Materials for solid state lighting and displays 9781119140580, 1119140587, 9781119140597, 1119140595, 9781119140610, 1119140617

Table of contents :
Content: List of Contributors xi Series Preface xiii Preface xv Acknowledgments xvii About the Editor xix 1. Principles of Solid State Luminescence 1 Adrian Kitai 1.1 Introduction to Radiation from an Accelerating Charge 1 1.2 Radiation from an Oscillating Dipole 4 1.3 Quantum Description of an Electron during a Radiation Event 5 1.4 The Exciton 7 1.5 Two-Electron Atoms 10 1.6 Molecular Excitons 16 1.7 Band-to-Band Transitions 19 1.8 Photometric Units 23 1.9 The Light Emitting Diode 28 References 30 2. Quantum Dots for Displays and Solid State Lighting 31 Jesse R. Manders, Debasis Bera, Lei Qian and Paul H. Holloway 2.1 Introduction 31 2.2 Nanostructured Materials 34 2.3 Quantum Dots 35 2.3.1 History of Quantum Dots 36 2.3.2 Structure and Properties Relationship 36 2.3.3 Quantum Confinement Effects on Band Gap 38 2.4 Relaxation Process of Excitons 41 2.4.1 Radiative Relaxation 42 2.4.2 Nonradiative Relaxation Process 45 2.5 Blinking Effect 46 2.6 Surface Passivation 47 2.6.1 Organically Capped QDs 47 2.6.2 Inorganically Passivated QDs 48 2.7 Synthesis Processes 49 2.7.1 Top-Down Synthesis 49 2.7.2 Bottom-Up Approach 50 2.8 Optical Properties and Applications 53 2.8.1 Displays 53 2.8.2 Solid State Lighting 73 2.8.3 Biological Applications 78 2.9 Perspective 81 Acknowledgments 82 References 82 3. Color Conversion Phosphors for Light Emitting Diodes 91 Jack Silver, George R. Fern and Robert Withnall 3.1 Introduction 91 3.2 Disadvantages of Using LEDs Without Color Conversion Phosphors 93 3.3 Phosphors for Converting the Color of Light Emitted by LEDs 95 3.3.1 General Considerations 95 3.3.2 Requirements of Color Conversion Phosphors 95 3.3.3 Commonly Used Activators in Color Conversion Phosphors 97 3.3.4 Strategies for Generating White Light from LEDs 97 3.3.5 Outstanding Problems with Color Conversion Phosphors for LEDs 98 3.4 Survey of the Synthesis and Properties of Some Currently Available Color Conversion Phosphors 99 3.4.1 Phosphor synthesis 99 3.4.2 Metal Oxide Based Phosphors 99 3.4.3 Metal Sulfide Based Phosphors 113 3.4.4 Metal Nitrides 117 3.4.5 Alkaline Earth Metal Oxo-Nitrides 120 3.4.6 Metal Fluoride Phosphors 121 3.5 Multi-Phosphor pcLEDs 122 3.6 Quantum Dots 123 3.7 Laser Diodes 124 3.8 Conclusions 125 Acknowledgments 125 References 126 4. Nitride and Oxynitride Phosphors for Light Emitting Diodes 135 Le Wang and Rong-Jun Xie 4.1 Introduction 135 4.2 Synthesis of Nitride and Oxynitride Phosphors 138 4.2.1 Solid State Reaction Method 138 4.2.2 Gas Reduction and Nitridation 139 4.2.3 Carbothermal Reduction and Nitridation 140 4.2.4 Alloy Nitridation 140 4.2.5 Ammonothermal Synthesis 141 4.3 Photoluminescence Properties of Nitride and Oxynitride Phosphors 142 4.3.1 Luminescence Spectra of Typical Activators 142 4.4 Emerging Nitride Phosphors and Their Synthesis 165 4.4.1 Narrow-Band Red Nitride Phosphors 165 4.4.2 Narrow-Band Green Nitride Phosphors 167 4.5 Applications of Nitride Phosphors 169 4.5.1 General Lighting 169 4.5.2 LCD Backlight 172 References 173 5. Organic Light Emitting Device Materials for Displays 183 Tyler Davidson-Hall, Yoshitaka Kajiyama and Hany Aziz 5.1 Introduction to OLEDs and Organic Electroluminscent Materials 184 5.2 OLED Light Emitting Materials 186 5.2.1 Neat Emitters 187 5.2.2 Guest Emitters 192 5.2.3 Aggregate-Induced Emission 201 5.3 OLED Displays 203 5.3.1 RGB Color Patterning Approaches 203 5.3.2 Display Addressing Approaches 204 5.3.3 FMM Technology 207 5.3.4 Alternative Fabrication Techniques 208 5.3.5 Outlook on OLED Display Commercialization 212 5.4 Quantum Dot Light Emitting Devices 213 5.4.1 QD Optimization by Core Shell Morphology 214 5.4.2 Organic Charge Transport QD-LEDs 215 5.4.3 Hybrid Organic Inorganic Charge Transport QD-LEDs 217 5.4.4 Energy Transfer Enhanced QD-LEDs 219 5.4.5 QD-LED Lifetime 220 References 220 6. White-Light Emitting Materials for Organic Light-Emitting Diode-Based Displays and Lighting 231 Simone Lenk, Michael Thomschke and Sebastian Reineke 6.1 Introduction 231 6.2 White Organic Light-Emitting Diodes 233 6.3 Photometry and Radiometry 236 6.3.1 OLED Efficiencies 239 6.3.2 Color Stimulus Specification 239 6.3.3 Color Correlated Temperature 240 6.3.4 Color Rendering Index 241 6.3.5 White Light 241 6.4 Device Optics 242 6.4.1 Optical Properties of Thin Films 242 6.4.2 Optical Outcoupling 245 6.4.3 Top-Emitting OLEDs 247 6.4.4 Simulation Tools 248 6.5 Materials for Efficient White Electroluminescence 248 6.5.1 Spin Statistics for Electroluminescence 248 6.5.2 Fluorescence-Emitting Molecules 249 6.5.3 Advanced Concepts Comprising Fluorescent Emitters 251 6.5.4 Phosphorescence-Emitting Molecules 251 6.5.5 Single White-Light Emitting Phosphorescent Materials 256 6.5.6 Thermally Activated Delayed Fluorescence-Based Emitters 257 6.5.7 Phosphorescence Versus Thermally Activated Delayed Fluorescence 261 6.5.8 TADF Assisted Fluorescence (TAF) Emitters 263 6.6 Polymer Concepts 263 6.6.1 Various Concepts Involving Polymer Materials 265 6.6.2 Learning from High Performance Small Molecules for High Efficiency Polymers 267 6.7 Summary and Outlook 268 References 269 7. Light Emitting Diode Materials and Devices 273 Michael R. Krames 7.1 Introduction 273 7.2 Light Emitting Diode Basics 273 7.2.1 Construction 273 7.2.2 Recombination Processes 275 7.2.3 Heterojunctions 277 7.2.4 Quantum Wells 278 7.2.5 Current Injection 278 7.2.6 Forward voltage 280 7.3 Material Systems 280 7.3.1 Ga(As,P) 280 7.3.2 Ga(As,P):N 281 7.3.3 (Al,Ga)As 282 7.3.4 (Al,Ga)InP 282 7.3.5 (Ga,In)N 283 7.3.6 White Light Generation 285 7.4 Packaging Technologies 288 7.4.1 Low Power 288 7.4.2 Mid Power 288 7.4.3 High Power 289 7.4.4 Chip-On-Board LEDs 290 7.4.5 Multi-Color LEDs 290 7.4.6 Electrostatic Discharge Protection 290 7.5 Performance 291 7.5.1 Light Extraction Efficiency 291 7.5.2 Monochromatic Performance 292 7.5.3 White-Emitting Performance 298 7.5.4 Temperature Effects 306 7.5.5 Reliability 306 References 307 8. Alternating Current Thin Film and Powder Electroluminescence 313 Adrian Kitai 8.1 Introduction 313 8.2 Background of TFEL 314 8.2.1 Thick Film Dielectric EL Structure 315 8.2.2 Ceramic Sheet Dielectric EL 316 8.2.3 Sphere-Supported TFEL 316 8.3 Theory of Operation 317 8.4 Electroluminescent Phosphors 324 8.5 Thin Film Double-Insulating EL Devices 325 8.6 Current Status of TFEL 327 8.7 Background of AC Powder EL 328 8.8 Mechanism of Light Emission in AC Powder EL 329 8.9 Electroluminescence Characteristics of AC Powder EL Materials 333 8.10 Emission Spectra of AC Powder EL 334 8.11 Luminance Degradation 335 8.12 Moisture and Operating Environment 336 8.13 Current Status and Limitations of Powder EL 336 8.14 Research Directions in AC Powder EL and TFEL 336 References 337 Index 339

Citation preview

Materials for Solid State Lighting and Displays

Wiley Series in Materials for Electronic and Optoelectronic Applications www.wiley.com/go/meoa Series Editors Professor Arthur Willoughby, University of Southampton, Southampton, UK Dr Peter Capper, formerly of SELEX Galileo Infrared Ltd, Southampton, UK Professor Safa Kasap, University of Saskatchewan, Saskatoon, Canada Published Titles Bulk Crystal Growth of Electronic, Optical and Optoelectronic Materials, Edited by P. Capper Properties of Group-IV, III–V and II–VI Semiconductors, S. Adachi Charge Transport in Disordered Solids with Applications in Electronics, Edited by S. Baranovski Optical Properties of Condensed Matter and Applications, Edited by J. Singh Thin Film Solar Cells: Fabrication, Characterization, and Applications, Edited by J. Poortmans and V. Arkhipov Dielectric Films for Advanced Microelectronics, Edited by M. R. Baklanov, M. Green, and K. Maex Liquid Phase Epitaxy of Electronic, Optical and Optoelectronic Materials, Edited by P. Capper and M. Mauk Molecular Electronics: From Principles to Practice, M. Petty CVD Diamond for Electronic Devices and Sensors, Edited by R. S. Sussmann Properties of Semiconductor Alloys: Group-IV, III–V, and II–VI Semiconductors, S. Adachi Mercury Cadmium Telluride, Edited by P. Capper and J. Garland Zinc Oxide Materials for Electronic and Optoelectronic Device Applications, Edited by C. Litton, D. C. Reynolds, and T. C. Collins Lead-Free Solders: Materials Reliability for Electronics, Edited by K. N. Subramanian Silicon Photonics: Fundamentals and Devices, M. Jamal Deen and P. K. Basu Nanostructured and Subwavelength Waveguides: Fundamentals and Applications, M. Skorobogatiy Photovoltaic Materials: From Crystalline Silicon to Third-Generation Approaches, G. Conibeer and A. Willoughby Glancing Angle Deposition of Thin Films: Engineering the Nanoscale, Matthew M. Hawkeye, Michael T. Taschuk and Michael J. Brett Spintronics for Next Generation Innovative Devices, Edited by Katsuaki Sato and Eiji Saitoh Physical Properties of High-Temperature Superconductors, Rainer Wesche Inorganic Glasses for Photonics, Animesh Jha Amorphous Semiconductors: Structural, Optical and Electronic Properties, Koichi Shimakawa, Sandor Kugler and Kazuo Morigaki

Materials for Solid State Lighting and Displays

Edited by ADRIAN KITAI Departments of Engineering Physics and Materials Science and Engineering, McMaster University, Hamilton, Canada

This edition first published 2017 © 2017 John Wiley & Sons Ltd All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions. The right of Adrian Kitai to be identified as the editorial material in this work has been asserted in accordance with law. Registered Office John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK Editorial Offices 111 River Street, Hoboken, NJ 07030, USA 9600 Garsington Road, Oxford, OX4 2DQ, UK The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK Boschstr. 12, 69469 Weinheim, Germany For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com. Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that appears in standard print versions of this book may not be available in other formats. Limit of Liability/Disclaimer of Warranty The publisher and the authors make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of fitness for a particular purpose. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for every situation. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. The fact that an organization or website is referred to in this work as a citation and/or potential source of further information does not mean that the author or the publisher endorses the information the organization or website may provide or recommendations it may make. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this works was written and when it is read. No warranty may be created or extended by any promotional statements for this work. Neither the publisher nor the author shall be liable for any damages arising herefrom. Library of Congress Cataloging-in-Publication Data applied for. 9781119140580 Set in 10/12pt, TimesLTStd by SPi Global, Chennai, India.

10 9 8 7 6 5 4 3 2 1

Contents

List of Contributors Series Preface Preface Acknowledgments About the Editor 1. Principles of Solid State Luminescence Adrian Kitai 1.1 Introduction to Radiation from an Accelerating Charge 1.2 Radiation from an Oscillating Dipole 1.3 Quantum Description of an Electron during a Radiation Event 1.4 The Exciton 1.5 Two-Electron Atoms 1.6 Molecular Excitons 1.7 Band-to-Band Transitions 1.8 Photometric Units 1.9 The Light Emitting Diode References 2. Quantum Dots for Displays and Solid State Lighting Jesse R. Manders, Debasis Bera, Lei Qian and Paul H. Holloway 2.1 Introduction 2.2 Nanostructured Materials 2.3 Quantum Dots 2.3.1 History of Quantum Dots 2.3.2 Structure and Properties Relationship 2.3.3 Quantum Confinement Effects on Band Gap 2.4 Relaxation Process of Excitons 2.4.1 Radiative Relaxation 2.4.2 Nonradiative Relaxation Process 2.5 Blinking Effect 2.6 Surface Passivation 2.6.1 Organically Capped QDs 2.6.2 Inorganically Passivated QDs

xi xiii xv xvii xix 1 1 4 5 7 10 16 19 23 28 30 31 31 34 35 36 36 38 41 42 45 46 47 47 48

vi

Contents

2.7

Synthesis Processes 2.7.1 Top-Down Synthesis 2.7.2 Bottom-Up Approach 2.8 Optical Properties and Applications 2.8.1 Displays 2.8.2 Solid State Lighting 2.8.3 Biological Applications 2.9 Perspective Acknowledgments References 3. Color Conversion Phosphors for Light Emitting Diodes Jack Silver, George R. Fern and Robert Withnall 3.1 Introduction 3.2 Disadvantages of Using LEDs Without Color Conversion Phosphors 3.3 Phosphors for Converting the Color of Light Emitted by LEDs 3.3.1 General Considerations 3.3.2 Requirements of Color Conversion Phosphors 3.3.3 Commonly Used Activators in Color Conversion Phosphors 3.3.4 Strategies for Generating White Light from LEDs 3.3.5 Outstanding Problems with Color Conversion Phosphors for LEDs 3.4 Survey of the Synthesis and Properties of Some Currently Available Color Conversion Phosphors 3.4.1 Phosphor synthesis 3.4.2 Metal Oxide Based Phosphors 3.4.3 Metal Sulfide Based Phosphors 3.4.4 Metal Nitrides 3.4.5 Alkaline Earth Metal Oxo-Nitrides 3.4.6 Metal Fluoride Phosphors 3.5 Multi-Phosphor pcLEDs 3.6 Quantum Dots 3.7 Laser Diodes 3.8 Conclusions Acknowledgments References 4. Nitride and Oxynitride Phosphors for Light Emitting Diodes Le Wang and Rong-Jun Xie 4.1 Introduction 4.2 Synthesis of Nitride and Oxynitride Phosphors 4.2.1 Solid State Reaction Method 4.2.2 Gas Reduction and Nitridation 4.2.3 Carbothermal Reduction and Nitridation 4.2.4 Alloy Nitridation 4.2.5 Ammonothermal Synthesis

49 49 50 53 53 73 78 81 82 82 91 91 93 95 95 95 97 97 98 99 99 99 113 117 120 121 122 123 124 125 125 126 135 135 138 138 139 140 140 141

Contents

4.3

Photoluminescence Properties of Nitride and Oxynitride Phosphors 4.3.1 Luminescence Spectra of Typical Activators 4.4 Emerging Nitride Phosphors and Their Synthesis 4.4.1 Narrow-Band Red Nitride Phosphors 4.4.2 Narrow-Band Green Nitride Phosphors 4.5 Applications of Nitride Phosphors 4.5.1 General Lighting 4.5.2 LCD Backlight References 5. Organic Light Emitting Device Materials for Displays Tyler Davidson-Hall, Yoshitaka Kajiyama and Hany Aziz 5.1 Introduction to OLEDs and Organic Electroluminscent Materials 5.2 OLED Light Emitting Materials 5.2.1 Neat Emitters 5.2.2 Guest Emitters 5.2.3 Aggregate-Induced Emission 5.3 OLED Displays 5.3.1 RGB Color Patterning Approaches 5.3.2 Display Addressing Approaches 5.3.3 FMM Technology 5.3.4 Alternative Fabrication Techniques 5.3.5 Outlook on OLED Display Commercialization 5.4 Quantum Dot Light Emitting Devices 5.4.1 QD Optimization by Core–Shell Morphology 5.4.2 Organic Charge Transport QD-LEDs 5.4.3 Hybrid Organic–Inorganic Charge Transport QD-LEDs 5.4.4 Energy Transfer Enhanced QD-LEDs 5.4.5 QD-LED Lifetime References 6. White-Light Emitting Materials for Organic Light-Emitting Diode-Based Displays and Lighting Simone Lenk, Michael Thomschke and Sebastian Reineke 6.1 Introduction 6.2 White Organic Light-Emitting Diodes 6.3 Photometry and Radiometry 6.3.1 OLED Efficiencies 6.3.2 Color Stimulus Specification 6.3.3 Color Correlated Temperature 6.3.4 Color Rendering Index 6.3.5 White Light 6.4 Device Optics 6.4.1 Optical Properties of Thin Films 6.4.2 Optical Outcoupling 6.4.3 Top-Emitting OLEDs

vii

142 142 165 165 167 169 169 172 173 183 184 186 187 192 201 203 203 204 207 208 212 213 214 215 217 219 220 220

231 231 233 236 239 239 240 241 241 242 242 245 247

viii

Contents

6.4.4 Simulation Tools Materials for Efficient White Electroluminescence 6.5.1 Spin Statistics for Electroluminescence 6.5.2 Fluorescence-Emitting Molecules 6.5.3 Advanced Concepts Comprising Fluorescent Emitters 6.5.4 Phosphorescence-Emitting Molecules 6.5.5 Single White-Light Emitting Phosphorescent Materials 6.5.6 Thermally Activated Delayed Fluorescence-Based Emitters 6.5.7 Phosphorescence Versus Thermally Activated Delayed Fluorescence 6.5.8 TADF Assisted Fluorescence (TAF) Emitters 6.6 Polymer Concepts 6.6.1 Various Concepts Involving Polymer Materials 6.6.2 Learning from High Performance Small Molecules for High Efficiency Polymers 6.7 Summary and Outlook References 6.5

7. Light Emitting Diode Materials and Devices Michael R. Krames 7.1 Introduction 7.2 Light Emitting Diode Basics 7.2.1 Construction 7.2.2 Recombination Processes 7.2.3 Heterojunctions 7.2.4 Quantum Wells 7.2.5 Current Injection 7.2.6 Forward voltage 7.3 Material Systems 7.3.1 Ga(As,P) 7.3.2 Ga(As,P):N 7.3.3 (Al,Ga)As 7.3.4 (Al,Ga)InP 7.3.5 (Ga,In)N 7.3.6 White Light Generation 7.4 Packaging Technologies 7.4.1 Low Power 7.4.2 Mid Power 7.4.3 High Power 7.4.4 Chip-On-Board LEDs 7.4.5 Multi-Color LEDs 7.4.6 Electrostatic Discharge Protection 7.5 Performance 7.5.1 Light Extraction Efficiency 7.5.2 Monochromatic Performance 7.5.3 White-Emitting Performance

248 248 248 249 251 251 256 257 261 263 263 265 267 268 269 273 273 273 273 275 277 278 278 280 280 280 281 282 282 283 285 288 288 288 289 290 290 290 291 291 292 298

Contents

7.5.4 7.5.5 References

Temperature Effects Reliability

ix

306 306 307

8. Alternating Current Thin Film and Powder Electroluminescence Adrian Kitai 8.1 Introduction 8.2 Background of TFEL 8.2.1 Thick Film Dielectric EL Structure 8.2.2 Ceramic Sheet Dielectric EL 8.2.3 Sphere-Supported TFEL 8.3 Theory of Operation 8.4 Electroluminescent Phosphors 8.5 Thin Film Double-Insulating EL Devices 8.6 Current Status of TFEL 8.7 Background of AC Powder EL 8.8 Mechanism of Light Emission in AC Powder EL 8.9 Electroluminescence Characteristics of AC Powder EL Materials 8.10 Emission Spectra of AC Powder EL 8.11 Luminance Degradation 8.12 Moisture and Operating Environment 8.13 Current Status and Limitations of Powder EL 8.14 Research Directions in AC Powder EL and TFEL References

313 313 314 315 316 316 317 324 325 327 328 329 333 334 335 336 336 336 337

Index

339

List of Contributors

Hany Aziz, Department of Electrical & Computer Engineering, University of Waterloo, Canada Debasis Bera, NanoPhotonica, Inc., USA and Department of Materials Science and Engineering, University of Florida, USA Tyler Davidson-Hall, Department of Electrical & Computer Engineering, University of Waterloo, Canada George R. Fern, Brunel University, London, UK Paul H. Holloway, NanoPhotonica, Inc., USA and Department of Materials Science & Engineering, University of Florida, USA Yoshitaka Kajiyama, Department of Electrical & Computer Engineering, University of Waterloo, Canada Adrian Kitai, Departments of Engineering Physics and Materials Science and Engineering, McMaster University, Hamilton, Canada Michael R. Krames, Arkesso, LLC, USA Simone Lenk, Dresden Integrated Center for Applied Physics and Photonic Materials (IAPP) & Institute for Applied Physics, Technische Universität Dresden, Germany Jesse R. Manders, Nanosys, Inc., USA Lei Qian, NanoPhotonica, Inc., USA and Department of Materials Science and Engineering, University of Florida, USA Sebastian Reineke, Dresden Integrated Center for Applied Physics and Photonic Materials (IAPP) & Institute for Applied Physics, Technische Universität Dresden, Germany Jack Silver, Brunel University, London, UK

xii

List of Contributors

Michael Thomschke, Dresden Integrated Center for Applied Physics and Photonic Materials (IAPP) & Institute for Applied Physics, Technische Universität Dresden, Germany Le Wang, College of Optical and Electronic Technology, China Jiliang University, China Robert Withnall (deceased), Brunel University, London, UK Rong-Jun Xie, National Institute for Materials Science (NIMS), Japan

Series Preface Wiley Series in Materials for Electronic and Optoelectronic Applications This book series is devoted to the rapidly developing class of materials used for electronic and optoelectronic applications. It is designed to provide much-needed information on the fundamental scientific principles of these materials, together with how these are employed in technological applications. The books are aimed at (postgraduate) students, researchers, and technologists, engaged in research, development, and the study of materials in electronics and photonics, and industrial scientists developing new materials, devices, and circuits for the electronic, optoelectronic, and communications industries. The development of new electronic and optoelectronic materials depends not only on materials engineering at a practical level, but also on a clear understanding of the properties of materials, and the fundamental science behind these properties. It is the properties of a material that eventually determine its usefulness in an application. The series therefore also includes such titles as electrical conduction in solids, optical properties, thermal properties, and so on, all with applications and examples of materials in electronics and optoelectronics. The characterization of materials is also covered within the series in as much as it is impossible to develop new materials without the proper characterization of their structure and properties. Structure–property relationships have always been fundamentally and intrinsically important to materials science and engineering. Materials science is well known for being one of the most interdisciplinary sciences. It is the interdisciplinary aspect of materials science that has led to many exciting discoveries, new materials, and new applications. It is not unusual to find scientists with a chemical engineering background working on materials projects with applications in electronics. In selecting titles for the series, we have tried to maintain the interdisciplinary aspect of the field, and hence its excitement to researchers in this field. Arthur Willoughby Peter Capper Safa Kasap

Preface

Luminescent materials play a key role in a vast range of products from luminaires to televisions to cell phones. We cherish well-illuminated indoor and outdoor spaces. We take for granted a wide range of spectacular flat panel displays and are actively developing next generation flexible materials for flexible displays and lighting products as well as a wider range of colors and higher quantum efficiencies in both display and lighting markets. The book begins with a very accessible treatment of the theory of luminescence. The first chapter is designed to target fundamental processes in inorganic semiconductors and other materials as well as in molecular solids. It also introduces the key metrics by which luminescence is measured and qualified. Subsequent book chapters then present the key categories of materials and the solid state devices they enable. The topics being addressed include organic light emitting diodes, more accurately referred to as organic light emitting devices, inorganic light emitting diodes, quantum dot wavelength conversion materials, a wide range of important phosphor down-conversion materials and electroluminescent materials and devices. Solid state luminescent materials are rapidly displacing more traditional luminescence processes in fluorescent and other gas-phase lamps in all but a few areas of application. This trend will continue due to the unprecedented power efficiency of solid state light emitters since global warming is a topic of international concern. The decreasing cost and increasing importance of a wide range of solid state luminescent materials and devices makes this book an essential resource for both industry and academia. Adrian Kitai Hamilton, Ontario, Canada

Acknowledgments

I would like to express my gratitude to the many people who contributed to this book. The significance of the chapter contributors is self-evident and their expertise in their respective areas of specialization is second to none. My thanks also extend to my assistant Dylan Genuth-Okonwho has made a big impact on my workload. Finally, it has been a great pleasure working with the staff at Wiley including Rebecca Stubbs, Emma Strickland and Ramya Raghavan who collectively guided me through the process of getting this book off the ground and continued doing so throughout the many stages of bringing the book to completion.

About the Editor

Adrian Kitai is Professor in the Departments of Materials Science and Engineering and Engineering Physics at McMaster University (Canada). He was educated at McMaster University and received his PhD in Electrical Engineering from Cornell University (USA). His research interests include nano-sized oxide phosphors for sunlight collection in fluorescent photovoltaic building windows, oxide phosphor electroluminescence and LED-based high resolution display systems. He has over 30 years of experience in solid state luminescence and has contributed to a few start-up companies. He holds several patents relating to display technology and is the Chapter Chair of the Society for Information Display in Canada. He has also authored an undergraduate-level textbook introducing the fundamentals of solar cells, LEDs and other p-n junction devices.

1 Principles of Solid State Luminescence Adrian Kitai McMaster University, Hamilton, Ontario, Canada

It is useful to understand the origin of luminescence. Solid state luminescent materials and devices all rely on a common mechanism of luminescence whether they are semiconductor light emitting diodes (LEDs) or phosphors or quantum dots, and whether they are organic or inorganic materials. This is introduced in Sections 1.1–1.3 and then this chapter presents a series of more specific luminescence processes.

1.1

Introduction to Radiation from an Accelerating Charge

Light is electromagnetic radiation which can be produced by an accelerating charge. Let us first consider a stationary point charge q in a vacuum. Electric field lines are produced from the point charge with electric field lines emanating radially out from the charge as shown in Figure 1.1. This stationary point charge does not produce electromagnetic radiation but since it does produce an electric field there is electric field energy surrounding the point charge. This energy is related to the electric field by: E𝜀 =

𝜖0 2  2

where 𝜖0 is the permittivity of vacuum and E𝜀 is the electric field energy density. If the charge q were to move with a constant velocity v an additional magnetic field B is produced. Lines of this magnetic field form closed loops that lie in planes perpendicular to the velocity vector of the moving change as shown in Figure 1.2.

Materials for Solid State Lighting and Displays, First Edition. Edited by Adrian Kitai. © 2017 John Wiley & Sons Ltd. Published 2017 by John Wiley & Sons Ltd.

2

Materials for Solid State Lighting and Displays

ε

q

Figure 1.1

Lines of electric field  produced by stationary point charge q

B

q

Figure 1.2 Closed lines of magnetic field B due to a point charge q moving with constant velocity into the page

Both magnetic and electric fields exist surrounding the charge moving with uniform velocity. The magnetic field also has an energy associated with it. The magnetic field energy density EB is given by: 1 2 EB = B 2𝜇0 where 𝜇0 is the magnetic permeability of vacuum. The total energy density due to both fields is now: E = E𝜀 + EB =

𝜖0 2 1 2 B  + 2 2𝜇0

The field strengths of both the electric and the magnetic fields fall off as we move further away from the charge and therefore the energy density falls off rapidly with distance from the charge. There is no radiation from the charge.

Principles of Solid State Luminescence

3

acceleration of charge

Figure 1.3

Lines of electric field emanating from an accelerating charge

The situation changes dramatically if the charge q undergoes acceleration. Consider a charge that rapidly accelerates as shown in Figure 1.3. The electric field lines further away from the charge are still based on the original position of the charge at position A before the acceleration occurred, however electric field lines after acceleration will now emanate from the new location at position B of the charge. The new electric field lines will expand outwards and replace the original field lines. The speed at which this expansion occurs is the speed of light c because it is not possible for information on the new location of the charge to arrive at any particular distance away from the charge faster than the speed of light. The kinks in the electric field lines in Figure 1.3 associated with this expansion must contain an electric field component ⟂ which is perpendicular to the radial field direction. ⟂ propagates outwards from the accelerating charge at velocity c. Notice that the biggest kink and therefore the largest magnitudes of ⟂ propagate in directions perpendicular to the acceleration of the charge. In the direction of the acceleration ⟂ = 0. In addition, there is a magnetic field B⟂ that is perpendicular to both the direction of acceleration as well as to ⟂ . This field is shown in Figure 1.4. This magnetic field ⟂ propagates outwards and is also a maximum in a direction normal to the acceleration. The combined electric and magnetic fields ⟂ and B⟂ form a propagating electromagnetic wave that travels away from the accelerating charge. B⊥ θ

A

B

acceleration of charge

Figure 1.4 Direction of magnetic field B⟂ that is perpendicular to both the direction of acceleration as well as to ⟂ from Figure 1.3

4

Materials for Solid State Lighting and Displays

The magnitudes of ⟂ and B⟂ are given by: ⟂ =

qa sin 𝜃 4𝜋𝜖0 c2 r

and B⟂ =

𝜇0 qa sin 𝜃 4𝜋cr

The electromagnetic radiation formed by these two fields propagates away from the accelerating charge and this radiation has a directed power flow per unit area (Poynting vector) given by: q2 a2 → − 1 ⟂ × B⟂ = sin2 𝜃 r̂ S = 𝜇0 16𝜋 2 𝜖0 c3 r2 where r̂ is a unit radial vector. The total radiated energy from the accelerated charge is calculated by integrating the magnitude of the Poynting vector over a sphere surrounding the accelerating charge and we obtain: 𝜋

2𝜋

P=

∫sphere

SdA =

∫0

∫0

Substituting for S(𝜃), P=

S(𝜃)r2 sin 𝜃d𝜃d𝜙 =

𝜋

∫0

S(𝜃)2𝜋r2 sin 𝜃d𝜃

𝜋 2q2 a2 sin3 𝜃d𝜃 3 16𝜋𝜖0 c ∫0

Upon integration we obtain: P=

1.2

2q2 a2 12𝜋𝜖0 c3

(1.1)

Radiation from an Oscillating Dipole

The manner in which a charge can accelerate can take many forms. For example, an electron orbiting in a cyclotron accelerates steadily towards the center of its orbit and radiation according to Equation 1.1 will be emitted most strongly tangentially to the orbit in a direction perpendicular to the acceleration vector. If energetic electrons are directed towards an atomic target, the rapid deceleration upon impact with atomic nuclei causes radiation called bremsstrahlung (radiation due to deceleration). The charge acceleration that is by far the most important in luminescent solids, however, is generated by an oscillating dipole formed by an electron oscillating in the vicinity of a positive atomic nucleus. This is known as an oscillating dipole and the radiation it produces is called dipole radiation. Dipole radiation can occur within, and be very effectively released from, solids such as semiconductors or insulators that are substantially transparent to the dipole radiation. Consider a charge q that oscillates about the origin along the x-axis having position given by: x(t) = A sin 𝜔t

Principles of Solid State Luminescence

The electron has acceleration a =

d2 x(t) dt2

5

or

a(t) = −A𝜔2 sin 𝜔t Substituting into Equation 1.1 we can write: P(t) =

2q2 A2 𝜔4 sin2 𝜔t 12𝜋𝜖0 c3

and averaging this power over one cycle we obtain average power P=

𝜔 2q2 A2 𝜔4 2𝜋 12𝜋𝜖0 c3 ∫0

which yields: P=

2𝜋 𝜔

sin2 𝜔t dt

q2 A2 𝜔4 12𝜋𝜖0 c3

(1.2)

In terms of the dipole moment p = qA this is written: P=

p2 𝜔4 12𝜋𝜖0 c3

Dipole radiation may take place from atomic orbitals inside a crystal lattice or it may take place as an electron and a hole recombine. We do not think of classical oscillating electron motion because we describe electrons using quantum mechanics. We are now ready to show that the quantum mechanical description of an electron can yield oscillations during a radiation event.

1.3

Quantum Description of an Electron during a Radiation Event

Solving Schroedinger’s equation for a potential V(r) in which an electron may exist yields a set of wavefunctions or stationary states that allow us to obtain the probability density function and energy levels of the electron. Examples of this include the set of electron orbitals of a hydrogen atom or the electron states in a potential well. These are called stationary states because the electron will remain in a specific quantum state unless perturbed by an outside influence. There is no time dependence of measurable electron parameters such as energy, momentum or expected position. As an example of this, consider an electron in a stationary state 𝜓n which is a solution of Schroedinger’s equation. 𝜓n may be written in terms of a spatial part of the wavefunction 𝜙n (r) as: ) ( −iEt (1.3) 𝜓n = 𝜙n (r) exp ℏ We can calculate the expected value of the position of this electron as: ⟨r⟩ (t) = ⟨𝜓n |r|𝜓n ⟩ =

∫V

V |𝜓n |2 r dV

6

Materials for Solid State Lighting and Displays

where V represents all space. Substituting the form of a stationary state we obtain: ) ( )] [ ( iEt −iEt 𝜙(r) exp rdV = r𝜙2 (r)dV 𝜙(r) exp ⟨r⟩(t) = ∫V ∫V ℏ ℏ which is a time-independent quantity. This confirms the stationary nature of this state. A stationary state does not radiate and there is no energy loss associated with the behavior of an electron in such a state. Note that electrons are not truly stationary in a quantum state from a classical viewpoint. It is therefore the quantum state that is described as stationary and not the electron itself. Quantum mechanics sanctions the existence of a charge that has a distributed spatial probability distribution function and yet that is in a stationary state. Classical physics fails to describe or predict this. Experience tells us, however, that radiation may be produced when a charge moves from one stationary state to another and we can show that radiation is produced if an oscillating dipole results from a charge moving from one stationary state to another. Consider a charge q initially in normalized stationary state 𝜓n and eventually in normalized stationary state 𝜓n′ . During the transition, a superposition state is created which we shall call 𝜓s : 𝜓s = a𝜓n + b𝜓n′ where |a|2 + |b|2 = 1 to normalize the superposition state. Here, a and b are timedependent coefficients. Initially a = 1 and b = 0 and after the transition, a = 0 and b = 1. If we now calculate the expectation value of the position of q for the superposition state 𝜓s we obtain: ⟨rs ⟩ = ⟨a𝜓n + b𝜓n′ |r|a𝜓n + b𝜓n′ ⟩ = |a|2 ⟨𝜓n |r|𝜓n ⟩ + |b|2 ⟨𝜓n′ |r|𝜓n′ ⟩ + a∗ b⟨𝜓n |r|𝜓n′ ⟩ + b∗ a⟨𝜓n |r|𝜓n′ ⟩ Of the four terms, the first two are stationary but the last two terms are not and therefore ⟨r⟩s (t) may be written using Equation 1.3 as: ( ( ) ) −i(En − En′ )t i(En − En′ )t ∗ ∗ ⟨r⟩s (t) = a b⟨𝜙n |r|𝜙n′ ⟩ exp + b a⟨𝜙n |r|𝜙n′ ⟩ exp ℏ ℏ Using the Euler formula eix + e−ix = 2 cos x we have: ( ( ) ) −i(En − En′ )t i(En − En′ )t ⟨r⟩s (t) = a∗ b⟨𝜙n |r|𝜙n′ ⟩ exp + b∗ a⟨𝜙n |r|𝜙n′ ⟩ exp ℏ ℏ ( ) (En – En′ )t = 2a∗ b⟨𝜙n |r|𝜙n′ ⟩ cos ℏ Defining |rnn′ | = a∗ b⟨𝜙n |r|𝜙n′ ⟩ and 𝜔nn′ =

(En – En′ ) ℏ

we finally obtain:

⟨r⟩s (t) = 2|rnn′ | cos(𝜔nn′ t)

(1.4)

Here, |rnn′ | is called the matrix element for the transition. It is seen that the expectation (E – E ′ ) value of the position of the electron is oscillating with frequency 𝜔nn′ = n ℏ n which is

Principles of Solid State Luminescence

7

1.0 a Wavefunction amplitude b

Time evolution

Figure 1.5 A time-dependent plot of coefficients a and b is consistent with the time evolution of wavefunctions 𝜙n and 𝜙n′ . At t = 0, a = 1 and b = 0. Next a superposition state is formed during the transition such that |a|2 + |b|2 = 1. Finally, after the transition is complete a = 0 and b = 1

the required frequency to produce a photon having energy E = En – En′ . The term |rnn′ | also varies with time, but does so very slowly compared with the cosine term. This is illustrated in Figure 1.5. We may also define a photon emission rate Rnn′ of a continuously oscillating charge q. We use Equations 1.2 and 1.4 and E = ℏ𝜔 to obtain: Rnn′ =

q2 𝜔3 P |r ′ |2 photons∕s = ℏ𝜔 3𝜋𝜖0 c3 ℏ nn

The photon emission rate is only an average rate. This is because of the Heizenberg Uncertainty Principle which states that the position and the momentum of an electron cannot be precisely measured simultaneously. It also means that we cannot predict the exact time of photon creation while simultaneously knowing its exact energy. Since the energy of the photon is defined without uncertainty there will be uncertainty about the precise time of release of each photon.

1.4

The Exciton

A hole and an electron can exist as a valence band state and a conduction band state. In this model the two particles are not localized and they are both represented using Bloch functions in the periodic potential of the crystal lattice. If the mutual attraction between the two becomes significant then a new description is required for their quantum states that is valid before they recombine but after they experience some mutual attraction. The hole and electron can exist in quantum states that are actually within the energy gap. Just as a hydrogen atom consists of a series of energy levels associated with the allowed quantum states of a proton and an electron, a series of energy levels associated with the quantum states of a hole and an electron also exists. This hole–electron entity is called an exciton, and the exciton behaves in a manner that is similar to a hydrogen atom with one important exception: a hydrogen atom has a lowest energy state or ground state when its quantum number n = 1, but a exciton, which also has a ground state at n = 1, has an opportunity to be annihilated when the electron and hole eventually recombine.

8

Materials for Solid State Lighting and Displays

For an exciton we need to modify the electron mass m to become the reduced mass 𝜇 of the hole–electron pair, which is given by: 1 1 1 = ∗+ ∗ 𝜇 me mh For direct gap semiconductors such as GaAs this is about one order of magnitude smaller than the free electron mass m. In addition the excition exists inside a semiconductor rather than in a vacuum. The relative dielectric constant 𝜖r must be considered, and it is approximately 10 for typical inorganic semiconductors. Adapting the hydrogen atom model, the ground state energy for an exciton is: Eexciton =

−𝜇q4 8𝜖o2 𝜖r2 h2

≅

ERydberg 1000

This yields a typical exciton ionization energy or binding energy of under 0.1 eV. The exciton radius in the ground state (n = 1) will be given by: aexciton =

4𝜋𝜖0 𝜖r ℏ2 ≅ 100a0 𝜇q2

which yields an exciton radius of the order of 50 Å. Since this radius is much larger than the lattice constant of a semiconductor, we are justified in our use of the bulk semiconductor parameters for effective mass and relative dielectric constant. Our picture is now of a hydrogen atom-like entity drifting around within the semiconductor crystal and having a series of energy levels analogous to those in a hydrogen atom. Just eV where quantum number n is an integer, as a hydrogen atom has energy levels En = 13.6 n2 the exciton has similar energy levels but in a much smaller energy range, and a quantum number nexciton is used. The exciton must transfer energy to be annihilated. When an electron and a hole form an exciton it is expected that they are initially in a high energy level with a large quantum number nexciton . This forms a larger, less tightly bound exciton. Through thermalization the exciton loses energy to lattice vibrations and approaches its ground state. Its radius decreases as nexciton approaches 1. Once the exciton is more tightly bound and nexciton is a small integer, the hole and electron can then form an effective dipole and radiation may be produced to account for the remaining energy and to annihilate the exciton through the process of dipole radiation. When energy is released as electromagnetic radiation, we can determine whether or not a particular transition is allowed by calculating the term |rnn′ | in Equation 1.4 and determining whether it is zero or non-zero. If |rnn′ | = 0 then this is equivalent to saying that dipole radiation will not take place and a photon cannot be created. Instead lattice vibrations remove the energy. If |rnn′ | > 0 then this is equivalent to saying that dipole radiation can take place and a photon can be created. We can represent the exciton energy levels in a semiconductor as shown in Figure 1.6. At low temperatures the emission and absorption wavelengths of electron–hole pairs must be understood in the context of excitons in all p-n junctions. The existence of excitons, however, is generally hidden at room temperature and at higher temperatures in

Principles of Solid State Luminescence n=3 n=2 n=1

9

Exciton levels

Eg Eminimum

Figure 1.6 The exciton forms a. series of closely spaced hydrogen-Iike energy levels that extend inside the energy gap of a semiconductor. If an electron falls into the lowest energy state of the exciton corresponding to n = 1 then the remaining energy available for a photon is Eminimum

inorganic semiconductors because of the temperature of operation of the device. The exciton is not stable enough to form from the distributed band states and at room temperature kT may be larger than the exciton energy levels. In this case the spectral features associated with excitons will be masked and direct gap or indirect gap band-to-band transitions occur. Nevertheless, photoluminescence or absorption measurements at low temperatures conveniently provided in the laboratory using liquid nitrogen (77 K) or liquid helium (4.2 K) clearly show exciton features, and excitons have become an important tool to study inorganic semiconductor behavior. An example of the transmission as a function of photon energy of a semiconductor at low temperature due to excitons is shown in Figure 1.7. In an indirect gap inorganic semiconductor at room temperature without the formation of excitons, the electron–hole pair can lose energy to phonons and be annihilated but not through dipole radiation. In a direct-gap semiconductor, however, dipole radiation can occur. The calculation of |rnn′ | is also relevant to band-to-band transitions. Since a dipole does not carry linear momentum it does not allow for the conservation of electron momentum during electron–hole pair recombination in an indirect gap semiconductor crystal and dipole radiation is forbidden. The requirement of a direct gap for a band-to-band transition that conserves momentum is consistent with the requirements of dipole radiation. Dipole radiation is effectively either allowed or forbidden in band-to-band transitions. Not all excitons are free to move around in the semiconductor. Bound excitons are often formed that associate themselves with defects in a semiconductor crystal such as vacancies and impurities. In organic semiconductors molecular exictons form, which are very important for an understanding of optical processes that occur in organic semiconductors. This is because molecular excitons typically have high binding energies of approximately 0.4 eV. The reason for the higher binding energy is the confinement of the molecular exciton to smaller spatial dimensions imposed by the size of the molecule. This keeps the hole and electron closer and increases the binding energy compared with free excitons. In contrast to the situation in inorganic semiconductors, molecular excitons are thermally stable

10

Materials for Solid State Lighting and Displays

2.12

2.13

Photon energy (eV) 2.14 2.15

2.16

n=3

n=5 n=4

–3

In (Transmission)

n=2 –2

–1

0 17100

17 200

17 300 Photon energy (cm–1)

17 400

Figure 1.7 Low-temperature transmission as a function of photon energy tor Cu2 O. The absorption of photons is caused through excitons, which are excited into higher energy levels as the absorption process takes place. Cu2 O is a semiconductor with a bandgap of 2.17 eV. Reprinted from Kittel, C., Introduction to Solid State Physics, 6e, ISBN 0-471-87474-4. Copyright (1986) John Wiley and Sons, Australia

at room temperature and they generally determine emission and absorption characteristics of organic semiconductors in operation. The molecular exciton is fundamental to organic light emitting diode (OLED) operation. We will first need to discuss in more detail the physics required to understand excitons and optical processes in molecular materials.

1.5

Two-Electron Atoms

Until now we have focused on dipole radiators that are composed of two charges, one positive and one negative. In Section 1.3 we introduced an oscillating dipole having one positive charge and one negative charge. In Section 1.4 we discussed the exciton, which also has one positive charge and one negative charge. However, we also need to understand radiation from molecular systems with two or more electrons, which form the basis of organic semiconductors. Once a system has two or more identical particles (electrons) there are additional and very fundamental quantum effects that we need to consider. In inorganic semiconductors, band theory gives us the tools

Principles of Solid State Luminescence

11

to handle large numbers of electrons in a periodic potential. In organic semiconductors electrons are confined to discrete organic molecules and “hop” from molecule to molecule. Band theory is still relevant to electron behavior within a given molecule provided it contains repeating structural units. Nevertheless, we need to study the electronic properties of molecules more carefully because molecules contain multiple electrons, and exciton properties in molecules are rather different from the excitons we have discussed in inorganic semiconductors. The best starting point is the helium atom, which has a nucleus with a charge of +2q as well as two electrons each with a charge of −q. A straightforward solution to the helium atom using Schrödinger’s equation is not possible since this is a three-body system; however, we can understand the behavior of such a system by applying the Pauli exclusion principle and by including the spin states of the two electrons. When two electrons at least partly overlap spatially with one another their wavefunctions must conform to the Pauli exclusion principle; however, there is an additional requirement that must be satisfied. The two electrons must be carefully treated as indistinguishable because once they have even a small spatial overlap there is no way to know which electron is which. We can only determine a probability density |𝜓|2 = 𝜓 ∗ 𝜓 for each wavefunction but we cannot determine the precise location of either electron at any instant in time and therefore there is always a chance that the electrons exchange places. There is no way to label or otherwise identify each electron and the wavefunctions must therefore not be specific about the identity of each electron. If we start with Schrödinger’s equation and write it by adding up the energy terms from the two electrons we obtain: ) ) ( ( 2 2 2 2 2 2 ℏ2 𝜕 𝜓 T 𝜕 𝜓 T 𝜕 𝜓 T ℏ2 𝜕 𝜓T 𝜕 𝜓T 𝜕 𝜓T − + VT 𝜓T = ET 𝜓T + + + + − 2m 2m 𝜕x12 𝜕y21 𝜕z21 𝜕x22 𝜕y22 𝜕z22 (1.5) Here 𝜓T (x1 , y1 , z1 , x2 , y2 , z2 ) is the wavefunction of the two-electron system, VT (x1 , y1 , z1 , x2 , y2 , z2 ) is the potential energy for the two-electron system and ET is the total energy of the two-electron system. The spatial coordinates of the two electrons are (x1 , y1 , z1 ) and (x2 , y2 , z2 ). To simplify our treatment of the two electrons we will start by assuming that the electrons do not interact with each other. This means that we are neglecting coulomb repulsion between the electrons. The potential energy of the total system is then simply the sum of the potential energy of each electron under the influence of the helium nucleus. Now the potential energy can be expressed as the sum of two identical potential energy functions V(x1 , y1 , z1 ) for the two electrons and we can write: VT (x1 , y1 , z1 , x2 , y2 , z2 ) = V(x1 , y1 , z1 ) + V(x2 , y2 , z2 ) Substituting this into Equation 1.5 we obtain: ) ) ( ( 2 2 2 2 2 2 ℏ2 𝜕 𝜓T 𝜕 𝜓T 𝜕 𝜓T ℏ2 𝜕 𝜓 T 𝜕 𝜓 T 𝜕 𝜓 T − − + + + + 2m 2m 𝜕x12 𝜕y21 𝜕z21 𝜕x22 𝜕y22 𝜕z22 + V(x1 , y1 , z1 )𝜓T + V(x2 , y2 , z2 )𝜓T = ET 𝜓T

(1.6)

12

Materials for Solid State Lighting and Displays

If we look for solutions for 𝜓T of the form Equation 1.6 becomes ( 𝜕2 ℏ2 𝜕2 + + 𝜓(x2 , y2 , z2 ) − 2m 𝜕x12 𝜕y21 ( ℏ2 𝜕2 𝜕2 − + 2 𝜓(x1 , y1 , z1 ) 2 2m 𝜕x2 𝜕y2

𝜓T = 𝜓(x1 , y1 , z1 )𝜓(x2 , y2 , z2 ) then 𝜕2 𝜕z21

)

𝜕2 + 2 𝜕z2

𝜓(x1 , y1 , z1 ) ) 𝜓(x2 , y2 , z2 )

+ V(x1 , y1 , z1 )𝜓(x1 , y1 , z1 )𝜓(x2 , y2 , z2 ) + V(x2 , y2 , z2 )𝜓(x1 , y1 , z1 )𝜓(x2 , y2 , z2 ) = ET 𝜓(x1 , y1 , z1 )𝜓(x2 , y2 , z2 )

(1.7)

Dividing Equation 1.7 by 𝜓(x1 , y1 , z1 )𝜓(x2 , y2 , z2 ) we obtain: ( ) ℏ2 𝜕2 𝜕2 1 𝜕2 𝜓(x1 , y1 , z1 ) − + + 2m 𝜓(x1 , y1 , z1 ) 𝜕x2 𝜕y2 𝜕z2 1 1 1 ) ( ℏ2 𝜕2 𝜕2 1 𝜕2 − 𝜓(x2 , y2 , z2 ) + + 2m 𝜓(x2 , y2 , z2 ) 𝜕x2 𝜕y2 𝜕z2 2 2 2 + V(x1 , y1 , z1 ) + V(x2 , y2 , z2 ) = ET Since the first and third terms are only a function of (x1 , y1 , z1 ) and the second and fourth terms are only a function of (x2 , y2 , z2 ), and furthermore since the equation must be satisfied for independent choices of (x1 , y1 , z1 ) and (x2 , y2 , z2 ) it follows that we must independently satisfy two equations, namely ) ( 𝜕2 𝜕2 ℏ2 1 𝜕2 𝜓(x1 , y1 , z1 ) + V(x1 , y1 , z1 ) = E1 + + − 2m 𝜓(x1 , y1 , z1 ) 𝜕x2 𝜕y2 𝜕z2 1 1 1 and ℏ2 1 − 2m 𝜓(x2 , y2 , z2 )

(

𝜕2 𝜕2 𝜕2 + + 𝜕x22 𝜕y22 𝜕z22

) 𝜓(x2 , y2 , z2 ) + V(x2 , y2 , z2 ) = E2

These are both identical one-electron Schrödinger equations. We have used the technique of separation of variables. We have considered only the spatial parts of the wavefunctions of the electrons; however, electrons also have spin. In order to include spin the wavefunctions must also define the spin direction of the electron. We will write a complete wavefunction [𝜓(x1 , y1 , z1 )𝜓(S)]a , which is the wavefunction for one electron where 𝜓(x1 , y1 , z1 ) describes the spatial part and the spin wavefunction 𝜓(S) describes the spin part, which can be spin up or spin down. There will be four quantum numbers associated with each wavefunction of which the first three arise from the spatial part. A fourth quantum number, which can be +1∕2 or −1∕2 for the spin part, defines the direction of the spin part. Rather than writing the full set of quantum numbers for each

Principles of Solid State Luminescence

13

wavefunction we will use the subscript a to denote the set of four quantum numbers. For the other electron the analogous wavefunction is [𝜓(x2 , y2 , z2 )𝜓(S)]b indicating that this electron has its own set of four quantum numbers denoted by subscript b. Now the wavefunction of the two-electron System including spin becomes: 𝜓T1 = [𝜓(x1 , y1 , z1 )𝜓(S)]a [𝜓(x2 , y2 , z2 )𝜓(S)]b

(1.8a)

The probability distribution function, which describes the spatial probability density function of the two-electron system, is |𝜓T |2 , which can be written as: ∗

|𝜓T1 |2 = 𝜓T 𝜓T1 1

∗

∗

= [𝜓(x1 , y1 , z1 )𝜓(S)]a [𝜓(x2 , y2 , z2 )𝜓(S)]b [𝜓(x1 , y1 , z1 )𝜓(S)]a [𝜓(x2 , y2 , z2 )𝜓(S)]b

(1.8b)

If the electrons were distinguishable then we would need also to consider the case where the electrons were in the opposite states, and in this case 𝜓T2 = [𝜓(x1 , y1 , z1 )𝜓(S)]b [𝜓(x2 , y2 , z2 )𝜓(S)]a

(1.9a)

Now the probability density of the two-electron system would be: ∗

|𝜓T2 |2 = 𝜓T 𝜓T2 2

∗

∗

= [𝜓(x1 , y1 , z1 )𝜓(S)]b [𝜓(x2 , y2 , z2 )𝜓(S)]a [𝜓(x1 , y1 , z1 )𝜓(S)]b [𝜓(x2 , y2 , z2 )𝜓(S)]a

(1.9b)

Clearly Equation 1.9b is not the same as Equation 1.8b and when the subscripts are switched the form of |𝜓T |2 changes. This specifically contradicts the requirement, that measurable quantities such as the spatial distribution function of the two-electron system remain the same regardless of the interchange of the electrons. In order to resolve this difficulty, it is possible to write wavefunctions of the two-electron system that are linear combinations of the two possible electron wavefunctions. We write a symmetric wavefunction 𝜓S for the two-electron system as: 1 𝜓S = √ [𝜓T1 + 𝜓T2 ] 2

(1.10)

and an antisymmetric wavefunction 𝜓A for the two-electron system as: 1 𝜓A = √ [𝜓T1 − 𝜓T2 ] 2

(1.11)

If 𝜓S is used in place of 𝜓T to calculate the probability density function |𝜓S |2 , the result will be independent of the choice of the subscripts, In addition since both 𝜓T 1 and 𝜓T 2 are valid solutions to Schrödinger’s equation (Equation 1.6) and since 𝜓S is a linear combination of these solutions it follows that 𝜓S is also a valid solution. The same argument applies to 𝜓A . We will now examine just the spin parts of the wavefunctions for each electron. We need to consider all possible spin wavefunctions for the two electrons. The individual electron

14

Materials for Solid State Lighting and Displays

spin wavefunctions must be multiplied to obtain the spin part of the wavefunction for the two-electron system as indicated in Equations 1.8 or 1.9, and we obtain four possibilities, namely 𝜓 1 𝜓− 1 or 𝜓− 1 𝜓 1 or 𝜓 1 𝜓 1 or 𝜓− 1 𝜓− 1 . 2

2

2

2

2

2

2

2

For the first two possibilities to satisfy the requirement that the spin part of the new twoelectron wavefunction does not depend on which electron is which, a symmetric or an antisymmetric spin function is required. In the symmetric case we can use a linear combination of wavefunctions: ( ) 1 (1.12) 𝜓 = √ 𝜓 1 𝜓− 1 + 𝜓− 1 𝜓 1 2 2 2 2 2 This is a symmetric spin wavefunction since changing the labels does not affect the result. The total spin for this symmetric system turns out to be s = 1, There is also an antisymmetric case for which ( ) 1 𝜓 = √ 𝜓 1 𝜓− 1 − 𝜓− 1 𝜓 1 (1.13) 2 2 2 2 2 Here, changing the sign of the labels changes the sign of the linear combination but does not change any measurable properties and this is therefore also consistent with the requirements for a proper description of indistinguishable particles. In this antisymmetric system the total spin turns out to be s = 0. The final two possibilities are symmetric cases since switching the labels makes no difference. These cases therefore do not require the use of linear combinations to be consistent with indistinguishability and are simply 𝜓 = 𝜓1 𝜓1 2

(1.14)

2

and 𝜓 = 𝜓− 1 𝜓 − 1 2

(1.15)

2

These symmetric cases both have spin s = 1. In summary, there are four cases, three of which, given by Equations 1.12, 1.14, and 1.15, are symmetric spin states and have total spin s = 1, and one of which, given by Equation 1.13, is antisymmetric and has total spin s = 0. Note that total spin is not always simply the sum of the individual spins of the two electrons, but must take into account the addition rules for quantum spin vectors. (See reference [1]) The three symmetric cases are appropriately called triplet states and the one antisymmetric case is called a singlet state. Table 1.1 lists the four possible states. We must now return to the wavefunctions shown in Equations 1.10 and 1.11. The antisymmetric wavefunction 𝜓A may be written using Equations 1.11, 1.8a and 1.9a as: 1 𝜓A = √ [𝜓T1 − 𝜓T2 ] 2 1 = √ {[𝜓(x1 , y1 , z1 )𝜓(S)]a [𝜓(x2 , y2 , z2 )𝜓(S)]b 2 −[𝜓(x1 , y1 , z1 )𝜓(S)]b [𝜓(x2 , y2 , z2 )𝜓(S)]a }

(1.16)

Principles of Solid State Luminescence Table 1.1 State

15

Possible spin states for a two-electron system

Prob- Total Spin ability Spin arrangement

Spin Spatial summetry symmetry

Singlet

25%

0

𝜓 1 𝜓− 1 − 𝜓− 1 𝜓 1

Antisymmetric

Triplet

75%

1

𝜓 1 𝜓− 1 + 𝜓− 1 𝜓 1

Symmetric Antisymmetric

2

2

2

2

2

2

or 𝜓1 𝜓1 2

2

2

Spatial Dipole-allowed attributes transition to/from singlet ground state

Symmetric Electrons Yes close to each other Electrons far apart

No

2

or 𝜓− 1 𝜓− 1 2

2

If, in violation of the Pauli exclusion principle, the two electrons were in the same quantum state 𝜓T = 𝜓T1 = 𝜓T2 which includes both position and spin, then Equation 1.16 immediately yields 𝜓A = 0, which means that such a situation cannot occur. If the symmetric wavefunction 𝜓S of Equation 1.10 was used instead of 𝜓A , the value of 𝜓S would not be zero for two electrons in the same quantum state. For this reason, a more complete statement of the Pauli exclusion principle is that the wavefunction of a system of two or more indistinguishable electrons must be antisymmetric. In order to obtain an antisymmetric wavefunction, from Equation 1.16 either the spin part or the spatial part of the wavefunction may be antisymmetric. If the spin part is antisymmetric, which is a singlet state, then the Pauli exclusion restriction on the spatial part of the wavefunction may be lifted. The two electrons may occupy the same spatial wavefunction and they may have a high probability of being close to each other. If the spin part is symmetric this is a triplet state and the spatial part of the wavefunction must be antisymmetric. The spatial density function of the antisymmetric wavefunction causes the two electrons to have a higher probability of existing further apart, because they are in distinct spatial wavefunctions. If we now introduce the coloumb repulsion between the electrons it becomes evident that if the spin state is a singlet state, the repulsion will be higher because the electrons spend more time close to each other. If the spin state is a triplet state, the repulsion is weaker because the electrons spend more time further apart. Now let us return to the helium atom as an example of this. Assume one helium electron is in the ground state of helium, which is the 1s state, and the second helium electron is in an excited state. This corresponds to an excited helium atom, and we need to understand this configuration because radiation always involves excited states. The two helium electrons can be in a triplet state or in a singlet state. Strong dipole radiation is observed from the singlet state only, and the triplet states do not radiate. We can understand the lack of radiation from the triplet states by examining spin. The total spin

16

Materials for Solid State Lighting and Displays

of a triplet state is s = 1. The ground state of helium, however, has no net spin because if the two electrons are in the same n = 1 energy level the spins must be in opposing directions to satisfy the Pauli exclusion principle, and there is no net spin. The ground state of helium is therefore a singlet state. There can be no triplet states in the ground state of the helium atom. There is a net magnetic moment generated by an electron due to its spin. This fundamental quantity of magnetism due to the spin of an electron is known as the Bohr magneton. If the two helium electrons are in a triplet state there is a net magnetic moment, which can be expressed in terms of the Bohr magneton since the total spin s = 1. This means that a magnetic moment exists in the excited triplet state of helium. Photons have no charge and hence no magnetic moment. Because of this a dipole transition from an excited triplet state to the ground singlet state is forbidden because the triplet state has a magnetic moment but the singlet state does not, and the net magnetic moment cannot be conserved. In contrast to this the dipole transition from an excited singlet state to the ground singlet state is allowed and strong dipole radiation is observed. The triplet states of helium are slightly lower in energy than the singlet states. The triplet states involve symmetric spin states, which means that the spin parts of the wavefunctions are symmetric. This forces the spatial parts of the wavefunctions to be antisymmetric, as illustrated in Figure 1.8 and the electrons are, on average, more separated. As a result, the repulsion between the ground state electron and the excited state electron is weaker. The excited state electron is therefore more strongly bound to the nucleus and it exists in a lower energy state. The observed radiation is consistent with the energy difference between the higher energy singlet state and the ground singlet state. Direct, dipole-allowed radiation from the triplet excited state to the ground singlet state is forbidden. See Figure 1.9. We have used helium atoms to illustrate the behavior of a two-electron system; however, we now need to apply our understanding of these results to molecular electrons, which are important for organic light emitting and absorbing materials. Molecules are the basis for organic electronic materials and molecules always contain two or more electrons in a molecular system.

1.6

Molecular Excitons

In inorganic semiconductors electrons and holes exist as distributed wavefunctions, which prevents the formation of stable excitons at room temperature. In contrast to this, holes and electrons are localized within a given molecule in organic semiconductors, and the molecular exiton is thereby both stabilized and bound within a molecule of the organic semiconductor. In organic semiconductors, which are composed of molecules, excitons are clearly evident at room temperature and also at higher operational device temperatures. An exciton in an organic semiconductor is an excited state of the molecule. A molecule contains a series of electron energy levels associated with a series of molecular orbitals that are complicated to calculate directly from Schrödinger’s equation because this is a multi-body problem. These molecular orbitals may be occupied or unoccupied. When a molecule absorbs a quantum of energy that corresponds to a transition from one molecular orbital to another higher energy molecular orbital, the resulting electronic excited state of the molecule is a molecular exciton comprising an electron and a hole within the molecule.

Principles of Solid State Luminescence

17

spin ψ

ψ x

x

ψs x (a) spin ψ

ψ x

x

ψA x

(b)

Figure 1.8 A depiction of the symmetric and antisymmetric wavefunctions and spatial density functions of a two-electron system. (a) Singlet state with electrons closer to each other on average. (b) Triplet state with electrons further apart on average

An electron is said to be found in the lowest unoccupied molecular orbital and a hole in the highest occupied molecular orbital, and since they are both contained within the same molecule the electron-hole state is said to be bound. A bound exciton results, which is spatially localized to a given molecule in an organic semiconductor. Organic molecule energy levels are relevant to OLEDs. These molecular excitons can be classified as in the case of excited states of the helium atom, and either singlet or triplet excited states in molecules are possible. The results from Section 1.5 are relevant to these molecular excitons and the same concepts involving

18

Materials for Solid State Lighting and Displays Excited singlet state ES Excited triplet state

ET

Dipole allowed transition

Ground state

Eg

Figure 1.9 Energy level diagram showing a ground state and excited singlet and triplet states. The excited triplet state is slightly lower in energy compared with the excited singlet state because two triplet state electrons are, on average, further apart than two singlet state electrons. Radiative emission from an electron in the excited singlet state to the ground singlet state is dipole-allowed. Radiative emission between the excited triplet state and the ground state requires an additional angular momentum exchange. See Section 1.6

electron spin, the Pauli exclusion principle, and indistinguishability are relevant because the molecule contains two or more electrons. If a molecule in its unexcited state absorbs a photon of light it may be excited forming an exciton in a singlet state with spin s = 0. These excited molecules typically have characteristic lifetimes on the order of nanoseconds, after which the excitation energy may be released in the form of a photon and the molecule undergoes fluorescence by a dipole-allowed process returning to its ground state. It is also possible for the molecule to be excited to form an exciton by electrical means rather than by the absorption of a photon which is the situation in OLEDs. Under electrical excitation the exciton may be in a singlet or a triplet state since electrical excitation, unlike photon absorption, does not require the total spin change to be zero. There is a 75% probability of a triplet exciton and 25% probability of a singlet exciton, as described in Table 1.1. The probability of fluorescence is therefore reduced under electrical excitation to 25% because the decay of triplet excitons is not dipole-allowed. Another process may take place, however. Triplet, excitons have a spin state with s = 1 and these spin states can frequently be coupled with the orbital angular momentum of molecular electrons, which influences the effective magnetic moment of a molecular exciton. The restriction on dipole radiation can be partly removed by this coupling, and light emission over relatively long characteristic radiation lifetimes is observed in specific molecules. These longer lifetimes from triplet states are generally on the order of milliseconds and the process is called phosphorescence, in contrast with the shorter lifetime fluorescence from singlet states. Since excited triplet states have slightly lower energy levels than excited singlet states, triplet phosphorescence has a longer wavelength than singlet fluorescence in a given molecule (see Chapters 4 and 5).

Principles of Solid State Luminescence

19

In addition, there are other ways that a molecular exciton can lose energy. There are three possible energy loss processes that involve energy transfer from one molecule to another molecule. One important process is known as Förster resonance energy transfer. Here a molecular exciton in one molecule is established but a neighboring molecule is not initially excited. The excited molecule will establish an oscillating dipole moment as its exciton starts to decay in energy as a superposition state. The radiation field from this dipole is experienced by the neighboring molecule as an oscillating field and a superposition state in the neighboring molecule is also established. The originally excited molecule loses energy through this resonance energy transfer process to the neighboring molecule and finally energy is conserved since the initial excitation energy is transferred to the neighboring molecule without the formation of a photon. This is not the same process as photon generation and absorption since a complete photon is never created; however, only dipole-allowed transitions from excited singlet states can participate in Förster resonance energy transfer. Förster energy transfer depends strongly on the intermolecular spacing, and the rate of energy transfer falls off as R16 where R is the distance between the two molecules, A simplified picture of this can be obtained using the result for the electric field of a static dipole. This field falls off as R13 Since the energy density in a field is proportional to the square of the field strength it follows that the energy available to the neighboring molecule falls of as R16 . This then determines the rate of energy transfer. Dexter electron transfer is a second energy transfer mechanism in which an excited electron state transfers from one molecule (the donor molecule) to a second molecule (the acceptor molecule). This requires a wavefunction overlap between the donor and acceptor, which can only occur at extremely short distances typically of the order 10–20 Å. The Dexter process involves the transfer of the electron and hole from molecule to molecule. The donor’s excited state may be exchanged in a single step, or in two separate charge exchange steps. The driving force is the decrease in system energy due to the transfer. This implies that the donor molecule and acceptor molecule are different molecules. This is relevant to a range of important OLED devices. The Dexter energy transfer rate is proportional to e−𝛼R where R is the intermolecular spacing. The exponential form is due to the exponential decrease in the wavefunction density function with distance. Finally, a third process is radiative energy transfer. In this case a photon emitted by the host is absorbed by the guest molecule. The photon may be formed by dipole radiation from the host molecule and absorbed by the converse process of dipole absorption in the guest molecule.

1.7

Band-to-Band Transitions

In inorganic semiconductors the recombination between an electron and a hole occurs to yield a photon, or conversely the absorption of a photon yields a hole–electron pair. The electron is in the conduction band and the hole is in the valence band. It is very useful to analyze these processes in the context of band theory for inorganic semiconductor LEDs. Consider the direct-gap semiconductor having approximately parabolic conduction and valence bands near the bottom and top of these bands, respectively, as in Figure 1.10. Two possible transition energies, E1 and E2 , are shown, which produce two photons having

20

Materials for Solid State Lighting and Displays

Ec E1

E2

Ev E

k (a)

ΔEc Ec E2 Ev

ΔEv E k

Δk (b)

Figure 1.10 (a) Parabolic conduction and valence bands in a direct-gap semiconductor showing two possible transitions. (b) Two ranges of energies ΔEv in the valence band and ΔEc in the conduction band determine the photon emission rate in a small energy range about a specific transition energy. Note that the two broken vertical lines in (b) show that the range of transition energies at E2 is the sum of ΔEc and ΔEv

two different wavelengths. Due to the very small momentum of a photon, the recombination of an electron and a hole occurs almost vertically in this diagram, to satisfy conservation of momentum. The x-axis represents the wavenumber k, which is proportional to momentum. Conduction band electrons have energy Ee = Ec +

ℏ2 k 2 ∗ 2me

Eh = Ev −

ℏ2 k 2 ∗ 2mh

and for holes we have:

In order to determine the emission/absorption spectrum of a direct-gap semiconductor we need to find the probability of a recombination taking place as a function of energy E. This transition probability depends on an appropriate density of states function multiplied by probability functions that describe whether or not the states are occupied.

