly harder with thin interlayers, requiring immersion in a water bath for at least 10 min for complete removal. Without an interlayer, the adhesion between Al and P3HT:PCBM was so strong that even immersion in a water bath did not result in significant cathode removal during the peel test due to metal diffusion into the active layer, as shown in Fig. 19.5. Phatak et al. (2012) also

610

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 19.5 Effect of interlayer thickness on interfacial interaction zone and peeling. (Left) Optical micrographs of complete devices before peeling for various interlayers. (Right) Optical micrographs of complete devices after peeling for various interlayers. (Bottom) Schematic models of the interfacial interaction zone and nature of the electrode interface. (Modified with permission from Turak, A., Hanisch, J., Ahlswede, E., Barrena, E., Dosch, H., 2009. Interfacial adhesion in polymer blend P3HT:PCBM solar cells. In: Fruhjahrstagung DPG, Dresden. Copyright 2009, DPG.)

saw that LiF resulted in faster growth of dark spots in high-humidity environments attributed to interfacial delamination. Inclusions in the active layer, such as nanoparticles for hybrid devices, also greatly increase the rate of electrode delamination. Using fracture mechanics, Momodu et al. (2014) showed that hybrid OLEDs, with titanium dioxide (TiO2) nanoparticles embedded in a 2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylenevinylene (MEH-PPV) active layer, has a higher degradation rate, due primarily to increased principal stresses and crack-driving forces, proportional to the nanoparticle size. The inclusions promote the formation of cracks and defective sites within the active layers, due to the high stress distributions around nanoparticle-induced voids. These make hybrid devices more prone to moisture-induced delamination and faster degradation than their neat counterparts.

19.3.1.3 Preventing delamination There are two approaches to preventing delamination: increase the adhesion strength by modification of the processing conditions or introduce nonsoluble interlayers. Preannealing has been shown to have a significant effect on OPV device performance

Device stability in organic optoelectronics

611

(Turak et al., 2010; Watts et al., 2009; Motaung et al., 2013; Reyes-Reyes et al., 2005), and on the resilience of the metal layer to an adhesive-tape-peel test (Ahn et al., 2005). Motaung et al. (2013) have established that 15 min of annealing at 140°C leads to an optimized P3HT:PCBM morphology with lower surface route mean square (RMS) roughness than 5 min at the same temperature, contributing to higher PCEs. Dupont et al. (2012, 2015) have also seen that longer annealing can improve the interfacial fracture energy and cohesive energy at the PEDOT:PSS surface. Typically a change from 5 to 15 min annealing results in a 14% improvement in PCE (Vasilak et al., 2017; Turak et al., 2010; Watts et al., 2009; Motaung et al., 2013; Reyes-Reyes et al., 2005), as shown in Fig. 19.6. Vasilak et al. (2017) showed that longer annealing time reduces the delamination percentage from the adhesive-tapepeel test by almost 60%. This suggests that the interfacial adhesion was stronger with greater annealing time. Ahn et al. (2005) reported similar behavior for polymer-based light-emitting diodes, though in that case, annealing for extended times completely suppressed delamination. Using a Weibull analysis, Vasilak et al. (2017) established a slight difference for the two annealing times, with m15min ¼ 1:1m5min , which is roughly equal to the ratio in the RMS roughness of the polymer surfaces after peeling (15.3 nm vs. 16.6 nm, respectively). A higher Weibull parameter suggests a lower distribution of flaws at the interface that can act as defects; therefore, a reduction in roughness is correlated with greater adhesion and a larger modulus (Vasilak et al., 2017).

Fig. 19.6 (A) Result of an interfacial adhesion adhesive-tape-peel test using an adhesive tape (3M Scotch). (B) Current-voltage characteristics for P3HT:PCBM devices subjected to various annealing times (best-performing cells). (Reprinted with permission from Vasilak, L., Tanu Halim, S.M., Das Gupta, H., Yang, J., Kamperman, M., Turak, A., 2017. Statistical paradigm for organic optoelectronic devices: normal force testing for adhesion of organic photovoltaics and organic light-emitting diodes. ACS Appl. Mater. Interfaces 9 (15), 13347–13356. Copyright 2017. American Chemical Society.)

612

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 19.7 Optical and representative ATMs of various interlayers after an interfacial adhesion adhesive-tape-peel test. Samples were immersed in water for at least 1 h prior to testing. (Modified with permission from Turak, A., Hanisch, J., Ahlswede, E., Barrena, E., Dosch, H., 2009. Interfacial adhesion in polymer blend P3HT:PCBM solar cells. In: Fruhjahrstagung DPG, Dresden. Copyright 2009, DPG.)

However, as annealing also results in compressive stresses in the device, prolonged annealing actually can lead to more rapid blistering and device failure. Akande et al. (2010) showed with 2 h annealing, the blisters took less time to initiate growth and achieve their final configuration than their nonannealed counterparts. Consequently, in line with their model, overannealing reduces the stress barrier required to cause blisters during the operation of the device (Akande et al., 2010). As moisture greatly diminishes the adhesion of the electrode/polymer interface, the introduction of insoluble interlayers also can help to improve the adhesion. Phatak et al. (2012) improved the adhesion and performance by using a metal-organicmixed-layer (MOML) in place of the more vulnerable LiF, which provides a continuous graded transition in composition between the active layer and the electrode (Aziz et al., 2004; Phatak et al., 2012). Yoon et al. (2008) observed that the reaction between Al and TiOx leads to better adhesion of the electrode on the active organic surface at lower pressures, improving performance. Turak et al. (2009) explored the resistance to peeling of a number of interlayers with different water solubilities, showing a correlation between water solubility and percent delamination (see Fig. 19.7 and Table 19.2).

19.3.2 Morphological degradation It is the interplay between molecule-molecule self-interaction and substrate-molecule interactions that determines the morphological stability of films on a given surface (Israelachvili, 2011). Heating above the glass transition temperature (Tg) of one of the components has been linked to irreversible device failure in small-molecule OLEDs (Fenter et al., 1997) and OPVs (Franke et al., 2008). However, even relatively mild elevated temperatures or even room temperature is sufficient to encourage metastable organic films to evolve into their equilibrium forms (shown schematically in Fig. 19.8 for a single-phase film). Many organic molecular crystals exhibit several distinct crystal structures, which are energetically very similar and may coexist (Burke et al., 2009; Krause et al., 2003). The amorphous films often used in devices

Interlayer LiCoO2 LiF NaF KF a

Island height of peeled regiona (nm) N/A 55  5 57  3 52  6

RMS roughness (nm) N/A 17.3  3 18.5  4 18.5  2

Correlation length (mm) N/A 3.1  1.3 3.4  1.1 2.5  1.5

PCE (%)

Solubilityb (g/100 g H2O)

Coverage (%)

3.5  3.2  3.3  2.7 

Insoluble 0.134 4.13 102

100 55 25 10

0.2 0.6 0.1 0.1

Device stability in organic optoelectronics

Table 19.2 Characteristics of devices with various interlayers after an interfacial adhesion adhesive-tape-peel test with water immersion

Determined by taking a line profile from the Al electrode across the peeled interface, as described in Vasilak et al. (2017). Lide (2010) and Barton (1991).

b

613

614

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 19.8 (A) Schematic of dewetting from metastable as-deposited layers to equilibrium island structures. (B) (Left) 3.9 mL of diindeoperylene deposited on superflat ITO at room temperature. (Right) The equilibrium morphology after 1 month of storage under low humidity, low oxygen conditions. (Reprinted with permission from Heidkamp, J., Maye, F., Turak, A.Z., 2013. Stabilization methods for small molecule dewetting on indium tin oxide substrates for organic photovoltaics. In: Cheben, P., Schmid, J., Boudoux, C., Chen, L.R., Del^age, A., Janz, S., Kashyap, R., Lockwood, D.J., Loock, H.-P., Mi, Z. (Eds.), Proceedings of SPIE, vol. 8915. SPIE, Ottawa, ON, p. 891508.)

do not have the benefit of lateral crystalline organization to stabilize against dewetting (Vix et al., 2000), and they also are very susceptible to extreme morphological instabilities. As BHJ devices require the spontaneous phase segregation of the donor and acceptor polymers, the interpenetrating morphology is highly metastable, with the morphology continuing to evolve throughout the device’s lifetime. There also often exists a large surface-energy mismatch at interfaces between hydrophilic transparent conducting metal oxides (contact angles 0-30 degrees) (Liu et al., 2010b; Granqvist, 2007) and hydrophobic active materials; between immiscible donor-acceptor or guesthost pairs (Watts et al., 2009; Ma et al., 2007; Peet et al., 2009); and between overlayers on organic surfaces (Abeysinghe et al., 2009). These interfaces are also very susceptible to moisture and light irradiation. As such, the molecule-molecule and substrate-molecule interactions can change over time, leading to a highly metastable environment within organic devices. Changes to the device morphology over time can throw off the delicate balance between charge percolation and absorption and lead to diminished device properties. Metastability of the active layers, therefore, can be a significant driving force for degradation, as the organic layers are subjected to thermal stress at a number of points during the device’s lifetime. During normal device operation, low mobility in the organic films can lead to high electric fields and local Joule heating (G€arditz et al.,

Device stability in organic optoelectronics

615

2007; Preezant and Tessler, 2006; Tyagi et al., 2016). Any local failures due to morphological inhomogeneities on the ITO surface (in spikes, thinner organic regions, and other elements), where the high local electric field would already encourage Joule heating, also would tend to both accelerate and be accelerated by morphological instabilities. Zhou et al. (2000) observed that the surface temperature for thieno[3,4-c]pyrrole4,6-dione (TPD)-based OLEDs can reach as high as 86°C, suggesting that the temperature inside the actual devices could be higher than 200°C. Akande et al. (2010) calculated that temperatures as high as 160°C are possible during device operation. Tessler et al. (1998) saw temperature variation during operation as high as 60°C in the recombination zone. As semiconducting organic molecules tend to show poor natural heat dissipation (Nakanotani et al., 2005; Zhou et al., 2000), such large temperature variations cannot be handled by the poor heat sink at the glass surface. Choi et al. (2008) were able to find an inverse correlation between the OLED device’s lifetime and its internal temperature, as measured by a scanning thermal microscope. OPVs have the additional burden of illumination-induced heating and cooling cycles, which can cause the internal temperatures to reach well beyond the Tg when coupled with the already high burden of Joule heating. Sullivan and Jones (2008) observed that ultraviolet (UV)-induced heating (through UV absorption by the glass substrate) lead to a structural reorganization of pentacene, decoupling it from the ITO surface, causing kinks to form in the J-V curves for PEN:C60 solar cells. Paci et al. (2008) observed that the metastable morphology of P3HT:PCBM solar cells was modified in a similar way under illumination as by deliberate thermal annealing. For many devices, the weakest link in the device lifetime is the intrinsic low Tg of layers immediately adjacent to the ITO (see Turak, 2013b and references therein). On the contrary, many electron-accepting and -transporting materials have relatively high Tg. Do et al. (1997) for MEH-PPV, Fenter et al. (1997) and Orita et al. (1997) for TPD, and Choi et al. (2008) for PFO observed film buckling where the organic layers completely detached from the ITO due to significant expansion of the least thermally stable material; the result was dead areas that eventually covered the whole surface. Such instability in the hole-transport layer (HTL) leads to inefficient hole injection and modification of the other layers deposited atop the dewetted surface (Xu et al., 2004). Even below Tg, significant structural changes have been observed in systems such as TPD, where at 60°C, thickness and density changes measured by XRR (Fenter et al., 1997) and even serious dewetting shown by AFM (Orita et al., 1997) were seen. Morphological instability that occurs within organic devices takes three forms: molecular dewetting, crystallization, and phase segregation.

19.3.2.1 Dewetting Resulting from the many dangling O bonds on the surface (Paramonov et al., 2008), many transparent metal oxides present a hydrophilic surface. However, many molecules deposited on electrode surfaces are hydrophobic. This surface-energy mismatch leads to inhomogeneous deposition, both during vacuum evaporation and deposition from solution, as seen in Fig. 19.9.

616

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 19.9 (A) AFM micrographs of 50 nm of DIP deposited on a bare-ITO substrate, showing strong island growth and incomplete coverage. (B) Optical micrograph of inhomogeneous and incomplete coverage of ITO surface with PEDOT:PSS deposited from solution.

Even when continuous films appear to be formed, they are highly metastable. The relatively low Tg and weak van der Waals interactions for many small-molecule holetransporting materials can lead to dewetting to their equilibrium structures under mild thermal treatments, or even with to storage over time at ambient temperatures (Turak, 2013a; Burke et al., 2009), as shown in Fig. 19.8. DIP, a novel material of interest due to its well-defined ordering, interesting growth behavior, promising electron transport properties, favorable electronic structure, and long exciton diffusion lengths has been shown to have tunable behavior in solar cells based on its morphology (Turak et al., 2011; Gruber et al., 2013). As seen in Fig. 19.8, films of DIP form large, flat islands on ITO, with a high degree of order upon initial deposition (Turak et al., 2011). After storage under vacuum at room temperature for 1 month, very strong dewetting into columnar structures was observed by Heidkamp et al. (2013, 2016). Dewetting is not as significant a problem for polymer system, due to the generally higher Tg of polymer materials, although sometimes it has been observed under high-temperature treatments. There has been much research into methods of counteracting dewetting, particularly from the ITO surface. Incorporation of the unstable film into a device, with a mutlilayer film stack, already significantly suppresses the dewetting of single films (Turak, 2013a). Deliberate use of a stable capping layers within the device also can greatly improve the stability of the underlying organic phases, observed for interlayer materials such as Al2O3 by Sellner et al. (2004), LiF nanoparticles by Heidkamp et al. (2013), or electrode metals such as silver (Ag) by Peumans et al. (2003) or gold (Au) by D€ urr et al. (2003a). Utilizing a rough substrate also can encourage wetting and stability (Barnes et al., 2000; Karapanagiotis et al., 2002; Netz and Andelman, 1997; Shen et al., 2001; Zhang et al., 2005; Heidkamp et al., 2013), but that can lead to undesirable morphologies in the original deposition (Heidkamp et al., 2013), as shown schematically in Fig. 19.10. Increasing the glass transition temperature of the HTL layer was seen as a promising approach (see the discussion in Turak, 2013a,b), but although new materials are synthesized regularly, with significantly higher glass transition temperatures than the

Device stability in organic optoelectronics

617

Fig. 19.10 Morphology of DIP for (A) superflat (sflat), ITO (B), rough ITO (C) with a PEDOT: PSS interlayer on ITOsflat (D) with thermally evaporated LiF nanoparticle decoration on DIP on ITOsflat. Top panels are schematics of the as-deposited DIP morphology, while the bottom panels show the schematic of DIP structure on ITOsflat and with various stabilization approaches. The top middle panels are AFM micrographs showing the as-deposited morphology, after at least 1 month of storage in an evacuated desiccator, and the bottom middle panels are AFM micrographs of the same samples after at least 1 month of storage in an evacuated desiccator. (Modified with permission from Heidkamp, J., Maye, F., Turak, A.Z., 2013. Stabilization methods for small molecule dewetting on indium tin oxide substrates for organic photovoltaics. In: Cheben, P., Schmid, J., Boudoux, C., Chen, L.R., Del^age, A., Janz, S., Kashyap, R., Lockwood, D.J., Loock, H.-P., Mi, Z. (Eds.), Proceedings of SPIE, vol. 8915. SPIE, Ottawa, ON, p. 891508.)

classically utilized molecules, the correct combination of high hole mobility, good energetic compatibility with electron-accepting materials, and good optical absorbance has proved elusive, and so many of these materials currently do not see widespread use in devices. The generally employed strategy that seems most compatible with device manufacturing is the introduction of stabilizing interlayers under the HTL in order to create a graded interface. A wide variety of interlayer materials, ITO treatments,

618

Handbook of Organic Materials for Electronic and Photonic Devices

self-assembled monolayers, and doped organic layers have been tried for the purpose of preventing morphological instabilities, as described in greater detail in Turak (2013a,b). The relative success of various approaches is summarized in Fig. 19.10, where Heidkamp et al. (2013, 2016) show that a PEDOT:PSS interlayer was the most effective in maintaining the desired film morphology after storage in vacuum for 1 month, with X-ray diffraction (XRD) by Maye (2011) and Turak (2013b) confirming that the crystal structure also is preserved during storage.

19.3.2.2 Crystallization One of the major mechanisms for degradation during dark storage, crystallization of the active layers is considered to be problematic because it lowers the ionization potential, fundamentally changing the electrical characteristics and exciton formation/dissociation ability. Though of greater concern for small-molecule-based devices, crystallization of both acceptors and donors have been linked to device failures in small-molecule- and polymer-based devices. Generally, it has been found that the higher the glass transition temperature (Tg), the lower the likelihood of crystallization (Adachi et al., 1995), and the higher the lifetime. In most devices, the material with the lowest Tg in the device is the most likely candidate to initiate device failure. Especially with the layers adjacent to the electrodes, the interface can have a major impact on the crystallization behavior of the organic layers, and as such, some crystallization also can be related to interfacial degradation (Turak, 2013b). One particularly problematic material is also one of the most widely used holeinjection and electron-donating materials for small-molecule OLEDs and OPVs, copper phthalocyanine (CuPc)(Claessens et al., 2008; De Oteyza et al., 2010; Walter et al., 2010). First introduced as a stabilizing buffer layer for N,N0 -bis-(1naphthyl)-N,N0 -diphenyl-1,10 -biphenyl-4,40 -diamine (NPB), the use of CuPc prevents the ambient dewetting observed for NPB directly deposited on ITO (Grozea et al., 2007), providing a metastable equilibrium structure for devices at room temperature. CuPc also was seen to increase the crystallization temperature of NPB to above 160°C (Lee et al., 2009; Xu et al., 2004, 2005), leading to more than a fivefold increase in the luminance t50 during room-temperature operation (Aziz et al., 2006). However, at even mildly elevated temperatures, the devices perform poorly. Moderate heating well below Tg (as low as 60°C) leads to CuPc crystallization (Cui et al., 2002b; Grozea et al., 2007), intermixing with other HTLs (Xu et al., 2005; Cui et al., 2002a), and dewetting. Ultimately, CuPc-buffered ITO does not prevent HTL crystallization and decohesion upon heating. It is highly susceptible to chemical reaction with the ITO surface itself (Peisert et al., 2002; Xu et al., 2005), introducing another unstable interface. Additionally, the use of CuPc leads to significant increase in the driving voltage for OLEDs (Kim et al., 2010). To overcome some of these difficulties, a separate buffer layer can be used under the CuPc layer, such as Pr2O3 (Qiu et al., 2002) or LiF (Lee et al., 2004). The most successful approach to suppressing crystallization has been doping of the HTL, particularly with other molecules (Turak, 2013b). Often referred to as alloying,

Device stability in organic optoelectronics

619

or more recently as a BHJ, the use of a composite system disrupts the periodic ordering needed for crystallization, leading in extreme cases to phase-segregated systems. Mori et al. (2002) used metal-free Pc to disrupt the crystallization of CuPc, showing about a twofold improvement in lifetimes for operation above 85°C. Chu et al. (2008) and Lee et al. (2005b) used higher Tg molecules N,N-bis(4-trifluoromethoxybenzyl)-1,4,5,8naphthalene-tetracarboxylic diimide (NTCDI) and perfluorinated ionomer (PFI) to stabilize NPB and PEDOT:PSS, respectively. Doping of the electron-transporting layer, Alq3, with other molecules also has been very successful in improving device stability (Turak, 2013b). The incorporation of PCBM into various polymers, including poly(3-hexylthiophene) (P3HT), PPV, and MDMO:PPV, as BHJs also stabilizes the morphology against heat treatment and improves long-term stability (Turak, 2013a,b). The addition of a diblock copolymer of P3HT-C60 to a P3HT:PCBM composite has been seen by Lee et al. (2010a,b) to promote even-greater stabilization against phase desegregation. In some cases, the hole-transport molecule was doped into a more stable matrix, such as TPD into high-Tg polymers (Santerre et al., 2001), which significantly suppressed the crystallization. Ruberene:TPD (Vestweber and Riels, 1997), MADN:NPB (Tsai et al., 2005), F4TCNQ:NPB (Aziz et al., 2006; Luo et al., 2007b), and DSA-Ph:NPB(Lee et al., 2005a) combinations have all been employed with various levels of doping, leading in most cases to about a twofold improvement in luminance t80 in small-molecule OLEDs. Doping with nanoparticles such as l

l

l

l

l

LiF (Grozea et al., 2007; Heidkamp et al., 2013; Gao et al., 2010) C60 (Barnes et al., 2000; Grozea et al., 2007; Holmes et al., 2007; Yuan et al., 2005, 2006) Sodium chloride (NaCl) (Kim et al., 2010) Au (Mukherjee et al., 2010 V2O5 (Rafique et al., 2016)

also has been very successful in stabilizing layers, although there is sometimes a tradeoff between stability and performance for such systems (Grozea et al., 2007). The concentration and layer thickness must be chosen such that the electrical performance is not adversely affected by the presence of the doped layer. There is a long history of nanoparticle inclusions for stabilization in nonconducting polymers (Barnes et al., 2000; Luo and Gersappe, 2004; Sharma et al., 2001). Interparticle or surface forces strongly influence the suspension behavior of nanoparticles; therefore, not every nanofiller works with every organic. Grozea et al. (2007) observed that LiF greatly enhanced the stability of NPB at 120°C, while having no impact on the crystallization of CuPc. Luo and Gersappe (2004) suggest that a combination of factors are responsible for the stabilization effects, such as the mobility of nanofillers, their size, interaction with the organic, and additional pinning effects at contact lines. Fillers work best if they are immobile; therefore, diffusion to and pinning at the substrate interface comprise one suggested stabilization mechanism (Barnes et al., 2000; Sharma et al., 2001). Chu et al. (2007) did see similar stabilization with a C60 layer deposited at the interface below NPB as Yuan et al. (2006) saw with C60 doping into NPB, supporting such a mechanism; Barnes et al. (2000), however, saw greatly enhanced dewetting with C60 at the Si surface for polystyrene thin films,

620

Handbook of Organic Materials for Electronic and Photonic Devices

refuting it. Additionally, there are a number of cases where diffusion to the substrate is unlikely, as no phase separation was observed (Grozea et al., 2007; Kim et al., 2010; Tokito and Taga, 1995). Mukherjee et al. (2010) observed a concentration dependence on the stabilization, where dewetting droplets form a core-shell structure rather than leaving behind nanoparticles, while the polymer layer retreats as expected for substrate segregation. In such cases, strong electrostatic or charge-transfer (CT) interactions between the particle and the organic layer leading to a cross-linked network are the most likely route to creating highly stable films (Grozea et al., 2007; Sharma et al., 2001). Pinning effects at grain boundaries in the organic layer also are observed when LiF is deposited on top of the HTL layer, as shown in Fig. 19.10. Other possible mechanisms for stabilization include changing the Tg with a high volume-to-surface area ratio (effectively modifying the film rheology), preventing heterogeneous nucleation, and relieving of residual stress in the film through desegregation (Mukherjee et al., 2010). Again, due to the complicated nature of degradation, approaches that improve morphological stability, such as nanoparticle doping, also were seen to increase the rate of electrode delamination greatly (see Section 19.3.1), resulting in reduced lifetimes. As with all approaches, it is important to check multiple degradation pathways to determine if a strategy will be successful.

19.3.2.3 Phase segregation/demixing For polymer-based OLEDs and OPVs, morphological instability of the active layers also can lead to undesirable intermixing (Fujihira et al., 1996) or phase segregation (Bertho et al., 2007), with significant consequences for the devices in operation. Optimal performance in organic devices is often the result of a complex balance of chargecarrier flow and emission/absorption locations. Changes to the device over time can throw off this balance and lead to diminished device properties. Although the development of BHJs was responsible for the rapid improvement of OPV-device performance in the mid-1990s (Forrest, 2005), and were even successfully applied to OLEDs (Aziz et al., 2002; Chen et al., 2005b; Choong et al., 2000; Chwang et al., 2002; Matsushima and Murata, 2008; Naka et al., 1994; Shi and Tang, 1997; Tang et al., 1989), it results in an inherently metastable film. The process involves the codeposition of donor and acceptor phases that spontaneously selfassemble into an optimally separated, phase-segregated film (Halls et al., 1995; Yu et al., 1995). The spontaneous phase separation of the immiscible donor and acceptor molecules leads to a random distribution of phase-separated sections of varying shapes and sizes, with a heterogeneous dispersion of components throughout the film (Watts et al., 2009; Ma et al., 2007; Peet et al., 2009). Overannealing, elevated temperature storage over long time scales or overexposure to solvent vapors can lead to significant phase segregation (Hoppe et al., 2004; Motaung et al., 2011; Yang et al., 2004b; K€astner et al., 2016; Wang et al., 2016; Chang et al., 2011). Fig. 19.11 shows micron-sized PBCM aggregates segregating from a P3HT:PCBM BHJ film. Separated domains beyond the exciton diffusion length remove the benefits of a BHJ and ultimately decrease performance over time

Device stability in organic optoelectronics

621

(A)

(B)

Fig. 19.11 Optical microscopy images of P3HT:PCBM films posttreated with xylene (A) and dichlorobenzene (B) vapors. All images have an identical size. (Adapted from Wang, W., Guo, S., Herzig, E.M., Sarkar, K., Schindler, M., Magerl, D., Philipp, M., Perlich, J., M€uller-Buschbaum, P., 2016. Investigation of morphological degradation of P3HT:PCBM bulk heterojunction films exposed to long-term host solvent vapor. J. Mater. Chem. A 4 (10), 3743–3753, published by the Royal Society of Chemistry.)

(Bertho et al., 2007, 2008; Conings et al., 2010; Yang et al., 2004b). Wang et al. (2016) observed an almost 100% decrease in PCE with the appearance of large aggregates, well beyond the typical exciton diffusion lengths. Even normal device operation can disrupt the optimal conditions for charge flow: Schaffer et al. (2013) showed direct evidence using micro-small-angle X-ray scattering (μSAXS) that the growth of intermediate-sized domains of PCBM, not visible by cross-sectional SEM or AFM, resulted in decay in the short-circuit current. Ray and Alam (2011), using FloryHuggins’s mean field theory coupled with Cahn-Hilliard (C-H) equation, described the expected effect of phase segregation on the short-circuit current during fabrication and operation for OPVs. They proposed a model for JSC decay based on the effective mutual diffusivity of the donor and acceptor phases:   n EA JSC ðt0 + ts Þ ts ¼ 1 + e kTs JSC ðt0 Þ te q

(19.5)

where n is the “Lifshitz-Sloyozov” power exponent, EA is the activation energy for mutual polymer diffusion, ts and ts are effective annealing temperature and time durEA

ing fabrication or under electrical stress, and teq  t0 e kT0 is the “equivalent anneal time” which translates the actual anneal time t0 to specific stress conditions. A variety of methods have been proposed to produce long exciton dissociating interfaces without relying on metastable phase separated domains including ordered BHJs based on imprint lithography (Yang et al., 2012), columnar growth of either donor or acceptor phases (Turak et al., 2011; Eccher et al., 2013; Yu et al., 2011; Zhang et al., 2007) and directed growth through spatial confinement of the polymer

622

Handbook of Organic Materials for Electronic and Photonic Devices

during deposition (Huck, 2005; Yang et al., 2005a; Himmelberger et al., 2013). Another method is to form a quasiplanar, bilayer structure by taking advantage of the miscibility of the donor and acceptor in the same solvent. The so-called interdiffused bilayer (ID-BL) is formed through sequential deposition from solution, where swelling and diffusion of fullerene into the underlying polymer layer is observed (Kim et al., 2016a; Seok et al., 2015; Zhang et al., 2014). High morphological stability is expected from such OPVs as the PCBM is deposited onto the highly ordered domains of the polymer bottom layer. Other approaches to morphological stability include introducing solvent additives (Liao et al., 2013), forming ternary BHJs rather than binary ones (Ameri et al., 2013; An et al., 2016; Baran et al., 2017; Lee et al., 2010a,b), and periodic polymerization (Watanabe et al., 2012).

19.3.3 Chemically induced degradation The interfaces at the electrodes are generally quite mobile, with significant reactivity and metal diffusion at both top and bottom contacts. At the ITO surface, depending on the organic molecule in contact, the indium (In) and tin (Sn) at the surface reacts with oxidizers in the organic layer (i.e., the S in PSS (de Jong et al., 2000b; Sharma et al., 2011a; Sˇkraba et al., 2011), chlorine (Cl) in PPV (Andersson et al., 1999; de Jong et al., 2000a)) to form various ions and compounds. The reaction to form metal ions at ITO has been attributed to the acidity of the organic layer (mostly with PEDOT: PSS; see Turak, 2013b) or the oxidation of the organic layer by dangling O at the ITO surface. These mobile metal ions then diffuse into the organic layers. Under extreme conditions, it can leave behind optically visible craters (see Fig. 19.12A; Chao et al.,1996). At the top contact, the interface depends on the reactivity of the electrode materials (Grozea et al., 2002; Huang et al., 2005; Turak et al., 2002, 2007; Turak, 2006). Typical low-work-function metals, such as magnesium (Mg), showed significant reaction/diffusion with destruction of the organic molecule (Turak et al., 2002; Turak, 2013b), as seen in Fig. 19.12B. A wide variety of polymers and oligomers have been found to be reactive with potential electrode materials, leading to device instability (see Turak, 2013b, for examples). On the other hand, if there is no chemical interaction, the interface can be diffuse, with penetration of the metal deep into the organic layers. Fig. 19.12C shows a transmission electron microscopy (TEM) cross-section of Au on DIP, showing major penetration into the metal (D€ urr et al., 2003b). The rate of diffusion is strongly correlated with the reactivity of the metals with the organic active layers (Turak, 2013b; Scholz et al., 2008; Turak et al., 2002; Huang et al., 2016), with less-reactive metals diffusing much faster in polymer and organic materials. Ag and Au typically show the largest diffusion depths, although significant metal penetration also has been observed for typical low-work-function electrode materials, such as Al and calcium (Ca) (Turak, 2013b; Greenbank et al., 2017; Huang et al., 2016). Most metal diffusion from both top- and bottom-contact interfaces occurs immediately upon deposition (for ITO, see Aristov et al., 2005; Crispin et al., 2003; de Jong et al., 2000a; Jo et al., 2008; Nguyen and de Vos, 2004; Sharma et al., 2011a; Wong et al., 2002; for top contact metals, see Scholz et al., 2008; Turak et al., 2002; Huang et al., 2016). High concentrations are observed through all the active layers (Lee et al.,

Device stability in organic optoelectronics

623

Fig. 19.12 (A) Scanning electron micrograph of the morphology of the cleaned ITO surface of a damaged device (ITO/C12OPPP/Ca/Ag, operating voltage 50 V). (B) Optical micrograph of degraded Mg/Alq3 interface, taken from the ITO side through the organic layer. (C) Crosssectional TEM images of the sample A4 (Tsub ¼ 70°C, Rate ¼ 0:35 A° /min). (Part A: Reprinted from Chao, C., Chuang, K., Chen, S., Chao, C.L., Chuang, K.R., Chen, S.A., 1996. Failure phenomena and mechanisms of polymeric light-emitting diodes: indium-tinoxide-damage. Appl. Phys. Lett. 69 (19), 2894–2896, with the permission of AIP Publishing. Part B: Reproduced with permission from Turak, A., 2013. Interfacial degradation in organic optoelectronics. RSC Adv. 3 (18), 6188, the Royal Society of Chemistry. Part C: Reprinted from D€urr, A.C., Schreiber, F., Kelsch, M., Carstanjen, H.D., Dosch, H., Seeck, O.H., 2003. Morphology and interdiffusion behavior of evaporated metal films on crystalline diindenoperylene thin films. J. Appl. Phys. 93 (9), 5201–5209, with the permission of AIP Publishing.)

1999; Greenbank et al., 2017). In and Sn have been observed up at the low-workfunction electrode interface (de Jong et al., 2000a; Gardonio et al., 2007; Nguyen et al., 2003; Sharma et al., 2011a; Sˇkraba et al., 2011), and penetration of top-contact atoms down to the substrate surface has led to secondary reactions, such as Al replacement of thiol bonds (i.e., self-assembled monolayers on Au; Fisher et al., 2002) or

624

Handbook of Organic Materials for Electronic and Photonic Devices

direct reaction with the carbonyls or hydroxides on the ITO surface, further limiting device performance. Although most diffusion occurs immediately upon deposition, significant electromigration (Cumpston and Jensen, 1996; Probst and Haight, 1997) is also possible under applied fields, destabilizing the interface even further. In defect areas at either electrode, where the electric field is particularly high, as electrons rush to the short-circuit region, metal atoms from the electrode also are carried along by the electron momentum. This leads to a metal atom pileup in the center of a short-circuit region, which is aggravated by high current and high temperature. The penetration of the metal into the organic layers is typically considered as detrimental to the device’s performance and lifetime. Although doping sometimes can have the beneficial effect of increasing injection (Parthasarathy et al., 2001; Scholz et al., 2008), or acting as an exciton-blocking layer (Mityashin et al., 2012; Wang et al., 2006) near the interface, the continued diffusion of metal atoms even at room temperature can limit device lifetime severely. The most obvious effect of metal diffusion is the development of electrical shorts when metal atoms penetrate across the organic film thickness (Cho et al., 2006; Cumpston et al., 1997; D€urr et al., 2003a; Probst and Haight, 1997; Scharnberg et al., 2008; Suemori et al., 2006). Although shorting has been directly observed with metal migration from the back-side contact, there has been only indirect evidence of In diffusion, causing shorting in diodes. Dark spots often occur above surface imperfections, such as In-rich spikes, where a higher electric field exists due to the thinner active layer. This increased electrical load causes significant local heating, which is thought to be high enough to decompose the ITO (Turak, 2013b). Even without full penetration, diffusion of metals into the active layer also has been correlated with PL quenching, with Ca seen to modify the EL spectra of PPV ( Janssen et al., 2004) and the PL spectra of 4PV (Choong et al., 1996); Al affecting the PL spectra for both Alq3 (Huang et al., 1998) and C60, and In yielding new features in the EL spectrum (Schlatmann et al., 1996) or completely quenching luminance in Alq3 (Lee et al., 1999). As it diffuses into the organic layer, Al catalyzes the reaction between O2 and C60 with illumination, leading to significant device degradation (Norrman and Krebs, 2006). Ca diffusion into P3HT:PCBM leads to the formation of organocalcium compounds that yield anion radicals that can act as recombination centers ( Jin et al., 2009). Sharma et al. (2011b) observed that In diffusion causes a 0.31-eV shift in the high-energy cutoff by UPS and a shift of 0.14 eV in the highest occupied molecular orbital (HOMO) position, suggesting dipole formation at the ITO/PEDOT:PSS interface. The primary mechanism for the loss of performance from metal diffusion is the formation of metal-induced gap states (MIGSs) (Turak, 2013b; Huang et al., 2016). Although generally considered necessary for charge exchange across the interface, they also can lead to trap states throughout the film thickness. These states trap electrons or quench radiation, reducing the mobility by up to two orders of magnitude (e.g., Ag in bathocuproine [BCP]; Huang et al., 2009). Al in particular interacts with a variety of organic materials, leading to CT states in C60 due to diffusion (Hebard et al., 1994; Peumans et al., 2000; Song et al., 2007), and the formation of dipoles at interface (Owens et al., 1995), with up to six electrons transferred to C60 from Al (Hebard et al., 1994). Similarly, up to four electrons were transferred to

Device stability in organic optoelectronics

625

OH-terminated SAMs after Al reaction with OH group (Fisher et al., 2002) and three for CH3O-SAMs (Fisher et al., 2002). These CT states significantly modify the doped layer conductivity. Although the role of metal dopants on degradation is somewhat controversial, given that they occur immediately upon deposition and often have beneficial effects (Turak, 2013b), it does appear that some degradation can be attributed to metal migration through the active layers, as the application of reverse bias to OLEDs has been seen to reverse degradation in the short term—metal migration back and forth has been observed as a function of the applied field direction (Kasim et al., 1997). Due to this mobility of the metal atoms in the organic matrix, shorting and healing cycles are sometimes seen as well, with performance fluctuating until the eventual formation of visible dark spots in the electrode.

19.3.3.1 Approaches to preventing chemical reactions Substrate modifications The chemical reactivity of the electrode surfaces with active layers was an early problem for organic electronics. For ITO, a variety of approaches have been adopted to prevent the breakdown of the electrode/organic interface, including acid treatments, plasma or UV ozone etchants, annealing, metal doping, and stabilizing interlayers (see Turak, 2013b for a comprehensive description of various approaches to stabilizing reactivity at interfaces). These approaches were geared toward modifying the surface of ITO to limit reactivity or to place a physical barrier at the interface. However, the choice of interlayer or treatment is very important; generally, approaches that increase the acidity of the surface (which is desired for high performance) lead to greater reactivity, whereas increasing the basicity of ITO increases the stability but limits performance (Turak, 2013b; Paniagua et al., 2016). PEDOT:PSS is an interesting case with regard to interfacial reactivity. Considered an essential component of most polymer-based devices (Wen and Xu, 2017), both PLEDs and OPVs, it is used in almost all polymer solar cells to improve device performance rather than for stability. Due to its high hole mobility, PEDOT:PSS is often actually classified as a conductor and referred to as a polymeric anode rather than a buffer layer. It was originally introduced to block the diffusion of In into MEH-PPV, but its high reactivity with ITO, unstable oxidation, and hygroscopic nature meant that it is often a primary source of degradation in both OLEDs and OPVs (Hains et al., 2010; Kawano et al., 2006; Nguyen and de Vos, 2004; Sharma et al., 2011a; Gevorgyan et al., 2016b). Gevorgyan et al. (2016b), in a metaanalysis of OPV stability studies, confirmed that the shortest lifetimes, particularly in the dark occur, in devices with a PEDOT:PSS interlayer. To continue the use of PEDOT:PSS for its electrical advantages, researchers either introduced another buffer layer beneath PEDOT, treated PEDOT:PSS postdeposition, or replaced PEDOT:PSS with a more stable interlayer (Turak, 2013b). Most interlayers are designed to block In diffusion into PEDOT:PSS. Wong et al. (2002) using short-chain alkylsiloxanes saw a drop of In concentration by a factor of 28–42.

626

Handbook of Organic Materials for Electronic and Photonic Devices

Allyltriethoxysilane had the most abrupt interface, attributed to cross-linking, which blocked mass transport through the SA. Other buffers including diamond like carbon, tetrahedral amorphous carbon, Au, and parylene, typically show a threefold improvement in the device’s lifetime (Turak, 2013b). Titanium also has been introduced to prevent reactions between PEDOT:PSS and other electrode materials such as Ag (Suh et al., 2012) in inverted type structures. These interlayers generally have the side benefit of further preventing the formation of dark spots by smoothing the surface (Chen et al., 2005a; Ke et al., 2004a,b), which often also improves the device performance (Turak et al., 2012; Heidkamp et al., 2016; Yoon et al., 2008). Another approach has been to treat the PEDOT:PSS after deposition to prevent or reverse adverse reactions. Although postdeposition annealing generally increases In diffusion, Nguyen and de Vos (2004) and Nguyen et al. (2004) were able to decrease the In content in the PEDOT:PSS films after treatment with 10% HCl. Chemical additives, such as an ethylene glycol, dimethyl sulfoxide (DMSO), and formic acid, improve the lifetime and performance of solar cells by removing the insulating and hygroscopic PSS components in the layer (Kim et al., 2011, 2016b; Chou et al., 2015; Mengistie et al., 2014). The PSS-depleted samples minimize residual water and decrease acidity, minimizing migration and corrosion. As these PSS components also can react with the other polymer layers (Norrman et al., 2006b), eliminating excess PSS leads to greatly increased lifetimes (1.3- to 5-fold improvement). Other treatments to improve wetting (Savva et al., 2015) or increase cross-linking (Duc et al., 2018) also have seen improvements in device stability. Most recently, there has been a move to replace PEDOT:PSS with transition metal oxides (Kim et al., 2005; Shrotriya et al., 2006a; Meyer et al., 2012), the most widely used of which is molybdenum trioxide (MoO3) for both OLEDs ( Jiang et al., 2007; Matsushima and Murata, 2008) and OPVs (Chen et al., 2012; Girotto et al., 2011; Jasieniak et al., 2012; Kanai et al., 2009; Kato et al., 2011; Liu et al., 2010a; Murase and Yang, 2012; Sun et al., 2011; Voroshazi et al., 2011; Zilberberg et al., 2012; Tan et al., 2013; Z€ ufle et al., 2015). The best improvement in device performance relative to PEDOT:PSS to date is seen for V2O5, with Z€ufle et al. (2015) reporting 187-fold improvement in t80 for P3HT:PCBM-based OPVs. Although the reactivity of PEDOT:PSS is often cited as the reason to explore new hole injection layers (HILs) in organic devices, researchers rarely examine the chemical stability of the newly proposed materials. Voroshazi et al. (2011), Jasieniak et al. (2012), Zilberberg et al. (2012), Z€ufle et al. (2015), and Girotto et al. (2011) all reported significant improvements in stability (>100 times) with MoOx interlayers, whether from vacuum or solution-deposition approaches. However, there has been some controversy with these interpretations. Recently, Bovill et al. (2015), using a multiplexer to collect multiple samples at once, and Kumar et al. (2016), under different ISOS testing protocols, have shown that PEDOT:PSS is more stable if the average device results are compared. Additionally, the shift to a new material systems means that new unexplored interfaces exist and new sites of instability are possible: Jasieniak et al. (2012) has reported that an unstable interface possibly exists between MoO3 and P3HT:PCBM, resulting in the appearance of an s-shape in the J-V characteristics.

Device stability in organic optoelectronics

627

Two-dimensional (2D) materials, such as graphene oxide (GO) and molybdenum disulfide (MoS2), are also a new promising PEDOT:PSS replacement. Murray et al. (2011) found a 13-fold improvement in t80 in a high-humidity environment for PTB7: PCBM solar cells using a GO interlayer in place of PEDOT:PSS, with an almost identical PCE of 7.5%. Liu et al. (2018) showed even-greater improvement for smallmolecule OLEDs with MoO3-doped MoS2 nanosheets, with about 24-fold improvement in luminance stability, at t90. Although there has been some success with new materials, unlike previous predictions (Turak, 2013b), PEDOT:PSS continues to be a significant part of next-generation devices.

Top-side contact modifications Similar to the ITO buffers, at the top-contact interface, the key approach to improving the stability of devices is to introduce a buffer layer. At the top-contact side, the buffer layer needs to block oxygen and moisture ingress through the electrode, prevent metal diffusion, enhance adhesion, and inhibit the reaction between the active organic and the electrode. To achieve these goals, a wide variety of interlayer materials have been attempted, including organic molecules, wide-band-gap inorganic dielectrics, and composites of organics with inorganic dopants, discussed in detail in Turak (2013b). There is typically an added benefit of increased performance with the introduction of such interlayers, which can act either as a dedicated electron-transporting layer (i.e., oxadiazole, triazole, quinoxaline, triazine-containing molecules, F-PCBM, Al-doped Alq3, LiFdoped C60) or an exciton-blocking layer. An exciton-blocking layer allows the transport of one charge carrier (typically the electron), but it also prevents the other from reaching the electrode and quenching charge injection. To accomplish this, it is typical to use a wide-band-gap organic insulator. Most widely used buffers are inorganic interlayers, originally introduced in OLEDs to increase performance. Among the many combinations attempted, those that were seen to have an impact on the stability of devices include CdSe, CsCl, CsF (with Yb or Al), MgF2, CaF2, NaF, Na3AlF6, AlF3, ZnF2, CrOx, Al2O3, Cu2O, TiOx, ZnO, and LiF (see Turak, 2013b and references therein). Although this is not an exhaustive list, it shows the breadth of materials that were tried in order to improve stability and performance. Often, very thin interlayers are sufficient to affect device stability significantly: Jin et al. (2009), for example, observed that only 5 nm of CdSe was needed to enhance shelf life, while preventing the development of a kink in the J-V curve. As an interlayer below Ag, ZnO allowed 1 month of continuous operation of a small-molecule OPV device (Hiramoto and Shiokawa, 2010). Al-doped ZnO allowed high-efficiency OPV cells (>10%) to maintain their performance with unencapsulated storage under ambient conditions for over 2 weeks (Liu et al., 2017). CsF with Yb (Chan et al., 2003) show very long lifetimes (500 h t80 at 500 Cd/m2). Submonolayer LiF increased the storage time for devices significantly (Turak, 2013b; Heidkamp et al., 2013). There are also a variety of mechanisms proposed for this improvement: CrOx (Wang et al., 2010a, 2011a) was observed to block the formation of interfacial oxides (Wang et al., 2011a), and prevent Al diffusion during deposition. CsCl and LiF

628

Handbook of Organic Materials for Electronic and Photonic Devices

prevents the direct reaction of Alq3 with Al (Yi et al., 2005; Grozea et al., 2002; Turak, 2006, 2013b). LiF pins grain boundaries in the active layer (Maye, 2011). TiO2, a promising interlayer that the Heeger group used to produce polymer OPVs with nearly 100% IQE (Park et al., 2009), has a very interesting impact on the stability. It is seen to initially increase performance, but it also introduces a new degradation mechanism if produced from vacuum sublimation: a reaction at the interface between TiO2 and the organic layer, with Al, or both simultaneously has been observed (Hansel et al., 2002). Lee et al. (2007) using a sol-gel process, and Wang et al. (2009), using a solution-based process to form polymeric TiOx, both saw substantial retardation of the degradation observed by vacuum evaporation, suggesting that oxygen doping leads to the loss of properties in the short term. Hayakawa et al. (2007) saw great improvement in stability over the first 100 h with a vacuum-evaporated TiOx layer when the device was not exposed to air or moisture. The Heeger group saw substantial oxygenblocking abilities with TiOx, which they attributed to a desiccant effect (Lee et al., 2007). TiOx layer acts as an oxygen scavenger, preventing O2 penetration into the organic layer, which results in an 11-fold improvement in shelf life. When used as nanoparticles, shelf life was seen to increase up to 100 times (Xiong et al., 2014). The polymeric TiOx, produced using a different precursor, appears to eliminate the reaction at the interface, although the mechanism remains unclear (Wang et al., 2009). The effects of TiOx, as well as CuPc and PEDOT:PSS (as described in previous sections of this chapter), introduces an important caveat in the use of interlayers to improve performance—the additional complexity with new layers sometimes can lead to new degradation mechanisms that were not observed without it (e.g., BCP and PBD crystallize; TiO2 undergoes an interfacial reaction). This underlines the importance of stability testing as a part of device design.

19.3.4 O2/moisture-induced degradation The largest external factor in device stability is oxygen- and moisture-induced degradation. All layers will react with molecular oxygen and water to varying degrees, causing degradation.

19.3.4.1 Active layers Within the organic active layers, CT complexes between oxygen and polymers act as exciton-quenching sites due to their strong electron affinity (i.e., electrons can hop to the next segment through the CT, but holes cannot due to the decrease in energy of both ground and excited states). Additionally, the triplet-exciton state leads to the formation of singlet oxygen. This highly reactive oxygen state attacks the molecule, causing cleavage of polymer chains or substitution on vinyl groups. While this greatly affects OLEDs, where triplets are formed directly during EL, this is also a significant problem for OPVs, where triplets are formed through interchain crossing from a singlet after photoexcitation. The composite layer in BHJ polymer devices aids in device stability (Neugebauer et al., 2000; Pacios et al., 2006); the fullerene, by dissociating singlet excitons before interchain crossing to the triplet state can occur, also prevents

Device stability in organic optoelectronics

629

the formation of the reactive singlet oxygen in the polymer. Although it is most damaging to polymers, singlet oxygen is also seen to affect small molecules such as pentacene (Sullivan and Jones, 2008). Both singlet-oxygen and CT formation result in decreased conjugation length and formation of local charged defects. As a result, oxidation generally leads to a longer tail of energetic trap states. This results in a higher trap density and higher saturation of traps (as seen with transient absorption kinetics). In a fresh sample, charge carriers are in shallow traps; in a degraded sample, charge carriers lie in deeper traps, and so charge emission from traps is curtailed. Further, with a higher density of traps, current flow slows. For most materials, the mobility is lowered and a shift to dispersive transport is observed. Pacios et al. (2006) provides a detailed description of these effects for interested readers. Oxygen incorporation without reaction is also detrimental to device performance (Khelifi et al., 2011; Kawano et al., 2006; Rusu et al., 2007), as oxygen itself quenches excitons, preventing EL in OLEDs and exciton dissociation in OPVs. The diffusion of oxygen into polymers and oligomers is quite rapid, but recovery of device performance is often possible with thermal annealing (Huang et al., 2005; Liao et al., 2008), nitrogen (N) flushing (Kaminorz et al., 1998), or chemical treatment (Oostra et al., 2015; Liang et al., 2011) to drive off the oxygen. The direct impact of ambient oxygen or moisture on device stability through organic layer oxidation has been observed for a wide variety of both donor and acceptor molecules, including PPV, P3HT, PEN, ZnPc, fullerenes, NPD, Alq3, BCP, and PEDOT:PSS, among many others (see Turak, 2013b and references therein). Deliberate doping of NPD with water at the NPD/emitting-layer interface, for example, decreases t80 by 82% (Yamamoto et al., 2012). C60 is particularly susceptible to oxidation (Heutz et al., 2004; Konenkamp et al., 1999; Rusu et al., 2007), becoming heavily p-doped by oxygen (Ng et al., 2009) or developing CT complexes or oxides (Tanaka et al., 2007). The oxygen impurity state in C60 lies 0.3 eV below the excited triplet state, acting as a deep trap with energy not far from the mid gap, quenching charge flow (Yang et al., 2004a). There has been much effort to produce electrically air-stable materials, especially acceptors, which are particularly sensitive to oxygen incorporation. Interested readers are directed to topical reviews by Usta et al. (2011) and Takimiya et al. (2007) for further details.

19.3.4.2 Bottom-contact surface Water is especially active at the ITO surface, enhancing any intrinsic reaction between the polymer and the surface. Sharma et al. (2011a) observed that the presence of interfacial water can increase the In concentration significantly in the active layers, enhancing any negative effects. Hydrolysis of residual water under an applied field is particularly damaging to the ITO surface and to any organic films in contact (Armstrong et al., 2009; Brumbach et al., 2007). Fig. 19.13 shows the process of water-activated liberation of In and Sn ions and compounds from the ITO surface under bias. Dark injection transient measurements (Khan et al., 2004) have shown that

630

Handbook of Organic Materials for Electronic and Photonic Devices

O H H

oxo hydroxyl In

O

(A)

In

O

In

OH Sn

(111) surface

O In In

In

O

+

In

+

In

Defects form in surface

O

O O O In Sn In In

Hydroxides/ OH OH carbonates In Sn form OH OH OH H H H H O O O OO O O In In In In Sn In In

In

H O

Damaged ITO surface HH H OO O O O In In Sn In

(B)

Fig. 19.13 (A) Schematic of the ITO (111) surface with hydroxyl and oxo terminations. (B) Schematic process for ITO damage, in which defects are introduced by the removal of electrons from the In or Sn atoms, couple with oxygen and water to form hydroxides and carbonates. The repulsion forces between the metal ions increase, causing an ejection of material from the surface.

electrical stress can change the ITO/PEDOT:PSS interface from Ohmic to blocking, with barriers as high as 0.8 eV formed. A likely explanation for these results is the decomposition of the ITO (Gardonio et al., 2007), where local electrical shorting leads to sufficiently high current to decompose the ITO surface from Joule heating. As mentioned previously, PEDOT:PSS is generally the layer most susceptible to degradation by external oxygen and water in both OLEDs and OPVs. Due to its highly hygroscopic nature (Duc et al., 2018; Lang et al., 2009) and high water solubility, PEDOT:PSS tends to attract and retain moisture, increasing the acidity of the interlayer and reactivity with ITO (Kim et al., 2011, 2016b; Chou et al., 2015; Mengistie et al., 2014). Unencapsulated P3HT:PCBM solar cell devices with PEDOT:PSS show rapid decay of the short-circuit current, with t80 essentially equal with or without ITO (Arora et al., 2011). The development of an s-shape in the illuminated diode curve for solar cells has been linked to PEDOT:PSS oxidation by Jin et al. (2009), Sharma et al. (2011a), and Voroshazi et al. (2011). Also, the metaanalysis of OPV stability studies by Gevorgyan et al. (2016b) has indicated that the PEDOT: PSS interlayer was the primary reason for the extremely short lifetimes. The transition metal oxides typically used to replace PEDOT:PSS are most effective in high-humidity environments. The prevention of humidity trapping and oxidation leads to significant improvements in high-humidity environments: Voroshazi et al. (2011) and Z€ ufle et al. (2015) report 100- to 200-fold improvement in P3HT: PCBM OPVs with metal oxides. To date, the best improvement in device performance relative to PEDOT:PSS is with V2O5, with Z€ ufle et al. (2015) reporting 380-fold improvement in high-humidity environments. Details on a variety of interlayer materials used to stabilize ITO interfaces are discussed in Turak (2013b).

19.3.4.3 Top contacts The oxidation of the low-work-function electrode is considered to be the main efficiency-limiting mechanism in both OLEDs and OPVs (Turak, 2013b). Although oxidation can be accelerated, during device operation or with illumination, the device is most vulnerable to oxidative attack during storage. The electrode, as the final layer in the device, generally undergoes preferential oxidation compared to the organic active layers. Wang et al. (2011b) observed that P3HT:PCBM/PEDOT:PSS/ITO solar cells being exposed to air for 2 days before

Device stability in organic optoelectronics

631

the deposition of electrode have the same performance as a fresh device prepared without air exposure. Even devices that have degraded can show partial or complete recovery of performance by removal and redeposition of the electrode (Cao et al., 2000; Wang et al., 2010b), or even just cleaning of the electrode surface with an etchant (Kaminorz et al., 1998). This preferential oxidation occurs partially because the electrode acts as a physical barrier (Pacios et al., 2006; Reese et al., 2008), but also because it tends to consist of highly reactive, low-work-function metals. Low-work function, compared to ITO, is generally preferred for the back contact in organic devices for high efficiency. These low-work-function electrodes, due to the ease of electron stripping necessary for high carrier injection/extraction efficiency, are also very unstable. An Alq3-based OLED with a pure Mg cathode, for example, can be operated in air for only a few seconds before the onset of irreversible degradation (Kiy et al., 2000). The initial rationale for the use of Al, a relatively high-workfunction material, as a counter electrode was its resistance to continuous oxidation through the formation of a self-passivating oxide (Turak et al., 2007). Unfortunately, Al itself is very porous to oxygen and moisture (Hermenau et al., 2011a; Norrman et al., 2006b, 2009). After a couple of days of storage, MEH-PPV OLED devices could not be operated due to Al oxidation (Kaminorz et al., 1998). With storage in dry N environments, Song et al. (2005) recorded the shelf life for small-molecule CuPc/C60/Alq3 devices as up to 3 months, and Hermenau et al. (2011a) observed as much as 67.5-fold improvement in t80 lifetime for P3HT:PCBM solar cells. Even just decreasing the water content can lead to substantial improvement in storage time, with Yang et al. (2010) reporting an eightfold improvement in t50 if RH is decreased from 65% to 33%, with the same O2 partial pressure. In OLEDs, the inhomogeneous loss of luminance is the main attribute associated with electrode oxidation; in OPVs, oxidation has been linked to increased series resistance, FF changes, and kinking in the current-voltage curves, in addition to the rapid decay of PCE (Turak, 2013b). Glatthaar et al. (2007) were able to model the kink by assuming that poor injection efficiency developed as a result of oxidation of the Al electrode. Similarly, increased series resistance can be explained as resulting from severely hindered charge extraction in the oxidized regions. Lloyd et al. (2009) observed a polarity reversal in OPVs with Ag electrodes exposed to ambient conditions for 11 days, suggesting that the Ag work function increased with oxidation beyond that of ITO. Similar behavior also was observed in small-molecule OLEDs with an Al/LiF electrode by Rocha et al. (2015), causing the sudden, catastrophic failure of the device. Generally, for both OLEDs and OPVs, changes in the barrier height at the electrode interface due to oxidation is considered a main influence on device performance degradation (Kawano and Adachi, 2010). A wide variety of techniques have shown evidence of oxidation at the electrode and, using a delamination technique to remove the electrode, a number of groups have examined both the top-side and internal interface of the electrode, showing evidence of oxidation at both surfaces (Turak, 2013b and references therein, and Vasilak et al., 2017). Although oxygen incorporation at the electrode/organic interface has been observed for a variety of active materials (including C60, Alq3, bathophenanthroline

632

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 19.14 (A) Schematic of oxygen and water ingress into organic layer, causing the formation of dark spots and edge growth of a dark front. (B) Light beam-induced current maps of as-prepared OPV devices with AZO showing “hot spots” that correspond to dark spots. (In part from Andersen, T.R., Almyahi, F., Cooling, N.A., Elkington, D., Wiggins, L., Fahy, A., Feron, K., Vaughan, B., Griffith, M.J., Mozer, A.J., Sae-Kung, C., Wallace, G.G., Belcher, W.J., Dastoor, P.C., 2016. Comparison of inorganic electron transport layers in fully roll-to-roll coated/printed organic photovoltaics in normal geometry. J. Mater. Chem. A 4 (41), 15986–15996, with permission of the Royal Society of Chemistry.) (C) Micrographs of OLEDs taken after 27 h of storage in 100% RH. (Adapted from Phatak, R., Tsui, T.Y., Aziz, H., 2012. Dependence of dark spot growth on cathode/organic interfacial adhesion in organic light emitting devices. J. Appl. Phys. 111 (5), 054512, with the permission of AIP Publishing.)

(Bphen), MDMO-PPV:PCBM, fluoropolymer, and P3HT:PCBM), oxidation appears to be inhomogeneous. These inhomogeneities are typically described by changes in local electrical properties (EL, PL, or photocurrent), as revealed by a 2D map of the electrode surface (see Fig. 19.14 and Turak, 2013b). These dark regions were among the first observed device failures. The typically observed exponential decay of electrical characteristics also support inhomogeneous oxidation, as described by Z€ufle et al. (2015) for OPVs. They determined that only a model consisting of multiple subcells of degraded and fresh sections fits the changing JV and transient photocurrent curves. A typical chemical profile through the electrode, such as a 3D time-of-flight secondary ion mass spectrometry (TOF-SIMS) profile of Al in a PLED (BulleLieuwma and van de Weijer, 2006 or an X-ray photoelectron spectroscopy (XPS) depth profile of Al in an OLED (Turak, 2006, 2013b), shows a layer of Al oxidation at top-side, metal Al, as well as islands of Al oxidation at the buried internal interface with the polymer. As a general rule, the smaller the active device area, the faster the degradation (Wang et al., 2011a), suggesting that some of the oxidation is controlled by the lateral diffusion of water and oxygen from the device edges. This can be observed directly as a dark front from the edges in EL/PL from the OLED surface or in the OPV photocurrent, as shown schematically in Fig. 19.14. Voroshazi et al. (2011) observed that the presence of PEDOT:PSS greatly accelerated electrode oxidation, promoting the infiltration of humidity from the device edges. Using this oxidation premise, they were able to establish an exponential relation between the device lifetime and increasing humidity levels.