Principles of Solid State Luminescence

21

We will first determine the appropriate density of states function. Any transition in Figure 1.10 takes place at a fixed value of reciprocal space where k is constant. The same set of points located in reciprocal space or k-space gives rise to states both in the valence band and in the conduction band. In our picture of energy bands plotted as E versus k, a given position on the k-axis intersects all the energy bands including the valence and conduction bands. There is therefore a state in the conduction band corresponding to a state in the valence band at a specific value of k. Therefore, in order to determine the photon emission rate over a specific range of photon energies we need to find the appropriate density of states function for a transition between a group of states in the conduction band and the corresponding group of states in the valence band. This means we need to determine the number of states in reciprocal space or k-space that give rise to the corresponding set of transition, energies that can occur over a small radiation energy range ΔE centred at some transition energy in Figure 1.10. For example, the appropriate number of states can be found at E2 in Figure 1.10b by considering a small range of k-states Δk that correspond to small differential energy ranges ΔEc and ΔEv and then finding the total number of band states that fall within the range ΔE. The emission energy from these states will be centred at E2 and will have an emission energy range ΔE = ΔEc + ΔEv producing a portion of the observed emission spectrum. The density of transitions is determined by the density of states in the joint dispersion relation, which will now be introduced. The available energy for any transition is given by: E(k) = hv = Ee (k) − Eh (k) and upon substitution we can obtain the joint dispersion relation, which adds the dispersion relations from both the valence and conduction bands. We can express this transition energy E and determine the joint dispersion relation from Figure 1.10a as: E(k) = h𝜐 = Ec − Ev + where

ℏ2 k 2 ℏ2 k 2 ℏ2 k 2 ∗ + ∗ = Eg + 2𝜇 2me 2mh

(1.17)

1 1 1 = ∗ + ∗ 𝜇 me mh

Note that a range of k-states Δk will result in an energy range ΔE = ΔEc + ΔEv in the joint dispersion relation because the joint dispersion relation provides the sum of the relevant ranges of energy in the two bands as required. The smallest possible value of transition energy E in the joint dispersion relation occurs at k = 0 where E = Eg from Equation 1.17 which is consistent with Figure 1.10. If we can determine the density of states in the joint dispersion relation, we will therefore have the density of possible photon emission transitions available in a certain range of energies. The density of states function for an energy conduction band is: 3 ( ∗ ) √ 2 2m 1 E D(E) = 𝜋 2 2 2 𝜋 ℏ ∗

We can formulate a joint density of states function by substituting 𝜇 in place of m .

22

Materials for Solid State Lighting and Displays

Recognizing that the density of states function must be zero for E < Eg we obtain: )3 ( 1 2𝜇 2 1 2 (E − E ) Djoint (E) = 𝜋 g 2 𝜋 2 ℏ2

(1.18)

This is known as the joint density of states function valid for E ≥ Eg . To determine the probability of occupancy of states in the bands, we use Fermi–Dirac statistics. The Boltzmann approximation for the probability of occupancy of carriers in a conduction band is: ] [ (Ee − Ef ) F(E) ≅ exp − kT and for a valence band the probability of a hole is given by: [ ] (Eh − Ef ) 1 − F(E) ≅ exp kT Since a transition requires both an electron in the conduction band and a hole in the valence band, the probability of a transition will be proportional to: ) ( ) ( (Ee − Eh ) E (1.19) = exp − F(E)[1 − F(E)] = exp − kT kT Including the density of states function, we conclude that the probability p(E) of an electron–hole pair recombination applicable to an LED is proportional to the product of the joint density of states function and the function F(E)[1 − F(E)], which yields: p(E) ∝ D(E − Eg )F(E)[1 − F(E)]

(1.20)

Now using Equations 1.18, 1.19, and 1.20, we obtain the photon emission rate R(E) as: ) ( E (1.21) R(E) ∝ (E − Eg )1∕2 exp − kT The result is shown graphically in Figure 1.11. If we differentiate Equation 1.21 with respect to E and set dR(E) = 0 the maximum is dE kT found to occur at E = Eg + 2 . From this, we can evaluate the full width at half maximum to be 1.8kT. If we were interested in optical absorption instead of emission for a direct gap semiconductor, the absorption constant 𝛼 can be evaluated using Equation 1.18 and we obtain: 1

𝛼(hv) ∝ (hv − Eg ) 2

(1.22)

We consider the valence band to be fully occupied by electrons and the conduction band to be empty. In this case the absorption rate depends on the joint density of states function only and is independent of Fermi–Dirac statistics. The absorption edge for a direct gap semiconductor is illustrated in Figure 1.12. This absorption edge is only valid for direct gap semiconductors, and only when parabolic band-shapes are valid. If hv ≫ Eg this will not be the case and measured absorption coefficients will differ from this theory. In an indirect gap semiconductor, the absorption 𝛼 increases more gradually with photon energy hv until a direct gap transition can occur.

Principles of Solid State Luminescence

23

Probability of occupancy Density of states Photon emission rate

1.8kT

Eg

Eg + kT/2

Energy

Figure 1.11 Photon emission rate as a function of energy for a direct gap transition of an LED. Note that at low energies the emission drops off due to the decrease in the density of 1

states term (E − Eg ) 2 and at high energies the emission drops off due to the Boltzmann term exp(−E/kT)

Absorption

Eg

Figure 1.12

1.8

Photon energy hv

Absorption edge for a direct gap semiconductor

Photometric Units

The most important applications of LEDs and OLEDs are for visible illumination and displays. This requires the use of units to measure the brightness and color of light output. The power in watts and wavelength of emission are often not adequate descriptors of light emission. The human visual system has a variety of attributes that have given rise to more appropriate units and ways of measuring light output. This human visual system includes the eye, the optic nerve and the brain, which interpret light in a unique way. Watts, for example, are considered radiometric units, and this section introduces photometric units and relates them to radiometric units. Luminous intensity is a photometric quantity that represents the perceived brightness of an optical source by the human eye. The unit of luminous intensity is the candela (cd).

24

Materials for Solid State Lighting and Displays

1

683

0.9

615

0.8

546

0.7

478

0.6

410

0.5

342

0.4

273

0.3

205

0.2

137

0.1

68

0

Luminous Efficacy lm/W

Relative Luminous Efficiency

One cd is the luminous intensity of a source that emits 1/683 watt of light at 555 nm into a solid angle of one steradian. The candle was the inspiration for this unit, and a candle does produce a luminous intensity of approximately 1 cd. Luminous flux is another photometric unit that represents the light power of a source. The unit of luminous flux is the lumen (lm). A candle that produces a luminous intensity of 1 cd produces 4𝜋 lumens of light power. If the source is spherically symmetrical then there are 4𝜋 steradians in a sphere, and a luminous flux of 1 lm is emitted per steradian. A third quantity, luminance, refers to the luminous intensity of a source divided by an area through which the source light is being emitted; it has units of cd m−2 . In the case of an LED die or semiconductor chip light source the luminance depends on the size of the die. The smaller the die that can achieve a specified luminous intensity, the higher the luminance of this die. The advantage of these units is that they directly relate to perceived brightnesses, whereas radiation measured in watts may be visible, or invisible depending on the emission spectrum. Photometric units of luminous intensity, luminous flux and luminance take into account the relative sensitivity of the human vision system to the specific light spectrum associated with a given light source. The eye sensitivity function is well known for the average human eye. Figure 1.13 shows the perceived brightness for the human visual system of a light source that emits a constant

0 350

400

450

500 550 600 Wavelength of Light in nm

650

700

750

Figure 1.13 The eye sensitivity function. The left scale is referenced to the peak of the human eye response at 555 nm. The right scale is in units of luminous efficacy. International Commission on Illumination (Commission Internationale de l’Eclairage, or CIE), 1931 and 1978

Principles of Solid State Luminescence

25

optical power that is independent of wavelength. The left scale has a maximum of 1 and is referenced to the peak of the human eye response at 555 nm. The right scale is in units of luminous efficacy (lm W−1 ), which reaches a maximum of 683 lm W−1 at 555 nm. Using Figure 1.13, luminous intensity can now be determined for other wavelengths of light. An important measure of the overall efficiency of a light source can be obtained using luminous efficacy from Figure 1.13. A hypothetical monochromatic electroluminescent light source emitting at 555 nm that consumes 1 W of electrical power and produces 683 lm has an electrical-to-optical conversion efficiency of 100%. A hypothetical monochromatic light source emitting at 450 nm that consumes 1 W of electrical power and produces approximately 30 lm also has a conversion efficiency of 100%. The luminous flux of a blue LED or a red LED that consumes 1 W of electric power may be lower than for a green LED; however, this does not necessarily mean that they are less efficient. Luminous efficiency values for a number of light sources may be described in units of lm W−1 , or light power divided by electrical input power. Luminous efficiency can never exceed luminous efficacy for a light source having a given spectrum. The perceived color of a light source is determined by its spectrum. The human visual system and the brain create our perception of color. For example, we perceive a mixture of red and green light as yellow even though none of the photons arriving at our eyes is yellow. The human eye contains light receptors on the retina that are sensitive in fairly broad bands centred at the red, the green and the blue parts of the visible spectrum. Color is determined by the relative stimulation of these receptors. For example, a light source consisting of a combination of red and green light excites the red and green receptors, as does a pure yellow light source, and we therefore perceive both light sources as yellow in color. Since the colors we observe are perceptions of the human visual system, a color space has been developed and formalized that allows all the colors we recognize to be represented on a two-dimensional graph called the colour space chromaticity diagram (Figure 1.14). The diagram was created by the International Commission on Illumination (Commission Internationale de l’Éclairage, or CIE) in 1931, and is therefore often referred to as the CIE diagram. CIE x and y color coordinates are shown that can be used to specify the color point of any light source. The outer boundary of this color space refers to monochromatic light sources that emit light at a single wavelength. As we move to the center of the diagram to approach white light the light source becomes increasingly less monochromatic. Hence a source having a spectrum of a finite width will be situated some distance inside the boundary of the color space. If two light sources emit light at two distinct wavelengths anywhere on the CIE diagram and these light sources are combined into a single light beam, the human eye will interpret the color of the light beam as existing on a straight line connecting the locations of the two sources on the CIE diagram. The position on the straight line of this new color will depend on the relative radiation power from each of the two light sources. If three light sources emit light at three distinct wavelengths that are anywhere on the CIE diagram and these light sources are combined into a single light beam, the human eye will interpret the color of the light beam as existing within a triangular region of the CIE diagram having vertices at each of the three sources. The position within the triangle of this

26

Materials for Solid State Lighting and Displays 530 nm 0.8 540 nm 510 nm 550 nm

0.7

560 nm 0.6 500 nm 580 nm

0.5

590 nm

3200 K A 0.4 ⎨⎝

Sunlight D65

2000 K ⎛

y - chromaticity coordinate

570 nm

E

0.3 490 nm

B

600 nm

4000 K

1000 K

C 10 000 K

620 nm

0.2 480 nm

Location of Planckian blackbody radiators (Planckian locus)

0.1 470 nm 450 nm 0.0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

x - chromaticity coordinate

Figure 1.14 Color space chromaticity diagram showing colors perceptible to the human eye. The center region of the diagram indicates a Planckian locus, which corresponds to the colors of emission from a blackbody source having temperatures from 1000 to 10 000 K. This locus includes the solar spectrum corresponding to a 5250 K blackbody. International Commission on Illumination (Commission Internationale de l’Eclairage, or CIE), 1931

new color will depend on the relative radiation power from each of the three light sources. This ability to produce a large number of colors of light from only three light sources forms the basis for trichromatic illumination. Lamps and displays routinely take advantage of this principle. It is clear that the biggest triangle will be available if red, green and blue light sources are selected to define the vertices of the color triangle. This colour triangle is often referred to as a color space that is enabled by a specific set of three light emitters. The Planckian blackbody locus is also shown in Figure 1.14. All the color points on this line represent a blackbody source of a specific given temperature. For example, a point at approximately 5000 K represents the color of the sun. A tungsten filament lamp with a filament temperature of 3000 K is also a point on this line.

Principles of Solid State Luminescence

Figure 1.15

Name

Appr. Munsell

Appearance under daylight

TUS01

7,5 R 6/4

Light grayish red

TUS02

5 Y 6/4

Dark grayish yellow

TCS03

5 GY 6/8

Strong yellow green

TCS04

2,5 G 6/6

Moderate yellowish green

TCS05

10 BG 6/4

Light bluish green

TCS06

5 PB 6/8

Light blue

TCS07

2,5 P 6/8

Light violet

TCS08

10 P 6/8

Light reddish purple

27

Swatch

Eight standard color samples used to determine the colour rendering index

The colors on this Planckian locus are important since they are used as reference spectra for non-blackbody light emitters such as LEDs. A measure used to quantify the closeness of an LED spectrum to a blackbody radiator is the color rendering index or CRI. Appropriately named, a CRI value indicates how well a given light source can substitute for a blackbody source in terms of illuminating a wide range of colored or pigmented objects. The highest CRI of a light source is 100 indicating a perfect blackbody spectrum. Lamps achieving CRI values between 90 and 100 are considered “museum grade” lamps and may be used to view art, color samples, and pigments with a high degree of accuracy. For more general illumination tasks a CRI between 80 and 90 is often considered adequate. The CRI is calculated by comparing the color rendering of the source in question to that of a blackbody radiator for sources with correlated color temperatures under 5000 K, and a phase of daylight otherwise (e.g., D65). The procedure makes use of a set of standard test color samples shown in Figure 1.15. A simplified summary of the steps used to determine the CRI are as follows: 1. Find the chromaticity coordinates of the test source in the CIE 1960 color space. The CIE 1960 color space is a modified version of the color space in Figure 1.14. 2. Determine the correlated color temperature (CCT) of the test source by finding the closest point to the Planckian locus on the chromaticity diagram. 3. If the test source has a CCT 0 Net diffusion current flows

In equilibrium, V = 0, I = 0 Reverse bias V < 0 Net drift current flows V

Figure 1.19 Resulting current–voltage characteristic of a diode showing an approximately exponentially increasing forward current and a saturated reverse bias current

np and pn are the minority carrier concentrations in the p-side and n-side of the diode, respectively. In an LED, a forward bias is applied which results in a net diffusion current. Once majority carriers diffuse across the junction into the opposite side of the diode they become minority carriers and are subject to recombination with majority carriers. This recombination process produces photons and since the 1960s, semiconductors and device structures have been developed to optimize this process resulting in today’s high performance LEDs (see Chapter 7).

References 1. Eisberg, R. and Resnick, R. (1985) Quantum Physics of Atoms, Molecules, Solids, Nuclei and Particles, Second edn, John Wiley & Sons, Inc. 2. Kittel, C. (2005) Introduction to Solid State Physics, Eighth edn, John Wiley & Sons, Inc. 3. Kitai, A.H. (2011) Principles of Solar Cells, LEDs and Diodes, John Wiley & Sons, Ltd. 4. Schubert, E.F. (2006) Light Emitting Diodes, Second edn, Cambridge University Press.

2 Quantum Dots for Displays and Solid State Lighting Jesse R. Manders1 , Debasis Bera2,3 , Lei Qian2,3 and Paul H. Holloway2,3 1 Nanosys, Inc., Milpitas, CA, USA 2 NanoPhotonica, Inc., Alachua, FL, USA 3 Department of Materials Science and Engineering, University of Florida, Gainesville, FL, USA

2.1

Introduction

The word “phosphor” comes from the Greek language and means “light bearer” to describe light emitting materials or luminescent material; barium sulfide is one of the earlier known naturally occurring phosphors [1]. A phosphor is luminescent, that is emits energy from an excited electron as light, where the excitation of the electron is caused by absorption of energy from an external source such as another electron, a photon, an electric field, or heat. An excited electron occupies a quantum state whose energy is above the minimum energy ground state. In semiconductors and insulators, the electronic ground state is commonly referred to as electrons in the valence band which is completely filled with electrons. The excited quantum state often lies in the conduction band which is empty and is separated from the valence band by an energy gap called the band gap, ΔEg in Figure 2.1. Therefore, unlike metallic materials, small continuous changes in electron energy within the band are not possible. Instead a minimum energy equal to the band gap is necessary to excite an electron in a semiconductor or insulator, and the energy released by de-excitation is often nearly equal to the band gap. For a semiconductor, the band gap energy is sufficiently large that at room temperature, very few electrons are promoted from the valence band to the conduction band, a process which leaves holes in the valence band. Figure 2.1(a) shows an energy band diagram (plot of allowed quantum state energy versus wave vector magnitude k) for a direct band gap semiconductor. In the direct band gap semiconductors, the

Materials for Solid State Lighting and Displays, First Edition. Edited by Adrian Kitai. © 2017 John Wiley & Sons Ltd. Published 2017 by John Wiley & Sons Ltd.

32

Materials for Solid State Lighting and Displays

Conduction Band

Energy

LUMO

Conduction Band

ħ2k2 EC = 2me

ΔEg

ΔEg Virtual State

EV = ħ k 2mh

2 2

HOMO Valence Band

0

k

Valence Band

0

k

Wave vector

Wave vector

(a)

(b)

Figure 2.1 Schematic band-energy diagrams for (a) direct band gap and (b) indirect band gap semiconductors

positions of the highest energy state of the valence band and the lowest energy state of the largely unoccupied conduction band are at the same k resulting in a high probability of emitting light. The case of an indirect band gap semiconductor is shown in Figure 2.1(b) and has the valence band maximum and conduction band minimum at different values of k. Therefore, electrons need to undergo a change of k-value followed by a change in energy. That is, the transition requires a change of both in energy and momentum. In other words, an indirect transition requires energy excitation of an electron simultaneous with an electron–phonon interaction to give the required momentum change. Therefore, the absorption and recombination efficiency of direct band gap materials is about four orders of magnitude larger than that of indirect material. Zinc sulfoselenide (ZnS1−x Sex ) and gallium phosphide (GaP) are examples of direct and indirect band gap compound semiconductors, respectively. As discussed above, luminescence from phosphors can be observed by exciting the electrons to higher energy states, for example into the conduction band. There are several approaches to provide this excitation, such as photoluminescence (PL), electroluminescence (EL), cathodoluminescence (CL), mechanoluminescence, chemiluminescence, and thermoluminescence [1]. In this chapter, only PL, EL, and CL (Figure 2.2) will be discussed, with most emphasis being on PL and EL. When an insulator or semiconductor absorbs electromagnetic radiation (i.e., a photon) an electron may be excited to a higher energy quantum state. If the excited electron returns (relaxes) to a lower energy quantum state by radiating a photon, the process is called PL. Some of the quantum state relaxation transitions are not allowed based on the spin and Laporte selection rules [2]. The PL intensity depends on the measurement temperature and the energy of the exciting light (known as photoluminescence excitation or PLE spectrum).

Quantum Dots for Displays and Solid State Lighting Low-energy photon emission

High-energy photon absorption

33

electrode electrode

Emission

(a) (b)

X-rays

High-energy electron Secondary electron Back-scattered electron

UV, Visible, IR emission

Auger electron

(c)

Figure 2.2 Schematic illustrations of (a) photoluminescence, (b) electroluminescence, and (c) cathodoluminescence

In general, peaks in the PLE spectrum are higher in energy than those in the PL spectrum. Figure 2.2(a) schematically illustrates the excitation and emission processes of PL. When a material emits electromagnetic radiation as a result of the application of an electric field, the process is called EL. The photon emission results from radiative recombination of electrons and holes created in - or injected into - the phosphor by the voltage between the two electrodes, as shown in Figure 2.2(b). One of the electrodes is transparent to the wavelength of the light emitted by the device. The first report of an EL device was in 1907 [3], when Henry Joseph Round observed that light was emitted from silicon carbide under the application of a high voltage. As discussed below, there is significant interest in inorganic nanophosphors combined with conducting organic materials to produce EL devices because of their potentially high efficiencies. Other advantages for EL devices in comparison with conventional lighting systems also include a large size range, flexible substrates and shapes, high brightness, long device lifetimes, lower operating temperatures, non-directionality and antiglare lighting. Depending on the applied bias, a thin-film electroluminescence (TFEL) device can be categorized as either a DC or AC (ACTFEL) device. Cathodoluminescence is emission of light from a material that is excited by energetic electrons. The exciting primary electrons can be focused into a beam and scanned across the surface (as in a scanning electron microscope) resulting in high spatial resolution CL. The CL process is shown schematically in Figure 2.2(c), along with other phenomena that result from primary electron–material interactions, for example X-ray and various electron emissions.

34

2.2

Materials for Solid State Lighting and Displays

Nanostructured Materials

Nanostructured materials, by definition, can exist as individual particles or clusters of nanoparticles with various shapes and sizes [4, 5] . Research has shown that nanostructured materials generally exhibit geometries that reflect the atomistic bonding analogous to the bulk structure. The nanomaterials are of interest because they can bridge the gap between the bulk and molecular levels and lead to entirely new avenues for applications. Nanostructured materials have a very high surface-to-volume ratio compared with their bulk counterparts. Therefore, a large fraction of atoms is present on the surface which makes them possess different thermodynamic properties. During the last two decades, a great deal of attention has been focused on the optoelectronic properties of nanostructured semiconductors with an emphasis on fabrication of the smallest possible particles. The research has revealed that many fundamental properties are size dependent in the nanometer range. For example, the density of quantum states (DOS) versus energy for periodic materials with three, two or one dimensions is shown in Figure 2.3. (See reference [6] for detailed explanations.) If the extent of the material is on the order of 1–10 nm in all three directions, the material is said to be quantum dots (QDs). A QD is zero dimensional relative to the bulk, and the DOS depends upon whether or not the QDs have aggregated, as shown in Figure 2.3. The DOS for a molecule and an atom are also shown in Figure 2.3 (see reference [6]). The density of electrons in a three-dimensional bulk crystal is so large that the energy of the quantum states becomes nearly continuous as shown in Figure 2.3. However, the limited number of electrons results in discrete quantized energies in the DOS for two-, one- and zero-dimensional structures (Figure 2.3). The presence of one electronic charge Diameter

1.2 A

Density of States (DOS)

MO theory

Dimension

zero

Atom Few molecules Single molecule

A

1-3 nm

LUM

A

HOM

A

zero

Molecule

Series of QDs 1-20 nm

Few QDs

One QD

zero

QDs 3D > 20 nm

Three, two or one

2D

1D

N(E)

Energy Bulk Material

Figure 2.3 Schematic illustration of the changes of the density of quantum states with changes in the number of atoms in materials (see text for a detailed explanation)

Quantum Dots for Displays and Solid State Lighting

35

in the QDs repels the addition of another charge and leads to a staircase-like I–V curve and DOS. The step size of the staircase is proportional to the reciprocal of the radius of the QDs. The boundaries as to when a material has the properties of bulk, QDs or atoms are dependent upon the composition and crystal structure of the compound or elemental solid. When a solid exhibits a distinct variation of optical and electronic properties with a variation of size, it can be called a nanostructure, and is categorized as (1) two dimensional, for example thin films or quantum wells, (2) one dimensional, for example quantum wires, or (3) zero dimensional, for example QDs. Although each of these categories shows interesting optical properties, our discussion will be focused on QDs. An enormous range of fundamental properties can be realized by changing the size at a constant composition or by changing the composition at constant size. In the following sections, we will discuss the history, structure and properties relationships, and the optical properties of QDs.

2.3

Quantum Dots

Nanostructured semiconductors or insulators have dimensions and numbers of atoms between the atomic-molecular level and bulk material with a band gap that depends in a complicated fashion upon a number of factors, including the bond type and strength with the nearest neighbors. For isolated atoms, there is no nearest-neighbor interaction. Therefore, sharp and narrow luminescent emission peaks are observed. A molecule consists of only a few atoms and therefore exhibits emission similar to that of an atom. A nanoparticle, however, is composed of approximately 100–10 000 atoms, and has optical properties distinct from its bulk counterpart. Nanoparticles with dimensions in the range of 1–20 nm are called QDs. Zero-dimensional QDs are often described as artificial atoms due to their 𝛿-function-like DOS which can lead to narrow optical line spectra. A significant amount of current research is aimed at using the unique optical properties of QDs in devices, such as light emitting diodes (LEDs), solar cells and biological markers. QDs are of interest in biology for several reasons including (1) higher extinction coefficients, (2) higher quantum yields, (3) less photobleaching, (4) absorbance and emissions can be tuned with size, (5) generally broad excitation window but narrow emission peaks, (6) multiple QDs can be used in the same assay with minimal interference with each other, (7) toxicity may be less than conventional organic dyes, and (8) the QDs may be functionalized with different bio-active agents [7–9]. The inorganic QDs are more photostable under ultraviolet excitation than are organic molecules, and their fluorescence is more saturated. The ability to synthesize QDs with narrow size distributions with high quantum yields [10, 11] has made QDs an attractive alternative to organic molecules in hybrid LEDs and solar cells. QDs can be broadly categorized into either elemental or compound systems. In this chapter, we will emphasize compound semiconductor based nanostructured materials. Compound materials can be categorized according to the columns in the periodic table, for example IB–VIIB (CuCl, CuBr, CuI, AgBr, etc.), IIB–IVB (ZnO, ZnS, ZnSe, ZnTe, CdO, CdS, CdSe, CdTe, etc.), IIIB–VB (GaN, GaP, GaAs, InN, InP, InSb, etc.), and IVB–VIB (PbS, PbSe, PbTe, etc.). Our discussion will be confined mainly to IIB–VIB, IIIB–VB and IVB–VIB QDs based phosphors, where frequently the B designations are dropped, that is the compounds are designated as II–VI or III–V.

36

2.3.1

Materials for Solid State Lighting and Displays

History of Quantum Dots

A process for synthesizing PbS QDs was developed more than 2000 years ago using low-cost natural materials such as PbO, Ca(OH)2 , and water [12]. The Romans and Greeks used these materials as cosmetics to dye their hair. In more recent history, control of the size of QDs in silicate glasses is one of the oldest and most frequently used techniques to control the color of glass. In the early 20th century, CdS and CdSe were incorporated into silicate glasses to get red-yellow colors. In 1932, Rocksby [13] used X-ray diffraction to determine that precipitates of CdS and CdSe caused the colors. Semiconductor doped glasses were used in optics as filters. A blue shift of the optical spectrum for nanometer sized CuCl in silicate glass was reported in 1981 by Ekimov and Onushchenko [14]. In 1982, Efros and Efros [15] advanced the postulate that quantum size effects (the change of optical and optoelectronic properties with size) could be used to control the color of glass by either changing the size or stoichiometry of CdSx Se1−x . In 1984, the change in color of colloidal solutions of semiconductor was discussed by Rosetti et al [16]. Over the last two decades, experimental and theoretical research on nanoparticles has increased significantly [17, 18]. 2.3.2

Structure and Properties Relationship

The most fascinating change of QDs with particle size kT restrict the electron and hole mobility in the crystal. Among the many properties that exhibit a dependence upon size in QDs, two are of particular importance for nanophosphors. The first is a blue shift (increase) of band gap energy when the nanoparticle diameters are below a particular value that depends on the type of semiconductor. This is called a quantum confinement effect [21, 22] and is discussed below

Quantum Dots for Displays and Solid State Lighting

37

in detail. This effect allows tuning of the energy gap with changes in the QD size. The band gap energy also depends on the composition of the semiconductor as well as the size. The second important property, as illustrated in Figure 2.3, is the observation of discrete, well separated energy states due to the small number of atoms in QDs compared with the bulk. This leads to the electronic states of each energy level exhibiting wave functions that are more atomic-like. Since the QDs solutions for Schrödinger wave equations are very similar to those for electrons bound to a nucleus, QDs are called artificial atoms, and atomic-like sharp emission peaks are possible. Typical intraband energy level spacings for QDs are in the range of 10–100 meV. QDs exhibit solid–solid phase transitions like bulk semiconductors, and these transitions have a substantial influence on the optical properties of QDs. Phase transitions in bulk materials can be induced by varying pressure, temperature, and composition [23, 24]. Bulk CdSe may exhibit either a hexagonal wurtzite or a zinc blende rock salt cubic structure with a direct or indirect band gap, respectively. Above a pressure of ∼3 GPa, the CdSe bulk semiconductor can be converted reversibly from low pressure wurtzite to the high pressure rock salt structure [25]. The low intensity optical emission from the rock salt form of CdSe is in the near-infrared spectral region at 0.67 eV (1.8 μm). Using high pressure X-ray diffraction and optical absorption, Tolbert and Alivisatos showed that the wurzite to rock salt structural transformation also occurred in CdSe QDs [23, 24], as shown in Figure 2.4. The ratio of oscillator strength between direct and indirect structures was unchanged with QD size. 1.50 indirect

1.25

direct

1.00 0.75

17.3 Å

0.50 ×10

0.25

O.D.

0.00 0.75 0.50

11.6 Å

×10

0.25 0.00 0.75 0.50 1.25

9.6 Å

×10

0.00 0.6

1.1

2.6 1.6 2.1 Photon energy (eV)

3.1

3.6

Figure 2.4 CdSe QDs exhibited direct and indirect band gaps at atmospheric and ∼9.3 GPa pressures. Arrows indicate the band gaps of the bulk CdSe at atmospheric pressure and 9 GPa, respectively. Source: Tolbert et al. 1994 [23]. Reproduced with permission of American Physical Society

38

Materials for Solid State Lighting and Displays

2.3.3

Quantum Confinement Effects on Band Gap

Quantum confinement generally results in a widening of the band gap with a decrease in the size of the QDs. The band gap in a material is the energy required to create an electron and a hole at rest (i.e., with zero kinetic energy) at a distance far enough apart that their Coulombic attraction is negligible. If one carrier approaches the other, they may form a bound electron–hole pair, that is an exciton, whose energy is a few meV lower than the band gap. This exciton behaves like a hydrogen atom, except that a hole, not a proton, forms the nucleus. Obviously, the mass of a hole is much smaller than that of a proton, which affects the solutions to the Schrödinger wave equation. The distance between the electron and hole is called the exciton Bohr radius (rB ). If m∗e and m∗h are the effective masses of electrons and holes, respectively, the exciton Bohr radius can be expressed by: ( ) 4𝜋𝜀o 𝜀r ℏ2 1 1 + (2.1) rB = m∗e m∗h e2 where 𝜀o , 𝜀r , ℏ, and e are the vacuum permittivity (vacuum dielectric constant), material relative dielectric constant, reduced Planck constant, and the charge of an electron, respectively. If the diameter (R) of a QD approaches 2rB , that is R ≈ 2rB, or R < 2rB, the motions of the electrons and holes are confined spatially to the dimension of the QD which causes an increase in the excitonic transition energy and the observed blue shift in the QD band gap and luminescence. The exciton Bohr radius is a threshold value, and the confinement effect becomes more important when the QD diameter is smaller. For small QDs, the exciton binding energy and bi-exciton binding energy (exciton–exciton interaction energy) is much larger than for bulk materials [26]. Note that for a material with a relatively large dielectric constant and small me and mh , the rB is large. Data for some semiconductors are summarized in Table 2.1. Two detailed theoretical approaches are used to better predict the exciton properties, specifically the effective mass approximation (EMA) model and linear combination of atomic orbital (LCAO) theory. 2.3.3.1

Effective Mass Approximation Model

This approach, based on the “particle-in-a-box” model, is the most widely used model to predict quantum confinement. It was first proposed by Efros and Efros [15] in 1982 and later modified by Brus [27]. It assumes a spherical particle in a potential well with an infinite potential barrier at the particle boundary. For a particle free to assume any position in the box the relationship between its energy (E) and wave vector (k) is given by: ℏ2 k 2 (2.2) 2m∗ In the EMA model, this relationship is assumed to hold for an electron or hole in the semiconductor, therefore the energy band is parabolic near the band edge. The shift of band gap energy (ΔEg ) due to confinement of the exciton in a QD with a diameter R can be expressed as: ( ) ℏ2 𝜋 2 1.8e2 1 1 1.78e2 ℏ2 𝜋 2 ∗ ΔEg = − + (2.3) − = − 0.248ERy 𝜀R 𝜀R 2𝜇R2 2R2 me mh E=

Quantum Dots for Displays and Solid State Lighting

39

Table 2.1 Band gap and Bohr radii data for selected semiconductors Group

Name

Band gap (eV) 4K

IIB–VIB

IIIB–VB

IVB–VIB IVB

300 K

Exciton binding energy (meV)

Relative dielectric constant

Crystal structure

Lattice constant (Å) a: 3.250 b: 5.207 a: 3.811 c: 6.234 5.406 5.667 6.104 5.832 a: 4.135 c: 6.749 a: 4.299 c: 7.010 6.4818

ZnO

3.44

3.37

59

8.1

Wurzite

ZnS

3.91

3.8

40

8.3

Wurzite

ZnSe ZnTe CdS

3.84 2.82 2.38 2.58 2.58

3.68 2.67 2.25 2.42 2.53

36 17 11 27 28

8.9 8.8 8.7 8.6

Zinc blende Zinc blende Zinc blende Zinc blende Wurzite

CdSe

1.84

1.714

15

9.4

Wurzite

CdTe GaN GaP GaAs InN

1.60 3.28 2.35 1.52 2.11

1.45 3.2 2.25 1.43 1.97

10 25.2 20.5 4.2 15.2

InP InAs InSb

1.42

1.35 0.354 0.17

6.0 0.6

10.3 9.3 11 13.2 9.3 15.3 12.6 15.2 16.8

Zinc blende Zinc blende Zinc blende

5.4505 5.653 a: 3.533 c: 5.693 5.867 6.0583 6.479

0.29 1.12 0.67

14.4 4.15

17.3 11.4 15.5

Diamond Diamond

5.431 5.658

PbS Si Ge

0.24 (0 K) 0.41 1.12 0.75

Zinc blende Zinc blende Zinc blende Zinc blende Wurzite

∗ is Rydberg energy. The where 𝜇 is the reduced mass of an electron–hole pair and ERy first term of Equation (2.3) represents a relation between “particle-in-a-box” quantum localization energy or confinement energy and the radius of the QD (R), whereas the second term shows the Columbic interaction energy with a R−1 dependence. The Rydberg energy term is size independent and is usually negligible, except for semiconductors with small dielectric constant [19]. Based on Equation (2.3), the first excitonic transition (i.e., the band gap) increases as the QD diameter decreases. However, the EMA model breaks down in the small QD regime [19, 28] because the E–k relationship can no longer be approximated as parabolic. Figure 2.5 shows such a deviation of theoretically predicted band gaps for CdS QDs from the experimental values.

2.3.3.2

Linear Combination of Atomic Orbital Theory and Molecular Orbital Theory

A model based on a linear combination of atomic orbitals and molecular orbitals (LCAO-MO) provides a more detailed basis for predicting the evolution of the electronic structure of clusters from atoms and/or molecules to QDs to bulk materials, and predicting the dependence of band gap on the size of the crystals. Figure 2.3 shows the results of this

40

Materials for Solid State Lighting and Displays 5.0 CdS

Exciton energy (eV)

4.5

4.0

3.5 3.0 2.5 0

20

40

60

80

100

Cluster diameter (Å)

Figure 2.5 Experimentally and theoretically determined band gap as a function of size of CdS QDs. Broken line: calculated parameters based on effective mass approximation; solid line: tight-bonding calculation; squares: experimental data. Source: Wang and Herron 1991 [19]. Reproduced with permission of American Chemical Society

approach pictorially. In a diatomic molecule, the atomic orbitals (AO) (Figure 2.3) of two individual atoms are combined, producing bonding and antibonding molecular orbitals. In this approach, nanosized QDs are considered as large molecules. As the number of atoms increases, the discrete energy band structure changes from large steps to small energy steps, that is to a more continuous energy band. The highest energy occupied (bonding) molecular orbital quantum state is called the highest occupied molecular orbital (HOMO). The lowest unoccupied antibonding orbital is called the lowest unoccupied molecular orbital (LUMO). Together, the HOMO and LUMO are the frontier orbitals for bands of energy states that are analogous to the valence band and conduction band in inorganic semiconductors, respectively. In other words, the HOMO is analogous to the valence band maximum, while the LUMO is analogous to the conduction band minimum in inorganic semiconductors. The energy difference between the HOMO and LUMO (equal to the band gap, ΔEg in Figure 2.1) increases and the bands split into discrete energy levels with reduced mixing of atomic orbitals for a small number of atoms. Therefore, the small size of the QDs results in quantized electronic band structures intermediate between the atomic/molecular and bulk crystalline molecular orbitals. Compared with the effective mass approximation, the LCAO-MO model provides a methodology to calculate the electronic structure of much smaller QDs. In contrast, this method cannot be used to calculate the energy levels of large QDs due to mathematical complexity and limitations of the computing systems. Nevertheless, the degree of quantum confinement is determined by the ratio of the diameter of a QD (R) to bulk excitonic Bohr radius (rB ). At crystal sizes less than the excitonic Bohr diameter (2rB ), semiconductor crystals exhibit translational motion confinement of the fully coupled exciton due to a strong Coulombic interaction between the electron and holes, that is, exhibit single-particle confinement behavior (sometimes called the strong confinement regime). In the intermediate size range (R ≤ rB ), the transition energies of photoexcited carriers in

Quantum Dots for Displays and Solid State Lighting

41

the crystal are determined by the relative strengths of the kinetic energy of confinement and the electron–hole interaction.

2.4

Relaxation Process of Excitons

After excitation by an external energy (photon, electric field, primary electron, etc.), electrons and holes possess high energies due to transitions from their ground state to an excited state. The energies associated with such optical absorptions are directly determined by the electronic structure of the material. The excited electron and hole may form an exciton as discussed above. The electron may recombine with the hole and relax to a lower energy state, ultimately reaching the ground state. The excess energy resulting from recombination and relaxation may be either radiative (emits photon) or nonradiative (emits phonons or Auger electrons). In PL, electrons in a material move to allowed excited quantum states upon absorption of a higher energy photon. Recombination of the electron and hole results in emission of a photon (i.e., radiative recombination leads to PL). Some radiative events from band edge, defects and nonradiative processes are depicted in Figure 2.6. These processes are discussed in detail below.

Conduction Band e

e e

e

e

h h

e

h

h

h Valence Band

Band Edge Recombination

Defect Recombination

Auger Recombination

(a)

(b)

(c)

Figure 2.6 A few radiative and nonradiative processes that can occur during luminescence. (a) Band edge recombination, (b) defect recombination, and (c) Auger recombination

42

Materials for Solid State Lighting and Displays

2.4.1

Radiative Relaxation

Radiative relaxation results in spontaneous luminescence from QDs. Such luminescence may result from band edge or near band edge transitions or from defect and/or activator quantum states. 2.4.1.1

Band Edge Emission

The most common radiative relaxation processes in intrinsic semiconductors and insulators are band edge and near band edge (exciton) emission. The recombination of an excited electron in the conduction band with a hole in the valence band is called band edge emission [Figure 2.6(a)]. As noted above, an electron and hole may be bound by a few meV to form an exciton. Therefore, radiative recombination of an exciton leads to near band edge emission at energies slightly lower than the band gap. Radiative emission may also be characterized as either fluorescence or phosphorescence, depending upon the path required to relax. Fluorescence exhibits very short radiative relaxation lifetimes (10−9 –10−6 s), but are still long compared with excitation times (10−15 –10−13 s) [29]. Radiative relaxation processes with lifetimes >10−6 s are called phosphorescence and operate under different spin-statistics than fluorescence. In a typical PL process (Figure 2.7), an electron in a phosphor is excited by absorption of an electromagnetic wave, hv, from its ground state to an excited state. Through a very fast vibrational (nonradiative) process, the excited electron relaxes to its lowest energy excited vibrational state. For electronic relaxation in molecules, nanoparticles or bulk solids, the emitted photon is red-shifted relative to the excitation photon energy/wavelength (i.e., Stokes shift as discussed below) because of the presence of vibrational levels in the

(a)

(b)

Triplet excited state

Singlet excited state

Intersystem crossing

t4

Emission Absorption hv

t1

Absorption

hv' hv

t2

t3

t5

molecular vibrational level

Ground state Tf = t1 + t2

hv' Emission

Tp = t3 + t4 + t5

hv' > hv' > hv''

Tf < Tp

Figure 2.7 Schematic diagram of (a) fluorescence (f) and (b) phosphorescence (p); hv is energy, and T and t are time. See reference [29] for detailed discussions. Source: Alexander Ross et al. 2013 [29]. Reproduced with permission of Springer

Quantum Dots for Displays and Solid State Lighting

43

excited as well as the lower energy (e.g., ground) states [Figure 2.7(a)]. Both organic and inorganic luminescent materials exhibit Stokes shifts. In organic materials, this relaxation process may be complicated by crossing from singlet to triplet excited states as shown in Figure 2.7(b) and discussed in reference [29]. When intersystem crossing occurs, the lifetime is long (10−6 –>10 s) and the emission is classified as phosphorescence. Radiative transitions in inorganic solids that are forbidden by the optical selection rules [2] may also exhibit phosphorescence. The full width at half maximum (FWHM) of a room-temperature band edge emission peak from QDs is typically in the range of 15–20 nm. Since the peak originates from a transition between the conduction and valence bands, the energy of the photons is equal to the band gap of the QD. The optical absorption spectrum reflects the band structure of the materials because of the large DOS. On the other hand, PL spectra can be dominated by the sparse lower energy defect and activator states because photoexcited carriers rapidly thermalize and are captured by states within ∼kT of the lowest energy levels [30]. Therefore, an increase in wavelength (decrease in energy) is observed between absorbed and emitted photons due to vibrations in molecular or atomic solids. As pointed out above, this red shift is called the Stokes shift and is illustrated in Figure 2.8. In a limited number of cases, the emission energy is shifted to a shorter wavelength (higher energy), and is called an anti-Stokes shift or ‘upconversion’. There is great interest in converting low-energy photons to higher-energy photons via processes such as second-harmonic generation and stimulated Raman [31]. 2.4.1.2

Defect Emission

Radiative emission from nanophosphors also comes from localized impurity and/or activator quantum states in the band gap, as shown in Figure 2.6(b). Defect states are called dark

Δ Stokes

Photoluminescence

Absorption FWHM

Wavelength

Figure 2.8

Spectral absorption and photoluminescence profile depicting the Stokes shift

44

Materials for Solid State Lighting and Displays

states when they lie inside the bands themselves [32]. Depending on the type of defect or impurity, the state can act as a donor (has excess electrons) or an acceptor (has a deficit of electrons). Electrons or holes are attracted to these sites of local charge due to Coulombic attraction. Similar to the case of excitons, trapped charge on defect/impurity sites can be modeled as a hydrogenic system where binding energy is reduced by the dielectric constant of the material [30]. These defects states can be categorized into either shallow or deep levels, where shallow level defect states have energies near the conduction band or valence band edge. In most cases, a shallow defect exhibits radiative relaxation at temperatures sufficiently low so that thermal energies (kT) do not excite the carriers out of the defect or trap states. Deep levels, on the other hand, are so long-lived that they typically experience nonradiative recombination. Luminescence from these defect levels can be used to identify their energy and their concentration is proportional to the intensity. Both PL spectral distribution and intensity change with changes of the excitation energy due to contributions from different defect energy levels and the band structure of the host. The excitation energy also determines the initial photoexcited states in the sample, but this state is short-lived because of thermalization of the photoexcited carriers via phonon emission, as discussed above. Relaxation to within kT of the lowest vibrational level of the excited states is usually orders of magnitude faster than the recombination event [30]. Defect states are expected at the surface of a QD despite the use of various passivation methods because of the large surface-to-volume ratio discussed above. The concentration of surface states on the QDs is a function of the synthesis and passivation processes. These surface states act as traps for charge carriers and excitons, which generally degrade the optical and electrical properties by increasing the rate of nonradiative recombination. However, in some cases, the surface states can also lead to radiative transitions, such as in the case of ZnO. Powders of ZnO have a green emission from defects along with a band edge near ultraviolet emission (the band gap of ZnO is 3.37 eV or 386 nm). It is also reported that the green emission will suppress the band edge emission. Theoretical and experimental studies [33, 34] showed that the defect states in a ZnO QD can be of several types including neutral, singly or doubly charged Zn vacancies (VZn ), neutral or singly charged oxygen vacancies (VO ), singly charged or neutral interstitial Zn (Zni ), interstitial O (Oi ), a complex of VO and Zni (VO Zni ), a complex of VZn and Zni (VZn Zni ), and substitution O at Zn position (OZn ). According to Aleksandra et al. [34], the singly charged oxygen vacancy (VO+ ) is located at 2.28 eV below the conduction band in the ZnO band gap and results in an emission at ∼540 nm. The most widely, but not universally, accepted mechanism for green luminescence from ZnO is the electron–hole recombination on singly ionized oxygen vacancies. In solution-based synthesis, the oxygen vacancies appear to be intrinsic and may result from heterogeneous nucleation and growth, enhanced by the large surface area. If the radiative center is associated in part with the surface, their concentration would be expected to decrease with aggregation of QDs as observed [35]. 2.4.1.3

Activator Emission

Luminescence from intentionally incorporated impurities is called extrinsic luminescence. These impurities are called activators and they perturb the band structure by creating local quantum states that commonly lie within the band gap. The predominant radiative

Quantum Dots for Displays and Solid State Lighting

45

mechanism in extrinsic luminescence is electron–hole recombination which can occur via transitions from conduction band to acceptor state, donor state to valence band, or donor state to acceptor state. In some cases, this mechanism is localized on the activator atom center. However, these localized transitions are only allowed when two selection rules are met [2]: (1) the spin selection rule, that is Δs = ± 1 where s is the spin quantum number; and (2) the Laporte selection rule (Δl = 0, where l is the orbital angular momentum quantum number). In many cases, the Laporte selection rule is relaxed due to mixing of orbitals such as d–p mixing in a crystal or ligand field where the orbitals are split into hyperfine structures. Therefore, the d–d transition is allowed in some cases for transition elements such as Mn2+ , Fe2+ , and Cr3+ . For Mn2+ , the lifetime of the luminescence [36] is in the order of milliseconds due to the forbidden d–d transition. Similarly, f–f transitions are also often observed for rare earth elements (e.g., Tm3+ , Er3+ , Tb3+ , and Eu3+ ), although the f levels are largely unaffected by the crystal field of the host due to shielding by the outer s- and p-orbitals [2]. Due to this shielding effect, f–f transitions typically have atomic-like sharp peaks in the emission spectrum. 2.4.2

Nonradiative Relaxation Process

As reported above, absorption of energy by a luminescent material may not result in emission of light. Electrons and holes in excited states may return to lower energy and ground states by radiative and/or nonradiative relaxations. Deep level traps have a tendency to undergo nonradiative recombination by emitting phonons. Experimental data show that the time required for nonradiative recombination is short (e.g., tens of picoseconds [37]). Nonradiative relaxation may be categorized as internal conversion, external conversion, or Auger recombination [38]. Nonradiative recombination though crystalline and/or molecular vibrations is a common phenomenon in internal conversion. The difference between the energy absorbed by QDs, h𝜈, and the band gap, Eg , is generally converted into heat by electron–phonon scattering processes and results in the Stokes shift. Even for a radiative process in an indirect semiconductor, a phonon is generated due to the change of k-value (Figure 2.1(b)). Furthermore, strain in a lattice can create a local potential well that can also trap electrons and holes and result in a nonradiative transition. Nonradiative recombination at surface states dominates the external conversion category. As reported above, 15–30% of atoms in QDs are at the surface and represent defects due to unsaturated dangling bonds. These defects are dominant channels for nonradiative decay of carriers. The electronic surface states are filled below the Fermi level with electrons from the core of the QDs. Accumulation of charge at the surface creates an electric field or a depletion region that leads to bending of the valence and conduction band edges. Electron and hole carriers generated in this region are swept in opposite directions by the electric field, prohibiting radiative recombination. This leads to the concept of a “dead layer” [30]. Capping of these defects with organic ligands or inorganic shells leads to an improvement in luminescent efficiency as discussed below. Strong carrier-to-carrier interaction can lead to an Auger nonradiative process. Rather than releasing the energy of recombination as a photon or phonon, the excess energy is transferred to another electron that is called an Auger electron [Figure 2.6(c)]. The Auger electron loses its surplus energy by creation of phonons or is ejected from the material.

46

Materials for Solid State Lighting and Displays

The Auger recombination process shown in Figure 2.6(c) involves two electrons and a hole in the conduction and valence bands, respectively. Auger recombination can also create a hole deep in the valence band or can be observed for electrons and holes on localized activator levels [38]. In an Auger transition, the momentum and energy must be conserved, therefore indirect semiconductors show much higher Auger recombination rates as compared with direct band gap materials. This is due to the fact that a momentum change is necessary for Auger recombination, and is also required for the transition in indirect band gap semiconductors.

2.5

Blinking Effect

Counts/10 ms

As mentioned above, QDs have a number of advantages over organic dyes in bioapplications, for example better photostability, wide absorption edges, and narrow, tunable emission. However, they may exhibit a random, intermittent luminescence which is called “blinking”. In blinking, a QD emits lights for a time followed by a dark period, as shown in Figure 2.9. In 1996, Nirmal et al. [39] observed this switching between an emitting and a nonemitting state from a single CdSe QD at room temperature. The postulated mechanism of blinking was a photoinduced ionization process [37] which leads to charged QDs that result in a separation between electrons and holes. Based on this model, the QDs 450 (a)

80 s range

300 150 0

Counts/10 ms

Counts/10 ms

0

450

20

(b)

40

60

80

10 s range

300 150 0 22

24

300 (c)

26

28

30

32

23.8

24.0

1 s range

150 0 23.0

23.2

23.4 23.6 Time (s)

Figure 2.9 Blinking effect from a single 2.9 nm CdSe QD. (a–c) Emission is shown in three expanded time scales. Source: Kuno et al. 2000 [43]. Reproduced with permission of AIP Publishing LLC

Quantum Dots for Displays and Solid State Lighting

47

would be dark for the lifetime of the ionized state. The nonradiative Auger recombination process would be expected to dominate the quenching of ionized QDs [37]. However, the experimental results do not completely support this model. For example, a photoinduced Auger process should exhibit a quadratic dependence of the average blinking time on excitation intensity, whereas the experimental result showed a linear behavior. In addition, the bright and dark periods follow an inverse power law [40] given by: P(t) = At−m

(2.4)

where P(t) is the probability of the blinking period, m is an exponent between 1 and 2, and A is a constant. Several additional mechanisms have been proposed to explain the blinking [32, 41–43], including thermally activated ionization, electron tunneling through fluctuating barriers or into a uniform distribution of traps, or resonant electron tunneling between the excited states of QDs and dark-trap states that vary in energy. Despite tremendous efforts, the blinking effect is still not completely understood.

2.6

Surface Passivation

To use QDs, stable emission with a high quantum efficiency must be achieved. As discussed above, surface defects in QDs act as temporary “traps” for the electron, hole or excitons, quenching radiative recombination and reducing the quantum yields. Therefore, capping or passivation of the surface is crucial for development of photostable QDs. In principle, a perfectly passivated surface of a QD has all dangling bonds saturated and therefore exhibits no surface states, and all near band edge states are quantum-confined internally. For a compound semiconductor, if the anion dangling bonds at the surface are not passivated, a band of surface states is expected in the gap just above the valence band edge. However, passivation of anions with surface cations would also leave dangling bonds that would lead to a broad band of surface states just below the conduction band edge. Therefore, surface modification of QDs is very demanding and is generally carried out by depositing an organic or inorganic capping layer on the QDs. 2.6.1

Organically Capped QDs

Generally, monodispersed QDs are developed by introducing organic molecules that absorb on the QD surface and act as capping agents [6]. A schematic illustration of an organically passivated surface is presented in Figure 2.10(a). Some advantages of organic capping layers include simultaneous achievement of colloidal suspension capping, and the ability to bio-conjugate the QDs. However, the selection of organic ligands that will bond with surface atoms of the QDs is a very delicate issue. In general, phosphines [e.g., tri-n-octyl phosphine oxide (TOPO)], carboxylic acids, thiols, amines, or mercaptans are the mostly widely used ligands. Most of the organic capping molecules are distorted in shape and larger than a surface site. As a result, coverage of surface atoms with the organic capping molecules may be sterically hindered. Another crucial issue is the simultaneous passivation of both anionic and cationic

48

Materials for Solid State Lighting and Displays (a)

(b)

organic molecule

(c) band offset Eg (core)

Eg (shell)

band offset

Figure 2.10 Schematic illustration of (a) an organically capped QD and (b) an inorganically passivated QD (core/shell structure of QD). (c) An energy diagram shows the band-gap difference of core and shell of inorganically passivated QDs

surface sites using such capping agents, which is extremely difficult. Therefore, some dangling bonds on the surface are always present when the surface is passivated by organic agents. Finally, the organic capped QDs are generally photo-unstable. The bonding at the interface between the capping molecules and surface atoms is generally weak, leading to the failure of passivation and creation of new surface states under ultraviolet irradiation. The surface states of nanocrystals are known to be sites of preferential photodegradation and luminescence quenching. 2.6.2

Inorganically Passivated QDs

A second approach to passivation of the QDs surface is the use of inorganic layers, particularly a material with a larger band gap. The passivating shell is grown either epitaxially [as depicted in Figure 2.10(b)] or as a nonepitaxial crystalline or amorphous layer on the core. The quantum efficiency of QDs is increased by a defect-free, uniform shell coating. When the shell material adapts the lattice parameters of the core during epitaxial growth, coherency strains result and can play an important role in the properties of these core/shell systems. For example, strain may cause the absorption and emission spectra of core/shell QDs to be red-shifted [44]. The maximum PL efficiency of core/shell QDs is also dependent upon the thickness of the shell layer, which is less than two monolayers for optimum properties of CdSe/CdS core/shell nanoparticles. Thicker capping layers lead to formation of misfit dislocations which are also nonradiative recombination sites that decrease the PL quantum yield. Generally, a wider band gap shell material is desired to create a potential barrier around the QD core to confine the exciton [Figure 2.10(c)]. Confinement of the charge carriers into the core region by the band offset potentials result in efficient and photostable luminescence from QDs. Additional factors to consider when selecting the QD inorganic shell material include whether it is hydrophobic or hydrophilic. Most inorganic core/shell QDs are not

Quantum Dots for Displays and Solid State Lighting

49

compatible with dispersion in water due to the hydrophobic surface property of the shell. For biological application of QDs, an appropriate water-compatible coating, such as an amorphous silica layer, is necessary [7]. For best passivation, the shell material should have a lattice parameter within 12% of the core to encourage epitaxy and minimize strain, and a thickness below the critical value that results in misfit dislocations [45]. Inverted core/shell QDs, for example ZnSe/CdSe with a larger band gap for the core, show very interesting optoelectronic properties. They exhibit either type I or type II interfacial band offsets depending on the core radius and the shell thickness [46]. Type I offset is shown in Figure 2.10(c), that is there is an opposite offset for both the valence and conduction bands. This is the case for bulk ZnSe/CdSe interfaces, where the ZnSe valence band edge is lower than that in CdSe (energy offset ∼0.14eV), while the conduction band edge is higher (energy offset ∼0.86eV). Such an energy alignment results in confinement of both electrons and holes inside the CdSe core which reduces their interactions with surface trap states and improves their quantum yields. However, the situation can change in the case of nanostructures in which the alignment of quantized energy states is determined not only by bulk energy offsets, but also by the confinement energies determined by the heterostructure dimensions. Core/shell QDs with type II offsets (valence and conduction band offsets in the same direction) can also provide “spatially indirect” states, in which electrons are spatially confined to the core (or shell) and holes confined to the shell (or core). The emission energy from type II core/shell nanostructures is smaller than the band gap of either the core or the shell material due to the interfacial energy offsets. Because of the reduced electron–hole wave function overlap, these structures show extended exciton lifetimes and are useful in photovoltaic and photocatalysis applications [47]. With a large red-shift in emission from type II core/shell QDs, near-infrared emission may be possible for in vivo bioanalytical and biomedical applications.

2.7

Synthesis Processes

Several routes have been used to synthesize phosphor QDs. Generally, the techniques for synthesis of QDs are categorized either as a top-down or bottom-up approach. 2.7.1

Top-Down Synthesis

In the top-down approaches, a bulk semiconductor is thinned to form the QDs. Electron beam lithography, reactive-ion etching, and/or wet chemical etching are commonly used to achieve QDs of ∼30 nm diameter. Controlled shapes and sizes with the desired packing geometries are achievable for systematic experiments on quantum confinement effects. Alternatively, focused ion or laser beams have also been used to fabricate arrays of zero-dimension dots. Major drawbacks with these processes include incorporation of impurities into the QDs and structural imperfections by patterning. Etching, known for more than 30 years, plays a very important role in these nanofabrication processes. In dry etching, a reactive gas species is inserted into an etching chamber and a radio frequency voltage is applied to create a plasma which breaks down the gas molecules to more reactive fragments. These high kinetic energy species strike the surface and form a volatile reaction product to etch a patterned sample. When the energetic species

50

Materials for Solid State Lighting and Displays

are ions, this etching process is called reactive ion etching (RIE). With a masking pattern, selective etching of the substrate is achieved. Fabrication of GaAs/AlGaAs quantum structures as small as 40 nm has been reported using RIE with a mixture of boron trichloride and argon [48]. This RIE process has been used to produce close-packed arrays for testing of lasing in QD semiconductors. Close packed arrays of ZnTe with interdot distance of 180–360 nm were produced by RIE using CH4 and H2 [49]. Focused ion beam (FIB) techniques also offer the possibility of etching QDs with extremely high lateral precision. Highly focused beams from a molten metal source (e.g., Ga, Au/Si, Au/Si/Be, or Pd/As/B) may be used directly to sputter the surface of the semiconductor substrate. The shape, size and inter-particle distance of the QDs depend on the size of the ion beam but a minimum beam diameter of 8–20 nm has been reported for both lab and commercial systems, allowing etching of QDs to dimensions of 𝜎 1 [69]. In other cases, formation of QDs was due to relaxation of strain required to maintain epitaxy. In the case of substrates with an overlayer with a large lattice mismatch and appropriately small surface and interface energies, initial growth of the overlayer occurs through a layer by layer FvdM growth. However, when the film is sufficiently thick (a few monolayers) to induce a large stain energy, the system lowers its total free energy by breaking the film into isolated islands or QDs (i.e., the VW). MBE has been used to deposit the overlayers and grow elemental, compound or alloy semiconductor nanostructured materials on a heated substrate under ultrahigh vacuum (∼10−10 torr) conditions [70, 71]. The basic principle of the MBE process is evaporation from an apertured source (Knudsen effusion source) to form a beam of atoms or molecules. The beams in the MBE process can be formed from solids (e.g., elemental gallium and arsenic are used to produce GaAs QDs) or a combination of solid plus gases (e.g., AsH3 , PH3 , or metal-organics such as trimethylgallium or triethylgallium). The metal-organic sources may leave high concentrations of carbon in the QDs. MBE has been mainly used to self-assemble QDs from III–V semiconductors using the large lattice mismatch, for example InAs on GaAs has a 7% mismatch and leads to SK growth as discussed above. Layer growth by physical vapor deposition (PVD) results from condensation of solid from vapors produced by thermal evaporation or by sputtering [51]. Different techniques have been used to cause evaporation, such as electron beam heating, resistive or Joule heating, arc-discharge and pulsed laser ablation. In any case, the factors discussed above (strain and surface energies) control the formation of QDs from the deposited thin films. Chemical vapor deposition (CVD) is another method to form thin films from which QDs can be self-assembled. In CVD, precursors are introduced in a chamber at a particular pressure and temperature and they diffuse to the heated substrate, react to form a film, followed by gas-phase byproducts desorbing from the substrate and being removed from the chamber. InGaAs and AlInAs QDs have been synthesized using either surface energy or strained-induced SK growth processes [72].

2.8

Optical Properties and Applications

2.8.1

Displays

2.8.1.1 2.8.1.1.1

Photoluminescent Applications of QDs Market and Product Overview

On account of the ability of QDs to achieve high photoluminescence quantum yield (PL QY) and narrow emission linewidth, they are becoming an increasingly important component in the display industry, especially in liquid crystal displays with LED backlights (LED-LCDs). This is particularly apparent in large size, high-end televisions, where wide color gamut and features such as high dynamic range (HDR) are of great importance. One

54

Materials for Solid State Lighting and Displays

display standard that contains an ultra-wide color gamut (in addition to color bit depth, chroma subsampling, display resolution, and frame rate) specifically for use in ultra-high definition (UHD) displays was developed by the International Telecommunication Union (ITU) and is called ITU-R Recommendation BT.2020 (colloquially, Rec. 2020) [73]. The color space within the Rec. 2020 specifications is the largest color space for broadcast content yet to be introduced. Additionally, computer monitors used for professional video and photography editing are incorporating QDs into their systems to meet these demands and because of the advantages described in this section. In practice, QDs are deployed in a component in the display system such as LCDs, and some parts of the LCD system are modified to optimize the use of the QDs. Most importantly, the backlight unit (BLU) composed of a white LED backlight – for example, a blue InGaN LED with a yellow Y3 Al5 O12 :Ce3+ (YAG:Ce) downconversion phosphor – is replaced with a plain blue InGaN LED BLU. These blue LEDs are then used as the optical pump for the QDs as photoluminescent phosphors. This is to ensure the red and green emission only originate from the narrow-band emitting QDs rather than a wide-band emitting rare-earth-doped phosphor such as YAG:Ce. Currently, Cd-based QDs such as CdSe/CdS, CdSe/ZnCdSe/ZnS, or CdSe/ZnCdS are the preferred materials for their narrow emission spectra and high quantum yield. However, there are regulatory issues that complicate using Cd in QDs for display applications. The European Union’s Restriction of Hazardous Substances (RoHS) Directive contains restrictions for heavy metals, including Cd, Pb, Hg, hexavalent Cr, and other substances. RoHS 2 (Directive 2011/65/EU) [74], is an updated version of the RoHS Directive and contains Annex III, Exemption 39, which grants an exemption to the RoHS 2 restrictions on Cd in “light control materials used for display devices”, and was adopted in 2011. This exemption extends the right to use Cd in QDs for displays until June 30, 2018. However, it is possible that in the future this exemption may be extended. Due to the uncertain future of the regulatory environment around Cd-containing QDs, some companies are moving toward including InP-based QDs such as InP/ZnSe or InP/ZnSeS/ZnS into their products. It should be noted that both Cd and InP are toxic to humans, but Cd must conform to a highly restrictive 95% [75], while state of the art InP-based QDs exhibit FWHM > 40 nm and PL QY ∼ 70–90% [76]. There are ongoing efforts to improve InP-based QDs enough to reach the performance specifications of Cd-based QDs, particularly in the competitive high-end display market. 2.8.1.1.2

Performance and Energy Advantages

Changing the backlight from white to blue and replacing traditional phosphors with QDs imparts multiple advantages. When the concentrations and color quality of the QDs (usually red- and green-emitting) are well controlled, narrow-band emission from the InGaN blue LED is combined with narrow-band emission from the red QDs and green QDs to

Quantum Dots for Displays and Solid State Lighting

55

give extremely pure colors and thus a wider color gamut than that of a broadband white LED being filtered through an equivalent LCD color filter array. Moreover, the narrower the QD emission spectrum, the wider the color gamut of the display. This can clearly be seen in Figure 2.11, where the simulated color gamut of a display with a variety of emission linewidths and without color filters is plotted on the CIE 1976 color space diagram. Additionally, a comparison of emission spectra of a LCD module with standard color filters with a blue BLU + QDs and a white LED (blue + YAG:Ce) is shown in Figure 2.12. It is clear that the emitted colors in a display are purer with a blue LED + QD configuration.

0.6

520

530

540

550

560 570

580

590

510

600

610

620

630 640

680

500

0.5

490

0.4

0.3 v' 480

0.2 R,G,B Primaries [467 nm, 532 nm, 630 nm] Full Widh at Half Maximum (FWHM) 470

0 nm (Rec. 2020) 20 nm 30 nm 40 nm 50 nm 60 nm

0.1 460 450

0

0.1

440 430 420

0.2

0.3

0.4

0.5

0.6

u'

Figure 2.11 Simulated color gamut coverage as full width at half maximum (FWHM) of QDs narrows from 60 to 20 nm. Monochromatic primaries (FWHM = 0 nm) representing the Rec. 2020 specifications are plotted as well. The peak wavelength of the simulated emission is the same as those of the Rec. 2020 primaries: 467 nm, 532 nm, and 630 nm. The CIE color space coverage increases with decreasing QD FWHM. The simulations assume a Gaussian emission profile from the QDs

Materials for Solid State Lighting and Displays

W-LED Backlight

Power Density

56

560

570

580

590

510

600

610

620 630 640

500

400

450

500

550

600

650

650

0.5

700

0.4

Transmission Coeffecient

Color Filter

520 530 540 550

0.6

490

0.3 480

v'

400

450

500

550

600

650

700

0.2

Power Density

470

LCM

0.1 460

500

600 550 Wavelength (nm)

650

440 430 420

0.1

0.2

0.3

0.4

0.5

0.6

u'

700

CIE 1976

0.6

520 530 540 550

560

570

580

590

510

600

610

620 630 640

500

400

450

500

550

600

650

650

0.5

700

0.4 490

Transmission Coeffecient

Color Filter

450

Power Density

400

Nanosys QDEF Backlight

W-LED Backlight

450

0

0.3 480

v' 400

450

500

550

600

650

700

0.2

Power Density

470

LCM

400

0.1 Nanosys QD Backlight W-LED Backlight

460

0 450

500

550 600 Wavelength (nm)

650

700

0.1

450 440 430 420

0.2

0.3

0.4

0.5

0.6

u' CIE 1976

Figure 2.12 Comparison of color emission spectra, color filter effect, final LCD emission spectra in the liquid crystal module (LCM), and color gamuts with a white LED BLU (blue LED + YAG:Ce) and blue BLU (blue LED + QD film). By replacing the white BLU with a blue LED + QD combination, the color gamut is greatly enhanced. Image courtesy of Nanosys, Inc.

In addition to the visual advantages of QDs in displays, there are three main energy efficiency gains when using QDs instead of traditional rare-earth phosphors. First, a large portion of the Stokes shift in traditional white LEDs goes to creating unwanted yellow-orange light that will be filtered out and lost, creating a double waste of energy. With QDs, all of the Stokes shift loss goes toward producing the desired red and green primaries. Compared with YAG:Ce, there is less Stokes loss to emit green from a QD than the strong yellow emission of YAG:Ce due to the wavelength, and thus the energy and photopic response difference. Additionally, the red QD primary is usually tuned to ∼630–640 nm and has a narrow FWHM of 95% Red QD Gen-3 ≥95% Red Phosphor-KSF Red Phosphor-SrLiAl3N4:Eu Red Phosphor-SCASN:Eu Red Phosphor-CaS:Eu

110% 105% 100% 95% 90% 85% 80% SS EQE at RT 75% 70% 65% 60% 25

Green QD Gen-1 ≥95% Green QD Gen-2 >95% Green QD Gen-3 >95% Green Phosphor-β-SiAION:Eu Green Phosphor-SrGaS:Eu 50

75

100

125

150

175

Temperature (ºC)

Figure 2.13 The PL intensity versus temperature relative to that at 25 ∘ C for three different generations of red (top) and green (bottom) emitting QDs from QD Vision and for narrow-band red (top) and green (bottom) emitting phosphors based on data found in the technical literature (K2 SiF6 :Mn4+ or KSF; SrLiAl3 N4 :Eu; Srx Ca1−x AlSiN3 :Eu or SCASN:Eu; CaS:Eu; 𝛽-SiAlON:Eu; SrGa2 S4 :Eu or SrGaS:Eu – for further details, see reference [78]). The solid state photoluminescence quantum yield at 25 ∘ C (SS EQE at RT) is shown for the QDs. Approximate temperature ranges that downconversion materials experience in film, edge optic, and on-chip configurations are separated with vertical broken lines. Order-of-magnitude approximation of blue LED BLU optical excitation intensities are also shown. Source: Steckel et al. 2015 [78]. Reproduced with permission of John Wiley and Sons

Quantum Dots for Displays and Solid State Lighting

Film

Edge Optic

59

On-chip

QD Optic QD Film

LGP

LED

LED

LGP

QD Film

LGP

LGP

QD LED

Panel LGP

QD Optic LGP LED

LED

LED

Source: Nanosys Online Media Kit Source: LumenMax Optoelectronics

Figure 2.14 Three- and two-dimensional representations of the three form factors for integrating QDs into displays from left to right: edge optic, film, and on-chip. The bottom row of images shows the individual components separated from the modules. Source: Steckel et al. 2015 [78]. Reproduced with permission of John Wiley and Sons

QDEF® QD Film in LCD

Standard LCD

LC Glass Enhancement Films QDEF® QD Film Diffuser LGP Reflector

Figure 2.15 Exploded view (simplified) of an LCD module showing the replacement of the diffuser sheet with the QD enhancement film. Image courtesy of Nanosys, Inc.

60

Materials for Solid State Lighting and Displays

with less LEDs and less electronic and software controls for driving them. A disadvantage is that in edge-lit displays, a feature called local dimming is practically impossible or at least not easily accomplished with high quality. Local dimming is critical for HDR imaging wherein certain areas, or zones, of the direct-lit backlight arrays are dimmed or turned off to create deeper or true blacks when the image is meant to be dark or black (e.g., in outer space, shadows, or other dark scenes). A graphic showing a direct-lit LCD module with a QD film is shown in Figure 2.16. This is in contrast to edge-lit systems wherein the BLU is always on across the area of the display, which leads to some light leakage through the liquid crystals and color filters, creating a dark gray where there should be black. This gray light leakage reduces the contrast ratio of the display, which is one of the critical factors in high quality products. By increasing the number of zones into which the display backlight is divided, finer control of local dimming and HDR can be achieved. The other critical component of HDR is being able to show extremely bright colors in the same scene as, and very close to, the blackest black levels. For this, the backlight needs to be driven at high current, emitting high intensity light in very specific areas of the display. Thus, the QD material and packaging must be able to withstand local and short-time exposures to high flux. However, on average, film-type components experience the lowest flux and temperature exposures

QDEF

Quantum Dots

Glass

Blue LEDs

Local Dimming Backlight

Figure 2.16 View of a typical LCD module where the BLU is a direct-lighting system containing an array of blue LEDs illuminating a QD-containing film. This direct-lit BLU enables HDR content display. Image courtesy of Nanosys, Inc.

Quantum Dots for Displays and Solid State Lighting

61

of all three architectures (∼0.1–1 W/cm2 and 25–75 ∘ C) because they are placed far away from the heat-generating backlights in the module stack. The benefit of this is that the packaging for films does not have to be hermetic, rather only provide enough protection to meet the display lifetime specifications. The disadvantage of film components is the cost. The barrier films are expensive to produce and the film coating lines currently require a nitrogen environment to avoid exposing the QDs and polymer resins to oxygen and moisture. Additionally, since the area of the QD film covers the entire display, more QDs and polymer resin are needed than in the edge optic configuration, further increasing the cost of materials. However, there are upcoming advances in QD formulations and coating manufacturing that will ease the stringent requirements on barrier film quality and manufacturing environment. It is anticipated that the water vapor transmission rate (WVTR) requirement will be reduced from 10−4 to 10−1 g/m2 /day in the next few generations of QD film development. These improvements in QD formulations will further drive down film component production costs and enable integration into a wider array of display products. Edge Optic In an edge optic component, the QDs are dispersed in an encapsulating resin and then loaded into a glass or transparent plastic tube and cured. Then, the tube component is inserted in front of the edge-lighting LED BLU. The rest of the LCD module remains mostly the same with the exception of fine-tuning color filters and the exact wavelength of the BLU LEDs to further improve the color gamut. Because the edge optic is placed directly in front of the LED BLU, the QDs experience a moderately high blue light flux, ∼1–10 W/cm2 and operating temperatures of 75–125 ∘ C. Of the three QD architectures, the edge optic experiences a medium amount of flux and heat exposure. A great advantage of this configuration is that the optic only has to be as large as the BLU, which is small and uses a small quantity of QDs. Thus, the cost of the edge optic is also lower. Comparatively, the edge optic should have a medium total QD cost, between the film and the on-chip. Additionally, it is relatively straightforward to create longer edge optic tubes or “drop in” more for longer edge-lighting BLUs in large size displays. Thus, edge optics are promising for scalable television sizes. Two major drawbacks of the edge optic are the blue flux and the incompatibility of edge optics with direct-lit displays that are necessary for emerging high quality HDR televisions with local dimming functionality. The moderate blue flux necessitates quite stable materials and packaging, while the position of the edge optic on the sides of the display limit its applicability across the display market. Despite these drawbacks, there is a market for lower cost, large size displays and even small and moderate size monitors that the edge optic fulfills. On-Chip The on-chip component is composed of a standard blue InGaN LED with the rare-earth-doped downconversion phosphor replaced by QDs. The QDs are loaded into the encapsulation resin or coated on the inside of the packaging, serving as both an encapsulation and a color conversion package. This small form factor for QD delivery would be advantageous for the lowest cost and most versatile implementation of QDs in displays. However, because of the requirement of the QDs to withstand extremely high flux and temperatures (>10 W/cm2 and >125 ∘ C) during operation, this technique has not been successful in the display marketplace. However, a number of research companies are making excellent progress on improving the reliability.

62

Materials for Solid State Lighting and Displays

2.8.1.1.4

Photoluminescent QD Products in the Display Market

There are many companies and countless academic and research institute groups developing QDs for a variety of applications. However, only Samsung, QD Vision and Nanosys have successfully mass produced QD materials that have been included in an edge optic or film product for sale on the consumer display market. The on-chip QD solutions are still being pursued by many companies, notably Pacific Light Technologies and others in the traditional solid state lighting space and may serve as a low-cost alternative to edge optics and films in the future. However, high light flux, high operating temperature, and LED packaging issues need to be solved before the on-chip QD solution becomes a viable market alternative. Even so, Pacific Light Technologies showed promising operational stability for red QDs in solid state lighting tests at the Quantum Dot Forum 2016. The first consumer display product in mass production to use QDs as downconversion phosphors was the Sony Triluminos television in the spring of 2013, which incorporated the Color IQTM edge optic component with Cd-based QDs from QD Vision. Since then, other producers have included QD Vision’s Color IQ component in their displays. Nanosys produces both Cd-based and InP-based QDs and film technology and partners with 3M for the development, production and distribution of a film component called Quantum Dot Enhancement Film (QDEF® ). The first non-television product to contain QDs was the Amazon Kindle Fire HDX, which incorporated the Nanosys/3M QDEF® film in the fall of 2013. The first computer monitor to incorporate QDs was the Dolby Maui PRM in 2014 with a Nanosys/3M QDEF® film. This monitor is used as a professional-grade reference monitor in high-end video and photo editing, such as wide color gamut movies. The first consumer grade computer monitor to incorporate QDs was the Philips 276E6 with an edge optic from QD Vision in 2016. In the spring of 2016, QD Vision concentrated its sales on its Color IQ edge optic with Cd-based QDs, while both Samsung and Nanosys sell their main product as a film. Currently, Samsung licenses Nanosys technology and produces InP-based QD material and films in-house and for its SUHD line of televisions. Currently, 3M is the main producer of QDEF® (using Nanosys QDs) sold in televisions and monitors on the market. Nanosys also previously engineered an edge optic component called QuantumRailTM that won an Innovation Award at CES 2011 with LG Innotek Co., Ltd. A list of companies with television display products or demos incorporating QDs as of spring 2016 is shown in Table 2.2. 2.8.1.2

Electroluminescent Display Applications of Quantum Dots

2.8.1.2.1 Advantages and Operating Principles of Emissive Displays (and Quantum Dot Light Emitting Diodes) Display Performance Considerations While deploying narrowband-emitting QDs in LCDs to enhance color is a major advancement and is sure to extend the LCD product lifetime, other display technologies with distinct advantages over LCDs are progressing quickly. Even with QDs to enhance color and local direct-backlighting to enable vivid HDR imagery, the problems with LCDs include narrow viewing angles, high power consumption, low contrast ratio in non-HDR modules, relatively slow switching speeds of ∼240 Hz maximum (∼4 ms), and only thick, non-flexible form factors. Some new LCD technology has been developed to help overcome these problems, such as in-plane

Quantum Dots for Displays and Solid State Lighting

63

Table 2.2 List of the television products from around the world incorporating QDs as of March 2016. This table shows the composition and form factor for each brand as well as if a product is available or a demonstration unit has been shown Quantum dot television display products - March 2016 Brand Samsung LG TCL Hisense Skyworth Sharp Changhong Konka Philips Vizio

Wide color gamut Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

Edge/Film

Cd/In

Product/Demo

Film Film Edge or film Edge or film Film Film Film Edge Edge Film

Indium Indium Cadmium Cadmium Cadmium Cadmium Cadmium Cadmium Cadmium Cadmium

Products Demo Products Products Demo Product announced Demo Demo Products Products

Source: Adapted from a lecture entitled “Quantum Dots – State of the Union” given by Dr Seth Coe-Sullivan, co-founder and Chief Technology Officer Emeritus of QD Vision, during QD Forum 2016.

switching (IPS) to increase viewing angle. Fine-tuning the dynamic contrast ratio in HDR content can be accomplished through adding more local dimming zones, but this increases the cost of the display due to improving the operating software, circuit drivers and adding more LED chips. This technology is only available in the highest-end displays at this time. To overcome these drawbacks, emissive EL technologies will be used. An organic light emitting diode (OLED) is one such emissive EL technology that does not suffer from these drawbacks. OLEDs are very thin (∼100 nm), making them flexible, can be switched on and off at the subpixel level for HDR applications, and are inherently nearly Lambertian emitters. OLEDs have quickly grown into the dominant display technology for high-end mobile phones and are beginning to gain market share in the high-end television market due to the vivid colors, perfect blacks for an infinite contrast ratio (individual pixels are switched on and off for black or “on” state), wide viewing angles, fast switching speeds and availability of flexible, foldable, and rollable form factors in the future. While OLED displays show vivid colors, this generally comes at a cost – either power loss or manufacturing and performance, as briefly discussed here. The native emission spectra of the organics in OLEDs is generally characterized by a doublet with a broad long wavelength tail extending more than 100 nm [79], and this broad emission peak limits the available color gamut. A wide color gamut can be achieved by narrowing the emission spectra with a microcavity [80] or color filters. However, OLED displays using side-by-side R-G-B pixels with microcavities suffer from narrow viewing angles and increased production costs and complexity, while OLED displays that use white OLEDs with color filters suffer from power loss by filtering. White OLEDs also use a fourth, white subpixel in addition to the R-G-B subpixels to increase the display luminance, but the unsaturated white color greatly diminishes the possible color gamut that makes OLEDs attractive over LCDs. Quantum Dot Light Emitting Diodes The technology that has the most promise to overcome the drawbacks of LCDs and OLEDs is the quantum dot light emitting diode (QLED).

64

Materials for Solid State Lighting and Displays

The fundamental mechanism, EL, is the same as in OLEDs except the organic emitting layer is replaced by QDs and the device structure is slightly modified accordingly. Electroluminescent devices such as OLEDs and QLEDs have an external quantum efficiency (EQE) of: (2.6) EQE = 𝜒 ∗ 𝜂PL ∗ 𝜂inj ∗ 𝜂out where 𝜒 is the fraction of spin-allowed transitions, 𝜂 PL is the emitter’s PL efficiency, 𝜂 inj is the charge injection efficiency (also charge balance), and 𝜂 out is the outcoupling efficiency. In both planar OLEDs and QLEDs, 𝜂 out is between 20% and 30%. Therefore, the maximum EQE without any light outcoupling techniques is 20–30%. OLEDs must use expensive triplet emitters doped into a matrix to achieve 𝜒 = 100% and EQE ≥ 20%, but QDs exhibit singlet–triplet thermal mixing [81] and so already achieve 𝜒 = 100%. By incorporating QDs into electroluminescent devices, one takes advantage of the narrow emission spectra of the QDs such that no color filtering or microcavities are needed to create a wide color gamut. Additionally, upon patterning of QLEDs into R, G or B subpixels, each can be switched off and on individually to reduce power consumption and create an infinite contrast ratio. QLEDs, like OLEDs, are nearly Lambertian emitters, so the viewing angle of displays with QLEDs will be wide. Since QDs are generally solution-processed, advanced printing manufacturing techniques can be used to fabricate these displays, reducing material waste and improving yield and throughput. Additionally, the promise of longer lifetimes should be realized by using inorganic materials rather than all organic materials, which have many intrinsic degradation pathways. Device Structures QLEDs have been fabricated using two different device structures in the bottom-emitting configuration. Namely, the “regular” and the “inverted” structure. Figure 2.17 compares the two structures. In the regular structure, an ITO anode-coated glass substrate is coated with a hole injection layer (HIL), hole transport layer (HTL), the QD emitting layer (EML), followed by the electron transport layer (ETL), and a reflective cathode. In the inverted structure, the polarity of the contacts is reversed and the transport layers are switched such that the ITO functions as the electron injecting cathode, followed by an ETL, QD EML, HTL, HIL, and reflective metal anode, such as Ag, Au, or Al. The transport materials can be organic or inorganic. When the device is composed of some organic and some inorganic transport layers, it is referred to as a hybrid device structure. The highest performing QLED devices reported to date have used a hybrid structure including organic hole injection and transport layers combined with an inorganic ZnO ETL in both the inverted and regular structure. The reasons for this being the highest performance configuration are unclear. Each structure can be used to create monochrome, multicolor, or white QLEDs as desired. Of particular interest to the QLED manufacturing community is creating a white QLED by using a QD EML that contains red-, green-, and blue-emitting QDs mixed into the same layer. This would simplify the patterning of the QLEDs – still seen as a major hurdle to mass production – but would introduce major challenges. The elimination of color filters is one of the motivations for using QLEDs rather than other EL technologies. However, with a white QLED, color filters would be needed. Additionally, charge transport, exciton formation, and radiative recombination in the EML would be extremely complicated by different size and composition QDs with overlapping absorption and emission spectra as

Quantum Dots for Displays and Solid State Lighting

REFLECTIVE CATHODE

REFLECTIVE ANODE

ELECTRON TRANSPORT LAYER

HOLE INJECTION LAYER

QD EMITTING LAYER HOLE TRANSPORT LAYER HOLW INJECTION LAYER ITO ANODE

TRANSPARENT SUBSTRATE

REGULAR

65

HOLE TRANSPORT LAYER QD EMITTING LAYER ELECTRON TRANSPORT LAYER ITO CATHODE

TRANSPARENT SUBSTRATE

INVERTED

Figure 2.17 The two bottom-emission QLED device structures. The left stack is the “regular” structure and the right is the “inverted” structure

well as offset energy levels existing in the same film. Likely, this opens many more Förster resonance energy transfer (FRET) pathways for quenching and charge transport barriers between QDs of different band gaps. One interesting feature of hybrid structure QLEDs is a sub-band gap turn-on of the device. That is, the device exhibits luminance turn-on at a voltage below the energy of the photon being emitted. One proposed mechanism to explain this observation is an Auger upconversion process that relies on charge buildup of one carrier at a charge injection barrier interface in the device [82]. Upon charge buildup and injection of the opposite charge, an exciplex state is formed which nonradiatively recombines and transfers energy to a third carrier, giving it enough energy to hop over the injection barrier at that interface. The extent to which sub-band gap emission is achieved can be controlled by the size of the injection barrier. This sub-band gap turn-on is also present in OLEDs composed of a rubrene/C60 interface that also has a charge injection barrier [83]. The explanation of the rubrene/C60 case is triplet–triplet annihilation, wherein two triplet excitons form directly and annihilate each other to make a high energy singlet exciton with enough energy to recombine radiatively near the band edge. Yet another plausible explanation is the existence of a locally high electric field caused by interfacial charge transfer that reduces the injection barrier only in the in-situ device stack. This sub-band gap turn-on is not present in high efficiency phosphorescent OLEDs used in displays or solid state lighting. However, it has been reported in numerous high efficiency QLEDs [84–86]. This sub-band gap turn-on allows for a lower drive voltage to be used, improving the power efficiency. This improvement in power efficiency is critical for acceptance as a display technology, particularly in the mobile market, where battery life is a key parameter.

66

Materials for Solid State Lighting and Displays

2.8.1.2.2

History of QLEDs

The first QLEDs were developed in the Alivisatos group in 1994 [87]. These devices were composed of a simple bilayer structure containing an ITO anode, plain CdSe QDs, poly(p-phenylene vinylene) (PPV) polymer, and Mg cathode. These devices emitted up to 100 cd/m2 and operated up to 50 mA/cm2 with an impressive operating voltage of only 4 V. At low voltage, emission from the QDs was dominant, while at high voltage, the PPV polymer emitted light strongly due to charge carrier overflow into the polymer layer. The devices showed an EQE of only 0.001–0.01%. Given that these were the first QLEDs ever reported, they had minimal device structure or synthesis optimization specifically tailored for QLEDs, and at the time, it was near the early stages of all solution processed thin film EL devices such as polymer light emitting diodes (PLEDs). Because of these facts, the efficiency levels showed promise for the future. Since the first report, QLEDs have increased in efficiency to levels comparable with OLEDs used in commercially available displays. This is because of improvement in both the synthesis, understanding of photophysics, and device structures used for these devices. The QDs used for the first QLEDs were simple CdSe with organic ligands, while QDs in modern, high-performance QLEDs are typically of the core/shell variety with some composition gradient from core to shell and even with exotic capping ligands or composite emitting layers [88]. The device structures can be considered to be in four different categories as described by Supran et al. [88]: 1. 2. 3. 4.

Type I: QLEDs with polymer charge transport layers. Type II: QLEDs with organic small molecule charge transport layers. Type III: QLEDs with inorganic charge transport layers. Type IV: QLEDs with both organic and inorganic charge transport layers.

The first QLED by Alivisatos et al. falls under Type I. The first Type II QLED was introduced in 2002, and was composed of a bilayer OLED structure with QDs at the p-n junction interface [89]. By spincasting QDs mixed in solution with the hole transporting small molecule N,N′ -diphenyl-N,N′ -bis(3-methylphenyl)-(1,1′ -biphenyl)-4,4′ -diamine (TPD), the QDs with their aliphatic ligands phase-separated from the TPD and formed a monolayer. The organic hole blocking layer and ETL were subsequently evaporated on top of the QD EML. This monolayer of QDs served to separate the responsibility of conduction from light emission in the QD layers that plagued earlier reports. By achieving this structure, QLEDs with a then-record 0.5% EQE were realized. There were drawbacks to this technique, however. There was still emission from the organic transport layers in contact with the EML because of current overflow, incomplete FRET from the organics to QDs, and/or possible pinholes in the QD monolayer. Several reports on Type III QLEDs were initially promising [90–92] because inorganic transport layers should be able to transport higher current densities and survive longer in the ambient environment than organics. However, the highest efficiency devices have all come from the Type IV class. With the growth of OLED development, QLED researchers have adopted many different OLED materials into their device stacks in Type IV QLEDs, particularly as hole injection layers (HILs) and hole transport layers (HTLs). The most common high efficiency device design using solution-processed HIL and HTL are PEDOT:PSS

Quantum Dots for Displays and Solid State Lighting

67

[poly(3,4-ethylenedioxythiophene) polystyrene sulfonate], TPD (or poly-TPD) (poly[N, N′ -bis(4-butylphenyl)-N,N′ -bis(phenyl)-benzidine]), TFB (poly[(9,9-dioctylfluorenyl-2,7diyl)-co-(4,4′ -(N-(p-butylphenyl))diphenylamine)]), and PVK [poly(9-vinylcarbazole)] [85, 89, 93]. When the HTL is evaporated, the most common materials are some variant of NPB [N,N′ -di(1-naphthyl)-N,N′ -diphenyl-(1,1′ -biphenyl)-4,4′ -diamine], CBP [4,4′ -bis(N-carbazolyl)-1,1′ -biphenyl], and TCTA [tris(4-carbazoyl-9-ylphenyl)amine] [86, 94, 95]. The overwhelming majority of devices with high performance use colloidal ZnO nanoparticles as the electron injecting and transporting layer (EIL or ETL), but the reasons that ZnO in particular creates such high performance devices are still unclear. 2.8.1.2.3

Modern Advancements in QLEDs

High Performance QLEDs Cadmium-containing QDs in QLEDs In recent years, the EQE of QLEDs has improved to be comparable with, or better than, OLEDs used in commercially available displays. In these cases, CdSe-based QDs have been used. InP- and ZnSe-based QDs have not been developed to the same extent as CdSe. There is a lot of chemistry and physics still to be discovered in these materials systems for EL. However, with the uncertainty in continuation of the RoHS exemptions described above, plus greater environmentally conscious display manufacturing, InP- and ZnSe-based QDs are being studied more thoroughly. In 2013, QD Vision reported a then-record high orange-red QLED EQE of 18% by using a CdSe/CdS-based core/shell QD in an inverted hybrid organic–inorganic device [86]. This inverted device structure was reported to be ITO/ZnO nanoparticles/QD EML/ spiro-2NPB/LG-101/Al, wherein the ZnO nanoparticles and QD EML were solution processed before loading the devices into the thermal evaporator to deposit the spiro-2NPB HTL, LG-101 HIL, and Al top anode. By thermally evaporating the transport layers on top of the QD layer, solvent damage to the QD EML is avoided and neat layers of organic molecules can be deposited with fine thickness control. The disadvantage to this approach is that it is no longer completely solution-processed, thus reducing the benefits of switching to QLEDs. The devices showed sub-band gap turn-on (1.5 V) and a high peak power efficiency of 25 lm/W with a peak wavelength ∼610 nm (∼2 eV photons). In one of the few papers to report operating lifetimes in the literature, the device lifetimes were shown to have a strong dependence on the EML thickness and correlate well with film PL yield of the QDs. The film PL yield of the dots is mediated by a short range, ultrafast QD/ZnO interfacial charge transfer phenomenon. Devices with thicker QD layers (45 and 60 nm) operated at higher efficiency, but had shorter operational lifetimes than those with thin QD layers (15 and 30 nm). As more photophysics of the QDs in their native QLED structure is investigated, degradation pathways should become more clear, leading to improvements in materials and device structures for longer lifetimes. In 2014, the Peng group reported the highest red QLED efficiency of 20.5% EQE [84]. The narrow emission peak with peak wavelength of 640 nm and FWHM of 28 nm produced very desirable CIE (x, y) coordinates of (0.71, 0.29), almost exactly matching the Rec. 2020 red primary requirement. The QDs used were CdSe/CdS-based core/ shell QDs in an all solution-processed regular hybrid device structure of ITO/PEDOT:PSS/ Poly-TPD/PVK/QD EML/PMMA/ZnO NPs/Ag. The peculiar addition of PMMA served

68

Materials for Solid State Lighting and Displays

as the key layer to obtaining the high EQE. It was shown that the electron current levels were more than an order higher than the hole current levels. This is likely caused by two QD material properties: the CdSe/CdS is a quasi-type II core/shell structure, meaning that the electron wavefunction extends from the CdSe core into the CdS shell while the hole wavefunction remains confined to the core, facilitating electron transfer between QDs. Additionally, the CdSe/CdS conduction band energy levels are well aligned to the ZnO ETL conduction band levels, creating a very small electron injection barrier, while the injection barrier from PVK into the CdSe/CdS remains hindered by a barrier. Because of this, the PMMA insulating layer was added between the ETL and QDs to inhibit electron injection and obtain charge balance in the EML. By depositing a very thin (∼6 nm) layer of PMMA, the turn-on voltage was kept at a low sub-band gap 1.7 V. It is also speculated that the PMMA layer reduces parasitic charge transfer between the QD layer and ZnO. The lifetime of the device was impressive, with a T95 lifetime at 10 600 cd/m2 of approximately 5 h. This translates to a T95 lifetime at 1000 cd/m2 of approximately 200 h, assuming an acceleration factor, n, of 1.5. The reported T50 lifetimes are also impressive, with T50 at 1000 cd/m2 = 3500 h, and T50 at 100 cd/m2 = ∼100 000 h [84]. In 2015, NanoPhotonica reported the first demonstration of all three primary colors (R, G, and B) having EQE over 10% [85]. These devices used the regular structure and were all solution-processed. By exploiting a one-pot QD synthesis with a smooth composition gradient from core to outer shell, the confinement potentials and outer shell thicknesses were optimized to enhance charge injection, reduce parasitic charge transfer and Auger recombination. For red and green, a Cd- and Se-rich core with a Zn- and Se-rich intermediate shell capped with a ZnS outer shell showed narrower FWHM and emission yields than that of QDs with a Cd- and S-intermediate shell. By fine-tuning the composition of these QDs, red QLEDs were fabricated with 12% EQE at a peak wavelength of 625 nm and FWHM of 25 nm. Green QLEDs following the same QD synthetic philosophy were fabricated exhibiting 14.5% EQE at a peak wavelength of 537 nm with a FWHM of 29 nm. Blue QLEDs containing a ZnCdS core and a ZnS shell showed extremely high EQE of 10.7%, with a peak wavelength of 455 nm and very narrow FWHM of 20 nm. Due to the sub-band gap turn-on, these QLEDs showed high power efficiencies of 18, 60, and 2.7 lm/W for red, green, and blue, respectively. Using spincoating and photolithographic patterning of the electrodes, prototype monochrome red and green Active Matrix-QLED displays were demonstrated in collaboration with the Shanghai Tianma Microelectronics Group. More recently, NanoPhotonica showed improved efficiency in blue and green QLEDs [96]. By further tuning the QD synthesis and device fabrication procedures, green QLEDs with a record high 21% EQE (approximately the theoretical maximum EQE) with a peak wavelength of 526 nm and FWHM of 27 nm, as well as blue QLEDs with 11.2% EQE with peak wavelength of 463 nm and FWHM of 24 nm were demonstrated. The 21% EQE for the green QLED represents the highest reported efficiency of any color QLED and is the same as vacuum-deposited red and green OLEDs used in commercially available mobile AMOLED displays. Additionally, by controlling the emission peak wavelength and FWHM, the QLEDs emitted extremely saturated colors, producing a record 90% of Rec. 2020 color gamut, as shown in Figure 2.18. The lifetimes of these devices were impressive, with green and red T95 at 1000 cd/m2 of ∼200 h. T50 lifetimes at 100 cd/m2 were reported to be >100 000 and >280 000 h for green and red, respectively.

Quantum Dots for Displays and Solid State Lighting 0.6

520 530 540 550

69

560 570

580

590

510

600 610

620 630

640

680

500

0.5

490

0.4

0.3 480

v' 0.2

NTSC 1987 470

Rec. 2020

0.1

NanoPhotonica QLED 460 450

0

0.1

440 430 420

0.2

0.3

0.4

0.5

0.6

u'

Figure 2.18 The champion color gamut of QLEDs as of spring 2016, produced by NanoPhotonica, Inc. NanoPhotonica’s QLEDs produce 90% of the Rec. 2020 color gamut, as well as ∼170% of the NTSC 1987 color gamut (also called SMPTE “C”, where SMPTE is the Society of Motion Picture and Television Engineers). Image courtesy of NanoPhotonica, Inc.

Other groups reporting high efficiency Cd-based QLEDs include Lee et al. [97], who reported 40 cd/A and 12.6% EQE green QLED by employing CdSe@ZnS/ZnS QDs in a regular structure. These QDs consisted of a CdSe core transitioning into a ZnS shell through alloying, followed by an additional 1.6 nm thick ZnS shell. Lee et al. [98] also reported 7.1% EQE deep blue QLEDs emitting at 452 nm using ZnCdS/ZnS QDs with an alloyed core/shell interface and a thick ZnS shell. Kwak et al. [95] have reported impressive work on all three primaries in an inverted structure. They achieved maximum EQEs of 7.3, 5.8, and 1.7% for red, green, and blue using Cd-based QDs with highly saturated emission. T50 lifetimes of the inverted devices were ∼600 h at 500 cd/m2 , while T50 lifetimes of the regular structure employing PEDOT:PSS as the HIL (a material known to be unstable in air) were only about 1 h.

70

Materials for Solid State Lighting and Displays

Shen et al. [93] reported an extremely high efficiency blue-violet QLED using ZnCdS/ZnS QDs in a regular device structure. By substituting relatively short 1-octanethiol ligands for long oleic acid ligands and fine-tuning the composition of the QDs, average EQEs over 10% were achieved, with a champion device reaching over 12%. The highest performing devices emitted at 443 nm with a FWHM of 21.5 nm and showed a low turn-on voltage of only 2.6 V – a sub-band gap turn-on. Ji et al. [94] reported an inverted orange QLED with a stepwise HIL scheme. By using materials with a consecutively deeper HOMO level between the ITO anode and the QD EML, hole injection was improved for orange-red CdSe/ZnS-based QLEDs, generating up to >16 cd/A. Castan et al. [99a] recently reported an all solution-processed inverted QLED. In this work, a surfactant was added into the top PEDOT:PSS layer, which improved the performance of the devices significantly over those without the surfactant due to improved wetting of the PEDOT:PSS on the underlying HTL. While this research is needed to fulfill the promise of an all solution-processed inverted QLED, the devices were plagued by low efficiencies (8% EQE) for an all solution-processed QLED with p-n junction charge generation layers serving on both the hole injection side and electron injection side. This work is promising for further improvement of the all solution-processed inverted structure. Cadmium-free QLEDs While CuInS2 , ZnS:Mn, and Si QDs have been used as Cd-free alternatives for electroluminescent devices, the most promising materials for a Cd-free QLED have been InP and ZnSe. Several authors have reported InP-based QLEDs, including Lim et al. [100], Ippen, et al. [101], and Yang et al. [102]. InP seems to be the most promising material at this time for Cd-free green, yellow, or red QLEDs in terms of efficiency. Lee’s group [100] obtained 3.46% EQE with a maximum luminance of 3900 cd/m2 , the highest reported in the literature for InP so far. The QDs were InP coated with a ternary ZnSeS graded composition shell for a smooth confinement potential and to isolate the core from the surface states that cause charge transfer and subsequent Auger recombination. Despite this promising result, InP is clearly far behind Cd-based QDs in efficiency, and no lifetime studies have been reported for InP-based QLEDs. Additionally, the narrowest emission spectra for InP QDs are approximately 40 nm; this is not narrow enough to generate the saturated colors needed to create a wide color gamut in displays. Much more work is needed to understand the nucleation and growth dynamics of InP QDs as well as the ligand dynamics and behavior in an electroluminescent device. For Cd-free blue QLEDs, ZnSe is the most promising material so far. Wang et al. [103] have reported a record high 7.8% EQE violet QLED using ZnSe/ZnS core/shell QDs in a regular structure. The emission of the QDs could be tuned from 400 to 450 nm and FWHM remained below 20 nm, with PL QY from 63 to 83%. As is the case for the InP, ZnSe-based dots suffer drawbacks that will need to be overcome before they are incorporated into displays. First, the bulk band gap of ZnSe is approximately 2.7 eV (460 nm photon). Therefore, ZnSe-based QDs in the quantum confinement regime are forced to emit with wavelengths

Quantum Dots for Displays and Solid State Lighting

71

shorter than approximately 450 nm, which is too far into the violet range to generate the pure blue hues required for complete Rec. 2020 color space coverage. Additionally, the efficiency of these QDs in devices has not been shown to approach the highest efficiency of the Cd-containing QLEDs created by the same group. It is unclear at this point why the efficiency is lower, but a likely factor is the lack of a ternary or quaternary composition gradient available to this system as well as the necessity to push the ZnSe size to the limit of strong quantum confinement. White QLEDs Multicolor and white QLEDs have been the subject of only a few reports so far. For displays, the appeal of a white QLED is not necessarily warranted due to the necessity to use color filters to separate R-G-B primaries in much the same way as LCDs and white OLEDs currently do. However, if a QLED display manufacturing cost analysis suggests this as the way forward, then it would be advantageous to use a white QLED for the following reasons. White QLEDs can be prepared in two ways: (1) in a tandem structure [104], wherein individual layers in the device stack are composed of only one color of emitting QDs. These individual color-emitting layers are then stacked on top of each other to create white light emitted from the pixel. (2) By directly mixing R-G-B-emitting QDs [105–107] (or other primaries) into the same emitting layer that would normally be reserved for a single color-emitting QD in a monochrome QLED. A graphic showing the device structure of white QLEDs of both types is shown in Figure 2.19. In both types of white QLEDs, the devices generally suffer from a voltage-dependent emission spectrum due to

Mixed QD EML

AI

ZnO NPs

QDs

PVK

PEDOT:PSS

ITO

Bilayered QD EMLs

Figure 2.19 A comparison of two methods of creating white QLEDs: mixing primary colors and tandem layer construction of primary colors. Source: Kim 2015 [104]

72

Materials for Solid State Lighting and Displays

the different band gaps and associated charge injection barriers associated with multiple sizes and compositions of QDs. At low voltages, the emitted color tends toward red, while at high voltages, the blue emission increases and the color can be tuned toward the blue part of the spectrum. In mixed layer white QLEDs, the devices also suffer from FRET-induced color shifting as excitons are transferred from larger band gap blue QDs to small band gap red QDs. Additionally, charge transport through a mixed layer QLED is difficult due to large barriers and changes in charge carrier mobilities between nearest neighbor dots of dramatically different composition and band gap. Significant work will need to be performed to overcome these challenges to realize high efficiency and long operating lifetime. 2.8.1.2.4

Challenges to High Performance and Reporting

High Performance Several obstacles remain in the way of realizing QLEDs in all three colors for high efficiency and long-lived displays. Chief among them are improving material purity, surface passivation with appropriate ligand choices [108], overcoming FRET, exciton diffusion or other energy transfer between QDs in the emissive layer [109], charge imbalance that opens nonradiative Auger recombination pathways [109–111] and choosing the appropriate smooth gradient confinement potentials to reduce Auger recombination in the case of charge imbalance or parasitic charge transfer and field-assisted exciton dissociation [112]. From a manufacturing standpoint, there has been work on printing QLEDs using several techniques, including various contact printing methods [113, 114], 3D printing [115], continuous flow nozzle jet printing [116], and standard piezoelectric inkjet printing. Typical piezoelectric printing has suffered from the coffee ring effect, printer head corrosion, and solution viscosity issues. However, progress has been made in proprietary printing methods using modified inkjet printers for OLEDs [117], which should translate into QLED printing progress in the future. Degradation Rate and Device Lifetime There are critical challenges if the QD and QLED communities are to realize commercially viable display technologies beyond the photoluminescent applications. While physics and chemistry principles are universal, it is often difficult to exactly reproduce the literature reports of QD synthesis or device fabrication because of the extreme sensitivity to individual researchers’ laboratory techniques. Additionally, compared with the small organic molecules present in OLEDs, it is difficult to model the photophysics, ligand dynamics, and charge transport through QDs and QLEDs using computational methods. This hinders progress through the lack of down-selection and rapid screening of synthesis recipes and QLED device structures. For QLEDs to be viable commercial products, there should be much more study on the operational lifetime of QLEDs. There are only a handful of papers published in the open literature that report operational lifetimes, fewer that investigate the failure mechanisms [86], and none that solely report on the failure mechanisms in-depth. Additionally, it is common to report the T50 lifetime at an initial luminance of 100 cd/m2 . While T50 gives a common comparison for all devices that include short lifetimes, this is a relic of the early days of electroluminescent OLEDs, and is no longer relevant to real displays or lighting. Humans can perceive changes in luminance as small as ∼3%, so T95 or T97 is actually the preferred metric in the display industry, while T50 represents a display decaying well beyond its useful

Quantum Dots for Displays and Solid State Lighting

73

life, especially when differential degradation times between blue, red, and green QLEDs (and OLEDs) are taken into account – Universal Display Corporation published T95 at 1000 cd/m2 lifetime data for red (50 000 h), green (18 000 h), and blue (700 h) phosphorescent OLEDs on their corporate website in 2012 [118]. Recent mass production-quality OLED lifetime data are proprietary and unavailable, but it is widely known that blue OLEDs (and QLEDs) degrade much quicker than red and green. The modern display industry standard is now the T95 lifetime reported with an initial luminance of 1000 cd/m2 . To illustrate the need to report the T95 lifetime with starting luminance of 1000 cd/m2 , consider the following. Modern mobile and television displays emit maximum luminance of ∼400–500 cd/m2 , with average luminance values of 200–300 cd/m2 . However, the actual R-G-B subpixels require much higher luminance during operation. To illustrate this, suppose a display pixel is divided into R − G − B subpixels such that each subpixel has one-third of the total pixel emitting area. Also, suppose the pixel has a generous aperture ratio (ratio of total emitting area of the R + G + B subpixels to total pixel area) of 50% (real AMOLED displays are often lower). To display a scene at 500 cd/m2 with only green (or red or blue) light (e.g., a grass field in a sporting event), the green subpixel would need to be driven at 500 cd/m2 × 3 (compensation for one-third of total emitting area inside the pixel being green) × 2 (50% aperture ratio) = 3000 cd/m2 . Additionally, there is often a circular polarizer in the display to reduce ambient light reflection, which has a transmission of ∼50%, further increasing the necessary light output from the subpixel by a factor of 2. Thus, the total luminance emitted from a green subpixel to create a 500 cd/m2 green image is actually 6000 cd/m2 . Following the same calculation, for an average display luminance output of 200 cd/m2 , a subpixel would need to be run at 2400 cd/m2 . Also, because displays are refreshed more than once per image with a duty ratio of less than 100%, the time-averaged luminance will need to be further increased by a factor of 1/(duty ratio) to maintain the display brightness. The duty ratio of a particular display is proprietary information for each manufacturer. It is apparent now that 100 cd/m2 is no longer a meaningful luminance level to drive and report lifetimes, and accelerated tests with starting luminance of 5000–10 000 cd/m2 or higher are needed in practical settings. 2.8.2 2.8.2.1

Solid State Lighting Motivation for QDs in Solid State Lighting

Solid state lighting (SSL) is another area in which QDs can have a significant cost and performance advantage over the current state-of-the-art. For many of the same reasons QDs are beneficial materials for displays, they can also be integrated into SSL technologies as downconversion phosphors or electroluminescent phosphors – high power efficiency, tunable emission wavelengths, relatively abundant, and low cost. At the time of writing, QDs are not as widely distributed into SSL products as in displays, but there are promising developments in stability and color performance. These developments should lead to more SSL products that include QDs in the coming years. One important driver for QD integration into SSL is limited material availability and poor supply chain stability of the rare-earth elements used in the current generation of downconversion phosphors for LEDs. Figure 2.20 depicts the supply risk and importance to clean energy of rare-earth and other elements. For example, it is noteworthy that Y and Eu are critical to clean energy and downconversion phosphors in particular [Y in Y3 Al5 O12 :Ce3+

74

Materials for Solid State Lighting and Displays

Neodymium Dysprosium

4 (high)

Importance to clean energy

Lithium Tellurium

Europium Yttrium

Terbium

3

Critical Near-Critical

Cerium Lanthanum Cobalt Manganese Gallium Praseodymium Indium

Nickel 2

Not Critical

Samarium

1 (low) 1 (low)

2

3

4 (high)

Supply risk

Figure 2.20 The U.S. Department of Energy assessment of materials critical to renewable energy and their supply risk. Of critical importance are Eu and Y, which are nearly ubiquitous in rare-earth downconversion phosphors for solid state lighting and displays. Source: http:// energy.gov/sites/prod/files/DOE_CMS2011_FINAL_Full.pdf . [119]

(YAG:Ce) and Eu in many different phosphors], but the supply risk for both is also very high [119, 120]. A counterintuitive reason for using QDs as light sources in SSL is the narrowband emission spectra of QDs. Ideal Planckian (blackbody) emitters emit broad spectra and by definition have a perfect 100 color rendering index (CRI), It is counterintuitive that narrowband emitters would be used in a high CRI application because most everyday objects have broad reflection spectra, and so could be subject to unfaithful color rendering with a nonideal (non-Planckian) light source. Man-made Planckian emitters, such as incandescent light bulbs have low luminous efficacy of radiation (∼15 lm/W), but nearly perfect CRI. Incandescent replacements with relatively high luminous efficacy of radiation (∼100 lm/W) and broad emission spectra such as fluorescent bulbs typically have medium–low CRI of 50–75. It has been shown that under the correct optimization conditions, luminous efficacy of radiation for a white light source can be maximized by using carefully selected multiple narrowband emitters [121]. Moreover, high CRI values can be obtained with relatively few narrow emitters while also maximizing the luminous efficacy of radiation. In fact, for CRI = 85, only three narrowband primaries are needed, while an extremely high CRI = 90 only requires four primaries to maximize luminous efficacy of radiation [121]. Commercially available LED bulbs with rare-earth-doped phosphors are typically 70–90 lm/W with CRI between 60 and 80, with the best LEDs

Quantum Dots for Displays and Solid State Lighting

75

well above 100 lm/W and CRI > 90. It is easy to realize that QDs, with their tunable emission spectra and narrowband emission, are ideal emitters to cover the visible spectrum, maximize luminous efficacy of radiation, and produce white light with a high CRI for faithful object illumination. QDs of various compositions, II–VI, III–V, and I–III–VI, core/shell, and core-only can all be used to create SSL-type devices. Each has its advantages and disadvantages. As discussed in Section 2.8.1, the II–VI materials such as CdSe compounds exhibit the highest PL QY and narrowest emission spectra, and thus can be used to create a high luminous efficacy of radiation devices while maintaining a high CRI. However, II–VI materials such as CdSe are coming under increasing regulatory and public relations scrutiny due to environmental and health concerns surrounding the use of Cd. III–V materials including InP, are not highly regulated in the RoHS Directive and can be tuned throughout the visible spectrum [102]. However, these materials still exhibit much broader emission spectra and lower PL QY than II–VI materials. Additionally, InP is a known carcinogen, so may be subject to future regulations. I–III–VI materials such as CuInS2 are relatively new and offer a lower toxicity than both the II–VI and III–V QDs. They have an interesting emission spectrum that is often extremely broad due to being mediated by a mid-gap defect state [122]. However, the PL QY is also far below that of the Cd-based QDs. While this may not be ideal for high luminous efficacy or color purity, they may see other applications in security and biological imaging in the future. 2.8.2.2

Quantum Dots as Electroluminescent Light Sources

QLEDs are promising electroluminescent light sources for SSL due to their low cost, large scale and high throughput manufacturing compatibility, flexible and versatile form factors, and ability to be a large area distributed light source rather than a point light source. As discussed in Section 2.8.1, white QLEDs can be created using a bilayer structure or mixed QD EMLs. It is well known that Cd-based QDs can be synthesized that span the entire visible spectrum, and thus could be combined to create white QLEDs from one simple synthesis and device fabrication method. Following the above discussion of narrowband phosphors creating high CRI values, this has indeed been demonstrated by coupling yellow-emitting ZnCdSeS QDs with blue emission from a ZnO-TiO2 hybrid in a multilayer QLED stack [123]. The device exhibited a very high CRI of 91 with CIE (x, y) coordinates of (0.33, 0.33). It should be noted that the blue emission in this device was rather broad covering blue and green, while the ZnCdSeS emission was positioned to cover a maximum amount of the remaining visible spectrum. In separate work by Lee [107], a mixed R-G-B QD layer with was used to create white QLEDs with CRI values of 71–75 and correlated color temperatures (CCT) of 6126–7719 K. The drawback of these devices is that the color shift due to the R-G-B mixing of the EML can dramatically change the color output as discussed above in Section 2.8.1. However, in some cases, this can be viewed as a desirable feature in SSL luminaires under well-controlled conditions. In other work published by Chen et al. [124], CdSe/ZnSe core/shell QDs were integrated with a InGaN blue LED to create a white LED. Two-band (only yellow QDs) emission spectra showed CRI of 50, while three-band LEDs showed a CRI of 91 with CIE (x, y) coordinates of (0.33, 0.33),

76

Materials for Solid State Lighting and Displays

extremely close to the D65 standard illuminant. These works are good proof of concept that a relatively narrow QD emission spectrum can be used to create a high CRI through careful band engineering, and more improvements should be realized with careful QD synthesis tuning and EML composition control. QLEDs are still in the early stages of development on the way toward commercialization. The low operational lifetimes in QLEDs designed for displays were discussed previously, and this requirement only increases for SSL applications. T95 lifetimes of several hundred hours at 1000 cd/m2 are considered good enough to begin commercialization efforts for OLEDs and QLEDs for displays. However, for SSL applications, T70 lifetimes of several tens of thousands of hours will need to be met for a viable commercial product [125]. Perhaps with the availability of large area, distributed planar light emission, the necessary operating luminance (luminous flux in lumens for SSL) for some indoor fixtures will be lower compared with some of the most stringent display requirements. Nevertheless, much more research on the failure modes and efficiency enhancements are needed to ensure QLEDs have a place in the SSL market in the future. 2.8.2.3

Quantum Dots as Downconversion Phosphors

Downconversion materials for SSL should meet the following criteria [120]: high PL QY, high luminous efficacy of radiation, enable a high CRI and low CCT for “warm” light, exhibit low or no emission quenching at elevated temperatures, have no significant absorption at wavelengths other than the excitation wavelengths, low reflectance and high absorbance at the excitation wavelength, chemical and photostability, fast PL decay (short exciton lifetimes), and long operating lifetimes. Table 2.3 compares traditional rare-earth-doped phosphors to QDs for these SSL applications. One particular area where QDs are expected to improve SSL applications is the red color rendering of LEDs. The saturated red test condition is not included in the standard CRI calculation as it only takes into account a certain set of primaries from the Munsell color system that do not include the saturated red [126]. Measuring the rendering of saturated red in SSL is increasingly important because of YAG:Ce downconversion phosphor issues with Table 2.3 Comparison of the qualities of rare-earth-doped phosphors and quantum dots Pros

Cons

Phosphors

Quantum dots

High quantum yield Thermal, chemical and photostability Low thermal quenching High oscillator strength (Eu2+ , Ce3+ ) Large absorption cross-section Fast PL decay (short radiative lifetimes) Localized emission at dopants Narrow blue absorption low oscillator strength (Eu3+ ) Deep red emission tail (Eu2+ )

High quantum yield Size and spectra-tunable, narrow emission Broadband absorption (especially blue) High oscillator strength Fast PL decay (short radiative lifetimes) Low scattering Thermal quenching at high temperatures Susceptible to ligand loss, restructuring

PL, photoluminescence. Source: McKittrick and Shea-Rohwer 2014 [120]. Reproduced with permission of John Wiley and Sons.

Quantum Dots for Displays and Solid State Lighting

77

properly rendering red, which of course is present in skin tone, blood, grocery stores, advertising, and other commercially relevant places. In a typical white LED, a blue InGaN die is covered in a YAG:Ce yellow phosphor to make white light. However, the YAG:Ce phosphor does not emit strongly in the red, but rather has a weak tail that extends into the deep red. To mitigate this, Eu2+ -doped phosphors such as Sr2 Si5 N8 :Eu2+ are included to increase the emission in the red spectral region. The drawback to this is that such Eu2+ -doped phosphors emit very broad peaks into the deep red and near-infrared where the human eye is not very sensitive or not at all sensitive [120]. This deep red and infrared emission offsets the very high QY of these materials. To find a compromise, K2 SiF6 :Mn4+ (KSF or PSF) phosphors have been developed that have extremely narrow (∼5 nm) FWHM emission peaks and high QY. However, this material has shown reliability issues under high flux testing. The reliability issues have greatly diminished due to additional commercial effort, but these improved KSF phosphors are not widely available as they are owned by GE Lighting and marketed under the name TriGainTM [127]. Nichia is also reported to be developing improved KSF phosphors. However, with restrictive use of these phosphors and lingering reliability issues, QDs are seen as an ideal solution to the red rare-earth-doped phosphor dilemmas. QDs are widely available, emit very narrow spectra so there is no light wasted in the deep red or infrared, and are rapidly improving in reliability. Figure 2.21 shows the three main methods to integrate QDs into existing LED products as downconversion phosphors utilizing PL [128]: 1. Directly deposited on the LED chip inside the package. 2. Uniformly dispersed in the LED packaging resin. 3. Deposited away from the chip on the packaging as a “remote phosphor”.

phosphor

encapsulant

(a) Uniform distribution of phosphor powders LED

reflector cup

(b) Conformal distribution

(c) Remote phosphor distribution transparent substrate

Figure 2.21 Schematics of the three methods for incorporating QDs or phosphors in the “on-chip” mode. (a) Distributed materials in the encapsulation resin, (b) deposition of materials directly on the LED die, and (c) deposition of the materials as a “remote” phosphor away from the die, but still within the packaging. Source: Choi et al. 2013 [128]. Reproduced with permission of The Electrochemical Society

78

Materials for Solid State Lighting and Displays

In all these methods, low photostability under extreme light flux (>10 W/cm2 ), and thermal instability due to high heat produced by the LED die (>125 ∘ C) are serious challenges that need to be overcome in order for QDs to be viable products [78]. In the first method, the QDs are directly placed on the LED die inside the packaging. This application demands the highest quality materials with extreme photostability and thermal stability. However, it also requires the least amount of material, driving the cost of the product down. In the second method, encapsulating resin above the LED chip is loaded with QDs. This creates some complications, as the surface of the QDs must be functionalized correctly to be dispersed into a proprietary resin. Often, the encapsulating resin, the LED die, and the QDs are all manufactured by different entities, making functionalization and composition tuning of the QDs and resins difficult. Once the QDs are dispersed in the resin, however, manufacturing the resin-filled packaging follows the standard procedure as if there were no QDs embedded. In the third method, the remote phosphor is perhaps the most lenient in terms of thermal stability and photostability requirements. However, when using this method, a larger amount of QD material must be used to coat the packaging or a remote plate, increasing the cost of the product. Additionally, by placing a mixture of red and green QDs to make a yellow remote phosphor, the aesthetics of the light bulb to the consumer may not be pleasant enough in the off-state for use in the home or in a commercial sales display. SSL is one of the most important industries in energy and environmental sustainability. It is clear that there are advantages to employing QDs in any number of methods for these technologies. In fact, there have been products released in the field with QDs. QD Vision’s first product was the Quantum LightTM remote phosphor plate that could be attached to a LED bulb for general SSL applications. QD Vision no longer manufactures this product, however. The Orion QD Quantum Dot Linear Lighting lamp incorporates Cadmium Free Quantum Dots from Nanoco, and the Zylight F8-100 LED Fresnel Light claims to incorporate QDs into its fixture. With increased research and development on electroluminescent QLEDs as well as the reliability of on-chip QDs, these materials should have the ability to dramatically improve the SSL industry. 2.8.3

Biological Applications

An important area in which QDs can play a critical role in providing a next-generation improvement is biological and biomedical applications. Biological systems are extremely complex and scientists rely on obtaining unambigious information from crowded environments, often at the molecular level, with many overlapping signals. QDs are a great material for inclusion into biological or biomedical systems due to their resistance to photobleaching, tunable emission wavelengths, broad absorption spectra, small size, and surface functionality. Because QDs can be synthesized by biological systems and external chemical methods [129], many different routes exist for creating and applying QDs tailored to specific biological environments. QDs in biological applications can act as static fluorescent labels, drug delivery systems, or dynamic system trackers. In fluorescent labels, the QDs can be delivered in four different ways [129]: (1) passive endocytosis, wherein the cell membrane engulfs the outside particle (QD) and captures it; (2) active endocytosis through ligand–receptor-mediated interactions with the cell membrane; (3) combinatorial, large parameter-space searches; and (4) delivery through physical treatments.

Quantum Dots for Displays and Solid State Lighting

79

Multicolor labeling for biological assays is an important method for resolving various components of a biological system in close proximity through relatively simple microscopy techniques compared with those used in semiconductor research and development. By carefully functionalizing QD surfaces, the use of 10 emission intensity levels and 6 colors could theoretically encode 1 million nucleic acid or protein sequences. Early work in biolabeling by Weiss and Alivisatos [130] was performed in which separate red and green emitting CdSe/CdS/SiO2 core/shell QDs were functionalized to label fibroblasts – cells that produce the extracellular matrix and collagen. Red-emitting QDs were functionalized to label actin filaments, which, along with myosin, are responsible for muscle contractions in sliding filament theory. Green-emitting QDs were used to label the cell nuclei. By employing QDs with broad, overlapping, absorption spectra and two different emission spectra, these filaments and nuclei could easily be resolved with single excitation confocal fluorescent microscopy. Han et al. [131] embedded different-sized QDs into highly uniform and reproducible polymeric microbeads, which yielded bead identification accuracies as high as 99.99%. The broad absorption spectra of the QDs also allowed single wavelength excitation and multicolor emission from different-sized QDs. This spectral encoding technology has opened up new opportunities in in-vivo and in-vitro diagnostics, gene expression studies, high-throughput screening, and medical diagnostics. Equally useful is multi-modal biolabeling [132], wherein a combination of fluorescence, electron microscopy, or magnetic imaging is used. For example, Foreman et al. [133] and Oey et al. [134] demonstrated that the CdS:Mn/ZnS QDs are good for multi-functional imaging. Here, Gd(III)-functionalized, silica-coated CdS:Mn/ZnS water soluble QDs were shown to be fluorescent, radio-opaque, paramagnetic, suitable for attaching biomolecules such as proteins, peptides and antibodies, and were stable in an in-vivo environment. The magnetic resonance imaging (MRI) properties of these Gd-functionalized QDs have been demonstrated [7] and good MRI contrast in both the longitudinal (T1 ) and transverse (T2 ) proton relaxation time-weighted images was demonstrated. Since the autofluorescence of live cells excited by ultraviolet is strong in the visible region and overlaps the emission of most labeling materials, it is difficult to get backgroundfree PL spectra. Compared with conventional organic dyes, QDs have longer lifetime (several tens of nanoseconds), therefore background-free PL images can be obtained by using time-gated fluorescent microscopy [135]. The photostability of QDs is better than that of organic dyes, allowing the use of QDs to monitor biological events over long times as dynamic system trackers. Another biological application of QDs is to build on–off switches by utilizing FRET between the QD donor and an organic acceptor. The emission from CdSe QDs can be quenched if they are connected to an organic acceptor because the energy absorbed by the QDs is transferred to the acceptor via FRET [136]. If the distance between the QD donor and organic acceptor is increased by the addition of foreign organic materials, emission from the QDs can be turned on again. This so called on–off switch has the potential to be used as a sensor in many important applications, including healthcare, environmental monitoring and biodefense systems. For biological and medical applications, it is of particular importance to study the photophysical properties of QDs in living cells [137]. The photoinduced optical properties of the intracellular QDs are of interest. After injecting thiol-capped CdTe QDs into living cells, the PL intensity increased with time and the emission peak blue-shifted [137]. Deoxygenation

80

Materials for Solid State Lighting and Displays

prevented the PL blue shift, suggesting that photoactivated oxygen was responsible. The activated oxygen is presumably formed from the oxygen that intercalates the thiol layer at the QD core surface. When QDs are used as fluorescence probes for cellular imaging, the effects of the PL blue shift and photobleaching must be considered. Because QDs are often synthesized in nonpolar, nonaqueous solvents, they must be made water soluble for biology application, for example by depositing a silica shell or organic layer as discussed above. Aqueous-based synthesis methods have been used to produce silica-capped, highly fluorescent CdTe QDs [138]. The emission from CdTe and QDs can be tuned to the near-infrared and the silica shell can prevent the leakage of toxic Cd2+ . This is important because tissue absorption is at a minimum in two biological windows of 700–900 and 1000–1400 nm, opened by low optical absorption of water and hemoglobin [139, 140]. Cytotoxicity and the potential interference of QDs with cellular processes are the subject of intensive studies [138, 141]. In addition to making QDs water soluble, a silica shell allows easy functionalization with biomolecules such as proteins [8] and results in greater photostability. Another application that takes advantage of infrared-emitting QDs is live tumor detection in deep tissue. Because the first biological window of 700–900 nm is more susceptible to interference from autofluorescence and Rayleigh scattering, the second window of 1000–1400 nm is preferred. Silver sulfide QDs were recently functionalized and injected into a mouse having a xenograft tumor on its leg [142]. As the QDs circulated through the body, the tumor uptake of the 6PEG-functionalized Ag2 S QDs was monitored with 10000

30 min

2h

4h

(a)

(b)

(c)

10 h

24 h

(d)

(e)

8000

6000

4000

2000

0

(f)

Figure 2.22 Near-infrared fluorescence image of a xenograft 4T1 tumor with uptake of 6PEG– functionalized Ag2 S quantum dots. The uptake of the QDs by the tumor was monitored for 24 h. (a–e) Near-infrared imaging of a time course of 6PEG-Ag2 S QD uptake by the xenograft tumor. (f) White light photograph showing the tumor grafted on to the leg of the mouse. Source: Hong et al. 2012 [142]. Reproduced with permission of John Wiley and Sons

Quantum Dots for Displays and Solid State Lighting

81

near-infrared imaging. Figure 2.22 shows the tracking of the QD uptake over the first 24 h after injection. The ability to track and diagnose tumors, cancers, viruses and other elements of biological systems live with simple fluorescent tools, combined with biopsy-type multi-modal diagnostics is indispensable. QDs play a key role in this field and will continue to lead to new healthcare and fundamental biological improvements.

2.9

Perspective

The optical properties of QDs and some of the current and upcoming applications have been discussed and detailed above. Among the various branches in nanotechnology, zero-dimension nanostructures (QDs) have paved the way for numerous advances in both fundamental and applied sciences. This is due to the fact that the QDs exhibit significantly different – and tunable – optical, electronic and physical properties as compared with bulk materials. With respect to the synthesis of QDs, significant progress has been made in studies of the growth kinetics through both theoretical models and experimental data. Procedures ranging from simple wet chemical methods to very sophisticated and equipment-intensive atomic layer deposition techniques are being used to synthesize QDs. The “bottom-up” approaches are being most widely explored with a wider variety of materials to generate QDs. The entry of photoluminescent applications of QDs into the display market over the last several years signals the arrival of these materials as high quality products in every-day consumer goods. In the future, QDs will see increased adoption into LCD systems as a way to maintain competitiveness with the color gamut and HDR available to OLED with the low cost structure of LCDs. As the electroluminescent efficiency and lifetimes of QLEDs improve to be as good as, or better than, OLEDs, QLEDs will take over a significant market share due to their inherent wide viewing angles, flexible form factors, extremely saturated colors, and low cost profile. Despite the large amount of research, there is still a lot to understand about the use of QDs in large-scale biological and SSL applications. Although preliminary experiments at the laboratory level have been successful, scaling up the production and retaining the properties of QDs in these demanding environments is not trivial. Continued research and development on QDs will provide further improvements in quantum efficiency, better device fabrication, longer operational lifetimes, and new materials systems. Nanobiotechnology is already a vibrant research and development area. However, several stumbling blocks need to be overcome to ensure a close marriage of biology and nanotechnology, particularly with respect to QDs. Bio-functionalizing the QDs and surface and interface engineering are major challenges in dynamic and complex biological systems. The requirements of applications such as displays, lighting, selective sensors, bio-imaging, MRI contrast agents, and bio-labels will require the scientific and engineering community to answer questions about QD synthesis, properties, ageing, and toxicity. As the future approaches, QDs are poised to enter new fields and improve upon old fields. QDs are promising materials for high-technology applications where demanding performance in color quality, functionalization, and novel form factors are key.

82

Materials for Solid State Lighting and Displays

Acknowledgments This work was partially supported by ARL grant W911 NF-04-200023, ARO grant W991 NF-07-1-0050, and NSF grant #1353411. Discussions with Drs Heesun Yang, Christian Ippen, Hyeokjin Lee and Jungsik Bang, and with Jeff Yurek and Sooyeon Seo, are gratefully acknowledged.

References 1.

Shionoya, S. and Yen, W.M. (1998) Phosphor Handbook, CRC Press, New York, p. 608. 2. Lee, J.D. (2005) Concise Inorganic Chemistry, Blackwell Science, Noida, India, p. 1032. 3. Round, H.J. (1907) A note on carborundum. Electrical World, 49(Feb. 9), 309. 4. Kuchibhatla, S.V.N.T. et al. (2007) One dimensional nanostructured materials. Progress in Materials Science, 52 (5), 699–914. 5. Bera, D., Kuiry, S.C. and Seal, S. (2004) Synthesis of nanostructured materials using template-assisted electrodeposition. JOM, 56 (1), 49–53. 6. Alivisatos, A.P. (1996) Perspectives on the physical chemistry of semiconductor nanocrystals. Journal of Physical Chemistry, 100 (31), 13226–13239. 7. Yang, H.S. et al. (2006) Gd-III-functionalized fluorescent quantum dots as multimodal imaging probes. Advanced Materials, 18 (21), 2890–2894. 8. Santra, S. et al. (2005) Synthesis and characterization of fluorescent, radio-opaque, and paramagnetic silica nanoparticles for multimodal bioimaging applications. Advanced Materials, 17 (18), 2165–2169. 9. Santra, S. et al. (2005) Rapid and effective labeling of brain tissue using TATconjugated CdS: Mn/ZnS quantum dots. Chemical Communications, 25, 3144–3146. 10. Yang, H. and Holloway, P.H. (2004) Efficient and photostable ZnS-passivated CdS:Mn luminescent nanocrystals. Advanced Functional Materials, 14 (2), 152–156. 11. Hines, M.A. and Guyot-Sionnest, P. (1996) Synthesis and characterization of strongly luminescing ZnS-capped CdSe nanocrystals. Journal of Physical Chemistry, 100 (2), 468–471. 12. Walter, P. et al. (2006) Early use of PbS Nanotechnology for an ancient hair dyeing formula. Nano Letters, 6 (10), 2215–2219. 13. Rooksby, H.P. (1932) The colour of selenium ruby glasses. Journal of the Society of Glass Technology, 16, 171. 14. Ekimov, A.I. and Onushchenko, A.A. (1981) Quantum size effect in 3-dimensional microscopic semiconductor crystals. JETP Letters, 34 (6), 345–349. 15. Efros, A.L. and Efros, A.L. (1982) Interband absorption of light in a semiconductor sphere. Soviet Physics Semiconductors-USSR, 16 (7), 772–775. 16. Rossetti, R. et al. (1984) Size effects in the excited electronic states of small colloidal CdS crystallites. The Journal of Chemical Physics, 80 (9), 4464–4469. 17. Sundaram, M. et al. (1991) New quantum structures. Science, 254 (5036), 1326–1335.

Quantum Dots for Displays and Solid State Lighting

18.

19.

20.

21.

22.

23.

24.

25. 26.

27.

28.

29.

30.

31. 32.

33.

83

Cingolani, R. and Ploog, K. (1991) Frequency and density dependent radiative recombination processes in III-V semiconductor quantum-wells and superlattices. Advances in Physics, 40 (5), 535–623. Wang, Y. and Herron, N. (1991) Nanometer-sized semiconductor clusters - materials synthesis, quantum size effects, and photophysical properties. Journal of Physical Chemistry, 95 (2), 525–532. Bang, J., Yang, H. and Holloway, P.H. (2006) Enhanced and stable green emission of ZnO nanoparticles by surface segregation of Mg. Nanotechnology, 17 (4), 973–978. Yoffe, A.D. (1993) Low-dimensional systems - quantum-size effects and electronicproperties of semiconductor microcrystallites (zero-dimensional systems) and some quasi-2-dimensional systems. Advances in Physics, 42 (2), 173–266. Yoffe, A.D. (2001) Semiconductor quantum dots and related systems: electronic, optical, luminescence and related properties of low dimensional systems. Advances in Physics, 50 (1), 1–208. Tolbert, S.H. et al. (1994) Comparison of quantum confinement effects on the electronic absorption-spectra of direct and indirect gap semiconductor nanocrystals. Physical Review Letters, 73 (24), 3266–3269. Tolbert, S.H. and Alivisatos, A.P. (1995) The wurtzite to rock-salt structural transformation in CdSe nanocrystals under high-pressure. Journal of Chemical Physics, 102 (11), 4642–4656. Woggon, U. (1997) Optical properties of semiconductor quantum dots, in Springer Tracts in Modern Physics, vol. 136 (eds J. Kuhn et al.), Springer-Verlag, Berlin. Klimov, V.I. (2006) Mechanisms for photogeneration and recombination of multiexcitons in semiconductor nanocrystals: Implications for lasing and solar energy conversion. Journal of Physical Chemistry B, 110 (34), 16827–16845. Brus, L.E. (1983) A simple-model for the ionization-potential, electron-affinity, and aqueous redox potentials of small semiconductor crystallites. Journal of Chemical Physics, 79 (11), 5566–5571. Murray, C.B., Norris, D.J. and Bawendi, M.G. (1993) Synthesis and characterization of nearly monodisperse CdE (E = S, Se, Te) semiconductor nanocrystallites. Journal of the American Chemical Society, 115 (19), 8706–8715. Alexander Ross, J.B. et al. (1991) Tyrosine fluorescence and phosphorescence from protiens and polypeptides, in Topics in Fluorescence Spectroscopy, vol. 3 (ed. J.R. Lakowicz), Springer, New York, p. 397. Gfroerer, T.H. (2000) Photoluminescence in analysis of surface and interfaces, in Encyclopedia of Analytical Chemisty (ed. R.A. Meyers), John Wiley & Sons, Ltd., Chichester, pp. 9209–9231. Auzel, F. (2004) Upconversion and anti-Stokes processes with f and d ions in solids. Chemical Reviews, 104 (1), 139–173. Issac, A., von Borczyskowski, C. and Cichos, F. (2005) Correlation between photoluminescence intermittency of CdSe quantum dots and self-trapped states in dielectric media. Physical Review B, 71 (16), 161302-1–161302-4. Xu, P.S. et al. (2003) The electronic structure and spectral properties of ZnO and its defects. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, 199, 286–290.

84

Materials for Solid State Lighting and Displays

34.

Djuriši´c, A.B. and Leung, Y.H. (2006) Optical properties of ZnO nanostructures. Small, 2 (8–9), 944–961. Spanhel, L. and Anderson, M.A. (1991) Semiconductor clusters in the sol-gel process - quantized aggregation, gelation, and crystal-growth in concentrated ZnO colloids. Journal of the American Chemical Society, 113 (8), 2826–2833. Yang, H. et al. (2004) CdS:Mn nanocrystals passivated by ZnS: Synthesis and luminescent properties. The Journal of Chemical Physics, 121 (20), 10233–10240. Efros, A.L. and Rosen, M. (1997) Random telegraph signal in the photoluminescence intensity of a single quantum dot. Physical Review Letters, 78 (6), 1110–1113. Williams, E.W. and Hall, R. (1977) Luminescence and the light emitting diode, in Science of the Solid State, 1 edn (ed. B.R. Pamplin), Pergamon Press, New York, p. 237. Nirmal, M. et al. (1996) Fluorescence intermittency in single cadmium selenide nanocrystals. Nature, 383 (6603), 802–804. Stefani, F.D. et al. (2005) Quantification of photoinduced and spontaneous quantumdot luminescence blinking. Physical Review B, 72 (12), 125304-1–125304-7. Stefani, F.D. et al. (2005) Memory in quantum-dot photoluminescence blinking. New Journal of Physics, 7, 197. van Sark, W. et al. (2002) Time-resolved fluorescence spectroscopy study on the photophysical behavior of quantum dots. Journal of Fluorescence, 12 (1), 69–76. Kuno, M. et al. (2000) Nonexponential “blinking” kinetics of single CdSe quantum dots: A universal power law behavior. Journal of Chemical Physics, 112 (7), 3117–3120. Chen, X.B. et al. (2003) Coherency strain effects on the optical response of core/shell heteronanostructures. Nano Letters, 3 (6), 799–803. Peng, X.G. et al. (1997) Epitaxial growth of highly luminescent CdSe/CdS core/shell nanocrystals with photostability and electronic accessibility. Journal of the American Chemical Society, 119 (30), 7019–7029. Balet, L.P. et al. (2004) Inverted core/shell nanocrystals continuously tunable between type-I and type-II localization regimes. Nano Letters, 4 (8), 1485–1488. Kim, S. et al. (2003) Type-II quantum dots: CdTe/CdSe(core/shell) and CdSe/ ZinTe(core/shell) heterostructures. Journal of the American Chemical Society, 125 (38), 11466–11467. Scherer, A., Craighead, H.G. and Beebe, E.D. (1987) Gallium-arsenide and aluminum gallium-arsenide reactive ion etching in boron-trichloride argon mixtures. Journal of Vacuum Science & Technology B, 5 (6), 1599–1605. Tsutsui, K., Hu, E.L. and Wilkinson, C.D.W. (1993) Reactive ion etched II-VI quantum dots - dependence of etched profile on pattern geometry. Japanese Journal of Applied Physics Part 1-Regular Papers Short Notes & Review Papers, 32 (12B), 6233–6236. Chason, E. et al. (1997) Ion beams in silicon processing and characterization. Journal of Applied Physics, 81 (10), 6513–6561. Burda, C. et al. (2005) Chemistry and properties of nanocrystals of different shapes. Chemical Reviews, 105 (4), 1025–1102.

35.

36. 37. 38.

39. 40. 41. 42. 43.

44. 45.

46. 47.

48.

49.

50. 51.

Quantum Dots for Displays and Solid State Lighting

52. 53.

54.

55.

56.

57. 58. 59.

60.

61. 62. 63.

64.

65. 66. 67. 68.

85

Bera, D. et al. (2008) Photoluminescence of ZnO quatum dots produced by a sol–gel process. Optical Materials, 30 (8), 1233–1239. doi: 10.1016/j.optmat.2007.06.001 Hoener, C.F. et al. (1992) Demonstration of a shell core structure in layered CdSe-ZnSe small particles by X-ray photoelectron and Auger spectroscopies. Journal of Physical Chemistry, 96 (9), 3812–3817. Kortan, A.R. et al. (1990) Nucleation and growth of CdSe on ZnS quantum crystallite seeds, and vice versa, in inverse micelle media. Journal of the American Chemical Society, 112 (4), 1327–1332. Yang, H.S., Santra, S. and Holloway, P.H. (2005) Syntheses and applications of Mndoped II-VI semiconductor nanocrystals. Journal of Nanoscience and Nanotechnology, 5 (9), 1364–1375. Yu, W.W., Wang, Y.A. and Peng, X.G. (2003) Formation and stability of size-, shape-, and structure-controlled CdTe nanocrystals: Ligand effects on monomers and nanocrystals. Chemistry of Materials, 15 (22), 4300–4308. Talapin, D.V. et al. (2001) A novel organometallic synthesis of highly luminescent CdTe nanocrystals. Journal of Physical Chemistry B, 105 (12), 2260–2263. Hines, M.A. and Guyot-Sionnest, P. (1998) Bright UV-blue luminescent colloidal ZnSe nanocrystals. Journal of Physical Chemistry B, 102 (19), 3655–3657. Singh, S., Rama, N. and Rao, M.S.R. (2006) Influence of d-d transition bands on electrical resistivity in Ni doped polycrystalline ZnO. Applied Physics Letters, 88 (22), 222111-1–222111-3. Lee, H., Holloway, P.H. and Yang, H. (2006) Synthesis and characterization of colloidal ternary ZnCdSe semiconductor nanorods. Journal of Chemical Physics, 125 (16), 164711-1–164711-7. Yang, H.S., Lee, H. and Holloway, P.H. (2005) Anisotropic growth of luminescent Eu3+ - or Er3+ -doped Gd2 O3 nanocrystals. Nanotechnology, 16 (12), 2794–2798. Battaglia, D. and Peng, X.G. (2002) Formation of high quality InP and InAs nanocrystals in a noncoordinating solvent. Nano Letters, 2 (9), 1027–1030. Zhu, J.J., Koltypin, Y. and Gedanken, A. (2000) General sonochemical method far the preparation of nanophasic selenides: Synthesis of ZnSe nanoparticles. Chemistry of Materials, 12 (1), 73–78. Qian, H.F., Li, L. and Ren, J.C. (2005) One-step and rapid synthesis of high quality alloyed quantum dots (CdSe-CdS) in aqueous phase by microwave irradiation with controllable temperature. Materials Research Bulletin, 40 (10), 1726–1736. Wang, L.C. et al. (2006) A hydrothermal method to prepare the spherical ZnS and flower-like CdS microcrystallites. Materials Letters, 60 (29-30), 3627–3630. Frank, F.C. and van der Merwe, J.H. (1949) One-dimensional dislocations. I. Static theory. Proceedings of the Royal Society London A, 198, 205. Volmer, M. and Weber, A. (1926) Keimbildung in übersättigten Gebilden (Nucleation in supersaturated structures). Zeitschrift für Physikalische Chemie, 119, 277. Stranski, I.N. and Krastanow, L. (1938) Zur Theorie der orientierten Ausscheidung von Ionenkristallen aufeinander (On the theory of oriented deposition of ionic crystals on top of each other)Abhandlungen der Mathematisch-Naturwissenschaftlichen Klasse IIb. Akademie der Wissenschaften Wien, 146, 797–810.