Device stability in organic optoelectronics

633

In addition to this encroaching dark front, the electrode surface typically is covered with circular oxidation areas (see Fig. 19.14A and C), sometimes called “dark spots” (see Turak, 2013b and references therein for a detailed description of the mechanisms of dark-spot formation). Fig. 19.14B and C shows light beam-induced current (LBIC) and EL maps of OPVs and OLEDs, respectively, after the formation of dark-spot regions (Andersen et al., 2016; Phatak et al., 2012). The number of dark spots increases as a function of storage time, not applied bias, and depends on the inhomogeneities and pin holes that exist during deposition (Turak, 2013b). The electrode defect acts as a route for ambient air and moisture to enter the device, suggesting that dark-spot formation is a result of local oxidation of the interface around the defect site. Removal and redeposition of the electrode often result in a different distribution of dark spots (Liew et al., 2000). Probing the chemical structure of these dark-spot regions (Turak, 2013b), typically only metal oxides or increased O with no evidence of the organic are observed. Outside the dark spots, only a thin passivation layer of metal oxides are visible, combined with strong signals from the organic layers. Schaer et al. (2001) suggested that the change in volume associated with the formation of oxides at the interface can act as a mechanism for electrode delamination. Cross-sectional TEM of a Ca/Al bilayer electrode by Lloyd et al. (2009) shows voids, which grow in size during storage; XPS of those void regions indicates oxidation of the Ca. By deliberately introducing defects in Ca electrodes using polystyrene balls, Lim et al. (2001) and Lin et al. (2001) were able to correlate a linear increase in dark-spot diameter with electrode oxidation. The growth rate is also related to the initial size of pinhole defects in the cathode surface (i.e., the amount of water/oxygen ingress and oxidation is directly related to the initial pinhole size). Through a detailed series of isotope tracer experiments, the group of Krebs in Denmark were able to establish that O2 diffuses vertically through pinholes in Al and expands in all lateral directions, reacting with the underlying organic (Krebs et al., 2009; Krebs and Norrman, 2007; Madsen et al., 2011; Norrman et al., 2006a, 2009; Norrman and Krebs, 2006), but H2O diffuses rapidly through grain boundaries, causing uniform oxidation of the entire electrode surface (Hermenau et al., 2011a; Madsen et al., 2011; Norrman et al., 2009, 2010). As the majority of the instabilities at the top contact are related to oxidation due to exposure to the ambient air, the most common approach is to prevent oxygen and moisture from reaching the device through encapsulation. Encapsulation, however, is a kinetic approach that slows degradation but does not prevent the nucleation of defects. For successful long-term stability, barrier materials must have water vapor transmission rates (WVTRs) of 10-6/g2 per day for small-molecule OLEDs (Burrows et al., 2001) and 10-3/g2 per day for polymer OPVs (Cros et al., 2011; Hauch et al., 2008b), in order to ensure minimal water ingress into the device even in very harsh conditions. With such strict requirements, the first (and still most common) approach for OLEDs and OPVs is to use a rigid glass lid sealed by a UV-cured epoxy, often with a desiccant. Although this does a good job of slowing down darkspot formation, rigid glass encapsulation greatly increases the size and complexity of devices without completely eliminating degradation (i.e., penetration through the epoxy or saturation of the desiccant leads to catastrophic device failure). Additionally,

634

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 19.15 Schematic diagram for conventional (A) and inverted (B) OPV device structure. ETL and HTL are adopted to form ohmic contact and extract charges at cathode and anode, respectively. The arrows indicate the direction of electron and hole transport. (Reprinted from Lai, T.H., Tsang, S.W., Manders, J.R., Chen, S., So, F., 2013. Properties of interlayer for organic photovoltaics. Mater. Today 16 (11), 424–432, with permission from Elsevier.)

it is totally unsuitable for flexible devices. Therefore, an encapsulation solution is required that is thin, defect-free, light weight, easy to process, without significant thermal expansion mismatch with the electrode, and applied without damage to the electrode. A comprehensive examination of encapsulation techniques is beyond the scope of this chapter; interested readers are directed to reviews by Lewis and Weaver (2004), Park et al. (2011), Krebs (2006), and Ahmad et al. (2013). The second major development in preventing degradation due to oxygen and moisture from attacking the top contact is to eliminate the need for a low-work-function metal entirely, which is accomplished in an inverted device architecture, as shown schematically in Fig. 19.15 (Lai et al., 2013). In such a configuration, the ITO substrate is typically covered by an n-type nanoporous metal oxide, such as ZnO, TiOx, CrOx, or Cs2O. The top contact, a high-work-function material, is typically Ag. Using this configuration, Krebs (2008) was able to produce OPVs that can be stored, unencapsulated, in the dark for up to 6 months. Inverted cells are also stable under irradiation compared to conventional structures, in which the efficiency drops to 0 after 20 h of irradiation (Kuwabara et al., 2008). In addition to the higher oxidation resistance of high-work-function materials, the major benefit to this approach is that the oxidation of the top contact actually improves both performance and lifetime. As Ag oxide is also a p-type semiconductor with a high-work function (Hau et al., 2008), devices in the inverted configuration with TiOx (Kim et al., 2009) or ZnO counter electrodes (Krebs, 2008; Lloyd et al., 2009) at ITO perform very well (even better than initially) in a high-oxygen environment. Gevorgyan et al. (2016b) recently compiled a comparison of inverted and regular cells, showing that the best performance and stability under both dark and illuminated conditions are found in inverted structures. Interested readers are directed to a review by Hau et al. (2010) on inverted polymer OPVs.

Device stability in organic optoelectronics

635

It is also common to use interlayers to prevent electrode oxidation with ambient air, the most successful of which has been LiF. Turak (2013b) provides a detailed description of the wide array of interlayer materials successfully applied to prevent interfacial oxidation.

19.3.5 Charge carrier-induced degradation The gradual decrease in brightness for OLEDs without obvious change in device appearance has been attributed to the degradation of the active layers via the passage of charge carriers. This effect is most commonly seen in Alq3-based devices, where Aziz and co-workers elucidated that hole current caused the irreversible ionization of Alq3 to an Alq3++ moiety (Aziz et al., 1999; Wang and Aziz, 2013; Luo et al., 2007a; Siboni et al., 2011). As the anion is not luminescent, a slow fade of luminance occurs during operation. Similar defect states also have been observed for the passage of electrons through OC1C10 (a poly(p-phenylenevinylene) derivative) (Parker et al., 1999) and PEDOT:PSS (Crispin et al., 2003). Even if the polaron itself does not result in irreversible ion formation, the internal polarization field developed by the passage of mobile ions does eventually lead to electrochemical decomposition of the molecules for both OLEDs and OPVs (Giebink et al., 2009; Shen et al., 2000). This appears to be one of the major degradation mechanisms for thermally activated delayed fluorescence (TADF) OLEDs (Sandanayaka et al., 2015). For many conjugated molecules, hydrocarbons undergo dehydrogenation reactions due to electronic excitation caused by exciton-polaron annihilation (Giebink et al., 2009). The semireversible oxidation/reduction reactions of the molecules for exciton formation and charge recombination eventually lead to charge traps and excited-statequenching defects (Gregg, 2009; Jarikov et al., 2006; Jarikov and Kondakov, 2009; Jarikov, 2006; Kondakov, 2008; Kondakov et al., 2010, 2007). Once formed, these defects often dominate the photoelectrical properties, overwhelming the intrinsic carrier density, inhibiting transport of both excitons and charge carriers (Gregg, 2009). The density of excitons, therefore, can have a profound effect on luminescence decay, and Bangsund et al. (2018) was able to tune the decay properties just by increasing the size of the emission zone. Liang et al. (2011) showed that chemically treating electrically charged defects in P3HT can lead to more stable electrical performance. Hermenau et al. (2011b) found that what appeared to be a light-induced degradation of P3HT:PCBM actually was due to the presence of excitons. They observed that the decay curves scale with light intensity, but if they are normalized with respect to the number of extracted charge carriers, all intensities fall on the same curve (Hermenau et al., 2011b). Tsang et al. (2016) found that inserting 3-nm hole-blocking layers (HBLs) in TADF-based OLEDs could enhance the device stability almost eightfold by preventing the formation of deep traps that would have acted as sites for bimolecular exciton-polaron annihilation. One very successful method of improving device stability focused on combating the intrinsic degradation of Alq3 by holes has been doping of the Alq3 layer in small-molecule OLEDs with other molecules, including TPD, NPB, NPD,

636

Handbook of Organic Materials for Electronic and Photonic Devices

quadricone, styrlamine, DMQA, rubrene, DNP, Bphen, and perylene, among many others (see Turak, 2013b and references therein). This mechanism, known as hole blocking, also was suggested as another mechanism for a few doped HTLs (MADN:NPB, Tsai et al., 2005; F4TCNQ:NPB, Aziz et al., 2006; Luo et al., 2007b; DSA-Ph:NPB, Lee et al., 2005a; Ir(piq)3:NPD, Meerheim et al., 2006; and Ir(piq)2(acac):BCFA, Hong et al., 2015).

19.3.6 Photoinduced degradation When organic materials are exposed to light, they are subject to photochemical reactions (i.e., intramolecular, intermolecular, or reactions between the active layers and oxygen and/or water), which leads to a slow decay of the electrical properties. Although this effect is most evident in OPVs, photodegradation also affects OLEDs (Colditz et al., 2005; Cumpston et al., 1997; Wang et al., 2010b). In OLEDs, external ambient light, as well as the light the device produces internally, are sources of degradation. Manceau et al. (2011) performed a comprehensive study of photochemical stability of electroactive polymers. The general rules developed from that study indicate that side chains, vinyl units, and ketones are highly susceptible to breakdown. Exocyclic double bonds in the main backbone (PPV derivatives), quaternary sites on moieties, and readily cleavable bonds anywhere in the molecule are all typically unstable. Aromatic polycyclical units (such as thiophenes), on the other hand, offer good stability. Side chains play a key role, as they increase solubility, but they are highly unstable; therefore, cleaving them off after deposition greatly improves stability. Thermoclevable polymers were developed specifically for this purpose (Petersen et al., 2008). Although many photon-driven reactions require external oxygen (e.g., PPV, MEHPPV, MDMO-PPV, P3HT:PCBM, pentacene (PEN), PEDOT:PSS; see Turak, 2013b and references therein), some materials, such as C60 (Norrman and Krebs, 2006; Rusu et al., 2007) and PCBM (Wong et al., 2014), are also susceptible to photoinduced phase transformation, photopolymerization, or photoinduced diffusion of molecular oxygen. Sullivan and Jones (2008) observed purely photoinduced degradation, leading to a kink in the current-voltage curve with positive bias, which occurs regardless of the presence of oxygen in a PEN/C60 planar heterojunction OPV. When exposed to light, traps in C60 form, leading to the formation of space charges within the active layers (Ghorashi et al., 2012). The induced electric field decreases the exciton dissociation efficiency. Typically, however, due to the lack of bulky side groups and readily cleavable bonds, small molecules appear to be less susceptible to photodegradation. Franke et al. (2008) observed that ZnPc:C60 tandem cell are stable under illumination, and Kobrin et al. (2004) saw that the photobleaching of all color small-molecule OLEDs is completely recoverable. Photodegradation seems antithetical to the use of organic molecules for solar applications. However, recovery of performance has been observed when the device is stored in the dark for some time (Katz et al., 2006; Krebs and Norrman, 2007;

Device stability in organic optoelectronics

637

Reese et al., 2008). Outdoor testing of solar cells under natural light, with dark and light cycles, shows almost no photobleaching (Hauch et al., 2008a). Additionally, organic materials appear most susceptible to UV light, which can be filtered for higher stability (Ryu et al., 2010; Schafer et al., 2011). Although photodegradation has been a hurdle for commercialization, it does not appear to be a significant factor in future applications.

19.4

Development of device lifetimes

Although OLEDs and OPVs were introduced at the same time, OLEDs matured first. Much of the early degradation research focused on rigid encapsulation methods, successfully improving lifetimes so that commercial products were possible. The current standard for OLEDs for display purposes is for a t50 of at least 10,000 h at an initial brightness of 100 Cd/m2 (Aziz and Popovic, 2004; Cao et al., 2000); white OLEDs for lighting have different requirements, with t70 of 100,000 h at an initial brightness of 1000 Cd/m2 and 100 L/W (Kamtekar et al., 2010). For OPVs, prior to 2005, the focus was on increasing efficiencies to develop viable organic devices instead of on establishing stability. Most researchers initially relied on OLED solutions being directly transferrable to OPVs. Although there has been some crossover, the unique challenges of OPVs has lead to an increase in research on stability within the last 10 years ( Jørgensen et al., 2012), which also has spurred research in OLED stability (So and Kondakov, 2010). There is a general assumption that 1000–10,000 h (1–10 years) is sufficient stability for most OPV applications (Krebs and Spanggaard, 2005). To enter the solar cell market, the produced operational hours multiplied by the energy costs must match the production costs of the device. Under the assumption that the $/Wp costs of organic solar cells will be less than one-quarter of inorganic solar cells, lifetimes on the order of at least 5 years are required for organic solar cells to compete as a low-cost alternative (Brabec et al., 2005; Schuller et al., 2004). The National Solar Technology roadmap for the United States in 2007 (Ginley, 2007) has set a goal of degradation of 105 cm1), which enable significant light absorption even in thin films that are merely tens of nanometers thick (Hoppe and Sariciftci, 2004; Głowacki et al., 2013). In addition, organic materials are excitonic, meaning that photoabsorption does not directly generate free electrons and holes, as it does in silicon, but rather results in a Coulombically bound electron-hole pair that requires additional help (and consequently, energetic costs) in order to separate and extract the charges. While it is unlikely that OPVs will surpass traditional crystalline inorganic PVs in terms of PCE, they nevertheless possess many unique and valuable properties that have made them an active topic of research and a promising contender for niche applications and commercial production. The strength of OPVs does not lie in the promise of superior performance, but rather in their potential for high-throughput, low-cost

Organic photovoltaics (OPVs): Device physics

667

manufacturing, coupled with their ability to achieve respectable efficiencies. Indeed, as recently as 2016, experts have predicted that OPV modules can reach median prices of 0.28 USD/Wp in the near term, making them competitive with commercially available technologies (e.g., crystalline Si, CdTe, and CIGS), which are currently in the range of 0.35–0.6 USD/Wp (Gambhir et al., 2016). Another feature that may help thin-film OPVs become commercially competitive is their compatibility with flexible and lightweight substrates like plastics and foils (Forrest, 2004). While amorphous Si, CdTe, and CIGS solar cells also utilize thin films and are flexible to a certain extent, they typically still require high-temperature steps, the use of scarce elements, or both. The high-temperature processing, in particular, means that these cells require a great deal of energy to fabricate, which must be taken into account when calculating the energy payback period of a commercial solar cell. In contrast, OPVs can be fabricated at low cost using room-temperature, high-throughput solution-processing techniques. The use of flexible and lightweight substrates not only allows the solar cells to conform to curved surfaces, but also serves to reduce the cost of PV system installation, which includes labor, land/surface acquisition, and the installation of supporting structural components. For modern commercial solar cells, these so-called balance of system or soft costs currently exceed the cost of the modules (Fu et al., 2017a).

20.3

Principles of operation

20.3.1 Charge photogeneration in OPVs Compared to inorganic semiconductors like silicon, organic semiconductors have relatively low dielectric constants (εr ¼ 3  4). Because of this, following absorption of a photon, the excited electron-hole pair is poorly screened and remains bound by a Coulombic energy on the order of hundreds of milli-electron volts, which is significantly greater than the thermal energy at room temperature, kBT (Clarke and Durrant, 2010; G
elinas et al., 2011). This bound electron-hole pair is known as an exciton. In order to generate current, the solar cell must be able to overcome this exciton binding energy to dissociate and extract the resulting charges. The earliest OPVs were based on a simple metal/organic/metal device structure that relied on the built-in electric field between metal electrodes of sufficiently different work function to dissociate the exciton into free charge (Kearns and Calvin, 1958; Tang and Albrecht, 1975). However, the exciton dissociation efficiency in these simple devices was extremely low because the built-in electric field was typically too weak to overcome the large binding energy of the photogenerated excitons (Głowacki et al., 2013). A major breakthrough occurred in 1986, when Ching W. Tang developed the donor-acceptor (DA) heterojunction concept (Tang, 1986). This device concept uses a two-layer structure in which an energetic offset in the frontier orbitals of the two materials drives exciton dissociation via a charge-transfer (CT) process. The DA heterojunction is illustrated in Fig. 20.1. Because of the relative alignment of the donor and acceptor lowest unoccupied molecular orbitals (LUMOs), electrons can lower their energy by transferring from the donor to the acceptor. Likewise, the

668

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 20.1 Energy-band diagram of an OPV with a DA heterojunction. The relative alignment of the donor and acceptor HOMOs and LUMOs enables CT at the interface.

relative alignment of the donor and acceptor highest occupied molecular orbitals (HOMOs) drives hole transfer from the acceptor to the donor. Consequently, following light absorption, an exciton at the DA interface experiences an energetic driving force for CT. After the exciton has undergone the CT process, the electron and hole are now located on different molecules, significantly increasing the efficiency of exciton dissociation. The DA heterojunction concept was responsible for the first OPVs that exceeded 1% PCE, and the concept has since become ubiquitous in the field. Photocurrent generation in OPVs can be described by four sequential processes (illustrated in Fig. 20.2A): 1. Photoabsorption: When light is absorbed by an organic semiconductor, an electron is promoted to the LUMO of the molecule, leaving a hole behind in the HOMO and forming an exciton. In an OPV, useful photoabsorption occurs in either the donor or acceptor layers of the device. When absorption occurs in materials other than the donor or acceptor layers, such events typically do not contribute significantly to photocurrent, and this is termed parasitic absorption. The optical band gaps of the absorbing materials ultimately limit the range of photon energies that can be absorbed by the solar cell. 2. Exciton diffusion: After the exciton is formed, it must travel to the dissociating DA interface. The driving force for exciton diffusion is an exciton concentration gradient. Far from the DA interface, there is a buildup of photogenerated excitons, while near the interface, the supply of excitons is depleted by the dissociation process. However, excitons have a finite lifetime and can only travel, on average, a diffusion length LD before recombining. In organic materials, typical LD are on the order of 10 nm, while typical absorption lengths are around 100 nm (Clarke and Durrant, 2010). This sets up a trade-off between using thick films that absorb many photons (but are unable to harvest many of them due to the short LD) and thin films that harvest nearly all the excitons (but are unable to generate many in the first place).

Organic photovoltaics (OPVs): Device physics

669

Fig. 20.2 (A) Charge photogeneration in OPVs. The sequence of events is (Perez and Perez, 2015) photoabsorption (Mints, 2015), exciton diffusion (Conibeer, 2007), exciton dissociation via CT, and (Rand and Richter, 2014) charge collection. (B) Schematic of a DA bulk heterojuncton, in which the active organic layer is an interpenetrating network of donor and acceptor phases. This reduces the distance between the dissociating interface and the site of photoabsorption. To address this trade-off, researchers have developed bulk heterojunction devices, in which a thick film of an interpenetrating DA network enables substantial light absorption while still ensuring that a DA interface lies within a diffusion length of any excitons that are generated. The bulk heterojunction device structure is illustrated schematically in Fig. 20.2B, although in reality, the structure is substantially more complicated and often hierarchical (Chen et al.,

670

Handbook of Organic Materials for Electronic and Photonic Devices

2011). As a convention, a bulk heterojunction between two materials is denoted by a colon (A:B) whereas a planar heterojunction is denoted by a slash (A/B). 3. CT: After an exciton has diffused to the DA interface, the HOMO/LUMO offset at the interface drives CT. The electron is driven to the acceptor, while the hole is driven to the donor. At this point, the exciton is now in a so-called intermediate CT state, and while substantially more delocalized than before, it nevertheless is still characterized by a Coulombic binding energy in the range of 0.1–0.5 eV (Clarke and Durrant, 2010). The exact mechanism for CT is a topic of ongoing research (Vandewal, 2016; Gao and Ingan€as, 2014). 4. Charge collection: Following exciton dissociation at the DA interface, the charges must be transported across the organic layers to be collected at the anode and cathode contacts of the solar cell. In an OPV, contacts with dissimilar work functions establish an internal electric field that drives free electrons and holes toward their respective contacts. The electron travels to the cathode (typically Ag or Al), and the hole travels to the anode (typically a transparent conducting oxide such as indium tin oxide, ITO).

The efficiency of photocurrent generation in an OPV is summarized by a metric called the external quantum efficiency (EQE). The EQE is a wavelength- and voltagedependent value that describes the efficiency with which incident photons (of a given wavelength) are converted into photocurrent: EQE ¼

electrons extracted photons incident on the device

The EQE is also equivalent to the product of the efficiencies of each of the four sequential processes discussed here: EQEðλ, V Þ ¼ ηA ðλÞ  ηED ðλ, V Þ  ηCT  ηCC ðV Þ where ηA(λ) is the photoabsorption efficiency, ηED(λ, V) is the exciton diffusion efficiency, ηCT is the CT efficiency, and ηCC(V) is the charge collection efficiency. As an example, if a solar cell has a short-circuit EQE of 25% at a wavelength of 550 nm, then one out of four 550 nm photons incident on the device is converted to an electron-hole pair. Note that EQE is based on the number of incident photons on the device and therefore incorporates optical losses, such as loss of photons due to reflection from the substrate or parasitic absorption in nonactive layers of the device. Another metric, the internal quantum efficiency (IQE), describes the efficiency of photocurrent production due to photons that are actually absorbed by (and not merely incident on) the device: IQE ¼

electrons extracted : photons absorbed by active layers of the device

It is equivalent to the product of the efficiencies of the three internal processes in the device (exciton diffusion, CT, and charge collection): IQEðλ, V Þ ¼ ηED ðλ, V Þ  ηCT  ηCC ðV Þ ¼

EQEðλ, V Þ : η A ðλÞ

Organic photovoltaics (OPVs): Device physics

671

The EQE spectrum of an OPV can be directly determined by measuring the photocurrent response from the device while exciting it with a monochromatic beam with a known photon flux. The IQE represents an upper limit to the EQE of the solar cell because the number of absorbed photons will never exceed the number of incident photons. Determining the IQE spectrum of a device requires knowledge of the absorption spectrum of the active layers of the device. This cannot be determined easily through experimentation, due to the complication of thin-film interference in multilayer device stacks, and is usually modeled instead using transfer matrix methods (Pettersson et al., 1999, 2001; Burkhard et al., 2010).

20.3.2 Current-voltage characteristics A solar cell has the current-voltage (I-V) characteristics of a diode. Because the amount of photocurrent generated by a solar cell scales with the active area of the device, current density (J), often with units of mA/cm2, is reported instead of current. From the J-V curve, shown in Fig. 20.3A, we can identify several different working regimes of a PV device. At point (I), the two terminals of the device are at equal potential and the device is under short-circuit conditions. In the dark, a device operating at this point is under equilibrium and no current flows; consequently, the dark J-V curve passes through the origin. As shown in Fig. 20.3B, when the two terminals of the device are placed under a short circuit, there is a built-in electric field in the device. Applying a reverse voltage bias to the device (i.e., the anode is at a negative potential with respect to the cathode), serves to increase the magnitude of this built-in field. Under these conditions, illustrated as region II in Fig. 20.3A, a solar cell in the dark may have a small number of thermally generated charge carriers that are swept through the device by the built-in electric field and collected at the contacts. The current that is generated by the mechanism is defined as negative (i.e., electrons flow toward the cathode, while holes flow toward the anode), and comprises the relatively small, quasi-voltageindependent reverse bias current that is shown in region II of the dark curve in Fig. 20.3A. However, when the device is illuminated, a large number of photogenerated charges are introduced into the device, undergoing transport to the contacts via the built-in electric field, and are collected as photocurrent. Note that, like the reverse bias dark current, the direction of photocurrent flow is negative and quasivoltage-independent. As such, the effect of illumination on a solar cell’s J-V curve is to shift the entire dark curve downward by an amount equivalent to the photocurrent (which, however, is often voltage dependent in OPVs) under illumination. Finally, region III designates the solar cell under forward voltage bias (i.e., anode is at a positive potential with respect to cathode). Under these conditions, the applied voltage reduces the magnitude of the built-in electric field in the device, as shown in the forward bias case in Fig. 20.3B. While this has a relatively minor effect on the photogenerated and thermally generated currents (because even a small built-in field is sufficient to collect these charges), under this reduced built-in field, a larger population of electrons and holes may overcome the energetic barrier to produce

672

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 20.3 (A) Current density-voltage (J-V) characteristics of a solar cell operating in the dark and under illumination. Three regions of operation are pictured: (I) the short-circuit condition, (II) reverse voltage bias, and (III) forward voltage bias. For the J-V curve under 1 Sun illumination, the short-circuit current density (JSC), open-circuit voltage (VOC), and maximum power point, with coordinates of (VMPP, JMPP), are pictured. (B) Simplified energy-band diagrams showing an OPV under (I) short circuit, (II) reverse voltage bias, and (III) forward voltage bias conditions. The unlabeled bands represent the HOMO and LUMO of the donor and acceptor layers (refer to Fig. 20.1) and the Fermi levels of the anode (A) and cathode (C).

current flow in the opposite direction; that is, electrons can flow more easily from the cathode toward the anode (and vice versa for holes). Due to Fermi-Dirac statistics, the number of electrons that can overcome the energy barrier increases exponentially as the barrier is reduced by the applied forward voltage bias. Consequently, the device

Organic photovoltaics (OPVs): Device physics

673

J-V curve shows exponential behavior in the forward bias region III. When illuminated, the J-V curve shows similar behavior, but the entire curve is shifted downward by an amount equivalent to the photocurrent, as discussed previously. Unlike the dark curve, the illuminated J-V curve passes through quadrant IV (positive x-axis and negative y-axis) where power, the product of current and voltage, is negative. Therefore, this is the power-generating regime of the solar cell. There are several important parameters to note in the J-V curve of a solar cell under illumination. Within quadrant IV, the maximum photocurrent is generated when the applied voltage is zero. The photocurrent at this point is called JSC, or the short-circuit photocurrent density. At this point, the overall power generation is zero because there is no photovoltage, and thus P ¼ I*V ¼ 0. Within quadrant IV, the maximum photovoltage is generated when the photocurrent is equal to zero. The photovoltage at this point is called VOC, or the open-circuit voltage, and likewise, the overall power generation is zero. At points along the J-V curve that fall between short-circuit and opencircuit conditions, the product of J and V is nonzero and negative, and there is a point where the amount of power generation is maximized. This is called the maximum power point, and it has coordinates of (JMPP, VMPP). The maximum generated power density, PMPP, is JMPP*VMPP and can be visualized as the area of the shaded rectangle in Fig. 20.3A. Because JSC represents the maximum photocurrent that can be produced by the solar cell and VOC represents the maximum photovoltage that can be produced by the solar cell, their product, JSC *VOC, represents the maximum theoretical power density output from the solar cell. However, this is a nonphysical value because the solar cell cannot operate simultaneously at short-circuit and open-circuit conditions. A metric called the fill factor (FF) describes the ratio between the maximum power generated by the solar cell and the product of JSC and VOC: FF ¼

JMPP  VMPP : JSC  VOC

The FF can be visualized as how rectangular the solar cell J-V curve looks in quadrant IV. A solar cell with the highest (albeit nonphysical) FF of 1 will have a J-V curve that looks like a perfect rectangle with a corner located at the coordinate (JSC, VOC). Finally, the PCE describes the ratio between the maximum power output from the solar cell versus the power input from the sun: PCE ¼

Pout JMPP  VMPP FF  JSC  VOC ¼ ¼ , Pin Pin Pin

where Pin is the power density of 1 Sun, which, using the AM1.5G standard, equals 1000 W/m2 or, equivalently, 100 mW/cm2. There is an important relationship between JSC and EQE (which was introduced in Section 20.3.1). Both metrics describe the photocurrent response of the solar cell, although under different excitation conditions. While JSC is the photocurrent output from the solar cell under the AM1.5G standard illumination (a broadband source designed to simulate the solar spectrum), EQE is the photocurrent output at a specific

674

Handbook of Organic Materials for Electronic and Photonic Devices

wavelength. It follows that JSC is equivalent to integrating the EQE spectrum with the solar photon flux: JSC ¼

ð q η SðλÞλdλ hc EQE

where h is Planck’s constant, c is the speed of light, q is the elementary charge, and S(λ) is the solar spectrum with units of W/m2/nm. Performing a comparison between the direct measurement of JSC and the value obtained by integrating the EQE spectrum is a simple and important method for cross-checking measurements. In the absence of nonlinear behaviors or spurious measurement artifacts like current spreading, the two values of JSC typically should be within 10% of each other. However, many organic and organic-inorganic hybrid PV devices have a nonlinear intensity dependence due to exciton charge annihilation, in which a higher density of charges (due to higher light intensity) destroys a fraction of excitons, reducing their diffusion efficiency. For this reason, organic and hybrid device performance often peaks around 10%–50% of 1 Sun (Zimmermann et al., 2014). Consequently, JSC (which is measured under 1 Sun illumination) is typically less than or equal to the value calculated from the EQE spectrum (which is measured under significantly lower-intensity monochromatic excitation).

20.3.3 Sources of performance loss in OPVs In this section, we explore the sources of performance loss in OPVs. As we discuss these losses, it is helpful to keep in mind how each source of loss will affect the various performance metrics (PCE, EQE, JSC, VOC, FF) that were defined in Sections 20.3.1 and 20.3.2. We begin the discussion by considering the theoretical limits to solar cell performance. In an early work, William Shockley and Hans Queisser sought to determine the maximum theoretical PCE of a p-n junction solar cell using basic physical principles (Shockley and Queisser, 1961). For example, their model requires that the solar cell be at equilibrium with its surroundings and therefore emit blackbody radiation by virtue of its nonzero temperature. This emitted radiation is one unavoidable source of energy loss. Another aspect of the model is the fact that a semiconductor can absorb only photons with energies above its band gap, such that the low-energy, subbandgap portion of the solar spectrum remains unabsorbed, creating another source of inefficiency. Furthermore, when photons with energy above the band gap are absorbed, rapid thermalization converts this excess energy into heat as the charge carriers relax to the band edge. For the last two reasons, the maximum achievable PCE is a function of the band gap of the absorber. Using the Shockley-Queisser model, along with the AM1.5G solar spectrum, the maximum PCE of a single-junction inorganic solar cell is 33%, and this maximum coincides with a band gap of 1.3 eV (R€uhle, 2016). The Shockley-Queisser model assumes that photon interactions (absorption and emission) occur through generation and recombination of free electrons and holes. While this assumption is appropriate in inorganic semiconductors, where electrons and holes are generated freely upon light absorption, the situation is fundamentally

Organic photovoltaics (OPVs): Device physics

675

different in excitonic organic semiconductors. Due to the typically weak intermolecular electronic coupling and low dielectric constant in organic materials, tightly bound excitons are formed upon absorption. Likewise, excitons are formed during the encounter between a free electron and hole during the recombination and emission processes. As discussed in Section 20.3.1, the frontier energy offset between donor and acceptor is designed to dissociate the exciton. The free energy change associated with the CT process, Δ GCT, at the DA heterojunction aids in exciton dissociation but represents a significant source of energy loss. While an empirical minimum energy loss of Δ GCT  0.1 eV is required to ensure efficient CT, typical small molecule and polymer cells today commonly use large frontier energy offsets at the DA interface in the range of Δ GCT  0.3  0.5 eV. In 2011, Giebink and coauthors developed a model, based on that of Shockley and Queisser, that incorporates the unique behaviors of excitonic solar cells, arriving at a maximum theoretical PCE of 22%–27%, depending on the value of the free energy loss at the DA interface, Δ GCT (Giebink et al., 2011). Real devices experience efficiency losses that reduce their performance far below these theoretical limits, but some of these losses can be avoided through intelligent engineering. For example, the performance of a solar cell is heavily influenced by two parasitic resistances—the series resistance Rs and the parallel resistance Rp. The value of Rs incorporates the effects of resistance at the interfaces between different layers of the device, the conductivity of the organic layers, and the conductivity of the metal contacts. Ideally, Rs should be zero. The value of Rp incorporates the effect of shunt pathways through the device stack and can be due to defects formed during the device fabrication process. Ideally, Rp should be infinite. A nonzero Rs or a finite Rp will negatively affect a solar cell’s performance. For example, FF has the following dependence on Rs and Rp (Rand et al., 2007):     JSC Rs VOC FF Rs , Rp  FFð0, ∞Þ 1   : VOC JSC Rp From this equation, it is evident that FF decreases as Rp decreases, or as Rs increases. The series and parallel resistances also affect the values of VOC and ISC through their relationship with FF. In the case when Rs and Rp have negligible effect on cell performance, the following empirical relationship holds (Green, 1981):   VOC VOC  ln + 0:72 nV nV t FFð0, ∞Þ ¼ FFmax ffi t , VOC 1+ nV t where Vt is the thermal voltage (Vt ¼ kT/q). If, however, Rp is close to its ideal value of infinity but Rs is significantly larger than its ideal value of zero, then (Green, 1981)   Rs ISC FFs ¼ FFmax 1  : VOC

676

Handbook of Organic Materials for Electronic and Photonic Devices

In the case where Rs is negligibly small but Rp is significantly smaller than its ideal value of infinity, then (Green, 1981)  3 VOC + 0:7 6 FFmax VOC 7 nV t 7: FFp ¼ FFmax 6 1  4 VOC Rp ISC 5 nV t 2



Finally, we consider the effects of series and parallel resistance on the J-V characteristics of a solar cell. The generalized Shockley diode equation is (Green, 1981)      Rp qðV  JRs Þ V J ðV Þ ¼ J0 exp  1 +  Jph ðV Þ , nkB T Rp Rs + Rp where n is the ideality factor, kB is the Boltzmann constant, T is the temperature, Jph is the photocurrent at a given voltage, and J0 is the reverse bias saturation current density (also known as the dark saturation current density). The presence of nonideal series and parallel resistance is one source of inefficiency in solar cells. As we saw in the previous discussion, nonideal Rs and Rp affect FF, VOC, and JSC—all of which are intimately linked to the overall PCE of the solar cell. While these principles are broadly applicable to all PV devices, in OPVs, there is an additional source of loss that, most notably, affects the maximum attainable VOC. Earlier, we had briefly discussed the exciton binding energy and the change in free energy (Δ GCT) associated with dissociating the exciton (i.e., CT). In a solar cell, VOC is dictated by the difference between the hole and electron quasi-Fermi levels at the anode and cathode contacts. In early single-layer OPV devices, the maximum quasi-Fermi level splitting is theoretically limited to the band gap of the absorber, but in practice, it is more tightly limited by the difference in the work functions of the two electrodes. In modern DA heterojunction OPVs, the VOC is generally understood as being limited by the energy difference between the HOMO of the donor and the LUMO of the acceptor minus the binding energy of the geminate (i.e., originating from the same photoabsorption event) electron-hole pair that is created during the CT process (Rand et al., 2007): max qVOC ¼ IPD  EAA 

q2 , 4πε0 εr rDA

where IPD is the ionization potential of the donor, EAA is the electron affinity of the acceptor, ε0 is the vacuum permittivity, εr is the dielectric constant of the DA interface, and rDA is the initial separation between the electron and hole immediately following CT at the interface. The third term here is simply the Coulomb interaction between the electron and the hole. Continuing, if we solve the generalized Shockley diode equation for VOC, assuming Rp ! ∞ and Rs ¼ 0, we obtain:   nkB T Jph ðVOC Þ ln VOC ¼ +1 : q J0

Organic photovoltaics (OPVs): Device physics

677

This result indicates that VOC also depends on the reverse bias saturation current density, J0. Larger values of J0 further limit VOC from approaching Vmax OC . Nevertheless, researchers have been able to demonstrate certain OPV systems where Vmax OC is attainable. One example is the boron subphthalocyanine chloride (SubPc)/C60 heterojunction, where Vmax OC  VOC  1 V (Mutolo et al., 2006). So far, we have considered VOC loss in terms of interface energetics (the donor HOMO/acceptor LUMO gap and Δ GCT) and process engineering (parasitic series and parallel resistances). In addition to these energetic and practical considerations, the value of the VOC is also highly dependent on recombination dynamics. Conceptually, a bimolecular recombination event can be broken down into two stages and their associated rates. In the first stage, an independent electron and hole meet and form a Coulombically bound pair; in the second stage, the electron-hole pair recombines and two neutral molecules are recovered (Liu et al., 2015). The Langevin model is the most commonly invoked description of bimolecular recombination in OPVs and assumes that the first step (the encounter) is rate-limiting, so that once the pair is formed, it will rapidly recombine. Such a model assumes direct recombination of two mobile charges, such that VOC is directly related to the difference between the transport levels at the DA interface (i.e., IPD  EAA) (Liu et al., 2015). However, the Langevin model consistently overpredicts experimental recombination rates (Lakhwani et al., 2014) and fails to model observed relationships between VOC, the value of IPD  EAA, and their temperature dependence (Burke et al., 2015). Consider, instead, if the species participating in the recombination event is an interfacial CT state (and not a pair of mobile charges, as in the Langevin model). Then, the VOC is expected to correlate with the CT state energy (and not the energy difference between the transport levels, IPD  EAA) (Liu et al., 2015). Because of these distinctions, the nature of bimolecular recombination at a DA interface (Langevin versus CT-facilitated) is intimately linked to the value of VOC. Recent theoretical work has proposed that, in contrast to the encounter-limited model in which every encounter between an electron and a hole at the interface results in recombination, an equilibrium may exist between intermediate CT states and the free charge states (Burke et al., 2015). In fact, the more accurate description in most OPVs may be that the redissociation rate of CT states back to the free charge states is much greater than the recombination rate of CT states to the neutral ground state. While this model represents a major reassessment of traditional models of bimolecular recombination in OPVs, it nevertheless was successful in explaining relationships between VOC, the CT state energy, as well as their temperature dependence in both experiment and device simulations (Liu et al., 2015; Burke et al., 2015). The authors subsequently derived a relationship between VOC and various device parameters and assessed the relative importance of different sources of voltage loss (Burke et al., 2015): qVOC ¼ E0  EB 

  σ 2CT qf N0 L  kB T log , τCT JSC 2 kB T

678

Handbook of Organic Materials for Electronic and Photonic Devices

where E0 is the difference between the transport levels at the interface (IPD  EAA), EB is the binding energy of the CT state, σ CT is the degree of interfacial energetic disorder, f is the volume fraction of the solar cell that is mixed or interfacial, N0 is the density of electronic states, L is the thickness of the solar cell, and τCT is the average lifetime of a CT state before it recombines. Notably, based on this result, the parameters that appear in the logarithm would have to be altered by orders of magnitude in order to substantially affect the value of VOC. Reducing the subtracted values outside of the logarithm—the CT state binding energy and the degree of interfacial energetic disorder—is more effective at improving VOC (Burke et al., 2015). Both of these parameters are highly sensitive to the morphology at the DA interface, so optimizing this interface represents a great opportunity for enhancing the voltage output from an OPV.

20.4

Device structures

The basic design of an OPV includes several organic layers sandwiched between two thin-film electrodes. Each layer is typically less than 100 nm thick, and at least one of the electrodes must be partially transparent in order to allow light penetration. The most common anode material is ITO, a highly doped transparent oxide, which is sputtered on a transparent supporting substrate such as glass or a flexible substrate such as polyethylene terephthalate (PET). The opposite electrode is usually highly reflective (e.g., Ag or Al), so any photons that are transmitted through the active layers of the device can reflect off the rear electrode to increase active layer absorption. Devices can be fabricated in the so-called normal (sometimes called conventional) and inverted geometries. In normal geometry, light enters through a transparent anode while a reflective contact serves as the cathode of the device. This geometry was invoked in early, high efficiency designs (Li et al., 2005) and is still commonly used today. However, the low work function metal cathode in this design is almost exclusively prepared through thermal deposition techniques, making the normal geometry less amenable to large-scale fabrication, which favors solution processing. Moreover, others have cited concerns about the long-term operational stability of devices in which the exposed top contact is a low work function metal because these oxidize easily in the presence of oxygen or water (Zhou et al., 2012). In the inverted geometry, light enters through a transparent cathode, while a reflective contact serves as the anode of the device. In this geometry, the reactive, low work function cathode is buried under several organic layers, and a less reactive, solution-processable metal (e.g., a sinterable Ag ink) can be used as the reflective anode contact (Krebs et al., 2009; Girotto et al., 2009). As discussed in Section 20.3.1, the earliest OPV designs used only a single active organic layer sandwiched between the two electrodes. However, these devices did not produce charge efficiently because the built-in electric field in these devices was insufficient to overcome the considerable binding energy of excitons. The introduction of the DA bilayer structure dramatically improved device performance. Further refinement of the DA morphology to create an interpenetrating network of the two

Organic photovoltaics (OPVs): Device physics

679

materials (bulk heterojunction), shown in Fig. 20.2B, yielded further improvements to device performance (Hoppe and Sariciftci, 2006; Park et al., 2009). While the electrodes and donor and acceptor materials form the basic OPV device structure, additional so-called buffer layers are often incorporated between the electrodes and active layers of the device. These layers can have several functions. For example, when a solar cell is illuminated with light, it is advantageous to maximize the electric field intensity near the DA heterojunction, so that a large concentration of excitons can be generated within a diffusion length of the dissociating interface. One strategy is to incorporate optical buffer layers that manipulate the electric field distribution within a device stack. Buffers can also be used as sacrificial layers to prevent damage to the active layers of the solar cell during the aggressive deposition of a metal electrode. Thin layers ( 2 V exceeding the leakage current regime, scales with the band gap of the emitters used. For this point in the j-V characteristics, often the term turn-on voltage is used (Tanaka et al., 2007; Su et al., 2008), and it is correlated to the peak emission wavelength of the EL. Even more reports suggest that EL takes place at or even below the theoretical limit (i.e., when Epeak  eVon). However, there are two points to consider: First, the value of Von cannot be determined experimentally properly, as it depends on a parasitic level of the leakage current; and second, emission at voltages below the peak energy Epeak is possible and can be explained with established semiconductor theory (W€urfel, 2005). Apart from the expected correlation of the onset of the exponential regime of the three example OLEDs with the emitter band gap, the overall shape of the curves differs. As the j-V characteristics are a consequence of bipolar transport through various layers and across various interfaces and recombination to neutral excitons, many material- and architecture-related parameters influence the overall transport of charge carriers. Here, sophisticated drift diffusion (Schober et al., 2010) or kinetic MonteCarlo (Coehoorn et al., 2015; Mesta et al., 2013) simulations are needed to explain the experimental data. The EL in Fig. 21.2B shows typical spectral distributions—one for a blue fluorescent emitter and two for phosphorescent green and red emitters. While all these

Organic light-emitting diodes

Fig. 21.2 Key device characteristics of three representative monochrome OLEDs (red, green, and blue). (A) Current density-voltage-luminance (j-V-L) characteristics. (B) Chemical structures of the emitter molecules and EL spectra of the same devices. Credit: PAW/IAPP.

699

700

Handbook of Organic Materials for Electronic and Photonic Devices

single-emitter devices are referred to as monochrome OLEDs, as mentioned previously, their spectrum strongly differs from a purely monochrome color stimulus. Typical full width at half maximum (FWHM) values are in the range of 50–100 nm (Reineke et al., 2010). For display applications, the FWHM of the EL spectrum is subject to optimization, where smaller values improve color purity, which is especially important for green emitters, as they can comprise blue and red parts of the spectrum. The narrowing of the FWHM can be achieved through the design of either the emitter molecules (Fleetham et al., 2014; Li et al., 2014) or the optical device (Meerheim et al., 2008). It is important to note that there are two distinct concepts for realizing the EML of a monochrome device. Either a host-guest system or a bulk emitter film is used. In the former, the guest is the emitter molecule and the host an energetically suited wideband-gap material embedding the emitter. The host material is often needed because many emitters do not show efficient luminescence at high concentrations due to emitter aggregation (Kawamura et al., 2006). To prevent this aggregation effect, the emitter is diluted into the host material. This concept also can serve the purpose to decouple the emitter from transport duties, which may prove beneficial for both the efficiency and stability of the OLEDs. Alternatively, monochrome OLEDs can be made from bulk films of emitters (Rosenow et al., 2010). This is possible whenever the emitter behaves different to the trend mentioned here (i.e., it shows efficient luminescence at high concentration—an effect called aggregation-induced emission (AIE)). These arguments hold equally true for the development of white OLEDs.

21.2.3 White light-emitting OLEDs White OLEDs were initially developed for their potential use in general lighting settings, but are currently mostly used in large-panel displays in the red, green, blue, white (RGBW) concept discussed in Section 21.4.1, where white EL is used as the primary color and converted into differently emitting pixels in combination with color filters. In general, white light can be realized even with a set of different monochromatic spectral lines distributed in the color space. At the same time, the color perception of objects illuminated by such a white light source—quantified by the so-called color-rendering index (CRI; Reineke et al., 2013)—would be poor. Here, the spectral broadness of organic emitters (FWHM typically between 50 and 100 nm) used in OLEDs are advantageous (Reineke et al., 2010). Fig. 21.3 shows the primary EL spectra of the RGB emitters Ir(piq)3, Ir(ppy)3, and 4P-NPD, respectively. Their spectra mark certain positions in the CIE color space, as shown in the center graph in the figure. The emitter contributions can be mixed in an additive fashion to generate white light. To realize high CRI values, the generated EL should be designed to match the emitted spectrum of a given black-body radiator with the corresponding color temperature on the Planckian locus ( Jou et al., 2013; Wu et al., 2016). There are two important color points on the Planckian locus: points E and A, which are widely accepted color standards. Point E is often referred to as the point of equal energy, which qualitatively fits with the very similar contributions of the composing monochrome emission bands (cf. Fig. 21.3). In contrast, point A, the warm white color point, contains

400

500

600

Ir(ppy)3

(0.294/0.620) 470 lm/W

0.5 0.0

500

600

700 (0.154/0.072) 74 lm/W

0.5 0.0

NTSC

0.6 0.5

400

500

600

Wavelength (nm)

700

0.4

T (K)

0.3

2.000

3.000

1.0

A

4.000

1.000

E

10.000

400

500

600

E

700 (0.333/0.333) 215 lm/W

Ir(piq)3

∞

0.2

0.0 0.0

(0.447/0.407) 210 lm/W

0.5

0.0

0.1 4P-NPD

A

Ir(ppy)3

6.000

400

1.0 Intensity

700

Intensity

0.7

Intensity

1.0

1.0

0.8

CIE y

Intensity

(0.682/0.318) 113 lm/W

0.5 0.0

Intensity

0.9 Ir(piq)3

Organic light-emitting diodes

1.0

0.5

4P-NPD

0.1

0.2

0.3

0.4 CIE x

0.5

0.6

0.7

0.8

0.0

400

500

600

700

Wavelength (nm)

Fig. 21.3 The CIE 1931 color space (center) and various representative EL spectra generated by OLEDs. The spectra of the monochrome OLEDs (left) are experimental data based on the emitters Ir(piq)3 (red), Ir(ppy)3 (green), and 4P-NPD (blue). Their associated color points are labeled in the CIE 1931 diagram. The shown white spectra (right) are calculated based on different compositions of these primary monochrome devices to demonstrate white spectra at color points A and E. Also shown here are the NTSC display standard (black triangle) and the Planckian locus (black curve; the different values represent perceived colors of a black body radiator at various temperatures). The values in each spectral plot are CIE (x/y) coordinates and the luminous efficacy of radiation Kr. Source: PAW/IAPP. 701

702

Handbook of Organic Materials for Electronic and Photonic Devices

much less blue light. The value Kr given in the panels of Fig. 21.3 for the various spectra is called luminous efficacy of radiation, and it represents a conversion of radiometric and photometric units. Its value is directly proportional to the luminous efficacy (given in lumens per watt) of the final device, which is the figure of merit for light sources. For instance, comparing the values Kr, E ¼ 215 lm/W and Kr, A ¼ 210 lm/W for the abovementioned color points and the very spectra shown in Fig. 21.3, one can directly derive that intrinsically the spectrum at color point E will yield slightly higher luminous efficacy values. The CRI values of the two white spectra are 69 and 61 for color points E and A, respectively. Clearly, these values are not very high, which is due to the significant gaps between the individual emission bands. Technologically, there are many ways to realize white EL from OLEDs (Reineke et al., 2013). Fig. 21.4 summarizes various white OLED designs. White light emission can be realized by an assembly of monochrome OLEDs, which can be fabricated either in a vertical (Bulovic et al., 1996; Rosenow et al., 2010) or horizontal (Krotkus et al., 2016) fashion (cf. Fig. 21.4A and B, respectively). In principle, the lateral concept does not differ much from an RGB display made of OLEDs—it only lacks the complex driving circuitry. The vertical stacking of OLEDs is feasible only for vacuum-based processes, as they allow the nondestructive additive deposition of as many layers as are needed. In contrast to these multiple-unit concepts, white light also can be generated in a single layer, as shown in Fig. 21.4C. Typically, an OLED EML has a thickness of 20–40 nm. Thicker layers are not feasible because nondoped layers would strongly increase resistive losses. Within these 20–40 nm, many concepts have been developed to realize white light. A crucial limitation of color mixing is the fact that the region within the EML where excitons are generated is usually only a few nanometers broad, and often close to one of the interfaces to adjacent layers (Reineke et al., 2007). Thus, often the volume in which the exciton distribution to various emitters needs to take place is severely restricted. Clearly, one way to realize white light is to divide the EML into sublayers, wherein each of them is containing one of the RGB emitters (Reineke et al., 2009). Alternatively, all emitters can be

W

R G

B R

G

B

W

B

(A)

R/G/B R+G+B

(B)

(C)

R

or

(D)

Fig. 21.4 Schematic illustrations of various concepts to realize white OLEDs. (A) Vertical stacking of RGB monochrome devices, (B) lateral structuring of RGB pixels, (C) single-unit white OLEDs, and (D) white-light generation through a combination of blue EL and at least one color conversion layer. For the single unit concept in (C), there are again various ways to generate white light in the EML. W, white-emitting material; R/G/B, sublayers for the primary colors; R + G + B, mixing of all emitters in one layer. Source: SR/IAPP.

Organic light-emitting diodes

703

mixed into one common layer (D’Andrade et al., 2004); however, here, the respective concentrations of the emitters need to be optimized carefully, as the distribution of excitons among the different emitters is strongly affected by nonradiative energy transfer processes. Through careful material design, single-emitter while EL can be realized. Here, one type of emitter actually shows two spectrally distinct emitting states: one monomeric and one excimeric. When carefully adjusting the ratio of monomer and excimer emission bands, white light is possible with high efficiency (D’Andrade et al., 2002; Cocchi et al., 2007, 2009). For completeness, white light can be generated from a combination of a high-energy EL (blue) and color-conversion layers, that down-convert part of the primary EL spectrum to compose white light (Gohri et al., 2011), as shown in Fig. 21.4D. It is important to note that, while the previous discussion centered on mixing the EL of RGB emitters, of course white light also can be generated by only two colors, which show CIE coordinates that cross the desired white color point on the Planckian locus on a straight connecting line (cf. Fig. 21.2; Fries et al., 2017; Springer et al., 2016; Su et al., 2008; Krotkus et al., 2016). Often, however, the CRI of two such color devices trails their three emitter analogs. On the other extreme, to generate white light with even higher CRIs, four emitter concepts have been proposed (Rosenow et al., 2010; Jou et al., 2013). The generation of white light based on polymeric materials cannot be realized easily using the vertical stacking approach (cf. Fig. 21.4A), as the fabrication of multiple units on top of each other currently is impossible with wet-processing techniques. Here cross-linking and orthogonal processing limit the device’s design freedom. On the other hand, the lateral concept (cf. Fig. 21.4B) is perfectly suited, especially for inkjet-printing techniques. Polymers with their inherent large physical size compared to small molecules have a number of advantages to realize white light “by (polymer) design.” Here, polymers are developed that either comprise various chromophores in a main polymer backbone (Tu et al., 2004; Lee et al., 2015) or utilize a copolymer architecture with chromophores in main- and side-chain positions (Liu et al., 2007). All these polymer concepts can be realized with precise stoichiometric compositions, which guarantees good control of the additive color mixing of various emitter subsystems (chromophores).

21.3

Efficiency considerations

Optimizing the efficiency of a given light source is of utmost importance for two reasons: (1) only respectful and conservative use of energy will allow a future that tries to avoid negative impacts on our environment, and (2) many electronic devices and applications become portable or small (or both), where an increased power consumption will strongly hamper their time of use in a disconnected fashion. The OLED efficiency, therefore, is one of the key parameters that triggered the decades-long research up to now and will remain to be the central driving force for more research efforts.

704

Handbook of Organic Materials for Electronic and Photonic Devices

21.3.1 Parameters influencing the efficiency The external quantum efficiency (EQE) of an OLED, which is the most representative performance indicator quantity when comparing differently emitting OLEDs (Reineke et al., 2013), as it gives the ratio of extracted photons over injected charges. The other often-used performance indicator quantities are current efficiency (given in candelas per ampere) and luminous efficacy (given in lumens per watt), which both factor in the human eye response function and thus cannot be used to compare OLEDs with different EL spectra directly. The EQE can be expressed as the product of four main influencing parameters: ηEQE ¼ γ  ηrad, eff  ηS=T  ηout

(21.1)

The first factor, γ, represents electrical efficiency. It accounts for injected charge carriers that do not contribute to recombination through the formation of excitons. Often, this factor is also referred to as charge balance, and it directly depends on the transport properties of the multilayer devices. Still, mostly determined indirectly from the discussion of the remaining parameters influencing ηEQE, the electrical efficiency is close to unity in optimized OLEDs (Adachi et al., 2001b; Erickson and Holmes, 2011, 2014). Furthermore, ηrad, eff—the second factor in Eq. (21.1)—accounts for the effectiveness of a certain emitter. It treats the competition between the radiative and nonradiative rates of the emitting state that is utilized in a given emitter, which can be either a singlet or a triplet state, and hence the actual definition of ηrad, eff depends on the system. This parameter is called effective radiative quantum efficiency because it also accounts for possible enhancements of the radiative rates through the optical environment (Amos and Barnes, 1997)—that is, the OLED forms a thin-film optical cavity (Furno et al., 2012; Br€ utting et al., 2013). The third factor, ηS/T, accounts for the share of excitons that are able to decay radiatively due to the quantum-mechanics spin selection rules. It will be discussed further in Section 21.3.2. The electrical efficiency, the effective radiative efficiency, and the spin factor can be summarized to the internal quantum efficiency as ηint ¼ γ  ηrad, eff  ηS/T, which describes the efficiency of photon generation of an emitter for the case of EL. The final factor, ηout, is the outcoupling efficiency. It depends on the optical environment and, as more recently was discovered, on the orientation of the transition dipole moment of the emitter molecules. A discussion of ηout will be given in Section 21.3.3.

21.3.2 Exciton spin EL in OLEDs is realized through the recombination of uncorrelated electrons and holes to form an intermediate, neutral state—the exciton. OSCs, with their low degree of order and their low dielectric constants, typically possess very localized, Frenkeltype excitons. Their localization leads to the occurrence of energetically distinct spin manifolds for the excitons in form of singlet and triplet states. These are energetically split, with typical energy splittings in the range of few 100 meV (Reineke and Baldo, 2014), and show radiative recombination typically only from the singlet state

Organic light-emitting diodes

705

(Turro et al., 2010). This process is referred to as fluorescence. The radiative recombination from the triplet state, called phosphorescence, is very unlikely, as it requires a spin flip (Turro et al., 2010). Before recombination, the respective electrons and holes are completely uncorrelated, so that the pairing of the two charge-carrier spins (fermions with spin-half ) to one bosonic exciton is a truly statistical process. As the triplet states of organic molecules are threefold degenerate, there are three possibilities to form a triplet for each singlet state, with the result that in such uncorrelated systems, about 75% of the excitons come to life as triplets (Baldo et al., 1998; Segal et al., 2003; Burrows et al., 2001; Reineke and Baldo, 2012). Thus, if a conventional fluorescent material is used in OLEDs as the emitter, the internal quantum efficiency ηint, F will be approximately only 25%, as all triplets formed decay nonradiatively (cf. Fig. 21.5). Already in the early years of OLED research, this limit has been considered a roadblock for the success of OLEDs. Nowadays, it is clear that this limitation has been overcome by innovative molecular designs particularly made for OLEDs. The first concept is based on the realization of emitters with efficient and device-compatible phosphorescence (Baldo et al., 1998). This is achieved using organometallic complexes, where central heavy metals like iridium, platinum, and osmium introduce strong spin-orbit coupling (SOC) (Thompson, 2007). This SOC significantly speeds up the radiative phosphorescence rate—outcompeting nonradiative channels—and at the same time, it accelerates the intramolecular intersystem crossing (ISC) from singlet to triplet states (Yersin, 2004). The latter effect leads to the complete population of the material’s triplet state,

Fluorescence

Phosphorescence Electrical excitation Electrical excitation

S1

75%

10 0 IS % C

T1

S1

T1

75%

75%

S1

% 25

ISC RISC

T1

Nonradiative

Energy

Electrical excitation

%

25

%

25

Thermally activated delayed fluorescence

S0

S0 ηint,F = 25%

S0 ηint,P = 100%

ηint,TADF = 100%

Fig. 21.5 Simplified energy diagrams to achieve equal emission energies for the important three emitter types used in OLEDs: fluorescence, phosphorescence, and TADF. S0 represents the molecular ground state, S1 and T1 are the respective first excited singlet and triplet states, and ηint, i represents the internal quantum efficiencies. The different colors indicate the respective energetic ranges, which are excited through the formation of excitons under electrical excitation. Source: SR/IAPP.

706

Handbook of Organic Materials for Electronic and Photonic Devices

which is an effective luminescent state. Consequently, under electrical excitation, this molecular concept leads to internal efficiencies topping at ηint, P ¼ 100% (cf. Fig. 21.5) (Adachi et al., 2001b; Kim et al., 2013b). As an alternative to organometallic phosphorescence, thermally activated delayed fluorescence (TADF) (cf. Fig. 21.5) allows for internal quantum efficiencies (ηint, TADF ¼ 100%) approaching unity as well (Endo et al., 2009; Uoyama et al., 2012). This excitonic concept does not need the complex organometallic chemistry; rather, it unfolds on a specifically optimized molecular design. The essence of TADF is the possibility of triplet states to upconvert to radiative singlet states via reverse ISC (RISC) with the assistance of thermal energy at room temperature (cf. Fig. 21.5). To do so, the energetic splitting between singlet and triplet states must be strongly reduced to be comparable to thermal energy (25 meV). As this is an energetically uphill process, it is slower than the radiative rates of the fluorescence, so TADF is accompanied by a delayed fluorescence component—hence the name—which shows identical spectral distribution compared to the prompt fluorescence. This is realized by constructing intramolecular donor-acceptor (DA) systems that lead to the formation of intramolecular charge-transfer excitons, where, due to the spatial separation, the energetic splitting strongly decreases and often reaches conditions where singlet and triplet configurations are energetically indistinguishable (Wong and Zysman-Colman, 2017). The same effect can be realized by constructing intermolecular interfaces of DA-type molecules, which leads to TADF emissions from the so-called exciplex state (Goushi et al., 2012). It is fair to note that while most of the efforts concentrate on purely organic TADF emitter development, the same excitonic properties can be achieved based on organometallic chemistry (Deaton et al., 2010). The internal limits of conventional fluorescence (ηS/T, F ¼ 25%) can be overcome by utilizing a nonlinear excitonic process. Triplet-triplet annihilation (TTA)—that is, the collision of two triplet excitons—can lead to a contribution of upconverted singlet states. This process can increase the internal quantum efficiency to ηS/T, F+TTA ¼ 62.5% (Kondakov et al., 2009), which is significantly higher than conventional fluorescence, but still far smaller than either phosphorescence or TADF. Fig. 21.5 compares schematically the energy diagram for fluorescent, phosphorescent, and TADF emitters, where the emitting state is always kept at the same energy (singlet state for fluorescence and TADF, triplet state for phosphorescence). The different-shaded boxes indicate the energetic spectra that are excited under electrical excitation. Clearly, the phosphorescent emitter will have to carry the highest energy states (25% by share as a consequence of the spin statistics), which are consequently redirected to the triplet manifold. By raising the triplet energy of a TADF emitter closer to the singlet state, the conventionally lost 75% triplet states can be harvested, but equally important, the maximum energetic load on the emitter does not differ compared to the fluorescent emitter. This is an important point, as the only industrially relevant blue emitters to date are fluorescent. Neither phosphorescence nor TADF has been convincingly reported to realize stable and high-efficiency OLEDs, but clearly based on the previous arguments, TADF seems to have the better energetics at play to realize long-lived blue OLEDs.