86

69.

Materials for Solid State Lighting and Displays

Eaglesham, D.J. and Cerullo, M. (1990) Dislocation-free Stranski-Krastanow growth of Ge on Si(100). Physical Review Letters, 64 (16), 1943–1946. 70. Tsukamoto, S., Bell, G.R. and Arakawa, Y. (2006) Heteroepitaxial growth of InAs on GaAS(001) by in situ STM located inside MBE growth chamber. Microelectronics Journal, 37 (12), 1498–1504. 71. Jiao, Y.H. et al. (2006) MBE InAs quantum dots grown on metamorphic InGaAs for long wavelength emitting. Physica E-Low-Dimensional Systems & Nanostructures, 35 (1), 194–198. 72. Lobo, C. and Leon, R. (1998) InGaAs island shapes and adatom migration behavior on (100), (110), (111), and (311) GaAs surfaces. Journal of Applied Physics, 83 (8), 4168–4172. 73. International Telecommunication Union, BT. 2020: Parameter values for ultrahigh definition television systems for production and international programme exchange. Document number BT.2020-2 (10/2015), Available at http://www.itu.int/ rec/R-REC-BT.2020/en. Accessed May 1, 2015. 74. European Union, Directive on the restriction of the use of certain hazardous substances in electrical and electronic equipment, RoHS 2 – Directive, Document 2011/ 65/EU, Available at http://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex: 32011L0065. Accessed May 1, 2015. 75. Chen, O. et al. (2013) Compact high-quality CdSe-CdS core-shell nanocrystals with narrow emission linewidths and suppressed blinking. Nature Materials, 12 (5), 445–451. 76. Tamang, S. et al. (2016) Chemistry of InP nanocrystal synthesis. Chemistry of Materials, 28 (8), 2491–2506. 77. Donofrio, R.L. (2011) Review paper: The Helmholtz-Kohlrausch effect. Journal of the Society for Information Display, 19 (10), 658–664. 78. Steckel, J.S. et al. (2015) Quantum dots: the ultimate down-conversion material for LCD displays. Journal of the Society for Information Display, 23 (7), 294–305. 79. Forrest, S.R. et al. (2003) Measuring the efficiency of organic light-emitting devices. Advanced Materials, 15 (13), 1043–1048. 80. Hofmann, S. et al. (2011) Top-emitting organic light-emitting diodes. Optics Express, 19 (56), A1250–A1264. 81. Anikeeva, P.O. et al. (2006) Photoluminescence of CdSe/ZnS core/shell quantum dots enhanced by energy transfer from a phosphorescent donor. Chemical Physics Letters, 424, 120–125. 82. Qian, L. et al. (2010) Electroluminescence from light-emitting polymer/ZnO nanoparticle heterojunctions at sub-bandgap voltages. Nano Today, 5, 384–389. 83. Xiang, C. et al. (2015) Origin of sub-bandgap electroluminescence in organic lightemitting diodes. Small, 11 (40), 5439–5443. 84. Dai, X. et al. (2014) Solution-processed, high-performance light-emitting diodes based on quantum dots. Nature, 515 (7525), 96–100. 85. Yang, Y. et al. (2015) High-efficiency light-emitting devices based on quantum dots with tailored nanostructures. Nature Photonics, 9, 259–266.

Quantum Dots for Displays and Solid State Lighting

86. 87.

88. 89. 90. 91.

92. 93. 94.

95. 96. 97.

98. 99.

100.

101.

102.

87

Mashford, B.S. et al. (2013) High-efficiency quantum dot light-emitting devices with enhanced charge injection. Nature Photonics, 7, 407–412. Colvin, V.L., Schlamp, M.C. and Alivisatos, A.P. (1994) Light-emitting diodes made from cadmium selenide nanocrystals and a semiconducting polymer. Nature, 370, 354–357. Supran, G.J. et al. (2013) QLEDs for displays and solid-state lighting. MRS Bulletin, 38 (9), 703–711. Coe, S. et al. (2002) Electroluminescence from single monolayers of nanocrystals in molecular organic devices. Nature, 420, 800–803. Caruge, J.M. et al. (2008) Colloidal quantum-dot light-emitting diodes with metaloxide charge transport layers. Nature Photonics, 2, 247–250. Mashford, B.S. et al. (2010) All-inorganic quantum-dot light-emitting devices formed via low-cost wet-chemical processing. Journal of Materials Chemistry, 20, 167–172. Wood, V. et al. (2009) Selection of metal oxide charge transport layers for colloidal quantum dot LEDs. ACS Nano, 3 (11), 3581–3586. Shen, H. et al. (2015) High-efficiency, low turn-on voltage blue-violet quantum-dotbased light-emitting diodes. Nano Letters, 15 (2), 1211–1216. Ji, W. et al. (2015) Highly efficient and low turn-on voltage quantum dot lightemitting diodes by using a stepwise hole-transport layer. ACS Applied Materials and Interfaces, 7 (29), 15955–15960. Kwak, J. et al. (2012) Bright and efficient full-color colloidal quantum dot lightemitting diodes using an inverted device structure. Nano Letters, 12 (5), 2362–2366. Manders, J.R. et al. (2015) High efficiency and ultra-wide color gamut quantum dot LEDs for next generation displays. Journal of the SID, 23 (11), 523–528. Lee, K.-H. et al. (2014) Over 40 cd/A efficient green quantum dot electroluminescence device comprising unique large-sized quantum dots. ACS Nano, 8 (5), 4893–4901. Lee, K.-H. et al. (2013) Highly efficient, color-pure, color-stable blue quantum dot light-emitting devices. ACS Nano, 7 (8), 7095–7302. (a) Castan, A. et al. (2014) All-solution processed inverted quantum-dot lightemitting diodes. ACS Applied Materials and Interfaces, 6 (4), 2508–2515; (b) Kim, H.-M. et al. (2015) Inverted quantum dot light emitting diode using solution processed p-type WOx doped PEDOT:PSS and Li doped ZnO charge generation layer. ACS Applied Materials and Interfaces, 7 (44), 24592–24600. Lim, J. et al. (2013) Highly efficient cadmium-free quantum dot light-emitting diodes enabled by the direct formation of excitons within InP@ZnSeS quantum dots. ACS Nano, 7 (10), 9019–9026. Ippen, C. et al. (2015) Color tuning of indium phosphide quantum dots for cadmiumfree quantum dot light-emitting devices with high efficiency and color saturation. Journal of the Society for Information Display, 23 (7), 285–293. Yang, X. et al. (2012) Full visible range covering InP/ZnS nanocrystals with high photometric performance and their application to white quantum dot light-emitting diode. Advanced Materials, 24 (30), 4180–4185.

88

Materials for Solid State Lighting and Displays

103. Wang, A. et al. (2015) Bright, efficient, and color-stable violet ZnSe-based quantum dot light-emitting diodes. Nanoscale, 7, 2951–2959. 104. Kim, J.-H. (2015) Fabrication of a white electroluminescent device based on bilayered yellow and blue quantum dots. Nanoscale, 7, 5363–5370. 105. Bae, W.K. et al. (2014) R/G/B/Natural white light thin colloidal quantum dot-based light-emitting devices. Advanced Materials, 26 (37), 6387–6393. 106. Anikeeva, P.O. et al. (2007) Electroluminescence from a mixed red-green-blue colloidal quantum dot monolayer. Nano Letters, 7 (8), 2196–2200. 107. Lee, K.-H. (2015) Highly efficient, color-reproducible, full-color electroluminescent devices based on red/green/blue quantum dot-mixed layer. ACS Nano, 9 (11), 10941–10949. 108. Talapin, D.V. et al. (2010) Prospects of colloidal nanocrystals for electronic and optoelectronic applications. Chemical Reviews, 110, 389–458. 109. Lim, J. et al. (2014) Influence of shell thickness on the performance of light-emitting devices based on CdSe/Zn1-XCdXS core/shell heterostructured quantum dots. Advanced Materials, 26 (47), 8034–8040. 110. Galland, C. et al. (2012) Lifetime blinking in nonblinking nanocrystal quantum dots. Nature Communications, 3, 1–7. 111. Bae, W.K. et al. (2013) Controlling the influence of Auger recombination on the performance of quantum dot light-emitting diodes. Nature Communications, 4, 1–7. 112. Bozyigit, D. and Wood, V. (2013) Challenges and solutions for high-efficiency quantum dot-based LEDs. MRS Bulletin, 38 (9), 731–736. 113. Kim, T.-H. et al. (2011) Full-colour quantum dot displays fabricated by transfer printing. Nature Photonics, 5, 176–182. 114. Cho, H. et al. (2015) Soft contact-transplanted nanocrystal quantum dots for light-emitting diodes: effect of surface energy on device performance. ACS Applied Materials and Interfaces, 7, 10828–10833. 115. Kong, Y.L. et al. (2014) 3D printed quantum dot light-emitting diodes. Nano Letters, 14, 7017–7023. 116. Masuichi, M., et al., 30.1 New nozzle-printing method for large-size organic EL devices, SID Display Week 2005 Digest, 2005, 1192-1195. 117. Chen, P.-Y. (2014) 65-inch inkjet printed organic light-emitting display panel with high degree of pixel uniformity, SID Display Week. Digest, 2014, 396–398. 118. Universal Display Corporation, Phosphorescent (PHOLED) Lifetimes published as of 2012. Available at http://www.udcoled.com/default.asp?contentID=604. Accessed May 1, 2016. 119. United States Department of Energy Critical Materials Strategy for 2011, December 2011. Available at http://energy.gov/sites/prod/files/DOE_CMS2011_FINAL_Full .pdf. Accessed May 1, 2016. 120. McKittrick, J. and Shea-Rohwer, L.E. (2014) Review: Down conversion materials for solid-state lighting. Journal of the American Ceramic Society, 97 (5), 1327–1352. 121. Hung, P.-C. and Tsao, J.Y. (2013) Maximum white luminous efficacy of radiation versus color rendering index and color temperature: exact results and a useful analytic expression. Journal of Display Technology, 9 (6), 405–412.

Quantum Dots for Displays and Solid State Lighting

89

122. Rice, W.D. et al. (2014) Magneto-optical properties of CuInS2 nanocrystals. The Journal of Physical Chemistry Letters, 5 (23), 4105–4109. 123. Chen, J. et al. (2014) All solution-processed stable white quantum dot light-emitting diodes with hybrid ZnO@TiO2 as blue emitters. Scientific Reports, 4, 1–6. 124. Chen, H.-S. et al. (2006) InGaN-CdSe-ZnSe quantum dots white LEDs. IEEE Photonics Technology Letters, 18 (1), 193–195. 125. United States Department of Energy Solid-State Lighting R&D Plan 2015, May 2015. Available at http://energy.gov/eere/ssl/downloads/solid-state-lightingrd-plan or http://energy.gov/sites/prod/files/2015/06/f22/ssl_rd-plan_may2015_0.pdf. Accessed May 1, 2016. 126. David, W. and Ohno, Y. (2010) Color quality scale. Optical Engineering, 49 (3), 033602-1–033602-16. 127. Lighting, G.E. General Electric, TriGainTM Phosphor: Simple, high-performance red for led backlighting, Corporate white paper, Document number GE 2015 MISC020 6/2015. Available at http://www.gelighting.com/LightingWeb/na/images/MISC020GE-TriGain-Phosphor-for-LED-Screens-Whitepaper_tcm201-100804.pdf. Accessed May 1, 2016. 128. Choi, J.I. et al. (2013) Electrophoretic deposition of phosphors for white solid state lighting using near UV-emitting LEDs. ECS Journal of Solid State Science and Technology, 2 (7), R153–R159. 129. Zhou, J., Yang, Y. and Zhang, C.-Y. (2015) Toward biocompatible semiconductor quantum dots: from biosynthesis and bioconjugation to biomedical application. Chemical Reviews, 115 (21), 11669–11717. 130. Bruchez, M. Jr., et al. (1998) Semiconductor nanocrystals as fluorescent biological labels. Science, 281, 2013–2016. 131. Han, M.Y. et al. (2001) Quantum-dot-tagged microbeads for multiplexed optical coding of biomolecules. Nature Biotechnology, 19 (7), 631–635. 132. Wegner, K.D. and Hildebrandt, N. (2015) Quantum dots: bright and versatile in vitro and in vivo fluorescence imaging biosensors. Chemical Society Reviews, 44, 4792–4834. 133. Foreman, J.V. et al. (2006) Time-resolved investigation of bright visible wavelength luminescence from sulfur-doped ZnO nanowires and micropowders. Nano Letters, 6 (6), 1126–1130. 134. Oey, C.C. et al. (2006) Polymer-TiO2 solar cells: TiO2 interconnected network for improved cell performance. Nanotechnology, 17 (3), 706–713. 135. Dahan, M. et al. (2001) Time-gated biological imaging by use of colloidal quantum dots. Optics Letters, 26 (11), 825–827. 136. Medintz, I.L. et al. (2003) Self-assembled nanoscale biosensors based on quantum dot FRET donors. Nature Materials, 2 (9), 630–638. 137. Zhang, Y. et al. (2006) Time-dependent photoluminescence blue shift of the quantum dots in living cells: Effect of oxidation by singlet oxygen. Journal of the American Chemical Society, 128 (41), 13396–13401. 138. Wolcott, A. et al. (2006) Silica-coated CdTe quantum dots functionalized with thiols for bioconjugation to IgG proteins. Journal of Physical Chemistry B, 110 (11), 5779–5789.

90

Materials for Solid State Lighting and Displays

139. He, X. and Ma, N. (2014) An overview of recent advanced in quantum dots for biomedical applications. Colloids and Surfaces B: Biointerfaces, 124, 118–131. 140. Smith, A.M. et al. (2006) Multicolor quantum dots for moleculr diagnostics of cancer. Expert Review of Molecular Diagnostics, 6 (2), 231–244. 141. Boldt, K. et al. (2006) Comparative examination of the stability of semiconductor quantum dots in various biochemical buffers. Journal of Physical Chemistry B, 110 (5), 1959–1963. 142. Hong, G. et al. (2012) In vivo fluorescence imaging with Ag2 S quantum dots in the second near-infrared region. Angewandte Chemie International Edition, 51 (39), 9818–9821.

3 Color Conversion Phosphors for Light Emitting Diodes Jack Silver, George R. Fern and Robert Withnall(deceased) Brunel University London, Uxbridge, UK

3.1

Introduction

In this chapter we have not sought to present an exhaustive review of the literature, but rather to present the reader with an overview of the problems that currently need to be addressed by the light emitting diode (LED) and phosphor manufacturers. To do this we will focus on some of the major phosphors being developed by some of the major industrial players and their present drawbacks as color conversion materials. First we will present a brief history of LED development and the needs for color conversion phosphors. We will then discuss the photoluminescent phosphors that are currently being developed. “Electroluminescence” was first reported by Henry Joseph Round in 1906, while he was experimenting with SiC (carborundum), and the first LED was effectively fabricated. Little progress followed in the next 50 years until concerted research into semiconductors began. In the 1950s, academic and industrial researchers carried out many experiments to generate light emission at the p-n junction of diodes using a number of materials. The results were disappointing as light emission was low in intensity, the LEDs were expensive and only a few colors were available. In his book, Light Emitting Diodes, Schubert [1] documents many of the historical and technical aspects of the progress of LED research, and this provides greater detail on the evolution of LEDs. Only a brief synopsis is given here to put this review of phosphors for LEDs into context with the still existing shortcomings of LEDs. During the early 1960s, Texas Instruments sold 870 nm LEDs for US$130, and GE launched red LEDs for around US$260. Hence these early LEDs were expensive and sold only in low volumes. In 1964, IBM used GaAsP LEDs on circuit boards in an

Materials for Solid State Lighting and Displays, First Edition. Edited by Adrian Kitai. © 2017 John Wiley & Sons Ltd. Published 2017 by John Wiley & Sons Ltd.

92

Materials for Solid State Lighting and Displays

early mainframe computer as on–off indicator lights: this was perceived as a significant breakthrough as it provided lighting in a new role to both manufacturer and end-user. The beauty of this innovation was that the LED could be mounted directly onto the circuit board, it used little power, and had a long lifetime thereby eliminating maintenance. These attributes are the ones used today to manoeuvre LEDs into replacing other light sources as the LEDs have become more efficient, cheaper, and brighter. Thus, although infra-red and red LEDs have been around for over 40 years, it was only the arrival of bright blue LEDs that heralded the genesis of the present solid state lighting (SSL) revolution. Research that led to the development of blue LEDs followed from reports on III–V nitrides that allowed the production of high-quality crystals of GaN, AlGaN, GaInN, and of p-type conduction in GaN and AlGaN [2, 3]. In the early 1990s, Nakamura et al. followed up this work to develop low resistance p-type GaN and AlGaN using Mg doping, which were produced after growth and thermal annealing in N2 (which facilitated the reactivation of the Mg acceptors) [4]. This then allowed the bright blue GaInN/AlGaN LEDs to be prepared, along with blue-green LEDs [5–9]. This was followed with the fabrication of bluish-purple laser diodes [10–18]. These innovations by Nakamura et al. led to advances in InGaN LED technology, which were to initiate the wide-scale commercialization of blue and green solid state sources as well as the development of white LEDs. A full range of colored LEDs was then available and this facilitated their use in many colored lighting applications displacing more traditional lighting methods. Around 1997, prototype white SSL was demonstrated at international conferences based on using red, green and blue LEDs. However, it was rapidly realized that the three different LEDs would age at different rates, causing the white light to take on colored hues. In addition, the color rendering index (CRI) that could be achieved from three narrow-band-emitting LEDs was poor and left much to be desired. A further problem was that the light was not bright, as the green LEDs were not bright (Figure 3.1). It was then proposed that better white light could be produced from a combination of a blue LED and either two wide-band phosphors (green and red) each excited by the blue, or by a combination of a blue LED and a broad-band yellow-emitting phosphor. These phosphor-converted LEDs (pcLEDs) have been rapidly developed over the last few years and are being widely used in a broad variety of lighting applications. Here the development of the phosphors used in the pcLEDs will be reviewed, the properties desired in these phosphors will be discussed, and the problems that have yet to be overcome to achieve them will be outlined. Since first writing this chapter in 2007 LED lighting has made incredible inroads into first commercial and industrial lighting, into white goods and motor vehicles as well as into aircraft and other moving platform applications. In the last few years it has become a major force in domestic lighting though many of the challenges set out in this chapter have not yet been solved. In the last few years the cost and supply of rare earth elements has been a problem with supplies controlled largely by China. This has led to phosphor manufacturers outside of China trying to use less rare earth metal activators in their phosphors or in fact looking for new phosphors that do not contain rare earth activators at all. In addition, many of the early LED phosphor patents are now beginning to expire or have expired, and with the fact that there are now more possible phosphors to choose from the cost of LED phosphors has been falling.

Color Conversion Phosphors for Light Emitting Diodes 60

93

(AIxGa1–x)0.52In0.48P

InxGa1–xN V(λ)

50

1

EQE (%)

40

2

30

20

10

0 350

400

450

500 550 Wavelength (nm)

600

650

700

Figure 3.1 Plot of the external quantum efficiency (EQE) of LEDs vs. emission wavelength. The data points for wavelengths ≤ 550 nm have been fitted to a polynomial function (dotted curve), and the photopic eye response function, V(x), is shown for comparison. Data for points 1 and 2 were reported in 2004 by Nichia Chemical Co. and Cree, respectively; the other data have been reported by Philips Lumileds Lighting Company for high power (>1 W) LEDs @ 350 mA, 70 A ∕ cm2 and 25 ∘ C. Adapted from N.F. Gardner, ECS 210th meeting, abstract 313, Cancun, Mexico, 2006

The most important current challenges are new red phosphors for lighting applications and better green phosphors for display applications.

3.2

Disadvantages of Using LEDs Without Color Conversion Phosphors

It is obvious to ask: “Why is it necessary to convert the beautiful colors emitted by LEDs?” This question is particularly relevant when it is realized that the whole palette of colors is available (Figure 3.1 and Figure 3.2). Although it is possible to obtain a full spectrum of colors from III–V LEDs and related compounds, there are a number of problems associated with LEDs: • These include different working lifetimes of LEDs having different emission wavelengths, leading to different ageing of different color LEDs. • Varying brightness of the different wavelength emitting LEDs. This is known as the Green Window problem, and is caused by the lack of good external quantum efficiency (EQE) values in the deep green to yellow region of the spectrum (Figure 3.1). • There is a temperature dependence of the emission wavelength of some LEDs. Although for some LEDs the shift of the emission wavelength is small with temperature, for others

94

Materials for Solid State Lighting and Displays 0.9 520 530

0.8 510 Green InGaN LED

0.7

550 Green GaP

0.6

570 Green AIInGaN LED

500 0.5

590

y 0.4

620 700

490

0.3

0.2

Red GaAlAs LED

0.1 470 450

0 0

0.1

400 0.2

0.3

0.4 x

0.5

0.6

0.7

Figure 3.2 CIE chromaticity diagram showing color triangles based on some of the currently available LEDs

Internal quantum efficiency (%)

100

80

60

40

20

0 280 320 360 400 440 480 520 560 600 640 Wavelength (nm)

Figure 3.3 Highest literature values of internal quantum efficiencies as a function of emission wavelength

Color Conversion Phosphors for Light Emitting Diodes

95

it is larger. This can be readily addressed by the use of a phosphor which has an excitation band that is wider than the emission band of the LED. Even if the LED emission undergoes a wavelength shift with temperature, the excitation band of the phosphor may span the change so that the phosphor’s emission band is not wavelength sensitive or at least much less so than that of the LED. Where the highest reported values [19–24; C. McAleese and C. Humphries, personal communication, 2006] of internal quantum efficiencies are plotted against wavelength for blue and near UV emitting LEDs, it is apparent that the highest efficiencies are in the 380–410 nm range (Figure 3.3). This can be compared with Figure 3.1; allowing for the fact that the latter is based on EQE values, it appears that in addition to the green window there is also a problem with LED performance at wavelengths shorter than 360 nm in the near ultraviolet region of the electromagnetic spectrum.

3.3 3.3.1

Phosphors for Converting the Color of Light Emitted by LEDs General Considerations

It can be seen from the foregoing discussion, that there are several reasons for needing phosphor color converters. Phosphors are a long established technology, with many display and lighting applications based on their use. Although there are a large number of phosphors that emit light in the visible region of the spectrum, there are still relatively only a few photoluminescent phosphors that can be excited efficiently by blue and near UV LEDs. This means that there are presently a limited set of colors and spectral widths available, but there is much scope for further research into new host lattices to extend the color palette. The use of phosphors will, of course, involve loss of energy due to a Stokes shift, arising from the fact that the energy of a photon of emitted light is less than that of a photon of the exciting light. The energy difference is known as the quantum deficit. Properties required by the phosphor for specific use with LEDs depend on the desired applications, which are exemplified by the following: • Conversion from blue to a specific color, for example green. • Single phosphor white, for example a blue-emitting LED in conjunction with a complementary yellow-emitting phosphor. • Multi-phosphor white, for example a UV-emitting LED in conjunction with red-, greenand blue-emitting phosphors. • The physical location of the phosphor converter, that is whether it is deposited directly onto the LED chip, for example as a suspension in a binder, or if it used remotely from the LED, for example printed on a screen. In the former case it is desirable that the phosphor emission is not thermally quenched at its operating temperature (which is higher than room temperature due to heating by the LED itself), whereas in the latter case the operating temperature of the phosphor is close to room temperature so thermal quenching is not an appreciable problem. 3.3.2

Requirements of Color Conversion Phosphors

When producing white light from an artificial source, it is important to match the light to the requirements of the user. If the white light is required to view colored objects in

96

Materials for Solid State Lighting and Displays

a commercial or domestic situation, then the better the CRI of the light the better the consumer will appreciate it. However, if the light is only to illuminate a black and white page of a book for reading, then color rendering is not so important. For 100% color rendering, the spectral output of the light needs to match the blackbody radiation of the Sun. This is difficult to achieve in practice and CRIs of 80% or more are usually acceptable for domestic use. In specific cases, such as light for illuminating a medical operation, the higher the CRI the better, as in most cases the surgeon wants to see as much contrast and detail of internal tissues as possible. Where the SSL may be replacing existing lighting for navigation lights on aeroplanes or in backlights for military displays, the properties of the lights have previously been defined and the customer will only accept SSL that matches the previously accepted specific color co-ordinates. However, in most cases, the customer will also require higher brightness. These are great challenges for the phosphor developer/manufacturer, as usually new phosphors have to be designed. The properties required by the phosphor may include the following attributes: • Its excitation spectrum must include bands that overlap the exciting emission from the blue LED. • Its emission should be in the desired part of the visible spectrum. • It should have high quantum efficiency when emitting light of the desired wavelength. • It needs to have its grain size optimized to reduce/eliminate scattering. • It should be chemically stable over its entire operating temperature range and over the lifespan of the pcLED. This might mean that if the host lattice of the phosphor is a metal sulfide or oxysulfide then it has to be protected with an oxide layer. Ideally it will be stable to at least 200 ∘ C. • Its emission intensity in the visible should not be drastically temperature dependent over the operating temperature range. Many LEDs that may have the phosphor applied in the LED package in front of the emitting surface reach temperatures of 125 ∘ C or more, and the phosphor must be an efficient emitter at such temperatures. Ideally it will be stable to at least 200 ∘ C. • It should have a good particle morphology and narrow particle size distribution for easy incorporation into silicones or epoxy resins. • For many of these properties, it is necessary to optimize the phosphor composition and crystallinity. • Low production cost would also be useful. • IP protection can also be desirable when commercialization is under consideration. For white light a choice of the wavelength of the blue exciting LED needs to be made. Currently the maximum efficiency reported is at 410 nm (Nichia Chemical Company conference presentation at the International Workshop on Nitride Semiconductors 2004, Pittsburgh, USA). However, depending on the phosphor(s) used, other blue wavelengths may be preferable to achieve a desired CRI and color temperature. The choice of the wavelength of the blue emission may ultimately be dictated by the wavelength range of the excitation spectra of the phosphor converter(s), or may be a compromise between this and the emission wavelengths of the most efficient exciting blue LEDs available. Therefore, it should be stated that not all the light emitting phosphors that

Color Conversion Phosphors for Light Emitting Diodes

97

can be excited by near UV or blue radiation will be excited with equal probability by all wavelengths of the exciting light (see below). 3.3.3

Commonly Used Activators in Color Conversion Phosphors

Much of the research on phosphors for color converting LEDs has concentrated on the use of two activators, Ce3+ and Eu2+ , because of a number of the attributes of their emission properties. Advantages of Ce3+ and Eu2+ activated phosphors include: • These ions emit visible light when they undergo d → f rather than f → f electronic transitions. • The emissions arising from d → f electronic transitions are generally more intense than those originating from f → f transitions, because the former are orbitally allowed and consequently have higher oscillator strengths than the latter, which are orbitally forbidden. • The emissions arising from d → f electronic transitions are broader than those originating from f → f transitions, because the d-orbital is more sensitive to the environment of the ion than the f-orbital and the local environment of the ion changes during metal–ligand vibrations. • The emissions due to d → f electronic transitions are more wavelength tuneable than those originating from f → f transitions because, as mentioned above, the d-orbital is more sensitive to its environment than the f-orbital. Thus, wavelength tuneability of the emission can be achieved by altering the crystal field splitting of the d-orbitals of the activator ion by changing the host lattice of the phosphor. 3.3.4

Strategies for Generating White Light from LEDs

There are a number of different approaches which can be used for generating white light from LEDs or a combination of color conversion phosphors and LEDs (pcLEDs). The approach that will be used will depend largely on the application, because there is a trade-off between the CRI and luminous efficacy. Four different strategies for obtaining white light are given below, along with an assessment of their relative CRI vs. luminous efficacy attributes: (a) A combination of red, green and blue LEDs. This approach has been discussed already in Section 3.2, where it was mentioned that drawbacks arise from differential ageing of the LEDs, poor CRI of the white light (due to narrow wavelength ranges of emissions) and the “green window”. On the other hand, the luminous efficacy is relatively high, because of the absence of color conversion phosphors and the quantum deficit that they necessarily introduce. (b) A pcLED consisting of a blue LED and two color conversion phosphors, which emit in the green and red regions of the visible spectrum. This approach uses only a single LED so there is no problem with differential ageing of LEDs. Furthermore, the greenand red-emitting phosphors can have broad emission wavelength ranges, so the CRI is higher than for (a). There is obviously a quantum deficit involved in converting the blue LED emission to green and red by means of the phosphors, so the overall luminous efficacy of the pcLED is lower than for (a).

98

Materials for Solid State Lighting and Displays

(c) A pcLED consisting of a blue LED and a single color conversion phosphor, which emits in the yellow-orange region. This approach uses only a single phosphor, so the phosphor emission needs to cover a broad wavelength range in the orange-yellow region in order to achieve a CRI in the 70–80% range. Nevertheless, there are wavelength regions in the red and between the LED and phosphor emissions where there is little or no light output, so a maximum CRI of only around 80% is achievable for a pcLED with a single phosphor. Although the CRI is less than what can be obtained for (b), the luminous efficacy can be higher because the quantum deficit is lower than for (b), where there is a large loss in energy in converting blue light into red. Consequently, pcLEDs with a single phosphor do not have sufficiently high CRIs for them to replace incandescent bulbs or fluorescent lights for general white light applications, but they are used instead for niche applications, for example for car headlamps. (d) A pcLED consisting of a UV LED and three phosphors, which emit in the blue, green and red regions. This “color by neutral” approach can achieve a higher CRI than (a)–(c), because the blue, green and red phosphor emissions can cover the whole visible region. However, there is a trade-off in luminous efficacy, which is lower than for (a)–(c), on account of the high quantum deficit in converting the UV radiation into visible light. 3.3.5

Outstanding Problems with Color Conversion Phosphors for LEDs

Although the general approach of using color conversion phosphors in combination with LEDs has been widely adopted, there are still outstanding problems with the phosphors that need to be addressed: • The excitation spectrum of the color conversion phosphor needs to overlap the emission wavelength range of the pump LED. At present there are a number of phosphors that can be excited in the blue/violet and emit efficiently at longer wavelengths in the visible (see Section 3.4). • The particle sizes of the phosphors need to be optimized in order to reduce light loss by scattering. • The light that is emitted back towards the LED chip needs to be re-directed in the forward direction in order to maximize the light output from the pcLED. • The LED emission is angle dependent, whereas the phosphor has an emission that is close to Lambertian. Thus phosphor coatings need to be deposited on the LED chip to minimize any angle dependent shift of the color point of the pcLED. • The phosphor needs to be deposited homogeneously on the LED chip or other remote substrate in a controlled way. • The binder in which the phosphor is suspended should not be photo-degraded by the LED light. Many epoxy resins yellow over time when illuminated by blue LED light and, even more so, by UV LED light. Silicone type binders are often used in preference due to their better photo-stability. • Y3 Al5 O12 ∶Ce (YAG:Ce) is commonly used for pcLEDs with a single phosphor, but it has a high color correlated temperature (CCT); other efficient, broad-band-emitting phosphors with lower CCTs are required for white light pcLEDs that have a “warmer” white in order to satisfy consumer preferences. It is now useful to review some individual phosphors, and discuss their properties.

Color Conversion Phosphors for Light Emitting Diodes

3.4

99

Survey of the Synthesis and Properties of Some Currently Available Color Conversion Phosphors

3.4.1

Phosphor synthesis

The synthesis of inorganic phosphors has been the subject of intensive research for well over 80 years. Much of the work has been carried out in industry and not in the public domain. Thus, literature on the manufacturing methods for industrial phosphors tend to be confined to specialist books rather than research journals, and the reader is referred to the well-known texts by Ropp [25, 26] that contain preparative details. Almost all were synthesized by solid state reactions between very pure inorganic compounds at high temperature. It is impossible here to discuss every aspect of synthesis from its inception to the present day, so priority has been given to discoveries and developments appearing in the last 5 years to new phosphors developed specifically for use as color converters for LEDs. The structures of most of the host lattices used for phosphors have been described and correlated by Wells [27] and others [28], and the host lattice will, of course, determine the coordination environment of the dopant guest, which can influence its emission behavior. The general properties of luminescent materials have been discussed in several texts [25, 26, 29, 30]. 3.4.2 3.4.2.1

Metal Oxide Based Phosphors Yttrium Aluminum Garnet Activated by Trivalent Cerium (YAG : Ce3+ or Y3 Al5 O12 : Ce3+ )

The first mention of this phosphor was in 1967 as a cathodoluminescent phosphor, though in the same paper its photoluminescent properties are described [31]. The use of this phosphor and many others for fluorescent lighting was reviewed in 1987 [32]. Methods of making this phosphor have been well documented elsewhere [25, 26]. This phosphor is referred to in many patents for use as a color converter using blue LED excitation and emitting yellow light, which when combined with more blue light from the LED gives a passable white light. Although many are referenced here, no attempt has been made to find them all [33–44]. In the last 10 years many of the cheaper “white” LEDs that have appeared for the commercial domestic market have been based on blue LEDs with YAG∶Ce3 type garnet phosphor as the color convertor. This combination of blue LED and YAG : Ce color converter to generate white light is now both cheap and acceptable, providing there is no requirement for a high CRI. The excitation spectrum of YAG∶Ce3+ has its peak at around 470 nm (Figure 3.4). In the same figure the emission spectrum is shown to be asymmetrical as it is made up of two transitions 5 D → 2 F5∕2 and 5 D → 2 F7∕2 ; both the asymmetry and the position of this band varies, depending on the Ce3+ concentration, though the center of the peak is usually around 550 nm. Also, Figure 3.4 shows the optimum position of an exciting blue emission (as from a blue LED emitting at 470 nm) [45–47]. In Figure 3.5 the position of the color coordinates of YAG∶Ce3+ are shown at point 7, whereas those of the exciting blue LED light are shown at point 1. Points 2–6 are generated by putting layers of YAG∶Ce3+ phosphor particles on glass and exciting them with blue light so some blue light also comes through the phosphor

100

Materials for Solid State Lighting and Displays

Emission

Intensity

Excitation

400

600 500 550 Wavelength (nm)

450

650

700

Figure 3.4 The excitation and emission spectra of YAG∶Ce3+ . The hatched peak depicts an exciting emission from a blue LED emitting at 470 nm. The match of the exciting blue radiation to the excitation band of the phosphor is excellent

520

530

0.8

540

510

550 560 0.6 500 y 0.4 + 3

490 0.2 480

+ + 6 5 + 4

+ 570 7

580 590 600 620

+ 2

+ 470 1

0 420 0

0.2

0.4 x

0.6

0.8

Figure 3.5 CIE chromaticity diagram showing the position of the emission of YAG∶Ce3+ phosphor (point 7) and the emission of a blue 470 nm LED (point 1). The other points (points 2–6) are explained in the text. The continuous curved line in the center of the diagram is known as the Planckian, which is the locus of blackbody emitters

Color Conversion Phosphors for Light Emitting Diodes

101

layer along with the yellow emission. Point 2 is created by one layer of YAG∶Ce3+ particles, point 3 is created by two layers and so on to point 6, which contains five layers [45–47]. Clearly, in Figure 3.5, points 3 and 4 are close to the white color point (x = 0.33,y = 0.33) and a line between these points represents the best “whites” that can be generated from a blue LED emitting at 470 nm and YAG∶Ce3+ phosphor particles. To improve these whites it is necessary to chemically modify the YAG∶Ce3+ phosphor particles. The line between points 1 and 7 is known as the mixing line for varying phosphor thickness and blue LED light [45–47]. 3.4.2.2

Yttrium Aluminum Garnet Co-activated by Trivalent Cerium and Praseodymium (YAG : Ce3+ , Pr3+ or Y3 Al5 O12 : Ce3+ , Pr3+ )

Intensity

There are several ways this can be done, as shown in Figure 3.6 [46]. It can be seen that by adding Pr3+ cations to YAG∶Ce3+ , an emission peak around 620 nm (in the red) is added to the emission spectrum of YAG∶Ce3+ . Also by adding some Gd3+ to replace some Y3+ , this moves the main YAG∶Ce3+ emission towards the red and as seen in the same figure; both these effects can be engineered in the same phosphor particles [46–48]. The color coordinates generated by the emission from the phosphors shown in Figure 3.6 are shown in the CIE chromaticity diagram in Figure 3.7 [48]. The color coordinates generated by the emission from the phosphors shown in Figure 3.6 are also shown in the CIE chromaticity diagram in Figure 3.7. The solid line crossing the Planckian from point 3 to point 4 (in Figure 3.7) shows how a real white can now

450

500

550

600 650 700 Wavelength (nm)

750

800

850

Figure 3.6 The emission spectra of YAG∶Ce3+ (thin black line), Y3 Al5 O12 ∶Ce3+ , Pr3+ (gray line) and (Y, Gd)3 Al5 O12 ∶Ce3+ , Pr3+ (thick black line). Adapted from Silver et al. 2005 [46] and Silver 2006 [48]

102

Materials for Solid State Lighting and Displays 520

530

0.8

540

510

550 560 0.6

+ 570 1++ 2 3 580

500

590 600

y 0.4

620 490 0.2 480

+ 470 4

0 420 0

0.2

0.4 x

0.6

0.8

Figure 3.7 CIE color chromaticity diagram showing the position of the emission of YAG∶Ce3+ phosphor (point 1), the emission of Y3 Al5 O12 ∶Ce3+ , Pr3+ (point 2) and of (Y, Gd)3 Al5 O12 ∶Ce3+ , Pr3+ (point 3). The emission of a blue 470 nm LED is also shown (point 4). The solid line crossing the Planckian from point 3 to point 4 shows how a real white can now be generated from the blue LED and the (Y, Gd)3 Al5 O12 ∶Ce3+ , Pr3+ phosphor. Adapted from Silver 2006 [48]

be generated from the blue LED and the (Y, Gd)3 Al5 O12 ∶Ce3+ , Pr3+ phosphor [48]. The particular white can be described by its color coordinates or by its color temperature. An empirical rule has been put forward for Ce3+ luminescence in Y3 Al5 O12 ∶Ce3+ : Increasing the diameter of the ion on the dodecahedral (Y3+ ) site increases the crystal field splitting, while increasing the diameter on the octahedral (Al3+ ) site has the reverse effect [49, 50]. Thus on the dodecahedral site there is a wavelength shift on going from 550 nm for Y3+ to 585 nm for the larger Gd3+ ion, and a shift from 550 nm for Y3+ to 510 nm for the smaller Lu3+ ion. This is apparent for the (Y, Gd)3 Al5 O12 ∶Ce3+ , Pr3+ in Figure 3.6. In contrast, when the cation is located on the octahedral site, there is a wavelength shift from 550 nm for Al3+ to 505 nm for the larger Ga3+ ion. However, substituting Ga3+ for Al3+ reduces the emission intensity. The basic YAG:Ce lattice can accommodate more than 60 different elements so many different combinations are possible. In order to consider some of the issues with blue LEDs exciting phosphors such as YAG∶Ce3+ , it is worth considering the calculation and measurement of some luminous efficacies.

Color Conversion Phosphors for Light Emitting Diodes

3.4.2.2.1

103

Theoretical Limit of Luminous Efficacies of YAG:Ce Phosphors

The luminous efficacy, 𝜂LE, for the conversion by the YAG:Ce phosphor of blue (470 nm) LED light into yellow emission is given by [51, 52]: 𝜂LE = 𝜂LO × 𝜂QD × 𝜂QE × 𝜂SC where 𝜂LO is the lumen equivalent, 𝜂QD is the quantum deficit. 𝜂QE is the quantum efficiency, and 𝜂SC is the screening efficiency. The quantum deficit is ca. 0.80 and the quantum efficiency is ca. 0.95.Thus: 𝜂LE = 456 × 0.80 × 0.95 × 𝜂SC = 347 ⋅ 𝜂SC lm∕W The luminous efficacies of YAG:Ce phosphor powders under 470 nm and 430 nm excitation were measured as a function of Ce concentration and also of firing temperature. Clearly, the luminous efficacies for YAG:Ce phosphors are Ce activator concentration dependent (Figure 3.8). Plotting the Ce activator concentration for YAG : Ce phosphors vs. luminous efficacy (see Figure 3.9, in which 470 and 430 nm excitations are compared), it is apparent that the luminous efficacies are also wavelength dependent [51–54]. The lower values for the 430 nm excitation are mainly due to the quantum deficit caused by the wavelength difference between the excitation and emission wavelengths. Hence for YAG:Ce phosphors the luminous efficacy values are concentration and excitation wavelength dependent [51–54]. The luminous efficacies also reach maximum values in the YAG:Ce phosphors at a firing temperature of around 1500 ∘ C, as seen in Figure 3.10 (where they are plotted for two different excitation wavelengths). 260

Luminous efficacy (lm/W)

250

240

Hence concentration dependent

230

220 210 1.5

Figure 3.8 tation)

2.0

2.5

3.0 3.5 4.0 mol% Ce in YAG

4.5

5.0

Plot of Ce activator concentration in YAG:Ce vs. luminous efficacy (470 nm exci-

104

Materials for Solid State Lighting and Displays 470 nm 430 nm

260 250

Luminous efficacy (lm/W)

240 230 220 210 200 190 180 170 1.5

2.0

2.5

3.0 3.5 mol% Ce

4.0

4.5

5.0

Figure 3.9 Plot of Ce activator concentration in YAG:Ce vs. luminous efficacy (comparison of 470 and 430 nm excitation) 220 200

Luminous efficacy (lm/W)

180 160 140 120 100 80 60 40 20 800

900

1000

1100 1200 1300 1400 Firing temperature (°C)

1500

1600

1700

Figure 3.10 Plot of firing temperature vs. luminous efficacy of YAG:Ce phosphors (▾ = 430 nm and ▾ = 470 nm excitation)

Color Conversion Phosphors for Light Emitting Diodes

320

470 nm 430 nm

300 Luminous efficacy (lm/W)

105

280 260 240 220 200 180 160 0

1

2

3

4

5

mol% Ce

Figure 3.11 Plot of Ce activator concentration in YAG:Ce phosphors vs. luminous efficacy (extended down to lower Ce concentrations)

At lower concentrations of Ce, the luminous efficacy approaches the theoretical value (Figure 3.11). However, phosphors made at these lower concentrations do not perform well for two reasons. The first is that more phosphor is necessary to convert the correct proportion of the incoming exciting light; and the second is that the low concentrations reflect more of the exciting radiation, as is apparent from Figure 3.12 [51, 55, 56]. Less blue light is absorbed for a Ce concentration of 0.1 mol%. 𝜂LE is higher for this Ce concentration since: 𝜂LE = no. lumens ∕ power absorbed Table 3.1 gives the luminous efficacies and quantum efficiencies for the low concentration Ce in YAG:Ce phosphors, compared with a commercial sample [51]. The values for the latter are similar to the values we obtained for 1.5 and 2.5 mol% Ce, as seen in Figure 3.11. Color points and luminous efficacies (𝜆exc = 470 nm) of YAG:Ce and YAG:Ce, Pr phosphors (Pr concentration range 1–5 mol%) are plotted in Figure 3.13. 𝜂LE values of the YAG:Ce, Pr phosphors are very poor [51]. This is probably due to one of two reasons: 1. The Pr takes the exciting light from the LED preferentially (instead of the Ce) and downconverts it less efficiently than the Ce. 2. The Pr uses the Ce emission as excitation and does not emit it efficiently [51].

106

Materials for Solid State Lighting and Displays 65

% reflectance at 470 nm

60

55

50

45

40 0

1

2 3 Ce concentration (mol%)

4

5

Figure 3.12 Percentage reflectance of YAG:Ce at 470 nm as a function of Ce concentration (mol%) in YAG:Ce phosphors Table 3.1 Luminous efficacies (LE) and quantum efficiencies (QE) for in-house synthesized YAG:Ce phosphors having different Ce concentrations, along with the values for a commercial sample Ce conc. (mol%)a (TB)0.1 0.5 1.0 Commercial sample 470 nm a The

LE (470 nm exc.) (lm/W)

QE (470 nm exc.) (%)

313 278 265 242

78 69 66 62

YAG:Ce phosphors were made via a solid state route and fired at 1500 ∘ C.

Hence adding Pr as an additional activator in Y3 Al5 O12 ∶Ce3+ to make Y3 Al5 O12 ∶Ce3+ , Pr3+ is not a sensible path to follow as the luminous efficacy falls dramatically although the CIE coordinates of the white may be better. 3.4.2.3

Yttrium/Gadolinium Aluminum Garnet Activated by Trivalent Cerium (YAG : Ce3+ or Y3 Al5 O12 : Ce3+ )

The alternate approach of preparing (Yx Gd1-x )AG∶Ce (0 ≤ x ≤ 1) phosphors is also problematic [51]. A plot of luminous flux vs. wavelength (𝜆exc = 470 nm) for the (Yx Gd1-x )AG∶Ce (0 ≤ x ≤ 1) phosphors is shown in Figure 3.14. The emission band shifts to the red on substituting Y for Gd and the emission intensity decreases for x = 0.50 and x = 0.75. No emission is observed from GdAG:Ce under 470

Color Conversion Phosphors for Light Emitting Diodes

107

300

Luminous efficacy (lm/W)

250

200

150

100

50 0

1

2 3 Pr concentration (mol%)

4

5

Figure 3.13 Plot of Pr co-activator concentration in YAG:Ce, Pr phosphors vs. luminous efficacy. The Ce concentration was 2.4 mol% in each case

(e)

Luminous flux × 10–3 (lm/nm)

(a)

0.0005

(d)

(a) 550 nm (b) 560 nm (c) 565 nm (d) 565 nm

(e)

0.0000 300

400

500 600 Wavelength (nm)

700

800

Figure 3.14 Plots of luminous flux vs. wavelength for (Yx Gd1-x )AG∶Ce phosphors with x equal to (a) 1.00, (b) 0.75, (c) 0.50, (d) 0.25, and (e) 0.00. The Ce concentration was 2.45 mol% and the excitation was centered on a wavelength of 470 nm (seen in the figure due to partial reflection by the phosphors)

108

Materials for Solid State Lighting and Displays 300

Luminous efficacy (lm/W)

250

200

150

100 0

Figure 3.15 phors

20

40 60 %Gd in (Y,Gd)AG lattice

80

100

Luminous efficacies (𝜆exc = 470 nm) of (Yx Gd1−x )AG∶Ce (0.25 ≤ x ≤ 1) phos-

nm excitation (Figure 3.14 and Figure 3.15). To investigate why the latter material does not emit, we examined the excitation bands of these phosphors [51, 57]. From Figure 3.16 it is apparent that all the (Yx Gd1-x )AG∶Ce (0.25 ≤ x ≤ 1) phosphors have an excitation band centered on ca. 470 nm, which does not extend below ca. 400 nm [51, 57]. It is evident that the drop in luminous efficacy, as Gd is progressively substituted into the lattice, does not affect the excitation spectra. The most likely explanation is that as Gd is added, the energy gap between the valence band and conduction band decreases so that for the pure GdAG:Ce the excitation promotes the electron from the Ce directly into the conduction band, and this relaxes nonradiatively [51, 57]. Using the phosphor system based on (Y1-a Gda )3 Al5 O12 ∶Ce3+ (where a is in the range 0.3–0.6) on screens in combination with two blue LEDs allows specific white points to be targeted [58]. The temperature dependence of YAG:Ce phosphors prepared by a solid state mixing method are presented in Figure 3.17. It is apparent that there is a drop in efficacy with temperature that is roughly parallel for the YAG phosphors presented by this method for all the Ce concentrations studied. In all but the lowest Ce concentration the effect of a rise of 160 ∘ C is a drop in efficacy of around 15%. This averages to a reduction in efficacy of ca. 1% with every 10 ∘ C rise [51]. However, if the YAG:Ce is prepared in other ways, such as being fired in a reducing atmosphere, then the temperature dependence can be different, as seen in Figure 3.18 [51]. In this case the YAG:Ce, which starts with a significantly higher efficacy, drops by 38% in the temperature range of 30–190 ∘ C. This is around a 2.5% drop in efficacy for every 10 ∘ C rise, although the drop is less (1.5% for each 10 ∘ C rise) in the range 30–100 ∘ C.

Color Conversion Phosphors for Light Emitting Diodes

109

1.2

Intensity (arbitrary units)

1.0

0.8

0.6

0.4 (d) (c)

0.2

(b) (a)

0.0 150

200

250

300

350

400

450

500

550

Wavelength (nm)

Luminous efficacy (lm/W)

Figure 3.16 Observed excitation spectra of (Yx Gd1−x )AG∶Ce (0.25 ≤ x ≤ 1) phosphors when monitoring emission at 550 nm. (a) x = 1.00, (b) x = 0.75, (c) x = 0.50, and (d) x = 0.25

280

0.1 mol% Ce 0.5 mol% Ce 1.0 mol% Ce

260

1.6 mol% Ce 2.4 mol% Ce

240

220

200

180

160 0

Figure 3.17 mol %)

50

100 Temperature (°C)

150

200

Plot of luminous efficacy vs. temperature for YAG:Ce phosphors (Ce = 0.1–2.4

110

Materials for Solid State Lighting and Displays 280 270

Luminous efficacy (lm/W)

260 250 240 230 220 210 200 190 20

Figure 3.18

40

60

80

100 120 140 Temperature (°C)

160

180

200

Plot of luminous efficacy vs. temperature for YAG:Ce phosphor (Ce 5.0 mol%)

These temperature dependent efficacies are important as LED junction temperatures and LED arrays can become very hot when in operation and phosphors that have poor thermal properties will not be useful as LED color converters [51, 59]. So, in summary, garnet phosphors based on YAG:Ce are currently the best color converters for combining with blue LEDs to generate white light, providing there is no requirement for a high CRI. The parent YAG:Ce is a very easily substituted lattice permitting wide variation in its elemental make-up. 3.4.2.4

Silicate Garnets and Related Phosphors

Ca3 Sc2 Si3 O12 ∶Ce is excited at 446 nm and emits a broad band similar to that of YAG:Ce around 570 nm [60]. A related phosphor system is BaY2 SiAl4 O12 ∶Ce. In a patent [61], a wide range of substitution is claimed based on the host lattice general formula MLn2 QR4 O12 , where: • M is at least one element selected from the group consisting of Mg, Ca, Sr, and Ba. • Ln is at least one rare earth element selected from the group consisting of Sc, Y, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu. • Q is at least one element selected from the group consisting of Si, Ge, Sn, and Pb. • R is at least one element selected from the group consisting of B, Al, Ga, In, and Tl [61]. The luminescence of the vanadate garnet, Ca2 NaMg2 V3 O12 ∶Eu3+ has been evaluated and discussed as a potential ultraviolet light emitting diode (UV-LED) phosphor [62]. This single phosphor has been claimed to be capable of converting the UV emission of a UV-LED into white light with good luminosity and CRI. The luminescence of this material at elevated temperatures is of interest, because the junction temperatures of typical

Color Conversion Phosphors for Light Emitting Diodes

111

LEDs can be greater than 100 ∘ C. Indeed, it has been reported [62] that there is significant thermal quenching of this phosphor and an emission color shift at temperatures greater than 100 ∘ C. This has been explained as energy migration and transfer to nonradiative traps and Eu3+ within the host lattice [62]. 3.4.2.5

(Y2 − x − y Eux Biy )O3

Red phosphors based on (Y2-x Eux )O3 have been developed where the Bi is used to change the excitation spectrum by introducing absorption bands in the 340–370 nm region (due to the Bi 6s2–6s6p transition), which with the 4f - 4f Eu3+ transition, gives a continuous excitation band extending between 340 nm and 410 nm [63]. The red emission is due mainly to energy transfer from Bi3+ to Eu3+ . Compositions that were studied include Y1.84 Eu0.16 O3 , Y1.82 Eu0.16 Bi0.02 O3 , Y1.78 Eu0.16 Bi0.06 O3 , and Y1.72 Eu0.16 Bi0.12 O3 , with the latter giving the best red emission [63]. 3.4.2.6

Na2 Gd2 B2 O7 : Ce3+ , Tb3+

This is a green emitting phosphor that transfers energy from Ce3+ to Tb3+ prior to emission. The Ce3+ is thermally quenched above 100 ∘ C [64]. 3.4.2.7

YCa3 M3 B4 O15 : Eu3+

The photoluminescence properties of the red-emitting YCa3 M3 B4 O15 ∶Eu3+ (M = Al or Ga) phosphors under UV (395 nm) irradiation have been reported. A strong emission at 622 nm and several weaker peaks near this emission were observed [65]. The emission intensity increased with increasing Eu3+ content up to x = 0.75 in Y1−x Eux Ca3 M3 B4 O15 . The authors claim these phosphors are better reds than CaS∶Eu3+ for LED applications [65]. 3.4.2.8

Alkaline earth metal silicates (Ba1 − x − y Srx Cay )SiO4 : Eu2+

Barry [66] has reported that the incorporation of Eu2+ in the compounds Ca3 MgSi2 O8 , Sr3 MgSi2 O8 , and Ba3 MgSi2 O8 produces phosphors of high luminescence yield. The peak of the emission band occurs at progressively shorter wavelength and narrows as the radius of the major alkaline earth ion (Ca, Sr, or Ba) is increased. A 1200 ∘ C isotherm on compositions intermediate to Sr3 MgSi2 O8 and Ba3 MgSi2 O8 shows a complete series of crystalline solutions to exist at this temperature. Using an orthorhombic cell, the d-spacings of the (222) reflections vary continuously from 1.956 Å for Sr3 MgSi2 O8 ∶Eu2+ to 2.012 Å for Ba3 MgSi2 O8 ∶Eu2+ . However, the emission spectra of these samples vary in a discontinuous manner. In general, broadening of spectral energy distributions is observed as compositions move into the ternary section. All the compositions prepared resulted in phosphors of relatively high efficiency [66]. Following on from this work, a complete series of solid solutions, for example, in the Ba2 SiO4 − Sr2 SiO4 system with a continuous shift of luminescent color from green (𝜆max = 505 nm) to nearly orange (𝜆max = 575 nm) for the europium-doped materials, Starick et al. [67] reported interesting phosphors in the ternary (Ba1-x-y Srx Cay )SiO4 ∶Eu2+ system. In contrast to Barry’s work [66], isothermal run for all prepared phosphors, they preferred an individual optimization of the preparation conditions for every promising member of the series [66].

112

Materials for Solid State Lighting and Displays

It was found that the luminescent properties of these phosphors strongly depend on the conditions of their synthesis. Critical factors for getting high product quality include: • • • • •

a careful selection of suitable raw materials (purity, reactivity); selection of good flux materials; control of the firing atmosphere (N2 ∕ H2 ); optimization of the heating and cooling regimes; the post-treatment.

Phosphors with high crystal perfection, a favorable particle size distribution, and with improved luminescent efficiencies are claimed. Their excitation spectra are broad and cover from below 300 nm to 500 nm, while their broad-band emission spectra range from 525 nm green to 595 nm orange. These alkaline earth metal silicates of the general formula M2 SiO4 ∶Eu2+ (where M = Ba, Sr, or Ca) are color tuneable (by suitable combinations of Ba, Sr, and Ca) green to yellow phosphors. Quantum efficiencies of more than 90% are said to be feasible. They have narrow emission bands for green phosphors but they suffer from low quenching temperatures and are not stable at ambient temperatures [68–71]. We have studied factors affecting the color of the green-emitting phosphors in the system Sr2 SiO4 -Ba2 SiO4 activated by divalent europium ions; the CIE coordinates of the phosphors in this system do not lie on a single line on the CIE diagram but in fact change dramatically with Ba2+ concentration. The emission colors change from green through yellow greens in a complex way as the Ba2+ concentration decreases; we have shown that this is due to the changes in the crystal structures of these phosphors [72; J. Silver et al., unpublished results]. White light-emitting phosphors with a nominal composition of (Sr2.55 Ba0.4 )MgSi1.7O8 ∶ Eu2+ that can be excited in the near UV have been reported recently. Their emission spectra consist of two bands in the visible region (one a blue band at 460 nm and the other a yellow band at 575 nm); these are thought to originate from Eu2+ in both Sr3 MgSi2 O8 and in Sr2 SiO4 , respectively. A white LED was fabricated from combining this phosphor with a near UV InGaN chip (𝜆em = 405 nm). The LED is claimed to manifest high CRI (> 80), high color stability, very good color coordinates (CIE, x = 0.34, y = 0.33) and high luminous efficiency (14.31 lm/W) [73]. In the last 10 years such alkaline earth metal silicate phosphors have seen widespread use in LEDs, but because of stability issues may well slowly disappear from the market in the next few years. 3.4.2.9

Metal Oxide Based Phosphors Doped with Eu3+

In addition to the metal oxide based phosphors commented on in Sections 3.4.2.5 and 3.4.2.7, a wide variety of Eu3+ metal tungstate and molybdate lattices have been studied in the last 12 years as possible red phosphors for LEDs. These phosphors based on the general formula AEu(MO4 )2 (where A = Li+ , Ag+ , Na+ , K+ , or Rb+ ; M = Mo, W, or both; and Eu3+ may be substituted by Y3+ and/or Sm3+ ; and/or the formulae are related but the cations or the ratio of the cations to anions differ) are good red emitters when excited by blue LED light [74–81]. (This list of references is not comprehensive but a full list is beyond the scope of this chapter). None of these references gave luminous efficacy measurements

Color Conversion Phosphors for Light Emitting Diodes

113

up to the time of our work [82]. Such measurements were reported for some representative red phosphors of general formula LiEu2-x Yx (MoO4 )2-x (WO4 )x [82, 83]. However, these phosphors are unlikely to be used due to the cost of the europium as well as other factors [82, 83], including the development/commercialization of a Mn4+ phosphor (that does not contain an expensive rare earth element), see Section 3.4.6. 3.4.3

Metal Sulfide Based Phosphors

Metal sulfide type phosphors have been used as cathodoluminescent and electroluminescent phosphors since the 1930s [84, 85]. More modern methods of making these types of phosphors have aided their development for use as color converters for LEDs. These methods are based on precipitation of the precursor phosphor from solution, and are important for both metal sulfides and selenides [86–90]. There are two general methods. The first involves treatment of aqueous metal salts, with or without added dopant, with hydrazine hydrate and either sulfur or selenium. This method has been successfully employed to synthesize metal chalcogenide phosphors, such as ZnS:Ag, ZnSe, Zn(S,Se), CdS, ZnS:Cu, ZnS:Cu, Al, Au [72], (Zn,Cd)S:Cu and Zn(S,Se):Cu [86–88] and this method was also extended to the synthesis of ternary metal sulfide phosphors, such as Sr0.97 Ga2 S4 ∶ Eu0.3 , Sr0.985 Ga2 S4 ∶Ce0.015 , Zn0.98 Ga2 S4 ∶Mn0.02 , Ca0.8 Ga2 S4 ∶Eu0.2 and Ba2 Zn0.985 S3 ∶ Mn0.015 [87]. The second method utilized the decomposition of thiourea dioxide in aqueous solution at elevated temperatures in the presence of metal acetates, and has been used to prepare ZnS phosphors [88]. The advantage of these new methods is that the elaborate techniques necessary for the removal of toxic gases or vapors (H2 S and CS2 ) common in solid state reactions are no longer required [85]. 3.4.3.1

MS : Eu2+ (Where M = alkaline Earth Metal)

The MS∶Eu2+ (where M = Alkaline earth metal) phosphors have been explored for their use with LEDs by a number of groups [91–95]. Mixed alkaline earth metal sulfide lattices have also been studied, for example, strontium calcium sulfide activated with divalent europium (Srx Ca1-x S∶Eu), which has been reported as a red-emitting phosphor for LED white light generation. The host lattice, a solid solution of alkaline earth metal sulfides, can take various values of x between 0 and 1 [91–94]. The emission wavelength is then adjustable in the range from 635 nm, which is the pure SrS:Eu, to 655 nm, which corresponds to the pure CaS:Eu. The phosphor can be efficiently excited with blue and green light (in the range from 400 to 580 nm). The phosphors with the formulation Srx Ca1-x S∶Eu have low photoluminescent efficiencies due to a thermalization process, and are therefore not useful for blue LED excitation. However, it has been reported that co-doping in Srx Ca1-x S∶Eu with a halide ion increases the emission efficiency. These phosphors are synthesized as follows: the precursor of the phosphor, strontium calcium sulfate, is precipitated out of a strong acid solution containing the constituent metal ions by neutralizing the solution with ammonium hydroxide [93]. Wet chemical preparation of precursors: xSr(SO4 ) + (1 − x)Ca(SO4 ) + NH4 OH → Srx Ca1-x SO4 ∶Eu(OH)3 ↓ After pulverization and drying, it is then fired at 600 ∘ C for 3 h.

114

Materials for Solid State Lighting and Displays

High temperaturee reduction: Srx Ca1-x S∶Eu(OH)3 + N2 ∕ H2 → Srx Ca1-x S∶Eu2+ (at 900∘ C for 3 hours) After grinding and milling, it is then fired in H2 S at 1000∘ C for 6 h. Then, activation with halide doping [93]: Srx Ca1-x S∶Eu2+ + NH4 Cl + N2 → Srx Ca1−x S∶Eu2+ , Cl (at 1100∘ C for 1 hour) The quantum efficiency was measured to be 80%. The dependence of the emission intensity on the halide content was found to be linear in the range from 0.1 to 2%. The stability of this phosphor in a moist environment is not good. Hydrogen sulfide is liberated when the phosphor reacts with trace water. The emission efficiency is said to deteriorate to a very low level within a day [93]. Hydrolytic decomposition: SrS + H2 O + CO2 → SrCO3 + H2 S ↑ Oxidative decomposition: SrS + 2O2 → SrSO4 To counter this, a SiO2 coating was added to the phosphor particles to protect them from degrading in moist air [94]. As this phosphor can be excited by both blue and green light its use in a mixed system to generate white light with a green-emitting phosphor and a blue-emitting LED is not straightforward. Careful control of the amount of the red-emitting phosphor present is important to balance whether the phosphor will be excited by blue and/or green light. Alternatively, the red phosphor could be covered with a blue or red filter to facilitate the balance of the system. SrS:Eu submicrometer particles have been synthesized by a solvothermal method using ethylenediamine as the solvent. The reaction was carried out in a high-pressure autoclave at relatively low temperatures (< 200 ∘ C). A range of capping groups were explored and cubic crystallites in the size range 100–1000 nm were prepared [95]. It should be noted that since the use of remote phosphors where the LED is set back from the phosphor, many of these metal sulfide phosphors can be used as temperature effects are not a problem. However, as noted above, these metal sulfide based phosphors are moisture sensitive and need to be protected. We have carried out accelerated ageing studies on SiO2 coated CaS∶Eu2+ and SrS∶Eu2+ phosphors [96] and shown that the coatings offer protection but only delay the onset of degradation due to moisture rather than eliminate it. We have shown that using various proportions of CaS∶Eu3+ and SrS∶Eu3+ phosphors on screens with one blue LED it is possible to select specific red points (CIE coordinates) [97]. However, the phosphors still need to be protected from moisture on the screens. 3.4.3.2

Alkaline Earth Metal Thiogallates, MGa2 S4 : Eu2+ (Where M = alkaline Earth Metal)

Alkaline earth metal thiogallates were disclosed in the past as cathodoluminescent phosphors for display screens, and for use in electroluminescent displays [98–105]. The phosphors were synthesized by solid state reaction from alkaline earth sulfide, gallium

Color Conversion Phosphors for Light Emitting Diodes

115

sulfide, and rare earth sulfide. These phosphors manifest good saturation properties, but their emission efficiency was said to be low (around 30% of other metal sulfide phosphors). In recent years, the photoluminescent properties of strontium thiogallate activated with divalent europium (SrGa2 S4 ∶Eu2+ ) have been developed targeting high efficiency LED white light. The SrGa2 S4 ∶Eu2+ phosphor has a broad absorption band centered at 470 nm and stretches into the near UV, and it emits a green light peaking around 535 nm, as shown in Figure 3.19 and Figure 3.20. Various solution preparation methods involving homogeneous precipitation followed by reduction firing have proved to significantly improve the emission efficiency [86–90]. One preparation uses a wet chemical reaction between a soluble strontium salt (SrSO4 ) and europium nitrate in a strong acid medium [93]. This is followed by neutralization with ammonium hydroxide, which results in a suspension of strontium sulfate fine particles mixed with europium hydroxide: Sr(SO4 ) + Eu(NO3 )3 ∕ Pr (NO3 )3 + NH4 OH → SrSO4 ↓ + Eu(OH)3 ∕ Pr (OH)3 ↓ The adding of a Pr3+ is optional, depending on whether a red component is desired. The reaction temperature, the concentration of the soluble salt solution and additives such as an organic solvent, controls the particle size of the precipitates. Heating the equivalent acidic solution containing gallium makes a solution of an acid soluble gallium salt: Ga(HNO3 ) + NH4 OH → Ga(OH)3 ↓ Since gallium oxide is difficult to convert to an oxide-free sulfide with hydrogen sulfide, it is not used. The gallium nitrate solution is added to the resultant precipitates in a 150

Intensity

100

50

0 400

Figure 3.19

420

440

460 480 Wavelength (nm)

500

520

Excitation spectrum of SrGa2 S4 ∶Eu2+ with 470 nm LED emission superimposed

116

Materials for Solid State Lighting and Displays 520

530

0.8

540

510

+

550 560

0.6

570

500

580 590 600

y 0.4

620 490 0.2 480 470 0 420 0

Figure 3.20

0.2

0.4 x

0.6

0.8

Chromaticity diagram showing CIE coordinates of SrGa2 S4 ∶Eu2+

sufficient amount so as to produce an excess of gallium sulfide product. After combining these solutions and bringing the mixture to a neutral pH with ammonia, or by carrying out a precipitation of gallium using urea, a fine white powder is precipitated. This phosphor precursor is dried, ground, and fired in hydrogen sulfide at about 800 ∘ C for approximately 5 h: SrSO4 .Eu(OH)3 (2 + x)Ga(OH)3 + H2 S → SrGa2 S4 ∶Eu2+ ∶xGa2 S3 A two-step firing was reported, with the second firing being at 900 ∘ C. The resultant phosphor had an average particle size of 7–9 mm and emits green light with a quantum efficiency of 80–97%. The parameter x = 0.7 − 7%. This high efficiency phosphor is achieved as a nonstoichiometric formulation, with gallium sulfide being a separate phase in the crystalline powder. It is claimed that if an organic solvent is added to the mixture during the precipitation, the average particle size will be smaller (∼4 − 6 mm). However, the size control method needs careful control of the firing conditions (temperature and duration) [93]. The emission of MGa2 S4 ∶Eu shifts to a longer wavelength when Sr is substituted by Ca and Ba, and to a shorter wavelength when Sr is substituted by Zn (Figure 3.21). The temperature dependence of a SrGa2 S4 ∶Eu phosphor prepared by a solid state mixing method is presented in Figure 3.22. It is apparent that there is a drop in luminous efficacy with temperature; the effect of a rise of 110 ∘ C is a drop of in efficacy of around 95%. This is about a 20.0% drop in luminous efficacy for the first 60 ∘ C rise, averaging to about 3.5% for every 10 ∘ C. However, above 90 ∘ C the luminous efficacy drops dramatically. So unlike YAG:Ce, this phosphor has very poor luminous efficacy above 100 ∘ C and such poor thermal properties will not make it a useful as an LED color converter above this temperature.