Organic light-emitting diodes

707

21.3.3 Light outcoupling The factor influencing the EQE of OLEDs with the largest margin for improvement is the outcoupling efficiency ηout. In very simple words, the thin-film-layered architecture of OLEDs turns them into weak optical cavities. The organic layers have refractive indices in the range of norg  1.6
1.8, which is higher than the respective values of the substrates used (glass or plastic, nsub  1.5), and especially the desired outcoupling medium, air (nair ¼ 1.0). This optical environment leads to a significant confinement of optical modes within the structure. As a rule of thumb, only about 20%–30% of the internally generated photons leave the device to the observer hemisphere in standard device layouts (bottom emission) (Greenham et al., 1994), clearly indicating that there is much room for improvement. The light is generated in a high-refractive-index region and, thus, the differences in refractive indices introduce optical interfaces where total internal reflection (TIR) will occur, trapping part of the light in so-called guided internal modes. To account properly for the complexity of state-of-the-art multilayer OLEDs, the light generated by the emitter molecules within the device is described by classical radiating dipoles (Neyts, 1998; Furno et al., 2012). Then their field propagation and power distribution in the optical structure (the OLED) are modeled by a driven damped oscillator to ultimately calculate the far-field emission pattern after passing the various optical interfaces. Now such models can even account for a combination of thin-film layers and thick optically transparent regions, which the substrates typically are. In contrast to the thin-film layered structure, which is treated as a coherent system, the latter exceeds the coherence length and is described by ray optics (Kovacˇicˇ et al., 2018). Due to the rotational symmetry of the planar cavity structure, the description of the classical radiating dipole model is developed as a function of the in-plane wavevector kk (cf. Fig. 21.6). Often, this vector is used in its normalized representation kk/k, where k is the wave vector within the emission layer. Consequently, kk/k ¼ 1 corresponds to a propagation of the light in the plane of the layered structure. Fig. 21.6 (center, right) shows the optical modes of a green phosphorescent OLED kk nair based on the emitter Ir(ppy)3. For < , light can escape the layered structure, k nemitter which represents the useful portion of the radiated power generated in the OLED. Due to the optical interfaces in the complete OLED stack, different waveguided kk nair nsub modes occur. First, for < < , light is trapped inside the substrate nemitter k nemitter (for nair < nsub < nemitter). With increasing in-plane wave vector spanning kk nsub < < 1, the so-called organic or waveguide modes occur, wherein light is nemitter k trapped in the organic layer stack. For normalized in-plane wave vectors exceeding k 1 (i.e., kk > 1), the light couples to evanescent modes in the form of excitation of metal surface plasmon polariton (SPP) modes with imaginary out-of-plane wave vector kz. It is important to consider the interaction of the emitting dipoles with the various modes available. The respective coupling strength affects the excited-state lifetimes of the emitter molecules. This phenomenon is known as the Purcell effect (Purcell

708

2.6

2.6

2.5

2.5

2.4 550

600 650

2.3 2.2

air

sub

wg

1

100

2.3 2.2 2.1

2.0

2.0

1.9

1.9

10

1

508 nm

eva

2.4

2.1

1.0 0.5 0.0 Normalized intensity

10

10–1

k||/k K (nm–1)

2.7

Energy E (eV)

500

2.7

100

10–1 air

0

5

10 15 20 25 In-plane-wave vector k|| (μm–1)

30

10

–2

10

sub

wg

eva

–2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Normalized in-plane-wave vector k||/k (μm–1)

Fig. 21.6 Calculated power dissipation spectra (center) of a typical bottom-OLED with Ir(ppy)3 as emitter. Here, the power dissipation spectra were weighted with the corresponding PL spectrum of the emitter (left). A cross section (right) at λ ¼ 508 nm shows dissipated power into air, substrate (sub), waveguided modes (wg), and evanescent modes (eva). For determining the outcoupling efficiency the power dissipation spectra is integrated over the wavelength contributions, and then the ratio of outcoupled to total power is calculated. Source: PAW/IAPP.

Handbook of Organic Materials for Electronic and Photonic Devices

Wavelength (nm)

450

Organic light-emitting diodes

709

et al., 1946; Furno et al., 2012; Br€ utting et al., 2013). The complete power dissipation spectrum is a function of the wavelength (cf. Fig. 21.6, center), so that the correct treatment of the OLED optics requires the integration of the power dissipation spectrum and its weighting with the emission spectrum of the emitter (cf. Fig. 21.6, left). Whenever photons are generated at various vertical positions within the OLED stack (e.g., stacked and white OLEDs), their respective positions need to be considered correctly. Based on this optical modeling of OLEDs, the outcoupling efficiency of OLEDs can be computed, and further, the relative power per mode can be determined. This analysis is of key importance for the development of designs for increased light outcoupling. For the green Ir(ppy)3-based device, Fig. 21.7 (top center) shows the simulated EQE in dependence on ETL and HTL thickness. Here, the calculation considers a realistic radiative efficiency of the emitter of ηrad ¼ 0.8 (Kawamura et al., 2005) (i.e., nonradiative losses of the emitter are considered). The four distinct positions with maximal EQE are a consequence of the optical cavity that supports different resonant wavelengths depending on the overall thickness between the reflecting interfaces. So clearly, optimum conditions for a respective emitter spectrum appear and will differ depending on the emission spectrum of the material. Consequently, the thickness optimization for differently emitting OLEDs always needs to be considered carefully. Furthermore, Fig. 21.7 (top right) shows the calculated photon fractions of the different modes as a function of the ETL thickness, while the HTL is set to approximately 43 nm. While the EQE exhibit strong thickness dependence, the coupled power into the substrate stays relatively constant over the ETL thickness. It is interesting to note that the optimal ETL thicknesses for the highest value of the summed contribution of air and substrate modes is increased, which becomes important for the optimal use of external outcoupling structures (Kovacˇicˇ et al., 2018). As expected, the coupling to the waveguide modes increases significantly with increasing thickness, as the cavity can support more and more waveguide modes. On the contrary, the evanescent modes significantly decrease with increasing ETL thickness. Here, the increasing space between emitting dipoles and metal layers strongly suppresses coupling to the SPP modes (Reineke et al., 2009). One effect on OLED optics not discussed so far is the emitter orientation. Recent reports have discovered that many organic material systems are not isotropically oriented in thin films, despite their typical amorphous nature (Yokoyama et al., 2010, 2008; Frischeisen et al., 2010). The orientation of the emitting dipoles has a strong impact on the overall device optics. Here, vertical oriented dipoles couple only weakly to the escape cone of an OLED, whereas dipoles oriented horizontally are the ideal photon sources when embedded in OLED stacks. The mode distribution of the Ir(ppy)3-based OLED is shown in the top panel of Fig. 21.7. Here, the calculation considers a realistic radiative efficiency of the emitter of ηrad ¼ 0.8 (Kawamura et al., 2005) (i.e., nonradiative losses of the emitter are considered) and isotropically distributed dipole sources (orientation factor a ¼ 1/3). The bottom panel of Fig. 21.7 is calculated based on two important changes: First, the nonradiative losses of the emitter have been turned off to demonstrate an upper limit for the mode distribution; and, second, the emitter orientation is set to be preferentially horizontal (a ¼ 0).

20

Nonradiative

80

Metal

10

100

EQE (%)

ETL (nm)

15

Photon fraction (%)

200 150

5

50

710

100

250

Absorption Evanescent

60

Waveguided

40 20

Substrate

ETL

Outcoupled

0

50

100

150

200

250

0

0

0

50

100

HTL (nm) 250

HTL

100

40

ITO

35

20 100

15 10

50

EQE (%)

ETL (nm)

25

50

100

150

HTL (nm)

300

200

250

300

Absorption

Waveguided

60 40

Substrate

20

Outcoupled

5 0

250

80

30

150

0

200

Evanescent

Photon fraction (%)

200

Substrate

150 ETL (nm)

200

250

0

0

0

50

100

150 ETL (nm)

Fig. 21.7 Calculated EQE in dependence on ETL and HTL thickness and optical loss channels for a Ir(ppy)3 bottom-OLED with realistic (top row) and ideal (bottom row) radiative efficiency and emitter orientation. Source: PAW/IAPP.

Handbook of Organic Materials for Electronic and Photonic Devices

0 EML

Organic light-emitting diodes

711

Comparing the outcoupled fractions for both scenarios shows almost a twofold increase for the horizontally oriented emitter, with the radiative efficiency set to unity. This corresponds to about 40% EQE, when the cavity is optimized. Note that the optimal ETL and HTL layer thickness is changed. Experimentally, such an EQE has been demonstrated with a crystalline EML material that ensures a very high degree of horizontal orientation (Kim et al., 2016).

21.4

Applications

Based on the first devices reported, OLEDs were initially seen as a highly promising next-generation display technology. Soon afterward, they were suggested for use in general lighting. While these uses are vastly different, they share a similar device platform to the one previously described. However, the requirements to their emitted color, their luminance during use, and their lateral dimensions strongly differ, so these major research and development efforts are separated these days.

21.4.1 Displays The technology to be replaced by OLEDs in the display sector are liquid crystal displays (LCDs), which have revolutionized the market penetration of displays as a whole and especially paved the way for the development of mobile displays. In contrast to LCDs, the biggest advantage of OLEDs is that they represent a self-emitting pixel technology (i.e., each pixel is generating the light individually). This difference allows for a significant reduction of the thickness (and thus the weight) of the overall display because no backlight is needed. It also supports true back content, as instead of blocking the backlight through cross-polarizers in case of LCDs the respective pixel is turned off. Ultimately, OLEDs are considered to be more power-conservative than LCD solutions. Displays need to reproduce as many of the possible colors that the human eye can capture to allow the most realistic depiction of natural objects. The perception of color can be described by a color chart—namely, the CIE 1931 color space, defined by the Commission International de l’Eclairage (CIE) as shown in Fig. 21.3 for a 2-degree standard observer. In general, the term light refers to electromagnetic radiation with a wavelength within 380–780 nm. Here, monochromatic emission of a single wavelength is defining the outer rim (i.e., a horseshoe) of the diagram. Mixed colors, depending on their spectral distribution, are localized in the space opened up by the monochrome colors. Interestingly, different spectral distributions can result in identical color perceptions, a phenomenon called metamerism. Therefore, display technologies are designed to come as close to reality as possible, where the best compromise jointly for various display technologies is the use of three primary colors: red, green, and blue, opening up a color triangle in the CIE 1931 diagram. Thus, the primary pixel colors define the composition of possible colors by mixing the RGB colors accordingly. Here, the increase of the color triangle formed by red, green, and blue pixels will lead to wider coverage of the color gamut and allow for better overall color

712

Handbook of Organic Materials for Electronic and Photonic Devices

reproduction. In comparison to the NTSC color standard (Livingston, 1954), typical EL spectra of OLEDs are shown in Fig. 21.3, with their respective locations in the CIE 1931 color space. The composition of images from individually addressable primary colors (RGB) is realized by laterally integrating RGB pixels with a high-enough density that ideally cannot be resolved by the observer, who in contrast views a seamless lateral distribution of colors. The power consumption of a display can be decoupled from the primary optimization of display parameters that are associated with the image quality. Here, energy efficiency of a display enters the space of specifications to be optimized whenever the display usage must cope with limited power supply, as is the case in the fields of mobile displays that are part of small, off-grid packages. Ultimately, when for all primary colors, the most efficient OLEDs, operating at internal quantum efficiencies close to unity (cf. Section 21.3), are arranged in a display, OLED displays are more energy efficient than LCDs. However, the benefit in energy efficiency alone is not enough to allow a general transition from LCDs to OLEDs. The self-emitting nature of OLED pixels are important not only for the abovementioned high-contrast achieved using true-black pixels, but also more naturally allow a wide viewing angle because the OLED emission approximately follows a Lambertian intensity distribution (Tanaka et al., 2007). In addition to these two important image quality aspects, the OLED fabrication is compatible with display architectures that support conformable and flexible substrates, which offers various new markets. Here, the OLED as the pixel technology is only a small part of the complete system, where the backplane is by far harder to develop for such flexible displays. The OLED itself is made of comparably soft materials (i.e., organic amorphous molecules), and is also very thin compared to the desired curvature radius of a flexible display, so that it is not representing the limiting factor. The quality of the established LCDs is already so high that OLEDs need to follow current developments in display resolution, while at the same time not compromising the manufacturing yield. Current small mobile displays made from OLEDs at the eye-to-display distances typical for handheld devices are at >500 pixels-per-inch (ppi) (Soneira, 2018), which is beyond the resolution that the human eye can perceive (Banks et al., 1987; Campbell and Green, 1965; Campbell and Gubisch, 1966), in the range of 300–400 ppi. These resolutions are achieved with shadowmask techniques with vacuum-deposited OLEDs. This trend to higher and higher resolutions put an increased focus on the OLED-to-backplane integration, while the OLED stacks become a smaller and smaller part of the whole process and system. This in turn calls for very robust and simplified stack designs. For instance, it is desired for OLED displays to develop RGB subpixels that share common layers (e.g., transport and blocking layers), as the complex lateral structuring into RGB subunits can be reduced to the color-defining layers only. Even higher pixel densities (exceeding 2000 ppi) are needed for virtual reality (VR) and augmented reality (AR) applications (Fujii et al., 2018), as those displays are located much closer to the eye. An alternative concept, used often for large display panels, is the combination of a uniform white OLED architecture that is complemented with RGB color filters with

Organic light-emitting diodes

713

spatial resolution to realize the primary colors. Often, a fourth segment per unit cell is left unfiltered to allow a white pixel, leading to an RGBW display design. Especially on large area panels, this approach has advantages in the process yield.

21.4.2 Solid-state lighting The challenges for OLEDs developed as general light sources (SSL) are quite contrary to the ones in the display sector. General lighting needs white-light emitting sources, so that here, the OLED development needs to find efficient solutions made from individual primary colors. This is because, while OLED emitters produce an EL spectra with large full-width at half-maximum (FHWM) in the range of 50–100 nm (cf. Fig. 21.3) (Reineke et al., 2010), that is not enough to span the visible part of the electromagnetic spectrum. Fig. 21.3 shows two important color compositions (color points E and A in the CIE 1931 diagram) from primary RGB spectra that resemble white light. In contrast to OLED displays, solutions for SSL are single, large-area pixel concepts that do not need lateral structuring. Here, the challenge is the effective scaling to large areas, which is needed to realize white OLEDs with high-enough lumen output to allow their use as light sources. Structuring of OLED systems for SSL can become interesting only whenever the light source should sport the ability to vary its color to accommodate different user settings (Krotkus et al., 2016). The target for high lumen output can be achieved through increases in surface area or surface brightness, which both come with disparate technological hurdles. The increase in active OLED surface increases the physical size of the light source, which face problems in their installment the further the active size deviates from spot lights like lightbulbs. Additionally, large OLEDs need sophisticated solutions to distribute the charge carriers in the electrodes laterally (e.g., by metal grids), as otherwise, the emitting surface would be inhomogeneous and even self-heating effects may arise (Fischer et al., 2014). Increasing the surface brightness on the contrary negatively affects the device stability. The latter scales inversely with the device brightness (cf. Section 21.5.3). In reality, a good compromise between both parameters often needs to be found and is highly dependent on the respective application scenario. What is currently one of the most important challenges for OLEDs for SSL is its price target. While displays are comparably high price sectors, especially in the television market, general lighting does not allow new technologies with a large premium in price. Considering a white LED retrofit lightbulb as a good price reference, it currently has a price of only a few cents per lumen in the retail market ($10–$20 for a 1000-lm lightbulb). In comparison, white OLEDs currently are more expensive by about a factor of 10 (Spindler et al., 2018). Here, the large area-light form factor of OLEDs introduces a secondary problem: With increasing surface area, the substrate scales accordingly, which may add significantly to the overall package price. Beyond the simple cost arguments, the form factor introduces barriers to a seamless market penetration. Whereas inorganic LEDs fit into widely accepted general lighting infrastructure in the form of retrofit bulbs and fluorescent tube form factors, OLEDs provide new opportunities in the SSL sector, especially in emerging niche markets.

714

Handbook of Organic Materials for Electronic and Photonic Devices

21.4.3 Automotive sector An emerging application field for OLEDs is the automotive sector, which suggests that it is an area with a high need for customization of light sources. Here, beyond delivering a high-quality user experience, the differentiation of the individual manufacturers through customized display and lighting solutions is key to future concept developments. There are four main areas, where OLEDs are intensively explored for use in automobiles: dashboard displays, head-up displays, interior ambient lighting, and exterior lights. Currently, many concepts are proposed for future dashboards, which heavily build on customized OLED displays. Here, the ability to produce such displays on conformable substrates offers a virtually seamless integration into the car interior, naturally following the topography of the dashboard design. The slim design of the OLED technology is important to allow the integration of displays on surfaces where the installation depth is limited. As OLEDs can be made transparent (Bulovic et al., 1996; Fries et al., 2017; G€ orrn et al., 2006), they are suggested for use in head-up displays for augmented reality functionality in the driver’s field of view. One important aspect here is the required brightness of such head-up solutions, which is much higher (>10,000 cd/m2) than many other applications to allow a contrast with the daylight background. While reaching these values is not a problem for OLEDs, their long-term stability is greatly reduced as the device’s lifetime scales inversely with its brightness (cf. Section 21.5.3). The use of OLEDs as interior ambient lights is conceivable, but the degree of customization is very high, as the physical dimensions of such lights need to be defined specifically for each use. Current developments of interior design suggest that such solutions need to allow color tunability (Burrows et al., 1996; Kanno et al., 2006; Fr€obel et al., 2015; Krotkus et al., 2016), including the generation of white light with different color temperatures. However, inorganic LEDs are much further developed currently and can be packaged to realize virtually similar effects and functionalities. OLEDs used as exterior lights have seen first demonstrations recently in small series and premium automobiles in form of rear lights (red color) (HELLA GmbH & Co. KGaA, 2018). These rear lights are currently supplemented with LED solutions for brake lights and turn indicators, as the brightness requirements are not met by OLEDs yet. Even with the current performance limitations, OLEDs are very attractive for the various automotive manufacturers, as they allow for a high degree of product differentiation. The automotive sector is very attractive for OLEDs, as it offers gradual market penetration, tolerates a certain price premium—either over existing technologies or because it generates an added value—and represents a large market. For modern automotive lights including LEDs and also OLEDs, replacing the actual light source is no longer an option, because they are highly integrated into the complete light body. Therefore, OLEDs must meet lifetime performance values including immunity toward sudden death exceeding the average lifespan of an automotive. Importantly, the conditions are harsher compared to other usage scenarios (displays or SSL), especially because of the large range of possible temperatures (
20°C to 110°C in the most extreme cases) and various humidity levels.

Organic light-emitting diodes

715

21.4.4 Niche markets In addition to the applications mentioned previously, OLEDs offer unique features that make them attractive for various other fields of use. In particular, the flexibility in the device processing and the overall low-temperature fabrication are crucial aspects. Paired with the material softness (White et al., 2013), many applications in the medical sector become attractive. For instance, OLEDs can be used for light treatment in bandagelike systems (Samuel et al., 2015), where their low physical impact is of key importance for comfortable use on patients. Similarly, sensoric systems, where one or more light sources are needed for a photonic-based readout, can be made effectively with OLEDs (Krujatz et al., 2016; Lochner et al., 2014; Bansal et al., 2014). Furthermore, OLED microdisplays have been used recently to simulate and manipulate cells. Here, OLEDs are investigated as a novel platform for optogenetic research (Steude et al., 2016). Also recently, OLEDs have been developed as an ultralow-cost technology for interactive signage solutions in the areas of print media and retail product packaging (INURU GmbH [WWW Document], 2018), where the overall lifespan specifications are comparably moderate. Here, printing is used as the processing of choice, which will allow for print-on-demand solutions and for lowest possible manufacturing costs. In summary, on the shoulder of giants in the form of display and SSL developments, OLEDs are constantly being explored as light sources in new settings. Here, their use both as complete displays and individual segmented pixels is possible. What is a niche application today soon may become a mainstream solution.

21.5

Current research frontiers

OLED research started with the seminal work of Tang and VanSlyke (1987), demonstrating efficient EL from thin-film, organic-layered systems. The technology has matured since then, and products in the form of small and large displays have entered the broad consumer market. Still, the research field around the OLED technology is very active, where further optimization is needed and seems realistic. Answers to various scientific research questions remain to be found. At the same time, more consolidation of the technology will take place in the industrial sector. In this section, current research efforts will be discussed briefly.

21.5.1 Material development Organic electronics in general is a research field with heavy efforts on material investigation and development, simply because there are so many possible compounds to consider (Reymond et al., 2011). Seen globally, this material freedom reduces the overall pace of progress, as it is a highly parallel effort where various classes of materials are developed independently. Machine learning is nowadays implemented into the material development process, especially for the purpose of accelerating the material screening (Go´mez-Bombarelli et al., 2016). For both mainstream display

716

Handbook of Organic Materials for Electronic and Photonic Devices

applications and SSL, the need for high efficiency and stable blue OLED material systems remains omnipresent. While blue phosphorescence clearly has the potential to realize the highest possible internal quantum efficiencies, the device stability is not sufficient for its full adoption, even after many years of research. One specific material development has recently emerged as an alternative to phosphorescence: TADF. Here, the hope is high that this material concept will allow fulfilling the specifications of high efficiency and stability, especially because the material space is much greater than with organometallic phosphorescence solutions (Wong and Zysman-Colman, 2017). Both material systems are being continuously investigated in parallel (Zhang et al., 2014; Nakanotani et al., 2013) to find a solution to one of the most pressing industrial problems. If future applications call for higher operating brightness values by orders of magnitude (e.g., head-up or microdisplays), device stability becomes a more severe problem (cf. Section 21.5.3). Hence, further material research is likely to be an integral part of future efforts. Currently, all mass-market and most pilot-production OLED systems are made using vacuum deposition techniques because this fabrication is easier to control and allows a higher degree of complexity with respect to the stack architecture. Yet at the same time, the production process as a whole is heavy on resources compared to solution-processing techniques like inkjet printing or other coating solutions. Here, all vacuum-processed devices in the respective applications (displays, signage, or SSL) serve as the performance standard, where the major material development task is to provide performance semiconductor materials that—together with their processing protocol (cf. Section 21.5.4)—allow for matching the current benchmarks. If this is achieved, it also will bring about a significant drop in production costs.

21.5.2 Scalable concepts for light-outcoupling As discussed in Section 21.3.3, the limited share of photons that escape the thin-film structure have great potential for much-needed efficiency improvements. Here, simple arguments always suggest that two to three times the number of improvements are possible by harvesting most of the internal optical modes (cf. Fig. 21.7). However, while high improvements certainly are possible (Reineke et al., 2009; Jeon et al., 2015; Kim et al., 2013a), the respective realistic settings are often neglected, leading to many concepts and results that do not comply with industrial needs. For displays, two aspects are crucial. First, most displays use OLEDs in topemitting configurations, which is a different and much more complex optical problem. For instance, there are no substrate modes that one can utilize, and most of the light is trapped in organic mode (Hofmann et al., 2010; Thomschke et al., 2012). Second, the display technology cannot accept solutions that negatively affect the sharp pixel definition on the microscale, as this would introduce blurry images. Consequently, most of the concepts that use scattering or textured layers are not suited for use in displays. Considering these tight requirements, the twofold-to-threefold potential of OLEDs in concepts for enhanced light outcoupling does not seem to be a realistic target in the display sector.

Organic light-emitting diodes

717

For large-area SSL solutions, the requirements for successful outcoupling concepts are vastly different than for the display sector. OLEDs for large-area lighting are actually bottom-emitting devices. Here, optical solutions can indeed use scattering and textured layers and systems, as the far-field appearance of lighting tiles does not differ, and often light sources are used only in illuminating (i.e., indirect) settings. If lateral resolution is part of the device concept (Krotkus et al., 2016), such blurring optics can even be beneficial, as lateral light inhomogeneities are effectively mixed. What is important for these large-area systems is the scalability of possible concepts for enhanced light outcoupling. Therefore, the effect must be scalable (i.e., the enhancement must be given not only for lab-size OLED pixels, but equally for large lateral dimensions). At the same time, the fabrication for the outcoupling concept must be scalable to large areas, while ideally not significantly affecting the cost of fabrication of the complete package. Section 21.3.3 showed that the emitter orientation can increase the EQE of an OLED significantly. This effect is so powerful because it is a truly intrinsic property of the materials used, so all the requirements for both displays and SSL are met. Clearly for displays, perfectly oriented emitters might prove to be the ideal target for future uses. Currently, understanding of the nanoscale effects leading to the orientation is developed, and—depending on the material and the processes involved—different explanations have been suggested (Mayr and Br€utting, 2015; Jurow et al., 2016; Moon et al., 2017; Gather and Reineke, 2015). What is more important is to transfer this knowledge into concepts that will allow for controlling the orientation to exploit its potential fully. This is a truly nanoscale molecular engineering problem asking for interdisciplinary efforts. Another route to increase the light outcoupling while maintaining the undistorted pixel appearance needed for displays is the incorporation of a low refractive index (Fuchs et al., 2015; Watanabe et al., 2018) or birefringent functional layers (Callens et al., 2015) into the device stack to manipulate the mode distribution in favor for higher extracted modes.

21.5.3 Lifetime and failure The lifetime of light sources has been an issue for users ever since artificial light entered our routines. Here, there are two clear cases: (1) the light intensity, quality, or both reduce over time, and (2) the light source abruptly fails. For OLEDs, this is no different. Fig. 21.8 (left) shows the luminance aging of a green-bottom-OLED comprising Ir(ppy)3. When the device is operated at a constant current density of j ¼ 10 mA/cm2, the luminance drops as a function of time and, often, the driving voltage increases at the same time. The device’s lifetime, defined as the time to reach a certain value from the initial luminance (e.g., LT50 ¼ time to reach 50% of the initial luminance), scales inversely with the device brightness (cf. Fig. 21.8, right) (Meerheim et al., 2006; Scholz et al., 2015). For displays, rather than the LT50 often used in academia, LT97 is the standard. For SSL, LT70 is mostly used by industry. Aging effects of OLEDs have been studied using various approaches, and the sources for improvement are the material development on the one side, and the fabrication process, including packaging of the devices, on the other. The origins of

718 Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 21.8 Lifetime characteristics of a simplified two-layer, green-bottom OLEDs with Ir(ppy)3 as emitting material under extended aging. Under a constant current density of j ¼ 10 mA/cm2, the luminance drops and the voltage increases over time (left). The lifetime (e.g., LT50) scales inversely proportional to the luminance defining application limits (right). Source: PAW/IAPP.

Organic light-emitting diodes

719

device instabilities are highly material and package related, so that generalizations cannot be made. Clearly, the lifetime of OLEDs is the most important device characteristic when it comes to possible use in commercial products. The lifetime specifications need to fit into the targeted application scenario (cf. Fig. 21.8, right), where limited stability currently disregards some potential fields of use.

21.5.4 Processing and integration The materials (small molecules and polymers) and their possible fabrication methods open up a diverse setting for possible upscaling efforts. In the end, materials and processes need to fit perfectly together to obtain the best-performing OLED technology. However, the uncertainty in deciding which path allows the highest performance and profit often complicates the strategy development of leading industry. At the current stage, the processes are mostly vacuum- and small molecule-based, as this combination sets the benchmark for device efficiencies (and, more important, stability). Clearly, the most interesting question will be how much and how fast solutionprocessing techniques that include the necessary materials can close the gap to vacuum technology. Being able to print high performance OLED products will be a game changer.

21.6

Outlook

OLEDs are the most successful functional device category of the diverse electronics based on OSC materials. This success is based on the fact that OLEDs build on a central property to which organic materials are very well suited: highly efficient luminescence. The transport properties needed to realize efficient EL are not demanding compared to other electronic devices like transistors. Conceptually, everything has been demonstrated using this device platform. The main challenge for the future is to overcome limitations in device stability—always seen in relation to the application in question—which can involve vastly different specifications. If this issue can be solved, the investment in broader production capacities will follow, which ultimately will lead to a significant cost reduction. Consequently, OLEDs will become a highquality commodity, unmatched in many fields of use, that will be found everywhere in daily life.

References Adachi, C., Baldo, M.A., Forrest, S.R., Lamansky, S., Thompson, M.E., Kwong, R.C., 2001a. High-efficiency red electrophosphorescence devices. Appl. Phys. Lett. 78, 1622–1624. https://doi.org/10.1063/1.1355007. Adachi, C., Baldo, M.A., Thompson, M.E., Forrest, S.R., 2001b. Nearly 100% internal phosphorescence efficiency in an organic light-emitting device. J. Appl. Phys. 90, 5048–5051. https://doi.org/10.1063/1.1409582.

720

Handbook of Organic Materials for Electronic and Photonic Devices

Amos, R.M., Barnes, W.L., 1997. Modification of the spontaneous emission rate of Eu3+ ions close to a thin metal mirror. Phys. Rev. B 55, 7249–7254. https://doi.org/10.1103/ PhysRevB.55.7249. Baldo, M.A., O’Brien, D.F., You, Y., Shoustikov, A., Sibley, S., Thompson, M.E., Forrest, S.R., 1998. Highly efficient phosphorescent emission from organic electroluminescent devices. Nature 395, 151–154. https://doi.org/10.1038/25954. Banks, M.S., Geisler, W.S., Bennett, P.J., 1987. The physical limits of grating visibility. Vision Res. 27, 1915–1924. https://doi.org/10.1016/0042-6989(87)90057-5. Bansal, A.K., Hou, S., Kulyk, O., Bowman, E.M., Samuel, I.D.W., 2014. Wearable organic optoelectronic sensors for medicine. Adv. Mater. 27, 7638–7644. https://doi.org/ 10.1002/adma.201403560. Br€ utting, W., Frischeisen, J., Schmidt, T.D., Scholz, B.J., Mayr, C., 2013. Device efficiency of organic light-emitting diodes: progress by improved light outcoupling. Phys. Stat. Sol. A 210, 44–65. https://doi.org/10.1002/pssa.201228320. Bulovic, V., Gu, G., Burrows, P.E., Forrest, S.R., Thompson, M.E., 1996. Transparent lightemitting devices. Nature 380, 29. https://doi.org/10.1038/380029a0. Burroughes, J.H., Bradley, D.D.C., Brown, A.R., Marks, R.N., Mackay, K., Friend, R.H., Burns, P.L., Holmes, A.B., 1990. Light-emitting diodes based on conjugated polymers. Nature 347, 539–541. https://doi.org/10.1038/347539a0. Burrows, P.E., Forrest, S.R., Sibley, S.P., Thompson, M.E., 1996. Color-tunable organic lightemitting devices. Appl. Phys. Lett. 69, 2959–2961. https://doi.org/10.1063/1.117743. Burrows, H.D., Seixas de Melo, J., Serpa, C., Arnaut, L.G., Monkman, A.P., Hamblett, I., Navaratnam, S., 2001. S1 >T1 intersystem crossing in π-conjugated organic polymers. J. Chem. Phys. 115, 9601–9606. https://doi.org/10.1063/1.1413969. Callens, M.K., Yokoyama, D., Neyts, K., 2015. Anisotropic materials in OLEDs for high outcoupling efficiency. Opt. Express 23, 21128–21148. https://doi.org/10.1364/ OE.23.021128. Campbell, F.W., Green, D.G., 1965. Optical and retinal factors affecting visual resolution. J. Physiol. 181, 576–593. https://doi.org/10.1113/jphysiol.1965.sp007784. Campbell, F.W., Gubisch, R.W., 1966. Optical quality of the human eye. J. Physiol. 186, 558–578. https://doi.org/10.1113/jphysiol.1966.sp008056. Cocchi, M., Kalinowski, J., Virgili, D., Fattori, V., Develay, S., Williams, J.a.G., 2007. Singledopant organic white electrophosphorescent diodes with very high efficiency and its reduced current density roll-off. Appl. Phys. Lett. 90, 163508. https://doi.org/10.1063/1.2722675. Cocchi, M., Kalinowski, J., Fattori, V., Williams, J.A.G., Murphy, L., 2009. Color-variable highly efficient organic electrophosphorescent diodes manipulating molecular exciton and excimer emissions. Appl. Phys. Lett. 94, 073309. https://doi.org/10.1063/ 1.3086900. Coehoorn, R., van Eersel, H., Bobbert, P., Janssen, R., 2015. Kinetic Monte Carlo study of the sensitivity of OLED efficiency and lifetime to materials parameters. Adv. Funct. Mater. 25, 2024–2037. https://doi.org/10.1002/adfm.201402532. D’Andrade, B.W., Brooks, J., Adamovich, V., Thompson, M.E., Forrest, S.R., 2002. White light emission using triplet excimers in electrophosphorescent organic light-emitting devices. Adv. Mater. 14, 1032–1036. https://doi.org/10.1002/1521-4095(20020805) 14:153.0.CO;2-6. D’Andrade, B.W., Holmes, R.J., Forrest, S.R., 2004. Efficient organic electrophosphorescent white-light-emitting device with a triple doped emissive layer. Adv. Mater. 16, 624–628. https://doi.org/10.1002/adma.200306670.

Organic light-emitting diodes

721

Deaton, J.C., Switalski, S.C., Kondakov, D.Y., Young, R.H., Pawlik, T.D., Giesen, D.J., Harkins, S.B., Miller, A.J.M., Mickenberg, S.F., Peters, J.C., 2010. E-type delayed fluorescence of a phosphine-supported Cu2(μ-NAr2)2 diamond core: harvesting singlet and triplet excitons in OLEDs. J. Am. Chem. Soc. 132, 9499–9508. https://doi.org/ 10.1021/ja1004575. Endo, A., Ogasawara, M., Takahashi, A., Yokoyama, D., Kato, Y., Adachi, C., 2009. Thermally activated delayed fluorescence from Sn4+–porphyrin complexes and their application to organic light emitting diodes—a novel mechanism for electroluminescence. Adv. Mater. 21, 4802–4806. https://doi.org/10.1002/adma.200900983. Erickson, N.C., Holmes, R.J., 2011. Relating charge transport and performance in single-layer graded-composition organic light-emitting devices. J. Appl. Phys. 110, 084515. https://doi. org/10.1063/1.3653285. Erickson, N.C., Holmes, R.J., 2014. Engineering efficiency roll-off in organic light-emitting devices. Adv. Funct. Mater. 24, 6074–6080. https://doi.org/10.1002/adfm.201401009. Fan, C., Zhu, L., Jiang, B., Li, Y., Zhao, F., Ma, D., Qin, J., Yang, C., 2013. High power efficiency yellow phosphorescent OLEDs by using new iridium complexes with halogensubstituted 2-phenylbenzo[d]thiazole ligands. J. Phys. Chem. C 117, 19134–19141. https://doi.org/10.1021/jp406220c. Fischer, A., Koprucki, T., G€artner, K., Tietze, M.L., Br€ uckner, J., L€ ussem, B., Leo, K., Glitzky, A., Scholz, R., 2014. Feel the heat: nonlinear electrothermal feedback in organic LEDs. Adv. Funct. Mater. 24, 3367–3374. https://doi.org/10.1002/adfm.201303066. Fleetham, T., Li, G., Wen, L., Li, J., 2014. Efficient “pure” blue OLEDs employing tetradentate Pt complexes with a narrow spectral bandwidth. Adv. Mater. 26, 7116–7121. https://doi. org/10.1002/adma.201401759. Forrest, S.R., 2004. The path to ubiquitous and low-cost organic electronic appliances on plastic. Nature 428, 911–918. https://doi.org/10.1038/nature02498. Fries, F., Fr€obel, M., Lenk, S., Reineke, S., 2017. Transparent and color-tunable organic lightemitting diodes with highly balanced emission to both sides. Org. Electron. 41, 315–318. https://doi.org/10.1016/j.orgel.2016.11.022. Frischeisen, J., Yokoyama, D., Adachi, C., Br€utting, W., 2010. Determination of molecular dipole orientation in doped fluorescent organic thin films by photoluminescence measurements. Appl. Phys. Lett. 96, 073302. https://doi.org/10.1063/1.3309705. Fr€ obel, M., Schwab, T., Kliem, M., Hofmann, S., Leo, K., Gather, M.C., 2015. Get it white: color-tunable AC/DC OLEDs. Light Sci. Appl. 4, e247. https://doi.org/10.1038/ lsa.2015.20. Fuchs, C., Will, P.-A., Wieczorek, M., Gather, M.C., Hofmann, S., Reineke, S., Leo, K., Scholz, R., 2015. Enhanced light emission from top-emitting organic light-emitting diodes by optimizing surface plasmon polariton losses. Phys. Rev. B. 92, 245306. https://doi.org/ 10.1103/PhysRevB.92.245306. Fujii, T., Kon, C., Motoyama, Y., Shimizu, K., Shimayama, T., Yamazaki, T., Kato, T., Sakai, S., Hashikaki, K., Tanaka, K., Nakano, Y., 2018. 4032 ppi high-resolution OLED microdisplay. SID 2018 DIGEST 46-3, 613–616. Furno, M., Meerheim, R., Hofmann, S., L€ussem, B., Leo, K., 2012. Efficiency and rate of spontaneous emission in organic electroluminescent devices. Phys. Rev. B. 85, 115205. https:// doi.org/10.1103/PhysRevB.85.115205. Gather, M.C., Reineke, S., 2015. Recent advances in light outcoupling from white organic lightemitting diodes. JPE 5, 057607. https://doi.org/10.1117/1.JPE.5.057607. Gohri, V., Hofmann, S., Reineke, S., Rosenow, T., Thomschke, M., Levichkova, M., L€ussem, B., Leo, K., 2011. White top-emitting organic light-emitting diodes employing

722

Handbook of Organic Materials for Electronic and Photonic Devices

a heterostructure of down-conversion layers. Org. Electron. 12, 2126–2130. https://doi. org/10.1016/j.orgel.2011.09.002. Go´mez-Bombarelli, R., Aguilera-Iparraguirre, J., Hirzel, T.D., Duvenaud, D., Maclaurin, D., Blood-Forsythe, M.A., Chae, H.S., Einzinger, M., Ha, D.-G., Wu, T., Markopoulos, G., Jeon, S., Kang, H., Miyazaki, H., Numata, M., Kim, S., Huang, W., Hong, S.I., Baldo, M., Adams, R.P., Aspuru-Guzik, A., 2016. Design of efficient molecular organic light-emitting diodes by a high-throughput virtual screening and experimental approach. Nat. Mater. Adv. 15, 1120–1127. https://doi.org/10.1038/nmat4717. G€ orrn, P., Sander, M., Meyer, J., Kr€oger, M., Becker, E., Johannes, H.-H., Kowalsky, W., Riedl, T., 2006. Towards see-through displays: fully transparent thin-film transistors driving transparent organic light-emitting diodes. Adv. Mater. 18, 738–741. https://doi.org/ 10.1002/adma.200501957. Goushi, K., Kwong, R., Brown, J.J., Sasabe, H., Adachi, C., 2004. Triplet exciton confinement and unconfinement by adjacent hole-transport layers. J. Appl. Phys. 95, 7798–7802. https://doi.org/10.1063/1.1751232. Goushi, K., Yoshida, K., Sato, K., Adachi, C., 2012. Organic light-emitting diodes employing efficient reverse intersystem crossing for triplet-to-singlet state conversion. Nat. Photon 6, 253–258. https://doi.org/10.1038/nphoton.2012.31. Greenham, N.C., Friend, R.H., Bradley, D.D.C., 1994. Angular dependence of the emission from a conjugated polymer light-emitting diode: implications for efficiency calculations. Adv. Mater. 6, 491–494. https://doi.org/10.1002/adma.19940060612. Greiner, M.T., Helander, M.G., Tang, W.-M., Wang, Z.-B., Qiu, J., Lu, Z.-H., 2012. Universal energy-level alignment of molecules on metal oxides. Nat. Mater. 11, 76–81. https://doi. org/10.1038/nmat3159. HELLA GmbH & Co. KGaA, 2018. HELLA lighting concept sets new standards in new Audi A8. Hofmann, S., Thomschke, M., Freitag, P., Furno, M., L€ussem, B., Leo, K., 2010. Top-emitting organic light-emitting diodes: influence of cavity design. Appl. Phys. Lett. 97, 253308. https://doi.org/10.1063/1.3530447. INURU GmbH [WWW Document], 2018. INURU Let there be light. https://www.inuru.de. (Accessed 26 May 2018). Jeon, S., Lee, J.-H., Jeong, J.-H., Song, Y.S., Moon, C.-K., Kim, J.-J., Youn, J.R., 2015. Vacuum nanohole array embedded phosphorescent organic light emitting diodes. Sci. Rep. 5. https:// doi.org/10.1038/srep08685. Jou, J.-H., Hsieh, C.-Y., Tseng, J.-R., Peng, S.-H., Jou, Y.-C., Hong, J.H., Shen, S.-M., Tang, M.-C., Chen, P.-C., Lin, C.-H., 2013. Candle light-style organic light-emitting diodes. Adv. Funct. Mater. 23, 2750–2757. https://doi.org/10.1002/adfm.201203209. Jurow, M.J., Mayr, C., Schmidt, T.D., Lampe, T., Djurovich, P.I., Br€ utting, W., Thompson, M.E., 2016. Understanding and predicting the orientation of heteroleptic phosphors in organic light-emitting materials. Nat. Mater. 15, 85–91. https://doi.org/10.1038/ nmat4428. Kanno, H., Giebink, N.C., Sun, Y., Forrest, S.R., 2006. Stacked white organic light-emitting devices based on a combination of fluorescent and phosphorescent emitters. Appl. Phys. Lett. 89, 023503. https://doi.org/10.1063/1.2219725. Kawamura, Y., Goushi, K., Brooks, J., Brown, J.J., Sasabe, H., Adachi, C., 2005. 100% phosphorescence quantum efficiency of Ir(III) complexes in organic semiconductor films. Appl. Phys. Lett. 86, 071104. https://doi.org/10.1063/1.1862777. Kawamura, Y., Brooks, J., Brown, J.J., Sasabe, H., Adachi, C., 2006. Intermolecular interaction and a concentration-quenching mechanism of phosphorescent Ir(III) complexes in a solid film. Phys. Rev. Lett. 96, 017404. https://doi.org/10.1103/PhysRevLett.96.017404.

Organic light-emitting diodes

723

Kim, J.-B., Lee, J.-H., Moon, C.-K., Kim, S.-Y., Kim, J.-J., 2013a. Highly enhanced light extraction from surface plasmonic loss minimized organic light-emitting diodes. Adv. Mater. 25, 3571–3577. https://doi.org/10.1002/adma.201205233. Kim, S.-Y., Jeong, W.-I., Mayr, C., Park, Y.-S., Kim, K.-H., Lee, J.-H., Moon, C.-K., Br€utting, W., Kim, J.-J., 2013b. Organic light-emitting diodes with 30% external quantum efficiency based on a horizontally oriented emitter. Adv. Funct. Mater. 23, 3896–3900. https://doi.org/10.1002/adfm.201300104. Kim, K.-H., Liao, J.-L., Lee, S.W., Sim, B., Moon, C.-K., Lee, G.-H., Kim, H.J., Chi, Y., Kim, J.-J., 2016. Crystal organic light-emitting diodes with perfectly oriented non-doped Pt-based emitting layer. Adv. Mater. 28, 2526–2532. https://doi.org/10.1002/ adma.201504451. Koch, N., 2007. Organic electronic devices and their functional interfaces. ChemPhysChem 8, 1438–1455. https://doi.org/10.1002/cphc.200700177. Kondakov, D.Y., Pawlik, T.D., Hatwar, T.K., Spindler, J.P., 2009. Triplet annihilation exceeding spin statistical limit in highly efficient fluorescent organic light-emitting diodes. J. Appl. Phys.. 106124510https://doi.org/10.1063/1.3273407. Kotadiya, N.B., Lu, H., Mondal, A., Ie, Y., Andrienko, D., Blom, P.W.M., Wetzelaer, G.-J.A.H., 2018. Universal strategy for Ohmic hole injection into organic semiconductors with high ionization energies. Nat. Mater. 1. https://doi.org/10.1038/s41563-018-0022-8. Kovacˇicˇ, M., Will, P.-A., Lipovsˇek, B., Topicˇ, M., Lenk, S., Reineke, S., Krcˇ, J., 2018. Coupled optical modeling for optimization of organic light-emitting diodes with external outcoupling structures. ACS Photonics 5, 422–430. https://doi.org/10.1021/ acsphotonics.7b00874. Krotkus, S., Kasemann, D., Lenk, S., Leo, K., Reineke, S., 2016. Adjustable white-light emission from a photo-structured micro-OLED array. Light Sci. Appl. 5, e16121. https://doi. org/10.1038/lsa.2016.121. Krujatz, F., Hild, O., Fehse, K., Jahnel, M., Werner, A., Bley, T., 2016. Exploiting the potential of OLED-based photo-organic sensors for biotechnological applications. Chem. Sci. J. 7. https://doi.org/10.4172/2150-3494.1000134. Lee, S.Y., Yasuda, T., Nomura, H., Adachi, C., 2012. High-efficiency organic light-emitting diodes utilizing thermally activated delayed fluorescence from triazine-based donor–acceptor hybrid molecules. Appl. Phys. Lett.. 101093306. https://doi.org/ 10.1063/1.4749285. Lee, S.K., Hwang, D.-H., Jung, B.-J., Cho, N.S., Lee, J., Lee, J.-D., Shim, H.-K., 2015. The fabrication and characterization of single-component polymeric white-light-emitting diodes. Adv. Funct. Mater. 15, 1647–1655. https://doi.org/10.1002/adfm.200500060. Lee, J., Chen, H.-F., Batagoda, T., Coburn, C., Djurovich, P.I., Thompson, M.E., Forrest, S.R., 2016. Deep blue phosphorescent organic light-emitting diodes with very high brightness and efficiency. Nat. Mater. 15, 92–98. https://doi.org/10.1038/nmat4446. Li, G., Fleetham, T., Turner, E., Hang, X.-C., Li, J., 2014. Highly efficient and stable narrowband phosphorescent emitters for OLED applications. Adv. Opt. Mater. 3, 390–397. https://doi.org/10.1002/adom.201400341. Liu, J., Xie, Z.Y., Cheng, Y.X., Geng, Y.H., Wang, L.X., Jing, X.B., Wang, F.S., 2007. Molecular design on highly efficient white electroluminescence from a single-polymer system with simultaneous blue, green, and red emission. Adv. Mater. 19, 531–535. https://doi. org/10.1002/adma.200601580. Liu, J., Jiang, M., Zhou, X., Zhan, C., Bai, J., Xiong, M., Li, F., Liu, Y., 2017. High-efficient sky-blue and green emissive OLEDs based on FIrpic and FIrdfpic. Synth. Met. 234, 111–116. https://doi.org/10.1016/j.synthmet.2017.10.005.

724

Handbook of Organic Materials for Electronic and Photonic Devices

Livingston, D.C., 1954. Colorimetric analysis of the NTSC color television system. Proc. IRE 42, 138–150. https://doi.org/10.1109/JRPROC.1954.274620. Lochner, C.M., Khan, Y., Pierre, A., Arias, A.C., 2014. All-organic optoelectronic sensor for pulse oximetry. Nat. Commun. 5, 5745. https://doi.org/10.1038/ncomms6745. Mayr, C., Br€utting, W., 2015. Control of molecular dye orientation in organic luminescent films by the glass transition temperature of the host material. Chem. Mater. 27, 2759–2762. https://doi.org/10.1021/acs.chemmater.5b00062. Meerheim, R., Walzer, K., Pfeiffer, M., Leo, K., 2006. Ultrastable and efficient red organic light emitting diodes with doped transport layers. Appl. Phys. Lett. 89, 061111. https://doi.org/ 10.1063/1.2268354. Meerheim, R., Nitsche, R., Leo, K., 2008. High-efficiency monochrome organic light emitting diodes employing enhanced microcavities. Appl. Phys. Lett. 93, 043310. https://doi.org/ 10.1063/1.2966784. Mesta, M., Carvelli, M., de Vries, R.J., van Eersel, H., van der Holst, J.J.M., Schober, M., Furno, M., L€ussem, B., Leo, K., Loebl, P., Coehoorn, R., Bobbert, P.A., 2013. Molecular-scale simulation of electroluminescence in a multilayer white organic light-emitting diode. Nat. Mater. 12, 652–658. https://doi.org/10.1038/nmat3622. Moon, C.-K., Kim, K.-H., Kim, J.-J., 2017. Unraveling the orientation of phosphors doped in organic semiconducting layers. Nat. Commun. 8791. https://doi.org/10.1038/s41467-01700804-0. Nakanotani, H., Masui, K., Nishide, J., Shibata, T., Adachi, C., 2013. Promising operational stability of high-efficiency organic light-emitting diodes based on thermally activated delayed fluorescence. Sci. Rep. 3,https://doi.org/10.1038/srep02127 srep02127. Neyts, K.A., 1998. Simulation of light emission from thin-film microcavities. J. Opt. Soc. Am. A 15, 962–971. https://doi.org/10.1364/JOSAA.15.000962. Olthof, S., Meerheim, R., Schober, M., Leo, K., 2009. Energy level alignment at the interfaces in a multilayer organic light-emitting diode structure. Phys. Rev. B. 79. 245308. https://doi. org/10.1103/PhysRevB.79.245308. Purcell, E.M., Torrey, H.C., Pound, R.V., 1946. Resonance absorption by nuclear magnetic moments in a solid. Phys. Rev. 69, 37–38. https://doi.org/10.1103/PhysRev.69.37. Reineke, S., Baldo, M.A., 2012. Recent progress in the understanding of exciton dynamics within phosphorescent OLEDs. Phys. Stat. Sol. A 209, 2341–2353. https://doi.org/ 10.1002/pssa.201228292. Reineke, S., Baldo, M.A., 2014. Room temperature triplet state spectroscopy of organic semiconductors. Sci. Rep. 4. https://doi.org/10.1038/srep03797. Reineke, S., Walzer, K., Leo, K., 2007. Triplet-exciton quenching in organic phosphorescent light-emitting diodes with Ir-based emitters. Phys. Rev. B. 75125328. https://doi.org/ 10.1103/PhysRevB.75.125328. Reineke, S., Lindner, F., Schwartz, G., Seidler, N., Walzer, K., L€ ussem, B., Leo, K., 2009. White organic light-emitting diodes with fluorescent tube efficiency. Nature 459, 234–238. https://doi.org/10.1038/nature08003. Reineke, S., Rosenow, T.C., L€ussem, B., Leo, K., 2010. Improved high-brightness efficiency of phosphorescent organic LEDs comprising emitter molecules with small permanent dipole moments. Adv. Mater. 22, 3189–3193. https://doi.org/10.1002/adma.201000529. Reineke, S., Thomschke, M., L€ussem, B., Leo, K., 2013. White organic light-emitting diodes: status and perspective. Rev. Mod. Phys. 85, 1245–1293. https://doi.org/10.1103/ RevModPhys.85.1245. Reymond, J.-L., Blum, L.C., van Deursen, R., 2011. Exploring the chemical space of known and unknown organic small molecules at www.gdb.unibe.ch. CHIMIA Int. J. Chem. 65, 863–867. https://doi.org/10.2533/chimia.2011.863.

Organic light-emitting diodes

725

Rosenow, T.C., Furno, M., Reineke, S., Olthof, S., L€ussem, B., Leo, K., 2010. Highly efficient white organic light-emitting diodes based on fluorescent blue emitters. J. Appl. Phys. 108, 113113. https://doi.org/10.1063/1.3516481. Samuel, I.D.W., Kulyk, O., McNeill, A., Moseley, H., Ferguson, J., Ibbotson, S., 2015. Ambulatory photodynamic therapy of skin cancer. Photodiagnosis Photodyn. Ther. 12, 331. https://doi.org/10.1016/j.pdpdt.2015.07.029. Schober, M., Olthof, S., Furno, M., L€ussem, B., Leo, K., 2010. Single carrier devices with electrical doped layers for the characterization of charge-carrier transport in organic thin-films. Appl. Phys. Lett. 97, 013303. https://doi.org/10.1063/1.3460528. Scholz, S., Kondakov, D., L€ussem, B., Leo, K., 2015. Degradation mechanisms and reactions in organic light-emitting devices. Chem. Rev. 115, 8449–8503. https://doi.org/10.1021/ cr400704v. Segal, M., Baldo, M.A., Holmes, R.J., Forrest, S.R., Soos, Z.G., 2003. Excitonic singlet-triplet ratios in molecular and polymeric organic materials. Phys. Rev. B. 68075211. https://doi. org/10.1103/PhysRevB.68.075211. Soneira, R.M., 2018. Galaxy S9 OLED Display Technology Shoot-Out [WWW Document]. http://www.displaymate.com/Galaxy_S9_ShootOut_1s.htm. (Accessed 26 May 2018). Spindler, J., Kondakova, M., Boroson, M., B€uchel, M., Eser, J., Knipping, J., 2018. Advances in high efficacy and flexible OLED lighting. SID 2018 DIGEST 84-1, 1135–1138. Springer, R., Kang, B.Y., Lampande, R., Ahn, D.H., Lenk, S., Reineke, S., Kwon, J.H., 2016. Cool white light-emitting three stack OLED structures for AMOLED display applications. Opt. Express 24, 28131–28142. https://doi.org/10.1364/OE.24.028131. Steude, A., Witts, E.C., Miles, G.B., Gather, M.C., 2016. Arrays of microscopic organic LEDs for high-resolution optogenetics. Sci. Adv. 2, e1600061. https://doi.org/10.1126/ sciadv.1600061. Su, S.-J., Gonmori, E., Sasabe, H., Kido, J., 2008. Highly efficient organic blue-and white-lightemitting devices having a carrier- and exciton-confining structure for reduced efficiency roll-off. Adv. Mater. 20, 4189–4194. https://doi.org/10.1002/adma.200801375. Tanaka, D., Sasabe, H., Li, Y.-J., Su, S.-J., Takeda, T., Kido, J., 2007. Ultra high efficiency green organic light-emitting devices. Jpn. J. Appl. Phys. 46L10. https://doi.org/10.1143/ JJAP.46.L10. Tang, C.W., VanSlyke, S.A., 1987. Organic electroluminescent diodes. Appl. Phys. Lett. 51, 913–915. https://doi.org/10.1063/1.98799. Thompson, M., 2007. The evolution of organometallic complexes in organic light-emitting devices. MRS Bull. 32, 694–701. https://doi.org/10.1557/mrs2007.144. Thomschke, M., Reineke, S., L€ussem, B., Leo, K., 2012. Highly efficient white top-emitting organic light-emitting diodes comprising laminated microlens films. Nano Lett. 12, 424–428. https://doi.org/10.1021/nl203743p. Tu, G., Zhou, Q., Cheng, Y., Wang, L., Ma, D., Jing, X., Wang, F., 2004. White electroluminescence from polyfluorene chemically doped with 1,8-napthalimide moieties. Appl. Phys. Lett. 85, 2172–2174. https://doi.org/10.1063/1.1793356. Turro, N.J., Ramamurthy, V., Scaiano, J.C., 2010. Modern Molecular Photochemistry of Organic Molecules. University Science Books, Sausalito. Uoyama, H., Goushi, K., Shizu, K., Nomura, H., Adachi, C., 2012. Highly efficient organic light-emitting diodes from delayed fluorescence. Nature 492, 234–238. https://doi.org/ 10.1038/nature11687. Walzer, K., Maennig, B., Pfeiffer, M., Leo, K., 2007. Highly efficient organic devices based on electrically doped transport layers. Chem. Rev. 107, 1233–1271. https://doi.org/10.1021/ cr050156n.

726

Handbook of Organic Materials for Electronic and Photonic Devices

Watanabe, T., Yamaoka, R., Ohsawa, N., Tomida, A., Seo, S., Yamazaki, S., 2018. Extremely high-efficient OLED achieving external quantum efficiency over 40% by carrier injection layer with super-low refractive index. SID 2018 DIGEST 26-2, 332–335. White, M.S., Kaltenbrunner, M., Głowacki, E.D., Gutnichenko, K., Kettlgruber, G., Graz, I., Aazou, S., Ulbricht, C., Egbe, D.A.M., Miron, M.C., Major, Z., Scharber, M.C., Sekitani, T., Someya, T., Bauer, S., Sariciftci, N.S., 2013. Ultrathin, highly flexible and stretchable PLEDs. Nat. Photon 7, 811–816. https://doi.org/10.1038/nphoton.2013.188. Wong, M.Y., Zysman-Colman, E., 2017. Purely organic thermally activated delayed fluorescence materials for organic light-emitting diodes. Adv. Mater. 29. https://doi.org/ 10.1002/adma.201605444. Wu, J.-H., Chi, C.-A., Chiang, C.-L., Chen, G.-Y., Lin, Y.-P., Chen, C.-C., Ho, S.-Y., Chen, S.P., Li, J.-Y., 2016. Dimmable sunlight-like organic light emitting diodes with ultra-high color rendering index. Opt. Mater. 55, 90–93. https://doi.org/10.1016/j.optmat. 2016.03.027. W€ urfel, P., 2005. Physics of Solar Cells: From Principles to New Concepts. WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim. Yersin, H., 2004. Triplet emitters for OLED applications. Mechanisms of exciton trapping and control of emission properties. In: Transition Metal and Rare Earth Compounds, Topics in Current Chemistry. Springer, Berlin, Heidelberg, pp. 1–26. https://doi.org/10.1007/ b96858. Yokoyama, D., Sakaguchi, A., Suzuki, M., Adachi, C., 2008. Horizontal molecular orientation in vacuum-deposited organic amorphous films of hole and electron transport materials. Appl. Phys. Lett. 93, 173302. https://doi.org/10.1063/1.2996258. Yokoyama, D., Setoguchi, Y., Sakaguchi, A., Suzuki, M., Adachi, C., 2010. Orientation control of linear-shaped molecules in vacuum-deposited organic amorphous films and its effect on carrier mobilities. Adv. Funct. Mater. 20, 386–391. https://doi.org/10.1002/ adfm.200901684. Zhang, Y., Lee, J., Forrest, S.R., 2014. Tenfold increase in the lifetime of blue phosphorescent organic light-emitting diodes. Nat. Commun. 5. https://doi.org/10.1038/ncomms6008.

Materials and physics of light-emitting electrochemical cells (LECs)

22

Shiyu Hu, Jun Gao Department of Physics, Engineering Physics and Astronomy, Queen’s University, Kingston, ON, Canada

22.1

Introduction

Organic semiconductors (OSCs) in their many forms have proved to be as valuable and versatile as their inorganic counterparts. After three decades of intensive research and development by academics and industries alike, organic photonic devices such as organic light-emitting diodes (OLEDs) have achieved great commercial success. Both small molecule–based OLEDs (SM-OLEDs) and polymer-based OLEDs (P-OLEDs) are poised to replace liquid crystal displays (LCDs) as the dominant flat-panel-display technology. Composed largely of hydrocarbons, OSCs are easily chemically tailored and compatible with low-cost processing techniques. The promise of low-cost, designer semiconductors also motivates the development of other organic devices, such as solar cells and thin-film transistors. OLEDs in particular, possess not only potential cost advantages, but also superior performance and a mechanical flexibility that is unrivaled by their inorganic or LCD counterparts. A key difference between inorganic LEDs and OLEDs is that the former are based on doped semiconductors, whereas the latter are typically undoped. In inorganic devices, chemical doping serves two main functions: (1) to form a homojunction or heterojunction by which to facilitate efficient junction electroluminescence (EL) via band-gap engineering and (2) to realize ohmic contact with the metal contact points via degenerate doping. While chemical doping also has been applied to OSCs, it is the light-emitting electrochemical cell (LEC) that epitomizes the concept of doping that makes it a unique organic EL device and an analog of an inorganic p-n junction LED. LECs now employ not only organic polymers, but also small molecules and ionic transition metal complexes. In this chapter, we briefly introduce the history, device characteristics, operating mechanism, and materials of LECs. A new type of LECs employing bipolar electrodes will also be introduced.

Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00022-X © 2019 Elsevier Ltd. All rights reserved.

728

Handbook of Organic Materials for Electronic and Photonic Devices

22.1.1 Organic light-emitting diodes (OLEDs) OLEDs are injection EL devices that date back to the 1960s, when visible light was first observed in anthracene single crystals under a large DC voltage bias (Pope et al., 1963). In the ensuing decade, research in organic EL mainly involved molecular crystals of anthracene (Helfrich and Schneider, 1965; Helfrich and Wg, 1966), tetracene (Kalinowski et al., 1976), pyrene (Gonzalezbasurto and Burshtein, 1975), and naphthalene (Lohmann and Mehl, 1969). These early SM-OLEDs, however, are plagued by low efficiencies (if reported at all) and very high driving voltages of 100 V or more. The inferior performance is mainly due to large specimen thicknesses of tens of microns or even millimeters, as well as inefficient and impractical liquid electrolyte electrodes. In the late 1970s and early 1980s, SM-OLEDs made with evaporated octaethylporphin or anthracene thin films of only several hundred nanometers emerged (Kampas and Gouterman, 1977; Vincett et al., 1982). When driven with evaporated silver/aluminum or gold/aluminum electrodes, EL is observed with a bias voltage of only 12–15 V. The efficiencies of these early thin-film SM-OLEDs, however, remained very low (e.g., 0.03%–0.06%). A breakthrough finally came when a bilayer SM-OLED was invented by Eastman Kodak in 1987 (Tang and Vanslyke, 1987). The Kodak OLED consisted of an ultrathin (60-nm), vapor-deposited layer of 8-hydroxyquinoline aluminum (Alq3) as the emitting layer and an 75-nm aromatic diamine layer as the hole transport layer (HTL). The bilayer was sandwiched between an indium tin oxide (ITO) anode and an Mg:Ag alloy cathode. Due to the mismatching molecular orbitals of the emitting layer and the hole-transport layer, the injected electrons and holes were effectively trapped at the molecular heterojunction, greatly improving the likelihood of recombination. The Kodak OLED exhibits high efficiency (1.5 lm/W or 1% ph/el) and record brightness (>1000 cd/m2) at a driving voltage below 10 V. This revolutionary Kodak bilayer OLED provided the impetus for three decades of intensive research and development that eventually led to the successful commercialization of the OLED display technology. The huge success of the Kodak OLED can be attributed to band-gap engineering, a technique widely used in conventional semiconductor devices to control the spatial distribution of electronic charge carriers. Nowadays, nearly all OLEDs made with vapor-deposited small molecules adopt a multilayer, band-gap engineered design. Fig. 22.1 shows a state-of-the-art SM-OLED with deep blue emissions (Lee et al., 2016). The OLED exhibits brightness exceeding 7800 cd/m2 and an external quantum efficiency higher than 10%. The multilayer stack consists of a hole injection layer (HIL), a HTL, an electron-blocking layer (EBL), an emission layer (EML), a holeblocking layer (HBL) and an electron transport layer (ETL). The various blocking and transport layers ensure effective injection of electrons and holes, as well as their subsequent transport to the EML, where they recombine radiatively. The iridium complex is a phosphorescence emitter that allows for highly efficient emissions by harvesting the energy of both singlet and triplet excitons. OLEDs made with a phosphorescent light emitter are often abbreviated as PHOLEDs. OLEDs can be made from not only vapor-evaporated small molecules, but also polymers or plastics. The appeal of a polymer-based OLED is immediate when the

Materials and physics of light-emitting electrochemical cells (LECs)

729

Fig. 22.1 An example of a multilayer OLED structure showing the active layer stack. HOMO and LUMO energies of comprising materials (in electrovolts) are shown at the top and bottom of each layer. HIL, hole injection layer; HTL, hole blocking layer; EBL, electron/exciton blocking layer; EML, emissive layer; HBL, hole/exciton blocking layer; ETL, electron-transport layer; the corresponding number below each abbreviation is the thickness of the layer in nanometers. CzSi stands for 9-(4-tert-butylphenyl)-3,6-bis(triphenylsilyl)-9H-carbazole, and TPBi stands for 1,3,5-tris(1-phenyl-1H-benzimidazol-2-yl)benzene. The EML consists of a phosphorescent emitter fac-Ir(pmp)3 doped into a host material TSPO1. Reproduced with permission from Lee, J., Chen, H.F., Batagoda, T., Coburn, C., Djurovich, P.I., Thompson, M.E., Forrest, S.R., 2016. Deep blue phosphorescent organic light-emitting diodes with very high brightness and efficiency. Nat. Mater. 15, 92–98.

light-emitting polymer (LEP) is also solution-processable and mechanically flexible. The first P-OLED with a polymeric emitter was demonstrated in 1990 at Cambridge University (Burroughes et al., 1990). The LEP used was regular poly(p-phenylene vinylene) (PPV), an intractable conjugated polymer thermally converted from its solution-processable precursor. The Cambridge P-OLED was a simple sandwich with an ultrathin PPV layer contacted by an indium oxide anode and an aluminum cathode. The P-OLED turned on at about 14 V, but had a low external quantum efficiency (EQE) of only 0.05%. Soon after, an orange/red-emitting P-OLED was demonstrated at the University of California, Santa Barbara (UCSB), using a modified PPV as emitter (Braun and Heeger, 1991). The LEP of UCSB P-OLED was poly(2-methoxy, 5-(20 -ethylhexoxy)-1, 4-phenylene vinylene) (MEH-PPV), a designer compound with excellent solubility in common organic solvents due to its flexible alkoxy side chains. The addition of the alkoxy groups also changed the energy gap of PPV. Critically, the UCSB P-OLED uses calcium as the cathode material. This led to an EQE of about 1%, a 20-fold improvement over the Cambridge P-OLED. EL was visible in a lighted room at a forward bias of only 4 V.

730

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 22.2 Band diagram of a polymer LED in forward bias for various cathode materials. Reproduced with permission from Parker, I.D., 1994. Carrier tunneling and device characteristics in polymer lightemitting-diodes. J. Appl. Phys. 75, 1656–1666 with permission. Copyright (1994) American Institute of Physics.

2.8 eV

4.9 eV

Ca 2.9 eV In AI Ag Cu

4.2 eV 4.3 eV 4.6 eV 4.7 eV

Au 5.2 eV

Polymer layer

ITO 4.7 eV

40 Å

Field 5 ¬ 107V/m

0.2 eV

Fig. 22.2 explains why the cathode work function plays such an important role in determining P-OLED efficiency (Parker, 1994). In the ITO/MEH-PPV/metal P-OLED shown, the work function of ITO anode is sufficiently high that its Fermi level is close to the highest occupied molecular orbital (HOMO) of MEH-PPV. Therefore, the injection of holes is not severely impeded on the anode side. On the cathode side, the work functions of the various metals shown differ greatly. With a high work function cathode such as aluminum, the large energy barrier between the metal Fermi level and the MEH-PPV lowest unoccupied molecular orbital (LUMO) prevents efficient electron injection. Calcium, on the other hand, is a good electron injector due to its low work function of only 2.9 eV. The balanced charge injection of ITO/MEHPPV/Ca P-OLED is responsible for its superior performance. It is interesting to note that the UCSB P-OLED has a similar EQE as the Kodak SM-OLED despite the former’s lack of any charge transport layer. Although an HIL such as polystyrene sulfonate (PEDOT:PSS), a doped conducting polymer, is often used, P-OLEDs generally have a much simpler structure than SM-OLEDs. The latter are based on successively evaporated layers of organic small molecules. The simpler structure of P-OLEDs makes them compatible with low-cost processing techniques such as spin coating, screen printing and ink-jet printing (Lee et al., 2017; Kajii et al., 2016; Ha et al., 2015; Takeshita et al., 2005).

22.1.2 Light-emitting electrochemical cells (LECs) The preceding brief overview of SM-OLEDs and P-OLEDs paints a picture of remarkable evolution in both materials and device engineering. From a scientific curiosity in the early days, OLEDs have emerged to be the leading candidate for next-generation lighting and display technologies. Major breakthroughs, such as the Kodak

Materials and physics of light-emitting electrochemical cells (LECs)

731

SM-OLEDs and the Cambridge and UCSB P-OLEDs, overcame the drawbacks of early OLEDs by adopting thinner active layers and more efficient injecting electrodes. While state-of-the-art OLEDs often contain multiple layers and carefully tuned interfaces akin to an inorganic double heterojunction LED in design, OLEDs are fundamentally different from inorganic LEDs. The prototypical inorganic LED, as well as many other inorganic diodes, is a p-n homojunction based on doped semiconductors. In a p-n junction LED, EL occurs as a result of minority carrier injection across the junction. The injected minority carriers, once across the junction depletion region, are effectively annihilated by the many majority carriers available within a short distance (a diffusion length). At the metal/semiconductor interface, ohmic contacts are realized via heavy doping rather than energy level matching. Heavy doping significantly reduces the width of the Schottky barrier, allowing an easy injection of change carriers via quantum mechanical tunnelling. An LEC is the organic analog of an inorganic p-n junction LED. The first LEC was demonstrated by Dupont Displays (formerly Uniax Corp., in Santa Barbara, California) in 1995 (Pei et al., 1995). The first LECs are polymer LECs (PLECs) that use the same LEPs from Cambridge and UCSB P-OLEDs. The active layer of a PLEC also contains an ion solvating/transport polymer, poly(ethylene oxide) (PEO), and a lithium salt, which together form a solid ionic conductor known as a polymer electrolyte. The presence of mobile ions in a PLEC makes it fundamentally different from a P-OLED. When a PLEC is operated with a sufficiently large DC voltage or current, injection of electronic charges and their subsequent interaction with mobile counter ions cause the in situ electrochemical p- and n-doping of the LEP. The doping initially occurs at the electrode/polymer interfaces; p-doping occurs on the anode side and n-doping occurs on the cathode side. With time, the doping fronts propagate and eventually make contact to form a p-n or p-i-n junction. Here, EL is generated through the recombination of electrons and holes within the vicinity of the formed junction. The PLEC, therefore, is a polymer homojunction device based on doped polymer semiconductors. Like OLEDs, LECs also have several varieties, as distinguished by the types of emitters used. In addition to PLECs, LECs have been demonstrated with small molecules (Shanmugasundaram et al., 2016; Lin et al., 2016; Wong et al., 2015; Hill et al., 2008), ionic transition metal complexes (iTMCs) (Costa et al., 2012; Ertl et al., 2017), host/guest systems and quantum dots (QDs) (Pertega´s et al., 2016; Tang et al., 2015; Nishikitani et al., 2017; Liao et al., 2012; Frohleiks et al., 2017; Ayguler et al., 2015; Zhang et al., 2015). Despite differences in materials, all LECs share some common features and characteristics. LECs have a simple device structure that typically consists of an emitting layer and a PEDOT:PSS HIL. Like P-OLEDs, LECs are highly compatible with low-cost solution-processing techniques. Unlike OLEDs, however, LECs do not require highly reactive, low-work-function cathodes and are very tolerant of variations in the thickness of the active layer. The very first PLECs, for example, exhibit a very low EL turn-on voltage (VEL-on) on the order of Eg/e, where Eg is the energy gap of the LEP and e the elementary charge. Under a 4 V driving voltage, the devices reached a luminance of about 200 cd/m2 and an EQE close to 2%, despite a single-layer construction and the use of an Al cathode. Although LECs cannot yet

732

Handbook of Organic Materials for Electronic and Photonic Devices

compete with OLEDs in terms of long-term stability, response time, and maximum luminance, the highly unique and desirable characteristics of LECs have motivated an ever-increasing effort to explore the fundamental science and potential application of these devices.

22.2

The physics of LECs

22.2.1 Device characteristics LECs can be easily identified by their distinctive device characteristics, despite their similar appearance to OLEDs, which are made with the same emitters. Fig. 22.3A displays the current-voltage-light intensity (I-V-L) characteristics of an ITO/MEH-PPV: PEO:LiCF3SO3/Al sandwich PLEC (Pei et al., 1996). The voltage bias was scanned from 4 V to 4 V. Following the conventions of OLEDs, the bias voltage is defined as positive (or forward) when the electrode with the higher work function (i.e., ITO) is biased positive relative to one with the lower work function (i.e., Al). Unlike an OLED or a conventional p-n diode, however, the PLEC does not exhibit any diode behavior. In fact, the PLEC is equally conductive under both forward and reverse bias. More strikingly, the PLEC emits strongly regardless of the bias polarity, exhibiting a very low VEL-on of about  2.5 V and reaching over 100 cd/m2 in brightness at less than  4V. Fig. 22.3B shows the time evolution of cell current and EL intensity of a planar PLEC made with identical Al electrodes separated by a distance of 0.6 mm (Zhang and Gao, 2006). Both cell current and EL did not stay constant, despite a constant applied bias voltage. In less than 25 min, the current increased by nearly four orders of magnitude. EL, after a delay of about 2 min, also increased by several orders of magnitude. Afterward, both the cell current and the EL decreased as the cell was cooled from 315 K to 200 K. The application of a constant bias voltage (or more commonly, a bias current to an LEC) activates the LEC from an OFF state to an ON state that is far more conductive. The large variation in cell current and luminance under a fixed bias voltage shows a significant change in cell resistance. But why is the resistance of a PLEC changing so drastically under a constant voltage bias? What is the cause of delay in the appearance of EL? If LECs are p-n junction diodes, as mentioned earlier, why do they not exhibit significant diode rectification? The answer to these questions lies with the very doping process that dominates the LEC processes and characteristics. Here, it should be emphasized that LECs are true p-n junctions, except that the LEC p-n junction is formed in situ and is not permanent. The diode-like I-V-L characteristics of an LEC can be revealed by two methods. Fig. 22.4A shows the I-V-L characteristics of an ITO/MEH-PPV:PEO:LiCF3SO3/Al sandwich PLEC that had been cooled to 100 K after being activated with either a 4 V or a 3 V bias (Gao et al., 1997). In both cases, the cooled PLEC exhibits current rectification and emits only when biased in the same polarity as that of the activation bias voltage. The PLECs are called frozen-junction LECs because the p-n junction formed are immobilized upon cooling. Because the light-emitting junction is already formed and immobilized, frozen-junction LECs exhibit submicronsecond response times, as shown in Fig. 22.4B.

Materials and physics of light-emitting electrochemical cells (LECs)

733

100

40 30

50

10 0

0 –10

Light intensity (cd/m2)

Current (mA)

20

–50

–20 –30 –40 –3

–2

103

–1

0 1 Bias (V)

2

315K

3

315–200K

102

Current (µA)

4

101

100

101 10–1

Current

100

EL intensity

10–2

10–1

10–2

(B)

EL intensity (a.u.)

(A)

–100 –4

0

500

1000 Time (sec)

1500

10–3 2000

Fig. 22.3 (A) Current (I) and light output (L) versus voltage (V) for an ITO/MEH-PPV:PEO: LiTf/Al sandwich LEC with the ITO contact wired as the anode. (B) Current and light output as a function of time for an Al/MEH-PPV:PEO:LiTf/Al planar PLEC biased at 150 V. The LEC was cooled from 315 K to 200 K after activation. (A) Reproduced with permission from Pei, Q.B., Yang, Y., Yu, G., Zhang, C., Heeger, A.J., 1996. Polymer light-emitting electrochemical cells: in situ formation of a light-emitting p-n junction. J. Am. Chem. Soc. 118, 3922–3929 with permission. Copyright (1996) American Chemical Society. (B) Reproduced with permission from Zhang, Y., Gao, J., 2006. Lifetime study of polymer light-emitting electrochemical cells. J. Appl. Phys. 100, 084501 with permission. Copyright (2006) American Institute of Physics.

734

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 22.4 (A) Current and light versus voltage (I-V-L) data measured at 100 K after cooling under 4 V bias upper curves. The LEC was subsequently heated to 300 K without external bias, then biased at 3 V and cooled after reaching steady state to 100 K again (lower curve). (B) Temporal response of the light emission of the frozen-junction LEC at 100 K (upper panel) and of the dynamic junction LEC at 300 K (lower panel). Reproduced with permission from Gao, J., Yu, G., Heeger, A.J., 1997. Polymer light-emitting electrochemical cells with frozen p-i-n junction. Appl. Phys. Lett. 71, 1293–1295 with permission. Copyright (1997) American Institute of Physics.

The second method involves a remarkable LEC device configuration (shown in Fig. 22.5A). Two planar PLECs sharing a common Al electrode were biased with an 800-V bias, generating visible EL from narrow regions between the electrodes (Gao and Dane, 2003). The two PLECs used different LEPs (hence the different emission colors). The combined interelectrode gap of the two LECs was 3 mm

Materials and physics of light-emitting electrochemical cells (LECs)

735

Fig. 22.5 (A) Photograph of two working 1.5 mm planar PLECs in series under 800 V. The inset shows the device configuration and voltage bias applied. (B) I-V-L characteristics of a 1 mm MEH-PPV planar PLEC after being turned on at 100 V. The inset shows forward I-V-L characteristics of a 1.5 mm MEH-PPV planar PLEC after being turned on at 400 V. Solid circles represent current; open circles represent light intensity. Reproduced with permission from Gao, J., Dane, J., 2003. Planar polymer light-emitting electrochemical cells with extremely large interelectrode spacing. Appl. Phys. Lett. 83, 3027–3029 with permission. Copyright (2003) American Institute of Physics.

in width—100 times larger than any planar LECs demonstrated previously. Fig. 22.5B shows the I-V-L scan of a similar planar PLECs after being activated with a 100 V or 400 V bias. The scan was performed at a rate of 20 V/s, which was fast enough so that the p-n junction that formed did not have time to dissipate. The activated, planar PLEC displays the unipolar conduction and light emission expected for a p-n junction.

736

Handbook of Organic Materials for Electronic and Photonic Devices

It is remarkable that the planar PLEC had a completely symmetric structure. Therefore, the strong asymmetry in the I-V-L curves of about zero bias can only be caused by the p-n junctions formed. The extremely large planar PLECs, first demonstrated in 2003, also led to the visualization of the PLEC doping process and many important discoveries, some of which will be discussed in the following sections.

22.2.2 Operating mechanism In explaining the LEC characteristics shown in Fig. 22.3, the creators of the original LECs put forth electrochemical doping (ECD) and p-n junction formation as constituting the fundamental operating mechanism of LECs (Pei et al., 1995, 1996). Fig. 22.6 shows a schematic representation of the LEC operating mechanism according to the ECD model (Gao, 2018). The LEC active layer is a mixed conductor containing both an LEP as the semiconductor and a polymer electrolyte as an ionic conductor and salt, as shown in Fig. 22.6A. The application of a sufficient DC voltage bias causes the injection of electrons and holes at the negative and positive electrodes, respectively. When the injected electronic charges are compensated for by the available mobile ions, the LEP is electrochemically doped in situ (Fig. 22.6B). In PLECs, ECD is a redox reaction of the LEP in the presence of mobile ions. This is shown in the following scheme with PPV and solvated LiCF3SO3 salt as an example. In p-doping, the PPV repeat units are oxidized with the loss of electrons. This is followed by the insertion of anions in the vicinity of the oxidation sites. In n-doping, the PPV repeat units are reduced with the injection of electrons and neutralized with the insertion of cations. Both p- and n-doped polymers are electrically neutral but contain extra electronic charge carriers balanced by the counter-ions. Once doped, the LEP becomes conductive and the doping fronts propagate, causing the doped regions to expand, as shown in Fig. 22.6C. Eventually, the p- and n-doping meet to form a p-n junction between the doped regions. EL occurs in the junction region where the injected electrons and holes recombine radiatively, as shown in Fig. 22.6D. The electrochemical p- and n-doping of PPV in the presence of mobile CF3SO3 and K+ ions are illustrated below: p  doping : n  doping :

h
 i m+  ðPV Þn + ðnmÞCF3 SO Þ CF3 SO 3  ðnmÞe $ ðPV 3 m n   ðPV Þn + ðnmÞK + + ðnmÞe $ ðPV m + ÞðK + Þm n

The ECD model explains the unique LEC device characteristics shown in Fig. 22.3. First, the initial injection of electronic charges, which is necessary for in situ ECD, is made easy by the presence of mobile ions. This allows the doping to initiate regardless of the polarity of the applied voltage bias. Once the polymer is doped, charge injection at the metal/polymer interface becomes ohmic, shifting the doping front to the interface between doped and undoped polymers. This explains why the LEC can be activated under either a forward or a reverse bias, as shown in Fig. 22.3A. Second, there is a delay between the application of a voltage bias and the onset of EL that corresponds to the time it takes for the p-n junction to form. This time delay is shown in

Materials and physics of light-emitting electrochemical cells (LECs)

737

Fig. 22.6 Schematic illustration of PLEC device structure and operation. (A) Pristine PLEC without any voltage bias; mobile ions are uniformly distributed throughout the active layer. (B) PLEC under an applied voltage bias; in situ ECD first takes place at the anode and cathode interfaces. (C) The expansion of the p- and n-doped regions and the propagation of the doping fronts. (D) The formation of a light-emitting p–n junction; the PLEC is functional as a lightemitting device. (E) The schematic of a sandwich PLEC. Reproduced with permission from Gao, J., 2018. Polymer light-emitting electrochemical cellsrecent advances and future trends. Curr. Opin. Electrochem. 7, 87–94 with permission. Copyright (2018) Elsevier.

Fig. 22.3B. Once the light-emitting junction is formed, the overall doping level of the LEC continues to increase, causing both cell current and EL intensity to increase. This behavior has also been displayed in Fig. 22.3B. The ECD model is also consistent with many other early experimental results. This includes the observation of a narrow emission zone in planar PLECs and the change in optical absorption and photoluminescence (PL) upon activation (Dick et al., 1996). Once activated, PLECs can be discharged to exhibit a significant open-circuit voltage or short-circuit current over an extended period (Pei et al., 1996; AlTal and Gao, 2016a). These behaviors are drastically different from OLEDs. Although alternative LEC models exist (deMello et al., 1998; Slinker et al., 2007), the demonstration of extremely large planar PLECs in 2003 offered the strongest support yet to the ECD model (Gao and Dane, 2003, 2004).

738

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 22.7 shows the time-lapse PL imaging of a 10.4-mm planar PLEC activated with a 400-V bias (Hu and Gao, 2011). The PLEC had a structure of Au/MEH-PPV: PEO:CsClO4/Al. Under ultraviolet (UV) illumination, the polymer film displayed the characteristic orange-red PL of MEH-PPV (Fig. 22.7A). Once biased, the PL was significantly quenched near both electrodes (more so on the anode side). The PL quenching is the direct result of ECD, which creates nonradiative bandgap states in the LEP. Indeed, doping initiated from the electrode interfaces are propagated toward the opposite electrodes (Fig. 22.1B and C). The darker p-doped region expanded at a faster rate than the n-doped region, leading to the formation of a jagged p-n junction closer to the cathode (Fig. 22.1D). EL became discernible against the background PL (Fig. 22.7E) and grew stronger with time (Fig. 22.7F). The planar PLEC was activated at 335 K and subsequently cooled to 200 K to create a frozen junction. The time evolution of the cell current and cell temperature is shown in Fig. 22.7F. Fig. 22.7G shows the cell was emitting at 200 K, despite a reduced cell current.

Fig. 22.7 Time-lapse fluorescence imaging of a 10.4-mm MEH-PPV:PEO:CsClO4 planar LEC during turn on and cooling. A fixed DC bias of 400 V was applied to turn on the cell, which was at 335 K and under UV illumination. The times since the DC bias was applied to the cell are as follows: (A) no bias, (B) 2 min, (C) 5 min, (D) 8 min, (E) 19 min, (F) 37 min, and (G) 54 min. (H) Cell current and temperature as a function of time since the DC bias was applied. Uniform enhancement (level adjustment in Adobe Photoshop) has been applied to images (A)–(G). Reproduced with permission from Hu, Y., Gao, J., 2011. Direct imaging and probing of the p-n junction in a planar polymer light-emitting electrochemical cell. J. Am. Chem. Soc. 133, 2227–2231 with permission. Copyright (2011) American Chemical Society.

Materials and physics of light-emitting electrochemical cells (LECs)

739

Time-lapse PL imaging of extremely large planar PLECs thus offered direct visual evidence of the dynamic LEC-doping process. The time-lapsed images in Fig. 22.7 bear remarkable resemblance to the schematics of Fig. 22.6, leaving no doubt as to the existence of doping and a doping-induced homojunction. Extremely large planar LECs, with their fully exposed interelectrode gaps, have proved to be a very versatile and valuable tool in elucidating the complex LEC operating mechanism, as well as the electronic structure of the LEC junction (Hu and Gao, 2006; Hu et al., 2006; Hohertz and Gao, 2008; Matyba et al., 2009; Shin et al., 2006). Electrical potential mapping of the frozen PLEC establishes the potential and conductivity profiles of frozen-junction PLECs. Spatially resolved PL and photocurrent mapping provide estimates for the depletion width and doping concentration of the frozen PLEC junction (AlTal and Gao, 2016b, 2015, 2017).

22.3

LEC materials

The first LECs were realized by adding a common polymer electrolyte to the same LEPs used in P-OLEDs. LECs now employ a wide variety of commercially available and custom-designed materials. Because LECs tend to have a very simple structure and are not sensitive to electrode work function, the LEC active layer has been the main focus of research and development. As explained in the previous section, the LEC active layer must possess both electronic and ionic conductivity. The conductive phases should be optimally arranged so that the doping process and the film quality are not adversely affected. Additionally, the active layer should be electrochemically stable so that unintended electrochemical reactions, such as the reduction or oxidation of the electrolyte material, do not occur. All the while, the LEC active layer should be optimized for good response time, color, efficiency, and operational stability. These requirements drive the choice and development of LEC materials.

22.3.1 Light emitters 22.3.1.1 Conjugated polymers (CPs) Fig. 22.8 shows some of the widely used LEPs make PLECs. PPVs and MEH-PPVs are LEPs of the original Cambridge P-OLED and UCSB P-OLED, respectively. Both have been used in the demonstration of the first PLECs. This is made possible by the existence of a common solvent between the LEP (or its precursor) and the polymer electrolyte. MEH-PPV, in particular, is arguably the most widely studied LEP in both P-OLEDs and PLECs due to its excellent solubility in common organic solvents and its low energy gap, which allows easy injection of charge carriers. Sandwich ITO/ PEDOT:PSS/MEH-PPV:PEO:KCF3SO3/Al LECs can last 1000 h at a brightness of 100–200 cd/m2 (Fang et al., 2009). ADS108GE, sold commercially by American Dye Source, Inc., is a green-emitting polymer used in many planar LECs, including the one shown in Fig. 22.5. Super Yellow (SY-PPV) is a soluble PPV copolymer with superior processing and film-forming properties (Spreitzer et al., 1998). The 17  1% quantum yield of SY-PPV in solid film is significantly higher than that of MEH-PPV

740

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 22.8 Chemical structures of some LEPs used in PLECs.

film, at 10  1% (deMello et al., 1997). SY-PPV PLECs holds the record for the longest lifetime of 1400 h during continuous operation at about 200 cd/m2 and exhibit peak efficiency of 18.1 lm/W (Mindemark et al., 2016). SY-PPV LECs also can be operated at a high peak brightness of 2200 cd/m2 and still exhibit a respectable T75 lifetime of 102 h. This translates to a lifetime of 27,000 h at 100 cd/m2 (Yu et al., 2011). mLPPP is a blue and green-emitting polymer with a broad emission spectrum (Tasch et al., 1999). Using a multifluorophoric LEP such as the one shown at the bottom of Fig. 22.8 (Tang et al., 2013a), it is possible to realize LECs with white light emission. The single-emitter white LECs are less prone to the color drifting commonly observed in devices containing multiple emitters (Sun et al., 2010; Tang et al., 2011; Xiong et al., 2015).

Materials and physics of light-emitting electrochemical cells (LECs)

741

22.3.1.2 Nonpolymer emitters LECs based on ionic transition metal complexes have attracted increasing attention in recent years due to their simplicity and excellent overall performance (Fresta and Costa, 2017; Tordera et al., 2012). Further, iTMCs are multifunction phosphorescent emitters that are simultaneously semiconducting and ion-conductive. They generally have very good thermal and electrochemical stabilities. The first iTMC LECs use the ruthenium polypyridyl complex shown in Fig. 22.9 (Lee et al., 1996). Despite having a very modest light output and efficiency, the single-layer, solution-processed ITO/Ru complex/Al device exhibited characteristic LEC behaviors such as a charging effect under a fixed voltage bias and thickness-independent turn-on. Later, iTMC-LECs based on Ru(bpy)2+ 3 complexes achieved much higher external quantum efficiencies in a single-layer structure. The EL turn-on response of the devices was observed to be strongly affected by the size of counter-ions used, signifying the importance of ion conductivity in these devices (Rudmann and Rubner, 2001; Rudmann et al., 2002). While ruthenium iTMCs emit only in the orange-red region, iridium-based iTMCs exhibit easy color tunability, as well as high efficiency. A single-layer Ir(ppy)2(dtbbpy))+(PF6)– LEC can have a brightness of 300 cd/m2 and a power efficiency of 10 lm/W at just 3 V (Slinker et al., 2004). The emission color of IrITMCs can be easily tuned through the selection of ligands coordinated to the Ir core. Green, red, blue, and white Ir-iTMC LECs have all been demonstrated, some with high efficiency and long lifetimes (Costa et al., 2010a,b; Shavaleev et al., 2013; Tordera et al., 2013; Bunzli et al., 2015; Zhang et al., 2013; Hu et al., 2013). Recently, Cu-based iTMCs have drawn some attention due to their lack of heavy metal (Wang et al., 2005; Costa et al., 2011; Elie et al., 2016; Weber et al., 2016). Although they cannot yet compete with Ir- and Ru-based ITMCs in device performance, Cu-based iTMCs show great promise as a new class of ionic LEC compounds with a wide range of emission colors. Emitters other than LEPs and iTMCs also have been introduced to LECs with great success. These include various organic small molecules, QDs, and perovskite nanoparticles. A red-emitting perylene derivative, shown in Fig. 22.9, was one of the first non-ionic, small-molecule emitters used in an LEC (Hill et al., 2008). Small-molecule emitters have wide emission range and are easy to synthesize or purify. Their availability as emitters greatly expanded the material choices for LECs. QDs and nanoparticles (NPs) are the latest emitter to be introduced to LEC. For example, CdSe/ZnS QDs were blended into a host polymer (along with an electrolyte), and an LEC was demonstrated with emission from the QDs (Bader et al., 2011). Red-, blue-, green-, and white-light QD-based LECs have all been realized by careful control of QDs/polymer host ratios. The devices exhibit a maximum brightness of 1000 cd/m2 and a current efficiency of 1.9 cd/A (Qian et al., 2014). LECs can even have perovskite NPs as emitters and display a maximum luminance of over 4000 cd/m2 at 5.5 V (Ayguler et al., 2015) (Zhang et al., 2015; Li et al., 2015). These LECs exhibit the same characteristic LEC transient behavior as PLECs. In particular, the PL imaging of functional planar cells confirms that iTMC-LECs (Meier et al., 2013), along with SM-LECs (Tang et al., 2013b), operate on the same basic principle as a PLEC—the electrochemical formation of a light-emitting p-n junction.

742 Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 22.9 Chemical structures of example Ru-, Ir-, and Cu-based iTMCs used in LECs. Also shown is a perylene emitter used in SM-LECs.

Materials and physics of light-emitting electrochemical cells (LECs)

743

22.3.2 Electrolyte materials All LECs contain mobile ions, which are needed for doping reactions. The type of ions, as well as the ion-transport medium, plays a critical role in the performance of LECs. In the prototypical PLECs, the mobile ions are introduced via the addition of a polymer electrolyte such as the PEO:LiCF3SO3 complex. High-molecular-weight PEOs, however, do not mix well with LEPs, which are often nonpolar high polymers. Large-scale phase separation between PEO and LEP adversely affects the device response time, emission uniformity, and other performance parameters. A simple solution is to thermally anneal the LEC film at temperatures higher than the melting temperature of PEO (Alem and Gao, 2008). This is followed by fast cooling to preserve the finer film morphology realized at high annealing temperatures. Alternatively, a surfactant such as octylcyanoacetate is added to create a fine, bicontinuous network morphology. PLECs with the octylcyanoacetate additive exhibit fast responses (in milliseconds), as well as a much-improved operational lifetime (Cao et al., 1996). PLECs with crown ether (CE) in place of PEO have been demonstrated as well (Cao et al., 1997). As a small molecule soluble in nonpolar solvents, CEs mix well with LEPs due to an increased entropy of mixing (Yang et al., 2003; Panozzo et al., 2002; Ouisse et al., 2002; Habrard et al., 2006). The introduction of ionic liquids (ILs), or molten salts, to LECs makes a separate ion solvent unnecessary. The resulting two-component LEC film can have superior optical quality and very high device performance (Sakanoue et al., 2017). An example of a high performance IL is the methyltrioctylammonium trifluoromethanesulfonate (MATS), shown in Fig. 22.10 (Shao et al., 2009). ILs can also be used as additives to improve the ionic conductivity of iTMC LECs. At high concentrations, however, ILs can phase-separate from the LEPs, leading to fast degradation and poor mechanical integrity. Polymeric ionic liquid (PIL), however, gives rise to excellent device performance, with PIL loading up to 50 wt% (Marcilla et al., 2010). Finally, a class of designer LEPs with built-in ion solvating/transporting functionality has been developed. This class of materials includes both graft copolymers or block copolymers containing oligo(ethylene oxide) or CE side groups or segments (Holzer et al., 1999; Morgado et al., 1999, 2001a,b; Kong et al., 2006; Stephan et al., 2000; Sun et al., 2013, 2002a,b, 2003; Ko et al., 2004). A salt is still needed to provide the necessary ions, but PLECs made with bifunctional polymers can have a smoother and finer surface morphology because the active layer contains a single polymer (Morgado et al., 1999; Kong et al., 2006). In addition to miscibility with emitters, LEC electrolytes must have excellent electrochemical stability and ion-solvating and -transporting properties. A star-branched trimethylolpropane ethoxylate (TMPE), with its superior cathodic stability, is an example of a high-performance electrolyte polymer. TMPE, with alkyl carbonate end-groups, gives rise to PLECs with a fast turn-on, record operational lifetime of 1400 h and a power efficiency exceeding 18 lm/W (Mindemark et al., 2016). Last but not least, LEC electrolytes have been developed for the realization of frozenjunction LECs that can operate at room temperature. As mentioned in Section 22.2.1, the LEC junction is frozen when the counter-ions are immobilized at temperatures

744

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 22.10 Chemical structures of selected electrolyte materials.

below the glass transition temperature (Tg) of the electrolyte material. For frozenjunction LECs to operate at room temperature (RT), a high Tg electrolyte is needed. A high Tg random copolymer-lithium complex, for example, has been developed and applied to PLECs (Wantz et al., 2012). In addition, RT frozen-junction behaviors have been observed in LECs made with certain CE electrolytes (Yu et al., 1998; Edman et al., 2004), a light-emitting polyelectrolyte (Edman et al., 2003), and an ionic liquid electrolyte (Shao et al., 2007). The Tg of LEC electrolytes also can be increased after or during the activation of LECs via the use of various cross-linkable or polymerizable materials (Yu et al., 2013; Leger et al., 2006, 2008; Kosilkin et al., 2010; Pingree et al., 2007). A highly effective approach is to polymerize both the counter-ions and the ion-transport materials using a radical-initiator compound (Tang et al., 2010). And with a photosensitive initiator compound, a frozen junction can be realized on demand with UV exposure (Tang and Edman, 2011).

22.4

PLECs with bipolar electrodes

22.4.1 Introduction to bipolar electrochemistry In an electrochemical cell, a bipolar electrode (BPE) is a floating conductor that can induce electrochemical reactions at its extremities when a sufficient voltage bias is applied to the opposing driving electrodes. Fig. 22.11 illustrates such an electrochemical cell containing a wireless BPE immersed in a liquid electrolyte. When the driving electrodes are biased, an interfacial potential difference develops between the BPE,

Materials and physics of light-emitting electrochemical cells (LECs)

745

Fig. 22.11 Schematic illustration of an electrochemical cell with a bipolar electrode (top). The bottom schematic shows the potential across the electrochemical cell. Interfacial potential difference between the BPE and the electrolyte solution, η+ or η, drives the bipolar electrochemical reactions.

which is an equipotential conductor, and the electrolyte solution, whose potential varies with position. The potential difference is the largest at the two ends of the BPE, where electrochemical reactions, such as redox reactions, will first occur in the presence of suitable chemical species. A BPE has enormous appeal due to its wireless nature and an interfacial potential difference that is variable along the BPE. A branch of electrochemistry called bipolar electrochemistry exploits the wireless nature of BPE for applications that are either inconvenient or impossible to achieve using conventional wired electrodes. For example, millions of micro-BPEs dispersed in an electrolyte solution can be addressed wirelessly to generate three-dimensional (3D) electrochemiluminescence (ECL) (Sentic et al., 2015; de Poulpiquet et al., 2016). Asymmetrical objects, known as Janus objects, can be synthesized in bulk with dispersed carbon tube or carbon bead BPEs (Loget et al., 2012, 2011). Patterned BPE arrays have been widely used for high-throughput detection or screening of biomolecules (Xiao et al., 2017; Khoshfetrat et al., 2015; Zhai et al., 2016) and catalysts (Termebaf et al., 2015; Lin et al., 2012; Fosdick et al., 2013; Zhang et al., 2016) or the large-scale, wireless generation of ECL or functionalized graphene (Chow et al., 2009; Koefoed et al., 2017). Materials with a compositional gradient also can be synthesized with a BPE because of the interfacial potential gradient that exists along the BPE surface (Ishiguro et al., 2011; Shida et al., 2012).

746

Handbook of Organic Materials for Electronic and Photonic Devices

22.4.2 PLECs with evaporated or dispersed bipolar electrodes Conventional bipolar electrochemistry research is typically conducted in a liquid electrochemical cell with metallic or carbon-based BPEs. PLECs with a mixed conductor active layer provide a simple, solid-state platform to study bipolar electrochemistry phenomena. The introduction of BPEs to LECs, in turn, offers performance enhancement that is otherwise impossible to achieve with only two driving electrodes. Although not realized at the time, the principle of bipolar electrochemistry was already at play in 1998 when two sandwich PLECs sharing the same underlying ITO electrodes were turned on by applying a voltage bias directly between the two Al cathodes (Gao et al., 1998). The ITO electrode, which was not directly wired, served as a BPE connecting the two activated LEC junctions in series. The resulting double PLECs exhibit an open-circuit voltage that is twice that of a single cell, confirming that the cells were indeed connected in series. This geometry was recently employed to turn on two sandwich LECs side by side (Gao and AlTal, 2014). The extremely large planar PLECs shown in Fig. 22.5 present another early example of a PLEC with a BPE. Note that the two PLECs were activated by applying an 800 V bias between the two outer Al electrodes. The middle electrode was not biased and is therefore a BPE in the closed configuration. To expand on this concept, a planar PLEC was created with 52 electrically floating aluminum BPEs coated between the two driving electrodes. The cell was turned on by biasing the outer driving electrodes, which was 5.1 mm apart between the inner edges (Tracy and Gao, 2005). This led to the simultaneous activation of 53 light-emitting junctions in series. When the activated PLEC was cooled and operated as a photovoltaic cell (PVC), a giant open-circuit voltage of 63.5 V was measured, again confirming the series connection between the activated junctions, which were connected by a linear array of BPE in the closed configuration. Finally, micrometer-sized conductive ITO particles were dispersed into the PLEC blend, creating a unique bulk homojunction PLEC that contained thousands of tiny, light-emitting p-n junctions formed between the ITO microparticles, which served as BPEs (Tracy and Gao, 2005, 2006). Bulk homojunction PLECs also have been demonstrated by depositing a thin layer (only a few nanometers thick) of gold on top of the polymer layer (Bonnet et al., 2008). Fig. 22.12 shows two 0.5-mm planar PLECs side by side, including one control device on the left without any gold coverage (A). The control device was activated to show a single light-emitting p-n junction closer to the negative electrode (B). The goldcovered PLEC on the right was activated to exhibit EL across the entire interelectrode gap. It is postulated that the ultrathin gold layer that formed isolated patches or islands that functioned as BPEs when the PLEC was biased with 1000 V between the driving electrodes. Bulk homojunctions formed in this fashion have the highest density of all light junctions.

22.4.3 PLECs with bipolar electrodes—operating principles Despite the numerous demonstrations of PLECs with shared or added floating electrodes, which essentially functioned as BPEs, the concept of bipolar electrochemistry was not formally established in LECs until recently (Chen et al., 2016). The PL

Materials and physics of light-emitting electrochemical cells (LECs)

747

Fig. 22.12 Turn-on of MEH-PPV 0.5 mm planar devices, both without and with thin gold film. (A), Image of devices under UV illumination before turn-on. The light orange of the left device is a control where no gold was deposited, while the darker device on the right has 3.4 nm of gold on top of the polymer film. (B) Control device 3 min after turn-on at 150 V, 315 K under UV illumination. (C)–(E) Time-lapse images of the device from the right side of (A) and (B), taken in the dark. These were imaged at >1 s, 6 s, and 36 s after turn-on with conditions of 1000 V and 280 K. Reproduced with permission from Bonnet, W., Tracy, C., Wantz, G., Liu, G., Gao, J., 2008. Bulk electroluminescence from conjugated polymer thin films via the formation of gold nanoislands. Small. 4, 707–1710 with permission. Copyright (2008) Wiley.

imaging of extremely large planar PLECs again provided direct visualization of redox-doping reactions involving BPEs. Fig. 22.13 shows a planar PLEC with two driving electrodes positioned 10.19 mm apart. Three aluminum disk BPEs of different diameters were thermally evaporated on top of PLEC film with a shadow mask. The smaller disks had an elongated shape due to a shadowing effect from the shadow mask. The three disks, however, differed significantly in size, which led to very different rates of reaction around them. The image in Fig. 22.13 was taken approximately 25 s after a bias voltage of 100 V was applied to the driving electrodes. Under UV illumination, dark p-doping and faint n-doping can be seen near the top and bottom

748

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 22.13 Fluorescence image of a 10.19 mm planar cell with three aluminum disk BPEs of different sizes. The various dimensions and distances are labeled in the image. The driving electrodes are gold (top) and aluminum (bottom), respectively. The cell was heated to 350 K, and a 100 V bias had been applied across the driving electrodes approximately 25 s earlier. Also shown are magnified views of the two smaller aluminum disks. Reproduced with permission from Chen, S.L., Wantz, G., Bouffier, L., Gao, J., 2016. Solid-state bipolar electrochemistry: polymer-based light-emitting electrochemical cells. Chem. Aust. 3, 392–398 with permission. Copyright (2016) Wiley.

driving electrodes. In addition, both p- and n-doping have propagated a significant distance from the top and bottom edges of the largest BPE on the left. The largest disk, therefore, was functioning as a BPE from which redox-doping reactions are induced. The two smaller disks, however, did not show any sign of doping in this image. Examination of the cell’s time-lapse PL images of the cell revealed that both p- and n-doping did eventually appear from the top and bottom edges of the middle disk. The cause of the delay (about 60 s for p-doping) was an insufficient potential difference across the middle disk due to its smaller size. The potential drop across the disk increased, however, as doping from both driving electrodes propagated, causing a greater potential drop across the region surrounding the middle disk. But the smallest BPE did not show any sign of doping and was eventually overrun by the propagating p-doping from the driving electrode. The results observed in the

Materials and physics of light-emitting electrochemical cells (LECs)

749

PLEC shown in Fig. 22.13 confirm that a floating conductor indeed can function as a BPE to induce redox reactions at its extremities, and that the size of the floating conductor determines the potential difference across the floating conductor. Introducing BPEs or BPE arrays to PLECs creates an exciting opportunity to conduct bipolar electrochemistry research in solid state. Solid-state bipolar electrochemistry realized in a planar PLEC has some inherent advantages: (1) PLECs have two built-in indicators in PL and EL that allow easy visualization of the bipolar electrochemical processes; and (2) the solid-state nature of the PLEC allows the in situ formation of BPEs in the form of doped polymers or easy deposition or blending of BPEs using a variety of materials and techniques that are not compatible with a conventional liquid electrochemical cell (Gao et al., 2017; Hu et al., 2017). Solid-state bipolar electrochemistry, although still in its infancy, promises new discoveries and applications that straddles fields of condensed matter physics, electrochemistry, and molecular photonic devices.

22.5

Conclusions and outlook

This chapter provides a concise overview of LECs. The background, materials, device characteristics, and device physics of various LECs have been introduced. In addition, PLECs with BPEs have been introduced as a new type of LECs and a brand new platform for bipolar electrochemistry research. LECs are unique among molecular photonic devices. The presence of mobile ions by design distinguishes LECs from purely electronic OLEDs in both device characteristics and operation mechanisms. LECs exhibit highly desirable characteristics despite their simple device structures. For practical applications, however, LECs lag OLEDs in both operational lifetime and efficiencies. Significant progress is being made toward more efficient and long-lasting LECs through the engineering of better materials and device configurations. Under continuous operation, both P-OLEDs and iTMC-LECs can achieve a luminance half-life exceeding 1000 h. This has been attributed to the optimized driving scheme, active layer composition, and device structure (Gao, 2017). To catch up with the lifetime of OLEDs and P-OLEDs, however, another 10- to 100-fold increase is necessary. To accomplish this goal, new electrolyte and light-emitting materials with a large electrochemical stability window need to be developed. In addition, more fundamental research is needed to better understand the complex doping processes and the electronic structures of the LEC junctions. An important recent discovery shows that the apparent luminance decay observed in PLECs was largely a result of luminescence quenching brought on by the doping process itself (Li et al., 2013a). As a result, the luminance decay is not permanent, and an intermittent driving scheme allowed the recovery of peak luminance (Li et al., 2013b). To avoid quenching-induced luminance decay, a very effective approach is to create a frozen p-i-n junction through controlled dedoping so that EL occurs in the less-quenched intrinsic region (Zhang et al., 2006). For LECs to compete with OLEDs in lifetime, response speed, and brightness, the LEC junction must be frozen at room temperature to create an analog of an inorganic p-n junction LED.

750

Handbook of Organic Materials for Electronic and Photonic Devices

References Alem, S., Gao, J., 2008. The effect of annealing/quenching on the performance of polymer lightemitting electrochemical cells. Org. Electron. 9, 347–354. AlTal, F., Gao, J., 2015. Scanning photocurrent and PL imaging of a frozen polymer p–i–n junction. Phys. Status Solidi RRL 9, 77–81. AlTal, F., Gao, J., 2016a. Charging and discharging of a planar polymer light-emitting electrochemical cell. Electrochim. Acta 220, 529–535. AlTal, F., Gao, J., 2016b. High resolution scanning optical imaging of a frozen polymer p-n junction. J. Appl. Phys. 120, 115501. Altal, F., Gao, J., 2017. High resolution scanning optical imaging of a frozen planar polymer light-emitting electrochemical cell: an experimental and modelling study. Sci. China Chem. 60, 497–503. Ayguler, M.F., Weber, M.D., Puscher, B.M.D., Medina, D.D., Docampo, P., Costa, R.D., 2015. Light-emitting electrochemical cells based on hybrid lead halide perovskite nanoparticles. J. Phys. Chem. C 119, 12047–12054. Bader, A.J.N., Ilkevich, A.A., Kosilkin, I.V., Leger, J.M., 2011. Precise color tuning via hybrid light-emitting electrochemical cells. Nano Lett. 11, 461–465. Bonnet, W., Tracy, C., Wantz, G., Liu, G., Gao, J., 2008. Bulk electroluminescence from conjugated polymer thin films via the formation of gold nanoislands. Small 4, 1707–1710. Braun, D., Heeger, A.J., 1991. Visible-light emission from semiconducting polymer diodes. Appl. Phys. Lett. 58, 1982–1984. Bunzli, A.M., Constable, E.C., Housecroft, C.E., Prescimone, A., Zampese, J.A., Longo, G., Gil-Escrig, L., Pertegas, A., Orti, E., Bolink, H.J., 2015. Exceptionally long-lived light-emitting electrochemical cells: multiple intra-cation pi-stacking interactions in Ir(C N)(2)(N N) PF6 emitters. Chem. Sci. 6, 2843–2852. Burroughes, J.H., Bradley, D.D.C., Brown, A.R., Marks, R.N., Mackay, K., Friend, R.H., Burns, P.L., Holmes, A.B., 1990. Light-emitting diodes based on conjugated polymers. Nature 347, 539–541. Cao, Y., Yu, G., Heeger, A.J., Yang, C.Y., 1996. Efficient, fast response light-emitting electrochemical cells: electroluminescent and solid electrolyte polymers with interpenetrating network morphology. Appl. Phys. Lett. 68, 3218–3220. Cao, Y., Pei, Q.B., Andersson, M.R., Yu, G., Heeger, A.J., 1997. Light-emitting electrochemical cells with crown ether as solid electrolyte. J. Electrochem. Soc. 144, L317–L320. Chen, S.L., Wantz, G., Bouffier, L., Gao, J., 2016. Solid-state bipolar electrochemistry: polymer-based light-emitting electrochemical cells. Chem. Aust. 3, 392–398. Chow, K.-F., Mavre, F., Crooks, J.A., Chang, B.-Y., Crooks, R.M., 2009. A large-scale, wireless electrochemical bipolar electrode microarray. J. Am. Chem. Soc. 131, 8364. Costa, R.D., Pertegas, A., Orti, E., Bolink, H.J., 2010a. Improving the turn-on time of lightemitting electrochemical cells without sacrificing their stability. Chem. Mater. 22, 1288–1290. Costa, R.D., Orti, E., Bolink, H.J., Graber, S., Housecroft, C.E., Constable, E.C., 2010b. Efficient and long-living light-emitting electrochemical cells. Adv. Funct. Mater. 20, 1511–1520. Costa, R.D., Tordera, D., Orti, E., Bolink, H.J., Schonle, J., Graber, S., Housecroft, C.E., Constable, E.C., Zampese, J.A., 2011. Copper(I) complexes for sustainable light-emitting electrochemical cells. J. Mater. Chem. 21, 16108–16118.

Materials and physics of light-emitting electrochemical cells (LECs)

751

Costa, R.D., Orti, E., Bolink, H.J., Monti, F., Accorsi, G., Armaroli, N., 2012. Luminescent ionic transition-metal complexes for light-emitting electrochemical cells. Angew. Chem. Int. Ed. 51, 8178–8211. de Poulpiquet, A., Diez-Buitrago, B., Milutinovic, M., Goudeau, B., Bouffier, L., Arbault, S., Kuhn, A., Sojic, N., 2016. Dual-color electrogenerated chemiluminescence from dispersions of conductive microbeads addressed by bipolar electrochemistry. Chem. Aust. 3, 404–409. deMello, J.C., Wittmann, H.F., Friend, R.H., 1997. An improved experimental determination of external photoluminescence quantum efficiency. Adv. Mater. 9, 230. deMello, J.C., Tessler, N., Graham, S.C., Friend, R.H., 1998. Ionic space-charge effects in polymer light-emitting diodes. Phys. Rev. B 57, 12951–12963. Dick, D.J., Heeger, A.J., Yang, Y., Pei, Q., 1996. Imaging the structure of the p-n junction in polymer light-emitting electrochemical cells. Adv. Mater. 8, 985–987. Edman, L., Pauchard, M., Liu, B., Bazan, G., Moses, D., Heeger, A.J., 2003. Single-component light-emitting electrochemical cell with improved stability. Appl. Phys. Lett. 82, 3961–3963. Edman, L., Summers, M.A., Buratto, S.K., Heeger, A.J., 2004. Polymer light-emitting electrochemical cells: doping, luminescence, and mobility. Phys. Rev. B. 70, 115212. https://doi. org/10.1103/PhysRevB.70.115212. Elie, M., Sguerra, F., Di Meo, F., Weber, M.D., Marion, R., Grimault, A., Lohier, J.F., Stallivieri, A., Brosseau, A., Pansu, R.B., et al., 2016. Designing NHC-copper(I) dipyridylamine complexes for blue light-emitting electrochemical cells. ACS Appl. Mater. Interfaces 8, 14678–14691. Ertl, C.D., Momblona, C., Pertegas, A., Junquera-Hernandez, J.M., La-Placa, M.G., Prescimone, A., Orti, E., Housecroft, C.E., Constable, E.C., Bolink, H.J., 2017. Highly stable red-light-emitting electrochemical cells. J. Am. Chem. Soc. 139, 3237–3248. Fang, J., Matyba, P., Edman, L., 2009. The design and realization of flexible, long-lived lightemitting electrochemical cells. Adv. Funct. Mater. 19, 2671–2676. Fosdick, S.E., Berglund, S.P., Mullins, C.B., Crooks, R.M., 2013. Parallel screening of electrocatalyst candidates using bipolar electrochemistry. Anal. Chem. 85, 2493–2499. Fresta, E., Costa, R.D., 2017. Beyond traditional light-emitting electrochemical cells – a review of new device designs and emitters. J. Mater. Chem. C 5, 5643–5675. Frohleiks, J., Wefers, F., Wepfer, S., Hong, A.R., Jang, H.S., Nannen, E., 2017. CuInS2-based quantum dot light-emitting electrochemical cells (QLECs). Adv. Mater. Technol. 2, 1700154. https://doi.org/10.1002/admt.201700154. Gao, J., 2017. Strategies toward long-life light-emitting electrochemical cells. ChemPlusChem 82, 1–15. Gao, J., 2018. Polymer light-emitting electrochemical cells-recent advances and future trends. Curr. Opin. Electrochem. 7, 87–94. Gao, J., AlTal, F., 2014. Decoupled luminance decay and voltage drift in polymer light-emitting electrochemical cells: forward bias vs. reverse bias operation. Appl. Phys. Lett. 104, 143301. Gao, J., Dane, J., 2003. Planar polymer light-emitting electrochemical cells with extremely large interelectrode spacing. Appl. Phys. Lett. 83, 3027–3029. Gao, J., Dane, J., 2004. Visualization of electrochemical doping and light-emitting junction formation in conjugated polymer films. Appl. Phys. Lett. 84, 2778–2780. Gao, J., Yu, G., Heeger, A.J., 1997. Polymer light-emitting electrochemical cells with frozen p-i-n junction. Appl. Phys. Lett. 71, 1293–1295.

752

Handbook of Organic Materials for Electronic and Photonic Devices

Gao, J., Yu, G., Heeger, A.J., 1998. Polymer p-i-n junction photovoltaic cells. Adv. Mater. 10, 692–695. Gao, J., Chen, S., AlTal, F., Hu, S., Wantz, G., Bouffier, L., 2017. Bipolar electrode array embedded in a polymer light-emitting electrochemical cell. ACS Appl. Mater. Interfaces 9, 32405–32410. Gonzalezbasurto, J., Burshtein, Z., 1975. Electroluminescence studies in pyrene single-crystals. Mol. Cryst. Liq. Cryst. 31, 211–217. Ha, J., Park, J., Ha, J., Kim, D., Chung, S., Lee, C., Hong, Y., 2015. Selectively modulated inkjet printing of highly conductive and transparent foldable polymer electrodes for flexible polymer light-emitting diode applications. Org. Electron. 19, 147–156. Habrard, F., Ouisse, T., Stephan, O., 2006. Conjugated polymer/molten salt blend optimization. J. Phys. Chem. B 110, 15049–15051. Helfrich, W., Schneider, W.G., 1965. Recombination radiation in anthracene crystals. Phys. Rev. Lett. 14, 229. Helfrich, W., WG, S., 1966. Transients of volume-controlled current and of recombination radiation in anthracene. J. Chem. Phys. 44, 2902. Hill, Z.B., Rodovsky, D.B., Leger, J.M., Bartholomew, G.P., 2008. Synthesis and utilization of perylene-based n-type small molecules in light-emitting electrochemical cells. Chem. Commun. 6594–6596. Hohertz, D., Gao, J., 2008. How electrode work function affects doping and electroluminescence of polymer light-emitting electrochemical cells. Adv. Mater. 20, 3298–3302. Holzer, L., Winkler, B., Wenzl, F.P., Tasch, S., Dai, L., Mau, A.W.H., Leising, G., 1999. Lightemitting electrochemical cells and light-emitting diodes based on ionic conductive poly(phenylene vinylene): a new chemical sensor system. Synth. Met. 100, 71–77. Hu, Y., Gao, J., 2006. Cationic effects in polymer light-emitting electrochemical cells. Appl. Phys. Lett. 89, 253514. https://doi.org/10.1063/1.2422877. Hu, Y., Gao, J., 2011. Direct imaging and probing of the p  n junction in a planar polymer lightemitting electrochemical cell. J. Am. Chem. Soc. 133, 2227–2231. Hu, Y., Tracy, C., Gao, J., 2006. High-resolution imaging of electrochemical doping and dedoping processes in luminescent conjugated polymers. Appl. Phys. Lett. 88, 123507. Hu, T., Duan, L., Qiao, J., He, L., Zhang, D., Wang, L., Qiu, Y., 2013. Efficient doped red lightemitting electrochemical cells based on cationic iridium complexes. Synth. Met. 163, 33–37. Hu, S.Y., Chi, X., Chen, S.L., AlTal, F., Gao, J., 2017. Visualizing the bipolar electrochemistry of electrochemically doped luminescent conjugated polymers. J. Phys. Chem. C 121, 8409–8415. Ishiguro, Y., Inagi, S., Fuchigami, T., 2011. Gradient doping of conducting polymer films by means of bipolar electrochemistry. Langmuir 27, 7158–7162. Kajii, H., Mizuno, K., Ohmori, Y., 2016. Improved characteristics of polymer light-emitting diode based on poly(alkylfluorene) with functional layer fabricated by contact printing method. In: India-Japan Expert Group Meeting on Biomolecular Electronics & Organic Nanotechnology for Environment Preservation (Ijegmbe, 2015), p. 704. Kalinowski, J., Godlewski, J., Signerski, R., 1976. Electroluminescence in tetracene crystals. Mol. Cryst. Liq. Cryst. 33, 247–259. Kampas, F.J., Gouterman, M., 1977. Porphyrin films–electroluminescence of octaethylporphin. Chem. Phys. Lett. 48, 233–236. Khoshfetrat, S.M., Ranjbari, M., Shayan, M., Mehrgardi, M.A., Kiani, A., 2015. Wireless electrochemiluminescence bipolar electrode array for visualized genotyping of single nucleotide polymorphism. Anal. Chem. 87, 8123–8131.