Color Conversion Phosphors for Light Emitting Diodes

117

Radiant flux (mW/nm)

0.0020

0.0015 535nm

(a)

0.0010

(b)

0.0005 515nm 0.0000

560nm

(c)

540nm

(d) 300

400

700 500 600 Wavelength (nm)

800

Figure 3.21 Photoluminescent emission spectra of thiogallate phosphors MGa2 S4 ∶Eu2+ . (a) M = Sr, (b) M = Ca, (c) M = Ba, and (d) M = Zn

350

Luminous efficacy (lm/W)

300 250 200 150 100 50 0 20

Figure 3.22 mol %)

3.4.4

40

60

80 100 Temperature (°C)

120

140

160

Plot of luminous efficacy vs. temperature for SrGa2 S4 ∶Eu phosphor (Eu = 3

Metal Nitrides

Some advantages of nitrides over oxides and sulfides are: • • • •

higher charge density between activator and nitride ligands; large red-shift (allows phosphor to emit towards the red end of the spectrum); highly condensed anionic networks (this leads to the next three advantages); high density;

118

Materials for Solid State Lighting and Displays Table 3.2 Various M-Si-N phases that have been studied as host lattices for color conversion phosphors (where M = Mg, Ca, Sr, Ba, Ln, Y) MSiN2 M5 Si2 N6 M2 Si4 N7 M3 Si6 N11

M4 SiN4 MSi3 N5 Y2 Si3 N6 Y6 Si3 N10 M2 Si5 N8 M4 Si6 N11 MSi7 N10 SrSi7 N10 M9 Si11 N23

• good chemical stability; • greater hardness (tough materials). A number of M-Si-N phases have been studied [106] in the quest to synthesize metal nitride host lattices for phosphors, many of which are listed in Table 3.2. 3.4.4.1

M2 Si5 N8 (where M = Ca, Sr, Ba, Eu)

These phosphors have aroused much interest for their use with LEDs and several major players in the LED marketplace (such as Osram [88]) have discussed their merits at conferences [106–108]. Eu2+ -doped M2 Si5 N8 (where M = Ca, Sr, Ba) has been investigated. Although X-ray diffraction analysis proved that Eu2+ -doped Ca2 Si5 N8 forms a limited solid solution with a maximum solubility of about 7 mol% having a monoclinic lattice, the Eu2+ ion can be totally incorporated into Sr2 Si5 N8 and Ba2 Si5 N8 , forming complete solid solutions with orthorhombic lattices. M2 Si5 N8 ∶Eu2+ (where M = Ca, Sr) manifests typical broad-band emission in the orange to red spectral range (600–680 nm), depending on the type of M and the europium concentration. Ba2 Si5 N8 ∶Eu2+ emits in the yellow to red with maxima from 580 to 680 nm with increasing Eu2+ content. The long-wavelength excitation and emission was attributed to the effect of a high covalency and a large crystal-field splitting on the 5d band of Eu2+ in the nitrogen environment. With increasing Eu2+ concentration, the emission band manifests a red-shift for all M2 Si5 N8 ∶Eu2+ compounds due to changing Stokes shift and the re-absorption by Eu2+ cations. The conversion (i.e., quantum) efficiency increases going from Ca to Ba and Sr under excitation at 465 nm. Sr2 Si5 N8 ∶Eu2+ is reported to have a quantum efficiency of 75–80% and a thermal quenching of only a few percent at 150 ∘ C; it is thought to be a highly promising red-emitting conversion phosphor for white-LED applications [108]. Indeed, this phosphor has been commercialized and is likely to continue to be used for some applications for some time to come. (Ba, Sr)2 Si5 N8 ∶Eu2+ (BSSNE, 2-5-8) is said to have extreme quantum efficiency and superior stability (both thermal and chemical). Its emission center is at 𝜆em = 621 nm and its full with at half maximum (fwhm = 85 nm); it has been commercialized by Lumileds [109]. Lumileds also make a ceramic from this phosphor that is said to have an amber emission for color-controlled LEDs, for use in warning signs, automotive turn signals, and smartphone dual flash. Its attributes include very high color purity, very high temperature stability, and very high EQE [110]. The crystal structures of Eu2 Si5 N8 and EuYbSi4 N7 , which both contain highly condensed Si-N networks, have been reported [111].

Color Conversion Phosphors for Light Emitting Diodes

119

Philips [112] reported red-emitting phosphors based on the M2 Si5 N8 (where M = Ca, Sr, Ba, Eu) structures; Eu2 Si5 N8 emits at 640 nm. However, they did not give the composition of their best phosphors, but suggested that their new nitride reds could be used in combination with SrGa2 S4 ∶Eu2+ and a blue LED to give a CRI of 85–95 and a range of color temperatures between 2700 ∘ C and 8000∘ C. The latest challenge in metal nitride type phosphors is to find narrow-band red-emitting LED-phosphors [110]. The attributes of such a phosphor are the high impact of the red LED-phosphor on both CRI and efficacy as a narrow-band red-emitting LED-phosphor affords +30% more lumens per watt by minimizing IR spillover at the red side of the spectrum [110, 113]. The search strategy for finding narrow-band red-emitting LED-phosphors is said to include the need for the following properties: • highly condensed host lattice; • high substitutional variability; • symmetrical coordination of activator ion (Eu2+ ) with high coordination number (CN = 8); • long Eu − N distances; • small Stokes shift; • sufficient band gap; • target fwhm ≈ 50 nm. The clue to finding such a phosphor was the identification of a narrow-band cyan emission in BaSi2 O2 N2 ∶Eu2+ [110, 114]. This phosphor has a very narrow cyan emission (fwhm = 33 nm) and a hindered relaxation around Eu2+ due to elongated Eu–L distances. In fact it has a comparable FWHM with that of direct cyan InGaN LED emission [110, 114]. The target structure was then identified as an isoelectronic nitride of the UCr4 C4 structure type. The first three structures targeted were Ba[Mg2 Ga2 N4 ]∶Eu2+ , Sr[Mg2 Al2 N4 ]∶Eu2+ , and Sr[Mg3 SiN4 ]∶Eu2+ . Unfortunately Ba[Mg2 Ga2 N4 ]∶Eu2+ and Sr[Mg2 Al2 N4 ]∶Eu2+ manifested FWHMs of 82 and 80 nm, respectively, and strong thermal quenching at room temperature [115]. The third target was the nitridomagnesosilicate Sr[Mg3 SiN4 ]∶Eu2+ ; this had ordered tetrahedral sites (Mg/Si), only one crystallographic Eu2+ site (cube-like coordination), and an extremely narrow-band red emission (43 nm). It was a good proof of concept for narrow-band red emission. However, it had a high thermal quenching due to an insufficient band gap (3.3 eV). The new target needed to have an increased band gap (> 4.5 eV) while retaining a similar crystal structure [116, 117]. The next target was Sr[LiAl3 N4 ]∶Eu2+ which proved to have ordered (Li+ ∕ Al3+ ) cations and was a variant of the UCr4 C4 structure type. In fact, it had a triclinic Cs[Na3 PbO4 ] structure with two similar Eu2+ sites (cube-like coordination). It manifested significantly reduced IR spillover as its 𝜆em was 650 nm and its fwhm = 50 nm. In addition, it has superior thermal behavior (> 95% rel. QE at 500 K) [118]. It is now the next generation red LED-phosphor and helps to produce an excellent white light. The next target will be to move the red line to around 620 nm(see Chapter 4). 3.4.4.2

CaAlSiN3 : 0.8 % Eu2+

Other nitride phosphors in the system Ca3 N2 - AlN - Si3 N4 have been studied and a red phosphor has been found, CaAlSiN3 ∶0.8%Eu2+ . This was compared with an earlier red

120

Materials for Solid State Lighting and Displays

Ca2 Si5 N8 ∶0.8% Eu2+ and found to have superior high temperature performance [119]. The excitation spectrum of CaAlSiN3 ∶0.8% Eu2+ spans the range 300–600 nm and the emission spectrum is broad and centered at 650 nm [119, 120]. Another red phosphor, CaAlSiN3 ∶Eu2+ , has a broad excitation band extending from the UV region to 590 nm [123]. At the optimum Eu2+ concentration of 1.6 mol%, the quantum output is seven times higher than for a conventional red phosphor, La2 O2 S∶Eu3+ , under 405 nm excitation. The phosphor is also reported to be more efficient than CaSiN2 ∶Eu2+ or Ca2 Si5 N8 ∶Eu2+ at any excitation wavelength. One reason for the high room temperature efficiency is the small thermal quenching, which is probably related to the rigid network of [SiN4 ] and [AlN4 ] tetrahedra. The phosphor is also chemically stable and said to be a promising material for warm white LEDs [121]. Cheetam [106] has also reported routes for synthesizing Cax Aly Siz N3 ; Ce [yellow emitter (585 nm)], and CaSiN2 ∶Ce [red emitter (630 nm)], but did not report performance data (see Chapter 4). 3.4.5

Alkaline Earth Metal Oxo-Nitrides

Alkaline earth metal oxo-nitrides are color tuneable (by suitable combinations of Ba, Sr, and Ca) green to orange and red phosphors [68]. They have good thermal behavior, broad excitation bands, and are efficient phosphors with QE values > 90% thought to be feasible [68]. Their drawbacks are that they are difficult to manufacture, and the broad reds have low luminous efficacies [68]. The crystal structure of one of the possible host lattices (SrSiAl2 O3 N2 ) has been reported [122] and from it a nine-coordinate Sr site is apparent, which could accommodate Eu2+ . The 𝛼-Sialon phosphor [(Sr, Ca)p∕ 2 Alp+q Si12−p−q Oq N16−q ∶Eu] has a broad excitation band stretching from below 350 nm to above 490 nm and emits a broad band centered around 575 nm. When this phosphor is combined with a blue LED and a red nitride phosphor, it is said to give a warm white at 3200 K on the Planckian locus [60]. A Sialon-type phosphor has been reported in the form of a powder (Cax My )(Si, Al)12 (O, N)16 [where M is at least one metal selected from the group consisting of Eu, Tb, Yb, and Er, 0.05 < (x + y) < 0.3, 0.02 < x < 0.27, and 0.03 < y < 0.3] and having a structure such that Ca sites are partially substituted by the other metal [123]. Another patent taken on this type of phosphor covers a Sialon based phosphor having a high photoluminescent intensity, which can realize a high brightness LED, particularly a white LED using a blue LED as the light source, and production of the Sialon based phosphor is also covered. The 𝛼-Sialon based phosphor of this invention has the formula Lix My Lnz Si12−(m+n) Al(m+n) On N16−n , where M is at least one metal selected from Ca, Mg, and Y, Ln is at least one lanthanide metal selected from Eu, Dy, Er, Tb, Yb, and Ce, x + ay + bz = m (assuming that the valence of metal M is a and the valence of lanthanide metal Ln is b), 0 < x ≤ 0.8, 0 < y, 0 < z, 0.3 ≤ m < 4.5, and 0 < n < 2.25 [124]. Just as there has been a push for a narrow-band red phosphor so there has for a narrow-band green-emitting oxonitride, particularly for display technology [110]. The targets for a narrow-band green phosphor for displays are high color saturation, with rapid decay of luminescence intensity < 10 ms (no saturation of the excited state). The challenge is to increase the gamut in the green spectral region with a 𝜆em ≈ 535 nm and FWHM ≈ 50 nm [110, 113]. Currently SrSi2 O2 N2 ∶Eu2+ (SSONE, Sr-2-2-2) is

Color Conversion Phosphors for Light Emitting Diodes

121

used by Lumileds; it has very high quantum efficiency and good stability (both thermal and chemical). Its 𝜆em is 532 nm and its FWHM is ∼82 nm. In addition, it has an internal quantum efficiency (𝜆exc = 420 nm) ≤ 98% [110, 114, 125–127]. A novel narrow-band green-emitting phosphor has now been reported Ba[Li2 (Al2 Si2 )N6 ]∶Eu2+ which is based on a pure nitride rather than an oxonitride formulation. It has narrow green, tuneable luminescence; its general formula is Ba[(Lix Mg2-x )(Al4-x Six )N6 ]∶Eu2+ (where x = 0 − 2), when x = 2.0 then 𝜆em = 532 nm and FWHM = ∼57 nm, and when x = 0.4 then 𝜆em = 562 nm. Depending on the value of x, it exhibits green to yellow luminescence [128]. A second narrow-band green-emitting nitride based phosphor Ba2 LiSi7 AlN12 ∶Eu2+ has been reported; its 𝜆em is 515 nm and its FWHM is 61 nm [129] (see Chapter 4). 3.4.6

Metal Fluoride Phosphors

In the last few years, narrow-band red-emitting phosphor have attracted widespread attention within the white LED and backlit display world as discussed above for both the Eu2+ and Eu3+ activated phosphor. This is partially due to renewed interest in the red-emitting phosphor, K2 SiF6 ∶Mn4+ (KSF/PFS). Although the material has been known in the literature for more than two decades, its application in white LEDs and display devices could only be realized after some of the key issues in the synthesis of materials close to this formulation were addressed. It has been shown that Mn4+ -doped complex fluoride based red phosphors have the potential to improve efficiency and color quality (CRI and R9) in LED lighting systems and enhance color gamut for LED displays [130]. Due to the low absorption of Mn4+ , a high Mn4+ doping concentration is necessary to enable low CCT applications. One of the issues with Mn4+ -doped phosphors is their high temperature and high humidity (HTHH) stability; this is even more important for high Mn4+ doping. Various post synthesis treatment techniques that help improve HTHH stability of Mn4+ -doped fluoride phosphors have been discussed [130]. Methods of controlling the particle size and luminescence properties of Mn4+ -doped K2 SiF6 have also been reported [131]. This material exists in the cubic crystal structure with narrow photoluminescence emission (with a main peak at 631 nm with a FWHM of 2 nm). Depending on the type of application within displays or general lighting, the particle size requirement for PFS varies. Reports of studies on the parameters that may influence particle growth during synthesis in order to obtain PFS of the desirable particle size have been given at a number of conferences. One such study reports the use of a variety of anti-solvents (solvents that do not dissolve K2 SiF6 ), including isopropyl alcohol (IPA), acetic acid, dimethyl sulfoxide (DMSO), acetone, ethanol, and so on, to initiate crystallization [131]. Controlling the quantity of anti-solvents and process of addition, it has been shown that good control of the particle size distribution can be achieved. The study also discusses the influence of anti-solvents on the luminescence properties and yield of PFS obtained [131]. The thermal properties of this and other Mn4+ activated phosphors have also been reported [132]. So what are potential issues for KSF/PFS? These include low absorption strength leading to concentration quenching at high Mn4+ contents, a slow decay time that causes phosphor saturation or photo-damage (i.e., a two-photon process), and the fact that the Mn4+

122

Materials for Solid State Lighting and Displays

fluoride host lattice can cause reliability concerns such as HTHH stability. However, it has been shown that not all KSF/PFS phosphors are equivalent! GE’s TriGainTM phosphor has overcome many of these limitations [133–135]. TriGainTM phosphor technology using Mn4+ -doped complex fluorides has been implemented for general illumination and display backlighting. It has been licensed to the majority of LED display makers [135].

3.5

Multi-Phosphor pcLEDs

Having looked at the most important phosphors used or potentially of use for color converters for LEDs, we now briefly consider some of the commercial approaches currently used in achieving white light from blue LEDs that involve the use of more than one phosphor. This allows the fabrication of single-chip white LEDs with higher color rendition than can be achieved by single phosphor pcLEDs. At this point we will briefly review the use of some diphosphor combinations for LEDs. Some approaches using a blue InGaN LED, with green- and red-emitting phosphors, are: Blue LED InGaN InGaN InGaN InGaN

+‘Green’ phosphor 2+

SrGa2 S4 :Eu Y(Al,Ga)G:Ce3+ YAG:Ce SrSi2 O2 N2 :Eu2+ ;

+‘Red’ phosphor 2+

MS:Eu M2 Si5 N8 :Eu2+ CaS:Eu2+ (Sr1−x−y Bax Cay )2 Si5 N8 :Eu2+

Company Lumileds [91, 136] OSRAM [107], Nichia [137] Lumileds [138] Lumileds [110, 138]

The first two systems are limited in the color temperatures and CRI values that can be achieved. The approach of Lumileds [138] for a warm white was to add a red emitter to the plain (cold) white of blue +YAG∶Ce. They chose CaS∶Eu2+ . More red gives lower CCT, and more blue gives higher CCT. With this combination, all CCTs from 2700 to 5500 K, with excellent CRI Ra > 90, can be achieved. An all nitride white diphosphor system has also been developed by Lumileds [138] that can be used to generate a wide range of whites by changing composition and/or Eu2+ concentration. The third approach is to use a UV LED and three phosphors (triphosphors) emitting in the visible region of the electromagnetic spectrum. This also allows the fabrication of single-chip white LEDs with high color rendition. However, when the UV is down-converted into the visible region, there is a large loss of energy (quantum deficit). This is particularly important for the red emission. However, this approach is still attractive if efficiency is not a problem. Three phosphors that emit, respectively, in the blue, cyan, and orange are: (Sr, Eu)5 (PO4 )3 Cl, (Sr, Eu)4 Al14 O25 , and (Ca, Eu, Mn)5 (PO4 )3 Cl (see Table 3.3 for excitation bands) [139, 140]. When these three phosphors are combined with a UV LED, color correlated temperatures between 3000∘ C and 6500∘ C can be achieved.

Color Conversion Phosphors for Light Emitting Diodes

123

Table 3.3 Color conversion phosphors that can be excited by UV or blue LEDs Phosphor

Activator

YAG SrGa2 S4 SrS (Ca, Mn)5 (PO4 )3 Cl Sr2 Si5 N8 ZnS (Zn,Cd)S SrAl2 O4 SrAl14 O25 (Y2−x−y Eux Biy )O3 Gd2 O2 S Sr5 (PO4 )3 Cl Sr5 (PO4 )3 Cl BaMgAl10 O17 YBO3 ,Y2 SiO5

Ce3+ Eu2+ Eu2+ Eu2+ Eu2+ Ag+ Cu+ , Al3+ Eu2+ Eu2+ Eu3+ Eu3+ Eu2+ Eu2+ Eu2+ 3+ 3+ Ce , Tb

3.6

Emission color

Excitation wavelength

Reference (nm)

Yellow Green Red Orange Red Blue Green Green Cyan Red Red Blue Blue Blue Green

470 460 450 415 400 400 < 400 400 < 400 360–410 380 375 375–400 375 350

[50] [50] [50] [139, 140] [50] [50] [50] [50] [139, 140] [58] [50] [50] [139, 140] [50] [50]

Quantum Dots

Many claims have been made and there has been much hype about the possible use of quantum dots (QDs) (semiconductor nanoparticles) as color converters for LEDs. The claims are because their band gaps can be tuned by size, resulting in the occurrence of size-dependent, tuneable emission (and absorption). The idea is to mix different sizes of QDs to increase spectral width across the visible region and so get “perfect white” [141]. However, a serious drawback arises from the overlap of absorption and emission bands. This causes serious re-absorption at finite concentrations, which in turn moves the emission further to the red, and as long as the QE is < 1 in highly diluted suspensions leads to a decrease of QE with concentration; this occurs at concentrations well below those desired for use in current LEDs [141]. A further drawback is that the number of compounds emitting in the visible is limited and many of the elements used in the fabrication of QDs are poisonous, such as Cd, Se, In, P, and so on. In addition, QDs are still very expensive, however they are now widespread in screens in ceramic or polymer strips for use in backlighting liquid crystal based displays, such as in Sony or in Samsung televisions [142–144]. Such items are still relatively expensive so the additional cost of the QDs is not such an issue. It should be realized that the light levels needed in television sets is much less than those needed in most lighting situations. Although high light intensities are probably detrimental to QDs there is some talk of them being used in some lighting applications. According to Osram, current QD efficiency cannot currently compete in lighting with phosphor where FWHM is not critical [145]. However, to achieve good color rendering there is still a need to populate the red end of the spectrum and this is inefficient with broad phosphor emitters. The idea is to use a narrow-band red QD plus an efficient green phosphor as a “hybrid“ solution which can

124

Materials for Solid State Lighting and Displays

increase the operating efficiency for high quality color rendering. In summary, Osram states that QD conversion technology provides a new and flexible alternative to phosphors traditionally used in LED applications. However significant challenges for integrating QD converters into the harsh LED component environment exist. In addition, competition from new narrow-band phosphor materials will continue. Still, they believe the potential for QDs remains and the future looks promising[145] (see Chapter 2).

3.7

Laser Diodes

At the time of writing this chapter, LEDs coupled with phosphors are the most efficient source of solid state white light production so long as they are operated at low input powers. The drop in the LED’s efficiency at higher input powers is associated with a phenomenon called the efficiency droop. Efficiency droop has been studied over the last 10 years and is thought to be linked to the Auger effect but as yet no definite solution has been suggested to combat this issue [146–149]. To reach more lumens per power consumed, LEDs are usually operated under higher input powers at the expense of efficiency. An alternative approach that has been suggested is to use laser diodes (LDs) in combination with the same phosphors that are used in LED lighting but packaged differently to allow high performance and efficiency (SSL). The LDs are operated under stimulated emission and the mechanisms of efficiency droop are held at their lasing threshold [150]. Thus by using LDs (which are more efficient because they are operated by stimulated emission), higher efficiencies at higher input power densities can be achieved. At the lasing threshold, all the recombination processes (including the Auger, Shockley–Read–Hall mechanism, and spontaneous) are clamped and additional carriers injected into the light quantum wells contribute only to stimulated emission. From this an argument can be put forward that unless a fix to the efficiency droop which is an intrinsic issue is found, the LDs will be competing with LEDs for the future of high brightness/high power SSL modules [151]. Future LEDs will require modules which have a shift in power conversion efficiency (PCE) to higher input power densities. Considering that Auger recombination is a fundamental process, any improvement in LED structural design such as thicker active regions and no polar substrates will only offset the efficiency droop. Unfortunately, LDs have issues related to their application which require addressing before they become the light engines for future SSLs. Their existing peak PCE of ∼ 30% is well below the 69% PCE of LEDs. Simulation studies have been carried out to give an indication of what the future LDs and LEDs may offer in terms of luminous efficacy of radiation [152].As well as the needed advance in LD technology, it is also important that the phosphor screens/targets (that are to be stimulated by high power LDs) can withstand the high temperatures resulting from exposure to the laser beam. This means that the phosphor targets need to be constructed in such a way that this produced thermal energy can be managed. Much work will also be required to check current phosphors for their application in LD based SSLs. The nature of each phosphor, its morphology, and crystal structure all need investigation to obtain efficient targets for the LDs. We have previously reported a

Color Conversion Phosphors for Light Emitting Diodes

125

solution for the thermal management of phosphor targets which can be exposed to laser beam powers of up to 5 W [153]. We have also shown the possibility of LDs coupled to phosphor targets being used as a SSL system with high power applications [154]. It was demonstrated that white light emitting modules with efficiency of up to 217 lumens per watt based on LDs can currently be made and upon further development of this technology and relevant phosphors there is room for further improvements. The report also demonstrates the ability of this technology to produce a tailored emission spectrum for a given specific requirement. Two test lamp prototypes were made using LDs and phosphor targets and their emission characteristics were investigated [154]. We believe that if the problems outlined above are overcome, then in the future LD lighting using phosphors will take a large market share wherever high power is needed. Such intense light sources could be hidden in buildings whereby they could in combination with optics for light distribution send light wherever it is needed without causing any danger from the laser beams. Similarly, by careful design automobile headlights and airplane lights could also be LD based.

3.8

Conclusions

We have explained many of the challenges still to be met before ideal color converting phosphors can be fabricated for use with UV and blue-emitting LEDs. We have chosen not to cover the many academic papers that refer to color conversion as a reason for their work on phosphors, and have only covered the systems that have been used or researched by the major industrial players. We believe that this exciting field will stimulate much academic work to find new color conversion phosphors for industrial applications in the next decade. Indeed, it must be stated that some of tomorrow’s practical color converting phosphors for LEDs may currently be being explored in academic laboratories, but as yet they have not been recognized for their future potential. We have also discussed the current state of the art for coupling LDs with phosphors for lighting applications and expect this to be an expanding field that may take over many of the high power lighting applications.

Acknowledgments We are grateful to the DTI (and many Industrial Partners) for substantial financial support in the form of DTI Technology program “Nanophosphor” under the Nano/Micro Initiative and the “Novelels” part of the Photonics and Electronics program. We are also grateful to the DTI and Enfis Ltd for a KTP program. We thank Brunel University London for allocation of funds from EPSRC grant no. EP/K504208/1. Thanks are due to present and past members of the Centre for Phosphors and Display Materials, including A. Lipman, E. Barrett, D.A. Davies, R. Ellis, M. Fathullah, P. Harris, T. Ireland, P.J. Marsh, B. Patel, J.A. Rose, R. Stone, A. Salimian, and N. Wilstead.

126

Materials for Solid State Lighting and Displays

References 1. 2.

3. 4. 5. 6. 7.

8.

9. 10. 11.

12. 13. 14. 15. 16.

17. 18. 19.

Schubert, E.F. (2003) Light Emitting Diodes, Cambridge University Press, Cambridge. Amano, H., Kito, M., Hiramatsu, K. et al. (1989) p-Type conduction in Mg-doped GaN treated with low-energy electron beam irradiation (LEEBI). Jpn. J. Appl. Phys., 28, L2112–14. Strite, S., Lin, M.E. and Morkoc, H. (1993) Progress and prospects for GaN and the III–V nitride semiconductors. Thin Solid Films, 231, 197–210. Nakamura, S., Iwasa, N., Senoh, M. et al. (1992) Hole compensation mechanism of p-type GaN films. Jpn. J. Appl. Phys., 31, 1258–66. Nakamura, S. (1994) Nichia I Cd blue LED paves way for full-color display, Nikkei Electron. Asia, 6, 65–9. Nakamura, S., Mukae, T., and Senoh, M. (1994) High-brightness InGaN/AlGaN doubleheterostructure blue-green-light-emitting diodes. J. Appl. Phys., 76, 81–9. Nakamura, S., Mukae, T., and Senoh, M. (1994) Candela-class high-brightness InGaN/AlGaN doubleheterostructure blue-light-emitting diodes. Appl. Phys. Lett., 64, 1687–9. Nakamura, S., Senoh, M., Iwasa, N. et al. (1995) High-brightness InGaN blue, green and yellow light-emitting diodes with quantum well structures. Jpn. J. Appl. Phys., 34, L797–9. Nakamura, S., Senoh, M., Iwasa, N. et al. (1995) Superbright green InGaN singlequantum-well-structure light-emitting diodes. Jpn. J. Appl. Phys., 34, L1332–5. Nakamura, S., Senoh, M., Nagaham, S. et al. (1996) InGaN-based multi-quantumwell-structure laser diodes. Jpn. J. Appl. Phys., 35, L74–6. Nakamura, S., Senoh, M., Nagaham, S. et al. (1996) InGaN multi-quantum-wellstructure laser diodes with cleaved mirror cavity facets. Jpn. J. Appl. Phys., 35, L217–20. Nakamura, S., Senoh, M., Nagaham, S. et al. (1996) InGaN multi-quantum-well structure laser diodes grown on MgAl2 O4 substrates. Appl. Phys. Lett., 68, 2105–7. Nakamura, S., Senoh, M., Nagaham, S. et al. (1996) Characteristics of InGaN multi-quantum-well structure laser diodes. Appl. Phys. Lett., 68, 3269–71. Nakamura, S., Senoh, M., Nagaham, S. et al. (1996) Ridge-geometry InGaN multi-quantum-well structure laser diodes. Appl. Phys. Lett., 69, 1477–9. Nakamura, S., Senoh, M., Nagaham, S. et al. (1996) Optical gain and carrier lifetime of InGaN multi-quantum well structure laser diodes. Appl. Phys. Lett., 69, 1568–70. Nakamura, S., Senoh, M., Nagaham, S. et al. (1996) Continuous-wave operation of InGaN multi-quantum-well-structure laser diodes at 233 K. Appl. Phys. Lett., 69, 3034. Nakamura, S., Senoh, M., Nagahama, S. et al. (1996) 9th Annual Meeting of IEEE Lasers and Electro-Optics Society, Boston, PD1.1. Nov. 18–21. Nakamura, S. (1996) Materials Research Society Fall Meeting, Boston, N1.1, Dec. 2–6. Harris, J.C., Someya, T., Hoshino, K. et al. (2000) Photoluminescence of GaN quantum wells with AlGaN barriers of high aluminium content. Phys. Stat. Sol. (a), 180, 339–43.

Color Conversion Phosphors for Light Emitting Diodes

20.

21.

22. 23.

24.

25. 26. 27. 28.

29. 30. 31. 32. 33. 34.

35. 36. 37. 38. 39.

127

Fuhrmann, D., Retzlaff, T., Rossow. U. et al. (2006) Large internal quantum efficiency of In-free UV-emitting GaN/AlGaN quantum-well structures. Appl. Phys. Lett., 88, 191108. Akasaka, T., Gotoh, H., Saito, T. et al. (2004) High luminescent efficiency of InGaN multiple quantum wells grown on InGaN underlying layers. Phys. Lett., 85 (15), 3089–91. Hangleiter, A., Fuhrmann, D., Grewe, M. et al. (2004) Towards understanding the emission efficiency of nitride quantum wells. Phys. Stat. Sol. (a), 201 (12), 2808–13. Sun, Y., Cho, Y.H., Kim, H.M. et al. (2005) Formation of Ru nanocrystals by plasma enhanced atomic layer deposition for nonvolatile memory applications. Appl. Phys. Lett., 87, 093115. Fuhrmann, D., Rossow, U., Netzel, C. et al. (2006) Optimizing the internal quantum effi ciency of GaInN SQW structures for green light emitters. Phys. Stat. Sol. (c), 3 (6), 1966–9. Ropp, R.C. (1993) The Chemistry of Artificial Lighting Devices, Elsevier Science, Amsterdam, pp. 1–664. Ropp, R.C. (1991) Luminescence and the Solid State, Elsevier Science Publishing, Amsterdam, pp. 1–453. Wells, A.F. (1975) Structural Inorganic Chemistry, 4th edn, Clarendon Press, Oxford, pp. 1–1095. Leskela, M. and Ninisto, L. (1986) Inorganic complex compounds 1’, in: Handbook on the Physics and Chemistry of Rare Earths, (eds K.A. Gschneidnerand L. Eyring), Chapter 56, Elsevier Science, Amsterdam, pp. 203–334. Garlick, G.F.J. (1949) Luminescent Materials, The Clarendon Press, Oxford, pp.1–254. Shinoya, S. and Yen, W.M. (eds) (1998) Phosphor Handbook, CRC Press, London, pp. 1–921. Blasse, G. and Brill, A. (1967) A new phosphor for flying-spot cathode-ray tubes for colour television: Yellow-emitting. Y3 Al5 O12 − Ce3+ , Appl. Phys. Lett., 11 (2), 53–5. Smets, B.M.J. (1987) Phosphors based on rare-earths, a new era in fluorescent lighting. Mater. Chem. Phys., 16, 283–99. Butterworth, M. and Helbing, R.P. (1997) Fluorescent dye added to epoxy of light emitting diode lens, US Pat. 5,847,507, July 14. Shimizu, Y., Sakano, K., Noguchi, Y. et al. (1997) Light-emitting device having a nitride compound semiconductor and a phosphor containing a garnet fluorescent material, US Pat. 5,998,925, July 29. Lowery, C.H. (1998) Multiple encapsulation of phosphor-LED devices, US Pat. 5,959,316, Sept. 1. Shimizu, Y., Sakano, K., Noguchi, Y. et al. (1999) Light-emitting device with blue light LED and phosphor components’, US Pat. 6,614,179, Dec. 10. Holn, K., Debray, A., Schlotter, P. et al. (2000) Wavelength-converting casting composition and its use, US Pat. 6,066,861, May 23. Holn, K., Debray, A., Schlotter, P. et al. (2000) Wavelength-converting casting composition and light-emitting semiconductor component, US Pat. 6,245,259, Aug. 29. Ellens, A. and Zwaschka, F. (2000) Phosphor for light sources, and associated light source, US Pat. 6,552,487, Oct. 23.

128

40. 41. 42.

43. 44. 45.

46.

47. 48. 49.

50.

51.

52.

53.

54.

55. 56.

57.

Materials for Solid State Lighting and Displays

Reeh, U., Holn, K., Stath, N. et al. (2000) Light-radiating semiconductor component with a luminescence conversion element, US Pat. 6,812,500 B2, Dec. 6. Reeh, U., Holn, K., Stath, N. et al. (2000) Light-radiating semiconductor component with a luminescence conversion element, US Pat. 6,576,930 B2, Dec. 7. Srivastava, A.M., Setlur, A.A., and Comanzo, H.A. (2001) Broad-spectrum terbiumcontaining garnet phosphors and white-light sources incorporating the same, US Pat. 6,596,195, June 1. Ellens, A., Jermann, F., Kummer, F. et al. (2002) LED-based white emitting illumination unit, US Pat. 6,504,179, Jan. 17. Reeh, U., Holn, K., Stath, N. et al. (2005) Light-radiating semiconductor component with a luminescence conversion element, US Pat. 7,126,162, March 15. Withnall, R., Silver, J., Rose, J. et al. (2004) Phosphor screens excited by blue LEDs as a white light source for full colour head up displays in automobiles, Proceedings of EL 2004, Sept. 20–23, Toronto, Canada, pp. 325–8. Silver, J., Withnall, R., Rose, J.A. et al. (2005) Broad-band green phosphor screens as a light source for head up displays in moving platforms, Technical Abstracts, Opto Ireland, April 4–6, Dublin, Ireland, p. 87. Silver, J., Withnall, R., Rose, J.A. et al. (2005) Broad-band green phosphor screens as a light source for head up displays in moving platforms, Proc. SPIE, 5826, 537–44. Silver, J. (2006) Phosphors for LEDs, current state of the art, problems, and needs. Invited lecture, The Future of LEDs, June 26–27, London. Robertson, J.M., van Tol, M.W., Smits, W.H. et al. (1981) Colourshift of the Ce3+ emission in monocrystalline epitaxially grown garnet layers. Philips J. Res., 36 (1), 15–30. Rossner, W. (2003) Trivalent rare earth activated phosphors for wavelength conversion in LEDs: Possibilities and limitations, Phosphor Global Summit, Scottsdale, AR, USA, Conference Paper 14, pp. 1–18. Silver, J., Withnall, R., Fern, G. et al. (2007) Efficiency measurements on phosphors for LEDs. (Efficacy measurements relative to blue exciting light), Phosphor Global Summit, Seoul, Korea, Conference Paper 7, pp. 1–18. Withnall, R., Silver, J., and Lipman, A. (2006) Measurements of luminous efficiencies of conversion phosphors for blue emitting LEDs, Proceedings of the 13th International Display Workshop (IDW), Otsu, Japan, Vol. 1, pp. 431–2. Silver, J., Withnall, R., Fern, G. et al. (2007) White light production from blue LEDs: the synthesis of small particle YAG : Ce type phosphors for solid state lighting. ASID, 07 (1), 306–10. Withnall, R., Silver, J., Fern, G.R. et al. (2008) Experimental and theoretical luminous efficacies of phosphors used in combination with blue-emitting LEDs for lighting and backlighting. J. Soc. Inform. Display, 16, 359–66. Withnall, R., Silver, J., and Lipman, A. (2007) Measurements of luminous efficiencies of conversion phosphors for blue emitting LEDs. ASID, 07 (1), 1074–8. Withnall, R. , Silver, J., Ireland, T.G. et al. (2009) Phosphor converted LEDs for backlighting and solid state white lighting applications: the current status and future prospects, Proceedings of Latin Display ’09, São Paulo, Brazil, pp. 99–108. Withnall, R., Silver, J., Fern, G.R. et al. (2008) (Y,Gd)AG:Ce conversion phosphors for blue LEDs, Proceedings of the 14th International Workshop on Inorganic and

Color Conversion Phosphors for Light Emitting Diodes

58.

59.

60. 61. 62.

63.

64. 65.

66. 67.

68. 69.

70. 71. 72.

73.

74.

129

Organic Elecroluminescence and 2008 International Conference on the Science and Technology of Emissive Displays and Lighting, Rome, Italy, pp. 189–91. Silver, J., Withnall, R., and Marsh, P. (2012) Selecting a white point using (Y1-a Gda )3 Al5 O12 ∶Ce3+ (where a is in the range 0.3 to 0.6) phosphors on screens and two blue LEDs, Proceedings of the 19th International Display Workshop (IDW), Kyoto, Japan, Vol. 2, pp. 1063–6. Withnall, R., Silver, J., and Ireland, T.G. (2011) Thermal quenching of the luminescence from some LED phosphors. Proceedings of the 18th International Display Workshop (IDW), Nagoya, Japan, Vol. 2, pp. 1417–20. Zachau, M. (2006) Phosphors, patents, and products for LEDs, Phosphor Global Summit, San Diego, CA, USA, Conference Paper 12, pp. 1–17. Hirakata, S.O. (2003) Inorganic oxide and phosphor, US Pat. 7,029,602 B2, Sept. 25. Setlur, A.A., Comanzo, H.A., Srivastava, A.M. et al. (2005) Evaluation of a white light phosphor for UV-LEDs–Ca2 NaMg2 V3 O12 ∶Eu3+ , J. Electrochem. Soc., 152, H205–8. Liu, R.S. (2005) Efficiency enhancement for Bi-doped Y2 O3 ∶Eu red phosphors of UV LEDs, Phosphor Global Summit, San Diego, CA, USA, Conference Paper 19, pp. 1–12. Srovastava, A. (2006) A new green emitting phosphor for UV LEDs, Phosphor Global Summit, San Diego, CA, USA, Conference Paper 13, pp. 1–10. Lee, D.-W., Kwak, C.H., and Jung, H.K. (2006) Photoluminescence properties of red-emitting YCa3 M3 B4 O15 ∶Eu3+ (M = Al or Ga) phosphors under UV irradiation: A recent advance in EL, displays and lighting. Tech. Digest, 363–4. Barry, T.L. (1968) Equilibria and Eu2+ luminescence of subsolidus phases bounded by Ba3 MgSi2 O8 , Sr3 MgSi2 O8 , and Ca3 MgSi2 O8 . J. Electrochem. Soc., 115, 733–8. Starick, D., Kempfert, W., and Limberg, H.-J. (2003) Silicate phosphors and their performance in white LEDs, Phosphor Global Summit, Scottsdale, AR, USA, Conference Paper 19, pp. 1–3. Meyer, J. (2006) Novel garnet phosphors for LEDs, Phosphor Global Summit, San Diego, CA, USA, Conference Paper 9, pp. 1–13. Levinson, L.M. and Srivastava, A.M. (1998) Light emitting device with Ba2 MgSi2 O7 ∶Eu2+ , Ba2 SiO4 ∶Eu2+ , or (SrX CaY Ba1-X-Y ) (AlZ Ga1-Z )2 Sr∶Eu2+ phosphors. US Pat. 6,429,583, Nov. 30. Ota, K., Hirano, A., and Ota, A. et al. (2000) Light emitting device having phosphor of alkaline earth metal silicate, US Pat. 6,943,380, Dec. 28. Tasch, S., Pacher, P., Roth, G. et al. (2003) Light source comprising a light-emitting element, US Pat. 6,809,347, Dec. 22. Silver, J., Fern, G.R., and Marsh, P. (2013) Factors affecting the colour of the green emitting phosphors in the system Sr2 SiO4 -Ba2 SiO4 activated by divalent europium ions, Proceedings of the 2nd International Display Workshop (IDW), Sapporo, Japan, Vol. 1, pp. 500–503. Jae, C.-K., Park, J.P., Kim, K.N. et al. (2006) White light-emitting diodes of InGaNbased (Sr, Ba)3 MgSi1.7O8 ∶Eu (2.0 ≤ x ≥ 1.7) and the luminescent properties: A recent advance in EL, displays and lighting. Tech. Digest, 334–5. Sivakumar, V. and Varadaraju, U.V. (2006) Intense red phosphor for white LEDs based on blue GaN LEDs. J. Electrochem. Soc., 153, H54–H57.

130

75.

Materials for Solid State Lighting and Displays

Sivakumar, V. and Varadaraju, U.V. (2005) Intense red-emitting phosphors for white light emitting diodes. J. Electrochem. Soc., 152 (3), H168–H171. 76. Hu, Y., Zhuang, W., Ye, H. et al. (2005) A novel red phosphor for white light emitting diodes. J. Alloy Compd, 390, 226–9. 77. Justel, T. (2008) Luminescent materials for phosphor-converted LEDs, in: Luminescence from Theory to Applications (ed. C. Ronda), Chapter 7, Wiley-VCH Verlag GmbH, Weinheim, pp.179–90. 78. Rosa, I.L.V., Marques, A.P.A., Tanaka, M.T.S. et al. (2008) Synthesis, characterization and photophysical properties of Eu3+ doped in BaMoO4 . J. Fluoresc., 18, 239–45. 79. Zhou, L.-Y., Wei, J.-S., Yi, L.-H. et al. (2009) A promising red phosphor MgMoO4 ∶Eu3+ for white light emitting diodes. Mater. Res. Bull., 44, 1411–14. 80. Zhou, L., Wei, J., Wu, J. et al. (2009) Potential red-emitting phosphor for white LED solid-state lighting. J. Alloy Compd, 476, 390–2. 81. Zhou, L.-Y., Wei, J.-S., Gong, F.-Z. et al. (2008) A potential red phosphor ZnMoO4 ∶Eu3+ for light-emitting diode application. J. Solid State Chem., 181, 1337–41. 82. Stone, R., Silver, J., Withnall, R. et al. (2009) The problems with using Eu3+ activated red emitting phosphors with blue light emitting LEDs to generate white light. Luminous efficiency measurements of red emitting phosphors based on the formula Li(Eux Y1-x ) -(MoO4 )y (WO4 )2-y (0 ≤ x ≤ 1,y = 0, 1, 2), Proceedings of the 16th International Display Workshop (IDW), Miyazaki, Japan, Vol. 1, pp. 359–61. 83. Silver, J., Lipman, A., Fathullah, M. et al. (2015) The synthesis, characterisation and potential of Eu3+ doped molybdate phosphors for white light emitting diodes, Proceedings of the 22nd International Display Workshop (IDW), Otsu, Japan, Vol. 1, pp. 500–503. 84. Leverenz, H.W. (1950) An Introduction to Luminescence of Solids, John Wiley & Sons, Inc. New York. 85. Destriau, G. (1936) Recherches sur les scintillations des sulfures de zinc aux rayons. J. Chem. Phys., 33, 587–625. 86. Vecht, A., Davies, D.A. and Smith, D. (1998) A novel method for the preparation of sulfides and selenides and its application for phosphor synthesis. Mater. Res. Innovat., 2, 17680. 87. Davies, D.A., Vecht, A., Rose, J.A. et al. (1999) Improved zinc-sulfide type phosphors by a novel synthetic method. SID 99 Digest, 30, 1018–21. 88. Davies, D.A., Vecht, A., Silver, J. et al. (2000) A novel method for the preparation of inorganic sulfides and selenides. Part 1: Binary materials and Group II–VI phosphors. J. Electrochem. Soc., 147, 765–71. 89. Marsh, P.J., Davies, D.A., Silver, J. et al. (2001) A novel method for the clean syntheses of sulfides/selenides. Part 2: Ternary sulfides/selenides. J. Electrochem. Soc., 148, D88–93. 90. Davies, D.A., Silver, J., Vecht, A. et al. (2001) A novel method for the synthesis of ZnS for use in the preparation of phosphors for CRT devices. J. Electrochem. Soc., 148, H143–8. 91. Mueller-Mach, R. and Mueller, G.O. (2000) White light-emitting diodes for illumination. Proc. SPIE, 3938, 30–41.

Color Conversion Phosphors for Light Emitting Diodes

92.

93. 94.

95.

96.

97.

98.

99. 100. 101. 102.

103. 104. 105.

106. 107. 108.

109.

131

Van Haecke, J.E., Smet, P.F., Michelena, E. et al. (2006) Solvothermal synthesis of europium-doped submicron SrS: A recent advance in EL, displays and lighting. Tech. Digest, 353–5. Tian, Y. (2003) Phosphor development for LED lighting devices, Phosphor Global Summit, Scottsdale, AR, USA, Conference Paper 10, pp. 1–7. Juestel, T., Schmidt, P.J., and Busselt, W. (2003) Novel phosphors and phosphor blends for white LEDs, Phosphor Global Summit, Scottsdale, AR, USA, Conference Paper 15, pp. 1–13. Kim, K.N., Park, J.K., Choi, K.J. et al. (2006) Luminescent properties of europium doped alkaline earth selenide phosphors: A recent advance in EL, displays and lighting. Tech. Digest, 334–5. Withnall, R., Silver, J., Ghazli, M. et al. (2012) Accelerated ageing studies of CaS∶Eu2+ and SrS∶Eu2+ phosphors, Proceedings of the 19th International Display Workshop (IDW), Kyoto, Japan. Vol. 3, pp. 1727–9. Silver, J., Withnall. R., and Marsh, P. (2012) Selecting a red point using CaS∶Eu3+ and SrS∶Eu3+ phosphors on screens and one blue LED, Proceedings of the 19th International Display Workshop (IDW), Kyoto, Japan, Vol. 3, pp. 1741–4. Peters, T.E. and Baglio, J.A. (1972) Luminescence and structural properties of thiogallate phosphors Ce+3 and Eu+2 -activated phosphors’, Part I, J. Electrochem. Soc., 119, 230–6. Donohue, P.C. and Hanlon, J.E. (1974) The synthesis and photoluminescence of MII M2 III (S, Se)4 , J. Electrochem. Soc., 121, 137–42. Davolos, M.R., Garcia, M.R., Fouassier, C. et al. (1989) Luminescence of Eu2+ in strontium and barium thiogallates. J. Solid State Chem., 83, 316–23. Kobayashi, T., Uheda, K., Naruke, H. et al. (1996) A novel preparation of thiogallatebased rare-earth phosphors. Chem. Lett., 26, 567. Eichenauer, L., Jarofke, B., Mertins, H.-C. et al. (1996) ‘Optical characterization of europium and cerium in strontium thiogallate thin films and powders. Phys. Stat. Sol. (a), 153, 515–27. Yang, S., Stoffers, C., Zhang, F. et al. (1998) Green phosphor for low-voltage cathodoluminescent applications: SrGa2 S4 ∶Eu2+ , Appl. Phys. Lett., 72, 158–60. Barthou, C. and Benoit, J. (1998) SrGa2 S4 ∶RE phosphors for full colour electroluminescent displays. J. Alloy Compd, 275–7, 709–15. Sastry, I.S.R., Bacalski, C.F., and McKittrick, J. (1999) Preparation of green-emitting Sr1−x Eux Ga2 S4 phosphors by a solid-state rapid metathesis reaction. J. Electrochem. Soc., 146, 4316–9. Cheetham, A.K. (2006) Novel phosphors for solid state lighting, Phosphor Global Summit, San Diego, CA, USA, Conference Paper 7, pp. 1–13. Bogner, G., Botty, G., Braune, B. et al. (2003) Light source using a yellow-to-redemitting phosphor, US Pat. 6,649,946, Nov. 18. Yi, Y.Q., van Steen, J.E.J., van Krevel, J.W.H. et al. (2006) Luminescence properties of red emitting M2 Si5 N8 ∶Eu2+ (M = Ca, Sr, Ba) LED conversion phosphors. J. Alloy Compd, 417, 273–9. Mueller-Mach, R., Mueller, G., Krames, M.R. et al. (2005) Highly efficient all-nitride phosphor-converted white light emitting diode. Phys. Stat. Sol. (a), 202, 1727–32.

132

Materials for Solid State Lighting and Displays

110. Schnick, W. (2016) Narrow or broad –exploratory discovery of novel nitride LED-phosphors, Phosphor Global Summit, Newport Beach, CA, USA, Conference Paper 12, pp. 1–39. 111. Huppertz, H. and Schnick, W. (1997) Eu2 Si5 N8 and EuYbSi4 N7 . The first nitridosilicates with a divalent rare earth metal. Acta Cryst., C53: 1751–3. 112. Meyer, J. (2006) From garnets to novel LED phosphors, Phosphor Global Summit, San Diego, CA, USA, Conference Paper 9, pp. 1–13. 113. Pust, P., Schmidt, P.J., and Schnick, W. (2015) A revolution in lighting. Nat. Mater., 14, 454–8. 114. Bachmann, V., Ronda, C., Oeckler, O. et al. (2009) Color point tuning for (Sr, Ca, Ba)Si2 O2 N2 ∶Eu2+ for white light LEDs. Chem. Mater., 21, 316–25. 115. Pust, P., Hintze, F., Hecht, C. et al. (2014)‘Group (III) nitrides M[Mg2 Al2 N4 ] (M = Ca, Sr, Ba, Eu) and Ba[Mg2 Ga2 N4 ] – structural relation and nontypical luminescence properties of Eu2+ doped samples. Chem. Mater,. 26, 6113–19. 116. Schmiechen, S., Schneider, H., Wagatha, P. et al. (2014) Toward new phosphors for application in illumination-grade white pc-LEDs: The nitridomagnesosilicates Ca[Mg3 SiN4 ]∶Ce3+ , Sr[Mg3 SiN4 ]∶Eu2+ , and Eu[Mg3 SiN4 ]∕. Chem. Mater., 26, 2712–19. 117. Tolhurst, T.M., Schmiechen, S., Pust, P. et al. (2016) Electronic structure, bandgap, and thermal quenching of Sr[Mg3 SiN4 ]∶Eu2+ in comparison to Sr[LiAl3 N4 ]∶Eu2+ . Adv. Opt. Mater., DOI: 10.1002/adom.201500615. 118. Pust, P., Weiler, V., Hecht, C. et al. (2014) Narrow-band red-emitting Sr[LiAl3 N4 ]∶ Eu2+ as a next-generation LED-phosphor material. Nat. Mater., 13, 891–6. 119. Uheda, K. and Yamamoto, H. (2005) Nitride phosphors in the system Ca3 N2 -AlN Si3 N4 , Phosphor Global Summit, San Diego, CA, USA, Conference Paper 21, pp. 1–5. 120. Miyamoto, Y., Yamamoto, H., Hirosaki, N. et al. (2006) Luminescence decay characteristics of a red nitride phosphor CaAlSiN3 ∶Eu2+ for white light-emitting diodes: A recent advance in EL, displays and lighting. Tech. Digest, 339–40. 121. Uheda, K., Hirosaki, N., Yamamoto, Y. et al. (2006) Luminescence properties of a red phosphor, CaAlSiN3 ∶Eu2+ , for white light-emitting diodes. Electrochem. Soc. State Lett., 9, H22–5. 122. Lauterbach, R. and Schnick, W. (1998) Synthese, Kristallstruktur und Eigenschafteneines neuen Sialons -SrSiAl2 O3 N2 , Z. Anorg. Allg. Chem, 624, 1154–8. 123. Mitomo, M., Endo, T., Ueda, K. et al. (2002) Oxynitride phosphor activated by a rare earth element, and sialon type phosphor, US Pat. 6,632,379, June 6. 124. Sakata, S.-I. and Yamada, T. (2004) Sialon-based phosphor and its production method, US Pat. EP 1,498,466, July 15. 125. Höppe, H.A., Stadler, F., Oeckler, O. et al. (2004) Ca[Si(2)O(2)N(2)] – a novel layer silicate. Angew. Chem. Int. Ed., 43, 5540–2. 126. Oeckler, O., Stadler, F., Rosenthal, T. et al. (2007) Real structure of SrSi2 O2 N2 . Solid State Sci. 9, 205–12. 127. Kechele, J.A., Oeckler, O., Stadler, F. et al. (2009) Structure elucidation of BaSi2 O2 N2 – A host lattice for rare-earth doped luminescent materials in phosphorconverted (pc)-LEDs. Solid State Sci., 11, 537–43.

Color Conversion Phosphors for Light Emitting Diodes

133

128. Strobel, P., Schmiechen, S., Siegert, M. et al. (2015) Narrow-band green emitting nitridolithoalumosilicate Ba[Li2 (Al2 Si2 )N6 ]∶Eu2+ with framework topology whj for LED/LCD-backlighting applications. Chem. Mater., 27, 6109–15. 129. Takeda, T., Hirosaki, N., Funahashi, S. et al. (2015) Blue-emitting Sr3 Si8−x Alx O7+x N8−x ∶Eu2+ discovered by a single-particle-diagnosis approach: crystal structure, luminescence, scale-up synthesis, and its abnormal thermal quenching behavior. Chem. Mater., 27, 5892–7. 130. Nammalwar, P. K., Porob, D. G., Hanumantha, R. et al. (2015)‘Controlling particle size and luminescence in Mn4+ activated K2 SiF6 , Paper presented at the 228th ECS Meeting, Phoenix, AZ, USA, October 13, 2015. 131. Sista, S.P., Murphy, J., Garcia-Santamaria, F. et al. (2015) Thermal & humidity stability of Mn4+ doped complex fluoride phosphors, Paper presented at the 228th ECS Meeting, Phoenix, AZ, USA, October 13, 2015. 132. Setlur, A.A., Murphy, J., Garcia-Santamaria, F. et al. (2015) Thermal Properties of Mn4+ doped complex fluoride phosphors’, Paper presented at the 228th ECS Meeting, Phoenix, AZ, USA, October 13, 2015. 133. Murphy, E., Garcia-Santamaria, F., Setlur, A.A. et al. (2015) PFS, K2SiF6∶ Mn4+: the red-line emitting LED phosphor behind GE’s TriGain Technology™ platform. SID Digest, 46 (1), 927–30. 134. Setlur, A.A., Garcia-Santamaria, F., Murphy, J.E. et al. (2015) Comparing the properties and performance of GE TrigainTM Phosphors versus typical K2 SiF6 ∶Mn4+ phosphors, Phosphor Safari 2015, Niigata, Japan. 135. Murphy, J. (2016) GE TriGainTM : Commercialized, narrow band, red emitting phosphors for LED lighting and display applications, Phosphor Global Summit, Newport Beach, CA, USA, Conference Paper 8, pp. 1–26. 136. Craford, G. (2003) Solid state lighting: The promise and challenges for LEDs, Phosphor Global Summit, Scottsdale, AR, USA, Conference Paper 13, pp. 1–36. 137. Yamada, M., Naitou, T., Izuno, K. et al. (2003) Red-enhanced white-light-emitting diode using a new red phosphor. Jpn. J. Appl. Phys., 42, L20–3. 138. Mueller, G. (2005) Set the pace in the white space: White LEDs for illumination and backlighting, Phosphor Global Summit, San Diego, CA, USA, Conference Paper 18, pp. 1–15. 139. Radkov, E. (2005) High quality white light with near-UV LED chips, Phosphor Global Summit, San Diego, CA, USA, Conference Paper 17, pp. 1– 9. 140. Radkov, E. (2003) White light with UV LEDs, Phosphor Global Summit, Scottsdale, AR, USA, Conference Paper 20, pp. 1–19. 141. Mueller, G. (2006) Colour conversion of LED light, Phosphor Global Summit, San Diego, CA, USA, Conference Paper 11, pp. 1–14. 142. Dalton, S. Color IQ™ Optics Deliver Best in Class Color Performance in Award Winning Sony TVs, Reuters, Oct. 8, 2013. 143. Fern, G.R., Silver, J., Coe-Sullivan, S. et al. (2015) Cathodoluminescence and electron microscopy of red quantum dots used for display applications. J. SID,, 23, 50–5.

134

Materials for Solid State Lighting and Displays

144. Samsung Joins the Quantum Dot Crowd at CES 2015 with Super SUHD TVs, Consumer Reports. http://www.consumerreports.org/cro/news/2015/01/samsungjoins-the-quantum-dot-crowd-at-ces-2015-with-super-suhd-tvs/index.htm. Accessed September 10, 2016. 145. Steegmueller, U. (2016) Spectrally narrow optical down converters for next generation LEDs, Quantum Dots Forum, Newport Beach, CA, USA, Conference Paper 8, pp. 1–27. 146. Shen, Y.C., Mueller, G.O., Watanabe, S. et al. (2007) Auger recombination in InGaN measured by photoluminescence. Appl. Phys. Lett., 91, 141101–3. 147. David, D. and Grundmann, M.J. (2010) Influence of polarization fields on carrier lifetime and recombination rates in InGaN-based light-emitting diodes. Appl. Phys. Lett., 97, 033501–3. 148. Iveland, J., Martinelli, L., Peretti, J. et al. (2013) Direct measurement of Auger electrons emitted from a semiconductor light-emitting diode under electrical injection: identification of the dominant mechanism for efficiency droop. Phys. Rev. Lett., 110 (17), 177406. 149. Mueller-Mach, R. and Mueller, G.O., (2016) Phosphors beyond linearity, Phosphor Global Summit, Newport Beach, CA, USA, Conference Paper 10, pp. 1–38. 150. Wierer, J.J., Tsao, J.Y., and Sizov, D.S. (2013) Comparison between blue lasers and light-emitting diodes for future solid-state lighting Laser. Photonics Rev., 7 (6), 963–93. 151. Wierer, J.J., Tsao, J.Y. and Sizov, D.S. (2014) The potential of III-nitride laser diodes for solid-state lighting. Phys. Stat. Sol. (c), 11 (3–4), 674–7. 152. Ohno, Y. (2005) Spectral design considerations for white LED color rendering. Opt. Eng., 44, 111302–9. 153. Salimian, A., Silver, J., Fern, G. et al. (2015) Evaluation of thermally stable phosphor screens for application in laser diode excited high brightness white light modules. ECS J. Solid State Sci. Technol., 5 (1), R3001–R3006. 154. Salimian, A., Fern, G.R., Upadhyaya, H. et al. (2016) Laser diode induced lighting modules. ECS J. Solid State Sci. Technol., 5 (3), R26–R33.

4 Nitride and Oxynitride Phosphors for Light Emitting Diodes Le Wang1 and Rong-Jun Xie2 1 College of Optical and Electronic Technology, China Jiliang University, Hangzhou, Zhejiang, China 2 National Institute for Materials Science (NIMS), Tsukuba, Ibaraki, Japan

In Chapter 3, oxide and sulfide phosphors have been introduced and discussed. In this chapter we will give an overview of nitride and oxynitride phosphors for white light emitting diodes (LEDs). These phosphors have emerged as a new type of luminescent materials that are quite suitable for use as down-conversion materials in white LEDs.

4.1

Introduction

Inorganic phosphors are widely used in lighting (e.g., fluorescent tubes), displays (e.g., cathode tube display, field emission display, plasma display), imaging (e.g., computed tomography), and so on [1–3]. For different applications, the requirements for phosphors are varied greatly. For example, the phosphors in fluorescent lamps are required to have strong absorption at 254 nm, whereas those in plasma display panels must show strong emission colors of blue, green, and red under the 147 or 174 nm excitation. This leads to a variety of material systems that can be used as phosphor hosts, such as oxide, sulfide, fluoride, phosphate, chloride, and so on [4]. For phosphors used in white LEDs, they should have strong absorptions of ultraviolet (360–410 nm) or blue (440–470 nm) light, and small thermal quenching/degradation [5–7]. For this reason, most phosphors used for traditional lighting are not applicable for white LEDs because of their low conversion efficiency or quantum efficiency (the efficiency of converting the LED emissions into other visible light), such as Y2 O3 ∶Eu3+ and Materials for Solid State Lighting and Displays, First Edition. Edited by Adrian Kitai. © 2017 John Wiley & Sons Ltd. Published 2017 by John Wiley & Sons Ltd.

136

Materials for Solid State Lighting and Displays

LaPO4 ∶Tb3+ ,Ce3+ . Therefore, in recent years searching for appropriate host crystals of LED phosphors has been a hot topic for material scientists [7–16]. In fact, extensive efforts have been made to discover novel phosphors for white LEDs, leading to a large number of luminescent materials being reported or published. On the other hand, there are several problems with these phosphors, such as large thermal quenching (e.g., orthosilicate Sr2 SiO4 ∶Eu2+ ), low chemical stability (e.g., CaS∶Eu2+ , Sr2 GaS4 ∶Eu2+ ), and low quantum efficiency, which unfortunately limit their practical use in lighting devices. A report on luminescent nitride compounds dates back to 40 years ago when Gaido et al. investigated the photoluminescence of the green-emitting MgSiN2 ∶Eu2+ in 1974 [17]. In 1993, Endo et al. observed the red emission of Mn2+ -doped ZnSiN2 prepared at high pressure [18]. Lee et al. reported the red emission of Eu2+ in CaSiN2 in 1997 [19]. Later, Krevel et al. and Uheda et al. studied the effect of the nitrogen content in host crystals on the luminescence, and found that the redshift of the photoluminescence spectra was enhanced with increasing the nitrogen concentration [20, 21]. Hoppe et al. also demonstrated that the significant redshift in both excitation and emission spectra of Ba2 Si5 N8 ∶Eu2+ was attributed to the large crystal field splitting and nephelauxetic effect as Eu2+ was coordinated to the nitrogen ligand [22]. The interesting photoluminescence observed in nitride compounds launched a research boom for this new type of luminescent materials. Typically, the excitation spectrum of some nitride phosphors matches well with the emission band of ultraviolet or blue LED chips, which enables nitride phosphors to be promising down-conversion luminescent materials in white LEDs. The first paper devoted to the application of nitride phosphor in solid state lighting was contributed by Xie et al. in 2004 [9]. Since then, extensive investigations on the luminescence in rare earth activated nitride or oxynitride compounds have been carried out, resulting in the discovery of a large number of new phosphors [5, 8, 14]. Nitride/oxynitride phosphors have been proven to be the most encouraging luminescent materials in solid state lighting. They have better chemical stability than previously used alkaline earth sulfide (e.g., SrS∶Eu) and thiogallate (e.g., SrGa2 S4 ∶Eu) phosphors that are high efficiency red and green phosphors, respectively. Furthermore, in comparison with garnet (i.e., YAG:Ce) and orthosilicate (i.e., Sr2 SiO4 ∶Eu, Sr1-x Bax SiO4 ∶Eu) phosphors that are commonly utilized in solid state lighting, nitride/oxynitride phosphors have more abundant emission colors and superior thermal stability due to their structural diversities. The major advantages or characteristics of typical nitride and oxynitride phosphors are: 1. Significantly redshifted excitation and emission spectra. Nitride and oxynitride compounds are structurally built up on three-dimensional SiN4 , Si(O, N)4 , AlN4 , or Al(O, N)4 tetrahedral networks. When activator ions (e.g., Eu2+ and Ce3+ ) are introduced into the nitride structures, they will be coordinated to nitrogen atoms, forming covalent chemical bonds with relatively short distances. As shown in Figure 4.1, the nephelauxetic effect will occur due to the strong covalence between the activators and nitrogen ligands, lowering the 5d energy levels of the activator ions. Further, as nitrogen has a higher formal charge than oxygen, a large crystal field splitting will be expected, which again reduces the 5d energy levels of Eu2+ or Ce3+ . Both of these effects therefore result in great redshifts in emission and excitation spectra, enabling the activators in the hosts to absorb and emit light at longer wavelengths. In addition, broad excitation spectra will be achieved due to the large crystal field splitting.

Nitride and Oxynitride Phosphors for Light Emitting Diodes

137

5d nephelauxetic effect crystal field splitting free ion (Eu2+, Ce3+,...) ions in a solid

4f higher covalence shorter bond length

higher formal charge smaller coordination number

Figure 4.1 Schematics of the nephelauxetic effect and crystal field splitting of Eu2+ or Ce3+ ions in a host

2. Small Stokes shifts. Stokes shift is defined as the difference (in wavelength or frequency units) between positions of the band maxima of the absorption and emission spectra of the same electronic transition. As nitride and oxynitride phosphors exhibit broad excitation spectra extending to the blue or even green spectral region, the Stokes shift becomes smaller compared with that observed in oxide phosphors. The small Stokes shift indicates less energy loss during the luminescence process, and thus leads to high quantum efficiency. 3. Small thermal quenching. Thermal quenching of luminescence is largely related to the structural rigidity and band structure of the host. The nitride phosphor hosts usually contain condense corner- or edge-shared (Si, Al)(O, N)4 tetrahedra, and thus have stiff structures that make them more thermally stable. In addition, the small Stokes shift also contributes to a small thermal quenching. 4. Abundant emission colors. The emission color of activators with the 5d → 4f electronic transitions is greatly dependent on the local structure and electronic environments because the 5d electrons are not shielded by the outer orbitals. Along with the nephelauxetic effect and large crystal field splitting, the structural diversity enables nitride phosphors to emit colors in a very broad spectral range. It thus makes sense to tune the emission color of phosphors by applying some strategies to change the environments (crystal field strength and nephelauxetic effect) surrounding activators. Materials synthesis and processing affect a material’s microstructure, phase purity and hence its functional properties. Therefore, the development of suitable, advanced synthetic methods for luminescent materials is of great importance for achieving better performance. In the next section, we are going to introduce several common methods for the synthesis of nitride/oxynitride phosphors, which include solid state reaction, carbothermal reduction and nitridation, gas reduction and nitridation, ammonothermal synthesis, and alloy nitridation. Following this, we will overview the photoluminescence properties of

138

Materials for Solid State Lighting and Displays

nitride/oxynitride phosphors, such as excitation and emission spectra, quantum efficiency, and thermal quenching. Finally, the applications of nitride/oxynitride phosphors in solid state lighting, such as general illumination and liquid crystal display backlight, will be demonstrated.

4.2

Synthesis of Nitride and Oxynitride Phosphors

In white LEDs, phosphors are often dispersed in organic resins and used in powder forms. So, the powder characteristics of phosphors, such as morphology, particle size, and particle size distribution, are largely dependent on how they are synthesized or treated. Besides these parameters, the synthetic method also has a great influence on their phase purity, microstructure, photoluminescence, luminous efficiency, stability/reliability, cost, and so on. As nitride and oxynitride phosphors contain nitrogen atoms in their crystal structure, nitride compounds will be used as starting materials or a nitriding gas will be introduced for nitridation during the powder synthesis. This leads to some specialized synthetic approaches to nitride and oxynitride phosphors along with the traditional solid state reaction method. 4.2.1

Solid State Reaction Method

The solid state reaction method is commonly used to produce many kinds of phosphor powders including nitride and oxynitride phosphors. It is also a generally accepted way for the mass production of phosphors. This method usually involves four sequential steps: (i) matter diffusion at interfaces between solid particles; (ii) chemical reactions at the atomic level; (iii) nuclei formation; and (iv) solid-phase transport and growth of new phase. For the solid state reaction to proceed smoothly and effectively, the starting powders are thus required to have high chemical reactivity, large surface area, good dispersion, or low agglomeration. Nitride phosphors are free of oxygen, and are also referred to as nitridosilicate or nitridoaluminosilicate phosphors. To prepare these phosphors, silicon nitride (Si3 N4 ), aluminum nitride (AlN), alkaline earth metal nitride (M3 N2 , M = Ca, Sr, Ba), alkali metal nitride (Li3 N), and rare earth metal nitride (LnN, Ln = Y, La, Eu, Ce) are often used as starting powders. For oxynitride phosphors, oxide starting powders, such as SiO2 , Al2 O3 , or alkaline earth metal carbonates are used together with the aforementioned nitride starting materials. The firing temperature, varying from 1400 to 2100 ∘ C, is exclusively dependent on the chemical composition of the phosphors as well as the starting materials used. In addition, high pressure nitrogen gas (>1.0 MPa) is not always necessary but only if the firing temperature is higher than ∼1800 ∘ C (the decomposition temperature of Si3 N4 is ∼1830 ∘ C at an ambient nitrogen pressure). In the case of using high nitrogen gas pressure, a gas pressure sintering furnace needs to be used to synthesize nitride and oxynitride phosphors, as shown in Figure 4.2. As Si3 N4 has a low diffusion coefficient, a higher firing temperature is thus required for it to react with other starting powders. Moreover, the commercially available Si3 N4 usually contains a few percent of SiO2 , which may introduce oxygen into the lattice. In addition, alkaline earth metal nitrides or lanthanide metal nitrides are either expensive or not easily available. Therefore, they are sometimes replaced by other alternative raw materials to

Nitride and Oxynitride Phosphors for Light Emitting Diodes

Figure 4.2

A gas pressure sintering furnace used for firing nitride phosphors

Table 4.1 Examples of nitride phosphors synthesized using silicon diimide and metals Phosphor host Chemical reaction Temperature (∘ C) Ba2 Si5 N8 SrAlSi4 N7 Ba2 AlSi5 N9

139

2Ba + 5Si(NH)2 → Ba2 Si5 N8 + 5H2 + N2 3Sr + 4Si3 N4 + 3AlN + 2N2 → 3SrAlSi4 N7 6Ba + 5Si3 N4 + 3AlN + 2N2 → 3Ba2 AlSi5 N9

1600 1630 1725

Ref. [22] [23] [24]

reduce the firing temperature or cost. Schnick and co-workers use more reactive silicon diimide, Si(NH)2 , to substitute Si3 N4 as the silicon source, as well as metals to replace metal nitrides (Table 4.1) [22–24]. The phosphors synthesized by the solid state reaction method are usually agglomerated due to the high firing temperature, and thus should be pulverized to control the particle size, particle size distribution, and morphology. The powder pulverization may induce some surface defects or damage, which would significantly reduce the photoluminescence properties. In addition to their high cost or unavailability, air-sensitive alkaline earth and lanthanide metal nitride raw powders must be handled in a glove box with oxygen and moisture purification systems, leading to a complex powder processing. To minimize these problems, more facile synthetic methods, such as gas reduction and nitridation, carbothermal reduction and nitridation, and alloy nitridation, are required and developed. 4.2.2

Gas Reduction and Nitridation

The gas reduction and nitridation method enables the direct synthesis of nitride fine powders from the multicomponent oxide systems by using a gaseous reduction–nitridation medium. It generally consists of two reaction processes: (i) reduction of oxide components into metal substances; and (ii) nitridation of metal substances into metal nitrides. These are both gas–solid phase reactions.

140

Materials for Solid State Lighting and Displays

Table 4.2 Examples of nitride and oxynitride phosphors prepared by the gas reduction and nitridation method Phosphor host

Oxide precursors

Method to oxide precursors

Temperature (∘ C)

Ref.