Materials and physics of light-emitting electrochemical cells (LECs)

753

Ko, H.C., Lim, D.K., Kim, S.H., Choi, W., Lee, H., 2004. Light-emitting electrochemical cells based on polyimide containing perylene and tri(ethylene oxide) moieties. Synth. Met. 144, 177–182. Koefoed, L., Pedersen, S.U., Daasbjerg, K., 2017. Bipolar electrochemistry-a wireless approach for eletrode reactions. Curr. Opin. Electrochem. 2, 13–17. Kong, F., Wu, X.L., Zheng, Y., Ou, C.G., Yuan, R.K., Yang, C.Z., Chu, P.K., 2006. Optical and electrical properties of single-ion transport light-emitting electrochemical cells. J. Appl. Phys. 99, 064505. https://doi.org/10.1063/1.2180452. Kosilkin, I.V., Martens, M.S., Murphy, M.P., Leger, J.M., 2010. Polymerizable ionic liquids for fixed-junction polymer light-emitting electrochemical cells. Chem. Mater. 22, 4838–4840. Lee, J.K., Yoo, D.S., Handy, E.S., Rubner, M.F., 1996. Thin film light emitting devices from an electroluminescent ruthenium complex. Appl. Phys. Lett. 69, 1686–1688. Lee, J., Chen, H.F., Batagoda, T., Coburn, C., Djurovich, P.I., Thompson, M.E., Forrest, S.R., 2016. Deep blue phosphorescent organic light-emitting diodes with very high brightness and efficiency. Nat. Mater. 15, 92–98. Lee, Y., Suh, M., Kim, K., Kim, H., Kim, D., Chang, H., Lee, D., Kim, Y., Kim, S.W., Jeon, D.Y., 2017. Conjugated polyelectrolyte-assisted vacuum-free transfer-printing of silver nanowire network for top electrode of polymer light-emitting diodes. Org. Electron. 43, 64–69. Leger, J.M., Rodovsky, D.B., Bartholomew, G.P., 2006. Self-assembled, chemically fixed homojunctions in semiconducting polymers. Adv. Mater. 18, 3130–3134. Leger, J.M., Patel, D.G., Rodovsky, D.B., Bartholomew, G.P., 2008. Polymer photovoltaic devices employing a chemically fixed p–i–n junction. Adv. Funct. Mater. 18, 1212–1219. Li, X., Gao, J., Liu, G., 2013. Reversible luminance decay in polymer light-emitting electrochemical cells. Appl. Phys. Lett. 102, 223303. Li, X., AlTal, F., Liu, G., Gao, J., 2013b. Long-term, intermittent testing of sandwich polymer light-emitting electrochemical cells. Appl. Phys. Lett. 103, 243304. Li, J.Q., Bade, S.G.R., Shan, X., Yu, Z.B., 2015. Single-layer light-emitting diodes using organometal halide perovskite/poly(ethylene oxide) composite thin films. Adv. Mater. 27, 5196–5202. Liao, C.T., Chen, H.F., Su, H.C., Wong, K.T., 2012. Improving the balance of carrier mobilities of host-guest solid-state light-emitting electrochemical cells. Phys. Chem. Chem. Phys. 14, 1262–1269. Lin, X., Zheng, L., Gao, G., Chi, Y., Chen, G., 2012. Electrochemiluminescence imaging-based high-throughput screening platform for electrocatalysts used in fuel cells. Anal. Chem. 84, 7700–7707. Lin, K.Y., Bastatas, L.D., Suhr, K.J., Moore, M.D., Holliday, B.J., Minary-Jolandan, M., Slinker, J.D., 2016. Influence of lithium additives in small molecule light-emitting electrochemical cells. ACS Appl. Mater. Interfaces 8, 16776–16782. Loget, G., Lapeyre, V., Garrigue, P., Warakulwit, C., Limtrakul, J., Delville, M.-H., Kuhn, A., 2011. Versatile procedure for synthesis of janus-type carbon tubes. Chem. Mater. 23, 2595–2599. Loget, G., Roche, J., Kuhn, A., 2012. True bulk synthesis of janus objects by bipolar electrochemistry. Adv. Mater. 24, 5111–5116. Lohmann, F., Mehl, W., 1969. Dark injection and radiative recombination of electrons and holes in naphthalene crystals. J. Chem. Phys. 50, 500. Marcilla, R., Mecerreyes, D., Winroth, G., Brovelli, S., Yebra, M.M.R., Cacialli, F., 2010. Light-emitting electrochemical cells using polymeric ionic liquid/polyfluorene blends as luminescent material. Appl. Phys. Lett. 96, 043308. https://doi.org/10.1063/1.3279155.

754

Handbook of Organic Materials for Electronic and Photonic Devices

Matyba, P., Maturova, K., Kemerink, M., Robinson, N.D., Edman, L., 2009. The dynamic organic p–n junction. Nat. Mater. 8, 672–676. Meier, S.B., van Reenen, S., Lefevre, B., Hartmann, D., Bolink, H.J., Winnacker, A., Sarfert, W., Kemerink, M., 2013. Dynamic doping in planar ionic transition metal complex-based lightemitting electrochemical cells. Adv. Funct. Mater. 23, 3531–3538. Mindemark, J., Tang, S., Wang, J., Kaihovirta, N., Brandell, D., Edman, L., 2016. Highperformance light-emitting electrochemical cells by electrolyte design. Chem. Mater. 28, 2618–2623. Morgado, J., Friend, R.H., Cacialli, F., Chuah, B.S., Moratti, S.C., Holmes, A.B., 1999. Lightemitting devices based on a poly(p-phenylene vinylene) derivative with ion-coordinating side groups. J. Appl. Phys. 86, 6392–6395. Morgado, J., Cacialli, F., Friend, R.H., Chuah, B.S., Rost, H., Holmes, A.B., 2001a. Lightemitting devices based on a poly(p-phenylenevinylene) statistical copolymer with oligo(ethylene oxide) side groups. Macromolecules 34, 3094–3099. Morgado, J., Friend, R.H., Cacialli, F., Chuah, B.S., Rost, H., Moratti, S.C., Holmes, A.B., 2001b. Light-emitting electrochemical cells based on poly(p-phenylene vinylene) copolymers with ion-transporting side groups. Synth. Met. 122, 111–113. Nishikitani, Y., Suga, K., Uchida, S., Nishimura, S., Oyaizu, K., Nishide, H., 2017. High-colorrendering-index white polymer light-emitting electrochemical cells based on ionic hostguest systems: utilization of blend films of blue-fluorescent cationic polyfluorenes and red-phosphorescent cationic iridium complexes. Org. Electron. 51, 168–172. Ouisse, T., Armand, M., Kervella, Y., Stephan, O., 2002. Fully transparent, organic lightemitting electrochemical cells. Appl. Phys. Lett. 81, 3131–3133. Panozzo, S., Armand, M., Stephan, O., 2002. Light-emitting electrochemical cells using a molten delocalized salt. Appl. Phys. Lett. 80, 679–681. Parker, I.D., 1994. Carrier tunneling and device characteristics in polymer light-emittingdiodes. J. Appl. Phys. 75, 1656–1666. Pei, Q.B., Yu, G., Zhang, C., Yang, Y., Heeger, A.J., 1995. Polymer light-emitting electrochemical-cells. Science 269, 1086–1088. Pei, Q.B., Yang, Y., Yu, G., Zhang, C., Heeger, A.J., 1996. Polymer light-emitting electrochemical cells: in situ formation of a light-emitting p-n junction. J. Am. Chem. Soc. 118, 3922–3929. Pertega´s, A., Wong, M.Y., Sessolo, M., Zysman-Colman, E., Bolink, H.J., 2016. Efficient lightemitting electrochemical cells using small molecular weight, ionic, host-guest systems. ECS J. Solid State Sci. Technol. 5, R3160–R3163. Pingree, L.S.C., Rodovsky, D.B., Coffey, D.C., Bartholomew, G.P., Ginger, D.S., 2007. Scanning kelvin probe imaging of the potential profiles in fixed and dynamic planar LECs. J. Am. Chem. Soc. 129, 15903–15910. Pope, M., Kallmann, H.P., Magnante, P., 1963. Electroluminescence in organic crystals. J. Chem. Phys. 38, 2042. Qian, G., Lin, Y., Wantz, G., Davis, A.R., Carter, K.R., Watkins, J.J., 2014. Saturated and multicolored electroluminescence from quantum dots based light emitting electrochemical cells. Adv. Funct. Mater. 24, 4484–4490. Rudmann, H., Rubner, M.F., 2001. Single layer light-emitting devices with high efficiency and long lifetime based on tris(2,2 ’ bipyridyl) ruthenium(II) hexafluorophosphate. J. Appl. Phys. 90, 4338–4345. Rudmann, H., Shimada, S., Rubner, M.F., 2002. Solid-state light-emitting devices based on the tris-chelated ruthenium(II) complex. 4. High-efficiency light-emitting devices based on derivatives of the tris(2,2 ’-bipyridyl) ruthenium(II) complex. J. Am. Chem. Soc. 124, 4918–4921.

Materials and physics of light-emitting electrochemical cells (LECs)

755

Sakanoue, T., Yonekawa, F., Albrecht, K., Yamamoto, K., Takenobu, T., 2017. An ionic liquid that dissolves semiconducting polymers: a promising electrolyte for bright, efficient, and stable light-emitting electrochemical cells. Chem. Mater. 29, 6122–6129. Sentic, M., Arbault, S., Bouffier, L., Manojlovic, D., Kuhn, A., Sojic, N., 2015. 3D electrogenerated chemiluminescence: from surface-confined reactions to bulk emission. Chem. Sci. 6, 4433–4437. Shanmugasundaram, K., Subeesh, M.S., Sunesh, C.D., Choe, Y., 2016. Non-doped deep blue light-emitting electrochemical cells from charged organic small molecules. RSC Adv. 6, 28912–28918. Shao, Y., Bazan, G.C., Heeger, A.J., 2007. Long-lifetime polymer light-emitting electrochemical cells. Adv. Mater. 19, 365–370. Shao, Y., Gong, X., Heeger, A.J., Liu, M., Jen, A.K.Y., 2009. Long-lifetime polymer lightemitting electrochemical cells fabricated with crosslinked hole-transport layers. Adv. Mater. 21, 1972–1975. Shavaleev, N.M., Scopelliti, R., Graetzel, M., Nazeeruddin, M.K., Pertegas, A., RoldanCarmona, C., Tordera, D., Bolink, H.J., 2013. Pulsed-current versus constant-voltage light-emitting electrochemical cells with trifluoromethyl-substituted cationic iridium(III) complexes. J. Mater. Chem. C 1, 2241–2248. Shida, N., Ishiguro, Y., Atobe, M., Fuchigami, T., Inagi, S., 2012. Electro-click modification of conducting polymer surface using Cu(I) species generated on a bipolar electrode in a gradient manner. ACS Macro Lett. 1, 656–659. Shin, J.H., Dzwilewski, A., Iwasiewicz, A., Xiao, S., Fransson, A., Ankah, G.N., Edman, L., 2006. Light emission at 5 V from a polymer device with a millimeter-sized interelectrode gap. Appl. Phys. Lett. 89, 013509. Slinker, J.D., Gorodetsky, A.A., Lowry, M.S., Wang, J.J., Parker, S., Rohl, R., Bernhard, S., Malliaras, G.G., 2004. Efficient yellow electroluminescence from a single layer of a cyclometalated iridium complex. J. Am. Chem. Soc. 126, 2763–2767. Slinker, J.D., DeFranco, J.A., Jaquith, M.J., Silveira, W.R., Zhong, Y.-W., MoranMirabal, J.M., Craighead, H.G., Abrun˜a, H.D., Marohn, J.A., Malliaras, G.G., 2007. Direct measurement of the electric-field distribution in a light-emitting electrochemical cell. Nat. Mater. 6, 894–899. Spreitzer, H., Becker, H., Kluge, E., Kreuder, W., Schenk, H., Demandt, R., Schoo, H., 1998. Soluble phenyl-substituted PPVs–new materials for highly efficient polymer LEDs. Adv. Mater. 10, 1340. Stephan, O., Collomb, V., Vial, J.C., Armand, M., 2000. Blue-green light-emitting diodes and electrochemical cells based on a copolymer derived from fluorene. Synth. Met. 113, 257–262. Sun, Q.J., Wang, H.Q., Yang, C.H., Wang, X.G., Liu, D.S., Li, Y.F., 2002. Polymer lightemitting electrochemical cells with the block copolymers containing PEO segments. Thin Solid Films 417, 14–19. Sun, Q.J., Wang, H.Q., Yang, C.H., He, G.F., Li, Y.F., 2002a. Blue-green light-emission LECs based on block copolymers containing di(alpha-naphthalene vinylene)benzene chromophores and tri(ethylene oxide) spacers. Synth. Met. 128, 161–165. Sun, Q.J., Wang, H.Q., Yang, C.H., Li, Y.F., 2003. Synthesis and electroluminescence of novel copolymers containing crown ether spacers. J. Mater. Chem. 13, 800–806. Sun, M., Zhong, C., Li, F., Cao, Y., Pei, Q., 2010. A fluorene-oxadiazole copolymer for white light-emitting electrochemical cells. Macromolecules 43, 1714–1718. Sun, M., Zhong, C., Yu, Z., Cao, Y., Pei, Q., 2013. Fluorene-benzothiadiazole copolymer for single component green light-emitting electrochemical cells. J. Disp. Technol. 9, 476–482.

756

Handbook of Organic Materials for Electronic and Photonic Devices

Takeshita, K., Kawakami, H., Shimizu, T., Kitazume, E., Oota, K., Taguchi, T., Takashima, I., 2005. Development of polymer light-emitting diode (PLED) displays using a printing method. In: Idw/Ad ’05: Proceedings of the 12th International Display Workshops in Conjunction with Asia Display 2005. vols. 1 and 2, pp. 597–600. Tang, S., Edman, L., 2011. On-demand photochemical stabilization of doping in light-emitting electrochemical cells. Electrochim. Acta 56, 10473–10478. Tang, C.W., Vanslyke, S.A., 1987. Organic electroluminescent diodes. Appl. Phys. Lett. 51, 913–915. Tang, S., Irgum, K., Edman, L., 2010. Chemical stabilization of doping in conjugated polymers. Org. Electron. 11, 1079–1087. Tang, S., Pan, J., Buchholz, H., Edman, L., 2011. White light-emitting electrochemical cell. ACS Appl. Mater. Interfaces 3, 3384–3388. Tang, S., Pan, J., Buchholz, H.A., Edman, L., 2013. White light from a single-emitter lightemitting electrochemical cell. J. Am. Chem. Soc. 135, 3647–3652. Tang, S., Tan, W.-Y., Zhu, X.-H., Edman, L., 2013b. Small-molecule light-emitting electrochemical cells: evidence for in situ electrochemical doping and functional operation. Chem. Commun. 49, 4926–4928. Tang, S., Buchholz, H.A., Edman, L., 2015. On the selection of a host compound for efficient host-guest light-emitting electrochemical cells. J. Mater. Chem. C 3, 8114–8120. Tasch, S., Gao, J., Wenzl, F.P., Holzer, L., Leising, G., Heeger, A.J., Scherf, U., Mullen, K., 1999. Blue and green light-emitting electrochemical cells with microsecond response times. Electrochem. Solid-State Lett. 2, 303–305. Termebaf, H., Shayan, M., Kiani, A., 2015. Two-step bipolar electrochemistry: generation of composition gradient and visual screening of electrocatalytic activity. Langmuir 31, 13238–13246. Tordera, D., Meier, S., Lenes, M., Costa, R.D., Orti, E., Sarfert, W., Bolink, H.J., 2012. Simple, fast, bright, and stable light sources. Adv. Mater. 24, 897. Tordera, D., Frey, J., Vonlanthen, D., Constable, E., Pertegas, A., Orti, E., Bolink, H.J., Baranoff, E., Nazeeruddin, M.K., 2013. Low current density driving leads to efficient, bright, and stable green electroluminescence. Adv. Energy Mater. 3, 1338–1343. Tracy, C., Gao, J., 2005. Polymer bulk homojunction photonic devices. Appl. Phys. Lett. 87, 143502. Tracy, C., Gao, J., 2006. Polymer bulk homojunction light-emitting electrochemical cells. J. Appl. Phys. 100, 104503. Vincett, P.S., Barlow, W.A., Hann, R.A., Roberts, G.G., 1982. Electrical-conduction and lowvoltage blue electro-luminescence in vacuum-deposited organic films. Thin Solid Films 94, 171–183. Wang, Y.M., Teng, F., Hou, Y.B., Xu, Z., Wang, Y.S., Fu, W.F., 2005. Copper(I) complex employed in organic light-emitting electrochemical cells: device and spectra shift. Appl. Phys. Lett. 87, 233512. https://doi.org/10.1063/1.2139987. Wantz, G., Gautier, B., Dumur, F., Phan, T.N.T., Gigmes, D., Hirsch, L., Gao, J., 2012. Towards frozen organic PN junctions at room temperature using high-Tg polymeric electrolytes. Org. Electron. 13, 1859–1864. Weber, M.D., Garino, C., Volpi, G., Casamassa, E., Milanesio, M., Barolo, C., Costa, R.D., 2016. Origin of a counterintuitive yellow light-emitting electrochemical cell based on a blue-emitting heteroleptic copper(I) complex. Dalton Trans. 45, 8984–8993. Wong, M.Y., Hedley, G.J., Xie, G.H., Kolln, L.S., Samuel, I.D.W., Pertegas, A., Bolink, H.J., Zysinan-Colman, E., 2015. Light-emitting electrochemical cells and solution-processed

Materials and physics of light-emitting electrochemical cells (LECs)

757

organic light-emitting diodes using small molecule organic thermally activated delayed fluorescence emitters. Chem. Mater. 27, 6535–6542. Xiao, Y., Xu, L.R., Qi, L.W., 2017. Electrochemiluminescence bipolar electrode array for the multiplexed detection of glucose, lactate and choline based on a versatile enzymatic approach. Talanta 165, 577–583. Xiong, Y., Li, L., Liang, J., Gao, H., Chou, S., Pei, Q., 2015. Efficient white polymer lightemitting electrochemical cells. Mater. Horiz. 2, 338–343. Yang, C.H., Sun, Q.J., Qiao, J., Li, Y.F., 2003. Ionic liquid doped polymer light-emitting electrochemical cells. J. Phys. Chem. B 107, 12981–12988. Yu, G., Cao, Y., Andersson, M., Gao, J., Heeger, A.J., 1998. Polymer light-emitting electrochemical cells with frozen p-i-n junction at room temperature. Adv. Mater. 10, 385–388. Yu, Z., Wang, M., Lei, G., Liu, J., Li, L., Pei, Q., 2011. Stabilizing the dynamic p  in Junction in polymer light-emitting electrochemical cells. J. Phys. Chem. Lett. 2, 367–372. Yu, Z., Li, L., Gao, H., Pei, Q., 2013. Polymer light-emitting electrochemical cells: recent developments to stabilize the p-i-n junction and explore novel device applications. SCIENCE CHINA Chem. 56, 1075–1086. Zhai, Q.F., Zhang, X.W., Han, Y.C., Zhai, J.F., Li, J., Wang, E.K., 2016. A nanoscale multichannel closed bipolar electrode array for electrochemiluminescence sensing platform. Anal. Chem. 88, 945–951. Zhang, Y., Gao, J., 2006. Lifetime study of polymer light-emitting electrochemical cells. J. Appl. Phys. 100, 084501. Zhang, Y.G., Hu, Y.F., Gao, J., 2006. Improving the efficiency of polymer light-emitting electrochemical cells by controlled doping relaxation. Appl. Phys. Lett. 88, 163507. Zhang, J., Zhou, L., Al-Attar, H.A., Shao, K.Z., Wang, L., Zhu, D.X., Su, Z.M., Bryce, M.R., Monkman, A.P., 2013. Efficient light-emitting electrochemical cells (LECs) based on ionic iridium(III) complexes with 1,3,4-oxadiazole ligands. Adv. Funct. Mater. 23, 4667–4677. Zhang, H.M., Lin, H., Liang, C.J., Liu, H., Liang, J.J., Zhao, Y., Zhang, W.G., Sun, M.J., Xiao, W.K., Li, H., et al., 2015. Organic-inorganic perovskite light-emitting electrochemical cells with a large capacitance. Adv. Funct. Mater. 25, 7226–7232. Zhang, X.W., Shang, C.S., Gu, W.L., Xia, Y., Li, J., Wang, E., 2016. A renewable display platform based on the bipolar electrochromic electrode. Chem. Aust. 3, 383–386.

This page intentionally left blank

Vertical organic transistors

23

Bjo€rn Lu€ssem*, Changmin Keum†, Gil Sheleg‡, Nir Tessler‡ *Department of Physics, Kent State University, Kent, OH, United States, † Organic Semiconductor Centre, SUPA, School of Physics and Astronomy, University of St Andrews, St Andrews, United Kingdom, ‡Sara and Moshe Zisapel Nano-Electronic Center, Department of Electrical Engineering, Technion-Israel Institute of Technology, Haifa, Israel

23.1

Introduction

The flexibility and versatility of organic field-effect transistors (OFETs) opens new applications, such as in the field of wearable systems (Gualandi et al., 2016; Bonfiglio et al., 2005; Barbaro et al., 2010), bioelectronics (Owens and Malliaras, 2010; Berggren and Richter-Dahlfors, 2007), or as driving elements for individual pixels in a flat-panel display (Steudel et al., 2012). Being intensively researched since the 1980s (Ebisawa et al., 1983), OFETs have achieved a remarkable maturity in terms of reproducibility, available material systems, processing technology, and fundamental understanding of the underlying physics (Sirringhaus, 2014; Klauk, 2010). However, they remain limited in their performance, particularly in their transit frequency fT. The best OFETs reach transit frequencies in the lower-megahertz range (Klauk, 2010; Ante et al., 2012) or in the 10-MHz vicinity, if increased voltages are tolerated (Kitamura and Arakawa, 2011). The transit frequency of thin-film transistors is given by fT


μðVGS  Vth Þ 2πLðL + 2LC Þ

(23.1)

where μ is the charge-carrier mobility inside the organic semiconductor (OSC), VGS  Vth is the gate-source voltage corrected by the threshold voltage, L is the channel length, and LC is the parasitic overlap length between source/drain electrode and gate electrode. Considering Eq. (23.1), the low transit frequencies reached in OFETs are caused by the relatively low charge-carrier mobility of OSCs. Although significant advances have been made in recent years, the best charge-carrier mobilities are in the range of 1…10 cm2/V s (Yuan et al., 2014; Lee et al., 2016; Bittle et al., 2016). Furthermore, severe contact resistance effects are observed in OFETs, which reduce the effective charge-carrier mobility, and hence reduce the transit frequency (Ante et al., 2012; L€ ussem et al., 2016). Instead of increasing the charge-carrier mobility of OSCs, the channel length L of OFETs can be reduced, which is expected to increase the transit frequency as well. Unfortunately, scaling the channel length of OFETs into submicrometer dimensions Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00023-1 © 2019 Elsevier Ltd. All rights reserved.

760

Handbook of Organic Materials for Electronic and Photonic Devices

not only defies the low-cost nature of OFETs, but also is often ineffective due to contact resistance limitations (Ante et al., 2012). Vertical organic transistors try to address these challenges (Lussem et al., 2015; Lin et al., 2015). In vertical transistors, the transistor channel is formed perpendicular to the substrate and the channel length is defined by the thickness of the organic layers. As the thickness of organic layer can be precisely controlled down to nanometer precision, the channel length of these transistors can be scaled easily into the submicrometer area. For inorganic semiconductors, vertical transistors have been discussed since the late 1960s (Sze et al., 1966; Bozler, 1986; Bozler and Alley, 1980). However, vertical inorganic transistors were not taken up by the industry, and the field remains small, especially because of the incredible success of the metal oxide semiconductor fieldeffect transistor (MOSFET) and the ease of integration of these horizontal transistors. For OSCs, though, vertical approaches might present some key advantages, particularly the possibility to reach increased driving currents and switching frequencies without the necessity of a high-resolution structuring technique. A first vertical transistor was proposed by Yang and Heeger in 1994 (Yang and Heeger, 1994). In their transistor, two organic semiconducting films are connected by three electrodes, which can be denoted as emitter, base, and collector electrodes (cf. Fig. 23.1A). Charge carriers injected at the emitter electrode are accelerated by an emitter-collector potential difference. However, as proposed by the authors of this study, most of these charge carriers do not leave the device at the base electrode but are transmitted through the base and collected by the electric field of the collector electrode. Depending on

Fig. 23.1 Schematic of the forms of vertical organic transistors: (A) the organic permeable base transistor (OPBT), (B) the vertical organic field-effect transistor (VOFET), and (C) the patterned source vertical field-effect transistor (PS-VFET).

Vertical organic transistors

761

the potential of the base electrode, the current flowing from emitter to collector thus can be modulated between the off state at a base–emitter potential of zero and the on state observed if the base–emitter potential is increased. Depending on the nature of the transmission of charges across the base electrodes, different names for this transistor type were coined, which makes it difficult to obtain an overview of the complete field. Assuming a transmission mechanism based on hot charge carriers, these transistors were sometimes referred to as organic hot carrier triodes (Chao et al., 2005, 2008a; Ou et al., 2006). Additionally, names such as metal base transistors (Meruvia et al., 2005; Huang et al., 2009, 2008; Umetsu et al., 2013; Nakayama et al., 2006), static induction transistors (Wang et al., 1999; Yamauchi et al., 2008; Watanabe et al., 2007; Ohashi et al., 2006), and space-charge-limited transistors (Chao et al., 2008b, 2011; Zan et al., 2012) were used. Currently, the highest performance is observed if the base electrode is kept very thin (i.e., so thin that small holes or pores form in the layer). Often, these transistors are called organic permeable base transistors (OPBTs) (Huang et al., 2009; Klinger et al., 2015; Chen et al., 2014; Kaschura et al., 2015). However, regardless of the name, all these variants operate along the same physical principles, which are discussed in Section 23.2. In 2006 Nakamura and Kido proposed a different kind of vertical transistor (sketched in Fig. 23.1B) (Nakamura et al., 2006, 2008). In contrast to OPBTs, these transistors operate on the basis of the field effect and the general structure of an OFET (consisting of source, drain, and gate electrodes) is retained. However, in contrast to OFETs, the drain electrode is stacked on top of the source, which reduces the channel length to the thickness of the OSC. Similarly, in some transistors, the channel is formed around a step edge (Stutzmann et al., 2003; Parashkov et al., 2003, 2004; Yutani et al., 2006a), which reduces the channel length to the height of the step. The working mechanism of these transistors is discussed in Section 23.3. Finally, some transistors rely on a modulation of charge-carrier injection at a metal/ organic Schottky junction by an external electric field to form the transistor structure sketched in Fig. 23.1C (Hlaing et al., 2015; Ma and Yang, 2004; Ben-Sasson et al., 2009, 2014; Ben-Sasson and Tessler, 2012). As indicated in this figure, the source electrode is patterned so that the gate electric field can access the source/OSC interface, which motivates the designation of this type of transistors as patterned source vertical field-effect transistors (PS-VFETs). The performance and working mechanism of these transistors are summarized in Section 23.4. All of these vertical organic transistor types have benefits and shortcomings. Their working mechanisms are presented next, followed by a short summary of the bestperforming devices (cf. Table 23.1) and an outlook on future research.

23.2

Organic permeable base transistors

The general setup of an OPBT is shown in Fig. 23.1A. The two organic layers sandwiched between the three electrodes form two organic Schottky diodes connected back to back, as shown in the DC equivalent circuit (Fig. 23.2A). The upper Schottky

762

Handbook of Organic Materials for Electronic and Photonic Devices

Emitter

Collector

IEC

IE IEB

IB

IC IBC

Base

(A)

(B)

Fig. 23.2 (A) DC-equivalent circuit of the OPBT. (B) Transfer characteristics of an OPBT plotting the emitter, collector, and base current as a function of the base emitter voltage.

diode is formed between the emitter and the base electrode, whereas the lower Schottky diode is formed between the collector and the base electrode. For both Schottky diodes, the base electrode is the blocking contact. For n-type OPBTs, a positive potential VCE is applied between collector and emitter (see Fig. 23.1A). The base–emitter potential VBE varies between negative voltages and the collector potential VCE (cf. Fig. 23.2B). If VBE is negative, both diodes (i.e., the emitter–base and the base–collector diode) are operated in reverse (i.e., no charges are injected at the emitter and the transistor is turned off ). However, if VBE becomes positive, the emitter–base diode is operated in the forward direction (i.e., electrons are injected at the emitter forming the emitter current IE). At the base, most of the emitter current IE is transmitted through the base forming the transmission current IEC (the precise mechanism of transmission is discussed later in this chapter; see Fig. 23.1A for a definition of all currents), and only a smaller fraction of the current leaves the device at the base (IEB). The transmitted electrons experience a large electric field inside the base–collector diode and are finally collected at the collector. Furthermore, although the base–collector diode is operated in reverse, some electrons are still injected at the base that forms the current IBC. Overall, the experimentally accessible external currents IE, IB, and IC are connected to the internal currents IEC, IEB, and IBC as follows: 0

1 0 10 1 1 1 0 IEC IE @ IB A ¼ @ 0 1 1 A@ IEB A 1 0 1 IC IBC

(23.2)

Unfortunately, the matrix of Eq. (23.2) cannot be inverted (i.e., the internal currents cannot be calculated from the terminal currents IE, IB, and IC). Based on the definition of internal and external currents in Fig. 23.2A, the transmission α and amplification β can be defined. The transmission α is given by

Vertical organic transistors

α¼

IEC IC  IBC ¼ IE IE

763

(23.3)

Unfortunately, the base–collector current IBC is not experimentally accessible, so only approximations for the transmission α can be given. If IBC is significantly small compared to the emitter current, the transmission α becomes α
IICE . Another approximation, which is precise if IBC does not vary strongly with the base current, is the C differential transmission α
dI dIE (Fischer et al., 2012). The transmission of an OPBT is closely related to its amplification β. The amplification is defined as β¼

IC  IBC IB  IBC

(23.4)

As for the transmission, the amplification can only be approximated, assuming that the base–collector current is negligible. In this case, the transmission β becomes β
IICB or C β
dI dIB . It follows that

β¼

α 1α

(23.5)

Finally, another important performance parameter, not only of OPBTs but of all other dIC transistors as well, is the transconductance gm ¼ dV : The transconductance is closely BE related to the transit frequency of transistors (i.e., a higher transconductance relates to increased switching speeds).

23.2.1 Nature of transmission across the base electrode The nature of transmission of charge carriers through the base electrode has been examined by researchers for a while. Whereas some scholars proposed that the transmission is due to hot charge carriers that can traverse a thin but closed metallic base (Chao et al., 2005, 2008a), others assumed that small pores or holes inside the base electrode were responsible for the transmission of charge carriers (Fischer et al., 2012; Yu et al., 2014). Currently, the fact that transmission values above 99% are observed (Klinger et al., 2015; L€ ussem et al., 2014a), the observation that OPBTs can be operated in both directions (i.e., if emitter and collector are swapped) (Fischer et al., 2012), and microscopic studies (Yu et al., 2014) all support the hypothesis that the transmission is caused by small holes inside the base electrodes. Several authors tried to control the size and density of the holes inside the base electrode, such as by using a monolayer of polystyrene spheres as a deposition mask (Chao et al., 2010, 2009; Li et al., 2013). However, the holes in the base have to be smaller than the overall thickness of the OPBT in order to retain a reasonable on/off ratio, which puts a stringent requirement on the structuring technology.

764

Handbook of Organic Materials for Electronic and Photonic Devices

23.2.2 Detailed working mechanism of OPBTs To understand the details of the working mechanism of OPBTs, two- or threedimensional drift-diffusion models have been developed by Chen et al. (2014), Kaschura et al. (2016a), and L€ ussem et al. (2014a). These three models are structurally similar. The general setup of a model (L€ ussem et al., 2014a) is shown in Fig. 23.3A. The OPBT is modeled assuming an emitter electrode on top (consisting of n-doped C60), a base electrode surrounded by a perfect native oxide layer, and a collector electrode on the bottom. In this model, it is assumed that the base electrode is surrounded by native oxide layer, which is motivated by the experiments of Nakayama and coworkers (Yutani et al., 2006b; Nakayama and Yokoyama, 2007). The researchers found that exposure of an aluminum base to ambient air led to a significant increase in the amplification and on/off ratio of the device, which was explained by the formation of such a native oxide layer. Indeed, in almost all reports on OPBTs, processing of the device is stopped after deposition of the base electrode and the device is exposed to air before continuing deposition of the remaining layers (Lussem et al., 2015). In Fig. 23.3, a cylindrical symmetry is assumed (i.e., the pore in the base electrode is located at r ¼ 0 mm). The pore size in this particular case is 10 nm. Fig. 23.3A shows the potential inside the device in the off state (i.e., at VBE ¼  3 V). The potential drops homogeneously between the base and emitter electrodes and between the collector and base electrodes. Most important, the emitter seems to be completely shielded from the collector potential by the base electrode. To understand the switching into the on state, the potential along the symmetry axis (i.e., at r ¼ 0) is shown in Fig. 23.1B. The potential is plotted from the emitter on the left (i.e., at z ¼ 0) to the collector on the right (i.e., at z ¼ 0.2 μm). Several potential lines are shown, representing a base potential that decreases from 3 to 1 V (i.e., from the off state to the on state). The collector potential is kept at VCE ¼ 3 V. At VBE ¼  3 V, the potential of the base presents an effective barrier for electrons, no electrons can be injected at the emitter, and the device is turned off. If the base potential is increased to less negative or slightly positive values, the potential barrier decreases, which leads to an effective electric field at the emitter increasing the injection of electrons, and hence switching on the device. This potential barrier imposed by the base electrode can explain the fundamental switching of the OPBTs shown in Fig. 23.3C, which plots the transfer characteristics of OPBTs with increasing size of the pores. As the insulator surrounding the base electrode is assumed to be perfect, the off current is given by direct leakage from the emitter to the collector electrode. As already seen in Fig. 23.3A, the base electrode effectively screens the collector potential for small pore sizes, which leads to an off current close to the noise level. However, the screening becomes imperfect for larger pores, leading to an increase in off current. More interesting is the fact that Fig. 23.3C shows that the on current level does not depend on the pore size, which is surprising if one assumes that the emitter current of OPBT is dominated by current flowing directly from emitter to collector through the pores in the base. This numerical result correlates with experiments showing that the

Vertical organic transistors

765

Fig. 23.3 Results of a drift-diffusion model used to explain the detailed working mechanism of OPBTs. All images show a cylindrical symmetry, and the hole in the base electrode is centered on the left of the image (i.e., at r ¼ 0). (A) Potential distribution in the off state. (B) Potential along the symmetry axis (i.e., through the hole in the base). (C) Transfer characteristics obtained from the model for increasing pore size. (D) Electron distribution inside the device showing charge-carrier accumulation on top of the base. (E) Current density inside the device in the on state. (C) Reprinted with permission from L€ussem, B., et al., 2014. Beyond conventional organic transistors: novel approaches with improved performance and stability. Proc. SPIE Int. Soc. Opt. Eng. 9185, 91850H.

766

Handbook of Organic Materials for Electronic and Photonic Devices

collector current is much larger than would be expected from the overall pore size and density (Kaschura et al., 2016b). To explain this observation, the charge-carrier density and the current density in the on state of the device are plotted in Fig. 23.3D and E. As seen in Fig. 23.3D, a large number of electrons accumulate on top of the base electrode, which resembles the formation of a horizontal channel in OFETs. This channel can transport a charge toward the pores in the base electrode (cf. Fig. 23.3E). In effect, electrons are injected along the whole emitter surface—that is, the on current does not depend on the particular size of the holes in the base electrode (Kaschura et al., 2016b).

23.2.3 OPBTs with highest performance Klinger et al. demonstrated a design of OPBTs that reach the highest possible performance at the present time (Klinger et al., 2015, 2017b). The structure of these devices and a representative transfer characteristic are shown in Fig. 23.4. The device operates at low voltages (below 2 V), and reach very high on-state currents (up to 1 kA/cm2), which corresponds to a very high transconductance (approximately 8 mS (Klinger et al., 2015)). Transit frequencies above 10 MHz were measured at a collector-emitter voltage VCE ¼ 1 V, and it was argued that switching frequencies of up to 100 MHz are in reach (Table 23.1) (Klinger et al., 2015; Chen et al., 2014).

23.2.4 Current limitations of OPBTs Current OPBTs show an impressive performance. However, a practical use of OPBTs is limited by substantial base currents observed even at DC conditions. Although amplification values of above 1000 are reached (Klinger et al., 2015), relatively large

Fig. 23.4 Structure and performance of OPBTs as published by Klinger et al. (2015, 2017a). The devices show a very small subthreshold swing and very high driving currents, which relate to large transconductance values, and reach transit frequencies above 10 MHz.

Current amplification (Eq. 23.4), IICB

Subthreshold swing, S (mv/ decade)

Transit frequency (Eq. 23.1), fT (MHz)

Transconductance, gm (mS)

Author

Type

Operation voltage (largest), Vop (V)

Klinger et al. (2015) Klinger et al. (2017a)

OPBT

1.5

106

2600 (max)

102

2.2 (at 1 V)

8.16 (4  104 cm2)

OPBT

1–2

108

105 (max)

85

11.8

Approximately 70 mS (4  104 cm2, pulsed)

OPBT

5–8

VOFET

5

>106

VOFET

VG  30 V

5  105

800

VOFET

VG  10 V

105

>260

PSVOFET

VG  40 V

103

PSVOFET VOFET/ PSVOFET

VG  15 V

104

2–3 V

6  104

Yu et al. (2014) Kleemann et al. (2013) Kwon et al. (2016) G€ unther et al. (2016) Ben-Sasson and Tessler (2012) Greenman et al. (2013) McCarthy et al. (2010)

Switching ON ratio, IIOFF

or

ID IG

Vertical organic transistors

Table 23.1 Comparison between the performance of various vertical organic transistors

476

>107

0.1–0.2 mS (50 μS/ mm edge length) 5 μS/mm edge length 67 μS/mm edge length

1000 τ < 2 μs 500

767

768

Handbook of Organic Materials for Electronic and Photonic Devices

base currents lead to power losses that are not acceptable for most applications. New approaches are needed to limit the base currents. However, it has to be emphasized that in contrast to bipolar junction transistors, the base currents observed in OPBT are not inherent to their working mechanism (i.e., they represent a parasitic leakage current that can be completely avoided, at least in theory). Additionally, OPBTs are currently limited by the need for short exposure of the device to ambient air after depositing the base electrode, which is commonly done to oxidize the base electrode. This exposure step leads to a degradation of the charge-carrier mobility inside the n-type OSC (e.g., C60)—that is, a charge-carrier mobility in the range of only 102 cm2/V s is observed. Charge-carrier mobilities in the range of 1 cm2/V s, usually observed in C60-based OFETs, would increase the saturation currents of OPBTs, and then also their transconductance and transit frequencies, by another two orders of magnitude.

23.3

Vertical organic field-effect transistors

In contrast to OPBTs, VOFETs rely on the accumulation of charge carriers at an insulator–semiconductor junction very similar to standard OFETs. However, in contrast to standard, horizontal OFETs, the predominant charge transport is orthogonal to the substrate (e.g., along a step edge between vertically stacked source and drain electrode or toward a drain electrode deposited on top of the channel material, as sketched in Fig. 23.1B and C, respectively). In a prototypical vertical OFET, from bottom to top, the transistor consists of a gate electrode, a gate insulator, a layer of an OSC, a source electrode, a second layer of an OSC, and finally a drain electrode (cf. Fig. 23.1B). The working mechanism of this type of transistor has been studied by various authors based on 2D drift diffusion models (Kwon et al., 2016; Greenman et al., 2017; Sheleg et al., 2017). Representative results of a similar p-type transistor model are shown in Fig. 23.5. The structure of the model is shown in Fig. 23.5A. In comparison to Fig. 23.1B, the gate electrode is located on top of the device and the drain electrode on the bottom. As can be seen in Fig. 23.5A, an insulator is added between source and drain, which covers the edge of the source electrode as well. This insulator is included in almost all reports of vertical OFETs (Kleemann et al., 2013; Kwon et al., 2016; She et al., 2017) to shield the source electrode from the electric field of the drain electrode, which otherwise would lead to a strong direct current flowing from the source to the drain, which cannot be controlled by the gate (Greenman et al., 2017; She et al., 2017; Lee et al., 2017). With a negative voltage applied to the gate, hole carriers are accumulated at the gate-oxide/OSC interface (cf. Fig. 23.5B, which shows the hole concentration between the source and gate electrode—that is, inside the area marked by a circle in Fig. 23.5A). This layer of accumulated charges forms a sheet of high conductivity, allowing the transistor to turn on. As shown in Fig. 23.5C, current is injected at the side

Vertical organic transistors

769

Fig. 23.5 Operation of the VOFET. (A) Structure of the 2D drift-diffusion model, including the electrodes (source, drain, and gate), the gate oxide, and the OSC (pentacene). (B) Hole accumulation at the insulator/semiconductor interface at negative voltage applied to the gate (VGS ¼  4 V). The image shows an enlarged part of the device indicated by the circle in (A). (C) Current density inside the transistor. Current is injected at the source, transported toward the accumulated channel of holes at the insulator/semiconductor interface, and finally drained by the electric field of the drain electrode (VDS ¼  10 V).

of the source facing the gate. This current flows vertically toward the gate-oxide/ semiconductor interface. Once it reaches the gate oxide, it flows horizontally along the interface until it reaches the edge of the source electrode, where it is exposed to the electric field imposed by the drain electrode. If this drain electric field is large enough, almost all current flows vertically toward the drain (i.e., along the shortest way toward the drain electrode). As seen in Fig. 23.5, the transistor combines sections of horizontal and vertical charge transport. Ways to minimize horizontal transport and to increase the performance of these transistors were studied by She et al. (2017). They found that vertical transport dominated for large drain potentials, which leads to a fast response of their device (in this case, a nonvolatile memory cell). However, in some cases, particularly for light-emitting transistors discussed in Section 23.5, a more horizontal current distribution might be beneficial.

770

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 23.6 VOFET as presented by Kleemann et al. (2013). The vertical transistor (shown as closed symbols in A) shows an increased performance compared to the horizontal OFET measured between the two source electrodes on the bottom, as shown in (B). Reprinted with permission from Kleemann, H., G€unther, A.A., Leo, K., L€ ussem, B., 2013. Highperformance vertical organic transistors. Small 9(21), 3670–3677.

23.3.1 Vertical organic field-effect transistors with highest performance Several vertical OFETs with excellent performance have been reported in the past couple of years (Kleemann et al., 2013; Kwon et al., 2016; Greenman et al., 2017; Sheleg et al., 2017; She et al., 2017; Lee et al., 2017). One of the best results is shown in Fig. 23.6 (Kleemann et al., 2013). By an optimization of this structure, the vertical transistor displays significantly increased current compared to comparable horizontal transistors. On/off ratios in the order of five to six orders of magnitude and a transconductance in the range of 67 μS/mm have been reported in recent publications (G€ unther et al., 2016).

23.3.2 Current limitations of vertical organic field-effect transistors As reported by G€ unther et al. (2016), vertical OFETs are currently limited by common short channel effects, in particular by a dominance of the contact resistance. Following the example of horizontal OFETs, doping was introduced at the source electrodes to form an Ohmic contact and minimize injection losses (G€unther et al., 2016), which resulted in an increase in device performance. Similarly, due to the short device channel, it is difficult to reach saturation in the output characteristic of vertical OFETs, which is indispensable for reliable circuit design. The loss in saturation can be explained by the fact that the source electrode is not completely screened from the electric field due to the potential applied to the drain electrode, even if the edge of the source electrode is covered by an insulator (Greenman et al., 2017; Sheleg et al., 2017). Tessler et al. proposed the use of an

Vertical organic transistors

771

additional electrode shielding the source and were able to show in their model that such a shield would improve the saturation of these devices in the future (Sheleg et al., 2017).

23.4

Patterned source vertical field-effect transistor

In this section, we describe the operation of the patterned source VFET (PS-VFET) (Ben-Sasson et al., 2014), also known as the Schottky barrier VFET (SB-VFET) (Lussem et al., 2015), Fig. 23.7 shows three-dimensional (3D) schematics of both lateral (A) and vertical (B) metal-oxide-semiconductor (MOS)–OFETs. The VFET is a patterned source type that is essentially composed of two parts stacked together. The first is a diodelike structure composed of a source, a semiconductor, and a drain. The second is a gate capacitor, which is introduced by placing the gate below the source and separating the two by an insulating layer. As depicted in Fig. 23.7B, it is essential that the source electrode is noncontinuous or perforated. The material used to form the source may vary from metal films (Ma and Yang, 2004; Ben-Sasson and Tessler, 2011a), carbon nanotubes (CNTs) (McCarthy et al., 2010; Liu et al., 2008), graphene (Lemaitre et al., 2012), and metallic nanowires (Ben-Sasson et al., 2015), to porous indium tin oxide (ITO) (Yu et al., 2016a). The noncontinuous or perforated metal film can be achieved by evaporating a thin granular film (Ma and Yang, 2004; Xu et al., 2007), or by lift-off, such as by using block copolymer lithography (Ben-Sasson et al., 2009; Ben-Sasson and Tessler, 2012; Greenman et al., 2013) or photolithography (Greenman et al., 2016). Before we delve into the operation mechanisms governing the PS-VFET, we need to mention that while the structure in Fig. 23.7B represents the earliest structures, it is by no means the latest or the most advanced. Fig. 23.8 shows the cross section of PS-VFET structures that have evolved over the years, with Fig. 23.8A being the cross section of Fig. 23.7B. We start by considering the simplest structure, Fig. 23.8A, and later in this chapter, we will discuss the improvements introduced by the more advanced structures.

Fig. 23.7 Illustration of MOS-OFETs. The labels S, G, and D indicate the source, gate, and drain electrodes, respectively. The blue and green regions indicate the insulating and the semiconductor layers, respectively. Electrodes are indicated with yellow and gray. (A) Lateral transistor with bottom gate and bottom contact configuration. (B) PS-VFET. Reprinted with permission from Ben-Sasson, A., Tessler, N., 2012. Unraveling the physics of vertical organic field effect transistors through nanoscale engineering of a self-assembled transparent electrode. Nano Lett. 12, 4729–4733. Copyright 2012 American Chemical Society.

772

Handbook of Organic Materials for Electronic and Photonic Devices Drain

VDS VGS

sou

Drain

Drain

Drain

rce

Gate dielectric Gate

(A)

Gate dielectric Gate

(B)

Gate dielectric Gate

(D)

(C)

Gate dielectric Gate

Fig. 23.8 Illustration of the designs for PS-VOFET that have evolved over the years. The patterned gray layers in (B), (C), and (D) represent the source insulator. Drain

Drain

VDS

VDS sou

(A)

rce

VGS

(B)

sou

rce

VGS

(C)

Fig. 23.9 Illustration of the parts that play a role in the transistor operation. (A) The sourcedrain diode; vertical arrows depict the path of the leakage current. (B) The source perforation presented as a bottom-gate, bottom-contact structure. (C) The full structure, in which the channel shown in (B) serves as a virtual electrode for the diode.

Fig. 23.9A shows the upper part of the PS-VFET, which has a form of a diode. For a good transistor operation, in the absence of a gate bias, the source-semiconductordrain diode is supposed to remain off even when the drain potential VDS is applied. In the current device structure, it implies that the source electrode has to form a Schottky contact with a relatively large barrier for injection, which justifies an alternative name for this structure—the SB-VFET (Lussem et al., 2015). The current flow through this contact-limited diode would form the transistor’s leakage current. For the structure in Fig. 23.9A, the off current would mainly flow from the top surface of the source electrode (as shown by the vertical arrows). In Fig. 23.9B, we consider the lower part of the PS-VFET, which resembles a bottom-gate, bottom-contact lateral FET. As we have already determined that the contact must have an injection barrier, this would be a Schottky-contact FET (Sze and Ng, n.d.). Following the standard operation of lateral FET, when a gate bias is applied, charges will be drawn into the semiconductor and charge the gate capacitance, thus forming the channel (Fig. 23.9B). Once the lateral channel is formed, it can serve as a virtual contact for the diode, allowing the current to flow vertically (Fig. 23.9C). For a high-enough gate bias, the virtual contact would behave as an Ohmic contact, and the diode current flow would be a space-charge-limited current (SCLC) with a ID ∝ V2DS dependence (Ben-Sasson and Tessler, 2011a, b). An alternative way to consider the switching of the PS-VFET from contact limited to bulk limited is to examine the actual contact barrier between the source electrode and the semiconductor. While it is initially high, applying a gate bias can induce a potential lowering of the barrier, thus allowing higher currents to flow. The mechanism by which the barrier is lowered depends on the material forming the perforated

Vertical organic transistors

773

Fig. 23.10 Charge-density distribution overlaid on the PS-VFET structure. (A) A global view showing the formation of the vertical channel (VDS ¼ VGS ¼ 5 V). (B) A zoom into one of the edges of the source.

source electrode. If it is metallic, the mechanism is the field-induced lowering of the barrier (Sze and Ng, n.d.; Ben-Sasson and Tessler, 2011b). If it is composed of less dense materials, such as CNTs or graphene, then a gate field can modulate the Fermi level of the contact, thus providing a different mechanism for controlling the contactsemiconductor barrier (Chen et al., 2013). To support this discussion of the mechanism governing the first-generation PS-VFETs operation, simulation results of the charge-density distribution, current distribution, and electric-field distribution, for such transistor, are presented in Fig. 23.10. Fig. 23.10A shows the 2D distribution of the charge density (log scale) overlaid on the PS-VFET structure (the bottom-gate insulator and gate are not shown). The results are plotted for VDS ¼ VGS ¼ 5 V. The main feature shown here is the vertical channel that occupies most of the hole (perforation) in the source electrode. We note the high carrier density at the OSC/gate-insulator interface, which supports the notion that a virtual electrode was created and supplies the current to the vertical channel. Fig. 23.10B zooms on the source-edge region close to the perforation. We note that there is a range over which the charge density builds toward the central part of the vertical channel. For the current device structure, if the source is 20 nm thick, a distance about 100 nm from the edges of the source needs to be secured in order for the vertical channel to form. For a structure with a 2-μm opening in the source, the loss of about 100 nm is not critical. However, if the hole is made smaller and on the order of 100 nm, this effect may shut down the PS-VFET. This was dubbed the tunnel effect, and it was shown that for an 80-nm opening, the source thickness has to be below 10 nm (Ben-Sasson and Tessler, 2012).

774

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 23.11 Current distribution overlaid on the PS-VFET structure. (A) A global view of the device and the vertical channel (VDS ¼ VGS ¼ 5 V). (B) A zoom close to the source edge and for VDS ¼ 7 V.

Based on the same simulation used in Fig. 23.10, the current distribution overlaid on the PS-VFET structure was also investigated, as shown in Fig. 23.11A. These graphs show the distributed nature of the VFET structure, which is pronounced when the source perforation is significantly larger than the device thickness. Close to the insulator interface (bottom part), the current is primarily composed of lateral current that flows from the source edges toward the center of the hole. Due to the symmetry of the structure, this lateral current has to be zero at the center. The gradual change in the current density toward the middle of the hole supports the idea that the current is divided into two vertical and lateral parts. In fact, if one carefully examines the current at the middle of the device, it seems that the current is not uniformly distributed and that it is stronger near the source edge. This nonuniformity is simply because the drain is pulling the charges upward before they reach the center of the perforation. The strength of this effect generally depends on the PS-VFET relative dimensions, as well as the drain bias. At a higher VDS, the effect would be stronger and the vertical channel would split in two, with most of the current flowing upward close to the source edges. To demonstrate this effect, we show a higher drain bias of 7 V in Fig. 23.11B. Fig. 23.11B, which zooms into the edge region of the source, illustrates that the vertical channel indeed splits at a high VDS. Moreover, it indicates that the current is being injected by the side edge of the source contact. In Fig. 23.12, we present the electric-field lines for the same working conditions as for Figs. 23.10 and 23.11. As the transistor switching relies on the injection barriers lowering at the source, the electric field plays a major role, and we can use Fig. 23.12 to demonstrate the PS-VFET operation. Close to the source edge, we can clearly see that the field lines are emanating from the gate, with the strongest being directed toward the edge of the source. This field drives the lowering of the barrier, which

Vertical organic transistors

775

Fig. 23.12 Electric field lines close to the source edge for the same operating conditions as in Fig. 23.10 (VDS ¼ VGS ¼ 5 V).

is commonly described using the formula for the image force lowering (Sheleg et al., 2017; Sze and Ng, n.d.): rffiffiffiffiffiffiffiffiffiffiffiffiffiffi qE? Δ¼ 4πεO εS The top surface of the source shows field lines that are due to the drain source bias. This field draws current from the source even when the gate is off. Namely, the top surface constitutes a leakage current path. A closer look at the field lines suggests that the drain also has an effect on the electric field at the vertical edge of the source. In the ideal case, the drain does not affect the injection from the source. If it does, the effect is called drain-induced barrier lowering (DIBL). Note that the DIBL not only affects the leakage current, but also prevents the device from entering saturation in the output characteristics. Indeed, most of the experimental reports show little to no saturation. With this explanation in mind, one can understand the motivation of moving to structures shown in Fig. 23.8B and C. In order to insulate the source leakage path but still draw current, the injection from the source has to be changed from one similar to bottom-gate, bottom-contact (Fig. 23.8A) to that of bottom-gate, top-contact (Fig. 23.8C) (Kleemann et al., 2013; Greenman et al., 2017). In fact, the changes leading to the structure shown in Fig. 23.8B and C make the PS-VFET very similar to the VOFET structure discussed in Section 23.3, indicating that the two lines of vertical transistor work will merge in the future. Examining Fig. 23.8C, it seems that there is no facet that is exposed to the draininduced field; hence, the DIBL is expected to be reduced, leading to ideal saturation characteristics. However, an analysis of the injection from bottom-gate, top-contact transistors (Tessler and Roichman, 2001) has shown that a good contact will inject

776

Handbook of Organic Materials for Electronic and Photonic Devices

Drain current (A/cm2)

0.24 0.2

0.12

0.008

0.006

0.006

0.004

0.004

0.002

0.002

Vg = 5 V Vg = 4 V Vg = 3 V Vg = 2 V Vg = 1 V Vg = 0 V

0.08 0.04 0

(A)

Vg = 5 V Vg = 4 V Vg = 3 V Vg = 2 V Vg = 1 V Vg = 0 V

0.16

0.008

0

2

4

6

8

Drain voltage (V)

10

0

(B)

0

2

4

6

Drain voltage (V)

8

10

0

(C)

0

5

10

15

20

Drain voltage (V)

Fig. 23.13 Output characteristics of the device structures shown in Fig. 23.8A (A), Fig. 23.8C (B), and Fig. 23.8D (C). Note that the vertical axis for (A) is different here.

only from the very edge that is close to the channel. As the contact barrier increases, the injection spreads to take a larger part of the contact’s bottom facet, not just from its edge. Unfortunately, while the bottom facet of the source should not be affected by the drain, this is not true for the very edge. Namely, this structure still contains the need to trade high current (low barrier) with quality of saturation (high barrier). Having identified the edge being susceptible to DIBL, it was suggested to insulate the very edge (Sheleg et al., 2017), as illustrated in Fig. 23.8D. To provide comparison of the various structures, the output characteristics of the devices, while keeping all other parameters equal, are presented in Fig. 13. Fig. 23.13A shows the output characteristics of the structure shown in Fig. 23.8A. In the ideal case, this structure should exhibit SCLC with V2DS dependence (Ben-Sasson and Tessler, 2011a, b). Indeed, the curves follow a power law of 2 at first, but with the DIBL kicking in, the slope becomes slightly higher. Fig. 23.13B shows the output characteristics of the structure shown in Fig. 23.8C. Two effects are noted when comparing to Fig. 23.13A. First, saturation is now visible; and second, the currents are about two orders of magnitude lower. The saturation implies that the current is largely limited by the amount of current drawn from the source by the gate. As has been discussed earlier in this chapter, the drain has relatively little effect due to the high injection barrier (0.7 eV). The relatively lower current is directly associated with the high barrier, as well as with the fact that charges are accumulating below the contact screening the gate field, thus limiting the gate-induced barrier-lowering effect. Hence, this structure exhibits a trade-off between quality of saturation and maximum current density. Fig. 23.13C shows that adding a shield to the source edge does not affect the current level. However, there is now no dependence on the drain bias, suggesting that one can use the reduced barrier height to enhance the current without compromising the level of saturation. Using such a design could allow one to bring the current above 1 A/cm2 while maintaining the saturation quality (Sheleg et al., 2017). To summarize this section, the performance of PS-VFET (or SB-VFET) depends on the relative dimension chosen for the various parts, which are source thickness, perforation size, total semiconductor thickness, and thickness of the layer below the source. As in most of the structures, the switching relies on lowering of the

Vertical organic transistors

777

injection barrier, and one has to keep in mind that the mechanism of barrier lowering may depend on the type of material making up the electrode (most of the discussion here dealt with a metallic source). The importance of the material and device structure parameters might have been the reason it took a while for this structure to become popular, but to date, several research groups have reported a wide range of materials and structures. This also includes various integrations with light emitting diodes.