Ca-α-sialon Y-α-sialon LaSi3 N5 (La, Ca)3 Si6 N11

SiO2 -Al2 O3 -CaO-Eu2 O3 Y2 O3 -Al2 O3 -SiO2 La2 O3 -SiO2 -CeO2 La(NO3 )3 ⋅ 6H2 O, Ca(NO3)2 ⋅ 4H2 O, SiO2 SrO-SiO2 -Eu2 O3 BaO-SiO2 -Eu2 O3 Al2 O3 -Eu2 O3 CaO-Al2 O3 -SiO2 -Eu2 O3

Co-precipitation Co-precipitation Mixing Mixing

1300–1500 1800 ∘ C) and in a high-pressure nitrogen atmosphere (∼190 MPa). Watanabe et al. used the ammonothermla synthesis method to prepare SrAlSiN3 ∶Eu2+ at a temperature as low as 500 ∘ C in a supercritical ammonia fluid with a pressure of 100 MPa [47]. In this method, NaNH2 was used as reaction medium. Li et al. synthesized the Eu-doped CaAlSiN3 phosphor at 500–800 ∘ C via the reaction of a CaAlSi alloy in supercritical ammonia (100 MPa) with the addition of sodium amide

142

Materials for Solid State Lighting and Displays

[48]. Well-crystallized samples with plate- and bar-like nanocrystals were successfully synthesized by first converting the alloy at 300–400 ∘ C into sodium ammonometallates and subsequently decomposing the ammonometallates up to 800 ∘ C into CaAlSiN3 . Instead of using sodium amide, the use of sodium azide, which was converted into sodium amide during heating, led to a product of plate-like crystals. Zeuner et al. obtained nanocrystalline M2 Si5 N8 ∶Eu2+ phosphors by a one-pot reaction of the corresponding elements (Sr/Ba, Eu, Si) in supercritical ammonia [49]. The nitridosilicate phosphors exhibited spherically shaped particles with crystallites which were 200 nm in size. The advantages of the one-pot precursor approach are: (i) the exact Sr/Ba content and the doping concentration of Eu2+ can easily be controlled in a wide range; (ii) thorough mixing of raw materials down to an atomic level (Sr, Ba, Eu) can be achieved; and (iii) no milling and pre-reaction steps are necessary which might give rise to contamination.

4.3

Photoluminescence Properties of Nitride and Oxynitride Phosphors

During the last 15 years, great efforts have been made to search for novel nitride and oxynitride phosphors for solid state lighting, and these led to the discovery of a variety of luminescent nitrides and oyxnitrides. Divalent europium (Eu2+ ) and trivalent cerium (Ce3+ ) are the two most effective and frequently used activators in phosphors, the luminescence of which is attributed to the 4f ↔ 5d electronic transitions. Unlike the 4f orbitals, the 5d orbitals are exposed to significant interaction with the surrounding atoms and ions, the luminescence due to the 5d → 4f transitions is therefore strongly dependent on the host lattice. This leads to the emission color of both Eu2+ and Ce3+ varying in a very broad range from ultraviolet (UV) to red. Besides these, other rare earth ions, such as Yb2+ , Pr3+ , and transition metal ions, such as Mn2+ , are also considered as activators in nitride phosphors. The luminescence of activators in a host depends on the local crystal and electronic structures, such as the symmetry, coordination number, bond length, bond angle, and cation size. Understanding the structure–property relationship is hence a must for the selection of hosts, materials design, and property tailoring. As phosphors used for white LEDs, photoluminescence properties, that is spectral shape, spectral position, band width, absorption, conversion efficiency (quantum efficiency, luminescence efficiency), and luminescence saturation, are critical parameters that determine their values and practical applications. Furthermore, the reliability of a phosphor, including thermal and chemical stabilities, is also a key factor that controls the life span of the white LEDs using it. In this section, we concentrate on the introduction of photoluminescence properties, the structure–property relationship, and thermal quenching of nitride and oxynitride phosphors. 4.3.1 4.3.1.1

Luminescence Spectra of Typical Activators Eu2+ -Doped Nitride and Oxynitride Phosphors

The Eu2+ ion has the ground state of 4f 7 (8 S7∕2 ) and the excited state of 4f 6 5d1 , and its luminescence arises from the spin-allowed electronic transition of 4f 6 5d1 → 4f 7 . The

Nitride and Oxynitride Phosphors for Light Emitting Diodes

143

Table 4.5 Crystal structure and photoluminescence properties of Eu2+ -doped nitride phosphors Phosphors

Crystal structure

Excitation (nm)

Peak emission (nm)

FWHM (nm)

IQE / EQE (%)

Ref.

AlN

Wurtzite

365

465

52

76/46

[50]

SrSi6 N8

Orthorhombic, Imm2

370

450

44

38/28 [51] (@ 405 nm)

BaSi7 N10

Monoclinic, Pc

297

475, 482–500

∼100

60/52

[52, 53]

Sr2 Si5 N8

Orthorhombic, Pmn21

450

609–680

88

82/71

[54, 55]

CaSiN2

Orthorhombic, Pbca

410

630

—

39.6/28.5

[56, 57]

SrYSi4 N7

Hexagonal, P63mc

390

548–570

—

—

[58]

Ba5 Si11 Al7 N25

Orthorhombic, Pnnm

450

568

90

36.7/—

[59]

SrAlSi4 N7

Orthorhombic, Pna21

460

635

116

70/49

[23, 60]

Ba2 AlSi5 N9

Triclinic, P1

450

584

100

22.3/—

[24]

BaSi4 Al3 N9

Monoclinic, P21/c

365

500

67

—

[59]

CaAlSiN3

Orthorhombic, Cmc21

460

660

90

86/70

[61]

SrAlSiN3

Orthorhombic, Cmc21

455

610

80

—

[62]

SrMg3 SiN4

Tetragonal, I41 ∕a

450

615

43

—

[63]

Ba2 LiAlSi7 N12

Orthorhombic, Pnnm

400

515

61

79/— [64] (@ 405 nm)

EQE, External quantum efficiency; FWHM, full width at half maximum; IQE, internal quantum efficiency.

luminescence of Eu2+ generally shows a broad emission band with the position and width determined by the crystal and electronic structure of the host. Table 4.5 and Table 4.6 list the photoluminescence properties of nitride and oxynitride phosphors reported in the literature [23, 24, 50–82]. As shown, the nitride and oxynitride phosphors exhibit a broad range of emission colors, different band width, and quantum efficiencies. Although the photoluminescence properties are dominantly linked to the crystal structure, they are also largely dependent on the chemical composition and processing conditions. In the following, several interesting and typical nitride and oxynitride phosphors will be discussed in detail.

144

Materials for Solid State Lighting and Displays

Table 4.6 Crystal structure and photoluminescence properties of Eu2+ -doped oxynitride phosphors Phosphors

Crystal structure Excitation Peak FWHM IQE/EQE (nm) emission (nm) (%) (nm)

Ref.

BaSi6 N8 O

Orthorhombic, Imm2

310

503

—

[65]

SrSi9 Al19 ON31

Rhombohedra, R-3

290

448–490 90

72/62 (@365 nm)

[66]

365

515–525 66

75/67

[67]

Sr3 Si13 Al3 O2 N12

102

BaSi3 Al3 O4 N5

Monoclinic, P21 ∕m

365

470

—

85/79 (@305 nm)

[68]

Sr(Al, Si)4 (N, O)6

Orthorhombic, Fdd2

334

490

97

—

[69]

Ba(Si, Al)5 (O, N)8

Orthorhombic, A21am

390

475

72

—

[70]

β-sialon

Hexagonal,

450

535

55

78/54

[10]

SrSi2 O2 N2

Triclinic, P1

450

544

83

86/69

[71]

CaSi2 O2 N2

Monoclinic, P21

450

563

98

87/72

[71]

BaSi2 O2 N2

Orthorhombic, Pbcn

450

492

36

53/41

[71]

Ba3 Si6 O12 N2

Trigonal, P-3

450

525

65

—

[72]

Sr5 Al5+x Si21−x N35−x O2+x

Orthorhombic, Pmn21

400

510

69

—

[73]

Sr14 Si68−s Al6+s Os N106−s

Monoclinic, P21

376

508

—

—

[74]

Li-α-sialon

Hexagonal, P31c

450

563–586 —

—

[75]

Ca-α-sialon

Hexagonal, P31c

400

583

94

—

[76]

Sr-α-sialon

Hexagonal, P31c

400

575

94

70.6/57.8

[77]

Ca15 Si20 N30 O10

Cubic, Pa3̄

460

650

—

—

[78]

Ba1.5 Ca0.5 Si5 N6 O3

Monoclinc, Cm

460

600

—

—

[79]

La4−x Cax Si24 O3+x N18−x

Monoclinic, C2

460

565

—

—

[80]

Sr3 Si2 O4 N2

Cubic, Pa3̄

460

600

80

—

[81]

Sr2 Si7 Al3 ON13

Orth, Pna21

450

615

106

87/73

[82]

EQE, External quantum efficiency; FWHM, full width at half maximum; IQE, internal quantum efficiency.

Nitride and Oxynitride Phosphors for Light Emitting Diodes

4.3.1.1.1

145

Blue-Emitting Phosphors

The blue LED-driven solid state lighting is the mainstream of white LEDs used for general illuminations and backlights, as blue LEDs have high efficiency, good safety, and cause less damage to the illuminated objects. However, the color rendering index of these white LEDs is generally smaller than 90, which makes them hard to be adopted in situations such as a photographer’s workroom, an art studio, and meat and vegetable markets. The near UV LED-driven solid state lighting is therefore an option to realize super-high color rendering white LEDs, in which blue phosphors are combined with other green and red phosphors. AlN:Eu, Si Hirosaki et al. reported an intense blue emission in Eu2+ and Si co-doped AlN, and demonstrated its use in field emission displays due to the high color purity, efficiency, and reliability [83]. This phosphor was prepared by firing the powder mixture of AlN, Si3 N4 and Eu2 O3 at 2050 ∘ C by using the gas-pressure sintering method. Later, Inoue et al. carried out an extensive investigation on the photoluminescence properties of AlN:Eu, Si [50]. As seen in Figure 4.3, the excitation spectrum shows a broad band covering the range of 250–425 nm, with the maximum at 290 nm. AlN:Eu, Si has an emission band centered at ∼465 nm and a FWHM of 52 nm. The optimal concentration of Eu2+ is about 0.1 mol%. At 150 ∘ C, the luminescence intensity maintains 90% of the initial value measured at room temperature, implying excellent thermal stability. Under 365 nm excitation, the absorption efficiency, IQE and EQE are 63, 76, and 46%, respectively. These results indicate that AlN:Eu, Si is a promising blue phosphor suitable for near UV LED chips. The blue emission cannot be achieved in the Eu2+ solely doped AlN without the addition of Si. Dierre et al. showed that Si played a key role in: (i) enhancing the solubility of Eu2+ in the AlN lattice; and (ii) forming an AlN polytypoid-like local structure that accommodates Eu2+ [84]. Takeda et al. directly observed the distribution of Eu2+ in the stacking faults by high-resolution transmission electron microscopy and energy-dispersive X-ray spectroscopy element mapping (Figure 4.4). A layered structure phosphor is thus proposed for AlN:Eu, Si [85].

Quantum efficiency and absorption

PL intensity (a.u.)

100

200 250 300 350 400 450 500 550 600 Wavelength (nm) (a)

Internal QE

80

Absortion efficiency

60 40

External QE

20 0 250

300

350 400 Wavelength (nm) (b)

450

500

Figure 4.3 Photoluminescence spectra (a) and quantum efficiency as a function of excitation wavelength (b) of the AlN:Eu, Si blue phosphor. Source: Inoue et al. 2009 [50]. Reproduced with permission of American Chemical Society

146

Materials for Solid State Lighting and Displays

Eu

SEM

Sr or Eu layer

10 nm

BF

10 nm

c

Eu L

Al

Si

c

>0.5 nm

a 10 nm

Si K

10 nm

Al K

b

(a)

(b)

Figure 4.4 (a) Elemental distribution mapping of Eu, Si, and Al, showing the stacking faults in AlN:Eu, Si. (b) Proposed layered structure of AlN:Eu, Si. Source: Takeda et al. 2010 [85]. Reproduced with permission of Royal Society of Chemistry

b a

c

Figure 4.5

Sr layer

(Si, Al)(O,N)4 layer

Layered structure of SrSi9 Al19 ON31

SrSi9 Al19 ON31 :Eu Grins et al. reported the crystal structure of SrSi9 Al19 ON31 [86]. It has an orthorhombic crystal system and the space group R-3. The lattice parameters are a = 5.335 Å and c = 79.1 Å. The Sr atoms are accommodated in the structure by forming layers containing Sr(O, N)12 cubo-octahedra and (Si, Al)(O, N)4 tetrahedra, with these layers being alternated with AlN-type blocks into which (Si, Al)3 (O, N)4.5 layers are incorporated (Figure 4.5). The SrSi9 Al19 ON31 ∶Eu phosphor shows a single and symmetric band centered at 471 nm and a FWHM of 90 nm, upon the 290 nm excitation (Figure 4.6) [66]. The excitation spectrum is also a broad band covering a spectral range of 250–450 nm, with the peak maximum at 290 nm. The crystal field splitting and the Stokes shift are 13 000, and 1500 cm−1 , respectively. Very uniquely, there is no concentration quenching for this blue phosphor. As known from its crystal structure, the neighboring two Eu atoms are separated by a distance of 5.335 Å when they are in the same layer, and by 26.37 Å when they are in different layers. Therefore, energy transfer between Eu2+ ions occurs at a very low probability, leading to the absence of concentration quenching. It is also evidenced by a quite small spectral overlap between the excitation and emission spectra.

Nitride and Oxynitride Phosphors for Light Emitting Diodes

2500

em. = 471 nm ex. = 290 nm

2000 1500 1000 500 0 250 300 350 400 450 500 550 600 Wavelength (nm) (a)

1.0 Relative intensity (a.u.)

PL intensity (a.u.)

3000

147

Absorption Inter. QE Exter. QE

0.8 0.6 0.4 0.2 0.0 300

350

400 450 500 Wavelength (nm) (b)

550

600

Figure 4.6 Photoluminescence spectra (a) and quantum efficiency (b) of SrSi9 Al19 ON31 ∶ 5 mol% Eu2+ . Source: Liu et al. 2013 [66]. Produced with permission of Elsevier

The blue SrSi9 Al19 ON31 ∶Eu phosphor has low thermal quenching. The luminescence of the sample with 5 mol% Eu declines by only 8% when measured at 150 ∘ C. Under the 365 nm excitation, the absorption efficiency, IQE, and EQE are 85, 72, and 62%, respectively. They are 66, 50, and 33% when excited at 405 nm. It therefore shows that the SrSi9 Al19 ON13 ∶Eu2+ phosphor is a promising blue phosphor suitable for near UV LEDs. 4.3.1.1.2

Green-Emitting Phosphors

𝛃-sialon:Eu β-sialon is derived from β-Si3 N4 by partially replacing Si—N bonds with Al—O ones, with the chemical composition of Si6−z Alz Oz N8−z (0 < z ≤ 4.2) [87]. It has a hexagonal crystal system and a space group of P63 . Hirosaki et al. first reported the photoluminescence of Eu2+ -doped β-sialon∶Eu that was prepared by firing the powder mixture of Si3 N4 , AlN, Al2 O3 , and Eu2 O3 at 1900–2100 ∘ C in a 1.0 MPa nitrogen atmosphere [10]. The emission spectrum shows a narrow band centered at 535 nm and a FWHM of 55 nm when the phosphor is excited at 450 nm. The excitation spectrum covers a broad spectral range of 280–480 nm, and has two maxima at 300 and 405 nm, respectively. The emission band can be shifted greatly by varying the z value, and a low z value leads to a blueshifted emission. Takahashi et al. developed a short-wavelength β-sialon∶Eu with the emission maximum at 525 nm and a FWHM of 47 nm, by using Si instead of Si3 N4 as the raw material (Figure 4.7) [88]. β-sialon∶Eu shows an extremely high reliability under the humidity test and blue light irradiation [89]. The commercially available β-sialon∶Eu has IQE and EQE of 78 and 54% under 450 nm excitation, respectively. It is recognized as one of the best green phosphors for LED backlights due to its narrow emission band (i.e., high color purity) and excellent thermal stability. It is generally accepted that the foreign metal atoms cannot be stabilized in β-sialon. However, the intense green emission of Eu2+ in β-sialon indicates that Eu2+ is indeed located in the host lattice. It was confirmed by several analytical techniques including electron energy loss spectroscopy, transmission electron microscopy, and cathodoluminescence. Kimoto et al. directly observed the Eu2+ atoms in the lattice of β-sialon by using scanning transmission electron microscopy (STEM) (Figure 4.8) [90]. They reside in the large voids along

148

Materials for Solid State Lighting and Displays Emission

Excitation Emission

Intensity

Normalized intensity

Excitation

z = 0.24 0.075 0.050 0.025

200 250 300 350 400 450 500 550 600 650 Wavelength (nm) (a)

250

300

350

400

450

500

550

650

Wavelength (nm) (b)

Figure 4.7 Photoluminescence spectra of 𝛽-sialon∶Eu2+ with (a) z = 1.0, 0.3 mol% Eu2+ and (b) z = 0.025 − 0.24, 0.1 mol% Eu2+ . Source: Takahashi et al. 2011 [88]. Reproduced with permission of The Electrochemical Society

the [001] direction. Moreover, Eu2+ is coordinated to six (O, N) anions with an equivalent distance of 2.4932 Å [91]. This is supported by the narrow and symmetric emission spectrum of β-sialon∶Eu. MSi𝟐 O𝟐 N𝟐 ∶Eu(M = Ca , Sr , Ba) Alkaline earth metal oxonitridosilicates, MSi2 O2 N2 (M = Ca, Sr, Ba), have the same chemical composition but different crystal structures and emission colors, as shown in Table 4.6. The crystal system of CaSi2 O2 N2 , SrSi2 O2 N2 , and BaSi2 O2 N2 is monoclinic, triclinic, and orthorhombic, respectively. The corresponding space group is P21, P1, and Pbcn [71, 92]. Their structures are closely related and built up of highly dense and corner-sharing SiON3 tetrahedral layers that are separated by alkaline earth metal layers. All the excitation spectra of MSi2 O2 N2 ∶Eu(M = Ca, Sr, Ba) show a broad band centered at 440 nm covering the spectral range of 200 – 500 nm. The emission band of BaSi2 O2 N2 ∶Eu is much narrower than the other two; the FWHM is 36 nm whereas it is 83 and 98 nm for SrSi2 O2 N2 ∶Eu and CaSi2 O2 N2 ∶Eu, respectively. The maximum emission is 563, 544, and 492 nm for CaSi2 O2 N2 ∶Eu, SrSi2 O2 N2 , and BaSi2 O2 N2 ∶Eu, respectively. Among these phosphors, SrSi2 O2 N2 ∶Eu has the smallest thermal quenching. The EQE is 72, 69, and 41% for CaSi2 O2 N2 ∶Eu, SrSi2 O2 N2 ∶Eu, and BaSi2 O2 N2 ∶Eu, respectively.5 Color tuning is easily achieved in MSi2 O2 N2 ∶Eu(M = Ca, Sr, Ba) by forming solid solutions [71]. By substituting Ca for Sr in SrSi2 O2 N2 , the maximum of the emission band can be tuned precisely to any wavelength between 537 nm and 560 nm. As CaSi2 O2 N2 is not isotypic to SrSi2 O2 N2 , a heterogeneous mixture of the two structures rather than solid solutions is observed from the X-ray diffraction patterns when the Ca substitution is larger than 50%. The quantum efficiency remains unchanged until the Ca fraction is larger than 75%. In addition, the quenching temperature of the emission gradually reduces upon raising the fraction of Ca, which is 600 and 440 K for SrSi2 O2 N2 ∶Eu2+ and CaSi2 O2 N2 ∶Eu2+ , respectively. A continuous redshift (from 538 to 564 nm) is found when the Ba fraction is smaller than 75%, whereas a sudden blueshift of the emission band to 495 nm is seen when Ba totally replaces Sr, as shown in Figure 4.9. This change is ascribed to the structural evolutions. In (Sr, Ba)Si2 O2 N2 ∶Eu, the color tuning can be obtained while preserving a high

Nitride and Oxynitride Phosphors for Light Emitting Diodes

149

(a)

(b)

Eu

(c)

Figure 4.8 (a) Standard crystal structure, viewed from the c-direction; (b) simulated structure, showing the location of Eu2+ in the channel along the c-axis; and (c) HAADF-STEM image, showing the presence of Eu2+ in the large voids parallel to the c-axis. Source: Kimoto et al. 2009 [90] and Li et al. [91]. Reproduced with permission of AIP Publishing LLC

150

Materials for Solid State Lighting and Displays Sr0.98–xBaxSi2O2N2:Eu x=0 x = 0.05 x = 0.1 x = 0.15 x = 0.25 x = 0.5 x = 0.75 x = 0.98

Emission intensity

1.0 0.8 0.6 0.4 0.2 0.0 450

500

550 600 650 Wavelength (nm)

700

750

800

Figure 4.9 Photoluminescence spectra of Sr0.98−x Bax Si2 O2 N2 ∶Eu2+ . Source: Bachmann et al. 2009 [71]. Reproduced with permission of American Chemical Society

quantum efficiency and a high quenching temperature for samples with a small amount of Ba. The maximum emission can be varied in the spectral range of 538–548 nm. 4.3.1.1.3

Yellow-Emitting Phosphors

Ca-𝛂-sialon:Eu2+ Ca-α-sialon is isostructural with α-Si3 N4 , and has the chemical formula Cax Si12−m−n Alm+n On N16−n (M, here Ca, is the “modifying” cation, x is the solubility of cation M, and m and n refer to the numbers of Al—N and Al—N bonds substituted for Si—N, respectively) [93]. It has trigonal crystal symmetry with the space group P31c. The structure is built up of highly dense and corner-sharing SiN4 tetrahedra, and consists of stacking layers of Si(Al) and N(O) in the sequence of ABCDABCD . . . . The Ca atoms are coordinated to seven nearest N(O) atoms at three distances. Ca-α-sialon∶Eu2+ shows both broad excitation and emission spectra that are ascribed to the 4f → 5 d electronic transitions of Eu2+ (Figure 4.10) [76]. The excitation spectrum extends from 250 to 550 nm, with two distinct bands centered at 290 and 400–450 nm, respectively. The emission spectrum covers a wide spectral range of 500–700 nm with the maximum at 582 nm and FWHM of 94 nm [76]. Ca-α-sialon∶Eu2+ exhibits a very small thermal quenching, the luminescence intensity of which only reducing by 10 % at 150 ∘ C. The absorption efficiency and EQE are 81 and 50% under the 450 nm excitation, respectively. Ba2 AlSi5 N9 :Eu2+ Ba2 AlSi5 N9 , reported by Kechele et al., was synthesized by firing the powder mixture of Si3 N4 , AlN, and Ba in a radio-frequency furnace at ∼1725 ∘ C [24]. It crystallizes in the triclinic space group P1 (no. 1), and has the lattice parameters a = 9.860(1) Å, b = 10.320(1) Å, c = 10.346(1) Å, 𝛼 = 90.37(2)∘ , 𝛽 = 118.43(2)∘ , 𝛾 = 103.69(2)∘ , and Z = 4. There are eight independent Ba sites in this new nitridosilicate that have different coordination numbers (CN = 6 − 10). The Ba—N distances are between 2.547 Å and 3.680 Å.

Nitride and Oxynitride Phosphors for Light Emitting Diodes

151

PL intensity (a.u.)

Absorption and quantum efficiency

1.0

300

400

500 600 Wavelength (nm) (a)

Absorption efficiency

0.8 0.6

Internal QE External QE

0.4 0.2 0.0 300

700

350

400 450 500 Wavelength (nm) (b)

550

600

Figure 4.10 (a) Photoluminescence spectra and (b) quantum efficiency of Ca-𝛼-sialon∶ Eu2+ (m = 2, n = 1, 7 mol% Eu)

1.0

Relative intensity

0.8 0.6 0.4 0.2 0.0 300

400

500

600

700

800

λ /(nm)

Figure 4.11 Excitation and emission spectra of Ba2 AlSi5 N9 ∶Eu2+ . Source: Kechele et al. 2009 [24]. Reproduced with permission of American Chemical Society

As seen in Figure 4.11, Ba2 AlSi5 N9 ∶Eu2+ (2 mol% Eu2+ ) reveals a broadband emission peaking at 584 nm and a FWHM of 100 nm under the 450 nm excitation. The quantum efficiency of this yellow phosphor is about 22.3%. Ba5 Si11 Al7 N25 :Eu2+ Ba5 Si11 Al7 N25 was discovered by Hirosaki et al. using a singleparticle diagnosis approach. It crystallizes in the orthorhombic system with the space group Pnnm (no. 58) [59]. The lattice parameters are a = 9.5923(2) Å, b = 21.3991(5) Å, c = 5.8889(2) Å and Z = 2. This new nitridosilicate compound consists of a highly condensed framework built up of SiN4 and AlN4 tetrahedra that are mostly liked via common corners. Both Si and Al atoms share the equivalent site with a disordered distribution. The framework of the Ba5 Si11 Al7 N25 host has three different Ba positions that are coordinated to 11 (Ba1), 10 (Ba2), and 8 (Ba3) nitrogen atoms, respectively.

Materials for Solid State Lighting and Displays

PL intensity (a.u.)

152

50 μm 300 (a)

400

500 600 Wavelength (nm) (b)

700

800

Figure 4.12 (a) A single particle and (b) excitation and emission spectra of Ba5 Si11 Al7 N25 ∶Eu2+ (5 mol%). Source: Hirosaki et al. 2014 [59]. Reproduced with permission of American Chemical Society

The Ba(Eu)—N bond lengths vary between 2.838(2) Å and 3.030 (3) Å for Ba1, between 3.003(2) Å and 3.354(2) Å for Ba2, and between 2.719(3) Å and 3.083(2) Å for Ba3. The photoluminescence of Ba5 Si11 Al7 N25 ∶Eu2+ was measured on a single phosphor particle (Figure 4.12). It shows a broad emission band centered at ∼570 nm and a FWHM of 98 nm. The luminescence intensity reduces by 35% at 150 ∘ C. The thermal quenching temperature (at which the luminescence reduces by 50%) is ∼190 ∘ C. The IQE of the particle is about 36.4%. 4.3.1.1.4

Red-Emitting Phosphors

Sr2 Si5 M8 :Eu Sr2 Si5 N8 was discovered by Schlieper et al. [94]. It has ane orthorhombic crystal system with the space group Pmn21 . The lattice parameters are a = 5.712(3) Å, b = 6.817(3) Å, c = 9.336(1) Å, and Z = 2. The Sr atoms are accommodated in the open channels that are formed by a highly dense framework of corner-sharing SiN4 tetrahedra. The Sr atoms occupy two different crystallographic sites, which are coordinated to eight and nine nitrogen atoms, respectively. The distances between Sr and N are 2.57–3.231 Å for Sr1 (CN = 8) and 2.542–3.181 Å for Sr2 (CN = 9), respectively. Sr2 Si5 N8 ∶Eu2+ shows both broad excitation and emission bands (Figure 4.13) [54, 55]. The excitation band covers the spectral range of 250–600 nm with the maximum at ∼450 nm. The emission band has a peak emission of 625 nm and a FWHM value of ∼100 nm. Under the 450 nm excitation, the absorption efficiency and EQE are 85 and 72%, respectively (Figure 4.13). Although Sr2 Si5 N8 ∶Eu2+ exhibits a very small thermal quenching, it has a serious problem of thermal degradation that needs to be minimized for practical applications [95].95 CaAlSiN3 :Eu2+ CaAlSiN3 crystallizes in the orthorhombic Cmc21 system [61]. The lattice parameters are a = 9.8020(4) Å, b = 5.6506(2) Å, and c = 5.0633(2) Å. The Ca atoms are located in the voids parallel to the c-axis, and are coordinated to five nitrogen atoms with an average distance of 2.50 Å.

Nitride and Oxynitride Phosphors for Light Emitting Diodes

PL intensity (a.u.)

Absorption and quantum efficiency (%)

100

200

300

400 500 600 Wavelength (nm) (a)

700

800

153

Internal QE

80

Absortption

60

External QE

40 20 0 300

350

400 450 500 Wavelength (nm) (b)

550

600

Figure 4.13 Photoluminescence spectra (a) and quantum efficiency (b) of Sr2 Si5 N8 ∶ 2 mol% Eu2+

CaAlSiN3 ∶Eu2+ is a red-emitting phosphor that shows both broad excitation and emission bands (Figure 4.14). The emission band has a maximum of 658 nm and a FWHM value of 94 nm. The absorption efficiency and EQE are 86 and 70% under the 450 nm excitation, respectively. CaAlSiN3 ∶Eu2+ also shows a small thermal quenching, the luminescence of which is only reduced by 10% at 150 ∘ C [96]. The emission band of CaAlSiN3 ∶Eu can be blue-shifted by replacing Ca with Sr (Figure 4.14c). For Ca1−x Srx AlSiN3 ∶Eu (0.8 mol% Eu), the peak emission is shifted from 650 nm (x = 0) to 610 nm (x = 1) [62]. This blueshift is attributed to the increased Eu—N distance with the Sr substitution. Under the 455 nm excitation, the IQE (80–83%) and EQE (70%) of the composition with x = 0.8 are comparable with that of the x = 0 composition. SrAlSi4 N7 :Eu2+ SrAlSi4 N7 was discovered by Hecht et al. [23]. It crystallizes in the orthorhombic system and the space group Pna21 (no. 33). The lattice parameters are a = 11.742(2) Å, b = 21.391(4) Å, c = 4.966(1) Å, and Z = 8. This new nitridoaluminosilicate compound contains a highly condensed network structure built up of SiN4 and AlN4 tetrahedra. These tetrahedra are not only linked via common corners but also edge-shared for tetrahedra which are presumably centered by aluminum. Two different Sr atoms are coordinated to six and eight nitrogen atoms, respectively. The Sr—N distances vary between 2.504(5) and 3.143(5) Å for Sr1 and between 2.653(7) and 3.057(6) Å for Sr2. SrAlSi4 N7 ∶Eu2+ exhibits both broadband excitation and emission spectra (Figure 4.15) [60]. The excitation spectrum consists of four bands peaking at 320, 355, 410, and 460 nm, respectively. For blue excitation, a single emission band centered at 620–650 nm (FWHM ∼116 nm) is observed in samples containing varying Eu concentrations. At 150 ∘ C the luminescence intensity remains 83% of the initial intensity measured at room temperature. The activation energy for thermal quenching is 0.235 eV (1 mol% Eu2+ ). The absorption efficiency, IQE and EQEe of this red phosphor (3 mol% Eu2+ ) are 70, 70 and 49% under 460 nm excitation, respectively. Fukuda et al. developed a new red Sr2 Si7 Al3 ON13 ∶Eu phosphor by replacing Si—N with Al—O in SrAlSi4 N7 [82]. Sr2 Si7 Al3 ON13 has an unchanged crystal structure but

154

Materials for Solid State Lighting and Displays

100

PL intensity (a.u.)

Absorption and quantum efficiency (%)

Absorption

300

400

60

External QE

40 20 0 300

800

500 600 700 Wavelengthh (nm)

Internal QE

80

350

400 450 500 Wavelength (nm)

(a)

550

600

(b)

PL intensity (a.u.)

Ca1–xSrxAlSiN3 Eu: 0.8 mol% x = 0.1 x = 0.4 x = 0.7 x = 0.8 x = 0.9

500

550

600

650 700 750 Wavelength (nm)

800

850

(c)

Figure 4.14 (a) Typical photoluminescence spectra, (b) quantum efficiency of CaAlSiN3 ∶ Eu (0.8 mol%), and (c) color tuning in (Ca1−x Srx )AlSiN3 ∶Eu (0.8 mol%)

Ex.410 nm

PL intensity (a.u.)

1.2 1.0 0.8

1% 2% 3% 5% 7.5% 10%

0.6 0.4 0.2 0.0 200

100 90 80 70

0

60

ΔE = 0.235eV

−1 In [Io /l-1]

Em.

Relative intensity (%)

1.4

50

−2

40

1% 2% 3% 5% 7.5% 10%

−3

30

−4

20

−5 15

10

20

25 1/T × 104

30

35

0 300

400

500

600

Wavelength (nm) (a)

700

800

25

50

75

100

125

150

175

200

Temperature (°C) (b)

Figure 4.15 (a) Photoluminescence spectra of SrAlSi4 N7 with varying Eu2+ and (b) thermal quenching of SrAlSi4 N7 ∶Eu2+ (1 mol%). Source: Ruan et al. 2011 [60]. Reproduced with permission of John Wiley and Sons

Nitride and Oxynitride Phosphors for Light Emitting Diodes

155

its unit cell is enlarged with the substitution. The lattice constants are a = 11.8033(13), b = 21.589(2), and c = 5.0131(6) Å. The increased unit volume reduces the crystal field splitting and thus blue-shifts the luminescence. The emission peak is at 615 nm for the sample with 10 mol% Eu. The absorption efficiency, IQE and EQE of this sample are 84, 87 and 73%, respectively, for the 450 nm excitation. From the large number of nitride phosphors that have been investigated, only a few of them can be practically used in solid state lighting depending on their luminescence spectra, quantum efficiency, thermal stability, and so on. Table 4.6 lists the quantum efficiency (IQE and EQE) of some nitride phosphors. It is seen that nitride phosphors generally have a lower quantum efficiency than YAG∶Ce3+ (IQE>90% and EQE>80%). This could be due to the fact that nitride phosphors do not have advantages over garnets in many aspects, such as the selection and treatment of raw materials, selection of synthetic methods, control of the processing conditions, addition of fluxes, and so on. Therefore, it is still a daunting challenge and an endless task to enhance the quantum efficiency of nitride phosphors. Of the phosphors listed in Table 4.5 and Table 4.6, β-sialon∶Eu2+ and CaAlSiN3 ∶Eu2+ are the most successful products. β-sialon∶Eu2+ has a very narrow emission band, high color purity, and very stable photoluminescence against temperature and irradiation, enabling it to be used in highly reliable and wide color gamut white LED backlights. CaAlSiN3 ∶Eu2+ is a highly efficient and reliable red-emitting phosphor, widely used in high color rendering and wide color gamut white LEDs. In addition, some other phosphors, such as Ca-α-sialon∶Eu2+ , MSi2 O2 N2 ∶Eu2+ (M = Ca, Sr, Ba), and M2 Si5 N8 ∶Eu2+ (M = Ca, Sr, Ba) can also be used practically if their quantum efficiency or thermal reliability is further improved. 4.3.1.2

Ce3+ −Doped Nitride and Oxynitride Phosphors

The Ce3+ ion has the simplest electron configuration among the rare earth metal ions. There are two sub-levels for the 4f 1 ground state configuration: 2 F5∕2 and 2 F7∕2 , which are separated by about 2000 cm−1 due to the spin-obit coupling [97]. This is the origin for the double emission bands of Ce3+ . The Ce3+ emission, strongly affected by the host lattice through the crystal field splitting and the nephelauxetic effect, usually ranges from UV to blue colors. However, yellow or even red emission colors can also be achieved in hosts having covalent chemical bonds and strong crystal fields (e.g., nitride or oxynitride compounds), as shown in Table 4.7 [20, 27, 28, 56, 58, 81, 98–106]. In the following, photoluminescence properties of several typical Ce3+ -doped nitride and oxynitride phosphors will be introduced. 4.3.1.2.1

LaAl(Si6−z Alz )(N10−z Oz )∶Ce3+

LaAl(Si6−z Alz )(N10−z Oz ) crystallizes in the orthorhombic system with the space group Pbcn. The La atoms, residing in the channels along the [001] direction, are coordinated by seven (O, N) atoms at an average distance of 2.70 Å. Takahashi et al. reported the photoluminescence of Ce3+ -activated LaAl(Si6−z Alz )(N10−z Oz ) (z ∼ 1) and its application in white LEDs (Figure 4.16) [100]. The phosphor was synthesized by firing the powder mixture of α-Si3 N4 , AlN, La2 O3 , and CeO2 at 1900 ∘ C for 2 h under 1 MPa nitrogen atmosphere. LaAl(Si6−z Alz )(N10−z Oz )∶Ce3+ shows a broad excitation band covering the spectral range of 250–450 nm with the maximum at about 370 nm. The excitation band can

156

Materials for Solid State Lighting and Displays

Table 4.7 Crystal structure and photoluminescence of some Ce3+ − doped nitride and oxynitride phosphors Phosphors

Crystal system

Excitation (nm)

Peak emission (nm)

Ref.

AlN Y2 Si3 O3 N4 Y5 Si3 O12 N Y4 Si2 O7 N2 YSiO2 N La5 Si3 O12 N La4 Si2 O7 N2 LaSiO2 N La3 Si8 O4 N11 LaSi3 N5 LaAl(Si6−z Alz ) (N10−z Oz )(z ∼ 1) SrYSi4 N7 BaYSi4 N7 Sr3 Si2 O4 N2 Ca-α-sialon β-sialon La3 Si6 N11 CaAlSiN3 La3 BaSi5 N9 O2 CaSiN2 Ca2 Si5 N8 Sr2 Si5 N8 Ba2 Si5 N8

Wurtzite Melilite-type Apatite-type Cuspidine-type Pseudowollastonite-type Hexagonal, P63 ∕m Monoclinic, P121 ∕C1 Hexagonal, P-6c2 Orthorhombic, C2/c Orthorhombic, P21 21 21 Orthorhombic, Pbcn

310 390 355 390 350 361 345 356 365 355–380 405

470 500 475 500 440 478 488 416 424 464–475 470–490

[98] [20] [20] [20] [20] [99] [99] [99] [99] [27] [100]

Hexagonal, P63 mc Hexagonal, P63 mc Cubic, Pa3̄ Hexagonal, P63 ∕m Hexagonal, P63 ∕m Tetragonal, P4bm Orthorhombic, Cmc21 Orthorhombic, Pmn21 Cubic Orthorhombic, Pbca Orthorhombic, Pmn21 Orthorhombic, Pmn21

340 338 300–400 383 410 450 450 450 535 395 420 410

445 420 520 500–518 486 577–581 570–603 578 625 470 495, 553 451, 497, 561

[58] [101] [81] [102] [103] [28] [104] [105] [56] [106] [106] [106]

be redshifted by increasing the Ce3+ concentration, thus enabling the phosphor to be excited efficiently by the near UV light (e.g., 405 nm). The emission spectrum of LaAl(Si6−z Alz )(N10−z Oz )∶Ce3+ exhibits a broad band centered at 460–500 nm depending on the Ce3+ concentration. The FWHM value of the emission band is 110 nm. The IQE and EQE are 62 and 50%, respectively, at the 405 nm excitation. 4.3.1.2.2

LaSi3 N5 :Ce3+

The crystal structure of LaSi3 N5 was revealed by Woike and Jeitschko [107]. LaSi3 N5 crystallizes in the orthorhombic crystal system with the space group P21 21 21 . The La atoms are located in the pentagonal voids along the c-axis, formed by five corner-shared SiN4 tetrahedra. The coordination number of La is nine, and the average La—N distance is 2.78 Å. Suehiro et al. synthesized the Ce3+ -activated LaSi3 N5 phosphor by using the gas reduction and nitridation method [27]. The raw materials of La2 O3 , CeO2 , and SiO2 were used. The powder mixture was first fired at 1350 ∘ C for 2 h in a flowing NH3 − 1.0 vol% CH4 gas mixture, then at 1450 ∘ C for 1 h in NH3 − 0.5 vol% CH4 gas, and finally treated at

Nitride and Oxynitride Phosphors for Light Emitting Diodes

Intensity (a.u.)

0.14 at% (X = 0.03) 0.55 at% (X = 0.1) 1.10 at % (X = 0.2)

250

300

157

405 nm

2.75 at % (X = 0.5) 5.5 at % (X = 1.0)

350 Wavelength (nm) (a)

400

450

Ce Concentration 0.55 at %(x = 0.1) Intensity (a.u.)

2.75 at %(x = 0.5) 5.5 at %(x = 1.0)

0.28 at %(x = 0.05) 400

450

500

550

600

650

700

Wavelength (nm) (b) Figure 4.16 Excitation (a) and emission (b) spectra of LaAl(Si6−z Alz )(N10−z Oz ) with varying Ce concentrations. Source: Takahashi et al. 2007 [100]. Reproduced with permission of AIP Publishing LLC

1500 ∘ C for 12 h in N2 . The excitation spectrum, covering the range of 240–400 nm, shows two distinct bands at 255 and 355 nm, respectively (Figure 4.17). The crystal field splitting is estimated as 1.10 × 104 cm−1 . LaSi3 N5 ∶Ce has a broad emission band centered at 422–444 nm, and a FWHM of 79–95 nm. The band gap calculated from the diffuse reflectance spectrum is 4.43 eV. Upon the 355 nm excitation, the luminescence intensity remains at 85% of that measured at room temperature, and the thermal quenching temperature is around 260 ∘ C. The absorption efficiency, IQE and EQE of the sample with 10 mol% Ce are 49.8, 68.9, and 34.3%, respectively, when excited at 380 nm. 4.3.1.2.3

Sr3 Si2 O4 N2 :Ce3+

The crystal structure and photoluminescence of Sr3 Si2 O4 N2 were reported by Wang et al. [81]. It has a cubic symmetry and space group of Pa3 (a = 15.6593 Å). The structure

Materials for Solid State Lighting and Displays

PL intensity

158

250

300

350

400 450 500 Wavelength (nm)

550

600

Figure 4.17 Excitation and emission spectra of LaSi3 N5 ∶Ce3+ (10 mol%). The emission spectrum was measured upon 355 nm excitation, and the excitation spectrum was monitored at 430 nm

RF

100

5000 EX

EM 80

λEM = 520 m

3000

λEX = 320 nm

60

2000 40

Reflection (%)

Intensity (a.u.)

4000

λEX = 400 nm

1000

20 0 200

300

600 400 500 Wavelength (nm)

700

800

Figure 4.18 Photoluminescence and diffuse reflectance spectra of Sr3 Si2 O4 N2 ∶Ce3+ . Source: Wang et al. 2012 [81]. Reproduced with permission of American Chemical Society

of Sr3 Si2 O4 N2 is built on 12 isolated and highly corrugated rings that consist of 12 vertex-sharing [SiO2 N2 ] tetrahedra with bridging N and terminal O. The Sr ions are accommodated in the three-dimensional tunnels formed by the 12 tetrahedra, and have 6 distinct crystallographic sites (CN=6–8). The bond lengths of Sr − O and Sr − N are in the range of 2.39 − 3.16 and 2.61 − 3.10 Å, respectively. Sr3 Si2 O4 N2 has an indirect band structure and a band gap of 2.84 eV. Sr3 Si2 O4 N2 ∶Ce3+ has a broad emission band with a maximum at 520 nm and a FWHM of 120 nm (Figure 4.18). The excitation spectrum shows two distinct bands at 280 − 350 nm and 380 − 420 nm. The concentration quenching occurs at 2 mol% for this green phosphor.

159

PL intensity (a.u.)

Nitride and Oxynitride Phosphors for Light Emitting Diodes

200

300

400 500 600 Wavelength (nm)

700

800

Figure 4.19 Excitation and emission spectra of CaAlSiN3 ∶Ce3+ . The emission spectrum was measured upon 450 nm excitation, and the excitation spectrum was monitored at 580 nm

4.3.1.2.4

CaAlSiN3 :Ce3+

Li et al. reported the photoluminescence properties of Ce3+ -activated CaAlSiN3 [104]. CaAlSiN3 ∶Ce3+ has an excitation spectrum showing two principle bands at 200–330 nm and 350–560 nm (Figure 4.19). The excitation band of Ce3+ consists of five Gaussian sub-bands at about 259, 313, 370, 421, and 483 nm with a larger crystal field splitting of 14 000 cm−1 . The fundamental absorption edge of CaAlSiN3 is estimated to be about 5.3 eV. The emission spectrum of CaAlSiN3 ∶Ce3+ shows a broadband peaking at 570–603 nm. Upon excitation in the blue light range, the absorption increases as the Ce3+ concentration increases, and reaches the saturation value of 78% with 2 mol% Ce3+ . The IQE and EQE of 80 and 56%, respectively, are maximized at 1 mol% Ce. CaAlSiN3 ∶Ce shows a small thermal quenching. The relative luminescence intensity maintains 87 and 64% at 150 and 300 ∘ C, respectively. The thermal quenching temperature is about 300 ∘ C for samples with Ce3+ concentration less than 2 mol%. The high quantum efficiency and thermal stability of the yellow CaSiAlN3 ∶Ce3+ enables it to be used in white LEDs. 4.3.1.2.5

CaSiN2 ∶Ce3+

Le Toquin and Cheetham synthesized Ce3+ -doped CaSiN2 by firing the powder mixture of Ca3 N2 , Si3 N4 , and CeO2 at 1300–1500 ∘ C under flowing N2 in a tube furnace [56]. Differing from the orthorhombic phase reported by Gal et al.,108 the Ce3+ -activated CaSiN2 has the cubic structure with cell parameter of a = 14.8822(5) Å [108]. CaSiN2 ∶Ce3+ has a broad emission spectrum extending from 550 to 700 nm with a FWHM of 80 nm. The emission band is centered at 625 nm, which is much longer than the Ce3+ emission in other hosts. The excitation band has a maximum at 535 nm and covers the spectral range of 425–575 nm. The Stokes shift is estimated as 2700 cm−1 . On the other hand, Wang et al. reported that the CaSiN2 phase crystallized in an orthorhombic unit cell with the space group Pbca (no. 61) and cell parameters a = 5.1334(3) Å, b = 10.3090(5) Å, c = 14.5756(5) Å, Z = 16 [109]. The CaSiN2 ∶Ce, Li phosphors were

160

Materials for Solid State Lighting and Displays 4000 EX EM Fitted

EM

EX 3000 Intensity (a.u.)

524

570

2000

1000

λEX = 440 nm

λEM = 530 nm

0 300

400

500 600 Wavelength (nm)

700

Figure 4.20 Excitation and emission spectra of CaSiN2 ∶Ce, Li with an orthorhombic structure. Source: Wang et al. 2013 [109]. Reproduced with permission of Royal Society of Chemistry

prepared at 1550 ∘ C in a N2 ∕H2 (6%) atmosphere by solid state reaction using Ca3 N2 , Si3 N4 , CeCl3 , and LiF as the starting materials. Ce3+ ∕Li+ co-doped CaSiN2 emits a yellow band peaking at 530 nm (FWHM ∼ 120 nm), which originates from the 5d → 4f transition of Ce3+ (Figure 4.20). The emission of CaSiN2 ∶Ce3+ can be tuned by cation substitutions. A larger cation on the Si site decreases the emission wavelength by ∼70 nm. For 10% Al substitution, the excitation spectrum is also blue-shifted from the peak maximum of 535 nm to 475 nm. The blue-shift of the emission band is also achieved by Mg substitution for Ca. On the contrary, the Sr substitution on the Ca site results in a red-shift of the emission (from 625 to 640 nm) [56]. 4.3.1.3 4.3.1.3.1

Other Rare Earth-Doped Nitride and Oxynitride Phosphors Yb2+ -doped

Phosphors Similar to Eu2+ and Ce3+ , Yb2+ also exhibits 4f–5d fluorescence spectra as its ground state has the xenon closed shell plus a 4f 14 configuration [97]. The observed photoluminescence spectra of Yb2+ are ascribed to the 4f 14 → 4f 13 5d interconfigurational transitions. Ca-𝛂-sialon:Yb2+ Xie et al. investigated the photoluminescence properties of Yb2+ doped Ca-α-sialon [110]. The excitation spectrum of Ca-α-sialon∶Yb2+ shows a number of bands centered at 219, 254, 283, 307, 342, and 445 nm, respectively, indicating that it can be excited by the blue light. An intense green emission is achieved for Ca-α-sialon∶Yb2+ that has a single broadband with a peak maximum at 549 nm. The concentration quenching occurs at 0.5 mol% Yb2+ . The CIE color coordinates are x = 0.323 and y = 0.601 for Ca-α-sialon∶Yb2+ (m = 2, n = 1).

161

PL intensity (a.u.)

Nitride and Oxynitride Phosphors for Light Emitting Diodes

250

300

350

400

450 500 550 Wavelength (nm)

600

650

700

Figure 4.21 Excitation and emission spectra of Ca-𝛼-sialon∶Yb2+ (m = 2, n = 1, 0.5 mol% Yb2+ ). The emission spectrum was measured upon 450 nm excitation, and the excitation spectrum was monitored at 550 nm

The emission intensity of Ca-α-sialon∶Yb2+ at 150 ∘ C is about 74% of that measured at room temperature. The thermal quenching temperature is 234 ∘ C, and the activation energy for thermal quenching is calculated as 0.23 eV (Figure 4.21). SrSi2 O2 N2 :Yb2+ Bachmann et al. studied the photoluminescence of Yb2+ -doped SrSi2 O2 N2 and compared it with that of the Eu2+ -doped version [111]. The excitation spectrum of SrSi2 O2 N2 ∶Yb2+ covers a large spectral range of 100–550 nm, and has several distinct bands centered at 180, 230, 275, 310, 400, and 455 nm, respectively. The emission band has a peak maximum at 620 nm. Compared with SrSi2 O2 N2 ∶Eu2+ , SrSi2 O2 N2 ∶Yb2+ shows a larger crystal field splitting, redshifted emission (620 nm vs 540 nm), larger Stokes shift (0.54 eV vs 0.42 eV) and FWHM value (0.32 eV vs 0.30 eV). The large crystal field splitting is due to the large effective nuclear charge and the smaller atomic radius of Yb2+ compared with Eu2+ . This type of anomalous red-shifted emission is attributable to the localization of the lowest energy excited state in the conduction band. When the excited state is at energies higher than the conduction band edge, excitation into the excited state is followed by photoionization and trapping of the electron close to the Yb2+ impurity forming an impurity-trapped exciton state. The assignment of the Yb emission to the impurity-trapped exciton emission is supported by the relatively short luminescence decay time. Ruan et al. prepared Eu2+ -Yb2+ co-doped SrSi2 O2 N2 and investigated the energy transfer between Eu2+ and Yb2+ [112]. In contrast with the single Yb2+ -doped sample, an additional five excitation bands centered at 260, 340, 360, 400, and 450 nm are observed in the Eu2+ -Yb2+ co-doped sample, which is due to the 4f–5d transition of Eu2+ in SrSi2 O2 N2 (Figure 4.22). Upon the excitation at 400 nm, the co-doped phosphor exhibits a broadband yellow emission peaking at 604 nm and a large FWHM of 170 nm. The highest energy

162

Materials for Solid State Lighting and Displays Eu

Eu-Yb

PL intensity (a.u.)

Yb

254 nm

No UV

(a) Ex.400 nm

(a) Em.542 nm

(b) Ex.400 nm

(b) Em.620 nm

(d) = (b)–(c) (c) Em.620 nm 200 300

365 nm

(c) Ex.400 nm

400 500 Wavelength (nm)

600

700

Figure 4.22 Excitation and emission spectra of (a) Sr0.98 Eu0.02 Si2 O2 N2 , (b) Sr0.96 Eu0.02 Yb0.02 Si2 O2 N2 , and (c) Sr0.98 Yb0.02 Si2 O2 N2 . (d) shows the difference between the excitation spectra of (b) and (c). The inset shows the samples with and without UV lamp irradiation at 254 and 365 nm. Source: Ruan et al. 2012 [112]. Reproduced with permission of John Wiley and Sons

transfer efficiency is estimated as ∼91.9%. The absorption efficiency and IQE are ∼75 and ∼31.9%, respectively. 4.3.1.3.2

β-sialon:Pr3+

The luminescence of Pr3+ , ranging from UV to infrared, originates from three excited states: 1 S0 , 3 P0 , and 1 D2 , and depends largely on the host crystals [97]. Liu et al. investigated the luminescence of Pr3+ in β-sialon: and developed a red-emitting phosphor for white LEDs [113]. The β-sialon∶Pr3+ phosphor was synthesized at 1900 ∘ C for 2 h under 1.0 MPa N2 . The emission spectrum consists of several sharp lines in the range of 600–-650 nm. Three narrow bands centered at 613, 624, and 641 nm are observed upon the 460 nm excitation, which are, respectively, ascribed to the 1 D2 → 3 H4 , 3 P0 → 3 H6 , and 3 P0 → 3 F2 transitions of Pr3+ . When monitored at 624 nm, the excitation spectrum is composed of five peaks at 460, 470, 488, 498, and 509 nm, respectively. These peaks correspond to the 3 H4 → 3 P (J = 2, 1, 0) transitions. J 4.3.1.4

Mn2+ -doped 𝛾-alon

The luminescence of Mn2+ is extensively investigated in many hosts including oxides, sulfides, fluorides, and phosphates. Its emission color can be green or orange/red if Mn2+ occupies the tetragonal or octahedral site. There are very few reports on the luminescence of Mn2+ in nitride or oxynitride compounds. Here we introduce interesting Mn2+ -doped oxynitride phosphors – γ-alon∶Mn2+ and γ-alon∶Mn2+ ,Eu2+ [114]. γ-alon is a solid solution in the binary system of Al2 O3 –AlN, and has a defective cubic spinel structure with Fd3m space group. The Mg2+ -Mn2+ co-doped γ-alon was prepared by firing the powder mixture of Al2 O3 , AlN, MgO, and MnCO3 at 1800 ∘ C for 2 h under

Nitride and Oxynitride Phosphors for Light Emitting Diodes

H4 → 3F2

3

3

4→

Intensity

Intensity

3H

3 P0 → 3H6 z=0 z = 0.05 z = 0.1 z = 0.2 z = 0.3 z = 0.4 1D2 → 3H4 z = 0.5

P0 → 3H6

λex = 460 nm

λem = 624 nm

1D → 3H 2 4 3P ,3F 1 1 3P

0→

3F

163

3P → 3F 0 2

2

3H → 1D 4 2

400

440

480

520 560 Wavelength (nm)

600

640

580

600

620 640 Wavelength (nm)

(a)

660

(b)

1.0 0.9

Relative Intensity

0.8

0.8

0.6

0.7 0.6

0.4 0.5 y

0.2

0.4

x

0.3

0.0 0.0

0.5

1.0 Z

(c)

1.5

2.0 0.2 0.1 0.0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

x

(d)

Figure 4.23 (a) Excitation (𝜆em = 624 nm) and emission spectra (𝜆ex = 460 nm) of Si5.9 Al0.1 O0.1 N7.9 ∶Pr0.016 . (b) Comparison of emission spectra at different z values. (c) Effect of Al and O substitution on luminescent properties of 𝛽-sialon∶Pr3+ . (d) CIE chromaticity coordinates of 𝛽-sialon∶Pr3+ in wavelength range 600–650 nm. Inset shows 𝛽-sialon∶Pr3+ irradiated under a 254 nm UV lamp box. Source: Liu et al. 2011 [113]. Reproduced with permission of American Chemical Society

a 0.5 MPa nitrogen atmosphere. The optimal concentration of Mg2+ and Mn2+ is 10 and 7 mol%, respectively. The excitation spectrum of γ-alon∶Mn2+ consists of five bands centered at 340, 358, 381, 424, and 445 nm, originating from the electronic transitions of 6 A to 4 T (4 P), 4 E(4 G), 4 T2 , [4 E(4 G), 4 A(4 G)], and 4 T (4 G), respectively (Figure 4.24). 1 2 2 An intense excitation peak at 445 nm enables γ-alon∶Mn2+ to be a suitable phosphor for white LEDs using blue LED chips. The emission spectrum shows a single band peaking at 520 nm, due to the electronic transition of 4 T1 (4 G) → 6 A1 . The FWHM value of the emission band is about 44 nm, indicating a narrow band and high purity green phosphor suitable for white LED backlights. The IQE of this green phosphor is about 62% under

164

Materials for Solid State Lighting and Displays 1500

4E (4G)

4

T2 (4G)

A1 (4G)

1000 4A 4T

500

2

1

(4F) (4F) 4

4 T2 (4D) E (4G)

PL Intensity

PL intensity

4

4

T1 (4G)

4T (4P) 2

0

200

250

300 350 400 Wavelength (nm) (a)

450

500

450

475

500

525 550 575 Wavelength (nm) (b)

600

625

Figure 4.24 (a) Excitation and (b) emission spectra of 𝛾-alon∶Mn, Mg phosphors (7 mol% Mn2+ , 10 mol% Mg2+ ). The emission spectrum was measured upon 445 nm excitation, and the excitation spectrum was monitored at 518 nm

PL intensity

QE, Absorption

0.8

γ-alon: 2% Eu γ-alon: 5% Mn γ-alon: 5% Mn,2% Eu

200

300

471 nm 517 nm

400

500

Wavelength (nm) (a)

600

Absorption Inter QE Exter QE

0.6

5% Mn 10% Mg 2% Eu

0.4

0.2

0.0 300

350

400

450

500

550

Wavelength (nm) (b)

Figure 4.25 (a) Photoluminescence spectra and (b) quantum efficiency of 𝛾-alon∶Mn2+ , Eu2+ . Source: Liu et al. 2015 [115]. Reproduced with permission of American Chemical Society

445 nm excitation. As the Mn2+ absorption is as low as 25%, the EQE of the as-synthesized sample is only about 15%. Great efforts are needed to enhance the absorption efficiency of this narrow band green phosphor. To improve the absorption efficiency of Mn2+ -doped γ-alon, the energy transfer between Eu2+ and Mn2+ was utilized [115]. When Eu2+ is co-doped with Mn2+ , the excitation spectrum of γ-alon∶Mn2+ , Eu2+ changes significantly, showing a very intense broadband covering the spectral range of 200–400 nm due to the 4f–5d transitions of Eu2+ and several weak bands of Mn2+ (Figure 4.25). γ-alon∶Mn2+ , Eu2+ is therefore suitable for near UV LEDs. Under 365 nm excitation, the emission spectrum of γ-alon∶Mn2+ , Eu2+ also shows a narrow band centered at 517 nm, and the emission intensity is nine times higher than the solely Mn2+ -doped γ-alon. This is evidence of the energy transfer between Eu2+ and Mn2+ . The mechanism of energy transfer is a resonant type via a dipole−dipole mechanism. The EQE of γ-alon∶Mn2+ , Eu2+ is 49% under 365 nm excitation, which is seven times higher than the Eu2+ -free γ-AlON∶Mn2+ .

Nitride and Oxynitride Phosphors for Light Emitting Diodes

4.4

165

Emerging Nitride Phosphors and Their Synthesis

Material scientists or chemists never stop the pursuit of novel nitride phosphors that have new crystal structures or promising photoluminescence properties. With rapid advances in solid state lighting technologies and applications, it is urgently required that white LEDs should have a much wider color gamut for LCD backlighting, and higher luminous efficacy and color rendering for general illumination. To meet these requirements, phosphors should have narrow-band red/green emissions (typically for backlighting applications), high quantum efficiency, and high thermal stability. To search for such phosphors requires finding a way to rapidly discover new host crystals, and the need for greater understanding of the relation between the photoluminescence properties and crystal/electronic structure. Recently, many efforts have been made to discover novel nitride and oyxnitride phosphors. Schnick and co-workers concentrated on the traditional single crystal method to find nitride compounds with interesting crystal structures [14, 22–24, 63, 73]. The material systems that they are now focusing on are alkaline earth nitridoaluminates [i.e., SrLiAl3 N4 , CaLiAl3 N4 , M[Mg2 Al2 N4 ] (M = Ca, Sr, Ba)] and double nitrides (i.e., Mg3 GaN3 and Ba3 Ga3 N5 ) [116–120]. Park et al. used a solid state combinatorial chemistry method in combination with a heuristic optimization approach to discover new nitride and oxynitride phosphors. These phosphors include Ca15 Si20 N30 O10 ∶Eu2+ , Ba1.5 Ca0.5 Si5 N6 O3 ∶Eu2+ , and La4−x Cax Si24 O3+x N18−x ∶Eu2+ [78–80]. Hirosaki at al. applied a single-particle-diagnosis approach to search for new nitride and oxynitride phosphors in a very rapid and efficient way [59]. This method enables the crystal structure to be determined and to characterize the photoluminescence of a very tiny single crystal (several tens of micrometers in size) distinguished from a randomly prepared powder mixture, saving the time for the large-size single crystal growth as well as the structural determination of the powder by X-ray diffraction. Several nitride and oxynitide phosphors, such as Ba5 Si11 Al7 N25 ∶Eu2+ , BaSi4 Al3 N9 ∶Eu2+ , and Ba2 LiSi7 AlN12 ∶Eu2+ , have been discovered by this approach [59, 64]. Although a number of nitride and oxynitride phosphors have been discovered as emerging phosphors, only a few of them will be introduced in this section on account of their unique crystal structure and promising photoluminescence. Other new luminescent nitride and oxynitride phosphors can be found in the literature. 4.4.1

Narrow-Band Red Nitride Phosphors

Pust et al. discovered novel red-emitting nitride phosphors with narrow bands, which are suitable for improving the luminous efficacy and color gamut of white LEDs [116–118]. These phosphors include Sr[LiAl3 N4 ]∶Eu2+ , M[Mg2 Al2 N4 ]∶Eu2+ (M = Ca, Sr, Ba), Sr[Mg2 Ga2 N4 ], and Sr[Mg3 GeN4 ]. All these phosphor hosts are isostructural to UCr4 C4 -type with a statistical distribution of the tetrahedral network cations Li/Al, Mg/Al, or Ga/Mg, respectively. Attractive narrow-band red emissions are obtained due to the highly symmetric cuboid-like coordination of the heavy atom site. Pust et al. developed different substitutional variants based on AB2 C2 X4 and ABC3 X4 structures containing Ga/Mg, Al/ Mg, Li/Al, or Mg/Si on the tetrahedrally coordinated sites, leading to red-emitting nitridomagnesosilicates M[Mg3 SiN4 ] (M = Ca, Sr, Eu) and nitridoaluminates Ca[LiAl3 N4 ]∶Eu2+ [116–118].

Materials for Solid State Lighting and Displays

1.0

1.0

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.0

PL intensity

Reflectance/Excitation

166

0.0 300

400

500 600 Wavelength (nm)

700

Figure 4.26 Diffuse reflectance, excitation and emission spectra of Sr[LiAl3 N4 ]∶Eu2+ . Source: Pust et al. 2104 [116]. Reproduced with permission of Nature Publishing Group

4.4.1.1

Sr[LiAl3 N4 ]∶Eu2+

Sr[LiAl3 N4 ]∶Eu2+ was synthesized by firing the stoichiometric mixture of LiAlH4 , AlN, SrH2 , and EuF3 at 1000 ∘ C for 2 h in a N2 − 5%H2 mixing gas atmosphere [116]. + Sr[LiAl3 N4 ]∶Eu2 is isotypic to the oxoplumbate Cs[Na3 PbO4 ], and crystallizes in the triclinic space group P1 (no. 2) with unit cell parameters a = 5.86631(12)Å, b = 7.51099 (15)Å, c = 9.96545(17)Å, α = 83.6028(12)o , β = 76.7720(13)o and γ = 79.5650(14)o . The structure consists of a highly condensed and rigid framework of ordered edge- and corner-sharing AlN4 and LiN4 tetrahedra that form channels of vierer rings along [011]. Strontium atoms occupy two distinct crystallographic sites, and both are coordinated to eight nitrogen atoms in a highly symmetric cuboid-like environment. Sr[LiAl3 N4 ]∶Eu2+ shows a broad excitation band having a maximum at 466 nm (Figure 4.26). Very interestingly, the emission spectrum displays an extremely narrow band centered at 654 nm with FWHM of 50 nm. This is very unusual for the Eu2+ luminescence, and is attributable to the cuboidal coordination of Eu2+ . The IQE and EQE is about 76 and 52%, respectively, upon 440 nm excitation. Similar narrow-band red emissions are also seen in nitridomagnesoaluminates M[Mg2 Al2 N4 ](M = Ca, Sr, Ba, Eu) and nitridomagnesogallate Ba[Mg2 Ga2 N4 ] [118]. 4.4.1.2

Sr[Mg3 SiN4 ]:Eu2+

The nitridomagnesosilicates M[Mg3 SiN4 ] (M = Ca, Sr, Eu) are isostructural to the lithosilicate Na[Li3 SiO4 ] [118]. They crystallize in space group I41/a (no. 88). The lattice parameters are a = 11.424(2), c = 13.445(3) Å for Ca, a = 11.495(2), c = 13.512(3) Å for Sr, and a = 11.511(4), c = 13.552(4) Å for Eu, respectively. Their crystal structure consists of SiN4 tetrahedra that are connected via two edges and one corner to MgN4 tetrahedra, leading to the formation of strands running along [001]. The strands of SiN4

Nitride and Oxynitride Phosphors for Light Emitting Diodes

167

and MgN4 tetrahedra are connected to each other by common corners in an up−down sequence, forming vierer rings with one-half of them centered by M2+ . This results in a rigid network and a maximum degree of condensation (i.e., atomic ratio (Mg, Si)∶N) κ = 1. Sr[Mg3 SiN4 ]∶Eu2+ has the most narrow emission band centered at 615 nm and a FWHM of 43 nm, upon 440 nm excitation. The Stokes shift of Sr[Mg3 SiN4 ]∶Eu2+ is calculated as small as ∼ 772 cm−1 , which accounts for its good thermal behavior. 4.4.2

Narrow-Band Green Nitride Phosphors

Narrow-band green phosphors are essential for wide color gamut LED backlights. β-sialon∶Eu2+ is a green phosphor that has the narrowest emission band, so it is commonly used to improve the color gamut of the LCD backlight. With advances in quantum dot (QD) technologies, the color gamut of devices using QDs is much larger than that using β-sialon∶Eu2+ because QDs have much narrower and tunable emission bands. To take advantage of the low cost, no toxicity, and high stability of inorganic phosphors, it is necessary to search for new green phosphors with much narrower bands and shorter emission wavelengths. 4.4.2.1

Ba2 LiAlSi7 N12 :Eu

Takeda et al. discovered an interesting narrow-band green phosphor, Ba2 LiSi7 AlN12 ∶Eu2+ [64]. It crystallizes in the orthorhombic crystal system with the space group Pnnm (no.58). The lattice parameters are a = 14.0993(2) Å, b = 4.8967(1)Å, and c = 8.0719(1) Å. The crystal structure consists of corner-sharing (Si, Al)N4 corrugated layers and edge-sharing (Si, Al)N4 and LiN4 tetrahedra. Ba(Eu) is located at the one dimensional channel in a zigzag way, and coordinated to 11 nitrogen atoms. The distance between Ba and N ranges from 2.93 to 3.32 Å with an average distance of 3.12 Å. Ba2 LiSi7 AlN12 ∶Eu2+ was synthesized by a gas pressure sintering method. The starting powder was Ba3 N2 , EuN, Si3 N4 , AlN, and Li3 N, mixed in a cation molar ratio of Ba∕Eu∕Si∕Al∕Li = 0.80∶0.20∶0.58∶6.42∶3.00 in a glove box under a nitrogen atmosphere. The powder mixture was then loaded in boron nitride crucibles and fired at 1800 ∘ C for 2 h under 1.0 MPa nitrogen atmosphere. Ba2 LiSi7 AlN12 ∶Eu2+ shows a narrow emission band with the maximum at 515 nm and a FWHM of 61 nm (Figure 4.27). This makes it to be potential green phosphor used for backlights. At 200 ∘ C, its luminescence maintains 86% of that measured at room temperature. The internal quantum efficiency is 79% when excited at 405 nm. 4.4.2.2

Ba[Li2 (Al2 Si2 )N6 ]∶Eu2+

The crystal structure and luminescence of Ba[Li2 (Al2 Si2 )N6 ]∶Eu2+ and its related Mg-substituted compounds Ba[(Mg2–x Lix )(Al4–x Six )N6 ]∶Eu2+ (x = 1.6, 1.8) were reported by Strobel et al. (Figure 4.28) [121]. All these compounds have the tetragonal crystal system and space group P4∕ncc (no. 130). The crystal structure of these compounds consists of a highly condensed anionic tetrahedra framework of disordered (Li, Mg)N4 and (Al, Si)N4 units that are connected to each other by common edges and corners. The Ba atoms are coordinated to eight nitrogen atoms, forming a truncated square pyramid.

Materials for Solid State Lighting and Displays

PL intensity (a.u.)

168

50 μm (a)

350 400 450 500 550 600 650 700 Wavelength (nm) (b)

Figure 4.27 (a) A single crystal and (b) excitation and emission spectra of Ba2 LiSi7 AlN12 ∶ Eu2+ . Source: Takeda et al. 2015 [64]. Reproduced with permission of American Chemical Society

1.0

Intensity (a.u.)

0.8 0.6 0.4 0.2 0.0 400

450

500

550 600 650 Wavelength (nm)

700

750

800

Figure 4.28 Excitation and emission spectra of Ba[Li2 (Al2 Si2 )N6 )∶Eu2+ . The inset shows the phosphor under UV irradiation. Source: Strobel at al. 2015 [121]. Reproduced with permission of American Chemical Society

Eu2+ -doped Ba[Li2 (Al2 Si2 )N6 ] was synthesized using high purity BaF2 , AlF3 , Si(NH)2 , Li3 N, and EuF3 . Li pieces were used as a flux. The powder was ground and filled into a reaction vessel. The vessel was placed in a silica tube and heated to 950 ∘ C for 24 h in a tube furnace after evacuation. Ba[Li2 (Al2 Si2 )N6 ]∶Eu2+ (1 mol%) exhibits a narrow-band emission centered at 532 nm and a FWHM of 57 nm. The color coordinates (CIE) are x = 0.30 and y = 0.64.

Nitride and Oxynitride Phosphors for Light Emitting Diodes

4.5

169

Applications of Nitride Phosphors

As mentioned before, nitride and oyxnitride phosphors are superior down-conversion luminescent materials for white LEDs, due to their interesting photoluminescence, high quantum efficiency, and small thermal quenching. In addition to the mostly used broadband garnet phosphors (i.e., YAG∶Ce3+ ), nitride phosphors are now commonly used in white LEDs to improve the reliability, color rendering index, or color gamut. There are basically two main applications for white LEDs: general illumination; and backlighting. In these two cases, the selection of phosphors is different. In this section, the use of nitride and oxynitride phosphors in white LEDs for these different applications will be summarized. 4.5.1 4.5.1.1

General Lighting High Efficiency White LEDs

The simplest way to fabricate white LEDs is to combine a yellow phosphor with a blue LED chip. The luminous efficacy of white LEDs is much higher than other combinations because the spectrum is closer to the human eye sensitivity curve. Yttrium aluminum garnet (YAG∶Ce3+ ) is the most popular yellow phosphor that has a wide band width and high quantum efficiencies. On the other hand, the use of YAG∶Ce3+ leads to high color temperatures (>4000 K) due to the lack of enough red spectral component in YAG∶Ce3+ . The red component of YAG∶Ce3+ can be enhanced by Gd- or Tb-substitution, but the thermal stability of these redshifted YAG∶Ce3+ is reduced. This trade-off can be solved by searching for alternative yellow phosphors with longer emission wavelength and small thermal quenching. Table 4.8 lists the optical properties of one-phosphor-converted white LEDs using a single yellow phosphor [28, 104, 122–124]. As seen, the orange phosphors, such as Ca-α-sialon∶Eu2+ , CaAlSiN3 ∶Ce3+ , and La3 Si6 N11 ∶Ce3+ , are able to prepare warm white, which can be used for in-home lighting. The yellow phosphors, including Li-α-sialon∶Eu2+ and (Sr0.5 Ba0.5 )Si2 O2 N2 ∶Eu2+ , enable cool white to be fabricated. In addition, the color rendering index of ∼60 is usually small for Eu2+ -doped phosphors due to their narrow emission band. For Ce3+ -doped yellow phosphor, the color rendering index is equal to that using YAG∶Ce3+ . The luminous efficacies shown in Table 4.8 are difficult to compare with each other because different blue LED chips are used in each case. A typical electroluminescence spectrum of one-phosphor-converted white LEDs is given in Figure 4.29. Two distinct emission bands are seen, originating from the LED chip and the yellow phosphor, respectively. 4.5.1.2

High Color Rendition White LEDs

The color rendering index of one-phosphor-converted white LEDs is always less than 80, making them unable to be used for high quality lightings for the home, medical operations, exhibitions, supermarkets, and so on. To achieve high color rendering index, multi-phosphors must be used to compensate for the spectral part that is lacking in a single phosphor. In addition, it is hard to tune the color temperature using a single phosphor, but it is easy to obtain tunable color temperatures by just changing the phosphor ratios.

170

Materials for Solid State Lighting and Displays

Table 4.8 Optical properties of white LEDs using a single yellow nitride or oxynitride phosphor Yellow phosphor

Maximum emission (nm)

FHWM (nm)

Luminous efficacy (lm/W)

CCT (K)

CRI

Ref.

Ca-α-sialon:Eu Li-α-sialon:Eu (Sr0.5 Ba0.5 )Si2 O2 N2 ∶Eu CaAlSiN3 ∶Ce La3 Si6 N11 ∶Ce

585 573–575 560 570–603 577–581

94 90 93 130 117

25.9 43 117 50 —

2750 6150 6000 3722 2600–3800

58 62 61 70 76

[122] [123] [124] [104] [28]

CCT, Correlated color temperature; CRI, color rendering index.

Total spectral flux (mW/nm)

0.8 CCT = 5,970 K η = 117 lm/W (0.323, 0.328) Ra = 61

0.6

0.4

0.2

0.0 400

500

600 Wavelength (nm)

700

800

Figure 4.29 Electroluminescence spectrum of the one-phosphor-converted white LEDs using (Sr0.5 Ba0.5 )Si2 O2 N2 ∶Eu2+ yellow phosphor. Source: Wang et al. 2105 [124]. Reproduced with permission of Royal Society of Chemistry

As shown in Table 4.9, the color rendering index is significantly improved by using multi-phosphors [67, 116, 125–129]. Super-high color rendering index can be obtained by using four phosphors (Figure 4.30), with the blue-green BaSi2 O2 N2 ∶Eu2+ filling the gap between the blue LED chip and the green phosphor. In addition, the R9 index, reflecting the red color, is also remarkably increased by using red phosphors. It has a negative value in one-phosphor-converted white LEDs. In addition, high color rendering white LEDs can also be attained by combining a near UV LED with red, green, and blue phosphors. Takahashi et al. reported high color rendering white LEDs by using a 405 nm LED to pump JEM∶Ce3+ , β-sialon∶Eu2+ , Ca-α-sialon∶Eu2+ , and CaAlSiN3 ∶Eu2+ phosphors [130]. The color rendering index was as high as 95, and the luminous efficacy was 20 lm/W. Wang et al. combined a 380 nm LED with AlN-SiC∶Eu2+ , Lu3 Al5 O12 ∶Ce3+ , and (Sr, Ca)AlSiN3 ∶Eu2+ , and achieved a color rendering index of 95 and R9 = 72 [131].

Nitride and Oxynitride Phosphors for Light Emitting Diodes

Table 4.9

171

Optical properties of white LEDs using multi-phosphors

Phosphor blend

Luminous efficacy (lm/W)

CCT (K)

CRI

R9

Ref.

Ca-α-sialon∶Yb2+ + Sr2 Si5 N8 ∶Eu2+

17–23

2744–6508

83

—

[125]

SrSi2 O2 N2 ∶Eu2+ + Sr2 Si5 N8 ∶Eu2+

25

3100

89

56

[126]

SrSi2 O2 N2 ∶Eu2+ + CaSiN2 ∶Ce3+

30

5206

90.5

—

[127]

Sr3 Si13 Al3 O2 N21 ∶Eu2+ + (Ca, Sr)2 SiO4 ∶Eu2+

56–62

3230–6450

82–88

—

[67]

β-sialon∶Eu2+ + Ca-α-sialon∶Eu2+ + CaAlSiN3 ∶Eu2+

25–28

2840–6580

81–88

77–99

[128]

2700

91

57

[116]

2900–6380

95–98

89

[129]

LuAG∶Ce3+ + Sr2 Si5 N8 ∶Eu2+ + Sr[LiAl3 N4 ]∶Eu2+ BaSi2 O2 N2 ∶Eu2+ + β-sialon∶Eu + Ca-α-sialon∶Eu2+ + CaAlSiN3 ∶Eu2+

28–35

CCT, Correlated color temperature; CRI, color rendering index.

D CCT (K) 6380 x y Ra lm/W

0.315 0.329 96 35

N 4900

W WW 4280 3540

L 2900

0.348 0.357 96 33

0.371 0.406 0.446 0.378 0.397 0.411 95 97 98 31 30 28

Intensity (a.u.)

Incadescent light (L) Warm white (WW) White (W) Normal white (N) Day white (D)

400

450

500

550 600 650 Wavelength (nm)

700

750

800

Figure 4.30 Electroluminescence spectrum of multi-band white LEDs, showing extra-high color rendering index. Source: Kimura et al. 2007 [129]. Reproduced with permission of AIP Publishing LLC

172

4.5.2

Materials for Solid State Lighting and Displays

LCD Backlight

Cold cathode fluorescent lamps (CCFLs), which have a color gamut of 65–75% of the National Television Standard Committee (NTSC) standard, are commonly used as standard backlights for large size LCDs. As there is much concern about backlight units with larger color gamut, higher brightness, smaller volume, lower power consumption, high dimming ratio, and Hg-free, white LEDs are thus considered a promising alternative to CCFLs. Compared with those utilizing the combination of individual red, green, and blue LED chips, phosphor-converted white LEDs are more reliable and simple. Currently, the two-band white LED using YAG:Ce is used practically as backlights for small size LCDs such as cell phone displays [3]. Although it has high luminous efficiency, its color gamut is narrow because of the color deficiency in the red- and blue-green spectral region. To solve this problem, three-band white LEDs are one of the key technologies used to enhance the color gamut of displays, that is, individual green and red phosphors are required. There are quite a few phosphors that can be used in LED backlights because narrow-band phosphors are required to achieve wide color gamut and high efficiency. Among the discovered nitride and oxynitride phosphors, β-sialon∶Eu2+ is recognized as the most promising green phosphor for backlights because it has the narrowest emission band (about 55 nm) and highest thermal robustness among Eu2+ -doped green nitride and oxynitride phosphors. CaAlSiN3+ is one of the best red phosphors for backlights as it has a higher quantum efficiency and thermal stability. Table 4.10 summarizes the color gamut, luminous efficacy, and color temperature of LED backlights using nitride and oxynitride phosphors [132–135]. In general, two-phosphor-converted LED backlights have wider color gamut than the YAG∶Ce3+ -based one. The color gamut of the latter is close to that of the traditional CCFLs (which is less than 75% of the NTSC standard). By careful selection of both green and red phosphors, much higher color gamut can be reached. For example, the color gamut (CIE 1976) of 107% NTSC standard is obtained by using sharp β-sialon∶Eu2+ and K2 SiF6 ∶Mn4+ . As seen in Figure 4.31, the role of β-sialon∶Eu2+ is very crucial as it is in between the blue and red spectra, so that it determines the color purity (i.e., color gamut) and luminous efficacy of LED backlights. Sharp β-sialon∶Eu2+ has a higher color purity than the standard β-sialon∶Eu2+ as it has a shorter emission wavelength and a narrower emission band. Compared with CaAlSiN3 ∶Eu2+ , KSF∶Mn4+ has a much Table 4.10 Optical properties of phosphor-converted white LEDs for LCD backlights Phosphors

Green

Red

β-sialon∶Eu2+ Sr3 Si13 Al3 O2 N21 ∶Eu2+ Sr2 SiO4 ∶Eu2+ β-sialon∶Eu2+ YAG∶Ce3+

CaAlSiN3 ∶Eu2+ CaAlSiN3 ∶Eu2+ CaAlSiN3 ∶Eu2+ K2 SiF6 ∶Mn4+

Color temperature (K)

8620 12723 8000 8611 8000

Luminous efficacy (lm/W)

38 41 103 94 105

Color gamut (% NTSC) CIE 1931

CIE 1976

82.1 83.8 74.7 85.9 67.9

91.9 92.4 — 96.2 —

Ref.

[132] [133] [134] [135] [134]

400

173

Intensity (a.u.)

Normalized intensity

Nitride and Oxynitride Phosphors for Light Emitting Diodes

450

500

550

600

650

700

400

450

500

550

600

Wavelength (nm)

Wavelength (nm)

(a)

(b)

650

700

0.6 0.5

v'

0.4 0.3 0.2

CaAlSiN3: Eu2+ + β-sialon:Eu2+ K2SiF6: Mn4+ + β-sialon:Eu2+ NTSC

0.1 0.0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

u' (c)

Figure 4.31 Electroluminescence spectra of LED backlights after filtering: (a) blue LED + 𝛽-sialon∶Eu2+ + CaAlSiN3 ∶Eu2+ ; (b) 𝛽-sialon∶Eu2+ + K2 SiF6 ∶Mn4+ ; and (c) their color gamut in CIE 1976. Source: Xie et al. 2009 [132]. Reproduced with permission of The Japan Society of Applied Physics

narrower emission band and almost no emission longer than 700 nm, so the backlight using KSF∶Mn4+ has higher color gamut and luminous efficacy. These results also emphasize the importance and need for the development of green and red phosphors having much narrower emission bands.

References 1. 2. 3. 4.

Jüstel, T., Nikol, H. and Ronda, C.R. (1998) New developments in the field of luminescent materials for lighting and displays. Angew. Chem. Int. Ed., 37, 3084–3103. Feldmann, C., Justel, T., Ronda, C.R. and Schmidt, P. (2003) Inorganic luminescent materials: 100 years of research and application. Adv. Funct. Mater., 12, 511–516. Zhao, J., Jin, D., Schartner, E.P. et al. (2013) Single-nanocrystal sensitivity achieved by enhanced upconversion luminescence. Nature Nanotechnol, 8, 729–734. Shionoya, S., Yen, W.M. and Yamamoto, H. (2006) Phosphor Handbook, 2nd edn, CRC Press.

174

5.

Materials for Solid State Lighting and Displays

Xie, R.-J., Li, Y.Q., Hirosaki, N. and Yamamoto, H. (2011) Nitride Phosphors and Solid State Lighting, CRC Press. 6. Smet, P.F., Parmentier, A.B. and Poelman, D. (2011) Selecting conversion phosphors for white light-emitting diodes. J. Electrochem. Soc., 158, R37–R54. 7. Ye, S., Xiao, F., Pan, Y.X. et al. (2010) Phosphors in phosphor-converted white light-emitting diodes: Recent advances in materials, techniques and propertie. Mater. Sci. Eng. R., 71, 1–34. 8. Xie, R.-J. and Hirosaki, N. (2007) Silicon-based oxynitride and nitride phosphors for white LEDs – A review. Sci. Tech. Adv. Mater., 8, 588–600. 9. Xie, R.-J., Hirosaki, N., Sakuma, K. et al. (2004) Eu2+ -doped Ca-alpha-SiAlON: A yellow phosphor for white light-emitting diodes. Appl. Phys. Lett., 84, 5404–5406. 10. Hirosaki, N., Xie, R.-J., Kimoto, K. et al. (2005) Characterization and properties of green-emitting β-SiAlON∶Eu2+ powder phosphors for white light-emitting diodes. Appl. Phys. Lett., 86, 211905. 11. Setlur, A.A., Heward, W.J., Gao, Y. et al. (2006) Crystal chemistry and luminescence of Ce3+ -doped Lu2 CaMg2 (Si, Ge)3 O12 and its use in LED based lighting. Chem. Mater., 18, 3314–3322. 12. Shimomura, Y., Honma, T., Shigeiwa, M. et al. (2007) Photoluminescence and crystal structure of green-emitting Ca3 Sc2 Si3 O12 ∶Ce3+ phosphors for white light emitting diodes. J. Electrochem. Soc., 154, J35–J38. 13. Im, W.B., Fellows, N.N., DenBaars, S.P. et al. (2009) LaSr2 AlO5 , a versatile host compound for Ce3+ -based yellow phosphors: Structural tuning of optical properties and use in solids state white lighting. Chem. Mater., 21, 2957–2966. 14. Zeuner, M., Pagano, S. and Schnick, W. (2011) Nitridosilicates and oxonitridosilicates: From ceramic materials to structural and functional diversity. Angew. Chem. Int. Edn. Engl., 50, 7754–7775. 15. Daicho, H., Iwasaki, T., Enomoto, K. et al. (2012) A novel phosphor for glareless white light-emitting diodes. Nature Commun., 3, 1132. 16. Son, K.H., Singh, S.P. and Sohn, K.S. (2012) Discovery of novel phosphors for use in light emitting diodes using heuristics optimization-assisted combinatorial chemistry. J. Mater. Chem., 22, 8505–8511. 17. Gaido, G.K., Dubrovskii, G.P. and Zykov, A.M. (1974) Photoluminescence of. europium-activated MgSiN2 . Izv. Akad. Nauk. SSSR Neorg. Mater., 10, 485. 18. Endo, T. (1993) High pressure synthesis of periodic compounds and its optical and electrical properties, in New Functionality Materials, Vol. C (eds T. Tsumura, M. Doyama, and M. Seno), Elsevier, pp. 107–112. 19. Lee, S.S., Lim, S., Sun, S.S. and Wager, J.F. (1997) Photoluminescence and electroluminescence characteristics of CaSiN2 ∶Eu phosphor. Proc. SPIE, 3241, 75. 20. van Krevel, J.W.H., Hintzen, H.T., Metselaar, R. and Meijerink, A. (1998) Long wavelength Ce3+ emission in Y-Si-O-N Materials. J. Alloy Compd, 268, 272–277. 21. Uheda, K., Takizawa, H., Endo, T. et al. (2000) Synthesis and luminescent property of Eu3+ -doped LaSi3 N5 phosphor. J. Lumin., 87–89, 967–969. 22. Hoppe, H.A., Lutz, H., Mory, P. et al. (2000) Luminescence in Eu2+ -doped Ba2 Si5 N8 : fluorescence, thermoluminescence, and upconversion. J. Phys. Chem. Solids, 61, 2001–2006.

Nitride and Oxynitride Phosphors for Light Emitting Diodes

23.

24.

25. 26. 27.

28.

29.

30.

31.

32.

33.

34.

35.

36. 37. 38.

39.

175

Hecht, C., Stadler, F., Schmidt, P.J. et al. (2009) SrAlSi4 N7 ∶Eu2+ – A nitridoalumosilicate phosphor for warm white light (pc)LEDs with edge-sharing tetrahedral. Chem. Mater., 21, 1595–1601. Kechele, J.A., Hecht, C., Oeckler, O. et al. (2009) Ba2 AlSi5 N9 – A new host lattice for Eu2+ -doped luminescent materials comprising a nitridoalumosilicate framework with corner- and edge-sharing tetrahedral. Chem. Mater., 21, 1288–1295. Suehiro, T., Hirosaki, N., Xie, R. and Mitomo, M. (2005) Powder synthesis of Ca-𝛼-SiAlON as a host material for phosphors. Chem. Mater., 17, 308–314. Suehiro, T., Onuma, H., Hirosaki, N. et al. (2010) Powder synthesis of Y-𝛼-sialon and its potential as a phosphor host. J. Phys. Chem. C, 114, 1337–1342. Suehiro, T., Hirosaki, N., Xie, R.-J. and Sato, T. (2009) Blue-emitting LaSi3 N5 ∶Ce3+ fine powder phosphor for UV-converting white light-emitting diodes. Appl. Phys. Lett., 95, 051903. Suehiro, T., Hirosaki, N. and Xie, R.-J. (2011) Synthesis and photoluminescent properties of (La, Ca)3 Si6 N11 ∶Ce3+ fine powder phosphors for solid state lighting. ACS Appl. Mater. Interfaces, 3, 811–816. Li, H.L., Xie, R.-J., Hirosaki, N. and Yajima, Y. (2008) Synthesis and photoluminescence properties of Sr2 Si5 N8 ∶Eu2+ red phosphors by a gas-reduction and nitridation method. J. Electrochem. Soc., 155, J378–J381. Li, H.L., Xie, R.-J., Zhou, G.H. et al. (2010) A cyan-emitting BaSi7 N10 ∶Eu2+ phosphor prepared by gas reduction and nitridation for UV-pumping white LEDs. J. Electrochem. Soc., 157, J251–J255. Yin, L.J., Zhu, Q.Q., Yu, W. et al. (2012) Europium location in the AlN:Eu green phosphor prepared by a gas-reduction-nitridation route. Appl. Phys. Lett., 111, 053534. Suehiro, T., Xie, R.-J. and Hirosaki, N. (2014) Gas-reduction-nitridation synthesis of CaAlSiN3 ∶Eu2+ fine powder phosphors for solid state lighting. Ind. Eng. Chem. Res., 53, 2713–2717. Piao, X.Q., Horikawa, T., Hanzawa, H. and Machida, K.-I. (2006) Characterization and luminescence properties of Sr2 Si5 N8 ∶Eu2+ phosphor for white light-emitting-diode illumination. Appl. Phys. Lett., 88, 161908. Zhang, H., Horikawa, T., Hanzawa, H. et al. (2007) Photoluminescence properties of α-sialon∶Eu2+ prepared by carbothermal reduction and nitridation method. J. Electrochem. Soc., 154, J59–J61. Kurashima, T., Gundiah, G., Shimomura, Y. et al. (2010) Synthesis of Eu2+ -activated MYSi4N7 (M = Ca, Sr, Ba) and SrYSi4−x Alx N7−x Ox (x = 0-1green phosphors by carbothermal reduction and nitridation. J. Electrochem. Soc., 157, J64–J68. Yin, L.J., Xu, X., Yu, W. et al. (2010) Synthesis of Eu2+ -doped AlN phosphors by carbothermal reduction. J. Am. Ceram. Soc., 93, 1702–1707. Liu, L.H., Xie, R.-J., Li, W. et al. (2013) Yellow-emitting Y3 Si6 N11 ∶Ce3+ phosphors for white light-emitting diodes (LEDs). J. Am. Ceram. Soc., 96, 1688–1690. Li, S.X., Peng, X., Liu, X.J. and Huang, Z.R. (2014) Photoluminescence of CaAlSiN3 ∶Eu2+ -based fine red-emitting phosphors synthesized by carbothermal reduction and nitridation method. Opt. Mater., 38, 242–247. Piao, X., Machida, K., Horikawa, T. and Hanzawa, H. (2008) Synthesis of nitriodosilicate CaSr1−x Eux Si5 N8 (x = 0-1) phosphor by calcium cyanamide reduction for white light-emitting diode applications. J. Electrochem. Soc., 155, J17–J22.

176

40.

41.

42.

43.

44.

45.

46. 47.

48. 49.

50.

51. 52. 53. 54.

55.

56.

Materials for Solid State Lighting and Displays

Kim, H.S., Machida, K., Horikawa, T. and Hanzawa, H. (2013) Carbothermal reduction synthesis using CaCN2 as calcium and carbon sources for CaAlSiN3 ∶Eu2+ phosphor and their luminescence properties. Chem. Lett., 43, 533–534. Piao, X., Machida, K.-i., Horikawa, T. and Hanzawa, H. (2007) Self-propagating high temperature synthesis of yellow-emitting Ba2 Si5 N8 ∶Eu2+ phosphors for white light-emitting diodes. Appl. Phys. Lett., 91, 041908. Paio, X., Machida, K.-i., Horikawa, T. et al. (2007) Preparation of CaAlSiN3 ∶Eu2+ phosphors by the self-propagating high-temperature synthesis and their luminescence properties. Chem. Mater., 19, 4592–4599. Watanabe, H., Imai, M. and Kijima, N. (2009) Nitridation of AEAlSi for production of AEAlSiN3 ∶Eu2+ nitride phosphors (AE = Ca, Sr). J. Am. Ceram. Soc., 92, 641–648. Li, H.L., Xie, R.-J., Hirosaki, N. and Takeda, T. (2009) Synthesis and luminescence properties of orange-red-emitting M2 Si5 N8 ∶Eu2+ (M = Ca, Sr, Ba) light-emitting diode conversion phosphors by a simple nitridation of MSi2 . Int. J. Appl. Ceram. Technol., 6, 459–464. Yang, J.J., Song, Z., Bian, L. and Liu, Q.L. (2012) An investigation of crystal chemistry and luminescence properties of Eu-doped pure-nitride 𝛼-sialon fabricated by the alloy-nitridation method. J. Lumin., 132 (9), 2390–2397. Wang, B. and Callahan, M.J. (2006) Ammonothermal synthesis of III-nitride crystals. Cryst. Growth Des., 6, 1227–1246. Watanabe, T., Nonaka, K., Li, J.W. et al. (2012) Low temperature ammonothermal synthesis of europium-doped SrAlSiN3 for a nitride red phosphor. J. Ceram. Soc. Jpn., 120, 500–502. Li, J.W., Watanabe, T., Sakamoto, N. et al. (2008) Synthesis of a multinary nitride, Eu-doped CaAlSiN3 , from alloy at low temperature. Chem. Mater., 20, 2095–2105. Zeuner, M., Schmidt, P.J. and Schnick, W. (2009) One-pot synthesis of single-source precursors for nanocrystalline LED phosphors M2 Si5 N8 ∶Eu2+ (M = Sr, Ba). Chem. Mater., 21, 2467–2473. Inoue, K., Hirosaki, N., Xie, R.-J. and Takeda, T. (2009) Highly efficient and thermally stable blue-emitting AlN∶Eu2+ phosphors for ultraviolet white light-emitting diodes. J. Phys. Chem. C, 113, 9392–9397. Shioi, K., Hirosaki, N., Xie, R.-J. et al. (2008) Luminescence properties of . J. Mater. Sci., 43, 5659–5661. Huppertz, H. and Schnick, W. (1997) Edge-sharing SiN4 tetrahedra in the highly condensed nitridosilicate BaSi7 N10 . Chem. Eur. J., 3 (2), 249–252. Li, Y.Q., Delsing, A.C.A., Metslaar, R. et al. (2009) Photoluminescence properties of rare-earth activated BaSi7 N10 . J. Alloy Compd, 487, 28–33 Li, Y.Q., van Steen, J.E.J., van Krevel, J.W.H. et al. (2006) Luminescence properties of red-emitting M2 Si5 N8 ∶Eu2+ (M = Ca, Sr, Ba) LED conversion phosphors. J. Alloy Compd, 417, 273–279. Xie, R.-J., Hirosaki, N., Suehiro, T. et al. (2006) A sinple, effiecient synthetic route to Sr2 Si5 N8 ∶Eu2+ -based red phosphors for white light-emitting diodes. Chem. Mater., 18, 5578–5583. Le Toquin, R. and Cheetham, A.K. (2006) Red-emitting cerium-based phosphor materials for solid state lighting applications. Chem. Phys. Lett., 423, 352–356.

Nitride and Oxynitride Phosphors for Light Emitting Diodes

57.

58.

59.

60.

61.

62. 63.

64.

65. 66.

67.

68.

69.

70.

71. 72.

177

Li, Y.Q., Hirosaki, N., Xie, R.-J. et al. (2010) Synthesis, crystal and local electronic structures, and photoluminescence properties of red-emitting CaAlz SiN2+z ∶Eu2+ with orthorhombic structure. Int. J. Appl. Ceram. Technol., 7 (6), 787–802. Li, Y.Q., Fang, C.M., de With, G. et al. (2004) Preparation, structure and photoluminescence properties of Eu2+ and Ce3+ -doped SrYSi4 N7 . J. Solid State Chem., 177, 4687–4694. Hirosaki, N., Takeda, T., Funahashi, S. et al. (2014) Discovery of new nitridosilicate phosphors for solid state lighting by the single-particle-diagnosis approach. Chem. Mater., 26, 4280–4288. Ruan, J., Xie, R.-J., Hirosaki, N. et al. (2011) Nitrogen gas pressure synthesis and photoluminescent properties of orange-red SrAlSi4 N7 ∶Eu2+ phosphors for white light-emitting diodes. J. Am. Ceram. Soc., 94 (2), 536–542. Uheda, K., Hirosaki, N., Yamamoto, Y. et al. (2006) Luminescence properties of a red phosphor, CaAlSiN3 ∶Eu2+ , for white light-emitting diodes. Electrochem. Solid State Lett., 9 (4), H22–H25. Watanabe, H., Yamane, H., and Kijima, N. (2008) Crystal structure and luminescence of Sr0.99 Eu0.01 AlSiN3 . J. Solid State Chem., 181, 1848–1852. Schmiechen, S., Schneider, H., Wagatha, P. et al. (2014) Toward new phosphors for application in illumination-grade white pc-LEDs: The nitridomagnesosilicates Ca[Mg3 SiN4 ]∶Ce3+ , Sr[Mg3 SiN4 ]∶Eu2+ , and Eu[Mg3 SiN4 ]. Chem. Mater., 26, 2712–2719. Takeda, T., Hirosaki, N., Funahashi, S. et al. (2015) Narrow-band green-emitting phosphor Ba2 LiSi7 AlN12 ∶Eu2+ with high thermal stability discovered by a single particle diagnosis approach. Chem. Mater., 27, 5892–5898. Xie, R.-J., Hirosaki, N., Li, Y.Q. et al. (2010) Photoluminecence of (Ba1−x Eux ) Si6 N8 O (0.005 ≤ x ≤ 0.2) phosphors. J. Lumin., 130, 266–269. Liu, L.H., Xie, R.-J., Hirosaki, N. et al. (2013) Luminescence properties of a blue-emitting phosphor: (Sr1−x Eux )Si9 Al19 ON31 (0 < x ≤ 1). J. Solid State Chem., 207, 49–54. Fukuda, Y., Ishida, K., Mitsuishi, I. et al. (2009) Luminescence properties of Eu2+ -doped green-emitting Sr-sialon phosphor and its application to white light-emitting diodes. Appl. Phys. Express, 2, 012401. Tang, J.-Y., Xie, Wen-Jie, Huang, Kai et al. (2011) A high stable blue BaSi3 Al3 O4 N5 ∶ Eu2+ phosphor for white LEDs and display applications. Electrochem. Solid State Lett., 14 (8), J45–J47. Yamane, H., Shimooka, S., and Uheda, K. (2012) Synthesis, crystal structure and photoluminescence of a new Eu-doped Sr containing sialon (Sr0.94 Eu0.06 )(Al0.3 Si0.7 )4 (N0.8 O0.2 )6 . J. Solid State Chem., 190, 264–270. Park, W.B., Singh, S.P., and Sohn, K.S. (2014) Discovery of a phosphor for light emitting diode applications and its structural determination. Ba(Si, Al)5 (O, N)8 ∶Eu2+ . J. Am. Chem. Soc., 136 (6), 2363–2373. Bachmann, V.M., Ronda, R.C., Meijerink, A. et al. (2009) Color point tuning for (Sr, Ca, Ba)Si2 O2 N2 ∶Eu2+ for white LEDs. Chem. Mater., 21, 316–325. Braun, C., Seibald, M., Borger, S.L. et al. (2010) Materials properties and structural characterization of M3 Si6 O12 N2 ∶Eu2+ (M = Ba, Sr) – A comprehensive study on a promising green phosphor for pc-LEDs. Chem. Eur. J., 16, 9646–9657.

178

73.

74.

75.

76. 77. 78.

79.

80.

81.

82.

83. 84. 85.

86.

87. 88.

Materials for Solid State Lighting and Displays

Oeckler, J.A., Kechele, H., Koss, P.J. et al. (2009) Sr5 Al5+x Si21−x N35−x O2+x ∶Eu2+ (x ≈ 0) – A novel green phosphor for white light pcLEDs with disordered intergrowth structure. Chem. Eur. J., 15, 5311–5319. Shioi, K., Michiue, Y., Hirosaki, N. et al. (2011) Synthesis and photoluminescence of a novel Sr-SiAlON∶Eu2+ blue-green phosphor (Sr14 Si68−s Al6+s Os N106−s ∶Eu2+ (s ≈ 7)). J. Alloy Compd, 509, 332–337. Xie, R.-J., Hirosaki, N., Mitomo, M. et al. (2006) Wavelength-tunable and thermal stable Li-α-sialon∶Eu2+ oxynitride phosphors for white light-emitting diodes. Appl. Phys. Lett., 89, 241103. Xie, R.-J., Hirosaki, N., Sakuma, K. et al. (2004) Eu2+ -doped Ca-alpha-SiAlON: A yellow phosphor for white light-emitting diodes. Appl. Phys. Lett., 84, 5404–5406. Shioi, K., Hirosaki, N., Xie, R.-J. et al. (2010) Synthesis, crystal structure, and photoluminescence of Sr-α-sialon∶Eu2+ . J. Am. Ceram. Soc., 93(2), 465–469. Park, W.B., Singh, S. P., Yoon, C. et al. (2012) Eu2+ luminescence from 5 different crystallographic sites in a novel red phosphor, Ca15 Si20 O10 N30 ∶Eu2+ . J. Mater. Chem., 22, 14068–14075. Park, W.B., Singh, S.P., Yoon, C. et al. (2013) Combinatorial chemistry of oxynitride phosphors and discovery of a novel phosphor for use in light emitting diodes. Ca15 Si20 O10 N30 ∶Eu2+ . J. Mater. Chem. C, 1, 1832–1839. Park, W.B., Shin, N., Hong, K.P. et al. (2012) A new paradigm for materials discovery: Heuristics-assisted combinatorial chemistry involving parameterization of material novelty. Adv. Funct. Mater., 22, 2258–2266. Wang, X.M., Wang, C.H., Kuang, X.J. et al. (2012) Promising oxonitridosilicate phosphor host Sr3 Si2 O4 N2 : Synthesis, structure, and luminescence properties activated by Eu2+ and Ce3+ ∕Li+ for pc-LEDs. Inorg. Chem., 51, 3540–3547. Fukuda, Y., Okada, A. and Albessard, A.K. (2012) Luminescence properties of Eu2+ -doped red-emitting Sr-containing Sialon phosphor. Appl. Phys. Express, 5, 062102. Hirosaki, N., Xie, R.-J., Inoue, K. et al. (2007) Blue-emitting AlN∶Eu2+ nitride phosphor for field emission displays. Appl. Phys. Lett., 91, 061101. Dierre, B., Yuan, X.L., Inoue, K. et al. (2009) Role of Si in the luminescence of AlN:Eu, Si phosphors. J. Am. Ceram. Soc., 92, 1272–1275. Takeda, T., Hirosaki, N., Xie, R.-J. et al. (2010) Anomalous Eu layer doping in Eu, Si co-doped aluminium nitride based phosphor and its direct observation. J. Mater. Chem., 20, 9948–9953. Grins, J., Esmaeilzadeh, S., Svensson, G. et al. (1999) High-resolution electron microscopy of a Sr-containing sialon polytypoid phase. J. Eur. Ceram. Soc., 19, 2723–2730. Jack, K.H. and Wilson, W.I. (1972) Ceramics based on the Si-Al-O-N and related systems. Nature Phys. Sci., 238, 28–29. Takahashi, K., Xie, R.-J. and Hirosaki, N. (2011) Toward higher color purity and narrower emission band β-silaon∶Eu2+ by reducing the oxygen concentration. Electrochem. Solid-State Lett., 14, E38–E40.

Nitride and Oxynitride Phosphors for Light Emitting Diodes

89. 90. 91.

92.

93. 94.

95.

96.

97. 98. 99. 100.

101. 102.

103.

104.

105.

179

Yamada, S., Emoto, H., Ibukiyama, M. et al. (2012) Properties of SiAlON powder phosphors for white LEDs. J. Eur. Ceram. Soc., 32, 1355–1358. Kimoto, K., Xie, R.-J., Matsui, Y. et al. (2009) Direct observation of single dopant atom in light-emitting phosphor of β-sialon∶Eu2+ . Appl. Phys. Lett., 94, 041908. Li, Y.Q., Hirosaki, N., Xie, R.-J. et al. (2008) Crystal and electronic structures, luminescence properties of Eu2+ -doped Si6−z Alz N8−z and My Si6−z Alz−y Oz+y N8−z−y (M = 2Li, Mg, Ca, Sr, Ba). J. Solid State Chem., 181, 3200–3210. Seibald, M., Oeckler, O., Celinski, V.R. et al. (2011) Real structure and diffuse scattering of Sr0.5 Ba0.5 Si2 O2 N2 ∶Eu2+ – A highly efficient yellow phosphor for pc-LEDs. Solid State Sci., 13, 1769–1778. Cao, G.Z. and Metselaar, R. (1991) 𝛼 ′ -Sialon Ceramics, A Review. Chem Mater., 3, 242–252. Schlieper, T., Milius, W., and Schnick, W. (1995) Nitrido silicates. II. High temperature synthesis and crystal structure of Sr2 Si5 N8 and Ba2 Si5 N8 . Z. Anorg. Allg. Chem., 621, 1380–1384. Zhang, C.N., Uchikoshi, T., Xie, R.-J. et al. (2015) Reduced thermal degradation of the red-emitting Sr2 Si5 N8 ∶Eu2+ phosphor via thermal treatment in nitrogen. J. Mater. Chem. C, 3, 7642–7651. Zhu, J., Wang, L., Zhou, T.L. et al. (2015) Moisture-induced degradation and its mechanism of (Sr, Ca)AlSiN3 ∶Eu2+ , a red-color-converter for solid state lighting. J. Chem. Mater. C, 3, 3181–3188. Blasse, G. and Grabmaier, B.C. (1994) Luminescent Materials, Springer-Verlag. Liu, T.C., Kominami, H., Greer, H.F. et al. (2012) Blue emission by interstitial site occupation of Ce3+ in AlN. Chem. Mater., 24, 3486–3492. Dierre, B., Xie, R.-J., Hirosaki, N. et al. (2007) Blue emission of Ce3+ in lanthanide silicon oxynitride phosphors. J. Mater. Res., 22, 1933–1941. Takahashi, K., Hirosaki, N., Xie, R.-J. et al. (2007) Luminescence properties of blue La1−x Cex Al(Si6−z Alz )(N10−z Oz )(z ∼ 1) oxynitrides phosphors and their application in white light-emitting diode. Appl. Phys. Lett., 91, 091923. Li, Y.Q., de With, G. and Hintzen, H.T. (2004) Synthesis, structure, and luminescence properties of Eu2+ and Ce3+ activated BaYSi4 N7 . J. Alloy Compd, 385, 1–11. Li, H.-L., Zhou, G.-H., Xie, R.-J. et al. (2011) Optical properties of green-blueemitting Ca-α-Sialon∶Ce3+ , Li+ phosphors for white light-emitting diodes (LEDs). J. Solid State Chem., 184, 1036–1042. Liu, L.H., Xie, R.-J., Hirosaki, N. et al. (2010) Optical properties of blue-emitting Cex Si6−z Alz−x Oz+1.5x N8−z−x for white light-emitting diodes. J. Electrochm. Soc., 157, H50–H54. Li, Y.Q., Hirosaki, N., Xie, R.J. et al. (2008) Yellow-orange-emitting CaAlSiN3 ∶Ce3+ phosphor: Structure, photoluminescence, and application in white LEDs. Chem. Mater., 20, 6704–6714. Durach, D., Neudert, L., Schmidt, P.J. et al. (2015) La3 BaSi5 N9 O2 ∶Ce3+ − A Yellow phosphor with an unprecedented tetrahedral network structure investigated by combination of electron microscopy and synchrotron X-ray diffraction. Chem. Mater., 27, 4832–4838.

180

Materials for Solid State Lighting and Displays

106. Li, Y.Q., de With, G. and Hintzen, H.T. (2006) Luminescence properties of Ce3+ -activated alkaline earth silicon nitride M2 Si5 N8 (M = Ca, Sr, Ba) materials. J. Lumin., 116, 107–116. 107. Woike, M., and Jeitschko, W. (1995) Preparation and crystal structure of the nitridosilicates Ln3 Si6 N11 (Ln = La, Ce, Pr, Nd, Sm) and LnSi3 N5 (Ln = Ce, Pr, Nd). Inorg. Chem., 34, 5105–5108. 108. Gal, Z.A., Mallinson, P.M., Orchard, H.J. et al. (2004) Synthesis and structure of alkaline earth silicon nitrides: BaSiN2 , SrSiN2 , and BaSiN2 . Inorg. Chem., 43, 3998–4006. 109. Wang, X.-M., Zhang, X., Ye, S. et al. (2013) A promising yellow phosphor of Ce3+ ∕Li+ doped CaSiN2−2𝛿∕3 O𝛿 for pc-LEDs. Dalton Trans., 42, 5167–5173. 110. Xie, R.-J., Hirosaki, N., Mitomo, M. et al. (2005) Strong green emission from 𝛼-SiAlON activated by divalent ytterbium under blue light irradiation. J. Phys. Chem. B, 109, 9490–9494. 111. Bachmann, V., Justel, T., Meijerink, A. et al. (2006) Luminescence properties of SrSi2 O2 N2 doped with divalent rare earth ions. J. Lumin., 121, 441–449. 112. Ruan, J., Xie, R.-J., Hirosaki, N. et al. (2012) Photoluminescence properties and energy transfer in Eu2+ -Yb2+ codoped SrSi2 O2 N2 oxynitride phosphor. J. Electrochem. Soc., 159 (2), H66–H71. 113. Liu, T.C., Cheng, B.M., Hu, S.F. et al. (2011) Highly stable red oxynitride beta-SiAlON∶Pr3+ phosphor for light-emitting diodes. Chem. Mater., 23, 3698–3705. 114. Xie, R.-J., Hirosaki, N., Liu, X.-J. et al. (2008) Crystal structure and photoluminescence of Mn2+ -Mg2+ codoped gamma aluminum oxynitride (𝛾-alon): A promising green phosphor for white light-emitting diodes. Appl. Phys. Lett., 92, 201905. 115. Liu, L.H., Wang, L., Zhang, C. et al. (2015) Strong energy-transfer-induced enhancement of luminescence efficiency of Eu2+ and Mn2+ -codoped gamma-AlON for near-UV-LED-pumped solid state lighting. Inorg. Chem., 54, 5556–5565. 116. Pust, P., Weiler, V., Hecht, C. et al. (2014) Narrow-band red-emitting Sr[LiAl3 N4 ]∶ Eu2+ as a next-generation LED-phosphor material. Nat. Mater., 13, 891–896. 117. Pust, P., Wochnik, A.S., Baumann, E. et al. (2014) Ca[LiAl3 N4 ]∶Eu2+ – A narrow-band red-emitting nitridolithoaluminate. Chem. Mater., 26, 3544–3549. 118. Pust, P., Hintze, F., Hecht, C. et al. (2014) Group (III) nitrides M[Mg2 Al2 N4 ] (M = Ca, Sr, Ba, Eu) and Ba[Mg2 Ga2 N4 ] – Structural relation and nontypical luminescence properties of Eu2+ doped samples. Chem. Mater., 26, 6113–6119. 119. Hintze, F., Hummel, F., Schmidt, P.J. et al. (2012) Ba3 Ga3 N5 – A novel host lattice for Eu2+ -doped luminescent materials with unexpected nitridogallate substructure. Chem. Mater., 24, 402–407. 120. Hintze, F., Johnson, N.W., Seibald, M. et al. (2013) Magnesium double nitride Mg3 GaN3 as new host lattice for Eu2+ doping: Synthesis, structural studies, luminescence, and band-gap determination. Chem. Mater, 25, 4044–4052. 121. Strobel, P., Schmiechen, S., Siegert, M. et al. (2015) Narrow-band green emitting nitridolithoalumosilicate Ba[Li2 (Al2 Si2 )N6 ]∶Eu2+ with framework topology whj for LED/LCD backlighting applications. Chem. Mater., 27, 6109–6115. 122. Sakuma, K., Omichi, K., Kimura, N. et al. (2001) Warm-white light-emitting diode with yellowish orange SiAlON ceramic phosphor. Opt. Lett., 29, 2001–2003.

Nitride and Oxynitride Phosphors for Light Emitting Diodes

181

123. Xie, R.-J., Hirosaki, N., Mitomo, M. et al. (2006) Highly efficient white-light-emitting diodes fabricated with short-wavelength yellow oxynitride phosphors. Appl. Phys. Lett., 88, 101104. 124. Wang, L., Zhang, H., Wang, X.-J. et al. (2015) Europium(II)-activated oxonitridosilicate yellow phosphor with excellent quantum efficiency and thermal stability – a robust spectral conversion material for highly efficient and reliable white LEDs. Phys. Chem. Chem. Phys., 7, 15797–15804. 125. Xie, R.-J., Hirosaki, N., Kimura, N. et al. (2007) 2-Phosphor-converted white light-emitting diodes using oxynitride/nitride phosphors. Appl. Phys. Lett., 90, 191101. 126. Mueller-Mach, R., Mueller, G., Krames, M.R. et al. (2005) Highly efficient all-nitride phosphor-converted white light emitting diode. Phys. Stat. Sol. (a), 202, 1727–1732. 127. Yang, C.C., Lin, C.M., Chen, Y.J. et al. (2009) Highly stable three-band white light from an InGaN-based blue light-emitting diode chip precoated with (oxy)nitride green/red phosphors. Appl. Phys. Lett., 90, 123503. 128. Sakuma, K., Hirosaki, N., Kimura, N. et al. (2005) White light-emitting diode lamps using oxynitride and nitride phosphor materials. IEICE Trans. Electron, E88-C, 2057–2063. 129. Kimura, N., Sakuma, K., Hirafune, S. et al. (2007) Extrahigh color rendering white light-emitting diode lamps using oxynitride nitride phosphors excited by blue light-emitting diode. Appl. Phys. Lett., 90, 051109. 130. Takahashi, K., Hirosaki, N., Xie, R.-J. et al. (2007) Luminescence properties of blue La1−x Cex Al(Si6−z Alz )(N10−z Oz ) (z = 1) oxynitride phosphors and their application in white light-emitting diode. Appl. Phys. Lett., 91, 091923. 131. Wang, L., Wang, X.J., Takeda, T. et al. (2015) Structure, luminescence, and application of robust carbidonitride blue phosphor (Al1−x Six Cx N1−x ∶Eu2+ ) for near UV-LED driven solid state lighting. Chem. Mater., 27, 8457–8466. 132. Xie, R.-J., Hirosaki, N. and Takeda, T. (2009) Wide color gamut backlight for liquid crystal displays using three-band phosphor-converted white light-emitting diodes. Appl. Phys. Express, 2, 022401. 133. Fukuda, Y., Matsuda, N., Okada, A. et al. (2012) White light-emitting diodes for wide-color-gamut backlight using green-emitting Sr-Sialon phosphor. Jpn., J. Appl. Phys., 51, 122101. 134. Oh, J.H., Kang, H., Ko, M. et al. (2015) Analysis of wide color gamut of green/red bilayered freestanding phosphor film-capped white LEDs for LCD backlight. Opt. Express, 23 (15), A791–A804. 135. Wang, L., Wang, X.J., Kohsei, T. et al. (2015) Highly efficient narrow-band green and red phosphors enabling wider color-gamut LED backlight for more brilliant displays. Opt. Express, 23, 28707–28717.

5 Organic Light Emitting Device Materials for Displays Tyler Davidson-Hall, Yoshitaka Kajiyama and Hany Aziz Department of Electrical & Computer Engineering, University of Waterloo, Ontario, Canada

In 1987, when the first efficient organic light emitting device (OLED) was demonstrated by Tang and Van Slyke, their potential for flat panel displays (FPDs) was immediately recognized [1]. Since then OLEDs have attracted significant interest and the focus of an enormous body of work for the purpose of realizing that potential. Although their performance was limited in the early stage, the unrelenting effort in developing better materials and device structures has led to a steady progress in device performance over the last 25 years. As a result, OLED displays are now used in a wide range of commercial products, and the technology has emerged as a serious competitor to liquid crystal displays (LCDs) [2–4]. OLEDs possess a unique combination of features that position them favorably relative to LCDs and other FPD technologies. These include their ability to emit bright, vivid colors, to provide higher contrasts and to lend themselves easily to large area fabrication processes including on flexible substrates [2, 5–7]. In this chapter, an introduction to OLEDs and organic electroluminescent materials is first provided in Section 5.1. This is followed, in Section 5.2, by a brief review that highlights some of the most important classes and types of organic electroluminescent materials developed to date and some of the important materials in each case. In Section 5.3, some of the important aspects of OLED display technology, its current status, and challenges are described. The chapter is concluded by an introduction to quantum dot (QD)-LEDs, which, because of their unsurpassed color purity signal and their use of organic semiconductors, are seen as an evolutionary extension of OLEDs.

Materials for Solid State Lighting and Displays, First Edition. Edited by Adrian Kitai. © 2017 John Wiley & Sons Ltd. Published 2017 by John Wiley & Sons Ltd.

184

Materials for Solid State Lighting and Displays

5.1

Introduction to OLEDs and Organic Electroluminscent Materials

In an OLED light is produced by a thin layer (∼10–100 nm) of an organic electroluminescent material sandwiched between two electrodes. Applying an electric bias across the electrodes causes electrons and holes to be injected from them into the LUMO (Lowest Occupied Molecular Orbital) and HOMO (Highest Occupied Molecular Orbital) states of the organic material. The electrons and holes then drift under the effect of the applied electric across the organic layer towards each other. They then recombine producing Frenkel-type (or localized) excitons, essentially electrostatically bound electron–hole (e–h) pairs that comprise an electron in a LUMO state and a hole in a HOMO state of the same molecule. The electron then relaxes from the higher energy state and fills the hole in the lower energy state emitting the energy radiatively as a photon and causing the exciton to vanish (or “decay”). In the first approximation, the energy of the emitted photon corresponds to the HOMO–LUMO difference, and the latter is treated as the energy band-gap Eg of the organic electroluminescent material. Figure 5.1 illustrates this process. In addition to the organic electroluminescent material layer, OLEDs normally include at least two additional organic semiconductor layers, interposed between it and the electrodes, that will help facilitate charge carrier injection from the electrodes and reduce the leakage of carriers to the counter electrodes, thereby increase e–h recombination probability and device electroluminescent efficiency. The layers are referred to as the hole and electron transport layers (HTL and ETL, respectively.). In this three-layer structure, the electroluminescent material layer is commonly referred to as the emissive layer (EML), in reference to its function. Figure 5.2 depicts schematic diagrams of the physical construction of a three-layer OLED and the electroluminescence process in this case. Assigning the charge injection and confinement roles of the transport layers to different materials often leads to better device performance. Therefore, most state-of-the-art devices now consist of at least five layers in addition to the electrodes [8].

LUMO

–

– – Al

Al g Organic EL material ITO Glass ITO + (a)

+

+ HOMO (b)

Figure 5.1 (a) Schematic diagram of the physical construction of a single layer OLED. (b) Energy band diagram of the OLED. The arrows illustrate the different steps of the electroluminescence process described in the text. EL, electroluminescent

Organic Light Emitting Device Materials for Displays

– AI

HTL

–

185

– –

EML

– 4.1 eV

ETL ETL

EML

Al

HTL ITO Glass

4.7 eV ITO +

+

+

+

+

(a)

(b)

Figure 5.2 (a) Schematic diagram of the physical construction of a three-layer OLED. (b) Energy band diagram of the OLED. The arrows illustrate the different steps of the electroluminescence process described in the text

As will be described in more detail later, using the electroluminescent material in the form of a dopant that is dispersed as a guest material into a wider band-gap host material at a low concentration (usually at ∼1–−5 mol%), as opposed to being used in the form of a “neat” layer in the fashion described above, often increases the quantum yield of the material. The quantum yield of a luminescent material, also commonly referred to as the photoluminescence quantum yield (𝜂PL ), is the probability that the excitons decay radiatively as opposed to through other non-radiative energy dissipative pathways such as internal conversion and thermal pathways, and what ultimately determines the luminescence efficiency of the material and hence the electroluminescence efficiency of the device. Therefore almost all state-of-the-art devices use this host–guest approach. In these systems, excitons are formed on the guest material through one of two possible mechanisms: (i) the recombination of the electrons and holes directly on the guest molecules, which, having a narrower Eg than that of the host, can act as deep charge traps and hence efficient e–h recombination centers; or (ii) e–h recombination on the host followed by the transfer of excitons from the host to the guest through dipole interactions (Förster-type process) or electron exchange (Dexter-type) energy transfer processes. These two mechanisms are illustrated schematically in Figure 5.3. One key advantage that organic electroluminescent materials have over their inorganic counterparts is the wide range of materials that can be relatively easily synthesized, and thus the opportunity to have materials with different bandgaps and emission colors. Electroluminescence that spans across the entire optical range (400–700 nm) can be readily obtained from organic materials. Moreover, the band-gap of any one material can be relatively easily modified via simple chemical substitution. This provides a tool to fine-tune the band-gap. As will become apparent from Section 5.1, the three primary colors required for FPD applications, red, green, and blue, can be readily obtained from a large selection of molecules in each of the three main OLED material classes: (i) fluorescent emitters; (ii) phosphorescent emitters; and (iii) thermally activated delayed fluorescence (TADF) emitters. As conjugated systems with a large number of carbon–carbon double bonds, the

186

Materials for Solid State Lighting and Displays EML

–

–

– –

– HTL

EML

– 4.1 eV Al

ETL Guest

+ Host

4.7 eV AI

ITO +

ETL

+

+

(b)

+

EML

– EML –

HTL

–

–

ITO

HTL

EML

Glass

– 4.1 eV Guest

ETL

Al

(a) +

4.7 eV

+

ITO +

+ +

+

(c)

Figure 5.3 (a) Schematic diagram of a host–guest three-layer OLED. (b) and (c) Energy band diagrams of the OLED illustrating the two possible emission mechanisms referred to in the text, respectively: (i) the recombination of the electrons and holes directly on the guest molecules; and (ii) e–h recombination on the host followed by the transfer of excitons from the host to the guest

HOMO and LUMO levels in organic semiconductors are typically 𝜋 (bonding) and 𝜋 ∗ (anti-bonding) molecular orbitals, respectively, produced from the quantum-mechanical overlap of the non-hybridized 2p atomic orbitals of the carbon atoms in the molecule. Increasing the number of the 2p atomic orbitals in a molecule (by simply increasing the number of carbon–carbon double bonds in it) results in an increase in the density of the 𝜋 and 𝜋 ∗ states, and therefore reduces the band-gap of the material. Therefore, increasing (decreasing) the extent of conjugation in these molecules will lead to a red (blue) shift in their electroluminescent spectra. This allows tuning of the band-gap.

5.2

OLED Light Emitting Materials

Since the breakthrough work of Tang and Van Slyke, a substantial amount of work has been spent on optimizing the materials used in OLEDs from the charge transport materials to the emissive molecules themselves. There is a significant number of qualities that can be addressed in the fabrication of these materials, but ultimately color and photoluminescence quantum yield (𝜂PL ) are the defining characteristics that arise from the coalescence of the electronic properties of OLEDs and the molecular interactions within.

Organic Light Emitting Device Materials for Displays

187

However, there are some properties that are universally desired in each layer of an OLED. Thermal stability of the molecules is very important because the heat losses in a device can impact the morphology of unstable layers. Thermal stability is also influential in the amorphous film forming ability of the deposited material which is important for uniform emission and device stability. Materials with large band-gaps are generally less stable than others, a behavior that is often attributed to the higher energy that the molecule possesses when in an excited state. This section will deal with the most popular molecular building blocks for OLED emitters and rather than a comprehensive list of emitters, we will instead focus on the reasoning behind the modifications to these elementary molecules while touching on some of the more popular materials. 5.2.1

Neat Emitters

The first efficient OLEDs were created using neat layers of organic semiconductors, which is to say that each layer consisted solely of a single material. This contrasts with emitters that have been doped into an organic host layer which have since become the leading OLED structure and will be discussed later. Provided the molecules have sufficient charge transport properties, the use of neat emitting layers allows for simple device structures consisting of hole and electron transporting layers only. In this case, the excitons are expected to form near the interface of the two layers and combine radiatively in the layer with the lowest band-gap. It is difficult to fabricate efficient red neat emitting layers because these molecules are either polar or extensively 𝜋-conjugated and prone to aggregation which results in strong non-radiative recombination of excitons [9]. 5.2.1.1

Metal Chelates

The metal chelate structure consists of a metal atom surrounded by ligands that form coordinate bonds. Unlike a traditional covalent bond, both electrons will be donated by the ligand with nitrogen or oxygen most likely to be the bonding atoms. Metal chelates were the first category of materials to be utilized in OLEDs, beginning with tris(8-hydroxyquinolinato)aluminum (Alq3 ) which has a green electroluminescent emission peak at 530 nm, quantum yield of 32% in solid state, and an electron mobility on the order of 10−5 cm2 ∕V∕s [10, 11]. Metal chelates typically have an electron mobility much greater than their hole mobility making them an option to use as an ETL for a simple neat emitter OLED. Although metal chelates may fluoresce in solution, this does not necessarily translate to a high solid state fluorescence due to the possibility of open coordination sites [12]. Open coordination sites in solution can be passivated by solvent molecules but lead to poor stability when deposited in thin films [12]. The metal ion used in chelate emitters has been shown to influence emission color, efficiency, and stability of the molecules. Paramagnetic ions such as Cr and Ni have high intersystem crossing rates which leads to non-radiative quenching of excitons [13]. Ions with increasing atomic number typically lead to an increase in intersystem crossing and ionic character of the coordination bond which redshifts the emission wavelength and

188

Materials for Solid State Lighting and Displays

(A)

(B)

(C)

(D)

λ = 530 nm

λ = 515 nm

λ = 580 nm

λ = 440 nm

N N O

O Al

N O

N

N

N O O B N

N

O

N O

O

Al N N

O

O Al

N

N O

N

N

N

N

Figure 5.4 Metal chelates: (A) tris(8-hydroxyquinolinato)aluminum (Alq3 ); (B) bis(10hydroxybenzo[h]quinolinato)beryllium (Be(bq)2 ); (C) tris(5-hydroxyquinoxaline)aluminum (AlX3 ); (D) tris(4-hydroxy-1,5-naphthyridine)aluminum (Al(ND)3 ) [12]

this can be seen in chelates of Al, Ga, and In which have emission wavelengths of 532, 545, and 558 nm, respectively [12, 14]. It is for this reason that Al3+ , Be2+ , and Zn2+ are the most common metal ions. Heavy metal chelates such as Eu3+ or Tb3+ , while increasing the rate of intersystem crossing, also increase the spin orbit coupling (SOC) character allowing for the emission of triplet excitons [15, 16]. However, the lifetime of these excitons is much longer than the singlets and using these chelates as a neat emitter leads to highly inefficient devices due to triplet–triplet annihilation (TTA) among other quenching processes. Instead, these emitters are better suited for use as dopant emitters. While chelates such as bis(10-hydroxybenzo[h]quinolinato)beryllium (Be(bq)2 ) which has a green peak emission of 515 nm and a greater electron mobility (10−4 cm2 ∕V∕s), Al chelates offer the best balance between efficiency and stability [12, 17, 18]. See Figure 5.4 for some examples of metal chelate emitters. The shifts that result from the previous rules still mostly result in primarily green emitting molecules but the ligand structure also has a significant influence on the HOMO and LUMO levels and can be modified to shift the emission wavelength further. Using Alq3 as the basis for such adjustments, substituting a hydrogen atom on the C5 or C7 of 8-hydroxyquinolinato with an electron withdrawing group such as cyano- or trihalides will cause a redshift in the emission spectrum [12, 19]. Adding an electron donating group such as methoxy- or amines to C2 or C4 will also cause a redshift in the emission spectrum. Conversely, substituting the opposite moiety on any of the carbons will result in a blueshift [12, 19]. This is possible because the filled 𝜋 orbitals which make up the HOMO are located on the phenoxide side of the ligand and the 𝜋 ∗ orbitals which make up the LUMO are located on the pyridyl side [20]. Thus, a change in the dipoles of the ligand will affect the band-gap of the molecule. However, the preferred modification method is the direct substitution of a carbon atom with a nitrogen atom or other moiety for large shifts in emission. Replacing C4 with nitrogen produces tris(5-hydroxyquinoxaline)aluminum (AlX3 ), redshifting the emission to 580 nm and replacing C5 with nitrogen produces tris(4-hydroxy-1,5-naphthyridine)aluminum (Al(ND)3 ), blueshifting the emission to 440 nm [12, 20]. The use of other ligands or modification of existing ligands therefore makes the creation of emitters in the red, green, and blue spectrum possible. However, deep red and deep blue metal chelate emitters are difficult to fabricate due to the large requisite peak shifts.

Organic Light Emitting Device Materials for Displays

5.2.1.2

189

Polycyclic Aromatic Hydrocarbon Oligomers

There are many semiconducting polycyclic aromatic hydrocarbons (PAHs) that can be used in OLEDs as emitters including naphthalene, anthracene, fluorene, and pyrene. PAHs are particularly useful because of their high hole mobility often greater than 10−3 cm2 V−1 s−1 which allows for the fabrication of efficient OLEDs if combined with an electron conducting molecule such as a metal chelate [21]. Acridine and carbazole are also of interest and have similar structures to anthracene and fluorene, respectively, where the only difference is in the substitution of the central carbon atom with nitrogen. The nitrogen substitution makes these molecules more electron rich and improves hole transport. The structure of these molecules as well as the most common derivatives are shown in Figure 5.5 [22–26]. Although the first efficient OLED was not reported until 1987, electroluminescence was first discovered in anthracene crystals in 1963 albeit with an applied bias of 400 V [27]. The emission wavelength depends highly on the conjugation character of the molecules where increasing conjugation leads to a redshift of the emission. For example, the peak fluorescence wavelength of benzene, naphthalene, and anthracene suspended in water increases from 280 to 320 to 400 nm, respectively [28]. It is clear that most PAHs emit blue and anthracene is a good molecule to build from because it emits visible light unlike many of the others mentioned previously. However, an increase in conjugation and planarity will lead to aggregation, redshifting the emission and decreasing quantum yield due to quenching when deposited as a film. For this reason, it is beneficial to add side groups to the core and reduce the molecular symmetry of the emitter. By adding m-terphenyl groups to C9 and C10 of anthracene, a highly efficient blue emitter 9,10-bis(3′′ ,5′′ -diphenylbiphenyl-4′ -yl)anthracene (TAT) with a peak emission wavelength of 444 nm and CIE coordinates of (0.156, 0.088) was created and has led to OLEDs with external quantum efficiency (EQE) of up to 7.18% [22]. Combining two PAHs as in 9, 10-bis-(𝛽-naphthyl)anthracene (ADN), which consists of an anthracene core and a naphthylene group on C9 and C10, is another method that has been used to fabricate stable neat emitting OLEDs. The performance of ADN can be improved simply by adding a tert-butyl group to C2 to create 2-tert-butyl-9, 10-bis-(𝛽-naphthyl)anthracene (TBADN) which decreases the aggregation of the emitters in a neat layer and shifts the emission wavelength from 460 to 452 nm, CIE coordinates from (0.154, 0.232) to (0.14, 0.10) and improves the quantum yield. There are countless modifications that have been made in an attempt to improve the properties of PAH-based emitters but most follow the scheme of reducing aggregation and improving quantum yield while tuning emission wavelength. In the interest of developing a full suite of colors, many green and red PAH emitters have been created. By adding diaryl groups to an anthracene core, the resultant molecule will exhibit green emission due to an increase in charge transfer [29]. In the case of N, N, N ′ , N ′ -tetrakis-(3,4-dimethyl-phenyl)-anthracene-9,10-diamine (TmpAD) where (3,4-dimethyl-phenyl)-amine side groups are used, a solid state quantum yield of 82%, a fluorescent peak of 535 nm, and CIE coordinates of (0.38, 0.59) were measured [26]. Extending upon the conjugation concept, it is possible to shift the emission wavelength of anthracene-based emitters into the orange-red region by introducing an ethylene compound between the diaryl and anthracene groups. This design was carried out to create 9,10-bis[4-(di-4-tert-butylphenylamino)styryl]anthracene (ATBTPA) which

190

Materials for Solid State Lighting and Displays

(A)

(B)

(C)

λ = 280 nm λ = 320 nm λ = 400 nm

(D)

(E)

(F)

(G)

λ = 476 nm

λ = 315 nm

λ = 425 nm

λ = 346 nm

N N

(H)

(I)

λ = 444 nm

λ = 579 nm

N N

(M) λ = 555 nm

(J)

(K)

(L)

λ = 460 nm

λ = 452 nm

λ = 656 nm

S OCH3

(N) λ = 535 nm OCH3 S N

N

Figure 5.5 Polycyclic aromatic hydrocarbons: (A) benzene; (B) naphthalene; (C) anthracene; (D) tetracene (adapted from [22]); (E) fluorene; (F) acridine; (G) carbazole (adapted from [23]); (H) 9,10-bis(3’’,5’’-diphenylbiphenyl-4’-yl)anthracene (TAT) [22]; (I) 9,10-bis[4-(di-4-tert-butylphenylamino)styryl]anthracene (ATBTPA) [24]; (J) 9, 10-bis-(𝛽naphthyl)anthracene (ADN) (adapted from [24]); (K) 2-tert-butyl-9, 10-bis-(𝛽-naphthyl) anthracene (TBADN) (adapted from [24]); (L) 5,12-dimethoxy-6,11-bis(5-triisopropylsilylthienylethynyl)tetracene (TACN) [25]; (M) rubrene [24]; (N) N,N,N’,N’-tetrakis-(3,4dimethyl-phenyl)-anthracene-9,10-diamine (TmpAD) [26]

Organic Light Emitting Device Materials for Displays

191

exhibited a quantum yield of 23% in 1,2-dichloroethene, a fluorescent peak of 579 nm, and CIE coordinates of (0.51, 0.48) but has not been incorporated as a neat OLED emitter [24]. Increasing the number of cyclic hydrocarbons from three in anthracene to four in tetracene is another technique that can be taken advantage of to further redshift the emission wavelength such as in rubrene with a peak emission of 555 nm or 5,12-dimethoxy-6,11-bis(5-triisopropylsilylthienylethynyl)tetracene (TACN) with a peak emission of 656 nm [30, 31]. However, devices with red neat emitting layers either have poor performance or are insufficiently characterized and presumed to have poor performance while exceeding as dopant emitters [9]. 5.2.1.3

Conjugated Polymers

Shortly after the small molecule OLED (SM-OLED) developed by Tang and VanSlyke, Burroughes et al. demonstrated a polymer OLED (P-OLED) in 1990 that used poly (p-phenylene vinylene) PPV [32]. However, P-OLEDs at the time suffered from comparatively poor performance due in part to the low hole and electron mobility of 10−5 cm2 ∕V∕s [33]. Poly(9,9′ -dioctylfluorene) (PFO)-based polymers have demonstrated hole mobility of up to 10−3 cm2 ∕V∕s but a polymer with high electron mobility remains difficult to design [33, 34]. The most common polymers since then have been poly(p-phenylene) (PPP), PFO, and PPV-derived molecules and examples of each are depicted in Figure 5.6 [33, 35–37]. Although the first PPV P-OLED exhibited poor performance at a fluorescent peak of 564 nm and a solid state quantum yield of only 8%, subsequent modifications of this polymer have shown promising growth [32]. Just as in the PAH-based emitters, it is possible to graft moieties to the polymer backbone to change the emission wavelength or electronic characteristics since the polymer emitters are themselves PAHs. Increasing the degree of conjugation in the polymers also redshifts the peak emission wavelength. Incorporating two different monomers into a copolymer is another technique that can be used to finely tune the

(A)

(B)

λ = 401 nm

(C)

λ = 550 nm

n

λ = 425 nm

n

C8H17

(D)

λ = 420 nm

Br–

C8H17 C8H17 C8H17

C8H17

n

(E)

N+ N+

λ = 648 nm Br–

C8H17

C8H17

C8H17

m S C8H17

Si

C8H17

n

m

S N

N S

n

Figure 5.6 Aromatic polymers: (A) poly(p-phenylene) (PPP) [35]; (B) poly(p-phenylene vinylene) (PPV) [35]; (C) poly(9,9′ -dioctylfluorene) (PFO) [35]; (D) poly(3,6-silafluoreneco-2,7-fluorene) (PSiFF) [36]; (E) poly(4,7-di-2-thienyl-2,1,3-benzothiadiazole-co-bis(4,4,5,5tetramethyl-1,3,2-dioxaborolan-2-yl)-9,9-dioctylfluorene) (PFNBr-DBT) [37]

192

Materials for Solid State Lighting and Displays

final emitter properties or even the properties within a polymer as a function of length by varying the polymerization parameters and monomers. Thus, beginning with a simple PPP polymer, substituting every sixth phenylene monomer with a vinylene monomer to create P5V shifts the emission wavelength from 401 to 446 nm and the quantum yield from 34 to 49% [38]. If instead vinylene is substituted for every third polymer to create P2V the emission wavelength will shift to 480 nm and quantum yield will decrease to 9% [38]. As evident from PPP, P5V, P2V, and PPV, the emission peak redshifts due to an increase in conjugation but there is also a corresponding decrease in solid state quantum yield which as a result of aggregation-induced non-radiative recombination [38]. Steric groups can be added to the polymers to reduce these effects, but solubility and performance may be adversely affected as well. The poor stability of PPP- and PPV-based polymers has led to the adoption of highly efficient and high energy PFO as the preferred backbone for P-OLEDs, particularly for blue devices [34, 39]. The same modifications discussed previously can also be applied to PFO and it is possible to create polymers that range from deep blue poly(3,6-silafluorene-co-2,7-fluorene) (PSiFF90) with an emission peak of 420 nm, CIE coordinates of (0.16, 0.07), and quantum yield of 84% to red poly(4,7-di-2thienyl-2,1,3-benzothiadiazole-co-bis(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-9,9dioctylfluorene) (PFNBr-DBT5) with an emission peak of 648 nm, CIE coordinates of (0.64, 0.35), and quantum yield of 29.9% [36, 37]. In the formative years of OLED research, the main advantage to P-OLEDs lie in their comparatively robust structure and solution processability which would ultimately lead to low cost roll-to-roll or ink jet fabrication of devices. SM-OLEDs required thermal evaporation of the organic materials but excelled in the fabrication of efficient devices with long lifetimes. There has been increasing interest in producing solution processable SM-OLEDs which could realize the benefits of low cost processing and efficient devices. While not all small molecule emitters have good solubility and thus poor film forming ability, adding alkyl or other side groups is one technique that can be used to improve solubility at the cost of efficiency or color shift [25]. The performance of solution processed SM-OLEDs depends on the solvents that are used which adds another layer of complexity to the OLED design [40]. Some of the main solution processed SM-OLED techniques and architectures are single layer OLEDs incorporating both charge transport and emissive molecules, cross-linkable layers that will not dissolve in the following solvents, and the use orthogonal solvents in subsequent layers [41, 42]. In some cases, the performance of solution processed devices can approach that of the vacuum deposited OLEDs but the lifetime remains an issue due to trapped solvent impurities [40]. 5.2.2

Guest Emitters

While fabrication of neat emitting OLEDs significantly simplifies the process, doping the emitters into a host matrix has been proven to produce superior devices. There are several benefits from transitioning to dopant emitters such as allowing for more optimization of the layer properties and suppressing aggregation of the emitters to maximize quantum

Organic Light Emitting Device Materials for Displays

193

yield and stability. First of all, the carrier transport properties of dopants are much less important than for neat emitters so the emissive properties can be improved independently of mobility. Instead, carrier transport is important in the host so the utilization of poorly emitting but excellent mobility molecules is possible here. Finally, by using a small ratio of dopant to host molecules in the emission layer it is possible to reduce interactions between emitters that would result in non-radiative or otherwise undesirable exciton recombination. It is important to take the host–guest interaction into consideration when optimizing OLED performance as both components can have a meaningful impact on device efficiency. Since the emission layer is predominantly made of the host molecules it is far more likely that excitons will form on the host rather than guest. The host must have an exciton energy larger than that of the dopant to facilitate energy transfer. For singlet excitons the main mode of energy transfer occurs through Förster resonance energy transfer (FRET) which requires spectral overlap of the donor molecule’s emission spectrum and the acceptor molecule’s absorption spectrum. Since radiative recombination of triplets is forbidden in most cases, the main mode of energy transfer of these excitons is through Dexter energy transfer which requires an overlap of the acceptor and donor electron wavefunctions. However, it is desirable for the host to have a larger exciton energy even for triplets to promote the trapping of excitons and carriers on the dopants. The HOMO and LUMO levels of the host molecules should be selected with the energy levels of the charge transport layers in mind to ensure charge injection is not inhibited. In order to improve charge balance within the device, a bipolar host material with similar hole and electron mobilities should be chosen although this may be difficult because in many cases the hole mobility will be orders of magnitude greater than the electron mobility. The neat emitting molecules discussed earlier can also be used as a host or guest. For example, Alq3 can be used as a host for red dopants or as a green dopant in a wider band-gap host such as N, N ′ -dicarbazolyl-4,4′ -biphenyl (CBP). CBP is one of the most popular host materials owing to its ambipolar transport characteristics with 3 × 10−4 cm2 ∕V∕s electron and 2 × 10−3 cm2 ∕V∕s hole mobility [43]. Polymer hosts such as poly(vinylcarbazole) (PVK), which has a 8 × 10−4 cm2 ∕V∕s hole mobility but negligible electron mobility, can also be used in OLEDs by either codepositing the emission layer or grafting the emitter to the polymer backbone [44]. Red emitters are typically the easiest to dope as there are many hosts available to use due to the low requisite energy level, and benefit the most from doping since many red emitters are prone to aggregation [9]. Conversely, blue emitters are difficult to use in a host–guest system because they require even higher energy hosts which are frequently less stable and necessitate higher corresponding HOMO and LUMO levels in the charge transport layers. The biphenyl core of CBP was substituted with benzene to make a wider band-gap host 1,3-bis(N-carbazolyl)benzene (mCP) to address these issues, but the thermal stability proved to be quite poor [45, 46]. By simply replacing the benzene core with pyridine, mCP becomes 2,6-bis(N-carbazolyl)pyridine (26mCPy) which has a HOMO decreased by 0.2 eV to further increase the band-gap and makes a host suitable for deep blue OLEDs [23]. Figure 5.7 depicts the structure of some of the more popular host materials [23, 35, 47].