23.5

Vertical organic light-emitting transistors

Organic light-emitting transistors (OLETs) are emerging multifunctional optoelectronic devices that combine light emissions with electrical switching. Typical OLETs have been developed in a planar geometry adopting a configuration of horizontal transistors. In this structure, both holes and electrons are injected into either a single ambipolar OSC layer or a multitude of layers from the source and drain contacts, and the radiative recombination occurs within the channel region (Muccini, 2006; Cicoira and Santato, 2007; Muccini et al., 2012). Horizontal OLETs are particularly attractive from a scientific point of view because they allow one to observe and image the position of charge recombination directly, which is greatly favorable to investigate the charge-carrier dynamics in OSCs, such as charge injection, charge transport, and electroluminescence (EL). However, intrinsic line or band-type light emission characteristics most likely prevent the horizontal OLETs from satisfying the requirements of commercial applications, such as display or lighting, where surface emission is essential. On the other hand, vertical OLETs (VOLETs) combine vertical transistors with organic light-emitting diodes (OLEDs) to form vertically stacked devices in which the light emission features resemble OLEDs. A basic device architecture can be regarded as an OLED stack sandwiched between the OSC and the drain electrode of any of the vertical transistors shown in Fig. 23.1. In this case, the source and drain electrodes of vertical transistors correspond to the anode and cathode electrodes of OLEDs (or vice versa) depending on the polarity of majority charge carriers of the underlying OSC and the layer stack of the OLED part. The fundamental current modulation mechanism is basically the same as in the vertical transistors described in Section 23.2.4. The primary benefits of this vertical approach are sufficiently large driving currents at relatively low voltage and with low power consumption, accomplished by taking advantage of the strengths of the vertical transistor. When it comes to active matrix circuits, VOLETs do not need an additional pixel area for the driving transistors due to their vertically stacked architecture, and therefore a high aperture ratio (i.e., the ratio of effective emissive area to whole pixel area) can be achieved compared to their respective OLEDs. In addition, the flexibility of designing the architecture and being able to use a wide range of materials for each layer allows one to yield many types of functioning devices. Furthermore, the device stack can be easily integrated and the light emission characteristics can be tuned in terms of color, EL spectrum, and outcoupling efficiency, depending on the optical cavity structure.

778

Handbook of Organic Materials for Electronic and Photonic Devices

The progress in research on VOFETs has led to more extensive studies on VOLETs as well. The Kudo group proposed several types of vertically stacked OLETs—namely, static-induction OLETs (Ohashi et al., 2006; Kudo, 2005) (cf. Section 23.2) and metal–insulator–semiconductor OLETs (Nakamura et al., 2006) (cf. Section 23.3). Xu et al. reported OPBT-type OLETs, where a thin and rough source electrode is situated in the middle of the stack (Xu et al., 2007). Considering that these transistors are at an early stage of development, they showed promising results for potential organic display applications, but the on/off ratios were limited to 5  103) (Lee et al., 2017), as well as drain current saturation in the output characteristics (Greenman et al., 2017), as discussed in Section 23.4. In these studies, the experimental results had good agreement with 2D simulation of modeled VOLET structures, which allows one to understand the charge dynamics in the devices in more detail, and thus can assist in the design of an optimum device stack for the high brightness and high efficiency of VOLETs.

23.6

Outlook

Vertical organic transistors are promising alternatives to conventional lateral fieldeffect transistors and hold the potential to increase the performance of organic transistors beyond the limits otherwise set by the relatively low charge-carrier mobility of OSCs. Several approaches were discussed in this chapter, particularly OPBTs, VOFETs, and PS-FETs. These three categories of vertical organic transistors have received increasing attention in the scientific literature, which has led to a better understanding of their working mechanisms. Although the results discussed here are promising, vertical organic transistors still face obstacles that must be solved before they can be used realistically in larger integrated circuits. OPBTs show extremely large transconductance values and an intriguing simple device setup, but currently they also suffer from sizable base currents, leading to unacceptably large static-power losses. Vertical OFETs, on the other hand,

Vertical organic transistors

779

do not show significant gate leakage, but they are limited by their contact resistance, and in some cases a lack of saturation in the output characteristic as well. However, despite these challenges, VFETs retain great potential for other thin-film transistor technologies. In particular, the possibility of combining these transistors with other functionalities (e.g., light emission or absorption) offers significant benefits and leads to interesting scientific questions.

Acknowledgments The authors acknowledge funding from the Binational Science Foundation (Grant No. 2014396) and the National Science Foundation (Grant No. 1639073).

References Ante, F., et al., 2012. Contact resistance and megahertz operation of aggressively scaled organic transistors. Small 8 (1), 73–79. Barbaro, M., Caboni, A., Cosseddu, P., Mattana, G., Bonfiglio, A., 2010. Active devices based on organic semiconductors for wearable applications. IEEE Trans. Inf. Technol. Biomed. 14 (3), 758–766. Ben-Sasson, A., Tessler, N., 2012. Unraveling the physics of vertical organic field effect transistors through nanoscale engineering of a self-assembled transparent electrode. Nano Lett. 12, 4729–4733. A. J. Ben-Sasson and N. Tessler, “Patterned electrode vertical OFET: analytical description, switching mechanisms, and optimization rules,” Proceedings of SPIE, vol. 8117, p. 81170Z, 2011a. Ben-Sasson, A.J., Tessler, N., 2011b. Patterned electrode vertical field effect transistor: theory and experiment. J. Appl. Phys. 110(4), 044501. Ben-Sasson, A.J., et al., 2009. Patterned electrode vertical field effect transistor fabricated using block copolymer nanotemplates. Appl. Phys. Lett. 95(21), 213301. Ben-Sasson, A.J., Greenman, M., Roichman, Y., Tessler, N., 2014. The mechanism of operation of lateral and vertical organic field effect transistors. Isr. J. Chem. 54, 568. Ben-Sasson, A.J., Azulai, D., Gilon, H., Facchetti, A., Markovich, G., Tessler, N., 2015. Selfassembled metallic nanowire-based vertical organic field-effect transistor. ACS Appl. Mater. Interfaces 7, 2149. Berggren, M., Richter-Dahlfors, A., 2007. Organic bioelectronics. Adv. Mater. 19 (20), 3201–3213. Bittle, E.G., Basham, J.I., Jackson, T.N., Jurchescu, O.D., Gundlach, D.J., 2016. Mobility overestimation due to gated contacts in organic field-effect transistors. Nat. Commun. 7, 1–7. Bonfiglio, A., et al., 2005. Organic field effect transistors for textile applications. IEEE Trans. Inf. Technol. Biomed. 9 (3), 319–324. Bozler, C.O., 1986. Current prospects for the permeable base transistor. Surf. Sci. 174 (1–3), 487–500. Bozler, C.O., Alley, G.D., 1980. Fabrication and numerical simulation of the permeable base transistor. IEEE Trans. Electron Devices 27 (6), 1128–1141. Chao, Y.-C., Yang, S.-L., Meng, H.-F., Horng, S.-F., 2005. Polymer hot-carrier transistor. Appl. Phys. Lett. 87(25), 253508.

780

Handbook of Organic Materials for Electronic and Photonic Devices

Chao, Y.-C., Xie, M.-H., Dai, M.-Z., Meng, H.-F., Horng, S.-F., Hsu, C.-S., 2008a. Polymer hotcarrier transistor with low bandgap emitter. Appl. Phys. Lett. 92(9), 093310. Chao, Y.-C., Meng, H.-F., Horng, S.-F., Hsu, C.-S., 2008b. High-performance solutionprocessed polymer space-charge-limited transistor. Org. Electron. 9 (3), 310–316. Y.-C. Chao et al., A 1-V operated polymer vertical transistor with high on/off current ratio. IEEE Electron Devices Meeting (IEDM), 2009, pp. 1–4. Chao, Y.-C., et al., 2010. Polymer space-charge-limited transistor as a solid-state vacuum tube triode. Appl. Phys. Lett.. 97(22)223307. Chao, Y.-C., Niu, M.-C., Zan, H.-W., Meng, H.-F., Ku, M.-C., 2011. High-mobility polymer space-charge-limited transistor with grid-induced crystallinity. Org. Electron. 12 (1), 78–82. Chen, W., Rinzler, A., Guo, J., 2013. Computational study of graphene-based vertical field effect transistor. J. Appl. Phys. 113(9), 094507. Chen, W., So, F., Guo, J., 2014. intrinsic delay of permeable base transistor. J. Appl. Phys. 116 (4), 44505. Cicoira, F., Santato, C., 2007. Organic light emitting field effect transistors: advances and perspectives. Adv. Funct. Mater. 17 (17), 3421–3434. Ebisawa, F., Kurokawa, T., Nara, S., 1983. Electrical-properties of polyacetylene polysiloxane interface. J. Appl. Phys. 54 (6), 3255–3259. Fischer, A., et al., 2012. Bidirectional operation of vertical organic triodes. J. Appl. Phys. 111 (4), 044507. Greenman, M., Ben-Sasson, A.J., Chen, Z., Facchetti, A., Tessler, N., 2013. Fast switching characteristics in vertical organic field effect transistors. Appl. Phys. Lett. 103(7), 073502. Greenman, M., Yoffis, S., Tessler, N., 2016. Complementary inverter from patterned source electrode vertical organic field effect transistors. Appl. Phys. Lett. 108, 043301. Greenman, M., Sheleg, G., Keum, C.-M., Zucker, J., Lussem, B., Tessler, N., 2017. Reaching saturation in patterned source vertical organic field effect transistors. J. Appl. Phys. 121(20). Gualandi, I., Marzocchi, M., Achilli, A., Cavedale, D., Bonfiglio, A., Fraboni, B., 2016. Textile organic electrochemical transistors as a platform for wearable biosensors. Sci. Rep. 6 (1), 33637. G€ unther, A.A., Sawatzki, M., Forma´nek, P., Kasemann, D., Leo, K., 2016. Contact doping for vertical organic field-effect transistors. Adv. Funct. Mater. 26, 768–775. Hlaing, H., Kim, C.-H., Carta, F., Nam, C.-Y., Barton, R.A., Petrone, N., Hone, J., Kymissis, I., 2015. Low-voltage organic electronics based on a gate-tunable injection barrier in vertical graphene-organic semiconductor heterostructures. Nano Lett. 15 (1), 69–74. https://doi. org/10.1021/nl5029599. Huang, J., Yi, M., Ma, D., H€ummelgen, I.A., 2008. Vertical structure p-type permeable metal-base organic transistors based on {N,N0 -diphentyl-N,N0 -bis(1-naphthylphenyl)1,10 -biphenyl-4,40 -diamine}. Appl. Phys. Lett. 92(23), 232111. Huang, J., Yi, M., H€ummelgen, I.A., Ma, D., 2009. Ambipolar permeable metal-base transistor based on NPB/C60 heterojunction. Org. Electron. 10 (1), 210–213. Kaschura, F., Fischer, A., Kasemann, D., Leo, K., L€ussem, B., 2015. Controlling morphology: a vertical organic transistor with a self-structured permeable base using the bottom electrode as seed layer. Appl. Phys. Lett. 107(3). Kaschura, F., et al., 2016a. Operation mechanism of high performance organic permeable base transistors with an insulated and perforated base electrode. J. Appl. Phys. 9, 120. Kaschura, F., et al., 2016b. Operation mechanism of high performance organic permeable base transistors with an insulated and perforated base electrode. J. Appl. Phys. 120(9).

Vertical organic transistors

781

C. M. Keum, I. H. Lee, S. H. Lee, G. J. Lee, M. H. Kim, and S. D. Lee, “Quasi-surface emission in vertical organic light-emitting transistors with network electrode,” Opt. Express, vol. 22, no. 12, pp. 14750–14,756, 2014. Kitamura, M., Arakawa, Y., 2011. High current-gain cutoff frequencies above 10 MHz in n-channel C 60 and p-channel Pentacene thin-film transistors. Jpn. J. Appl. Phys. 50, 01BC01. Klauk, H., 2010. Organic thin-film transistors. Chem. Soc. Rev. 39 (7), 2643–2666. Kleemann, H., G€unther, A.A., Leo, K., L€ussem, B., 2013. High-performance vertical organic transistors. Small 9 (21), 3670–3677. Klinger, M.P., et al., 2015. Advanced organic permeable-base transistor setup with superior performance. Adv. Mater. 27 (47), 7734–7739. Klinger, M.P., et al., 2017a. Organic power electronics: transistor operation in the kA/cm2 regime. Sci. Rep. 7, 3–8. Klinger, M.P., et al., 2017b. Organic power electronics: transistor operation in the kA/cm2 regime. Sci. Rep. 7, 3–8. Kudo, K., 2005. Organic light emitting transistors. Curr. Appl. Phys. 5 (4), 337–340. Kwon, H., Kim, M., Cho, H., Moon, H., Lee, J., Yoo, S., 2016. Toward high-output organic vertical field effect transistors: key design parameters. Adv. Funct. Mater. 26 (38), 6888–6895. Lee, B.H., Bazan, G.C., Heeger, A.J., 2016. Doping-induced carrier density modulation in polymer field-effect transistors. Adv. Mater. 28 (1), 57–62. Lee, S., et al., 2017. Vertical organic light-emitting transistor showing a high current on/off ratio through dielectric encapsulation for the effective charge pathway. J. Appl. Phys. 121, 024502. Lemaitre, M., et al., 2012. Improved transfer of graphene for gated Schottky-junction, vertical, organic, field-effect transistors. ACS Nano 6 (10), 9095–9102. Li, C.-H., et al., 2013. Achieving saturation in vertical organic transistors for organic lightemitting diode driving by nanorod channel geometric control. Appl. Phys. Lett. 102 (16), 163305. Lin, H.C., Zan, H.W., Chao, Y.C., Chang, M.Y., Meng, H.F., 2015. Review of a solutionprocessed vertical organic transistor as a solid-state vacuum tube. Semicond. Sci. Technol. 30 (5), 1–14. Liu, B., et al., 2008. Carbon-nanotube-enabled vertical field effect and light-emitting transistors. Adv. Mater. 20 (19), 3605–3609. L€ ussem, B., et al., 2014. Beyond conventional organic transistors: novel approaches with improved performance and stability. Proc. SPIE Int. Soc. Opt. Eng. 9185, 91850H. Lussem, B., et al., 2015. Vertical organic transistors. J. Phys. Condens. Matter 27 (44), 443003. L€ ussem, B., et al., 2016. Doped organic transistors. Chem. Rev. 116 (22), 13714–13751. Ma, L.P., Yang, Y., 2004. Unique architecture and concept for high-performance organic transistors. Appl. Phys. Lett. 85 (21), 5084–5086. McCarthy, M.A., Liu, B., Rinzler, A.G., 2010. High current, low voltage carbon nanotube enabled vertical organic field effect transistors. Nano Lett. 10 (9), 3467–3472. McCarthy, M.A., et al., 2011. Low-voltage, low-power, organic light-emitting transistors for active matrix displays. Science 332 (6029), 570–573. Meruvia, M.S., Benvenho, A.R.V., H€ummelgen, I.A., Pasa, A.A., Schwarzacher, W., 2005. Pseudo-metal-base transistor with high gain. Appl. Phys. Lett. 86(26), 263504. Muccini, M., 2006. A bright future for organic field-effect transistors. Nat. Mater. 5 (8), 605–613.

782

Handbook of Organic Materials for Electronic and Photonic Devices

Muccini, M., Koopman, W., Toffanin, S., 2012. The photonic perspective of organic lightemitting transistors. Laser Photon. Rev. 6, 258–275. Nakamura, K., Hata, T., Yoshizawa, A., Obata, K., Endo, H., Kudo, K., 2006. Metal-insulatorsemiconductor-type organic light-emitting transistor on plastic substrate. Appl. Phys. Lett. 89(10), 103525. Nakamura, K., Hata, T., Yoshizawa, A., Obata, K., Endo, H., Kudo, K., 2008. Improvement of metal–insulator–semiconductor-type organic light-emitting transistors. Jpn. J. Appl. Phys. 47 (3), 1889–1893. K. Nakayama and M. Yokoyama, “Very high on/off ratio in vertical-type metal-base organic transistors,” in Materials Research Society Symposium Proceedings, 2007, vol. 965. Nakayama, K., Fujimoto, S., Yokoyama, M., 2006. High-current and low-voltage operation of metal-base organic transistors with LiF /Al emitter. Appl. Phys. Lett. 88(15), 153512. Ohashi, N., Nakamura, M., Muraishi, N., Sakai, M., Kudo, K., 2006. Fabrication and device simulation of single nano-scale organic static induction transistors. IEICE Trans. 89 (12), 1765–1770. Ou, T.-M., et al., 2006. All-organic hot-carrier triodes with thin-film metal base. Appl. Phys. Lett. 89(18), 183508. Owens, R.M., Malliaras, G.G., 2010. Organic electronics at the Interface with biology. MRS Bull. 35 (06), 449–456. Parashkov, R., et al., 2003. Vertical channel all-organic thin-film transistors. Appl. Phys. Lett. 82 (25), 4579–4580. Parashkov, R., et al., 2004. Organic vertical-channel transistors structured using excimer laser. Appl. Phys. Lett. 85 (23), 5751. She, X.J., Gustafsson, D., Sirringhaus, H., 2017. A vertical organic transistor architecture for fast nonvolatile memory. Adv. Mater. 29(8). Sheleg, G., Greenman, M., Lussem, B., Tessler, N., 2017. Removing the current-limit of vertical organic field effect transistors. J. Appl. Phys. 122(19). Sirringhaus, H., 2014. 25th anniversary article: organic field-effect transistors: the path beyond amorphous silicon. Adv. Mater. 26 (9), 1319–1335. Steudel, S., et al., 2012. Design and realization of a flexible QQVGA AMOLED display with organic TFTs. Org. Electron. 13 (9), 1729–1735. Stutzmann, N., Friend, R.H., Sirringhaus, H., 2003. Self-aligned, vertical-channel, polymer field-effect transistors. Science 299 (5614), 1881–1884. S. M. Sze and K. K. Ng Physics of Semiconductor Devices, 3rd ed. John Wiley Sons New Jersey, 682. Sze, S.M., Crowell, C.R., Carey, G.P., LaBate, E.E., 1966. Hot-electron transport in semiconductor-metal-semiconductor structures. J. Appl. Phys. 37 (7), 2690–2695. Tessler, N., Roichman, Y., 2001. Two-dimensional simulation of polymer field-effect transistor. Appl. Phys. Lett. 79, 2987. Umetsu, K., Akiba, R., Nakayama, K., Kido, J., 2013. Polymer material dependence in the polymer/small molecule metal-base organic transistors. Mol. Cryst. Liq. Cryst. 580 (1), 117–124. Wang, D.X., Tanaka, Y., Iizuka, M., Kuniyoshi, S., Kudo, K., Tanaka, K., 1999. Device characteristics of organic static induction transistor using copper phthalocyanine films and {Al} gate electrode. Jpn. J. Appl. Phys. 38 (Part 1A), 256–259. Watanabe, Y., Iechi, H., Kudo, K., 2007. Improvement in on/off ratio of pentacene static induction transistors by controlling hole injection barrier. Jpn. J. Appl. Phys. 46 (4B), 2717–2721. Xu, Z., Li, S.-H., Ma, L., Li, G., Yang, Y., 2007. Vertical organic light emitting transistor. Appl. Phys. Lett. 91 (9), 92911.

Vertical organic transistors

783

Yamauchi, H., Watannabe, Y., Iizuka, M., Nakamura, M., Kudo, K., 2008. Characterization of organic static induction transistors with nano-gap gate fabricated by electron beam lithography. IEICE Trans. Electron. E91-C (12), 1852–1855. Yang, Y., Heeger, A.J., 1994. A new architecture for polymer transistors. Nature 372 (6504), 344–346. Yu, H., Kim, J.H., Chen, W., Kim, D., Guo, J., So, F., 2014. Effect of nano-porosity on high gain permeable metal-base transistors. Adv. Funct. Mater. 24 (38), 6056–6065. Yu, H., Dong, Z., Guo, J., Kim, D., So, F., 2016a. Vertical organic field-effect transistors for integrated optoelectronic applications. ACS Appl. Mater. Interfaces 8 (16), 10430–10435. Yu, H., et al., 2016b. High-gain infrared-to-visible upconversion light-emitting phototransistors. Nat. Photonics 10 (2), 129–134. Yuan, Y., et al., 2014. Ultra-high mobility transparent organic thin film transistors grown by an off-centre spin-coating method. Nat. Commun. 5, 3005. Yutani, K., Nakayama, K., Yokoyama, M., 2006a. Fabrication of vertical organic field effect transistor at the edge of patterned photoresist. Mol. Cryst. Liq. Cryst. 444 (1), 197–202. Yutani, K., Fujimoto, S., Nakayama, K., Yokoyama, M., 2006b. Role of oxidation layer of aluminum base electrode in metal-base organic transistors. Mol. Cryst. Liq. Cryst. 462, 51–57. Zan, H.-W., Hsu, Y.-H., Meng, H.-F., Huang, C.-H., Tao, Y.-T., Tsai, W.-W., 2012. High output current in vertical polymer space-charge-limited transistor induced by self-assembled monolayer. Appl. Phys. Lett. 101 (9), 93307.

This page intentionally left blank

Vapor sensing using organic, polymer, and nanomaterial field-effect transistors

24

Hui Li*, Wei Shi*,†, Jennifer Dailey*, Hyun-June Jang*, Jian Song*, Junsheng Yu†, Howard E. Katz* *Department of Materials Science and Engineering, Johns Hopkins University, Baltimore, MD, United States, †State Key Laboratory of Electronic Thin Films and Integrated Devices, School of Optoelectronic Information, University of Electronic Science and Technology of China, Chengdu, People’s Republic of China

24.1

Introduction

24.1.1 Field-effect transistor principles A field-effect transistor (FET, also termed a “thin film transistor,” TFT) includes source and drain electrodes contacting a semiconductor (Fig. 24.1, device schematic). An insulator (dielectric) touches one interface of the semiconductor, and a gate electrode is placed at the other insulator interface (Islam, 2016; Petti et al., 2016). Different configurations are possible for the source-drain electrodes around the dielectric. The voltage on the gate controls the charge density in the semiconductor by serving as one plate of a capacitor, with the semiconductor acting as the other plate. Charging this capacitor changes the charge density in a region of the semiconductor near the dielectric, called the channel, along which current can flow between the source and the drain (Fig. 24.2). This situation is called the “on” state, while the converse low-charge-density condition is called the “off” state. The ratio of currents at the two states is called the “on/off ratio.” The presence of traps or dopants determines the gate voltage at or above which the device becomes decidedly “on”, called the “threshold voltage” (VT or Vth). Scattering sites and barriers decrease the charge-carrier mobility, the speed at which charge is induced to move by an applied voltage per unit device length. The mobility is ideally a solid-state property of the semiconductor, but an empirical mobility often includes extrinsic factors such as semiconductor-electrode contact resistance and the effect of electronic states conferred by the dielectric. For any given gate voltage (Vg), there is a maximum drain-source voltage (Vd) beyond which the current no longer increases; this phenomenon is called “saturation”. There are two related equations relating current to voltages in the low-Vd regime (the “linear” region) and the saturation region, where current is independent of Vd. Organic and polymeric materials are used in FETs (thus termed organic field-effect transistors, or OFETs), where mechanical flexibility, low processing temperatures, Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00024-3 © 2019 Elsevier Ltd. All rights reserved.

786

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 24.1 Four configurations of FETs. Features include electronic contacts in correct positions, but the material domain dimensions are not to scale.

0V –15 V –30 V –45 V –60 V

–1.0 × 10–6

IDS (A)

–8.0 × 10–6 –6.0 × 10–6 –4.0 × 10–6 –2.0 × 10–6 0.0 0

(A)

–10

–20

–30 –40 VDS (V)

–50

–60

–8.0 × 10–6 0V –15 V –30 V –45 V –60 V

IDS (A)

–6.0 × 10–6

0.0 –10

–20

V2D (VG – VT)VD – — 2

ID: drain current C: capacitance of dielectric m: carrier mobility VG: gate voltage VT: threshold (gate) voltage VD: drain voltage

–2.0 × 10–6

0

WCm L

~ wcmsat (VG – VT)2 ID,sat = 2L

–4.0 × 10–6

(B)

ID =

–30 –40 VDS (V)

–50

–60

Fig. 24.2 I versus Vd for (A) poly(bisdodecylquaterthiophene) and (B) poly(bisdodecylthio quaterthiophene). Vg are indicated in the legends. Voltages are negative because holes are induced in the organic semiconductors (OSCs). The dodecyl device is off at Vg ¼ 0 (Vth negative), while the dodecylthio device is on (Vth positive) at Vg ¼ 0. The equations for linear (Vd-dependent, low-Vd) and saturation region (Vd-independent, high-Vd) currents are also shown.

and print deposition are needed, or additional chemical functionality, such as utilized in vapor sensors, is desired (Di et al., 2009; Guo et al., 2010; Kim et al., 2014; Sirringhaus, 2014). OSCs contain conjugated subunits that stabilize charge density and allow charges to drift with applied voltage. This movement is usually dominated by a thermally activated hopping mechanism. Fig. 24.3 shows a small but

Vapor sensing using organic and polymer FET

787

Fig. 24.3 Structures used in recent high-performance OSC materials.

representative sample of OSC molecular structures (Dong et al., 2013); others are mentioned in the vapor sensor examples given later in this chapter. For the highest mobility and on/off ratio, OSC molecules need to be arranged so that charge transport is as smooth as possible.

24.1.2 OFETs in vapor sensors Because vapor-phase molecules (as well as dissolved small analyte molecules that also could have a vapor pressure) can interact with OSC molecules and the morphological features in OSC solids, the small molecules can be detected by the changes in electronic properties that they induce in the OSCs as a result of these interactions. The vapor to be analyzed may exert an electronic effect by creating traps, acting as dopants, imposing resistive interfacial barriers, and changing the existing intermolecular interactions between molecular subunits in OSCs and dielectrics. A trap can be formed by a dipole where one end attracts a charge carrier, thus slowing it, or by a chemical reagent that annihilates a carrier. The use of OSCs in vapor sensing has been covered in numerous prior reviews (Lv et al., 2017; Someya et al., 2008; Torsi et al., 2013; Wang et al., 2006; Zhang et al., 2015). The vapor analytes usually interact with

788

Handbook of Organic Materials for Electronic and Photonic Devices

polycrystalline OSC molecules at the grain boundaries (except for the receptorfunctionalized devices). Typical polymeric OSCs have a lamellar packing structure in which molecular packing is relatively looser than in small molecule semiconductors and where the spaces between crystalline domains are not empty, but rather are filled with amorphous, higher-free-volume polymers; thus vapor can permeate and interact within the grain as trapping sites and doping or dedoping agents (Torsi et al., 2013). OFETs are also an emerging, real-time biomolecule sensor platform with the ability to detect a variety of biomolecules, including deoxyribonucleic acid (DNA), glucose, and various proteins and enzymes (Hammock et al., 2013; Khan et al., 2011, 2010; Magliulo et al., 2013; Buth et al., 2011; Kergoat et al., 2010, 2012; Park et al., 2012; Park and Salleo, 2009; Lin et al., 2010; Bobbert et al., 2012; Spijkman et al., 2010). This latter case is covered only briefly in this chapter, for molecules small enough to exert a vapor pressure. This discussion explores the use of OFETs in vapor sensors. The emphasis is on broad coverage of advances from the past 5 years, where designed and well-explained interactions between analytes and device materials (OSCs, gate dielectrics) are utilized. Earlier studies are cited to put the more recent ones in context.

24.2

Active OSCs in vapor sensors

Numerous OSCs for OFETs have been developed over the past few decades, aiming primarily for high charge-carrier mobility and, more recently, environmental stability. These OSCs may be improper for sensor applications because sensing primarily emphasizes the change of parameters, such as μ, Vth, Id, and on/off ratio, induced by analytes. Maximization of these changes is in some sense contrary to the goal of stability. These changes are strongly dependent on the morphologies and redox properties of OSCs and the interaction between OSCs and analytes.

24.2.1 Small molecules Among OSCs, phthalocyanines (Pcs) have particularly good thermal and chemical stability. Laurs and Heiland fabricated devices based on various phthalocyanines to study their responses to oxidizing gases. This work predates the terminology of “transistors” being applied to OSC-based devices. Current changes were observed in response to oxidizing gases (Laurs and Heiland, 1987). Adding a strong Lewis acid, tris-(pentafluorophenyl)borane (TPFB), to a CuPc or CoPc semiconductor layer can reduce the mobility of an OFET and also act as an ammonia (NH3) receptor; both of these actions can increase NH3 sensitivity. Using this strategy, high NH3 sensitivity was achieved from an OFET (Huang et al., 2012). Phthalocyanine layers can be made thinner by depositing them with the Langmuir– Blodgett technique (Lu et al., 1994; Granito et al., 1996; Hu et al., 2000) or preparing a nanostructured semiconductor (Kim et al., 2011; Zhang and Hu, 2009). CuPc nanowires were used in a gas-dielectric FET sensor to detect sulfur dioxide (SO2)

N2

600 400

2

Au/Ti Substrate

PMMA

800

200

2

Au

PMMA

(A)

SO2

2

Au

CuPc

789

(PSO – PN ) / PN × 100 (%)

Vapor sensing using organic and polymer FET

0

(B)

m VT Ioff Ion SS 5 10 15 20 0 Concentration (ppm)

Fig. 24.4 (A) Schematic of the gas dielectric OFET sensor based on CuPc single-crystalline nanowires. (B) Parameter changes with increasing SO2 concentration extracted from transfer curves. Reprinted with permission from Shaymurat, T., Tang, Q., Tong, Y., et al. 2013. Gas dielectric transistor of CuPc single crystalline nanowire for SO₂ detection down to sub-ppm levels at room temperature. Adv. Mater. 25(16), 2269–2273. Copyright 2013 Wiley.

molecules (Fig. 24.4A) (Shaymurat et al., 2013). On the one hand, the surfaces of a CuPc nanowire can be exposed to SO2 molecules, which is partly responsible for the dramatically improved sensing properties. On the other hand, polar SO2 molecules can interact directly with the shallow traps at the OSC/dielectric interface, inducing significantly increased Ion and mobility (Fig. 24.4B). The large increase in mobility (from 0.005 to 0.05 cm2/V s in 20 ppm) is in contrast to the more frequent case of Vth being used to detect gases. Dielectric surface treatment can change the interfacial trap density of transistors (Salleo et al., 2002). Recently, Marks et al. reported nitrogen dioxide (NO2) sensing based on a CuPc transistor with an ultraviolet ozone (UVO)–treated polymeric gate dielectric (Huang et al., n.d.). NO2 can diffuse through the grain boundaries of CuPc and be absorbed by oxygen-containing groups at the surface of UVO-treated polystyrene. The mobility of devices increases as the time of UVO treatment increases, with increasing concentration of NO2. For example, the mobility has a negligible increase over its original value under 30-ppm NO2 exposure without UVO treatment, while it doubles with 300-s UVO treatment under the same exposure, indicating shallow traps are created by the UVO processing. Pentacene OSC morphologies are highly adjustable, making sensitivity/stability trade-offs tunable (Han et al., 2014). Malliaras (Zhu et al., 2002) prepared pentacene OFETs, for which the saturation current decreased with the increase of humidity. For example, the mobility drops to 9  104 cm2/V s after a 30-min exposure to 72% relative humidity (RH). The grain size of pentacene is strongly dependent on the dielectric layer, thus resulting in different charge-carrier trap density at the semiconductordielectric interface (Han et al., 2014; Yu et al., 2012; Shi et al., 2017). A low polarity of dielectric layers contributes to the relatively smaller grain size of pentacene and low surface-trap density, improving the performance of NH3 sensors. Using an ultrathin-film transistor with a multilayer of pentacene, Chao et al. increased the sensing response to NO2 by three orders of magnitude over a thick-layer

790

Handbook of Organic Materials for Electronic and Photonic Devices

sensor device because grain boundaries were more accessible and more electronically active (Mirza et al., 2015). The mobility changes from 0.038 to 0.045 cm2/V s with 10 ppm NO2. In addition, pentacene-based TFTs can be operated in water and still keep moderate mobility and on/off ratio (Cramer et al., 2012). An OFET was fabricated to detect NH3 by spray-coating bis (triisopropylsilylethynyl)-pentacene (TIPS-pentacene) (a solution-deposited but still highly crystalline pentacene derivative) on a poly(methyl methacrylate) (PMMA) dielectric layer (Yu et al., 2013). NH3 adsorption decreases the saturation current. The average mobility of original devices falls from 0.068 to 0.05 cm2/V s, with increasing NH3 gas concentration to 100 ppm. The on/off ratio also falls from 104 to 103. Dinaphtho[2,3-b:20 ,30 -f]thieno[3,2-b]thiophene (DNTT) can be regarded as an analog of pentacene in which one benzene is replaced by the thieno[3,2-b]thiophene group. Huang et al. prepared a porous DNTT film by a simple vacuum freeze-drying template method (Lu et al., 2017). In this case, the pores added accessibility of analyte gases. Modification of the DNTT structure by adding side chains makes this compound solution-processable. Chan et al. reported monolayer crystals of 2,9-didecyl-dinaphtho[2,3-b:20 ,30 -f]thieno[3,2-b]thiophene (C10-DNTT) (with solubilizing alkyl chains (Kang et al., 2011)) with an exemplary carrier mobility of up to 10.4 cm2/V s by using a new dual solution-shearing method (Peng et al., 2017). The high initial mobility and thinness of the layers were beneficial. Oligothiophene materials show wide tunability in their electrical properties from varying grain boundary or thickness (Tian et al., 2005; Huang et al., 2008; Torsi et al., 2002). These variations make it possible to design arrays of OFETs to respond differently to various analytes and to create unique response patterns (Crone et al., 2001). By changing the channel length of α,ω-dihexylquarterthiophene (DHα4T)– based OFETs on the same film, Katz et al. adjusted the number of grain boundaries per device systematically and found that the vapor sensing occurs mainly at grain boundaries because the response increases with more grain boundaries in the channel (Someya et al., 2002). An early ambient-stable n-type semiconducting naphthalene diimide (NDI) was reported by Katz et al. (2000) with the derivative H,H-NDI-CH2C7F15 (μ ¼ 0.1 cm2/V s; on/off ¼ 105) (Katz et al., 2000a). The mobility can be brought to as high as 0.7 cm2/V s with an on/off ratio of 107 via additional surface modification (Katz et al., 2000b). An OFET based on a NDI precursor, 1,4,5,8-naphthalene tetracarboxylic dianhydride (NTCDA), can be used as a novel O2 and H2O sensor. This n-type material shows stable mobility of 1  104 cm2/V s with on/off ratio of 104. The on- and off-current values of transistor decrease on exposure to oxygen because chemisorbed oxygen acts as a deep trap for electrons (Torsi et al., 2000). NH3-adsorbed NDI derivatives (NDI(2OD)(4tBuPh)-DTYM2) surprisingly provide responses to HCl (Fig. 24.5A) (Zang et al., 2014). Such devices also can be flexible (Fig. 24.5B) (Zang et al., 2016). Perylene tetra-carboxylic diimides (PDIs) are n-type OSCs that can have mobility comparable to their p-type counterparts (Wurthner and Stolte, 2011). Adding an electron-withdrawing, conjugated nitrile group allows PTCDI-CN2C12 to be highly sensitive to NH3 molecules, even during exposure to air for 14 days (Huang et al.,

Vapor sensing using organic and polymer FET

791

Fig. 24.5 (A) The device structure and sensing mechanism. (B) Transparent and flexible ultrathin-film transistor sensors based on NDI(2OD)(4tBuPh)-DTYM2 thin film. (A) Reprinted with permission from Zang, Y., Zhang, F., Huang, D., et al. 2014. Specific and reproducible gas sensors utilizing gas-phase chemical reaction on organic transistors. Adv. Mater. 26(18), 2862–2867. Copyright 2014 Wiley. (B) Reprinted with permission from Zang, Y., Huang, D., Di, C.-A., et al. 2016. Device engineered organic transistors for flexible sensing applications. Adv. Mater. 28(22), 4549–4555. Copyright 2016 Wiley.

2011). Decreased electrical conductivity, associated with the change in either chargecarrier mobility μ or the charge carrier concentration, was observed when O2, ethanol, acetone, or n-butane was adsorbed on the PDI derivatives MePTCDI or Cl4MePTCDI (Graaf and Schlettwein, 2006). The diffusion of gas molecules into the bulk increases the distance of hopping centers between two grains, thus increasing the energy barrier and finally leading to a decrease in μ. Selectivity can be gained using the array and response pattern strategy mentioned above. Katz’s group made a modest sensor array combining the CuPc transistor with two other organic transistors to recognize eight analytes by analyzing the direction and different responses of current, mobility, and threshold voltage changes (Huang et al., 2013). They also fabricated a transistor with a two-layer blend as the active layer (Fig. 24.6), in which a blend layer mixed hydroxylated and nonhydroxylated semiconductors as an upper layer on top of a 5,50 -bis(4-n-hexyl-phenyl)-2,20 -bithiophene (6PTTP6) layer (Huang et al., 2007). Compared with the device with single 6PTTP6 or single blend as the active layer, this heterostructure shows higher sensitivity to the analyte dimethyl methylphosphonate (DMMP). Rather than a microstructural effect, the trapping charge carriers diffusing from the top layer probably penetrate to the conductive channel between grains, where DMMP adsorbs, which enhances the sensing performance. Furthermore, the response times of these sensors were dramatically shortened by reducing the film thickness (Huang et al., 2008). Torsi et al. reported that a different kind of bilayered OFET (Torsi et al., 2008), featuring L-phenylalanine amino acid groups or β-D-glucosidic units as chiral elements in the outer layer, can differentiate between optical isomers. A double heterojunction made by Yan’s group ( Ji et al., 2013) contained vanadyl phthalocyanine (VOPc) grown on the ultra-thin heterojunction surface of N,N0 -diphenyl perylene tetra-carboxylic diimide (PTCDI-Ph) and para-hexaphenyl (p-6P). On the one hand,

792

Handbook of Organic Materials for Electronic and Photonic Devices S S

Electrodes

6PTTP6

6PTTP6

Gate Device 1

Dielectric

HO

O

S S

OH

HO6OPT 6PTTP6+HO6OPT

Electrodes

6PTTP6

6PTTP6+HO6OPT

Gate Device 2

O

Dielectric

Gate Device 3

Dielectric

Fig. 24.6 The device structures and semiconductor molecular structures. Single-6PTTP6 (device 1), single-blend (device 2), and two-layer blend (device 3). Reprinted with permission from Huang, J., Miragliotta, J., Becknell, A., et al. 2007. Hydroxyterminated organic semiconductor-based field-effect transistors for phosphonate vapor detection. J. Am. Chem. Soc. 129(30), 9366–9376. Copyright 2007 ACS Publishing.

the highly ordered conductive heterojunction layer of PTCDI-Ph/p-6P improves the charge-carrier transport; on the other hand, VOPc is a very sensitive p-type material to NO2. These combinations make the relative response five times larger than that of the single-heterojunction device. Another organic p–n junction achieved direct detection of solid chemical analytes by OFETs for the first time (Huang et al., 2017). This principle could be applied to vapor phase analytes as well. With this horizontal sideby-side diode structure, chemical solids can interact directly with charge carriers near the surface of the OSC layer. Fig. 24.7 summarizes the small-molecule OSCs mentioned here.

24.2.2 Polymers Polymer semiconductors show good solution-processability, film-forming properties, and high flexibility in thin-film devices relative to many low-molecular-weight compounds, so long as there is some conformational freedom around some of its linking bonds. Because their domain boundaries are not as pronounced as those of molecular solid OSCs, they are not necessarily as sensitive to vapors as the molecular solids are. On the other hand, their general insolubility in polar solvents, especially water, makes them attractive for detecting solutes in solution, an application mostly outside the scope of this discussion. The most commonly used polymer OSCs for gas sensing are polythiophenes such as poly(3-hexylthiophene) (P3HT), which can have functional groups along the P3HT backbone or as the end-capping groups (Chang et al., 2006). Functionalized P3HT can be applied to gas-sensor arrays to provide a “fingerprint” (response-pattern) response to volatile organic solvents, including alcohols, organic acids, and amines, with a sensitivity at the low part-per-million level. Lambeth et al. reported a nanostructured, regioregular P3HT and proposed the sensing mechanisms are mainly the intragrain effect and a grain boundary effect existing at the same time, which show contrary responses (Li and Lambeth, 2008). Additives such as polystyrene (Han et al.,

Vapor sensing using organic and polymer FET

793

Fig. 24.7 Small-molecule semiconductors.

2016a) and nanocomposite (Xu et al., 2010; Yang and Katz, 2017), make the blend active layers more sensitive to the analytes. The addition of palladium particles to P3HT film induce additional pores to increase the overall sensing area. Furthermore, these particles act as ethylene receptor sites where ethylene has an affinity to bind with

794

Handbook of Organic Materials for Electronic and Photonic Devices

transition metals, thus resulting in an increased sensing response (Besar et al., 2017). This concept will be revisited in the next section. Because of its deeper HOMO level, poly(quaterthiophene)s (PQTs; 5.2 eV) is less sensitive to air and humidity than P3HT (4.9–5.0 eV) and shows good OFET performance (Ong et al., 2004). However, PQTs are also more expensive. A flexible NH3 sensor platform based on polydidodecylquaterthiophene (PQT12), with a sensitivity of 0.5 ppm and a limit of detection of 0.1 ppm, was achieved (Besar et al., 2014). Katz et al. modified the structure of PQT12 by inserting sulfur atoms into the side chains increasing trap density and promoting redox interactions between the polymer and NO2 molecules (Fig. 24.8) (Li et al., 2017). This gave higher sensitivity of PQTS12 to NO2 gas compared to PQT12. Notably, the ratio of response from these two polymers can distinguish concentration detection from long-exposure time detection. PBTTT is also more stable in air and has higher mobility relative to P3HT (Brocorens et al., 2009). The responses of PBTTT to alcohols and acetone are larger than those to P3HT due to large-crystalline domains in PBTTT films, where larger numbers of analyte molecules can be adsorbed between the grains (Manoli et al., 2014). In this respect, it approaches the morphology of a molecular solid. Responses can be contact dependent. Sensitivity to NH3 and humidity of donor-acceptor (DA) polymers, where thiophene units alternate with electron-demanding subunits in conjugated chains, was improved by the introduction of microporous structures in DA polymer layers by washing away one additive (Wang et al., 2016a; Nketia-Yawson et al., 2017) or removing specific groups of polymers (Yang et al., 2016). Chi et al. reported a spirobifluorene-based OFET sensor for H2S and achieved detection limitation as low as 1 ppb (Lv et al., 2016). According to the different and changing rates of absorption and desorption of H2S molecules in the sensor, they propose that thinning the active layer does not simply favor an improved sensitivity. Noh et al. reported the precise control of thickness of DPPT-TT, where mobility decreased with the thinner film (Khim et al., 2016). For instance, the mobility is 1.78 (0.39) cm2/V s for 12.8 (1.0)–nm-thick film, while 1.1 ( 0.26) for 2.2 (0.3)–nm ultrathin film. The gas sensitivity to NH3 is as high as 82% for 2-nm-thick film and decreases with thicker film. Even ethylene detection was achieved without any additives based on this ultrathin-film OFET. Other p-type polymers for chemical sensing can be found in this review (McQuade et al., 2000). The principle of porosity-induced enhancement was demonstrated on the more rarely used n-type polymer (PBIBDF-BT in this case) by mixing poly(1,4-butylene adipate) (PBA) into a polymer semiconductor, followed by washing with acetone (Fig. 24.9) (Wu et al., 2017). The polymer semiconductors mentioned in this section are shown in Fig. 24.10.

24.3

Morphological and additive enhancements

For FET-based vapor sensors where the semiconducting layer is interacting with the analyte of interest, it is necessary to ensure sufficient specificity and avoid signals from random atmospheric interactions. Additionally, it is beneficial to create as

Vapor sensing using organic and polymer FET

795

Fig. 24.8 (A) The molecular structure of PQT12 and PQTS12 and the device structure based on these polymers. (B) Sensitivity ratio of PQT12 and PQTS12 varies based on NO2 concentration. Reprinted with permission from Li, H., Dailey, J., Kale, T. et al. 2017. Sensitive and selective NO2 sensing based on alkyl- and alkylthio-thiophene polymer conductance and conductance ratio changes from differential chemical doping. ACS Appl. Mater. Interfaces 9(24), 20501–20507. Copyright 2017 ACS Publishing.

responsive a layer as possible to provide maximum detection potential. Several OSC modifications lead to improvements in these qualities, including the addition of pores, embedded particles, or other chemicals that react as desired with the surrounding vapor. The sensing ability of a semiconducting layer correlates with the number of potential binding sites of the material, so that the highest number of analyte molecules can

Fig. 24.9 The preparation of porous film based on the n-type polymer PBIBDF-BT. Reprinted with permission from Wu, S., Wang, G., Xue, Z., et al. 2017. Organic field-effect transistors with macroporous semiconductor films as high-performance humidity sensors. ACS Appl. Mater. Interfaces 9(17), 14974–14982. Copyright 2017 ACS Publishing.

Fig. 24.10 The polymer semiconductors.

Vapor sensing using organic and polymer FET

797

interact electronically. It may not be preferable to increase the overall area of a device because of geometric or power efficiency considerations, but it may be possible to add roughness or porosity to the material by chemical means, thereby creating more binding areas. The use of rough or porous morphology already has been mentioned as a means of increasing sensitivity. This was examined thoroughly by Yu et al., who examined the results of morphological changes to a polymer semiconductor layer, where an improvement in NH3 detection was noted (Yu et al., 2016). The results support the hypothesis that the additional surface area caused by the roughness of the layer increases gas-sensing ability. Wu et al. created a macroscopic semiconductor layer using a facilitated phaseseparation method and increased the functionality of a humidity sensor (Wu et al., 2017). Lu et al. used a simple vacuum-freeze drying template method to create pores for an NH3 sensor, with similarly positive results (Lu et al., 2017). This is shown in Fig. 24.11A–D (Lu et al., 2017). Whatever the method of formation, increased surface area to allow to more binding sites is a common method to increase sensitivity in organic FET vapor sensors. In 2014, Cho et al. developed an approach of evaporating pentacene on top of m-bis (triphenylsilyl)benzene (TSB3), resulting in extraordinary morphology with far fewer grain boundaries and myriad nanometer-sized pores in pentacene, as shown in Fig. 24.11E (Kang et al., 2014). These peculiar structures were formed by differences in molecular interactions between the organic layers and the substrate surface. The pore-rich structure improved the sensitivity of the OFET to methanol gas. The improvement in sensing ability was noted for NH3 detection (Yu et al., 2016). A completely different means of enhancement is to add metal nanostructures to OSCs. The enhancements can be from electronic effects, increased analyte binding ability, or effects of the particles on OSC morphology. Other additives to semiconductor films can similarly improve sensitivity or selectivity, through a variety of mechanisms. Metal particles such as platinum can be added to the sensing film in order to increase the frequency of charge transfer (CT), as demonstrated by Zheng et al. in the development of a DNT sensor (Zheng et al., 2013). Surya et al. similarly increased hydrogen sulfide detection through the addition of tin oxide nanoparticles (Surya et al., 2016). It should be noted that while the addition of metal particles with particular affinity can frequently increase sensor response, at very high concentrations, it is possible to create constant conductive pathways through the particles inadvertently, so there will likely be an optimum concentration of particles that provides sufficient binding, but enough separation to not ruin the semiconductor. Zinc oxide (ZnO) was added next to pentacene by Han et al., and it was found that this decreased the average grain size of the organic semiconductor. This allowed for increased diffusion of CuPc molecules along boundaries and resulted in a significantly increased sensing signal (Han et al., 2016b). These types of additives also can act in synergy, as demonstrated by Besar et al. in their development of an ethylene vapor sensor, which used both palladium particles and chemically induced porosity to achieve a substantially higher specific sensing response (Besar et al., 2017). In this study, a film containing a chemical component that degraded and released gas at high temperature was heated to create the pores.

798 Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 24.11 Schematic illustration of the porous OFET-based sensors. (A) The device structure of the porous OFET-based sensors and the molecular structure of DNTT. Optical images of a porous DNTT film (B) with polystyrene microspheres and (C) after removing polystyrene microspheres. (D) The fabrication procedure of the porous OFET-based sensors. (E) Device schematic and the molecular structure of TSB3. (D) Reprinted with permission from Lu, J., Liu, D., Zhou, J., et al. 2017. Porous organic field-effect transistors for enhanced chemical sensing performances. Adv. Funct. Mater. 27(20), 1700018. Copyright 2017 Wiley. (E) Reprinted with permission from Kang, B., Jang, M., Chung, Y., et al. 2014. Enhancing 2D growth of organic semiconductor thin films with macroporous structures via a small-molecule heterointerface. Nat. Commun. 5, 4752. Copyright 2014 Nature Publishing Group.

Vapor sensing using organic and polymer FET

24.4

799

Biomaterial incorporation

Many biomaterials are readily available, nontoxic, suitable for high-throughput processes, and biodegradable (Lebreton et al., 2017; Braun et al., 1998; Keren et al., 2003; Seeman, 2003; Yan et al., 2015). Applications of biomaterials in OFET gas sensors can be for adjustments in electronic parameters and as recognition elements. They also can be used as dielectrics or semiconductors in their own right, though charge-carrier mobilities are generally marginal. Proteins, caramelized glucose, edible hard gelatin, and commercially available plastics (organic polymeric molecules) made of potato or cornstarch can be fabricated as the metabolizable or biodegradable substrates (Bardea et al., 1997; Katz et al., 1998; Irimia-Vladu et al., 2010a). Nucleic acids and their individual bases and sugars can play roles in dielectrics (Steckl, 2007; Stadler et al., 2007). Beta carotene and indigo, which are p-type and n-type, respectively, show some activity as semiconductors, yielding a charge mobility of about 104 cm2/V s (Irimia-Vladu et al., 2010b, 2012). Mihai’s group has summarized the biomaterials which have been used in OFETs, as shown in Fig. 24.12.

Perylene diimide O N O

Beta carotene

O N O

Adenine

Guanine

Indigo O

Br

Vat Yellow 1 N

Br

O

gl

CH2OH

Ec

O

mi

uc

te

du

Die Sm

e Ha rd

CH OH 2 OH O

CH2OH

O 30 0–

OH O

ge

la

tin

ec

ap

NH

N H

OH

N

HO

urc

e

NH2

CH2OH O OH HO

OH

cto r tric

nin

Sucrose O

CH2OH O HO

CH OH 2

OH

Lactose CH2OH O OH

O

CH2OH O OH

OH

HO

HO

lec

g la ye r

bs

os

CH2OH O OH

N N

HO

So

co n

Su

of lex

HO

HO

Drain

oo the

OH

OH O

Glucose

se

Ga

HO

HO

N

O

Ca ra m CH OH eli 2 ze HO OH d O

HO

O

2

N H

N

O H N N H O

Vat Orange 3

NH

N

tra

te

rin Au O

OH

su

le HO

60 HO HO 0

Fig. 24.12 Utilization of natural materials or materials inspired by nature in OFETs. Reprinted with permission from Irimia-Vladu, M., Troshin, P. A., Reisinger, M., et al. 2010. Biocompatible and biodegradable materials for organic field-effect transistors. Adv. Funct. Mater. 20(23), 4069–4076. Copyright 2010 Wiley.

800

Handbook of Organic Materials for Electronic and Photonic Devices

24.4.1 Biomaterials at interfaces Most OFET current flows in the first few molecular layers of the semiconductor resting on the dielectric in OFETs (Horowitz, 1998; Stadlober et al., 2005; Sirringhaus, 2005; Stassen et al., 2004). The nucleobase guanine has physical properties that allow its deposition at that location and effective modification of the electronic properties of those layers (Irimia-Vladu et al., 2010a; Lee et al., 2014). It has low dielectric losses of 102 at 10 mHz and high breakdown strength between 1.5 and 3.5 MV/cm (IrimiaVladu et al., 2010a). Katz’s group employed guanine on top of silicon dioxide (SiO2) to enhance the performance of pentacene OFETs, as illustrated in Fig. 24.13A and B (Shi et al., 2016). The higher ionization energy of guanine molecule indicated that it will not trap the mobile holes in the pentacene film as shown in Fig. 24.13D. A tripling of the fieldeffect mobility, from 0.13 to 0.42 cm2/V s, was achieved by introducing a 2-nm guanine layer, as presented in Fig. 24.13E. They demonstrated that the increased fieldeffect mobility was mainly attributed to the hydrogen bonding capacity of otherwise

Fig. 24.13 Device architecture of the OFET and representation of the location of hydrogen atoms in an OFET (A) without guanine layer and (B) with guanine layer. (C) OFET with a guanine/pentacene/guanine/pentacene structure. (D) Schematic energy-level diagrams of pentacene and guanine base pairs with vacuum-level (EVAC) alignment. (E) Field-effect mobilities of OFETs with different thicknesses of the guanine layer. (B) Reprinted with permission from Shi, W., Zheng, Y., Yu, J. et al. 2016. Mobility enhancement of organic field-effect transistor based on guanine trap-neutralizing layer. Appl. Phys. Lett. 109(14), 769. Copyright 2016 AIP Publishing. (C) Reprinted with permission from Shi, W., Zheng, Y., Taylor, A. D., et al. 2017. Increased mobility and on/off ratio in organic field-effect transistors using low-cost guanine-pentacene multilayers. Appl. Phys. Lett. 111(4), 043301. Copyright 2017 AIP Publishing. (E) Reprinted with permission from Shi, W., Zheng, Y., Yu, J. et al. 2016. Mobility enhancement of organic field-effect transistor based on guanine trap-neutralizing layer. Appl. Phys. Lett. 109(14), 769. Copyright 2016 AIP Publishing.

Vapor sensing using organic and polymer FET

801

unassociated guanine molecules, which enabled them to neutralize trapping sites on the SiO2 surface. The Katz group inserted one more guanine layer into pentacene layers, as shown in Fig. 24.13C (Shi et al., 2017). This modification resulted in useful trap-reducing properties and a cost reduction, as guanine replaced some of the pentacene.

24.4.2 Biomaterial incorporation Indigo, a colorant that has been used for thousands of years, has an extremely low solubility and a high melting point (390–392°C) arising from strong intermolecular interactions (Steingruber, 2000). Indigo was used as a sensing material detecting ozone in water (Bader and Hoign
e, 1981) at the level of 2% or 3 μg/L Much more recently, Dubois’s group adsorbed indigo on carbonaceous nanomaterials as a chemical filter for the selective detection of ambient NO2 (Dubois et al., 2013). The surfaces of multiwalled carbon nanotubes (MWCNTs) were modified with indigo molecules held onto the CNTs by π-stacking. A distinction between ozone and NO2 was achieved, a particularly challenging chemical selection, as shown in Fig. 24.14A. Katz’s group fabricated a guanine and pentacene OFET pair, which were sensitive and selective to NO2 and five other vapor analytes, including acetone, isopropyl alcohol (IPA), ethyl acetate, water, and acetic acid (Shi et al., 2018). OFETs with two configurations were fabricated, one of which used thermally deposited guanine and pentacene in alternating layers, as shown in Fig. 24.13C, which was more sensitive to NO2, and another with only pentacene as OSC. Patterns of responses from NO2 and the other analytes were distinguishable, as shown in Fig. 24.14B and C. Torsi’s group fabricated an OFET integrating polymeric phospholipid (PL) membranes (quite different from molecular crystalline solids) between the dielectric and P3HT semiconductor to detect the volatile general anesthetic molecules such as diethyl ether and halothane, as shown in Fig. 24.15A and B (Daniela et al., 2013). Another polymeric material, silk fibroin (SF; Fig. 24.15C) was employed by Yu’s group on top of the PMMA gate dielectric to improve NO2 detection (Li et al., 2015). The enhancement of the sensing property was owing to the interaction between the hydroxyl and the amidogen of the SF biomaterial and NO2 molecules at the OSCdielectric interface.

24.5

Health-related vapor sensors using organic and polymer semiconductors

Many of the vapor-sensing examples presented so far are directed toward environmental monitoring. A second emerging application is for health-related assessments, either of triggers for health disorders or for the diagnosis of health conditions based on volatile compounds collected from patients. Alcohols are one commonly monitored class of health-related volatile compounds (Veenstra et al., 2009). Katz et al. reported a set of oligothiophenes used in OFET sensors responding to a series of primary alcohols

80 60 40

O

20

N H

5.0

99.5 3.6

99.0 4.9

Indigo

MWCNTs

Hybrid

Biphasic

η O3

60 40 20 0 –20 –40 –60

m

Ids 5 ppm

(B)

10 ppm

4.9

98.0 5.0

80

–80

3.6

99.3 4.7

(A) Parameters change ratio or value

O

Vth IPA

Acetone

η NO2 190 171 152 133 114 95 76 57 38 19 0

Ethyl acetate

(C)

Ids Water

m Acetic acid

Vth

Handbook of Organic Materials for Electronic and Photonic Devices

0

4.7

H N

Parameters change ratio or value

Filtering yield (%)

802

100

Vapor sensing using organic and polymer FET

Fig. 24.14 (A) Filtering yields toward O3 and NO2 for the indigo and MWCNTs, as well as their associations, the inset shows the molecular structure of indigo. Drain current and mobility change ratio and threshold voltage change value of (B) pentacene, and (C) guanine-pentacene OFETs to 5 ppm NO2, 10 ppm NO2, IPA, acetone, ethyl acetate, water, and acetic acid under 5 min exposure. (A) Reprinted with permission from Dubois, M., Brunet, J., Pauly, A., et al. 2013. Indigo molecules adsorbed on carbonaceous nanomaterials as chemical filter for the selective detection of NO₂ in the environment. J. Colloid Interface Sci. 407(10), 39–46. Copyright 2013 Elsevier. (C) Reprinted with permission from Shi, W., Zheng, Y., Taylor, A. D., et al. 2017. Increased mobility and on/off ratio in organic field-effect transistors using low-cost guanine-pentacene multilayers. Appl. Phys. Lett. 111(4), 043301. Copyright 2017 Elsevier.

803

804

Handbook of Organic Materials for Electronic and Photonic Devices

–600

S

0.015

D OSC

PLs 2 Highly doped silkon

G

–300 –200 –100 0

(A)

Diethyl ether Acetone

SiO

–400

Halothane

N2

Sensitivity

Io (nA)

–500

0.010

0.005

Diethyl ether 40

20

0

–20

–40

–60

–80 –100

Vg (V)

o

PL FBI-OFET

3

P3H-OFET

OH

o

CH

H N

H N N H

Glysine

(B) o

CH3

H N

0.000

N H

N H

O o o Alanine Glysine Alanine Glysine Serine

n

(C) Fig. 24.15 (A) Transfer characteristics for a PL OFET measured in N2 and in a controlled atmosphere of diethyl ether at a concentration of 3.0 wt% in N2 atmosphere; the inset is the side view of the PL OFET device. (B) Differential analytical sensitivities (ΔI  [saturated vapor fraction]–1) of PL-OFET and P3HT OFET sensors exposed to diethyl ether, halothane, and acetone. For all the compounds, the maximum concentration tested was half the saturated vapor. (C) Molecular structure of SF. (B) Reprinted with permission from Daniela, A. M., Magliulo, M., Cotrone, S., et al. 2013. Volatile general anesthetic sensing with organic field-effect transistors integrating phospholipid membranes. Biosens. Bioelectron. 40(1), 303–307. Copyright 2013 Elsevier.

(Crone et al., 2001). All the devices show obvious response to alcohols, while the degree of odor response increases as the length of the semiconductor hydrocarbon end group increases. Such effects of the side chain of the polymers on the response to alcohols also were observed by Torsi, who reported that ether oxygen groups on semiconducting polymer side chains altered responses of polymer OSCs to alcohol vapors (Torsi et al., 2003). Molecular solids also can be used for alcohol detection, where a pore-rich structure improved the sensitivity of the OFET to methanol gas (Kang et al., 2014). OFETs with two hydroxy-functionalized semiconductors, 5,50 -bis(4-n-hexylphenyl)-2,20 -bithiophene (6PTTP6) and 5,50 -bis(4-hydroxyhexyloxyphenyl)-2,20 bithiophene (HO6OPT), were used by Katz’s group to selectively detect as low as tens of ppm of the nerve gas model compound DMMP, as shown in Fig. 24.16 (Huang et al., 2007). The sensitivity was brought even lower, to 5 ppm, by making the OSC layers as thin as possible (Huang et al., 2008). An n-type OSC counterpart responding to DMMP was also demonstrated, an NDI with a fluorinated side chain (See et al., 2007). Ammonia is suspected as a trigger of various respiratory disorders (Brautbar et al., 2003). Katz et al. used potentially printable poly (3,3000 -didodecylquaterthiophene)

Vapor sensing using organic and polymer FET

805

0

Id-sat (μA)

S S

6PTTP6 HO

O

S S

HO6OPT

O

–4 –8 Vg = –20 V Vg = –30 V Vg = –40 V Vg = –50 V Vg = –60 V

OH

–12

0

100 200 300 400 500 600 700 800

Time (s)

Fig. 24.16 Molecular structure illustration of 6PTTP6 and HO6OPT and reversible responses of OFET integrating two-layer blend films upon exposure to 150 ppm of DMMP vapor. Reprinted with permission from Huang, J., Miragliotta, J., Becknell, A., et al. 2007. Hydroxyterminated organic semiconductor-based field-effect transistors for phosphonate vapor detection. J. Am. Chem. Soc. 129(30), 9366–9376. Copyright 2007 ACS Publishing.

(PQT-12) as sensitive OSC to detect NH3 at concentrations as low as 0.5 ppm (Besar et al., 2014). The detection limit was brought down to 0.01 ppm by Sung, who employed single-crystal poly(3-hexylthiophene) (P3HT) nanowires, as shown in Fig. 24.17A (Mun et al., 2017). Porous dinaphtho [2,3-b:20 ,30 -f] thieno [3,2-b] thiophene (DNTT) films showed a relative sensitivity of as high as 340%/ppm upon exposure to 10 ppb NH3 (Lu et al., 2017). A TIPS-pentacene and polystyrene blend (Feng et al., 2016) was incorporated into a battery-powered NH3 detector useful to 5 ppm (as shown in Fig. 24.17B) in ambient air while consuming only about 50 nW of power. NO2 is associated with bronchitis, emphysema, and respiratory irritation at low concentrations (Wang et al., 2016b). Das et al. detected NO2 using amorphous semiconducting polymers with methoxy-substituted poly(triarylamine) (PTA-OMe), as shown in Fig. 24.18A (Das et al., 2010). The PTA-OMe-based OFET responded to 10 ppb of NO2, as presented in Fig. 24.18B. Yan’s group adopted a highly ordered organic ultra-thin-film heterojunction of N,N0 -diphenyl perylene tetracarboxylic diimide (PTCDI-Ph) and para-hexaphenyl (p-6P) ultra-thin film for ambient detection of NO2 ( Ji et al., 2013). A more elaborate double-heterojunction was made by further addition of VOPc on top of the PTCDI-Ph layer, giving a 90% relative response to 5 ppm of NO2. Tang et al. used a dinaphtho[3,4-d:30 ,40 -d0 ]benzo[1,2-b:4,5-b0 ] dithiophene (Ph5T2)–modified, copper phthalocyanine (CuPc), single-crystal nanowire with gas dielectric to selectively detect NO2, NO, and H2S down to the sub–parts per million level, as depicted in Fig. 24.18C and D (Song et al., 2017). This OFET exhibited high response and excellent controllable selectivity at room temperature, as shown in Fig. 24.18D. Katz’s group considered the NO2-sensing properties of hybrid (Han et al., 2016a) P3HT and ZnO-graphene oxide shell-core nanoparticles (Yang and Katz, 2017), obtaining a 210% sensing response to 5 ppm of NO2 gas exposure for 5 min at room temperature. Zhu showed that preadsorption of NH3 assisted the detection of NO2 at the 10-ppm level (Zang et al., 2014).

806

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 24.17 (A) Schematic illustration of a P3HT single-crystal nanowire OFET and its interaction with NH3; the inset is the SEM perspective image of the P3HT nanowires. (B) The measured relative change of Vout over time upon NH3 exposure of different estimated concentrations from 5 to 25 ppm. (A) Reprinted with permission from Mun, S., Park, Y., Lee, Y. K., et al. 2017. Highly sensitive ammonia gas sensor based on single-crystal poly(3-hexylthiophene) (P3HT) organic field effect transistor. Langmuir 33(47), 13554–13560. Copyright 2017 ACS publications. (B) Reprinted with permission from Feng, L., Tang, W., Zhao, J., et al. Unencapsulated air-stable organic field effect transistor by all solution processes for low power vapor sensing. Sci. Rep. 6, 20671. Copyright 2016 Nature Publishing Group.

Tang and Liu’s group used CuPc nanowire OSCs and pioneered the use of gas as the OFET dielectric for room-temperature detection of SO2 at sub–parts per million levels (0.5 ppm) with high sensitivity (119%) and high resolution (100 ppb). Chi’s group employed a spirobifluorene-based polymeric OFET in an investigation of the thickness dependence of the OFET to the detection of H2S (Lv et al., 2016). When the optimal thickness was used, the greatest sensitivity (1 ppb) reported for OFET detection of this gas was reached, with promising selectivity as well, as shown in Fig. 24.19A. Zhang’s group developed a new cruciform DA p-type molecule for OFETs—namely, 2,20 -((5,50 -(3,7-dicyano-2,6-bis (dihexylamino)benzo[1,2-b:4,5b0 ]difuran-4,8-diyl)bis (thiophene-5,2-diyl))bis (methanylylidene)) dimalononitrile (BDFTM), which showed highly sensitive and selective detection of H2S gas (down to 10-ppb levels) (Fig. 24.19B) (Luo et al., 2014).

24.6

Small biomolecule detection

While most biosensors have focused on assaying dissolved molecules with little or no vapor pressure, there are a few examples where volatile molecules were detected using methods with biochemical origins, or they were generated as products of reactions intended to help analyze for larger biomolecules. For example, oxidase enzymes used to functionalize channels (Welch et al., 2015) or gate electrodes (Pappa et al., 2016) catalyze reactions generating products such as hydrogen peroxide (H2O2)

Vapor sensing using organic and polymer FET

807

(A) (B) 600 400 200 10 min

S (%)

0 –200 –400

NO2

NO

5 min H2S

Unmodified

–600 –800 –1000 –1200

(C)

(D)

Fig. 24.18 (A) Molecular structures of the PTA-OMe. (B) The required resistance change in the gain trim potentiometer (ΔRtrim) response of PTA-OMe based OFETs to NO2; the inset shows the same data on a log–log scale. (C) Schematic diagram of the OFET with Ph5T2 on the CuPc nanowire, the inset presents the typical SEM image of one CuPc nanowire under 10 min of Ph5T2 modification by vacuum evaporation deposition. (D) Effect of Ph5T2 deposition time on the response of 10 ppm of NO2, NO, and H2S. (A) Reprinted with permission from Das, A., Dost, R., Richardson, T., et al. 2010. A nitrogen dioxide sensor based on an organic transistor constructed from amorphous semiconducting polymers. Adv. Mater. 19(22), 4018–4023. Copyright 2010 Wiley. (D) Reprinted with permission from Song, Z., Liu, G., Tang, Q., et al. 2017. Controllable gas selectivity at room temperature based on Ph5T2-modified CuPc nanowire field-effect transistors. Org. Electron. 48, 68–76. Copyright 2017 Elsevier.

(Liao et al., 2013). Similarly, an NH3 sensor quantified urea from a pH change caused by OH produced from urea catalyzed by urease (UOx) (Werkmeister et al., 2016). Communication between enzymes and electrodes can be enhanced using a redoxactive additive, known as a mediator. Braendlein et al. investigated the detection of lactate (which has a measurable vapor pressure) directly secreted by cancer cells using lactate oxidase (LOx) and the mediator chitosan-ferrocene, immobilized at the gate (Braendlein et al., 2017). Horseradish peroxidase (HRP) osmium-redox polymer coated onto the extended Au gate electrode achieved a low limit of detection (LOD) of 66 nM for lactate (Minami et al., 2015). Utilizing enzymes for the sensors has conferred enhanced specificity and selectivity for targeting molecules. Several drawbacks, however, exist, such as the limited

808

Handbook of Organic Materials for Electronic and Photonic Devices

2S

30 Et h

60 40 20 0

(A)

1

2

H ex an Is e op ro pa no l C H 2C H 2l 2O To lu en Ac e et on H C e H O

3

N H

ΔISD,a (%)

80

3

4

5 6 7 Analytes

8

9 10

ID decreased (%)

an

ol

H

100

25 20 15 10 5 0

(B)

0

1

2

3

4 5 6 Gas type

7

8

Fig. 24.19 (A) Selectivity test of the OFET under 30-s exposure, the concentrations were 1 ppm for H2S, 50 ppm for NH3, 470 ppm for HCHO, and > 10,000 ppm for other solvent analytes. (B) Variation of the drain current after exposed to gas vapors: (1) H2 (pure); (2) CO2 (pure); (3) hexane (52,000 ppm); (4) CH2Cl2 (301,000 ppm); (5) CH3OH (6500 ppm); (6) acetone (1000 ppm); (7) NH3 (150 ppb); (8) H2S (100 ppb). (A) Reprinted with permission from Lv, A., Wang, M., Wang, Y., et al. 2016. Investigation into the sensing process of high-performance H2S sensors based on polymer transistors. Chem. Eur. J. 22(11), 3654–3659. Copyright 2016 Wiley. (B) Reprinted with permission from Luo, H., Chen, S., Liu, Z., et al. 2014. A cruciform electron donor-acceptor semiconductor with solidstate red emission: 1D/2D optical waveguides and highly sensitive/selective detection of H₂S gas. Adv. Funct. Mater. 24(27), 4250–4258. Copyright 2014 Wiley.

categories of usable enzymes, resultant limited targeting analytes, degradation of enzymes/mediators, and unsatisfactory sensitivity in physiological conditions.

24.7

Conclusions and outlook

The wide variety of OSCs and polymer semiconductors and their increases in functionality, mobility, and morphology control have opened up new opportunities for vapor sensing, especially for vapors that cause adverse health effects. Additives to OSCs provide further tunability and alternative response mechanisms. The tuning of OSC-dielectric interfaces and control of charge injection from electrodes allow the presetting of electronic properties so that the greatest responses to vapors are achieved. These interfaces can even contribute to the responses. However, significant challenges remain when it comes to achieving sufficient discrimination among vapors with similar chemical functional groups and baseline stability during exposure to ambient environments. There is a trade-off among high mobility, high stability, and high responsiveness to analytes; optimizing this trade-off can be considered the broad goal for future research in this area.

Acknowledgments The review of sensor technologies and mechanisms was funded by the national Institutes of Health R21 Grant No. R21-EB018426. The content is solely the responsibility of the authors

Vapor sensing using organic and polymer FET

809

and does not necessarily represent the official views of the National Institutes of Health (NIH). The review of semiconductor polymer properties was funded by the National Science Foundation (NSF), Division of Chemistry, Grant No. 1708245.

References Bader, H., Hoign
e, J., 1981. Determination of ozone in water by the indigo method. Water Res. 15 (4), 449–456. Bardea, A., Katz, E., B€uckmann, A.F., et al., 1997. NAD+-dependent enzyme electrodes: electrical contact of cofactor-dependent enzymes and electrodes. J. Am. Chem. Soc. 119 (39), 9114–9119. Besar, K., Yang, S., Guo, X., et al., 2014. Printable ammonia sensor based on organic field effect transistor. Org. Electron. 15 (11), 3221–3230. Besar, K., Dailey, J., Katz, H.E., 2017. Ethylene detection based on organic field-effect transistors with porogen and palladium particle receptor enhancements. ACS Appl. Mater. Interfaces 9 (2), 1173–1177. Bobbert, P.A., Sharma, A., Mathijssen, S.G., et al., 2012. Operational stability of organic fieldeffect transistors. Adv. Mater. 24 (9), 1146–1158. Braendlein, M., Pappa, A.M., Ferro, M., et al., 2017. Lactate detection in tumor cell cultures using organic transistor circuits. Adv. Mater. 29 (13), 1605744. Braun, E., Eichen, Y., Sivan, U., et al., 1998. DNA-templated assembly and electrode attachment of a conducting silver wire. Nature 391 (6669), 775–778. Brautbar, N., Wu, M.P., Richter, E.D., 2003. Chronic ammonia inhalation and interstitial pulmonary fibrosis: a case report and review of the literature. Arch. Environ. Health 58 (9), 592–596. Brocorens, P., Van Vooren, A., Chabinyc, M.L., et al., 2009. Solid-state supramolecular organization of polythiophene chains containing thienothiophene units. Adv. Mater. 21 (10 11), 1193–1198. Buth, F., Kumar, D., Stutzmann, M., et al., 2011. Electrolyte-gated organic field-effect transistors for sensing applications. Appl. Phys. Lett. 98 (15), 153302. Chang, J.B., Liu, V., Subramanian, V., et al., 2006. Printable polythiophene gas sensor array for low-cost electronic noses. J. Appl. Phys. 100(1), 014506. Cramer, T., Kyndiah, A., Murgia, M., et al., 2012. Double layer capacitance measured by organic field effect transistor operated in water. Appl. Phys. Lett. 100 (14), 143302. Crone, B., Dodabalapur, A., Gelperin, A., et al., 2001. Electronic sensing of vapors with organic transistors. Appl. Phys. Lett. 78 (15), 2229–2231. Daniela, A.M., Magliulo, M., Cotrone, S., et al., 2013. Volatile general anesthetic sensing with organic field-effect transistors integrating phospholipid membranes. Biosens. Bioelectron. 40 (1), 303–307. Das, A., Dost, R., Richardson, T., et al., 2010. A nitrogen dioxide sensor based on an organic transistor constructed from amorphous semiconducting polymers. Adv. Mater. 19 (22), 4018–4023. Di, C.A., Liu, Y., Yu, G., et al., 2009. Interface engineering: an effective approach toward highperformance organic field-effect transistors. Acc. Chem. Res. 42 (10), 1573–1583. Dong, H., Fu, X., Liu, J., et al., 2013. 25th anniversary article: Key points for high-mobility organic field-effect transistors. Adv. Mater. 25 (43), 6158–6183. Dubois, M., Brunet, J., Pauly, A., et al., 2013. Indigo molecules adsorbed on carbonaceous nanomaterials as chemical filter for the selective detection of NO₂ in the environment. J. Colloid Interface Sci. 407 (10), 39–46.

810

Handbook of Organic Materials for Electronic and Photonic Devices

Feng, L., Tang, W., Zhao, J., et al., 2016. Unencapsulated air-stable organic field effect transistor by all solution processes for low power vapor sensing. Sci. Rep. 6, 20671. Graaf, H., Schlettwein, D., 2006. Influence of gas molecules on the charge carrier mobility in thin films of semiconducting perylene tetracarboxylic imides. J. Appl. Phys. 100 (12), 126104. Granito, C., Wilde, J.N., Petty, M.C., et al., 1996. Toluene vapour sensing using copper and nickel phthalocyanine Langmuir-Blodgett films. Thin Solid Films 284–285 (Suppl. C), 98–101. Guo, Y., Yu, G., Liu, Y., 2010. Functional organic field-effect transistors. Adv. Mater. 22 (40), 4427–4447. Hammock, M.L., Knopfmacher, O., Naab, B.D., et al., 2013. Investigation of protein detection parameters using nanofunctionalized organic field-effect transistors. ACS Nano 7 (5), 3970–3980. Han, S., Huang, W., Shi, W., et al., 2014. Performance improvement of organic field-effect transistor ammonia has sensor using ZnO/PMMA hybrid as dielectric layer. Sens. Actuators B Chem. 203 (Suppl. C), 9–16. Han, S., Zhuang, X., Shi, W., et al., 2016a. Poly(3-hexylthiophene)/polystyrene (P3HT/PS) blends based organic field-effect transistor ammonia gas sensor. Sens. Actuators B Chem. 225 (Suppl. C), 10–15. Han, S., Cheng, J., Fan, H., et al., 2016b. Achievement of high-response organic field-effect transistor NO(2) sensor by using the synergistic effect of ZnO/PMMA hybrid dielectric and CuPc/pentacene heterojunction. Sensors 16 (10), 1763. Horowitz, G., 1998. Organic field-effect transistors. Adv. Mater. 10 (5), 365–377. Hu, W., Liu, Y., Xu, Y., et al., 2000. The gas sensitivity of a metal-insulator-semiconductor field-effect-transistor based on Langmuir–Blodgett films of a new asymmetrically substituted phthalocyanine. Thin Solid Films 360 (1), 256–260. Huang, J., Miragliotta, J., Becknell, A., et al., 2007. Hydroxy-terminated organic semiconductor-based field-effect transistors for phosphonate vapor detection. J. Am. Chem. Soc. 129 (30), 9366–9376. Huang, J., Sun, J., Katz, H.E., 2008. Monolayer-dimensional 5,50 -bis(4-hexylphenyl)-2,20 bithiophene transistors and chemically responsive heterostructures. Adv. Mater. 20 (13), 2567–2572. Huang, J., Zhang, G., Zhao, X., et al., 2017. Direct detection of dilute solid chemicals with responsive lateral organic diodes. J. Am. Chem. Soc. 139 (36), 12366–12369. Huang, W., Besar, K., LeCover, R., et al., 2012. Highly sensitive NH3 detection based on organic field-effect transistors with tris(pentafluorophenyl)borane as receptor. J. Am. Chem. Soc. 134 (36), 14650–14653. Huang, W., Sinha, J., Yeh, M.-L., et al., 2013. Diverse organic field-effect transistor sensor responses from two functionalized naphthalenetetracarboxylic diimides and copper phthalocyanine semiconductors distinguishable over a wide Analyte range. Adv. Funct. Mater. 23 (33), 4094–4104. W. Huang, X. Zhuang, F. S. Melkonyan, et al. UV–ozone interfacial modification in organic transistors for high-sensitivity NO2 detection. Adv. Mater.: 29(31) 1701706. Huang, Y., Fu, L., Zou, W., et al., 2011. Ammonia sensory properties based on single-crystalline micro/nanostructures of perylenediimide derivatives: core-substituted effect. J. Phys. Chem. C 115 (21), 10399–10404. Irimia-Vladu, M., Troshin, P.A., Reisinger, M., et al., 2010a. Biocompatible and biodegradable materials for organic field-effect transistors. Adv. Funct. Mater. 20 (23), 4069–4076.

Vapor sensing using organic and polymer FET

811

Irimia-Vladu, M., Troshin, P.A., Reisinger, M., et al., 2010b. Environmentally sustainable organic field effect transistors. Org. Electron. 11 (12), 1974–1990. Irimia-Vladu, M., Głowacki, E.D., Troshin, P.A., et al., 2012. Indigo-A natural pigment for high performance ambipolar organic field effect transistors and circuits. Adv. Mater. 24 (3), 375–380. Islam, A.E., 2016. Current status of reliability in extended and beyond CMOS devices. IEEE Trans. Device Mater. Reliab. 16 (4), 647–666. Ji, S., Wang, H., Wang, T., et al., 2013. A high-performance room-temperature NO2 sensor based on an ultrathin heterojunction film. Adv. Mater. 25 (12), 1755–1760. Kang, B., Jang, M., Chung, Y., et al., 2014. Enhancing 2D growth of organic semiconductor thin films with macroporous structures via a small-molecule heterointerface. Nat. Commun. 5, 4752. Kang, M.J., Doi, I., Mori, H., et al., 2011. Alkylated dinaphtho[2,3-b:20 ,30 -f]thieno[3,2-b] thiophenes (Cn-DNTTs): organic semiconductors for high-performance thin-film transistors. Adv. Mater. 23 (10), 1222–1225. Katz, E., Heleg-Shabtai, V., Bardea, A., et al., 1998. Fully integrated biocatalytic electrodes based on bioaffinity interactions. Biosens. Bioelectron. 13 (7–8), 741–756. Katz, H.E., Lovinger, A.J., Johnson, J., et al., 2000a. A soluble and air-stable organic semiconductor with high electron mobility. Nature 404 (6777), 478–481. Katz, H.E., Johnson, J., Lovinger, A.J., et al., 2000b. Naphthalenetetracarboxylic diimide-based n-channel transistor semiconductors: structural variation and thiol-enhanced gold contacts. J. Am. Chem. Soc. 122 (32), 7787–7792. Keren, K., Berman, R.S., Buchstab, E., et al., 2003. DNA-templated carbon nanotube fieldeffect transistor. Science 302 (5649), 1380–1382. Kergoat, L., Herlogsson, L., Braga, D., et al., 2010. A water-gate organic field-effect transistor. Adv. Mater. 22 (23), 2565–2569. Kergoat, L., Piro, B., Berggren, M., et al., 2012. Advances in organic transistor-based biosensors: from organic electrochemical transistors to electrolyte-gated organic field-effect transistors. Anal. Bioanal. Chem. 402 (5), 1813–1826. Khan, H.U., Roberts, M.E., Johnson, O., et al., 2010. In situ, label-free DNA detection using organic transistor sensors. Adv. Mater. 22 (40), 4452–4456. Khan, H.U., Jang, J., Kim, J.J., et al., 2011. In situ antibody detection and charge discrimination using aqueous stable pentacene transistor biosensors. J. Am. Chem. Soc. 133 (7), 2170–2176. Khim, D., Ryu, G.-S., Park, W.-T., et al., 2016. Precisely controlled ultrathin conjugated polymer films for large area transparent transistors and highly sensitive chemical sensors. Adv. Mater. 28 (14), 2752–2759. Kim, C.H., Bonnassieux, Y., Horowitz, G., 2014. Compact DC modeling of organic field-effect transistors: review and perspectives. IEEE Trans. Electron Devices 61 (2), 278–287. Kim, F.S., Ren, G., Jenekhe, S.A., 2011. One-dimensional nanostructures of π-conjugated molecular systems: assembly, properties, and applications from photovoltaics, sensors, and nanophotonics to nanoelectronics. Chem. Mater. 23 (3), 682–732. Laurs, H., Heiland, G., 1987. Electrical and optical properties of phthalocyanine films. Thin Solid Films 149 (2), 129–142. Lebreton, L.C.M., Zwet, J.V.D., Damsteeg, J.W., et al., 2017. River plastic emissions to the World’s oceans. Nat. Commun. 8, 15611. Lee, J., Park, J.H., Lee, Y.T., et al., 2014. DNA-base guanine as hydrogen getter and chargetrapping layer embedded in oxide dielectrics for inorganic and organic field-effect transistors. ACS Appl. Mater. Interfaces 6 (7), 4965–4973.

812

Handbook of Organic Materials for Electronic and Photonic Devices

Li, B., Lambeth, D.N., 2008. Chemical sensing using nanostructured polythiophene transistors. Nano Lett. 8 (11), 3563–3567. Li, H., Dailey, J., Kale, T., et al., 2017. Sensitive and selective NO2 sensing based on alkyl- and alkylthio-thiophene polymer conductance and conductance ratio changes from differential chemical doping. ACS Appl. Mater. Interfaces 9 (24), 20501–20507. Li, X., Shi, W., Yu, X., et al., 2015. Performance improvement of organic field-effect transistor based nitrogen dioxide gas sensor using biocompatible PMMA/silk fibroin bilayer dielectric. J. Mater. Sci. Mater. Electron. 26 (10), 7948–7954. Liao, C., Zhang, M., Niu, L., et al., 2013. Highly selective and sensitive glucose sensors based on organic electrochemical transistors with graphene-modified gate electrodes. J. Mater. Chem. B 1 (31), 3820–3829. Lin, T.W., Hsieh, P.J., Lin, C.L., et al., 2010. Label-free detection of protein-protein interactions using a calmodulin-modified nanowire transistor. Proc. Natl. Acad. Sci. U. S. A. 107 (3), 1047–1052. Lu, A., Zhang, L., Jiang, D., et al., 1994. Study on the stability of a copper phthalocyanine Langmuir-Blodgett film gas-sensitive element. Thin Solid Films 244 (1), 955–957. Lu, J., Liu, D., Zhou, J., et al., 2017. Porous organic field-effect transistors for enhanced chemical sensing performances. Adv. Funct. Mater. 27(20), 1700018. Luo, H., Chen, S., Liu, Z., et al., 2014. A cruciform electron donor-acceptor semiconductor with solid-state red emission: 1D/2D optical waveguides and highly sensitive/selective detection of H₂S gas. Adv. Funct. Mater. 24 (27), 4250–4258. Lv, A., Wang, M., Wang, Y., et al., 2016. Investigation into the sensing process of highperformance H2S sensors based on polymer transistors. Chem. Eur. J. 22 (11), 3654–3659. Lv, A., Pan, Y., Chi, L., 2017. Gas sensors based on polymer field-effect transistors. Sensors 17 (1), 213. Magliulo, M., Mallardi, A., Mulla, M.Y., et al., 2013. Electrolyte-gated organic field-effect transistor sensors based on supported biotinylated phospholipid bilayer. Adv. Mater. 25 (14), 2090–2094. Manoli, K., Dumitru, L., Mulla, M., et al., 2014. A comparative study of the gas sensing behavior in P3HT- and PBTTT-based OTFTs: the influence of film morphology and contact electrode position. Sensors 14 (9), 16869. McQuade, D.T., Pullen, A.E., Swager, T.M., 2000. Conjugated polymer-based chemical sensors. Chem. Rev. 100 (7), 2537–2574. Minami, T., Sato, T., Minamiki, T., et al., 2015. A novel OFET-based biosensor for the selective and sensitive detection of lactate levels. Biosens. Bioelectron. 74, 45–48. Mirza, M., Wang, J., Wang, L., et al., 2015. Response enhancement mechanism of NO2 gas sensing in ultrathin pentacene field-effect transistors. Org. Electron. 24 (Suppl. C), 96–100. Mun, S., Park, Y., Lee, Y.K., et al., 2017. Highly sensitive ammonia gas sensor based on singlecrystal poly(3-hexylthiophene) (P3HT) organic field effect transistor. Langmuir 33 (47), 13554–13560. Nketia-Yawson, B., Jung, A.R., Noh, Y., et al., 2017. Highly sensitive flexible NH3 sensors based on printed organic transistors with fluorinated conjugated polymers. ACS Appl. Mater. Interfaces 9 (8), 7322–7330. Ong, B.S., Wu, Y., Liu, P., et al., 2004. High-performance semiconducting polythiophenes for organic thin-film transistors. J. Am. Chem. Soc. 126 (11), 3378–3379. Pappa, A.M., Curto, V.F., Braendlein, M., et al., 2016. Organic transistor arrays integrated with finger-powered microfluidics for multianalyte saliva testing. Adv. Healthc. Mater. 5 (17), 2295–2302.

Vapor sensing using organic and polymer FET

813

Park, S.J., Kwon, O.S., Lee, S.H., et al., 2012. Ultrasensitive flexible graphene based field-effect transistor (FET)-type bioelectronic nose. Nano Lett. 12 (10), 5082–5090. Park, Y.M., Salleo, A., 2009. Dual-gate organic thin film transistors as chemical sensors. Appl. Phys. Lett. 95 (13), 266. Peng, B., Huang, S., Zhou, Z., et al., 2017. Solution-processed monolayer organic crystals for high-performance field-effect transistors and ultrasensitive gas sensors. Adv. Funct. Mater. 27(29), 1700999. Petti, L., M€unzenrieder, N., Vogt, C., et al., 2016. Metal oxide semiconductor thin-film transistors for flexible electronics. Appl. Phys. Rev. 3 (2), 021303–021357. Salleo, A., Chabinyc, M.L., Yang, M.S., et al., 2002. Polymer thin-film transistors with chemically modified dielectric interfaces. Appl. Phys. Lett. 81 (23), 4383–4385. See, K.C., Becknell, A., Miragliotta, J., et al., 2007. Enhanced response of n-channel naphthalenetetracarboxylic diimide transistors to dimethyl methylphosphonate using phenolic receptors. Adv. Mater. 19 (20), 3322–3327. Seeman, N.C., 2003. Feature DNA in a material world. Nature 6921, 427–431. Shaymurat, T., Tang, Q., Tong, Y., et al., 2013. Gas dielectric transistor of CuPc single crystalline nanowire for SO₂ detection down to sub-ppm levels at room temperature. Adv. Mater. 25 (16), 2269–2273. Shi, W., Zheng, Y., Yu, J., et al., 2016. Mobility enhancement of organic field-effect transistor based on guanine trap-neutralizing layer. Appl. Phys. Lett. 109 (14), 769. Shi, W., Zheng, Y., Taylor, A.D., et al., 2017. Increased mobility and on/off ratio in organic field-effect transistors using low-cost guanine-pentacene multilayers. Appl. Phys. Lett. 111(4), 043301. Shi, W., Yu, J., Katz, H.E., et al., 2018. Sensitive and selective pentacene-guanine field-effect transistor sensing of nitrogen dioxide and interferent vapor analytes. Sens. Actuators B Chem. 254, 940–948. Sirringhaus, H., 2005. Device physics of solution-processed organic field-effect transistors. Adv. Mater. 17, 2411–2425. Sirringhaus, H., 2014. 25th anniversary article: organic field-effect transistors: the path beyond amorphous silicon. Adv. Mater. 26 (9), 1319–1335. Someya, T., Katz, H.E., Gelperin, A., et al., 2002. Vapor sensing with α,ωdihexylquarterthiophene field-effect transistors: the role of grain boundaries. Appl. Phys. Lett. 81 (16), 3079–3081. Someya, T., Pal, B., Huang, J., et al., 2008. Organic semiconductor devices with enhanced field and environmental responses for novel applications. MRS Bull. 33 (7), 690–696. Song, Z., Liu, G., Tang, Q., et al., 2017. Controllable gas selectivity at room temperature based on Ph5T2-modified CuPc nanowire field-effect transistors. Org. Electron. 48, 68–76. Spijkman, M.J., Brondijk, J.J., Geuns, T.C.T., et al., 2010. Dual-gate organic field-effect transistors as potentiometric sensors in aqueous solution. Adv. Funct. Mater. 20 (6), 898–905. Stadler, P., Oppelt, K., Singh, T.B., et al., 2007. Organic field-effect transistors and memory elements using deoxyribonucleic acid (DNA) gate dielectric. Org. Electron. 8 (6), 648–654. Stadlober, B., Zirkl, M., Beutl, M., et al., 2005. High-mobility pentacene organic field-effect transistors with a high-dielectric-constant fluorinated polymer film gate dielectric. Appl. Phys. Lett. 86 (24), 169. Stassen, A.F., De Boer, R.W.I., Iosad, N.N., et al., 2004. Influence of the gate dielectric on the mobility of rubrene single-crystal field-effect transistors. Appl. Phys. Lett. 85 (17), 3899–3901. Steckl, A.J., 2007. DNA—A new material for photonics? Nat. Photonics 1 (1), 3–5.

814

Handbook of Organic Materials for Electronic and Photonic Devices

Steingruber, E., 2000. Indigo and indigo colorants[M]. Wiley-VCH, Germany. Surya, S.G., Ashwath, B.S.N., Mishra, S., et al., 2016. H2S detection using low-cost SnO2 nanoparticle bi-layer OFETs. Sens. Actuators B Chem. 235 (Suppl. C), 378–385. Tian, H., Wang, J., Shi, J., et al., 2005. Novel thiophene-aryl co-oligomers for organic thin film transistors. J. Mater. Chem. 15 (29), 3026–3033. Torsi, L., Dodabalapur, A., Sabbatini, L., et al., 2000. Multi-parameter gas sensors based on organic thin-film-transistors. Sensors Actuators B Chem. 67 (3), 312–316. Torsi, L., Lovinger, A.J., Crone, B., et al., 2002. Correlation between oligothiophene thin film transistor morphology and vapor responses. J. Phys. Chem. B 106 (48), 12563–12568. Torsi, L., Tanese, M.C., Cioffi, N., et al., 2003. Side-chain role in chemically sensing conducting polymer field-effect transistors. J. Phys. Chem. B 107 (31), 7589–7594. Torsi, L., Farinola, G.M., Marinelli, F., et al., 2008. A sensitivity-enhanced field-effect chiral sensor. Nat. Mater. 7, 412–417. Torsi, L., Magliulo, M., Manoli, K., et al., 2013. Organic field-effect transistor sensors: a tutorial review. Chem. Soc. Rev. 42 (22), 8612–8628. Veenstra, G., Webb, C., Sanderson, H., et al., 2009. Human health risk assessment of long chain alcohols. Ecotoxicol. Environ. Saf. 72 (4), 1016–1030. Wang, L., Fine, D., Sharma, D., et al., 2006. Nanoscale organic and polymeric field-effect transistors as chemical sensors. Anal. Bioanal. Chem. 384 (2), 310–321. Wang, L., Zhong, B., Vardoulakis, S., et al., 2016b. Air quality strategies on public health and health equity in Europe—a systematic review. Int. J. Environ. Res. Public Health 13 (12), 1196. Wang, Q., Wu, S., Ge, F., et al., 2016a. Solution-processed microporous semiconductor films for high-performance chemical sensors. Adv. Mater. Interfaces. 3(22)1600518. Welch, M.E., Doublet, T., Bernard, C., et al., 2015. A glucose sensor via stable immobilization of the GOx enzyme on an organic transistor using a polymer brush. J. Polym. Sci. A Polym. Chem. 53 (2), 372–377. Werkmeister, F., Koide, T., Nickel, B., 2016. Ammonia sensing for enzymatic urea detection with organic field effect transistors and a semipermeable membrane. J. Mater. Chem. B 4 (1), 162–168. Wu, S., Wang, G., Xue, Z., et al., 2017. Organic field-effect transistors with macroporous semiconductor films as high-performance humidity sensors. ACS Appl. Mater. Interfaces 9 (17), 14974–14982. Wurthner, F., Stolte, M., 2011. Naphthalene and perylene diimides for organic transistors. Chem. Commun. 47 (18), 5109–5115. Xu, M., Zhang, J., Wang, S., et al., 2010. Gas sensing properties of SnO2 hollow spheres/ polythiophene inorganic–organic hybrids. Sensors Actuators B Chem. 146 (1), 8–13. Yan, H., Park, S.H., Finkelstein, G., et al., 2015. DNA-templated self-assembly of protein arrays and highly conductive nanowires. Science 301 (5641), 1882–1884. Yang, Y., Katz, H.E., 2017. Hybrid of P3HT and ZnO@GO nanostructured particles for increased NO2 sensing response. J. Mater. Chem. C 5 (8), 2160–2166. Yang, Y., Zhang, G., Luo, H., et al., 2016. Highly sensitive thin-film field-effect transistor sensor for ammonia with the DPP-bithiophene conjugated polymer entailing thermally cleavable tert-butoxy groups in the side chains. ACS Appl. Mater. Interfaces 8 (6), 3635–3643. Yu, J., Yu, X., Zhang, L., et al., 2012. Ammonia gas sensor based on pentacene organic fieldeffect transistor. Sens. Actuators B Chem. 173 (Suppl. C), 133–138. Yu, S.H., Cho, J., Sim, K.M., et al., 2016. Morphology-driven high-performance polymer transistor-based ammonia gas sensor. ACS Appl. Mater. Interfaces 8 (10), 6570–6576.

Vapor sensing using organic and polymer FET

815

Yu, X., Zhou, N., Han, S., et al., 2013. Flexible spray-coated TIPS-pentacene organic thin-film transistors as ammonia gas sensors. J. Mater. Chem. C 1 (40), 6532–6535. Zang, Y., Zhang, F., Huang, D., et al., 2014. Specific and reproducible gas sensors utilizing gasphase chemical reaction on organic transistors. Adv. Mater. 26 (18), 2862–2867. Zang, Y., Huang, D., Di, C.-a., et al., 2016. Device engineered organic transistors for flexible sensing applications. Adv. Mater. 28 (22), 4549–4555. Zhang, C., Chen, P., Hu, W., 2015. Organic field-effect transistor-based gas sensors. Chem. Soc. Rev. 44 (8), 2087–2107. Zhang, Y., Hu, W., 2009. Field-effect transistor chemical sensors of single nanoribbon of copper phthalocyanine. Sci. China B Chem. 52 (6), 751–754. Zheng, H., Ramalingam, B., Korampally, V., et al., 2013. Large sensitivity enhancement in semiconducting organic field effect transistor sensors through incorporation of ultra-fine platinum nanoparticles. Appl. Phys. Lett. 103 (19), 193305. Zhu, Z.-T., Mason, J.T., Dieckmann, R., et al., 2002. Humidity sensors based on pentacene thinfilm transistors. Appl. Phys. Lett. 81 (24), 4643–4645.

This page intentionally left blank

Processing and patterning of conducting polymers for flexible, stretchable, and biomedical electronics

25

Tom Kitto*, Come Bodart-Le Guen†, Nicolo Rossetti†, Fabio Cicoira† *Department of Chemical Engineering, University of Bath, Bath, United Kingdom, † Polytechnique Montr
eal, Montreal, QC, Canada

25.1

Introduction

Bioelectronics deals with the applications of electronics in the field of healthcare for the realization of monitoring, diagnostic, and therapeutic devices (Rivnay et al., 2013b). Examples of common bioelectronic devices are pacemakers, pulse oximeters, and electroencephalographic and electromyographic (EMG) systems. In the case of electronic-neural interfaces, current systems make use of materials such as metal alloys and noble metals, which are not suitable for biological interfaces ( Jorfi et al., 2015; Polikov et al., 2005). These materials, which are stiff and undergo potentially harmful chemical reactions once in contact with biological tissues, are often the cause of mechanical and electrical device failure and tissue inflammation (Cogan, 2008; Polikov et al., 2005). Conductive polymers, such as polypyrrole, poly(3,4-ethylenedioxythiophene) (PEDOT), and polyaniline, which are a particular class of organic polymers characterized by the presence of a conjugated-system backbone and the possibility to conduct current upon oxidative doping (Balint et al., 2014; Malliaras and Friend, 2005), can be the best alternative to these materials (Aqrawe et al., 2018). Organic bioelectronic devices made of conductive polymers exploit their ability to conduct both electrons and ions, enabling them to create a suitable interface for the interaction with the biological environment (Aqrawe et al., 2018; Balint et al., 2014; Jorfi et al., 2015; Wilks et al., 2009). Furthermore, conductive polymers are soft and biocompatible, and they can be deposited on flexible and stretchable substrates, which makes them perfect candidates for bioelectronic applications (Rivnay et al., 2013b). It should be noted that the electrical conductivity of conducting polymers is typically on the order of 10
5 to 10
8 S cm
1 and, to achieve conductivities comparable to metals (about 10
4 to 10
5 S cm
1), doping is required (Balint et al., 2014; Bredas and Street, 1985; Malliaras and Friend, 2005). The doping of conductive polymers is achieved by their oxidation/reduction during polymerization, accompanied by the introduction of negative or positive ions, called dopants. Dopants yield enhanced electron/electron hole conductivity (Balint et al., 2014; Bredas and Street, 1985; Malliaras and Friend, 2005). Handbook of Organic Materials for Electronic and Photonic Devices. https://doi.org/10.1016/B978-0-08-102284-9.00025-5 © 2019 Elsevier Ltd. All rights reserved.

818

Handbook of Organic Materials for Electronic and Photonic Devices

One of the most interesting conductive polymers is PEDOT, characterized by good electrochemical stability and low oxidation potential (Aqrawe et al., 2018). Moreover, PEDOT films show exceptionally high conductivity, high charge-storage capacity, and biocompatibility, which are mandatory requirements for biological signal recording and stimulation (Aqrawe et al., 2018; Balint et al., 2014). The most diffused and commercially available doped version of PEDOT is polystyrene sulfonate (PEDOT: PSS), which uses polystyrene sulfonate (PSS
) as the anionic dopant, which is characterized by an enhanced mechanical stability with respect to smaller dopants, water stability, and very high conductivity (exceeding 4600 S cm
1; Worfolk et al., 2015). PEDOT already had been used for the realization of organic thin-film transistors (OTFTs), solar cells, and organic light-emitting diodes (OLEDs), but it also had been largely studied for biomedical applications such as coating for electrodes used in neural recording and stimulation, EMG and surface electromyography (sEMG), and retinal implants (Campana et al., 2014; Frost et al., 2014; Hu et al., 2011; Khodagholy et al., 2011; Kim et al., 2016; Maya-Vetencourt et al., 2017; Samba et al., 2015; Wang et al., 2008; Wilks et al., 2009; Yan et al., 2009; Yi et al., 2016). Among the examples of conductive polymer applications, one of the most interesting and most often studied is the realization of organic field-effect transistors (OFETs) and organic electrochemical transistors (OECTs). The research work on OECTs has focused on understanding device behavior. Biomedical applications of OECTs range from enzyme sensors, such as glucose (Tang et al., 2011) and lactate (Khodagholy et al., 2012) sensors, to cardiac signal biosensors (Campana et al., 2014) and neural actuators (Williamson et al., 2015). A large field of study in organic bioelectronics deals with the processing of conducting polymers to realize highly efficient, scalable, and reproducible devices. Relevant techniques are electropolymerization, spin coating, vapor phase polymerization, and electrospinning. The former allows a fine control of the electrochemical properties of the resulting films, thanks to the wide range of available dopants, solvents, and electrochemical techniques (Balint et al., 2014). Spin coating is the simplest approach; it allows the easy addition of other substances such as conductivity enhancers, plasticizers, and cross-linking agents to enhance film properties (Zhang et al., 2016, 2015). In vapor phase polymerization, a film of the oxidant is exposed to monomer vapor and polymerization is localized to the interface, which allows the formation of thin, uniform, and highly conductive films ( Jimison et al., 2012; Lawal and Wallace, 2014). Electrospinning allows the creation of microfibers and nanofibers, with high conductivity and unique mechanical properties suitable as scaffolds and stretchable mats (Balint et al., 2014). A key challenge has been the application of state-of-theart device-development techniques, such as photolithography, to organic devices. Patterning organic materials raises new challenges, though, as the choice of solvents and photoresists may damage the materials. For this reason, new patterning routes, such as the use of parylene and orthogonal solvents, have been explored (Zhang and Cicoira, 2017; Zhang et al., 2016). Finally, novel studies have been conducted on the self-healing properties of conducting polymers. Recent discoveries on the topic of PEDOT:PSS film self-healing may lead to the development of highly robust and efficient biomedical devices (Zhang and Cicoira, 2017).

Processing and patterning of conducting polymers

25.2

819

Organic conducting polymer-based transistors

In OTFTs, the amplifying and switching properties of transistors can be localized to the biological interface in ways that are not achievable by traditional, rigid, siliconbased electronics (Khodagholy et al., 2013). The application of small gate voltages (about 1 V) can allow the detection of weak biosignals generated by analytes (Liao et al., 2015; Zhang et al., 2015). OTFTs can be divided into two classes: OFETs and OECTs. In OFETs, a dielectric material separates the gate electrode and channel. Polymer doping is achieved by the effects of an electrostatic double layer (EDL) on the channel material. In OECTs, an electrolyte medium separates the two elements, and ions can penetrate the channel, subsequently doping or dedoping it (Lin and Yan, 2012; Simon et al., 2016).

25.2.1 Working principle of organic electrochemical transistors (OECTs) OECT devices consist of a gate electrode, electrolyte solution, polymer semiconductor channel, and source and drain electrodes. Fig. 25.1 shows the schematic diagram of a typical OECT, where the channel is defined by the overlapped region of the polymer and electrolyte. Extensive work has been performed in the recent years in order to gain a greater understanding of the OECT working mechanism. The influence and role of polymer channel material and electrolytes on OECT performance have been some of the most-investigated aspects. In addition, studies have investigated the influence of electrode material and device geometry on behavior (Kumar et al., 2015). A reference to the previous handbook edition will give the governing equations for OECT behavior (Tarabella et al., 2013). The following is the equation for transistor transconductance (Khodagholy et al., 2013):

Gate Vg

Electrolyte t Source

Organic active material

Drain electrode

Substrate Vds

Fig. 25.1 Schematic diagram of an OECT device. The overlap of electrolyte with the organic active material channel, of thickness t, defines the channel length, and Vg and Vds are the gate and source-drain voltages, respectively.

820

Handbook of Organic Materials for Electronic and Photonic Devices

gm ¼

ΔIds ΔVg

(25.1)

where, transconductance, gm, gives the signal amplification (i.e., the modulation in the source-drain current arising from a change in gate voltage); and △ Ids and △ Vg are changes in the source-drain current and gate voltage, respectively. Dimension characteristics are t, W, and L, standing for the OECT channel thickness, width, and length, respectively. Other key parameters include response time and on/off ratio. The response time is the delay between a change in the applied Vg and the successive modulation in source-drain current, Ids. This is typically slower in OECTs, as their operation relies on ion penetration and diffusion out; it also depends on the speed and efficiency of the redox reactions occurring within the bulk polymer and the electrolyte characteristics (Giovannitti et al., 2016; Mracek et al., 2015). The on/off ratio expresses the ratio of Ids when the device is on or off. Depletion mode transistors are in the on state in the absence of gate bias and in the off state when minimum Ids is achieved (Andersson Ersman et al., 2013). The transfer characteristics of a PEDOT:PSS micro-OECT are shown in Fig. 25.2. The operation regime of an OECT also determines device behavior. Two regimes are possible: a Faradic regime, where oxidative doping of the polymer channel results from volumetric ion penetration in the bulk; and a non-Faradic regime, where oxidative doping is coupled to the formation of an EDL at the gate/electrolyte and electrolyte/channel interfaces. Depending on the operation regime, the device is referred to as either an OECT or an EDL field effect transistor (EDL FET), respectively (Giovannitti et al., 2016; Lin and Lonergan, 2006).

Fig. 25.2 Electrical characteristics of PEDOT:PSS micro-OECTs on PDMS with channel width of 4000 μm, length of 10 μm, and thickness of 400 nm (shown in optical micrograph). (A) Transfer characteristics (with Vg varying from
0.2 to 0.8 V) and (B) transfer characteristics with associated transconductance (Vds ¼ 0.2 V). Data from Zhang, S., Hubis, E., Tomasello, G., Soliveri, G., Kumar, P., Cicoira, F., 2017. Patterning of stretchable organic electrochemical transistors. Chem. Mater. 29, 3126–3132. Copyright 2017 American Chemical Society.

Processing and patterning of conducting polymers

821

25.2.2 PEDOT:PSS-based organic electrochemical transistors (OECTs) PEDOT:PSS has seen extensive use in OECT devices as the channel material. As a p-type material, PEDOT:PSS OECTs operate in depletion mode: Upon the application of a positive Vg, cations enter the polymer bulk and compensate for the PSS
anions, decreasing hole density in the PEDOT (Khodagholy et al., 2013). Considerable work on OECT behavior has been performed by the groups of Malliaras, Someya, Berggren, Jivnay, Yan, and Cicoira, among others. Khodagholy et al. reported the fabrication of a PEDOT:PSS-based OECT with a peak transconductance of 2.7 mS, outperforming devices from both traditional and emerging technologies. The device operated with a near-constant transconductance from Vg between DC and a frequency of 1 kHz. The OECT was processed onto a flexible parylene-C substrate and showed minimal changes in transconductance and response times after crumpling (Khodagholy et al., 2013). The same group also reported the development of an OECT with maximum transconductance at zero gate bias. Minimizing the value of Vg that is required to achieve this is important at cell interfaces, as they can be sensitive to prolonged bias application (Rivnay et al., 2013a). The use of different electrolytes has given insight into the operating mechanism of OECTs. Common electrolytes are ionic liquids (ILs) and aqueous ionic solutions. ILs are exciting media due to their low volatility, thermal stability, and wide windows of electrochemical stability. Mracek et al. investigated the use of three ILs [1-ethyl-3-methylimidazolium-bis(trifluoromethyl-sulfonyl)imide (EMIM TFSI), 1-butyl-3-methylimidazoliumtrifluoromethanesulfonate (BMIM OTf ), and 1-ethyl-3-methylimidazoliumtetrafluoroborate (EMIM BF4)] as the electrolyte in PEDOT:PSS devices. The results showed that higher-viscosity ILs gave higher OECT on/off ratios. On/off switching times, tON/OFF, were consistently slower than off/on switching times, tOFF/ON, due to the need for ions to diffuse out of the channel material, although this process could be sped up with the application of a negative Vg. This behavior was independent of electrolyte viscosity and conductivity, although smaller ions exhibited faster current kinetics due to higher mobility (Mracek et al., 2016). Yi et al. investigated the use of triisobutyl(methyl)phosphonium tosylate (Cyphos IL 106) as the electrolyte. The authors applied the IL in its pure form and as H2O-IL binary mixtures and ion gels. The ion gels were prepared with the triblock copolymer polystyrene-b-poly(methylmethacrylate)-b-polystyrene (SMMAS) at SMMAS: Cyphos IL 106 mass ratios of 1:10 and 1:5. In all cases, H2O-IL binary mixtures and ion gels exhibited increased current modulation with respect to pure Cyphos IL 106, although performance was relatively similar for both ion gels. In contrast, the water content of H2O-IL mixtures influenced performance heavily; experiments utilizing activated carbon (aC) gate electrodes for PEDOT:PSS-channel devices suggested a weak dependence of current modulation on electrolyte viscosity and ionic conductivity. For H2O-IL mixtures, the on/off ratio strongly depended on water content and peaked at  5000 for 10% v/v H2O-IL. A further increase saw a fall in the ratio: at 99% v/v on/off ratio was identical to the pure IL at  40. These results do not show a substantial relationship between on/off ratio and viscosity or ionic conductivity, from which the group concluded that

822

Handbook of Organic Materials for Electronic and Photonic Devices

the complex ion associations within the electrolyte influenced the doping/dedoping process (Yi et al., 2015). Ersman et al. demonstrated the use of carbon paste drain/source electrodes extending into the OECT channel, on top of the channel, and in contact with both the channel and electrolyte, to improve tON/OFF. During operation, a reduction front (i.e., the front of reduced PEDOT:PSS in the polymer bulk) would extend outside the channel. By extending the carbon paste into the channel, particularly over the negative drain electrode, the authors found that the reduction front could be prevented. Their devices exhibited tON/OFF of 10 ms—more than 10 times faster than devices with no paste— and could simply be fabricated as all-printed OECTs (Andersson Ersman et al., 2013). Stavrinidou et al. investigated the mobilities of several ions (H+, K+, Na+, and C5H14NO+) as aqueous electrolytes in PEDOT:PSS films. Larger ions exhibited lower mobility and, additionally, water uptake by PEDOT:PSS was found to swell the film and allow easier ion transport (Stavrinidou et al., 2013). Kumar et al. investigated the use of two electrolytes [0.01 M of sodium chloride (NaCl) and 0.001M of cetyltrimethyl ammonium bromide (CTAB)] in a PEDOT:PSS channel and source/drain electrode devices with an aC gate electrode. Work detailed in the previous edition of this handbook demonstrated the high current modulation achieved by CTAB above its critical micelle concentration (Tarabella et al., 2013). Cyclovoltammetry (CV) of PEDOT:PSS films in the two electrolytes showed no dependence of the PEDOT:PSS redox peak position on the electrolytes, although obtained voltammograms were more distorted for CTAB, indicating slower doping/dedoping speeds compared to sodium chloride (NaCl). Further investigation with electrochemical impedance spectroscopy (EIS) confirmed a hindered ionic charge transport through the polymer film in the case of CTAB. Calculated time constants, τ, for the doping/dedoping processes were 0.8 s and 1.7 s at 50-nm film thickness and 2.4 s and 5.0 s for 500-nm film thickness for NaCl and CTAB, respectively. This was attributed to the lower ionic conductivity and larger ion size of CTAB. Under air purging, CV also was employed to investigate the effect of dissolved oxygen in the electrolyte. Increased cathodic current (between
0.5 V and
0.8 V) indicated a chemical reoxidation of the PEDOT:PSS film by dissolved O2 after reduction. For the CTAB electrolyte, reoxidation was partially prevented due to lower O2 solubility (Kumar et al., 2015). The effect of device dimensions has been shown to allow extensive control over OECT behavior. A reference to the OECT governing equations in the previous handbook edition will show that Ids increases with both W/L and (Tarabella et al., 2013). Experimentally, it has been shown that a higher value of W/L and a larger thickness both increase maximum transconductance, gm, max, and thus the on current, although a larger Vg is needed to achieve this. Alternatively, this required value of Vg can be independent of channel area, given that the ratio of W/L to t is constant. Thinner channels also tend to yield faster response times, as fewer ions are required to dedope the channel and, thus, less time is required for complete ion penetration into and subsequent diffusion out of the film (H€ utter et al., 2013; Khodagholy et al., 2013; Kumar et al., 2015; Rivnay et al., 2013a). When investigating the influence of CTAB and NaCl electrolytes, Kumar et. al. also investigated the effect of channel thickness on OECT performance. Device Ids

Processing and patterning of conducting polymers

823

was observed to increase with thickness, confirming the observation that the PEDOT: PSS charge transport process occurs within the entire channel volume. The on/off ratio simultaneously decreased, but this depended on the electrolyte: The on/off ratios for CTAB were 50 and 220, respectively; for NaCl, they were 7.0 and 7, respectively. CV showed an increased current with film thickness, which attributed to the increased volume of PEDOT:PSS available for doping/dedoping. The authors calculated a sixfold increase in film capacitance between 50-nm and 500-nm films, although volumetric capacitance (and thus doping/dedoping effects) were higher in thinner films. Finally, EIS results indicated lower electronic resistance for thick films and higher τ for the doping/dedoping process, as mentioned previously (Kumar et al., 2015). The effect on device dimensions on signal-to-noise ratio (SNR) was investigated by Stoop et al. Large channel surface areas maximized the SNR (Stoop et al., 2017). Tang et al. compared the performance of aC and PEDOT:PSS as the gate electrolyte. The devices were fabricated with a PEDOT:PSS channel, source and drain electrodes, and a 0.01-M NaCl electrolyte. Due to its high specific surface area and, thus, high specific double-layer capacitance and electrostatic storage capacity, aC is a material of specific interest. Tests showed that aC gates exhibited better current modulation: At a source/drain voltage, Vds, of –0.4 V, Ids varied by factors of 10 compared to 3, for aC and PEDOT:PSS gates, respectively, over a Vg range of
0.4 to 0.6 V. On/off ratios were also larger, at 500 and 15, respectively. CV was employed to further investigate the behavior of devices with either gate material. Results indicated that in PEDOT:PSS-gated devices, channel dedoping would be accompanied by gate doping (and vice versa), and the channel potential would not reflect the Vg experienced by a reference electrode (RE). Conversely, the high double-layer capacitance of aC would allow aC-gated device channel potential to be unaffected by Faradic effects at the gate electrode (Tang et al., 2015). A successive study reported nanostructured carbon (nsC)-gated OECTs with a PEDOT:PSS channel and poly(sodium 4-styrenesulfonate) (PSSNa) gel electrolyte on flexible polyethylene terephthalate (Mylar) substrates. These devices exhibited good current modulation with an on/off ratio of 60, but, more important, also exhibited joint transistor-capacitor (TransCap) behavior. By coupling a lateral, purely capacitive nsC electrode to a pseudocapacitive PEDOT:PSS electrode, a hybrid supercapacitor was fabricated. When a Vg was applied, each electrode was charged with a △Q and, during subsequent rest, partial discharge occurred. The devices showed voltage retentions, or lower self-discharge, of 80%. This device would be of great interest for autonomous systems in flexible/stretchable electronics (Sayago et al., 2014; Yi et al., 2016).

25.3

Processing methods for PEDOT

PEDOT can be processed in numerous ways. In these sections, we will review three methods used to process PEDOT in different forms: solution-processed PEDOT:PSS films, electrodeposited PEDOT coatings, and electrospun PEDOT fibers.

824

Handbook of Organic Materials for Electronic and Photonic Devices

25.3.1 Solution processing for PEDOT:PSS Spin-coating or drop-casting of PEDOT:PSS from commercial suspensions (typically Hareus Clevios or Agfa Orgacon, followed by baking) is a common method for PEDOT:PSS film deposition in organic electronics. An advantage of PEDOT:PSS solution processing is the possibility of mixing it with additives or treating the deposited film with a large variety of conductivity and flexibility enhancers and crosslinkers. The low conductivity of PEDOT:PSS films obtained from pristine commercial solutions (around 1 S cm
1) has pushed researchers to find ways to increase the conductivity of these films with additives, referred to as secondary dopants or, more appropriately, conductivity enhancers. Their most probable influence on PEDOT: PSS is to alter film morphology during drying, leading to a lower energy barrier for charge transport (Elschner and L€ ovenich, 2011). The conductivity enhancers for PEDOT:PSS should have good solubility in water and possess highly polar groups that can act as effect dopants. Wang et al. studied the use of stretchability and electrical conductivity (STEC) enhancers and identified a number of effective chemicals, such as dioctyl sulfosuccinate sodium salt, sodium dodecylbenzenesulfonate, dodecylbenzenesulfonic acid (DBSA) and ionic liquids (Wang et al., 2017). Zhang et al. also studied the effect of various solvents on the conductivity and stretchability of their PEDOT:PSS films. The highest conductivities, 600–700 S cm
1, were obtained for glycerol (5 v/v %), dimethylsulfoxide (DMSO; 5 v/v %) and sorbitol (2.5 wt.%). The surfactant DBSA was also found to significantly influence the film conductivity, as adding 2 v/v % of DBSA lead to a conductivity of 500 S cm
1, but a concentration higher than 0.5 v/v % induced a phase separation which resulted in difficult spin-coating (Zhang et al., 2015). Acid treatment appears to be extremely relevant: PEDOT:PSS acid posttreatment can reach around 2400 S cm
1 after the treatment with 1.5 M H2SO4 and 3065 S cm
1 after treated with 1 M H2SO4 three times (Xia et al., 2012). Ouyang et al. obtained similar high-conductivity PEDOT:PSS (higher than 3300 S cm
1) after treatment by methanesulfonic acid, with the relevant advantage of not being corrosive (Ouyang, 2013). Rinsing with alcohols also can significantly increase the conductivity of PEDOT:PSS films. Oh et al. found methanol to be the most suitable solvent because it dissolves excess PSS without changing the PEDOT morphology (Oh et al., 2014). The record conductivity for PEDOT:PSS films after combined treatments (Zonyl, methanol and sulfuric acid) was 4600 S cm
1 (Worfolk et al., 2015). Some authors used cross-linking agents, such as 3-glycidoxypropyltrimethoxysilan e (GOPS), to improve the mechanical stability of their PEDOT:PSS films. Zhang et al. observed that GOPS lead to an increased film thickness and a decreased conductivity, most likely due to nonconductive species being trapped in the PEDOT:PSS film. However, the PEDOT:PSS films combined with GOPS showed no decrease in film thickness after immersion in water (Zhang et al., 2015). The use of divinylsufone (DVS) as a cross-linker was also reported, with the advantage over GOPS that it improves the mechanical stability of the films without decreasing their conductivity (Mantione et al., 2017).

Processing and patterning of conducting polymers

825

For flexible electronics, flexibility is required not only for the substrate, but also for the PEDOT film. This can be achieved through the addition of flexibility enhancers such as Zonyl-FS300 (Lipomi et al., 2012), PEG (Li et al., 2015) and Triton-X100 (Oh et al., 2014), although the latter’s toxicity has been reported (Dayeh et al., 2004). Still, a compromise needs to be reached as some stretchability enhancers, such as DBSA, can influence the conductivity. Below a threshold concentration, and DBSA leads to an improved conductivity; above it, and conductivity falls (Zhang et al., 2015) Functionalization of PEDOT films with carboxylic groups for successive protein attachment has been explored. Berezhetska et al. reported the formation of COOHfunctionalized PEDOT:PSS film with excellent conductivity and high stability by mixing a commercial PEDOT:PSS suspension with carboxymethylated dextran (CMD), GOPS, DBSA, and glycerol (Berezhetska et al., 2015). The availability of dCOOH groups proved relevant for covalent coupling of biomolecules on the PEDOT:PSS/CMD film.

25.3.2 Electrochemical deposition of PEDOT Another deposition technique for PEDOT is electrochemical polymerization of EDOT, although it is limited to conductive substrates. Fig. 25.3 gives the schematic diagram of the electrodeposition setup. Potentiostatic a

t os

n lva

a

G

tic

Po t

en

tio

Voltage source

dy n

am

ic

V

Working electrode – The component/substrate to be coated with the conductive polymer. The synthesis will take place upon the surface of the component.

Control electrode

Polymer solution Containing the monomers of the polymer, dopant ions/molecules and the solvent.

Fig. 25.3 Schematic diagram of the electropolymerization setup. From Balint, R., Cassidy, N.J., Cartmell, S.H., 2014. Conductive polymers: towards a smart biomaterial for tissue engineering. Acta Biomater. 10, 2341–2353. Copyright 2014 Elsevier.

826

Handbook of Organic Materials for Electronic and Photonic Devices

The typical setup for this electropolymerization is a three-electrode cell (working, counter, and reference electrodes) (Elgrishi et al., 2017). The working electrode (WE) exchanges electrons with the solution, hence activating the monomers. This electrode’s material should be conductive and electrochemically inert within the working potential window, and the surface should be cleaned or “pretreated” by mechanical polishing, sonication and other methods (Fischer et al., 2009). The counterelectrode (CE) closes the electrical circuit, allowing current to flow between the WE and the CE. Usually, CEs are made of platinum. The RE possesses a well-known electrochemical potential, which can be used as a reference for the potentials. Common REs are the saturated carbon electrode (SCE) and Ag/AgCl electrode. The monomer, EDOT, is dissolved in the electrolytic solution of an electrochemical cell. When a potential is applied at the WE, the monomer is oxidized by a current flowing through the solution. Once oxidized, the monomers, now radical cations, can react to form an insoluble polymer, physically adsorbed at the surface of the working electrode. This polymer will continue to grow, so long as current is supplied or dissolved monomers are available. There are three different electrochemical deposition routes, known as potentiostatic, galvanostatic, and potentiodynamic (PD). The potentiostatic technique consists of the application of a constant potential at the WE for a certain amount of time. The galvanostatic technique is similar, but with the application of a constant current instead of a constant potential. In the PD technique, also called CV, the potential is swept across a potential window at a specific scan rate multiple times. Some reports suggested that the PD technique provides a smoother and more homogeneous coating of PEDOT (Castagnola et al., 2014b; Vomero et al., 2017). Electrodeposition possesses several advantages over chemical deposition. The electrolyte used in the solution is also the dopant of the conducting polymer, usually anions like PSS
(Ganji et al., 2017), ClO4
(Green et al., 2012), BF4
(Wang et al., 2010), and pTS
(Green et al., 2012). During the electrodeposition, ions from the electrolyte (typically anions) enter the polymer matrix and become the dopants of the deposited coating. This control of the doping level contributes to the high conductivity of electrodeposited films of PEDOT (Aregueta-Robles et al., 2014). Electrodeposition also provides a better control of the deposited layer’s structure (i.e., thickness, morphology, and targeted area). Electropolymerization nonetheless also has some drawbacks. It is limited to monomers that can be oxidized electrochemically, fortunately the case for most conducting polymers (i.e., PPy, PEDOT, and PANI), and the substrate must be conductive. The more recent concern for electrochemical deposition is its adhesive capacity. Apart from some substrates like gold (Castagnola et al., 2015) and glassy carbon (Vomero et al., 2017), conducting polymer electrodeposited layers tend to detach from the substrate surface after suffering mechanical stress, sonication or sterilization (Boehler et al., 2016; Green et al., 2012; Vomero et al., 2017; Yang et al., 2005). It also has been reported that PEDOT:PSS is prone to delamination and cracks when polymerized via galvanostatic route (Cui and Zhou, 2007). Some solutions have been developed to address this issue: EDOT-NH2 anchoring, through electrografting of the amine moieties (Ouyang et al., 2017), EDOT-acid chemical deposition via

Processing and patterning of conducting polymers

827

chemisorption of the carboxylic groups (Wei et al., 2015), electrochemical copolymerization with EDOT (Povlich et al., 2013), and diazonium salt anchoring through electrografting (Blacha et al., 2012). These methods lead to the formation of a strongly bonded organic layer on which PEDOT can subsequently grow. Finally, roughening the substrate surface, for instance gold-etching using iodine solution, has been proven relevant (Pranti et al., 2017).

25.3.3 Electrospinning Polymer nanofibers are a high potential application of conducting polymers. The intrinsic high surface area-to-volume ratio, porosity, simple surface functionalization, and good mechanical properties of fibers have enabled their use in a diverse range of applications from tissue scaffolds to nanoelectronics and smart clothing. Many established techniques exist for their fabrication, though electrospinning has emerged as the dominant process of choice. This is due to its simplicity, flexibility, and inexpensive and scalable nature (Boub
ee de Gramont et al., 2017; Teo and Ramakrishna, 2006; Tiwari and Venkatraman, 2012). The electrospinning apparatus, in its simplest form, consists of a syringe, containing a polymer solution or melt, with a charged metallic needle tip, a grounded or oppositely charged collector, and a high-voltage power supply. As the polymer solution is ejected from the needle tip, an electric charge is imparted onto its surface. As electric field strength is increased, this electrostatic charge reaches equilibrium with the surface tension of the solution and a Taylor cone is formed. Increasing voltage past this point causes a jet to eject from the tip of the Taylor cone. This jet passes through a region of instability, where the solution’s solvents evaporate, and is deposited onto the substrate as a fiber. Although the process itself is simple, fiber morphology depends heavily on numerous interrelated parameters. These can be broadly classified into one of three groups: solution, process, and ambient parameters. Key parameters include solution viscosity and polymer concentration, applied voltage, needle-to-collector distance, and ambient humidity. All the factors influence final fiber morphology and diameter through their impact on Taylor cone formation, instability development, and solvent evaporation (Boub
ee de Gramont et al., 2017; Thompson et al., 2007; Tiwari and Venkatraman, 2012). Electrospinning of conducting polymers has been a major challenge in their processing due in part to low molecular weight, limited solubility in solvents, and the presence of a rigid conjugated backbone. Often, conducting polymers are spun in blends in the presence of a carrier polymer, commonly poly(ethylene oxide) (PEO), poly(vinylpyrrolidone) PVP, or poly(vinyl alcohol) (PVA; Bessaire et al., 2017). PEDOT:PSS has seen extensive use in blends. A work by the group of Kara and Frey investigated the effect of numerous solvents, common to film processing, on the morphology and conductivity of electrospun PEDOT:PSS-PVA fibers. The role of these solvents in electrospinning, however, is more complex than in film counterparts because solvent properties play a major role in fiber production. The authors spun blends of an aqueous dispersion of PEDOT:PSS, 4% wt/wt PVA, the surfactant

828

Handbook of Organic Materials for Electronic and Photonic Devices

Triton X-100 at 0.5% wt/wt, and 5% wt/wt solvent. The solvents DMSO and EG provided the optimum fiber morphologies, which was attributed to the action of the solvents during electrospinning. When solution spinnability (i.e., the ease of electrospinning fibers) was low, as with dimethylformamide (DMF), fibers possessed a heterogeneous distribution of PEDOT:PSS, concentrated within beads; when it was high, as with EG, distribution was homogeneous. Further Raman spectroscopy, X-ray photoelectron spectroscopy (XPS), and X-ray diffraction (XRD) analysis indicated that solvent addition, particularly EG and DMSO, drove a transformation in the PEDOT:PSS chain structure from random to extended coil. Substantial increases in fiber conductivities were observed over fibers spun in the absence of a solvent. The presence of DMSO- and EG-enhanced conductivity by the factors of 15 and 30, respectively, which was attributed to this change in chain structure (Kara and Frey, 2014). An alternative method for PEDOT:TOS fiber fabrication was developed by Laforgue and Robitaille. Here, a two-step, in situ polymerization process was employed to avoid the need for a carrier polymer, thus improving fiber conductivity. First, an oxidant-containing solution with sacrificial polymer PVP was electrospun onto a substrate as a fiber mat. The mat then underwent vapor-phase polymerization, during which a layer of PEDOT:TOS polymerized on the fiber surfaces. The PVP was then removed with a solvent, causing a collapse in the PEDOT:TOS coating and giving the final fibers a rough surface morphology. XRD suggested a high ordering of PEDOT chains, which was attributed to the templating of tosylate ions, and a mat conductivity  60 S cm
1 was obtained (Laforgue and Robitaille, 2010). The conformability and stretchability of PEDOT fibers have been demonstrated since. Bessiare et al. electrospun PEDOT:PSS blends containing the carrier polymer PEO and the solvents DMF and EG onto nonplanar surfaces. The fiber mat sheet resistances were constant across all structures at  20,000 Ohms, whereas PEDOT:PSS thin films spin-coated onto the same structures showed near-200-fold increases. After electrospinning, the group soaked the fiber mats in ethanol and EG to remove the carrier PEO and further dope the PEDOT:PSS (Bessaire et al., 2017). Boub
ee de Gramont adapted the process developed by Laforgue et al. to fabricate highly stretchable PEDOT:TOS mats on polydimethylsiloxane (PDMS) substrates. Upon stretching, initial, permanent current losses were observed due to the breaking of fibers. However, subsequent stretches to the same length caused minimal change in resting currents, in both relaxed and stretched states, thus demonstrating the high stretchability of PEDOT:TOS fiber mats (Boub
ee de Gramont et al., 2017).

25.4

Patterning organic devices

The main way to fabricate organic electronics that currently exists is patterning. This term comprises many routes of fabrication, with their own advantages and weaknesses. New, emerging techniques, such as elastomer stamps and ink-jet printing (Roberts et al., 2016), are promising but lack the high resolution obtainable by other industry-standard methods.

Processing and patterning of conducting polymers

829

The most used technique for ultrahigh resolution in patterning devices is photolithography. This technique employs photoresists, compounds that can be rendered soluble or insoluble by ultraviolent (UV)-irradiation, and a developer solution to remove them and shape a pattern onto a substrate through a mask. Conventionally, this method is unsuitable for organic devices, as most photolithography chemical photoresists are miscible with organic semiconductors and thus can alter their electronic properties. Moreover, photoresists suffer from poor adhesion on elastomer substrates and swelling due to organic solvents. PEDOT:PSS films also suffer from deterioration during conventional lithography, as it can be damaged or dissolved in developer and aqueous alkaline solutions. A simple approach is to use a protection layer. Ouyang et al. deposited a thin layer of evaporated silver on a PEDOT:PSS film to shield it from the solvents and photoresists employed (Ouyang et al., 2014b). The etching of the silver interlayer was realized using nitric, phosphoric, and acetic acids and deionized water. As PEDOT:PSS films are stable in acids, no damage was caused to the film; a common PEDOT:PSS posttreatment is done by acid treatment to enhance film conductivity (Ouyang, 2013). Alternative solvents that are nonaggressive to organic materials (namely, orthogonal solvents) are relevant for patterning involving PEDOT:PSS processing. Supercritical carbon dioxide (CO2) is one of them: it is a fluid with poor solubility for ionic and high-molecular-weight organic compounds, combined with so-called green chemical properties (i.e., it is nonflammable and environment friendly). Other commonly used orthogonal solvents are fluorinated, and thus they are immiscible with water and organic chemicals and poor solvents for nonfluorinated materials. Parylene transfer patterning is a promising technique for the transfer of patterns formed using a Parylene mask onto a plastic substrate as Parylene is nearly inert to most chemicals. Zhang et al. showed that it was possible to pattern a Parylene film, deposit it onto a surfactant-treated polyethylene terephthalate (PET) substrate, and subsequently transfer it onto a “stickier” PDMS substrate, where it was used as a mask for the deposition of metal electrodes. This technique greatly facilitated the patterning of metals on PDMS, which is otherwise challenging with conventional lithography. The process is performed at room temperature and does not require water. The authors then finalized the fabrication of PEDOT:PSS-channel OECTs with 5-μm channel lengths using orthogonal lithography (Zhang and Cicoira, 2017). This process is shown in Fig. 25.4. A one-step photolithographic method for PEDOT:PSS/PEG films, avoiding the use of photoresist or a protection layer, was reported by Zhu et al. A solution containing poly(ethylene glycol) diacrylate (PEGDA) was spread between a PDMS layer and a PEDOT:PSS-coated glass substrate. After photopolymerization through UV exposure of the PEGDA, the glass substrate was removed, leaving behind a PDMS layer covered with PEDOT:PSS linked to the PEG matrix. The PEDOT:PSS/PEG film on PDMS featured high conductivity and flexibility (Zhu et al., 2017). An alternative technique is silk protein lithography, which was developed by Pal et al. The researchers developed a photopatternable and water-soluble ink composed of a sericin protein photoresist (SPP) mixed with PEDOT:PSS in suspension. This SPP-PEDOT:PSS ink (10%–20% wt/wt) could be spin-coated or casted onto various

830

Handbook of Organic Materials for Electronic and Photonic Devices

Fig. 25.4 The Parylene transfer process for the fabrication of a PEDOT:PSS microelectrode array. (A) Detaching CTAB-treated, PET-carried Parylene patterns from glass substrate; (B) laminating PET/Parylene onto PDMS; (C) depositing Parylene patterns onto PDMS, detaching PET; (D) metal deposition; (E) lift-off of Parylene film; (F) electrode arrays in PDMS. From Zhang, S., Hubis, E., Tomasello, G., Soliveri, G., Kumar, P., Cicoira, F., 2017. Patterning of stretchable organic electrochemical transistors. Chem. Mater. 29, 3126–3132. Copyright 2017 American Chemical Society.

substrates, such as PDMS, glass, indium tin oxide (ITO), and silicon, and then it was photopatterned and subsequently developed with water. The extensive use of water improves the simplicity of the process and reduces the environmental impact. The same authors processed the SPP-PEDOT:PSS ink onto photofibroin substrates. Photofibroin possesses exposed residual functional groups on its surface which, during photopatterning, allowed covalent conjugation between the substrate and SPPPEDOT:PSS ink through silane chemistry, improving PEDOT:PSS film adhesion (Pal et al., 2017, 2016). For patterning flexible or stretchable electronics, the need for a specific substrate that can be prestretched, like PDMS (Zhang and Cicoira, 2017), or features high flexibility, such as PDMS (Lipomi et al., 2012; Zhu et al., 2017), polyimide (Vomero et al., 2016), and PET (Yi et al., 2017; Zhang et al., 2016), is paramount.

25.5

Biomedical applications

The applications for PEDOT are numerous in bioelectronic medicine (Aqrawe et al., 2018; Aregueta-Robles et al., 2014; Balint et al., 2014; Fang et al., 2015; Green and Abidian, 2015; Jorfi et al., 2015; Martin and Malliaras, 2016; Rivnay et al., 2013b; Simon et al., 2016). Indeed, besides being highly conductive compared to other conducting polymers, PEDOT complies with most criteria: It is stretchable, mechanically soft, biocompatible, chemically stable, and able to sustain millions of electrical pulses. These properties are highly customizable and can be extensively influenced by processing conditions, as well as functionalization and the dopants used

Processing and patterning of conducting polymers

Solvent

Temperature pH

831

Conductivity

Synthesis

Method

Doping

“Small” dopant

“Large” dopant

Electric potential

Biodegradability Drug release Functionalization

Physical properties

Biocompatibility

Application

Color

Structure

Porosity

Mechanical properties

Fig. 25.5 The interconnected world of organic conducting polymers. From Balint, R., Cassidy, N.J., Cartmell, S.H., 2014. Conductive polymers: towards a smart biomaterial for tissue engineering. Acta Biomater. 10, 2341–2353. Copyright Elsevier 2014.

(Castagnola et al., 2014a; Green et al., 2013; Green et al., 2012; Vomero et al., 2017). Fig. 25.5 shows the interconnected nature of conducting polymer properties and their applications.

25.5.1 PEDOT for neural implants and electromyography PEDOT coatings can be deposited on implantable electrodes for recording (Anastassiou et al., 2016; Castagnola et al., 2015; Einevoll et al., 2013; Ludwig et al., 2006) and stimulation (Bronstein et al., 2011; Kringelbach et al., 2007; Stefani et al., 2007; Weaver et al., 2009). Some studies also have indicated that electropolymerization of PEDOT can be achieved in vivo after injection of the dissolved monomer near the implanted electrode (Ouyang et al., 2014a;

832

Handbook of Organic Materials for Electronic and Photonic Devices

Richardonson-Burns et al., 2006). Arrays used for electromyography can be coated with PEDOT (Kim et al., 2016; Roberts et al., 2016; Tian et al., 2016). OECTs could be used for recording purposes, although the requirement of an additional wire remains a practical problem. It has been reported repeatedly that because of their porous and rough morphology, PEDOT coatings increase the charge storage capacity (Castagnola et al., 2015; Cui and Zhou, 2007; Ganji et al., 2017; Green et al., 2013, 2012; Venkatraman et al., 2011) and charge transfer capacity (Vomero et al., 2017; Wilks et al., 2009) of electrodes and contributes to a drastic reduction of impedance (Castagnola et al., 2014a; Cui and Martin, 2003; Green et al., 2013; Ludwig et al., 2006; Pranti et al., 2017; Venkatraman et al., 2011; Vomero et al., 2017; Wilks et al., 2009). All of this contributes to improved charge transfer on the abiotic-biotic frontier and reduced power consumption (Ganji et al., 2017). Conducting polymers’ unique capacity for ionicelectronic current conversion is extremely interesting for signal transduction (Cui and Zhou, 2007). The body’s electrical conduction is primarily the result of cation movements, and these ionic fluxes lead to changes in PEDOT’s doping level that can be monitored through current modulation. However, the main concern of chronically implanted devices is about the foreign body reactions (FBRs) that arise from various causes (Biran et al., 2005; Liao et al., 2015; Lind et al., 2013; Nolta et al., 2015; Polikov et al., 2005; Prodanov and Delbeke, 2016; Saxena et al., 2013; Skousen et al., 2015; Szarowski et al., 2003; Turner et al., 1999; Winslow et al., 2010; Woolley et al., 2013). The FBR is the simultaneous action of immune cells and fibroblasts that leads to the formation of a protective layer around the intruding device, which is called encapsulation. Encapsulation prevents direct interaction with neural cells and provokes neuronal death due to released toxins. PEDOT possesses several advantages in this matter. Its soft and organic nature renders it more tolerable for biological tissues, even if more in vivo experiments are needed to assert its biocompatibility as some dopants (e.g., PF6
Þ have been reported to generate toxic compounds (Siedlecka et al., 2011; Swatloski et al., 2003; Visser et al., 2000). PEDOT also can be biofunctionalized (Bhagwat et al., 2016; Cui and Martin, 2003; Povlich et al., 2013; Vallejo-Giraldo et al., 2014), as well as its dopants (Asplund et al., 2009; Green et al., 2010).

25.5.2 PEDOT for lab-on-skin, flexible and stretchable electronics A lab-on-skin is an electronic device conformable to the skin that can interact in various ways with the epidermis for biosensing purposes (Liu et al., 2017; Simon et al., 2016). These devices require high flexibility and stretchability (Someya et al., 2016). Indeed, the maximum deformation for the human skin is 30%, with an elastic modulus between dozens and hundreds of kilopascal. A good deal of research has been done on developing PEDOT films for stretchable and flexible electronics that can be applied over various parts of the body, such as pressure sensors (Lipomi et al., 2012), glucose sensors (Pal et al., 2016), organic amplifiers for weak biosignals recording (Sekitani et al., 2016), electromyography arrays (Roberts et al. 2016; Kim et al. 2016), moisture sensors for perspiration tracking near the nose (Zhu et al., 2017), dopamine sensors

Processing and patterning of conducting polymers

833

(Pal et al., 2017), and tubular microelectrodes with fluidic drug delivery (Tian et al., 2016).

25.5.3 Drug delivery The principle of drug delivery with conducting polymers (Svirskis et al., 2010; Yue et al., 2013) is linked to the doping process that occurs during electrochemical polymerization. During electropolymerization, the electrolyte, which is also the dopant, gets trapped in the polymer matrix. If the dopant has previously been biofunctionalized by molecules, growth factors, drugs, or other substances, the electropolymerization process can be a drug-loading step for the polymer, used here as cargo. Upon the application of a potential, the doping molecules are expulsed from the polymer matrix, delivering the drug. This drug-delivery method is very localized, so it is better than drugs that work on a systematic scale. The main disadvantage is the loss of electroactivity of the conducting polymer. Indeed, dopant recovery is difficult because the dopant, bound to the drugs, will be consumed or dispersed into the body. Conducting polymers hence have limited usefulness for drug-delivery, as they become nonconductive and the drug load diminishes. Organic electronic ion pumps (OEIPs) may solve this problem (Uguz et al., 2017).

25.5.4 Self-healing of PEDOT:PSS The development of self-healing, conducting polymers would represent a major stepping-stone in increasing the lifetime and efficiency of devices. These materials would be characterized by their ability to repair their functionality automatically after the damage. Many self-healing polymers, composite materials, hydrogels, or ionic conductors have been reported, but their low conductivity and complex fabrication limit their use in electronic devices (Guofa et al., 2017; Jiheong et al., 2018; Kunmo et al., 2018; Ying et al., 2018; Yue et al., 2017). Self-healing of PEDOT:PSS films has been reported when high percentages (up to 80% wt/wt) of the plasticizer Triton X-100 were added (Young et al., 2015). Stretchable and conductive hydrogels containing PEDOT:PSS also have been reported to self-heal mechanically and electrically after the compression of the two separated surfaces and immersion into a 90°C water bath for 3 h (Wu et al., 2017). A great discovery for PEDOT:PSS films was their mechanical and electrical selfhealing property in aqueous environments. This could lead to the development of implanted devices with simple and highly efficient self-repair characteristics and long lifetimes (Zhang and Cicoira, 2017). PEDOT:PSS thin films (above 1 μm) were able to self-restore their electrical conductivities (as high as 500 S cm
1) almost instantaneously (150 ms) by wetting a cut of 40 μm with a droplet of water. After baking the healed region at 140°C for 1 h, the current remained stable, indicating that the presence of water is not necessary for the electric conduction. Interestingly, exposure to water vapor achieved similar results. Self-repair was observed at relative humidity (RH) levels of 80%, and complete electrical repair occurred at 90%, while cuts performed

834

Handbook of Organic Materials for Electronic and Photonic Devices

on prewetted films showed no drop in current. This was confirmed to be an intrinsic property of PEDOT:PSS, as films processed with additives such as conductivity enhancers (glycerol) and plasticizers (Capstone FS-30) on substrates such as flexible PET and polyimide continued to demonstrate self-healing. The authors tentatively attributed this phenomenon to the swelling of the PSS- shells, which enabled the PEDOT:PSS grains to move into the damaged area and finally restore the initial hydrogen bounds between the PSS- shells (Zhang and Cicoira, 2017).

25.6

Conclusion and outlook

PEDOT is one of the most exciting organic conductive polymers for biomedical applications. The polymer has seen extensive use in biosensors, drug-delivery devices, neural implants, and stretchable electronics, all of which exploit the polymer’s excellent electronic properties and processing versatility. Recent discoveries, such as selfhealing and new processing/patterning techniques to achieve flexible and stretchable devices, likely will lead to an even wider range of applications. Substantial research has been undertaken to understand PEDOT-based OECT behavior, while studies in processing methods have aided in improving device performance, biofunctionalization, and stability. However, a large gap still exists between the current technology and its successful application in medical techniques. Further efforts are needed to address stability and biocompatibility. For all organic devices, chronic evaluations of in vivo applications are lacking, while for OECTs, device noise is poorly understood. Despite this, the research described in this chapter is paving the way for the integration of organic polymers into our everyday lives.

References Anastassiou, C.A., Koch, C., Neuroscience, B., 2016. The origin of extracellular fields and currents—EEG, ECoG, LFP and spikes. Nat. Rev. Neurosci. 13, 407–420. Andersson Ersman, P., Nilsson, D., Kawahara, J., Gustafsson, G., Berggren, M., 2013. Fastswitching all-printed organic electrochemical transistors. Org. Electron. 14, 1276–1280. Aqrawe, Z., Montgomery, J., Travas-Sejdic, J., Svirskis, D., 2018. Conducting polymers for neuronal microelectrode array recording and stimulation. Sensors Actuators B Chem. 257, 753–765. Aregueta-Robles, U.A., Woolley, A.J., Poole-Warren, L.A., Lovell, N.H., Green, R.A., 2014. Organic electrode coatings for next-generation neural interfaces. Front. Neuroeng. 7, 15. Asplund, M., Thaning, E., Lundberg, J., Sandberg-Nordqvist, A.C., Kostyszyn, B., Ingan€as, O., von Holst, H., 2009. Toxicity evaluation of PEDOT/biomolecular composites intended for neural communication electrodes. Biomed. Mater. 4, 45009. Balint, R., Cassidy, N.J., Cartmell, S.H., 2014. Conductive polymers: towards a smart biomaterial for tissue engineering. Acta Biomater. 10, 2341–2353. Berezhetska, O., Liberelle, B., De Crescenzo, G., Cicoira, F., 2015. A simple approach for protein covalent grafting on conducting polymer films. J. Mater. Chem. B 3, 5087–5094.

Processing and patterning of conducting polymers

835

Bessaire, B., Mathieu, M., Salles, V., Yeghoyan, T., Celle, C., Simonato, J.-P., Brioude, A., 2017. Synthesis of continuous conductive PEDOT:PSS nanofibers by electrospinning: a conformal coating for optoelectronics. ACS Appl. Mater. Interfaces 9, 950–957. Bhagwat, N., Murray, R.E., Shah, S.I., Kiick, K.L., Martin, D.C., 2016. Biofunctionalization of PEDOT films with laminin-derived peptides. Acta Biomater. 41, 235–246. Biran, R., Martin, D.C., Tresco, P.A., 2005. Neuronal cell loss accompanies the brain tissue response to chronically implanted silicon microelectrode arrays. Exp. Neurol. 195, 115–126. Blacha, A., Koscielniak, P., Sitarz, M., Szuber, J., Zak, J., 2012. Pedot brushes electrochemically synthesized on thienyl-modified glassy carbon surfaces. Electrochim. Acta 62, 441–446. Boehler, C., Oberueber, F., Schlabach, S., Stieglitz, T., Asplund, M., 2016. Long-term stable adhesion for conducting polymers in biomedical applications: IrOx and nanostructured platinum solve the chronic challenge. ACS Appl. Mater. Interfaces 9 (1), 189–197. Boub
ee de Gramont, F., Zhang, S., Tomasello, G., Kumar, P., Sarkissian, A., Cicoira, F., 2017. Highly stretchable electrospun conducting polymer nanofibers. Appl. Phys. Lett. 111, 093701. Bredas, J.L., Street, G.B., 1985. Polarons, bipolarons, and solitons in conducting polymers. Acc. Chem. Res. 18, 309–315. Bronstein, J.M., Tagliati, M., Alterman, R.L., Lozano, A.M., Volkmann, J., Stefani, A., Horak, F.B., Okun, M.S., Foote, K.D., Krack, P., Pahwa, R., Henderson, J.M., Hariz, M.I., Bakay, R.A., Rezai, A., Marks, W.J., Moro, E., Vitek, J.L., Weaver, F.M., Gross, R.E., DeLong, M.R., 2011. Deep brain stimulation for parkinson disease. Arch. Neurol. 68, 165–171. Campana, A., Cramer, T., Simon, D.T., Berggren, M., Biscarini, F., 2014. Electrocardiographic recording with conformable organic electrochemical transistor fabricated on resorbable bioscaffold. Adv. Mater. 26, 3874–3878. Castagnola, E., Ansaldo, A., Maggiolini, E., Ius, T., Skrap, M., Ricci, D., Fadiga, L., Hofmann, U.G., Furlanetti, L.L., 2014a. Smaller, softer, lower-impedance electrodes for human neuroprosthesis: a pragmatic approach. Front. Neuroeng. 7, 8. Castagnola, V., Bayon, C., Descamps, E., Bergaud, C., 2014b. Morphology and conductivity of PEDOT layers produced by different electrochemical routes. Synth. Met. 189, 7–16. Castagnola, V., Descamps, E., Lecestre, A., Dahan, L., Remaud, J., Nowak, L.G., Bergaud, C., 2015. Parylene-based flexible neural probes with PEDOT coated surface for brain stimulation and recording. Biosens. Bioelectron. 67, 450–457. Cogan, S.F., 2008. Neural stimulation and recording electrodes. Annu. Rev. Biomed. Eng. 10, 275–309. Cui, X., Martin, D.C., 2003. Electrochemical deposition and characterization of poly (3,4ethylenedioxythiophene) on neural microelectrode arrays. Sensors Actuators B Chem. 89, 92–102. Cui, X.T., Zhou, D.D., 2007. Poly (3,4-ethylenedioxythiophene) for chronic neural stimulation. IEEE Trans. Neural Syst. Rehabil. Eng. 15, 502–508. Dayeh, V.R., Chow, S.L., Schirmer, K., Lynn, D.H., Bols, N.C., 2004. Evaluating the toxicity of Triton X-100 to protozoan, fish, and mammalian cells using fluorescent dyes as indicators of cell viability. Ecotoxicol. Environ. Saf. 57, 375–382. Einevoll, G.T., Kayser, C., Logothetis, N.K., Panzeri, S., 2013. Modelling and analysis of local field potentials for studying the function of cortical circuits. Nat. Rev. Neurosci. 14, 770–785.

836

Handbook of Organic Materials for Electronic and Photonic Devices

Elgrishi, N., Hammon, K., McCarthy, B., Eisenhart, T., Dempsey, J., 2017. A practical beginner’s guide to cyclic voltammetry. J. Chem. Educ. 95 (2), 197–206. Elschner, A., L€ovenich, W., 2011. Solution-deposited PEDOT for transparent conductive applications. MRS Bull. 36, 794–798. Fang, Y., Li, X., Fang, Y., 2015. Organic bioelectronics for neural interfaces. J. Mater. Chem. C 3, 6424–6430.
Fischer, L.M., Tenje, M., Heiskanen, A.R., Masuda, N., Castillo, J., Bentien, A., Emneus, J., Jakobsen, M.H., Boisen, A., 2009. Gold cleaning methods for electrochemical detection applications. Microelectron. Eng. 86, 1282–1285. Frost, C.M., Cederna, P.S., Martin, D.C., Shim, B.S., Urbanchek, M.G., 2014. Decellular biological scaffold polymerized with PEDOT for improving peripheral nerve interface charge transfer. In: 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, Chicago, IL, 2014, pp. 422–425. Ganji, M., Tanaka, A., Gilja, V., Halgren, E., Dayeh, S.A., 2017. Scaling effects on the electrochemical stimulation performance of Au, Pt, and PEDOT:PSS electrocorticography arrays. Adv. Funct. Mater. 1703019. Giovannitti, A., Sbircea, D.-T., Inal, S., Nielsen, C.B., Bandiello, E., Hanifi, D.A., Sessolo, M., Malliaras, G.G., McCulloch, I., Rivnay, J., 2016. Controlling the mode of operation of organic transistors through side-chain engineering. Proc. Natl. Acad. Sci. U. S. A. 113, 12017–12022. Green, R., Abidian, M.R., 2015. Conducting polymers for neural prosthetic and neural interface applications. Adv. Mater. 27 (46), 7620–7637. Green, R.A., Lovell, N.H., Poole-Warren, L.A., 2010. Impact of co-incorporating laminin peptide dopants and neurotrophic growth factors on conducting polymer properties. Acta Biomater. 6, 63–71. Green, R.A., Hassarati, R.T., Bouchinet, L., Lee, C.S., Cheong, G.L.M., Yu, J.F., Dodds, C.W., Suaning, G.J., Poole-Warren, L.A., Lovell, N.H., 2012. Substrate dependent stability of conducting polymer coatings on medical electrodes. Biomaterials 33, 5875–5886. Green, R.A., Matteucci, P.B., Hassarati, R.T., Giraud, B., Dodds, C.W.D., Chen, S., ByrnesPreston, P.J., Suaning, G.J., Poole-Warren, L.A., Lovell, N.H., 2013. Performance of conducting polymer electrodes for stimulating neuroprosthetics. J. Neural Eng. 10, 16009. Guofa, C., Jiangxin, W., Kai, Q., Jingwei, C., Shaohui, L., See, L.P., 2017. Extremely stretchable strain sensors based on conductive self-healing dynamic cross-links hydrogels for human-motion detection. Adv. Sci. 4, 1600190. Hu, Z., Zhang, J., Hao, Z., Zhao, Y., 2011. Influence of doped PEDOT:PSS on the performance of polymer solar cells. Sol. Energy Mater. Sol. Cells 95, 2763–2767. H€ utter, P.C., Rothl€ander, T., Haase, A., Trimmel, G., Stadlober, B., 2013. Influence of geometry variations on the response of organic electrochemical transistors. Appl. Phys. Lett. 103, 043308. Jiheong, K., Donghee, S., Nathan, W.G., Yuxin, L., Jeffrey, L., Yeongin, K., Young, O.J., Toru, K., Jaewan, M., Yeongjun, L., Lihua, J., Tok, J.B.-H., Zhenan, B., 2018. Tough and waterinsensitive self-healing elastomer for robust electronic skin. Adv. Mater. 30, 1706846. Jimison, L.H., Hama, A., Strakosas, X., Armel, V., Khodagholy, D., Ismailova, E., Malliaras, G.G., Winther-Jensen, B., Owens, R.M., 2012. PEDOT:TOS with PEG: a biofunctional surface with improved electronic characteristics. J. Mater. Chem. 22, 19498–19505. Jorfi, M., Skousen, J.L., Weder, C., Capadona, J.R., 2015. Progress towards biocompatible intracortical microelectrodes for neural interfacing applications. J. Neural Eng. 12, 11001.

Processing and patterning of conducting polymers

837

Kara, M.O.P., Frey, M.W., 2014. Effects of solvents on the morphology and conductivity of poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) nanofibers. J. Appl. Polym. Sci. 131, 40305. Khodagholy, D., Doublet, T., Gurfinkel, M., Quilichini, P., Ismailova, E., Leleux, P., Herve, T., Sanaur, S., Bernard, C., Malliaras, G.G., 2011. Highly conformable conducting polymer electrodes for in vivo recordings. Adv. Mater. 23, H268–H272. Khodagholy, D., Curto, V.F., Fraser, K.J., Gurfinkel, M., Byrne, R., Diamond, D., Malliaras, G.G., Benito-Lopez, F., Owens, R.M., 2012. Organic electrochemical transistor incorporating an ionogel as a solid state electrolyte for lactate sensing. J. Mater. Chem. 22, 4440–4443. Khodagholy, D., Rivnay, J., Sessolo, M., Gurfinkel, M., Leleux, P., Jimison, L.H., Stavrinidou, E., Herve, T., Sanaur, S., Owens, R.M., Malliaras, G.G., 2013. High transconductance organic electrochemical transistors. Nat. Commun. 4, 2133. Kim, N., Lim, T., Song, K., Yang, S., Lee, J., 2016. Stretchable multichannel electromyography sensor array covering large area for controlling home electronics with distinguishable signals from multiple muscles. ACS Appl. Mater. Interfaces 8 (32), 21070–21076. Kringelbach, M.L., Jenkinson, N., Owen, S.L.F., Aziz, T.Z., 2007. Translational principles of deep brain stimulation. Nat. Rev. Neurosci. 8, 623–635. Kumar, P., Yi, Z., Zhang, S., Sekar, A., Soavi, F., Cicoira, F., 2015. Effect of channel thickness, electrolyte ions, and dissolved oxygen on the performance of organic electrochemical transistors. Appl. Phys. Lett. 107, 053303. Kunmo, C., Gwon, S.B., Hye-In, Y., Dong-Min, K., Seung, L.C., Minwoo, P., Chan-Moon, C., 2018. Smart passivation materials with a liquid metal microcapsule as self-healing conductors for sustainable and flexible perovskite solar cells. Adv. Funct. Mater. 28, 1800110. Laforgue, A., Robitaille, L., 2010. Production of conductive PEDOT nanofibers by the combination of electrospinning and vapor-phase polymerization. Macromolecules 43, 4194–4200. Lawal, A.T., Wallace, G.G., 2014. Vapour phase polymerisation of conducting and nonconducting polymers: a review. Talanta 119, 133–143. Li, P., Sun, K., Ouyang, J., 2015. Stretchable and conductive polymer films prepared by solution blending. ACS Appl. Mater. Interfaces 7, 18415–18423. Liao, C., Zhang, M., Yao, M.Y., Hua, T., Li, L., Yan, F., 2015. Flexible organic electronics in biology: materials and devices. Adv. Mater. 27, 7493–7527. Lin, F., Lonergan, M.C., 2006. Gate electrode processes in an electrolyte-gated transistor: nonfaradaically versus Faradaically coupled conductivity modulation of a polyacetylene ionomer. Appl. Phys. Lett. 88, 133507. Lin, P., Yan, F., 2012. Organic thin-film transistors for chemical and biological sensing. Adv. Mater. 24, 34–51. Lind, G., Linsmeier, C.E., Schouenborg, J., 2013. The density difference between tissue and neural probes is a key factor for glial scarring. Sci. Rep. 3, 2942. Lipomi, D.J., Lee, J.A., Vosgueritchian, M., Tee, B.C.K., Bolander, J.A., Bao, Z., 2012. Electronic properties of transparent conductive films of PEDOT:PSS on stretchable substrates. Chem. Mater. 24, 373–382. Liu, Y., Pharr, M., Salvatore, G.A., 2017. Lab-on-skin: a review of flexible and stretchable electronics for wearable health monitoring. ACS Nano 11, 9614–9635. Ludwig, K.A., Uram, J.D., Yang, J., Martin, D.C., Kipke, D.R., 2006. Chronic neural recordings using silicon microelectrode arrays electrochemically deposited with a poly(3,4ethylenedioxythiophene) (PEDOT) film. J. Neural Eng. 3, 59–70. Malliaras, G., Friend, R., 2005. An organic electronics primer. Phys. Today 58, 53–58.

838

Handbook of Organic Materials for Electronic and Photonic Devices

Mantione, D., Del Agua, I., Schaafsma, W., Elmahmoudy, M., Uguz, I., Sanchez-Sanchez, A., Sardon, H., Castro, B., Malliaras, G.G., Mecerreyes, D., 2017. Low-temperature crosslinking of PEDOT:PSS films using divinylsulfone. ACS Appl. Mater. Interfaces 9, 18254–18262. Martin, D.C., Malliaras, G.G., 2016. Interfacing electronic and ionic charge transport in bioelectronics. ChemElectroChem 3, 686. Maya-Vetencourt, J.F., Ghezzi, D., Antognazza, M.R., Colombo, E., Mete, M., Feyen, P., Desii, A., Buschiazzo, A., Di Paolo, M., Di Marco, S., Ticconi, F., Emionite, L., Shmal, D., Marini, C., Donelli, I., Freddi, G., Maccarone, R., Bisti, S., Sambuceti, G., Pertile, G., Lanzani, G., Benfenati, F., 2017. A fully organic retinal prosthesis restores vision in a rat model of degenerative blindness. Nat. Mater. 16, 681–689. Mracek, L., Pretl, S., Syrovy´, T., Hama´cek, A., 2015. Ionic liquid as an electrolyte for organic electrochemical transistor–semantic scholar. In: 2015 38th Intenational Spring Seminar on Electronics Technology, pp. 11–15. Mracek, L., Syrovy, T., Pretl, S., Nespurek, S., Hamacek, A., 2016. Influence of topology on performance of organic electrochemical transistor. Period. Polytech. Electr. Eng. Comput. Sci. 60, 211–216. Nolta, N.F., Christensen, M.B., Crane, P.D., Skousen, J.L., Tresco, P.A., 2015. BBB leakage, astrogliosis, and tissue loss correlate with silicon microelectrode array recording performance. Biomaterials 53, 753–762. Oh, J.Y., Shin, M., Lee, J.B., Ahn, J.H., Baik, H.K., Jeong, U., 2014. Effect of PEDOT nanofibril networks on the conductivity, flexibility, and coatability of PEDOT:PSS films. ACS Appl. Mater. Interfaces 6, 6954–6961. Ouyang, J., 2013. Solution-processed PEDOT: PSS films with conductivities as indium tin oxide through a treatment with mild and weak organic acids. ACS Appl. Mater. Interfaces 5, 13082–13088. Ouyang, L., Shaw, C.L., Kuo, C., Griffin, L., Martin, D.C., 2014a. In vivo polymerization of poly(3,4-ethylenedioxythiophene) (PEDOT) in living rat hippocampus does not cause a significant loss of performance in a delayed alternation (DA). J. Neural Eng. 6, 2166–2171. Ouyang, S., Xie, Y., Zhu, D., Xu, X., Wang, D., Tan, T., Fong, H.H., 2014b. Photolithographic patterning of PEDOT:PSS with a silver interlayer and its application in organic light emitting diodes. Org. Electron. physics, Mater. Appl. 15, 1822–1827. Ouyang, L., Wei, B., Kuo, C., Pathak, S., Farrell, B., Martin, D.C., 2017. Enhanced PEDOT adhesion on solid substrates with electrografted P(EDOT-NH 2). Sci. Adv. 3, e1600448. Pal, R.K., Farghaly, A.A., Collinson, M.M., Kundu, S.C., Yadavalli, V.K., 2016. Photolithographic micropatterning of conducting polymers on flexible silk matrices. Adv. Mater. 28, 1406–1412. Pal, R.K., Kundu, S.C., Yadavalli, V.K., 2017. Biosensing using photolithographically micropatterned electrodes of PEDOT:PSS on ITO substrates. Sensors Actuators B Chem. 242, 140–147. Polikov, V.S., Tresco, P.A., Reichert, W.M., 2005. Response of brain tissue to chronically implanted neural electrodes. J. Neurosci. Methods 148, 1–18. Povlich, L.K., Cheol, J., Leach, M.K., Corey, J.M., Kim, J., Martin, D.C., 2013. Synthesis, copolymerization and peptide-modification of carboxylic acid-functionalized 3,4ethylenedioxythiphene (EDOTacid) for neural electrode interfaces. BBA-Gen. Subjects 1830, 4288–4293. Pranti, A.S., Schander, A., B€odecker, A., Lang, W., 2017. Highly stable PEDOT:PSS coating on gold microelectrodes with improved charge injection capacity for chronic neural stimulation. Proceedings 1, 492.

Processing and patterning of conducting polymers

839

Prodanov, D., Delbeke, J., 2016. Mechanical and biological interactions of implants with the brain and their impact on implant design. Front. Neurosci. 10, 11. Richardonson-Burns, S.M., Hendricks, J.L., Martin, D.C., 2006. Electrochemical polymerization of conducting polymers in living neural tissue. J. Neural Eng. 4, L6–L13. Rivnay, J., Leleux, P., Sessolo, M., Khodagholy, D., Herv
e, T., Fiocchi, M., Malliaras, G.G., 2013a. Organic electrochemical transistors with maximum transconductance at zero gate bias. Adv. Mater. 25, 7010–7014. Rivnay, J., Owens, R.M., Malliaras, G.G., 2013b. The rise of organic bioelectronics. Chem. Mater. 26 (1), 679–685. Roberts, T., De Graaf, J.B., Nicol, C., Herv
e, T., Fiocchi, M., Sanaur, S., 2016. Flexible inkjetprinted multielectrode arrays for neuromuscular cartography. Adv. Healthc. Mater. 5, 1462–1470. Samba, R., Herrmann, T., Zeck, G., 2015. PEDOT–CNT coated electrodes stimulate retinal neurons at low voltage amplitudes and low charge densities. J. Neural Eng. 12, 16014. Saxena, T., Karumbaiah, L., Gaupp, E.A., Patkar, R., Patil, K., Betancur, M., Stanley, G.B., Bellamkonda, R.V., 2013. The impact of chronic blood-brain barrier breach on intracortical electrode function. Biomaterials 34, 4703–4713. Sayago, J., Shafique, U., Soavi, F., Cicoira, F., Santato, C., 2014. TransCap: a monolithically integrated supercapacitor and electrolyte-gated transistor. J. Mater. Chem. C 2, 10273–10276. Sekitani, T., Yokota, T., Kuribara, K., Kaltenbrunner, M., Fukushima, T., Inoue, Y., Sekino, M., Isoyama, T., Abe, Y., Onodera, H., Someya, T., 2016. Ultraflexible organic amplifier with biocompatible gel electrodes. Nat. Commun. 7, 11425. Siedlecka, E.M., Czerwicka, M., Stolte, S., Stepnowski, P., 2011. Stability of ionic liquids in application conditions. Curr. Org. Chem. 15, 1974–1991. Simon, D.T., Gabrielsson, E.O., Tybrandt, K., Berggren, M., 2016. Organic bioelectronics: bridging the signaling gap between biology and technology. Chem. Rev. 116, 13009–13041. Skousen, J.L., Bridge, M.J., Tresco, P.A., 2015. A strategy to passively reduce neuroinflammation surrounding devices implanted chronically in brain tissue by manipulating device surface permeability. Biomaterials 36, 33–43. Someya, T., Bao, Z., Malliaras, G.G., 2016. The rise of plastic bioelectronics. Nature 540, 379–385. Stavrinidou, E., Leleux, P., Rajaona, H., Khodagholy, D., Rivnay, J., Lindau, M., Sanaur, S., Malliaras, G.G., 2013. Direct measurement of ion mobility in a conducting polymer. Adv. Mater. 25, 4488–4493. Stefani, A., Lozano, A.M., Peppe, A., Stanzione, P., Galati, S., Tropepi, D., Pierantozzi, M., Brusa, L., Scarnati, E., Mazzone, P., 2007. Bilateral deep brain stimulation of the pedunculopontine and subthalamic nuclei in severe Parkinson’s disease. Brain 130, 1596–1607. Stoop, R.L., Thodkar, K., Sessolo, M., Bolink, H.J., Sch€ onenberger, C., Calame, M., 2017. Charge noise in organic electrochemical transistors. Phys. Rev. Appl. 7, 014009. Svirskis, D., Travas-Sejdic, J., Rodgers, A., Garg, S., 2010. Electrochemically controlled drug delivery based on intrinsically conducting polymers. J. Control. Release 146, 6–15. Swatloski, R.P., Holbrey, J.D., Rogers, R.D., 2003. Ionic liquids are not always green: hydrolysis of 1-butyl-3-methylimidazolium hexafluorophosphate. Green Chem. 5, 361. Szarowski, D.H., Andersen, M.D., Retterer, S., Spence, A.J., Isaacson, M., Craighead, H.G., Turner, J.N., Shain, W., 2003. Brain responses to micro-machined silicon devices. Brain Res. 983, 23–35.

840

Handbook of Organic Materials for Electronic and Photonic Devices

Tang, H., Yan, F., Lin, P., Xu, J., Chan, H.L.W., 2011. Highly sensitive glucose biosensors based on organic electrochemical transistors using platinum gate electrodes modified with enzyme and nanomaterials. Adv. Funct. Mater. 21, 2264–2272. Tang, H., Kumar, P., Zhang, S., Yi, Z., De Crescenzo, G., Santato, C., Soavi, F., Cicoira, F., 2015. Conducting polymer transistors making use of activated carbon gate electrodes. ACS Appl. Mater. Interfaces 7, 969–973. Tarabella, G., Coppede`, N., Iannotta, S., Cicoira, F., Kumar, P., Santato, C., 2013. Organic bioelectronics. In: Ostroverkhova, O. (Ed.), Handbook of Organic Materials for Optical and (Opto)electronic Devices: Properties and Application. Woodhead Publishing Ltd. Teo, W.E., Ramakrishna, S., 2006. A review on electrospinning design and nanofibre assemblies. Nanotechnology 17, R89–R106. Thompson, C.J., Chase, G.G., Yarin, A.L., Reneker, D.H., 2007. Effects of parameters on nanofiber diameter determined from electrospinning model. Polymer 48 (23), 6913–6922. Tian, H.C., Liu, J.Q., Kang, X.Y., Tang, L.J., Wang, M.H., Ji, B.W., Yang, B., Wang, X.L., Chen, X., Yang, C.S., 2016. Enhanced flexible tubular microelectrode with conducting polymer for multi-functional implantable tissue-machine interface. Sci. Rep. 6, 26910. Tiwari, S.K., Venkatraman, S.S., 2012. Importance of viscosity parameters in electrospinning: of monolithic and core–shell fibers. Mater. Sci. Eng. C 32, 1037–1042. Turner, J.N., Shain, W., Szarowski, D.H., Andersen, M., Martins, S., Isaacson, M., Craighead, H., 1999. Cerebral astrocyte response to micromachined silicon implants. Exp. Neurol. 156, 33–49. Uguz, I., Proctor, C.M., Curto, V.F., Pappa, A.M., Donahue, M.J., Ferro, M., Owens, R.M., Khodagholy, D., Inal, S., Malliaras, G.G., 2017. A microfluidic ion pump for in vivo drug delivery. Adv. Mater. 29, 1701217. Vallejo-Giraldo, C., Kelly, A., Biggs, M.J.P., 2014. Biofunctionalisation of electrically conducting polymers. Drug Discov. Today 19, 88–94. Venkatraman, S., Hendricks, J., King, Z.A., Sereno, A.J., Richardson-burns, S., Martin, D., Carmena, J.M., Member, S., 2011. In vitro and in vivo evaluation of PEDOT microelectrodes for neural stimulation and recording. IEEE Trans. Neural Syst. Rehabil. Eng. 19, 307–316. Visser, A.E., Swatloski, R.P., Reichert, W.M., Griffin, S.T., Rogers, R.D., 2000. Traditional extractants in nontraditional solvents: groups 1 and 2 extraction by crown ethers in room-temperature ionic liquids. Ind. Eng. Chem. Res. 39, 3596–3604. Vomero, M., Van Niekerk, P., Nguyen, V., Gong, N., Hirabayashi, M., Cinopri, A., Logan, K., Moghadasi, A., Varma, P., Kassegne, S., 2016. A novel pattern transfer technique for mounting glassy carbon microelectrodes on polymeric flexible substrates. J. Micromech. Microeng. 26, 025018. Vomero, M., Castagnola, E., Ciarpella, F., Maggiolini, E., Goshi, N., Zucchini, E., Carli, S., Fadiga, L., Kassegne, S., Ricci, D., 2017. Highly stable glassy carbon interfaces for long-term neural stimulation and low-noise recording of brain activity. Sci. Rep. 7, 40332. Wang, G.-F., Tao, X.-M., Wang, R.-X., 2008. Fabrication and characterization of OLEDs using PEDOT:PSS and MWCNT nanocomposites. Compos. Sci. Technol. 68, 2837–2841. Wang, X., Sj€oberg-Eerola, P., Eriksson, J.E., Bobacka, J., Bergelin, M., 2010. The effect of counter ions and substrate material on the growth and morphology of poly(3,4ethylenedioxythiophene) films: towards the application of enzyme electrode construction in biofuel cells. Synth. Met. 160, 1373–1381. Wang, Y., Zhu, C., Pfattner, R., Yan, H., Jin, L., Chen, S., Molina-Lopez, F., Lissel, F., Liu, J., Rabiah, N.I., Chen, Z., Chung, J.W., Linder, C., Toney, M.F., Murmann, B., Bao, Z., 2017. A highly stretchable, transparent, and conductive polymer. Sci. Adv. 3, e1602076.

Processing and patterning of conducting polymers

841

Weaver, F.M., Follett, K., Stern, M., Hur, K., Harris, C., Marks Jr., W.J., Rothlind, J., Sagher, O., Reda, D., Moy, C.S., Pahwa, R., Burchiel, K., Hogarth, P., Lai, E.C., Duda, J.E., Holloway, K., Samii, A., Horn, S., Bronstein, J., Stoner, G., Heemskerk, J., Huang, G.D., 2009. Bilateral deep brain stimulation vs best medical therapy for patients with advanced parkinson disease. JAMA Otolaryngol. Head Neck Surg. 301 (1), 63–73. Wei, B., Liu, J., Ouyang, L., Kuo, C.C., Martin, D.C., 2015. Significant enhancement of PEDOT thin film adhesion to inorganic solid substrates with EDOT-acid. ACS Appl. Mater. Interfaces 7, 15388–15394. Wilks, S.J., Richardson-burns, S.M., Hendricks, J.L., Martin, D.C., Otto, K.J., 2009. Poly (3,4ethylenedioxythiophene) as a micro-neural interface material for electrostimulation. Front. Neuroeng. 2, 7. Williamson, A., Ferro, M., Leleux, P., Ismailova, E., Kaszas, A., Doublet, T., Quilichini, P., Rivnay, J., Ro´zsa, B., Katona, G., Bernard, C., Malliaras, G.G., 2015. Localized neuron stimulation with organic electrochemical transistors on delaminating depth probes. Adv. Mater. 27, 4405–4410. Winslow, B.D., Christensen, M.B., Yang, W.K., Solzbacher, F., Tresco, P.A., 2010. A comparison of the tissue response to chronically implanted Parylene-C-coated and uncoated planar silicon microelectrode arrays in rat cortex. Biomaterials 31, 9163–9172. Woolley, A.J., Desai, H.A., Otto, K.J., 2013. Chronic intracortical microelectrode arrays induce non-uniform, depth-related tissue responses. J. Neural Eng. 10, 26007. Worfolk, B.J., Andrews, S.C., Park, S., Reinspach, J., Liu, N., Toney, M.F., Mannsfeld, S.C.B., Bao, Z., 2015. Ultrahigh electrical conductivity in solution-sheared polymeric transparent films. Proc. Natl. Acad. Sci. 112, 14138–14143. Wu, Q., Wei, J., Xu, B., Liu, X., Wang, H., Wang, W., Wang, Q., Liu, W., 2017. A robust, highly stretchable supramolecular polymer conductive hydrogel with self-healability and thermoprocessability. Sci. Rep. 7, 41566. Xia, Y., Sun, K., Ouyang, J., 2012. Solution-processed metallic conducting polymer films as transparent electrode of optoelectronic devices. Adv. Mater. 24, 2436–2440. Yan, H., Kagata, T., Okuzaki, H., 2009. Micrometer-scaled OFET channel patterns fabricated by using PEDOT/PSS microfibers. Synth. Met. 159, 2229–2232. Yang, J., Kim, D.H., Hendricks, J.L., Leach, M., Northey, R., Martin, D.C., 2005. Ordered surfactant-templated poly(3,4-ethylenedioxythiophene) (PEDOT) conducting polymer on microfabricated neural probes. Acta Biomater. 1, 125–136. Yi, Z., Natale, G., Kumar, P., Di Mauro, E., Heuzey, M.-C., Soavi, F., Perepichka, I.I., Varshney, S.K., Santato, C., Cicoira, F., 2015. Ionic liquid–water mixtures and ion gels as electrolytes for organic electrochemical transistors. J. Mater. Chem. C 3, 6549–6553. Yi, Z., Bettini, L.G., Tomasello, G., Kumar, P., Piseri, P., Valitova, I., Milani, P., Soavi, F., Cicoira, F., Bettini Luca, G., Tomasello, G., Kumar, P., Piseri, P., Valitova, I., Milani, P., Soavi, F., Cicoira, F., 2016. Flexible conducting polymer transistors with supercapacitor function. J. Polym. Sci. B Polym. Phys. 55, 96–103. Yi, Z., Bettini, L.G., Tomasello, G., Kumar, P., Piseri, P., Valitova, I., Milani, P., Soavi, F., Cicoira, F., 2017. Flexible conducting polymer transistors with supercapacitor function. J. Polym. Sci. B Polym. Phys. 55, 96–103. Ying, G., Kaiqiang, Z., Pengbo, W., 2018. A flexible stretchable hydrogel electrolyte for healable all-in-one configured supercapacitors. Small 14, 1704497. Young, O.J., Sunghee, K., Hong-Koo, B., Unyong, J., 2015. Conducting polymer dough for deformable electronics. Adv. Mater. 28, 4455–4461. Yue, Z., Moulton, S.E., Cook, M., O’Leary, S., Wallace, G.G., 2013. Controlled delivery for neuro-bionic devices. Adv. Drug Deliv. Rev. 65, 559–569.

842

Handbook of Organic Materials for Electronic and Photonic Devices

Yue, C., Timothy, G.M., Eric, A., Sarah, I.A., Bryan, M.W., Christoph, K., Chao, W., 2017. A transparent, self-healing, highly stretchable ionic conductor. Adv. Mater. 29, 1605099. Zhang, S., Cicoira, F., 2017. Water-enabled healing of conducting polymer films. Adv. Mater. 29, 1–6. Zhang, S., Kumar, P., Nouas, A.S., Fontaine, L., Tang, H., Cicoira, F., 2015. Solvent-induced changes in PEDOT:PSS films for organic electrochemical transistors. APL Mater. 3, 014911. Zhang, S., Hubis, E., Girard, C., Kumar, P., DeFranco, J., Cicoira, F., 2016. Water stability and orthogonal patterning of flexible micro-electrochemical transistors on plastic. J. Mater. Chem. C 4, 1382–1385. Zhu, Z., Yang, G., Li, R., Pan, T., 2017. Photopatternable PEDOT:PSS/PEG hybrid thin film with moisture stability and sensitivity. Microsyst. Nanoeng. 3, 17004.

Further reading Zhang, S., Hubis, E., Tomasello, G., Soliveri, G., Kumar, P., Cicoira, F., 2017. Patterning of stretchable organic electrochemical transistors. Chem. Mater. 29, 3126–3132.

Organic electronic memory devices

26

Michael C. Petty Department of Engineering and Centre for Molecular and Nanoscale Electronics, Durham University, Durham, United Kingdom

26.1

Introduction

The past 20 years has seen an upsurge in academic and commercial interest in the field of organic electronics (Canatore, 2013; Cicoira and Santato, 2013; Cuevas and Scheer, 2017; Petty, 2018). Devices such as organic light-emitting displays and chemical sensors are already in the marketplace, while others, including transistors, photovoltaic cells (PVCs), and smart cards, are developing fast. Much of this work concerns the replacement of silicon and other inorganic semiconductors, which are currently used in electronic devices, with organic compounds. The motivation is the reduced cost and large-area capability of many organic thin-film technologies (Street et al., 2015). Tremendous advances have been achieved, but much more of the laboratory research needs to be translated into reliable and reproducible products that can be manufactured cheaply. The current progress on organic transistors is particularly impressive (Bao and Locklin, 2007; Sirringhaus, 2014; L€ ussem et al., 2015; Quinn et al., 2017). Fieldeffect mobility values in the range 0.1–1 cm2/V/s are routinely demonstrated, and field-effect transistors (FETs) can be fabricated conveniently using cost-effective techniques such as inkjet printing. Low-voltage operation (