194

Materials for Solid State Lighting and Displays (A)

(B)

λ = 400 nm

λ = 405 nm

(C)

λ = 345 nm

(D)

λ = 335 nm

n

N N

N

N

N

N

N

N

Figure 5.7 Host materials: (A) N,N’-dicarbazolyl-4,4’-biphenyl (CBP) [47]; (B) poly (vinylcarbazole) (PVK) [35]; (C) 1,3-bis(N-carbazolyl)benzene (mCP) [23]; (D) 2,6-bis(Ncarbazolyl)pyridine (26mCPy) [23]

5.2.2.1

Fluorescent Emitters

The molecular structure of fluorescent aromatic dyes looks very similar to that of the PAH neat emitters discussed previously. Indeed, many of the same rules that apply to the tailoring of emission character of the PAHs are applicable to fluorescent aromatic dyes as well. In 1989, Tang et al. doped their Alq3 device with 3-(2benzothiazolyl)-7-diethylaminocoumarin (C540), 4-(dicyanomethylene)-2-methyl-6-(4dimethylaminostyryl)-4H-pyran (DCM), and 4-(dicyanomethylene)-2-methyl-6-(julolidin4-ylvinyl)-4H-pyran (DCM2) green and red dyes with pure fluorescent emission peaks of 510, 620, and 650 nm, respectively [48]. From these results, it was clear that increasing the dopant loading led to a redshift in the emission spectrum but decreased the efficiency of the devices likely due to the same phenomenon that leads to poor neat red emission. However, at sufficiently low dopant concentrations there would be an observable host contribution to the emission spectrum which blueshifts the peak wavelength due to insufficient FRET from host to guest. These dyes would become the base structures from which more efficient fluorescent dyes would be created and are shown in Figure 5.8 [9, 49–52]. The pyran family of dopants typically contain an electron donating and electron accepting moiety to produce a rigid, polar molecule that tends to aggregate but emits red light. In such molecules, the 4-(dicyanomethylene)-4H-pyran group is the electron acceptor and different donors are used to tailor the emitter properties [49]. Of the DCM-based emitters, perhaps the derivative with the best performance is 4-(dicyanomethylene)2-tert-butyl-6(1,1,7,7-tetramethyljulolidyl-9-enyl)-4H-pyran (DCJTB) with an optimized fluorescent peak at 624 nm, CIE coordinates of (0.632, 0.363), and 78% quantum yield [49, 53]. The steric hindering methyl groups reduce the interactions between doped DCJTB molecules and while increasing stability and the optimum dopant loading from 0.5% for DCM2 to 2% for DCJTB [9]. In the interest of developing deep red dopants other highly conjugated electron acceptors have been investigated, such as a 4-(dicyanomethylene)-chromene in 4-(dicyanomethylene)-2-tert-butyl-6(1,1,7,7tetramethyljulolidyl-9-enyl)-chromene (RED2) which exhibits a fluorescent peak at 645 nm, CIE coordinates of (0.66, 0.33), and 15% quantum yield [49]. The deepest red emitter 5,10,15,20-tetraphenyl-21H,23H-porphine (TPP) based on a porphine structure has a very narrow, but weakly emitting fluorescent peak at 655 nm and CIE coordinates of (0.70, 0.28) [9, 54].

Organic Light Emitting Device Materials for Displays

(A) λ = 620 nm NC

(B) λ = 650 nm NC CN

195

(C) λ = 624 nm NC CN

CN

O

O O

N

N

N

(D) λ = 645 nm

(E) λ = 655 nm NC

CN

NH O

N

N HN

N

(F) λ = 510 nm

(G) λ = 514 nm CH3

S

N

H3C N Et

O

(H) λ = 462 nm O

O

S

N

Et

Et

Et N

O

H3C

N

O CH3

O

CH3

CH3

Figure 5.8 Fluorescent aromatic dyes: (A) 4-(dicyanomethylene)-2-methyl-6-(4-dimethylaminostyryl)-4H-pyran (DCM) [9]; (B) 4-(dicyanomethylene)-2-methyl-6-(julolidin-4-ylvinyl)-4H-pyran (DCM2) [9]; (C) 4-(dicyanomethylene)-2-tert-butyl-6(1,1,7,7-tetramethyljulolidyl9-enyl)-4H-pyran (DCJTB) [9]; (D) 4-(dicyanomethylene)-2-tert-butyl-6(1,1,7,7-tetramethyljulolidyl-9-enyl)-chromene (RED2) (adapted from [9, 49]); (E) 5,10,15,20-tetraphenyl-21H,23H-porphine (TPP) [50]; (F) 3-(2-benzothiazolyl)-7-diethylaminocoumarin (C540) [51]; (G) 10-(2-benzothiazolyl)-1,3,3,7,7-pentamethyl-2,3,6,7-tetrahydro-1H,5H,11H-benzo[l]-pyrano[6,7,8-ij]-quinolizin-11-one (C545P) [52]; (H) 7-(diethylamino)-4-(methyl)-coumarin (C47) [51]

196

Materials for Solid State Lighting and Displays

Likewise, many green and blue dopants have been derived from the coumarin moiety. For instance, decreasing the conjugation of the emitters leads to an emission blueshift as seen in 10-(2-benzothiazolyl)-1,3,3,7,7-pentamethyl-2,3,6,7-tetrahydro-1H,5H,11H-benzo[l]pyrano[6,7,8-ij]-quinolizin-11-one (C545P) with a fluorescence peak at 514 nm, CIE coordinates of (0.31, 0.65), and 99% quantum yield in dichloroethane and 7-(diethylamino)4-(methyl)-coumarin (C47) with a fluorescence peak at 462 nm and CIE coordinates of (0.16, 0.10) [51, 52]. Just like the methyl groups added to DCJTB, the side groups in C545P and C47 improve stability and efficiency while decreasing the effect of concentration quenching. The downside to using fluorescent dyes in OLEDs lies in the inherent limitation of device efficiency. Statistically, in electrically driven OLEDs for each singlet exciton there will be three triplet excitons. However, triplet exciton recombination is forbidden in fluorescent emitters so the maximum possible internal quantum efficiency of these devices amounts to a maximum of only 25% if all of the singlets combine radiatively. Interestingly, it is possible to utilize the triplet excitons by converting them to singlets through TTA at high current density which increases the theoretical maximum internal quantum efficiency (IQE) to 62.5% [55]. 5.2.2.2

Phosphorescent Emitters

Perhaps the most significant development in the performance of OLEDs has been the use of phosphorescent dyes capable of utilizing triplet excitons for radiative emission to reach a theoretical IQE of 100% in an optimized device. In organic molecules, the relaxation of triplet excitons is generally forbidden but heavy and transition metals increase the spin-orbit coupling of molecules which leads to some mixing of the singlet and triplet states. This was first practically employed by Baldo et al. in 1998 with a (2,3,7,8,12,13,17,18-octaethyl-21H,23H-porphine) platinum(II) (PtOEP) emitter which exhibited an EQE of 4% at a 650 nm wavelength and phosphorescence quantum yield of 50%, greater than any red fluorescent device at the time [56]. While the phosphorescence quantum yield of the device is inferior to some of the values discussed previously, by taking spin statistics into account such a quantum yield would be equivalent to 200% for a fluorescent emitter. However, the rate of phosphorescent radiative recombination is on the order of microseconds and is much slower than fluorescent recombination which occurs within nanoseconds [55]. The long lifetime of triplet states increases the probability of two triplets diffusing through the emission layer and interacting creating a significant non-radiative recombination pathway through TTA, particularly at high operation current density. Non-radiative recombination by TTA contributes to an efficiency roll-off whereby the OLED will become less efficient at high currents due to the increased concentration of triplets in the device. Although not the only contributor to efficiency roll-off as charge balance, aggregation and polaron quenching among others are believed to also induce the effect, a direct reduction in TTA is possible by reducing the phosphorescent lifetime of the emitters [55, 57]. PtOEP has a phosphorescent lifetime of approximately 300 μs which partially explains the cause of underwhelming performance in the initial phosphorescent OLED (phOLED) [56]. Soon after, the identification of fac-tris(2-phenylpyridine) iridium (Ir(ppy)3 ) with a phosphorescent lifetime of 0.5 μs, peak emission wavelength of 510 nm,

Organic Light Emitting Device Materials for Displays

197

CIE coordinates of (0.27, 0.63), and phosphorescent quantum yield of 97% demonstrated the potential of phosphorescent emitters with a phOLED capable of 8% EQE [58, 59]. Since then, (Ir(ppy)3 ) has been held as the standard by which all other phosphorescent emitters are compared and by optimizing the structure of Ir(ppy)3 -based devices the EQE has reached values up to 26.3% [60]. In general, Ir complexes are the most attractive phosphorescent dyes because of their high quantum yield and short exciton lifetime. The efficient metal to ligand charge transfer (MLCT) and its effect on SOC is believed to be the reason for this remarkably short lifetime so the interaction between the metal and ligands must be considered when designing new phosphorescent emitters [35]. With this in mind, it is possible to tune the emission color and reduce aggregation of these emitters by functionalizing the ligands or substituting the ligands altogether similar to metal chelate emitters. For example, the emission wavelength will be red-shifted if an electron donating group is added to the phenyl group of ppy or an electron withdrawing group is added to the pyridyl group [61]. Furthermore, simply increasing the conjugation of the ligands should also result in a bathochromic shift in peak emission wavelength. Some work has proposed that a redshift in the emission is also possible by stabilizing the MLCT state through functionalization however there is some disagreement regarding this in literature [61]. Ir(ppy)3 and many other phosphorescent dyes are homoleptic [i.e., in the form (ĈN)3 Ir] having all three ligands the same, but it is also possible to switch the ancillary ligand to another moiety to further customize the dopant properties in the heteroleptic form (ĈN)2 Ir(LX). Despite the fact that modifying the ancillary ligand has only shown a modest peak wavelength shift of 10 nm, it may prove to be the most important component for the overall device efficiency by influencing the molecular orientation of the emitters [62, 63]. Homoleptic Ir(ppy)3 has an isotropic orientation where the horizontal-to-vertical transition dipole moment ratio is 0.67:0.33 and a horizontal dipole factor of 67% whereas heteteroleptic bis(2-phenylpyridine) iridium(III) acetylacetonate (Ir(ppy)2 (acac)) is comparable with Ir(ppy)3 with respect to phosphorescent properties except for a non-isotropic horizontal dipole factor of 77% due to its lower permanent dipole moment [63]. This one difference can significantly impact the EQE of an OLED without any additional outcoupling layers and performance for displays as a greater fraction of light will be emitted perpendicular to the substrate. Currently, the highest performance OLED without an additional outcoupling structure has 34.1% EQE and was fabricated with bis(2-(3,5-dimethylphenyl)-4-methylpyridine) iridium(III) (2,2,6,6tetramethylheptane-3,5-diketonate) (Ir(3′ , 5′ , 4-mppy)2 (tmd)) which has been optimized for a horizontal dipole factor of 80%, having a phosphorescence quantum yield of 97% and peak emission at 531 nm [64]. The molecular structures of these OLED emitters are highlighted in Figure 5.9 [47, 55, 60, 65, 66]. Highly efficient Ir complexes have been developed spanning the entire color spectrum from tris(1-phenylisoquinolinato-C2,N)iridium(III) (Ir(piq)3 ) which has achieved an EQE of 17%, emits at a peak wavelength of 620 nm, quantum yield of 26%, phosphorescent lifetime of 0.75 μs, and CIE coordinates of (0.68, 0.32) to bis[(4,6-difluorophenyl)pyridinato-N,C2′ ] iridium(III) picolinate (FIrpic) which has achieved an EQE of 30.1%, emits at a peak wavelength of 460 nm, quantum yield of 99%, phosphorescent lifetime of 1 μs, and CIE coordinates of (0.17, 0.34) [19, 59, 67–70]. While Ir has been the most well-studied transition metal, it is possible to fabricate emitters with other atoms such as

198

Materials for Solid State Lighting and Displays (A) λ = 510 nm

(B) λ = 520 nm

O

N

N

(C) λ = 531 nm

Ir

O

N

Ir O

3

Ir O 2

2

(D) λ = 620 nm

(E) λ = 460 nm

(F) λ = 650 nm

N

N Ir O

N

Pt

F

O

N

3

F

(G) λ = 552 nm

N

N

Ir

N

2

(H) λ = 451 nm

N N N F3C

N Pt

N N N

CF3

N N

N Pt

O

Figure 5.9 Phosphorescent dyes: (A) fac tris (2-phenylpyridine) iridium (lll) (Ir(ppy)3 ) [60]; (B) bis(2-phenylpyridine) iridium(III) acetylacetonate (Ir(ppy)2 (acac)) [60]; (C) bis(2-(3,5dimethylphenyl)-4-methylpyridine) iridium(III) (2,2,6,6-tetramethylheptane-3,5-diketonate) (Ir(3′ , 5′ , 4-mppy)2 (tmd)) [60]; (D) tris(1-phenylisoquinolinato-C2,N)iridium(III) (Ir(piq)3 ) [47]; (E) bis[(4,6-difluorophenyl)-pyridinato-N,C2’] iridium(III) picolinate (FIrpic) [47]; (F) (2,3,7,8,12,13,17,18-octaethyl-21H,23H-porphine) platinum(II) (PtOEP) [55]; (G) bis[3trifluoromethyl-5-(2-pyridyl)-1,2-pyrazolato] platinum(II) (Pt-A) [65]; (H) [6-(1,3-dihydro-3methyl-2H-imidazol-2-ylidene-𝜅C2 )-4-tert-butyl-1,2-phenylene-𝜅C1 ]oxy[9-(4-tert-butyltpyridin-2-yl-𝜅N)-9H-carbazole-1,2-diyl- 𝜅C1 ] platinum(II) (PtON7-dtb) [66]

Organic Light Emitting Device Materials for Displays

199

Pt, Cu, Os, Eu, Re, or Ru [19]. Some of these emitters have shown interesting properties such as peak emission around 700 nm for Os or very sharp emission peaks for Eu (4 nm full width at half maximum, FWHM), but the performance of such molecules is typically unsatisfactory due to either long exciton lifetimes, degradation, or otherwise poor quantum yield [19, 44, 71]. Recently Pt complexes have become competitive with Ir as the most efficient emitters with a wide color gamut. Although Pt emitters initially showed poor performance because of their planar geometry which was susceptible to aggregation, the use of suitable bulky ligands has been able to improve the emitters in this respect [72]. The first phOLED made with PtOEP had a peak emission wavelength of 650 nm indicating the affinity of such emitters for use as red dopants. To this effect bis[3-trifluoromethyl5-(2-pyridyl)-1,2-pyrazolato] platinum(II) (Pt-A) has been developed with a quantum yield of 95%, peak wavelength of 552 nm at low doping concentrations, and CIE coordinates of (0.38, 0.54) [65, 73]. Interestingly, despite having green emission when doped, Pt-A has been surprisingly effective as a neat emitter in a 31.1% EQE device with a peak wavelength of 616 nm and CIE coordinates of (0.56, 0.39) [65, 73]. In this case, Pt-A does not appear to experience the effects commonly seen in aggregated molecules where intermolecular quenching by TTA leads to a decrease in quantum yield. Instead, the phosphorescence by Pt-A was proposed to be through a Pt–Pt metal–metal-to-ligand charge transfer state [65]. As a result of the excimer-based emission, the emitter spectrum is very wide resulting in an orange-red emission. Moreover, the best performance in a deep blue emitter has also been from a Pt complex, [6-(1,3dihydro-3-methyl-2H-imidazol-2-ylidene-𝜅C2)-4-tert-butyl-1,2-phenylene-𝜅C1]oxy[9(4-tert-butyltpyridin-2-yl-𝜅N)-9H-carbazole-1,2-diyl-𝜅C1] platinum(II) (PtON7-dtb) with an EQE of 24.8%, peak wavelength of 451 nm, and CIE coordinates of (0.148, 0.079) [66]. By creating a rigid tetradentate complex as opposed to the common bidentate commonly seen in Pt emitters, the vibronic sidebands are suppressed which results in a decrease in the phosphorescence FWHM and improved color saturation [66]. Despite recent innovations in the field, highly efficient deep red and blue emitters comparable with green are still desired for OLED displays. 5.2.2.3

Thermally Activated Delayed Fluorescence Emitters

According to Hund’s rule, a triplet state will always have lower energy than its comparable singlet state because triplet electrons have the same orbital angular momentum. In most organic electroluminescent molecules, the energy separating the triplet and singlet states (ΔEST ) was assumed to be between 0.5 eV and 1.0 eV because of the significant exchange energy between the two states [74]. This large ΔEST leads to a one-sided intersystem crossing (ISC) process that converts singlet excitons to triplets and allows for efficient phosphorescence in transition metal-containing dyes. However, if ΔEST is reduced to values > kr,P ). Clearly, kRISC is the rate-limiting step in this excitonic scheme, mostly because it is energetically up-hill but also because the coupling from T1 to S1 is of moderate strength. The design rule of minimizing the energetic splitting between the spin states ΔEST induces one general problem: Typically, the strength of the fluorescence decreases with decreasing singlet–triplet splitting [63]. The reason for this is the molecular design rule (see later) that allows for minimal splittings. For the development of efficient TADF emitters it is therefore the quest to find †† In contrast to P-type delayed fluorescence, which represents TTA.

Organic White Light-Emitting Diode-Based Displays and Lighting S1

kISC kR

ISC

knr,F

ence fluoresc TADF

tion excita

kr,F

T1

RIS

C

knr,P

Rates (S−1)

ΔEST Energy (arb. units)

259

flu

0

S0 (a)

e

nc

sce ore

TADF

n kBT Singlet-triplet splitting (b)

Figure 6.18 (a) Qualitative term scheme of a thermally activated delayed fluorescence (TADF) emitter molecule. S1 and T1 denote the first excited singlet and triplet state, respectively. ki represents various rates. ISC, intersystem crossing; RISC, reverse ISC; r, radiative; nr, non-radiative; F, fluorescence; P, phosphorescence. ΔEST indicates the singlet–triplet splitting. (b) Qualitative representation of how RISC and the fluorescence rate depend on the singlet–triplet splitting ΔEST

a good balance between a large RISC and a strong fluorescence channel, as schematically illustrated in Figure 6.18(b), which show opposing dependencies on the splitting ΔEST . Figure 6.18 shows the steady-state absorption and PL of the high performance TADF emitter 4CzIPN [98]. It has a PL quantum yield of 94%, a prompt fluorescence lifetime of 17 ns, and a delayed fluorescence time constant of 5 μs [98]. This emitter can be used exemplarily for discussion, as it still represents one of the best performing TADF emitters.‡‡ In comparison with highly efficient fluorescent emitters [99], the prompt lifetime is about one order of magnitude slower, owing to the fact that ISC does reduce the singlet state lifetime accordingly. Furthermore, the delayed part shows comparable time constants with high performance phosphorescent emitters in the single digit microsecond time range [83, 86]. The experimentally determined singlet–triplet splitting is 83 meV, that is < 4 kB T at room temperature [98]. On an electronic level of the molecule, the key criteria to realize small singlet–triplet splittings is to minimize the spatial overlap of the highest occupied and lowest unoccupied molecular orbitals [100]. Figure 6.19 shows the time-dependent density functional theory (TD-DFT) calculated orbitals of 4CzIPN in Figure 6.19(b) and (c), respectively. One can clearly see that the majority of the respective orbitals is distributed either on the carbazole ligands (HOMO) or on the central dicyanobenzene (LUMO). This design breaks the conjugation within the molecule, such that intramolecular charge-transfer states (CT states) are formed between the donor (here: carbazole) and acceptor (here: dicyanobenzene) units. At the same time, it is important to maintain a certain degree of overlap between the orbitals, as otherwise the fluorescence from the formed CT state would vanish (cf. Figure 6.18). ‡‡ This emitter 4CzIPN has been used to demonstrate OLED EQEs of about 19%, which approaches the theoretical maximum

for EQE based on 20–30% outcoupling efficiency [98].

260

Materials for Solid State Lighting and Displays 5

Photoluminescence intensity (a.u.)

Molar extinction coefficient (105 1/mol/cm)

4

3

2

1

0 300

(b)

400

500 Wavelength (nm) (a)

600

700

(c)

Figure 6.19 (a) Absorption and photoluminescence of a highly efficient TADF emitter 4CzIPN based on carbazole donor and cyano acceptor groups centered around a benzene ring. (b) and (c) show the calculated (TD-DFT) orbital densities of HOMO and LUMO wave functions, respectively. Source: Uoyama et al. 2012 [98]. Reproduced with permission of Nature Publishing Group

The separation of HOMO and LUMO can also be introduced artificially at a donor–acceptor hetero-interface between two separate molecules. If a hole resides on a donor molecule and an electron on an adjacent acceptor molecule, the charge separation is virtually complete, giving rise to almost degenerate singlet and triplet states. Importantly, these states are CT states, that is electron–hole pairs bound across the hetero-interface. Thermally activated delayed fluorescence from these kind of CT states has been observed. Qualitatively, the exciton dynamics are similar to intramolecular TADF. Emitting donor–acceptor interfaces have been used and studied for decades – the respective CT

Radiative rate

Control of torsion angle Increasing fluorescence

Reducing singlet-triplet splitting

Organic White Light-Emitting Diode-Based Displays and Lighting

D

Control of strength

261

Control of spatial separation

A

D = donor, A = acceptor

Figure 6.20 Scheme illustrating the molecular design rules that allow for efficient separation of HOMO and LUMO wave functions and thus small singlet–triplet splittings

state is commonly called exciplex. It is remarkable that the effect of exciplex-based TADF had not been discovered until recently [100], when the intramolecular concept for TADF was already known [97]. The exciplex-type material combinations have become very attractive host systems [101], as they minimize the general band gap of the host and allow for effective mixing between singlet and triplet states of the host, which is considered beneficial for various emitter types (TADF and phosphorescence). While the charge separation and the resulting small singlet–triplet splitting comes naturally for exciplex-based TADF, one has to carefully design intramolecular CT states to obtain the desired properties. Figure 6.20 illustrates the three most common concepts one can follow on the molecular design level to achieve small ΔEST values. One way of achieving the decoupling of conjugation and thus allow for separate hole and electron densities is the introduction of a large dihedral angle between donor and acceptor units. This can be intensified by adding electron donating or withdrawing character to the units through chemical modification. For example, a cyano group can be added to intensify the electron accepting character [98]. The control of HOMO and LUMO distribution can also solely be set also by a controlled implementation of such electron pushing and pulling functional groups. Finally, the molecular design can also be altered by physical spacing of non-conjugated linkers to induce an intramolecular CT state. Recently, blue TADF-based OLEDs with high performance have been reported as well [102]; however, at this stage, demonstration of superior device lifetime and color purity (which is important for display applications) have not been reported. 6.5.7

Phosphorescence Versus Thermally Activated Delayed Fluorescence

It is worth comparing TADF and phosphorescence briefly, as both of them offer highly efficient OLEDs with 𝜂int ∼ 100%. The concepts to achieve this remarkable result however vastly differ. While phosphorescent emitters are designed based on strong spin-orbit coupling by introduction of heavy transition metals to the molecule’s center, TADF molecules can be made from purely organic building blocks [103].§§ Figure 6.21 compares §§ It must be noted for completeness that TADF materials can also be made from metal-containing molecules. Here, copper

complexes are prototypical material systems. As the copper atoms also introduce SOC, such materials are often accompanied

Materials for Solid State Lighting and Displays Phosphorescence S1

k IS

ΔEgap

C

ΔEST

TADF

Energy (eV)

T1

S1

kRISC

T1

kISC kr,F

kr,PP

kr,F

kISC ˜ kr,F kR RISC >> knr,P

kISC >> kr,F S0

knr,P

S0 (a)

(b)

ΔEST

Intensity (a.u.)

262

Extra blue emission Wavelength (nm) (c)

Figure 6.21 Comparison of the excitonic mechanisms of (a) phosphorescence and (b) thermally activated delayed fluorescence (TADF). The schemes are drawn for equal emission state energies (gray shading). ki represents various rates. ISC, intersystem crossing; RISC, reverse ISC; r, radiative; nr, non-radiative; F, fluorescence; P, phosphorescence. ΔEST indicates the singlet–triplet splitting and ΔEgap is the gap between respective singlet states of phosphorescent and TADF emitters. (c) Qualitative shapes of phosphorescence and TADF-type photoluminescence. Source: Reineke 2014 [104]. Reproduced with permission of Nature Publishing Group

schematically the energetic differences between phosphorescence and TADF. While in TADF emitters, the 75% share of triplet excitons access a slightly energetically up-hill state for emission (S1 ), phosphorescent emitters collect all excitons on the energetically lowest state, where emission takes place (phosphorescence). Consequently, excitons may cycle within their lifetime many times between singlet and triplet manifold, whereas in phosphorescent emitters, excitons are trapped in the phosphorescent states. By comparing the schemes of Figure 6.21, it is easy to see that TADF emitters are conceptually lower band gap materials compared with phosphorescent ones, when setting both concepts to a similar vertical transition energy, as indicated by the shaded arrows. This is again due to the fact that in TADF, the higher lying singlet state is defining the emitting energy, whereas it is the triplet state for the case of phosphorescence [104]. Consequently, the band gap of the molecule itself can be significantly higher [e.g., 600 meV for the archetypical phosphorescent emitter Ir(ppy)3 ] [82]. While it has not been demonstrated yet, this difference of the material’s gap might allow TADF emitters to provide a pathway for long-term stable blue OLEDs. Finally, it is worth comparing the typical emission line shapes of TADF and phosphorescence. This difference is qualitatively illustrated in Figure 6.21(c). While phosphorescence shows a significant asymmetry and also often distinct vibronic substructure, TADF sports a very comparably symmetric, featureless emission band. The latter is due to the CT character of the emitting states. When now comparing TADF and phosphorescence with similar PL maxima, it can be seen that the TADF has a notable high energy shoulder in the spectrum that cannot be observed in phosphorescent emitters. This TADF phenomenon is very with substantial competition between fluorescent and phosphorescent channels. The true origin of the resulting emission band cannot be assigned easily as both possible bands are very close in energy, which again is a prerequisite for effective RISC.

Organic White Light-Emitting Diode-Based Displays and Lighting

263

attractive for solid-state lighting, as such a broad blue emitter fits well into the spectral distribution of a white spectrum. In constrast, the broad emission spectra of TADF OLEDs often reaching values up to 100 nm FWHM is a disadvantage for display applications, as those require a high color purity, which is obtained best with small FWHM emission bands. Here, phosphorescence might be more suitable. 6.5.8

TADF Assisted Fluorescence (TAF) Emitters

Mostly motivated to enhance the color purity of TADF OLEDs, Adachi and co-workers introduced an energy transfer scheme pairing TADF and fluorescent emitters [74, 105]. This concept is similar to the phosphorescence-sensitized fluorescence reported by Baldo et al. for the case of phosphorescence [73]. In such a system, the TADF emitter is used only to mix the singlet and triplet states effectively after initial generation. By adding a fluorescent emitter which is energetically accessible from the TADF emitter’s singlet state via FRET, the singlet states are directed to a fluorophore that sports fast fluorescence lifetimes, high PLQY, and comparably narrow emission line width. Figure 6.22(a) compares TADF with this hybrid approach, which has been termed TADF assisted fluorescence (TAF) [74, 105]. Figure 6.22(b) shows the PL of the donor TADF material, the absorption of the fluorescent acceptor, as well as its respective fluorescence spectrum. The large spectral overlap and the high PLQY of the TADF mixing agent allow for long Förster radii, such that effective energy transfer to the fluorescent acceptor can be guaranteed [74, 105]. This concept has one very significant change in the exciton dynamics compared with the sole TADF emitter. By introducing the FRET mechanism from TADF S1 to acceptor S1 , an additional and fast rate for depopulation of the TADF singlet state is introduced. This artificially increases the fluorescence rate in the system, not in the TADF singlet manifold, but rather through the singlet state of the acceptor. Thus, this mechanism compensates for the rather weak fluorescence rates in TADF emitters. Clearly, this concept becomes more and more complex with the transition from red to blue terminal emission color, simply because blue emission requires the respective TADF mixing materials to be even larger band gap materials to work properly.

6.6

Polymer Concepts

So far, this chapter has solely discussed material concepts based on small molecules, which can be fabricated via physical vapor deposition (PVD) techniques. The goal so far was to discuss all the various molecular and excitonic concepts in a clear and straightforward fashion. For this purpose, small molecules can be seen as the prototypical material class, as PVD allows for fabrication of devices of highest complexity, purity, and therefore efficiency. Thus, the concepts and device results obtained in such systems can serve well for polymeric material and device concepts as benchmark and upper limit. Without doubt, polymer-based OLEDs [9, 106] (often called PLEDs) have very high potential to ultimately offer superior fabrication processes such as wet-coating or printing that will allow for much lower fabrication costs. This is imperative for white OLEDs for

264

Materials for Solid State Lighting and Displays

+ −

+ − 25%

25%

75%

75% FRET RISC

RISC

Dexter

VS S1

ISC

S1

T1

ISC

T1

S1 T1

4CzIPN-Me

4CzIPN-Me

TBRb 0.10

TBRb Abs.

4CzIPN-Me PL 0.08 TBRb PL

0.06 0.04

Absorbance

Emission intensity (arb. unit)

(a)

0.02 500

600 Wavelength (nm) (b)

0.00 700

Figure 6.22 (a) Left: General scheme for thermally activated delayed fluorescence (TADF). Right: TADF paired with a fluorescent acceptor molecule giving rise to TADF assisted fluorescence (TAF). (b) Emission spectra of a TADF donor 4CzIPN-Me, and absorption and emission of a fluorescent acceptor molecule TBRb indicating substantial spectral overlap between donor emission and acceptor absorption needed for Förster resonant energy transfer. Source: Furukawa et al. 2015 [105]. Reproduced with permission of Nature Publishing Group

solid-state lighting, as the competition of inorganic LEDs is strong [1] and their pricing very aggressive. Generally, polymer systems need to fulfill the same criteria as small molecules to realize high performance electroluminescence. Most importantly, polymer white OLEDs need to allow for: • 𝜂int ∼ 100%; • a balanced, high color quality white emission; • long-term stability. The last criterion mentioned for PVD-based OLEDs, that is low cost, can be realized by wet-processing much easier, where the ultimate cost reduction is only awaiting substantial up-scaling of the desired process. In order to ramp up production however, a viable route to meet the three points give above needs to be found. The discussion of polymer concepts in this chapter will be limited to highlighting some molecular designs that are unique to this very material class.

Organic White Light-Emitting Diode-Based Displays and Lighting

6.6.1

265

Various Concepts Involving Polymer Materials

Figure 6.23 shows the major approaches to realize white light in PLEDs that already illustrates the diversity found in possible concepts. One early approach was the use of polymeric host materials that were doped with fluorescent small molecules to realize white-light emission [cf. Figure 6.23(a)] [107]. Here, the polymeric host needs to assure that charges of both sign can be transported effectively and that the polymer itself has a substantially large band gap not to introduce energetic traps (loss channels). Concepts based on monochrome fluorescent polymers are shown in Figure 6.23(b) and (c). Here, differently emitting polymers are either blended to form a bulk hetero-structure or deposited consecutively forming a planar interface between the two materials. Figure 6.23(d) illustrates a molecular concept for white PLEDs, which is unique to polymer material design. Here polymer structures are developed that sport the various color chromophores needed for white emission in hierarchical structures. These polymers can either be designed to have various chromophores in chains with different repeating frequency or be made of main and side chains. In the latter approach, chromophore units can be fitted into the main chain and/or added in dedicated side chain positions. Liu et al. designed white-light emitting polymers based on fluorescent chromophores as shown in Figure 6.24 [108]. In their concept, the polymer backbone (or main chain) serves as host material and blue chromophore simultaneously [108]. While the green chromophore is placed on a dedicated side chain only, they discuss two polymer variants with the red

blue blue green

yellow

red (a) yellow

(b) blue

blue green red (c)

(d)

Figure 6.23 Various types of concepts for white polymer-based OLEDs (white PLEDs). (a) Polymer matrix materials (gray) doped with small molecules (either of fluorescent or phosphorescent; filled symbols, different shade and shape indicate different emitters), (b) two or more light-emitting polymers blended as a bulk film, (c) a hetero-junction made of two light-emitting polymers, and (d) white-light emission based on multicomponent copolymers. Source: Reineke 2013 [4]. Reproduced with permission of American Physical Society

266

Materials for Solid State Lighting and Displays excitation

blue polymer host

red dopant

green dopant

White electroluminescence

x

1-x-y C8H17

C8H17

y

n

C6H13

C6H13

WRGB-P1, x = 0.0002, y = 0.0002 P1-30, x = 0, y = 0.0030, O

O

O

N

N

S N

S

N

N

C8H17

C8H17

N

x

1-x-y

n

C6H13

WRGB-P2, x = 0.0002, y = 0.0002 P2-30, x = 0, y = 0.0030,

S O

O

N

N S N

O

N

C10H21 O N

N

OCH3

H3CO

S N

GMC

S

N

RMC

N

n C8H17

C8H17

PF

Figure 6.24 Working principle of a white-light emitting copolymer made of a blue-emitting polymer backbone decorated with green- and red- emitting side-chain chromophores. The blue-emitting backbone is a poly(fluorene-co-benzene) (PF), DPAN and MB-BT-ThTPA serve as the green (GMC) and red (RMC) model compound, respectively. Source: Liu et al. 2007 [108]. Reproduced with permission of John Wiley and Sons

Organic White Light-Emitting Diode-Based Displays and Lighting

267

chromophore either being added to the main chain in a stoichiometric fashion or added to a side chain similar to the green chromophore. The key advantage of such an approach is clearly its simplicity. If the ratios of the different chromophores are met on the level of material synthesis, white-light electroluminescence can be as simple as casting one of these polymers between suitable transport layers or even just electrodes. This approach will also reduce color shifts as a function of driving conditions to a minimum. Still, one has to note that this concept, while realizing white light, is only made from fluorescent chromophores. Therefore, the maximum internal quantum efficiency one can expect from such a material will similarly be limited to 𝜂int, fl ∼ 25%. Those limitations can only be overcome by utilizing similar approaches as discussed in Section 6.5. 6.6.2

Learning from High Performance Small Molecules for High Efficiency Polymers

In this section, some highlights of polymer material concepts will be discussed that focus on the implementation of highly efficient excitonic concepts, similar to those discussed for small molecules in Section 6.5. Based on the hybrid main chain/side chain concept introduced above (cf. Figure 6.24), efforts are made to replace the fluorescent chromophores by phosphorescent counterparts [109]. Here, the main challenge is to make sure that the complex polymer made of various decoupled electronic units does not include sites that act as exciton quenchers. For instance, it is important to have all differently emitting chromophores to sport phosphorescent character. Otherwise, it is very likely that triplet states can still be quenched. With the back-bone part of such a hybrid molecule co-acting as fluorescent emitter, the most viable strategy might be to mimic the triplet harvesting concept, where the fluorescent main chain has significantly high triplet level to pass on the formed triplet states to phosphorescent emitters in the green to red spectral range, assuring overall 100% internal quantum efficiency [70, 91]. Phosphorescent emitters have seen another very interesting development. The molecules similar to the ones shown in Figure 6.3 are embedded in the center of polymeric branched networks called dendrimers [110–112]. Here, organic super structures are linked covalently to the ligand structures of the phosphorescent emitters in a way that the electronic structure of the emitter remains effectively unperturbed. Those structures can be added in various lengths with different functionalities. Those extended organic systems can support charge transport so that dendrimeric systems can be used as hierarchical OLED materials. In addition, the dendrimeric systems effectively separate the phosphorescent emitters in a solid film, avoiding emitter aggregation [6, 113] and thus, showing improved high brightness performance via reduced TTA. Recently, triggered by the dawn of TADF in small molecules and exciplex systems, the first polymer TADF concepts have been reported [114, 115]. In general, the design rules for achieving TADF in polymers is very similar to small molecules. Donor and acceptor moieties need to be linked in a way that HOMO and LUMO are spatially separated to assure small singlet–triplet splittings ΔEST . Figure 6.25 shows a recent polymer concept showing efficient TADF. For this purpose, a polymer backbone is designed with alternating donor and acceptor units [114]. The latter

268

Materials for Solid State Lighting and Displays

N

O

N

N

O

N

N

N n

R

pCzBP

n

R

R

pAcBP

R=

R

(a)

LUMO –1.94 eV

LUMO

3.04 eV ΔEST = 0.16 eV

–2.06 eV

2.46 eV ΔEST = 0.004 eV

–4.52 eV

HOMO –4.98 eV

HOMO (b)

Figure 6.25 (a) Chemical structures of two different repeat units of polymer materials pCzBP and pAcBP showing thermally activated delayed fluorescence. The donor and acceptor units of these materials are situated along the polymeric backbone. (b) TD-DFT calculations for the smallest representative systems, that is (D-A-D)-D’-(D-A-D), showing the respective HOMO and LUMO orbitals and energies. D, donor; A, acceptor; D is carbazole for pCzBP and acridan for pAcBP; A is benzophenone. Source: Lee et al. 2016 [114]. reproduced with permission of John Wiley and Sons

is benzophenone, which is either supplemented with carbazole or acridan donor units. All those building blocks are well known from small molecule reports, so that the main task is to effectively transfer the intramolecular charge transfer (ICT) character induced by the individual building blocks into an extended polymer chain. So far, only single emission band TADF-type polymers have been reported, but it can be easily imagined that multiple, differently emitting ICT configurations can be implemented in more complex, stoichiometric polymers to realize white light emission.

6.7

Summary and Outlook

From their first presentation in 1987 by Tang and VanSlyke [3], OLEDs look back on almost three decades of intensive research and development. While many products, mainly in the

Organic White Light-Emitting Diode-Based Displays and Lighting

269

display sector, have already reached the customers, it still seems that the research is far away from seeing a saturation. This observation is mostly due to the fact that the material designer’s toolbox is virtually unlimited with uncountable numbers of possible organic molecules [116]. This is in stark contrast to the field of inorganic LEDs, where optimization is rather on the fine-tuning and engineering of the limited high-performance semiconductors available [1, 10]. Therefore, OLED material research will carry on as long as not all of the device performance parameters are met with complete satisfaction. This includes color purity, efficiency, long-lifetime, and so on, paired with the ability to produce all these devices on flexible substrates, allowing for freeform solutions in the display and solid-state lighting sectors. Definitely, future OLED solutions do not allow for internal quantum efficiencies smaller that 100% in order to remain competitive with other technologies such as inorganic LEDs for solid-state lighting and colloidal quantum dot enhanced white LED backlights for display applications. The aforementioned variety in the molecular design [116] clearly represents a large potential for the OLED field as a whole, but at the same time also risks hampering the rate of progress of the technology. While OLEDs in the display sector are at the turning point of becoming one of the future, high quality and high volume technologies even for large TV size panels, the future for white OLEDs in the lighting sector is not set. Here, the cost of existing lighting products put a lot of pressure on the possible market penetration of white OLEDs for lighting. The key reason is that the consumer will not be able to afford a premium for a daily use lighting product, which is completely different for the high-end display sector. An additional risk for white OLED panels is the missing infrastructure that is needed to allow their wide penetration into the general lighting sector. With the ultrathin, large-area form factor of OLEDs paired with their possible compatibility with many (flexible) substrates, OLEDs will be an integral part of future display and lighting products sporting both high color quality and high efficiency. The future of artificial lighting will be dominated by a complementary use of inorganic LEDs and OLEDs [10], jointly replacing ineffective, low color-quality light sources of today.

References 1. 2. 3. 4.

Narukawa, Y., Narita, J., Sakamoto, T. et al. (2006) Jpn. J. Appl. Phys., 45, L1084. Scholz, S., Kondakov, D., Lüssem, B. and Leo, K. (2015) Chem. Rev., 115, 8449. Tang, C.W. and VanSlyke, S.A. (1987) Appl. Phys. Lett., 51, 913. Reineke, S., Thomschke, M., Lüssem, B. and Leo, K. (2013) Rev. Mod. Phys., 85, 1245. 5. Walzer, K., Maennig, B., Pfeiffer, M. and Leo, K. (2007) Chem. Rev., 107, 1233. 6. Kawamura, Y., Goushi, K., Brooks, J. et al. (2005) Appl. Phys. Lett., 86, 071104. 7. Kawamura, Y., Brooks, J., Brown, J.J. et al. (2006) Phys. Rev. Lett., 96, 017404. 8. He, G., Pfeiffer, M., Leo, K. et al. (2004) Appl. Phys. Lett., 85, 3911. 9. Seidler, N., Reineke, S., Walzer, K. et al. (2010) Appl. Phys. Lett., 96, 093304. 10. Reineke, S. (2015) Nat. Mater., 14, 459. 11. Gohri, V., Hofmann, S., Reineke, S. et al. (2011) Org. Electron., 12, 2126.

270

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

26.

27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45.

Materials for Solid State Lighting and Displays

D’Andrade, B.W., Holmes, R.J. and Forrest, S.R. (2004) Adv. Mater., 16, 624. Thomschke, M., Reineke, S., Lüssem, B. and Leo, K. (2012) Nano Lett., 12, 424. Freitag, P., Reineke, S., Olthof, S. et al. (2010) Org. Electron., 11, 1676. Kanno, H., Sun, Y. and Forrest, S.R. (2005) Appl. Phys. Lett., 86, 263502. Bulovic, V., Gu, G., Burrows, P.E. et al. (1996) Nature, 380, 29. Smith, G.S. (1997) An Introduction to Classical Electromagnetic Radiation, Cambridge University Press, Cambridge. Lambert, J.H. (1892) Lamberts Photometrie, Wilhelm Engelmann Verlag, Leipzig. Diras, H.H. (2011) Human Biology, 7th edn, Jones & Bartlett Learning, Sudbury. (1931) Commission Internationale de l’Eclairage Proceedings, Cambridge University Press, Cambridge. Smith, T. and Guild, J. (1931) Trans. Opt. Soc., 33, 73. Wright, W.D. (1929) Trans. Opt. Soc., 30, 141. Flesch, P. (2006) Light and Light Sources High-Intensity Discharge Lamps, Springer, Berlin. Tsujimura, T. (2012) OLED Display Fundamentals and Applications, John Wiley & Sons, Inc, Hoboken, NJ. (1995) Technical Report - Method of Measuring and Specifying Colour Rendering Properties of Light Sources Commission Internationale de l’Eclairage, Vienna. Available at http://www.cie.co.at/index.php/Publications/index.php?i_ca_id=303. Accessed September 12, 2016. (2008) ENERGY STAR Program Requirements for Solid State Lighting Luminaires Eligibility Criteria - Version 1.1, US Enviromental Protection Agency and US Department of Energy, Washington, DC. MacAdam, D.L. (1943) J. Opt. Soc. Am., 33, 18. Yokoyama, D., Sakaguchi, A., Suzuki, M. and Adachi, C. (2009) Org. Electron., 10, 127. Yokoyama, D. and Adachi, C. (2010) J. Appl. Phys., 107, 123512. Stenzel, O. (2005) The Physics of Thin Film Optical Spectra, Springer-Verlag, Berlin. Yeh, P. (2005) Optical Waves in Layered Media, 2nd edn, Wiley-Interscience, Hoboken, NJ. Katsidis, C.C. and Siapkas, D.I. (2002) Appl. Opt., 41, 3978. Benisty, H., Stanley, R. and Mayer, M. (1998) J. Opt. Soc. Am. A, 15, 1192. Neyts, K.A. (1998) J. Opt. Soc. Am. A, 15, 962. Sullivan, K.G. and Hall, D.G. (1997) J. Opt. Soc. Am. B, 14, 1149. Schmidt, T.D., Setz, D.S., Flämmich, M. et al. (2011) Appl. Phys. Lett., 99, 163302. Kim, S.-Y., Jeong, W.-I., Mayr, C. et al. (2013) Adv. Funct. Mater., 23, 3896. Purcell, E.M., Torrey, H.C. and Pound, R.V. (1946) Phys. Rev., 69, 37. Furno, M., Meerheim, R., Hofmann, S. et al. (2012) Phys. Rev. B, 85, 115205. Brütting, W., Frischeisen, J., Schmidt, T.D. et al. (2013) Phys. Status Solidi A, 210, 44. Saxena, K., Jain, V.K. and Mehta, D.S. (2009) Opt. Mater., 32, 221. Gather, M.C. and Reineke, S. (2015) J. Photon. Energy, 5, 057607. Reineke, S., Lindner, F., Schwartz, G. et al. (2009) Nature, 459, 234. Greiner, H. (2007) Jpn. J. Appl. Phys., 46, 4125. Meerheim, R., Nitsche, R. and Leo, K. (2008) Appl. Phys. Lett., 93, 043310.

Organic White Light-Emitting Diode-Based Displays and Lighting

46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81.

271

Peng, H., Zhu, X., Sun, J. et al. (2005) Appl. Phys. Lett., 87, 173505. Puzzo, D.P., Helander, M.G., O’Brien, P.G. et al. (2011) Nano Lett., 11, 1457. Callens, M.K., Yokoyama, D. and Neyts, K. (2015) Opt. Express, 23, 21128. Becker, H., Burns, S.E., Tessler, N. and Friend, R.H. (1997) J. Appl. Phys., 81, 2825. Meerheim, R., Furno, M., Hofmann, S. et al. (2010) Appl. Phys. Lett., 97, 253305. Smith, L.H., Wasey, J.A.E. and Barnes, W.L. (2004) Appl. Phys. Lett., 84, 2986. Nowy, S., Krummacher, B.C., Frischeisen, J. et al. (2008) J. Appl. Phys., 104, 123109. Setz, D.S., Schmidt, T.D., Flämmich, M. et al. (2011) J. Photonics Energy, 1, 011006. Hobson, P.A., Wedge, S., Wasey, J.A.E. et al. (2002) Adv. Mater., 14, 1393. Hung, L.S., Tang, C.W., Mason, M.G. et al. (2001) Appl. Phys. Lett., 78, 544. Riel, H., Karg, S., Beierlein, T. et al. (2003) J. Appl. Phys., 94, 5290. Huang, Q., Walzer, K., Pfeiffer, M. et al. (2006) J. Appl. Phys., 100, 064507. Hsu, S.-F., Lee, C.-C., Hwang, S.-W. and Chen, C.H. (2005) Appl. Phys. Lett., 86, 253508. Forrest, S.R. (2004) Nature, 428, 911. Baldo, M.A., O’Brien, D.F., You, Y. et al. (1998) Nature, 395, 151. Reineke, S. and Baldo, M.A. (2012) Phys. Status Solidi A, 209, 2341. Congreve, D.N., Lee, J., Thompson, N.J. et al. (2013) Science, 340, 334. Turro, N.J., Ramamurthy, V. and Scaiano, J.C. (2010) Modern Molecular Photochemistry of Organic Molecules, University Science Books, Sausalito, CA. Okumoto, K., Kanno, H., Hamaa, Y. et al. (2006) Appl. Phys. Lett., 89, 063504. Mattoussi, H., Murata, H., Merritt, C.D. et al. (1999) J. Appl. Phys., 86, 2642. Mei, J., Leung, N.L.C., Kwok, R.T.K. et al. (2015) Chem. Rev., 115, 11718. Kondakov, D.Y., Pawlik, T.D., Hatwar, T.K. and Spindler, J.P. (2009) J. Appl. Phys., 106, 124510. Meerheim, R., Walzer, K., Pfeiffer, M. and Leo, K. (2006) Appl. Phys. Lett., 89, 061111. Sun, Y., Giebink, N.C., Kanno, H. et al. (2006) Nature, 440, 908. Schwartz, G., Reineke, S., Rosenow, T.C. et al. (2009) Adv. Funct. Mater., 19, 1319. Broussard, J.A., Rappaz, B., Webb, D.J. and Brown, C.M. (2013) Nat. Protoc., 8, 265. Braslavsky, S.E., Fron, E., Rodríguez, H.B. et al. (2008) Photochem. Photobiol. Sci., 7, 1444. Baldo, M.A., Thompson, M.E. and Forrest, S.R. (2000) Nature, 403, 750. Nakanotani, H., Higuchi, T., Furukawa, T. et al. (2014) Nat. Commun., 5, 4016. Thompson, M. (2007) MRS Bull., 32, 694. Pope, M. and Swenberg, C.E. (1999) Electronic Processes in Organic Crystals, Oxford University Press, New York. Adachi, C., Baldo, M.A., Thompson, M.E. and Forrest, S.R. (2001) J. Appl. Phys., 90, 5048. Yersin, H. (2004) in Transition Metal and Rare Earth Compounds Springer, Berlin, pp. 1–26. Reineke, S., Seidler, N., Yost, S.R. et al. (2013) Appl. Phys. Lett., 103, 093302. Reineke, S. and Baldo, M.A. (2014) Sci. Rep., 4. Kobayashi, T., Ide, N., Matsusue, N. and Naito, H. (2005) Jpn. J. Appl. Phys., 44, 1966.

272

82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116.

Materials for Solid State Lighting and Displays

Tsuboi, T. (2006) J. Lumin., 119–120, 288. Reineke, S., Walzer, K. and Leo, K. (2007) Phys. Rev. B, 75, 125328. Murawski, C., Leo, K. and Gather, M.C. (2013) Adv. Mater., 25, 6801. ̇ Kalinowski, J., Stampor, W., M¸ezyk, J. et al. (2002) Phys. Rev. B, 66, 235321. Baldo, M.A., Adachi, C. and Forrest, S.R. (2000) Phys. Rev. B, 62, 10967. Dexter, D.L. (1953) J. Chem. Phys., 21, 836. Su, S.-J., Gonmori, E., Sasabe, H. and Kido, J. (2008) Adv. Mater., 20, 4189. Giebink, N.C., Sun, Y. and Forrest, S.R. (2006) Org. Electron., 7, 375. Wünsche, J., Reineke, S., Lüssem, B. and Leo, K. (2010) Phys. Rev. B, 81, 245201. Rosenow, T.C., Furno, M., Reineke, S. et al. (2010) J. Appl. Phys., 108, 113113. Andrade, B.w.D.’., Brooks, J., Adamovich, V. et al. (2002) Adv. Mater., 14, 1032. Connick, W.B., Henling, L.M., Marsh, R.E. and Gray, H.B. (1996) Inorg. Chem., 35, 6261. Ma, B., Djurovich, P.I., Garon, S. et al. (2006) Adv. Funct. Mater., 16, 2438. Cocchi, M., Kalinowski, J., Fattori, V. et al. (2009) Appl. Phys. Lett., 94, 073309. Cocchi, M., Kalinowski, J., Virgili, D. et al. (2007) Appl. Phys. Lett., 90, 163508. Endo, A., Ogasawara, M., Takahashi, A. et al. (2009) Adv. Mater., 21, 4802. Uoyama, H., Goushi, K., Shizu, K. et al. (2012) Nature, 492, 234. Chang, C.-W., Kao, Y.-T. and Diau, E.W.-G. (2003) Chem. Phys. Lett., 374, 110. Goushi, K., Yoshida, K., Sato, K. and Adachi, C. (2012) Nat. Photonics, 6, 253. Shin, H., Lee, S., Kim, K.-H. et al. (2014) Adv. Mater., 26, 4730. Zhang, Q., Li, B., Huang, S. et al. (2014) Nat. Photonics, 8, 326. Hofbeck, T., Monkowius, U. and Yersin, H. (2015) J. Am. Chem. Soc., 137, 399. Reineke, S. (2014) Nat. Photonics, 8, 269. Furukawa, T., Nakanotani, H., Inoue, M. and Adachi, C. (2015) Sci. Rep., 5, 8429. Burroughes, J.H., Bradley, D.D.C., Brown, A.R. et al. (1990) Nature, 347, 539. Kido, J., Kimura, M. and Nagai, K. (1995) Science, 267, 1332. Liu, J., Xie, Z.Y., Cheng, Y.X. et al. (2007) Adv. Mater., 19, 531. Jiang, J.X., Xu, Y.H., Yang, W. et al. (2006) Adv. Mater., 18, 1769. Namdas, E.B., Ruseckas, A., Samuel, I.D.W. et al. (2004) J. Phys. Chem. B, 108, 1570. Burn, P.L., Lo, S.-C. and Samuel, I.D.W. (2007) Adv. Mater., 19, 1675. Levell, J.W., Zhang, S., Lai, W.-Y. et al. (2012) Opt. Express, 20, A213. Reineke, S., Schwartz, G., Walzer, K. et al. (2009) Appl. Phys. Lett., 94, 163305. Lee, S.Y., Yasuda, T., Komiyama, H. et al. (2016) Adv. Mater. doi: 10.1002/adma .201505026 Nobuyasu, R.S., Ren, Z., Griffiths, G.C. et al. (2016) Adv. Opt. Mater., 4, 597. Reymond, J.-L., Blum, L.C. and van Deursen, R. (2011) Chim. Int. J. Chem., 65, 863.

7 Light Emitting Diode Materials and Devices Michael R. Krames Arkesso LLC, Palo Alto, CA, USA

7.1

Introduction

Ignited by the demonstration of visible-spectrum emission from a mixed III–V compound semiconductor device in 1962 [1], light emitting diode (LED) technology advanced dramatically in subsequent decades and now devices are available at all wavelengths of the visible spectrum and, in almost all cases, with efficiencies that exceed that of conventional light sources. These advances have led to penetration into various important commercial markets, first signaling and automotive applications, then mobile, and now into general lighting and large-area displays. In this chapter, we review the key material systems for LEDs and practical aspects that drove their development, as well as summarize state-of-the-art performance (circa 2016).

7.2 7.2.1

Light Emitting Diode Basics Construction

LEDs, like purely electronic diodes, comprise p-type and n-type layers to support a functional p-n junction (as discussed in Chapter 1 and reference [3] of Chapter 1). In addition, LEDs typically have specially selected active region materials inserted at or near the junction, to control the light generation, especially its color. A typical (simplified) LED chip construction is shown in Figure 7.1. A substrate wafer supports the epitaxial (atomic, typically crystal-lattice-matched) growth process for the

Materials for Solid State Lighting and Displays, First Edition. Edited by Adrian Kitai. © 2017 John Wiley & Sons Ltd. Published 2017 by John Wiley & Sons Ltd.

274

Materials for Solid State Lighting and Displays

anode p layer(s)

+

active region n layer(s) buffer layer(s)

substrate

cathode (a)

Figure 7.1 diode

-

(b)

(a) Simplified structure and (b) electrical schematic symbol for a light emitting

subsequent layers. Often, a “buffer” layer is grown on the substrate to distance the sensitive active region from impurities or defects that may be associated with the substrate interface. Then, the n-type layers are grown, designed to inject electrons efficiently into the active region, which follows next. The p-type layers are grown above the active layers, designed to efficiently inject holes. One or more heavily doped contact layers (not shown in Figure 7.1) are often provided to ensure low Ohmic contact resistances. Electrodes are formed via Ohmic contacts at the p- and n-type layers, or (as shown in Figure 7.1) on the substrate, providing that is conductive. In some cases, the ordering of the n-type and p-type layers is reversed. Also, in some cases electrodes are formed on the same “side” of the LED wafer, for example, in a “flip-chip” configuration. The epitaxial growth processes used for LEDs are mostly vapor phase epitaxy (VPE), liquid phase epitaxy (LPE), and metal organic chemical vapor deposition (MOCVD). The former two enjoy high growth rates but are challenging for thin layers and doping control. MOCVD, with its precise layer control and high volume capability, is by far the preferred growth method for the visible spectrum LED industry today. Molecular beam epitaxy (MBE) is not commonly used for the production of visible-spectrum LEDs but it is used in research and development here and there. Wafer fabrication is similar to that for conventional, that is, Si, semiconductor devices, and involves the steps of Ohmic contact formation, texturing or shaping for light out-coupling, and singulation. The number of mask layers (typically five to six) is much lower than that required for an integrated circuit. However, the more complicated materials combinations in LEDs can make wafer fabrication challenging. Also, substrate materials are more exotic than for Si. For these reasons the wafer diameters for LED processes in industry today are between 2” and 6”, as opposed to the 8” process which is mainstream for Si wafer fabrication facilities today. After singulation, LED chips are mounted in packages – of which there are wide varieties – tested, binned (in terms of forward voltage, color or peak emission wavelength, and light output), and then shipped in reels as for other semiconductor devices. Phosphor

Light Emitting Diode Materials and Devices

275

materials, if used, may be applied at various steps in the process, including at the wafer level, or after the chips are package mounted. 7.2.2

Recombination Processes

Charge carriers in a semiconductor can recombine and release energy in a number of different ways. While many complicated combinations are possible, the main recombination pathways that dominate LED performance are those illustrated in Figure 7.2. The first, and most desirable, recombination process is the direct recombination of an electron in the conduction band with a corresponding hole in the valence band to produce a photon of energy roughly equal to the semiconductor energy bandgap, sometimes referred to as bimolecular radiative recombination. A second set of processes is similar but wherein recombination involves a donor or acceptor site, that is, donor-level-to-valence-band or conduction-band-to-acceptor-level. This recombination can be radiative and reasonably efficient, but is slower than direct radiative recombination and produces photons with energies smaller than the bandgap. Donor-level-to-acceptor-level is also a possible transition within this set of recombination processes. A third process is one that involves traps for electrons and/or holes, in which case a charge carrier is localized to an impurity and recombines there with its corresponding recombination partner, a process which is predominantly non-radiative and inherently lossy. A fourth process is Auger recombination, wherein the recombination of an electron–hole pair imparts energy to a nearby charge carrier, elevating its energy level by an amount equivalent to the recombining electron–hole pair. The resulting “hot” carrier thermalizes back to its resting conduction or valence band state, such that the direct recombination of the original electron–hole pair is effectively non-radiative. While Figure 7.2 shows an electron–electron–hole (eeh)

Conduction band

photon emission

photon emission

deep level

Valence band Bimolecular (Radiative)

Figure 7.2 ductor

Donor or Acceptor (Radiative)

Shockley-Reed-Hall (Non-Radiative)

Auger (Non-Radiative)

Illustration of various recombination pathways for charge carriers in a semicon-

276

Materials for Solid State Lighting and Displays

Energy, E

Conduction band

Energy, E

process, the counterpart hole–hole–electron (hhe) Auger process is also a possibility. Auger processes can involve phonons and/or crystal defects, which can increase their probability of occurrence due to relaxed constraints on momentum conservation. First principles calculations of coefficients due to these indirect Auger processes for (Ga, In)N are in the range necessary to explain LED efficiency “droop” [2, 3], as will be discussed later. In general, such processes may be speculated to be the reason behind relatively high Auger coefficients observed for wide-bandgap semiconductors for which direct momentum-matched Auger recombination is expected to be highly improbable. Another non-radiative process worth consideration, and not illustrated in Figure 7.2, is surface recombination. The optical properties of a semiconductor are governed by its bandgap, which is only in effect as long as the crystal atomic spacing remains extended and uninterrupted from the point of view of the charge carrier. Clearly, at the edges of the semiconductor, this is no longer the case. The surface dangling bonds at such interfaces represent, effectively, a very high density of trap states to which charge carriers can recombine non-radiatively. This surface recombination can thus be extremely significant, and is often quantified in terms of a surface recombination velocity [4]. As discussed in reference [3] of Chapter 1, the configuration of atoms in a unit cell of a semiconducting crystal results in energy dispersion diagrams that are either direct or indirect. Referring to Figure 7.3, direct bandgap materials allow electron–hole recombination at the lowest energy point in the band diagram with immediate momentum conservation. In contrast, indirect bandgap materials have a momentum offset between the lowest conduction band valley and the valence band, requiring conduction band electrons

(1) (2) Eg photon emission

Valence band

Momentum, k (a)

photon emission

(1) phonon- or (2) impurityassisted

Momentum, k (b)

Figure 7.3 Energy-momentum diagram for (a) a direct and (b) an indirect bandgap semiconductor. For a direct bandgap semiconductor, charge carriers may radiatively recombine quickly. In an indirect bandgdap semiconductor, momentum transfer from the crystal lattice, or via impurities, is required for radiative recombination, which slows down the process

Light Emitting Diode Materials and Devices

277

to interact with the crystal lattice (i.e., phonons) or defects in order to satisfy momentum conservation for the emission of light. Thus, the radiative recombination rates of direct bandgap semiconductors are much faster than for indirect bandgap materials, and the resulting radiative efficiencies are significantly higher. While some early LED materials systems, such as GaP:N, are indirect bandgap and in fact successfully commercialized (by employing impurity-assisted recombination), the vast majority of LEDs produced today are based on direct bandgap systems, predominantly (Al, Ga)InP for red to amber emitters, and (Ga, In)N for violet to blue to green emitters. 7.2.3

Heterojunctions

Considering the recombination processes described above and the charge-carrier interactions required, we can write a total recombination rate, R, for charge carriers in the active layer(s), namely: (7.1) R = An + Bnp + Cn n2 p + Cp np2 cm3 s−1 where n and p are the electron and hole concentrations, A is the Shockley–Read–Hall recombination coefficient, B is the bimolecular radiation coefficient, and Cn and Cp are Auger coefficients for the eeh and hhe processes, respectively. The radiative efficiency is then simply: (7.2) 𝜂rad = Bnp∕(An + Bnp + Cn n2 p + Cp np2 ) The recombination rate is associated with various decay times for carriers generated in the active region. It is sometimes helpful to think of the various processes in terms of these decay times, as they are directly measurable in experiments such as time-resolved photo- and electroluminescence. Assuming high-level injection, for which n ∼ p, the radiative and non-radiative lifetimes may be written: 𝜏r = (Bn)−1 𝜏nr = [A + (Cn + Cp )n2 ]−1

(7.3) (7.4)

and are related to the total carrier lifetime by: −1 𝜏 −1 = 𝜏r−1 + 𝜏nr

(7.5)

in which case the radiative efficiency may be written: −1 𝜂rad = 𝜏r−1 ∕(𝜏r−1 + 𝜏nr )

(7.6)

One can immediately see from Equation 7.2 that a strong balance between electron and hole concentrations is required in order for the radiative process to dominate and result in high light emission efficiency. This presents a significant problem in the case of the simplest LED structure, the p-n homojunction. As described in reference [3] of Chapter 1, electron and hole distribution overlap is governed by the balance of carrier drift and diffusion and varies widely throughout the depletion region. This makes it very hard to optimize the device for maximum and stable light output efficiency. The solution to this problem is the utilization of high-bandgap heterojunctions which allow the confinement of carriers to localized layers, thus increasing the electron–hole

278

Materials for Solid State Lighting and Displays

overlap. The exploitation of heterojunctions was a watershed event not just in LEDs but in semiconductor devices in general, and the Nobel Prize in Physics in 2000 was awarded to Zhores Alferov and Herbert Kroemer for its development [5]. While single-heterojunction (SH) structures were investigated initially, double-heterojunction (DH) LEDs had the advantage of confining both electrons and holes to the active light emission layer, resulting in better radiative efficiency. 7.2.4

Quantum Wells

While DH LEDs helped solve the problem of electron–hole overlap, they still had the drawback, like for homojunctions, that light generated would be reabsorbed within the active layer. Absorption coefficients for photons with energies near the bandgap are on the order of 105 cm−1 , corresponding to an absorption length of 0.1 μm. For active layers of similar thicknesses, unless the radiative efficiency is extremely high, “photon recycling” of this light is negligible and thus light reabsorption by the active layer is a significantly lossy process. In 1978, work on thin-active-layer laser diode structures resulted in DH devices with layers so thin that distinct electron levels within the active layer could be observed. These “quantum well” (QW) devices [6] provided a path to tuning the emission wavelength of LEDs by varying the QW thickness, rather than material composition, while minimizing the amount of reabsorbing material. In addition, provided carriers are injected efficiently into the QW, very high electron–hole overlap is achieved, providing for very high radiative efficiency. High power devices are realized by having many QWs as active layers, the so-called Multiple-Quantum-Well (MQW) structure, which is the basis for the vast majority of state-of-the-art LEDs today. Comparative representations of homojunction, DH, and MQW structures are shown in Figure 7.4. 7.2.5

Current Injection

While the MQW structure provides an efficient structure for light generation for carriers already in the QW layers, it does not itself guarantee high performance under electrical injection. Indeed, for optimum performance, electrons and holes must be injected from nand p-doped confining layers, and distributed across and into the QW layers, without being lost due to de-confinement under forward bias. If the potential barriers of the confining layers are too low, the injection efficiency into the active region can be compromised, as carriers can drift, thermalize, or tunnel through to the counterpart majority-carrier area, thereby recombining non-radiatively. This problem is usually associated with electrons, with their effective masses typically much lighter (and mobilities much higher) than that of holes. The various loss mechanisms that can result in lower than unity injection efficiency can be captured in the following relations: Jtotal = Jrec + Jleak

(7.7)

Jleak = Jtherm + Jdrift + Jtunn

(7.8)

𝜂inj = Jrec ∕Jtotal

(7.9)

where Jtotal is the total current density flowing in the device, Jrec is the total (both radiative and non-radiative) recombination current in the active region, and Jleak is the total leakage

Light Emitting Diode Materials and Devices

279

Energy

Conduction band

n

p

Valence band Depletion (active) region (a)

Energy

Confining layers

n

p

Active layer (n, p, or non-doped) (b)

Energy

Barriers

n

p

Quantum wells (c)

Figure 7.4 Cartoon illustrations of (a) homojunction, (b) double-heterojunction (DH), and (c) multiple-quantum-well (MQW) heterojunction LED structures. The DH structure is advantageous over the homojunction structure for its ability to confine charge carriers. The MQW structure is advantageous over the DH because it further improves carrier confinement, while also minimizing the required amount of active layer material, thus preventing light reabsorption by the active layer

current (i.e., current due to recombination outside of the active region or at electrodes). As mentioned above, Jleak can comprise different mechanisms, including thermionic emission (Jtherm ), and tunneling leakage (Jtunn ). The key to addressing all of these mechanisms, and thus achieving high injection efficiency, is to engineer high potential barriers to the injected carriers, which is driven by

280

Materials for Solid State Lighting and Displays

the bandgap of the confining layers and their doping levels. In fact, a common optimization problem for LEDs in industry is balancing the high doping required for high injection efficiency while avoiding high impurity (dopant) concentrations near the active regions, which can negatively affect the radiative efficiency of the device. With injection efficiency defined, the internal quantum efficiency of the LED may be described: (7.10) 𝜂int = 𝜂inj 𝜂rad While formal treatments for current injection are available theoretically [2], in practice it is extremely difficult to measure leakage current at all, much less the detailed mechanisms behind it. One approach that is sometimes employed is to compare photo-injected vs. electrically injected luminescence efficiency for an LED structure. While such measurements can be difficult to carry out, if performed correctly, they can give an assessment of the injection efficiency of an LED. The goal of course is to achieve near unity injection efficiency, in which case simply 𝜂int ∼ 𝜂rad . 7.2.6

Forward voltage

As discussed in detail in reference [3] of Chapter 1, the forward voltage for the ideal diode turns on near the bandgap voltage with current growing exponentially with further increasing voltage. In reality, a practical LED includes resistive layers and electrodes, and potential barriers at various interfaces, such that it is most convenient to describe the LED forward voltage in an equivalent circuit model, so that: Vf = V0 + 𝜌 Jf p

V0 = Efn − Ef + ΔV

(7.11) (7.12)

where V0 is the “turn on voltage” which is the ideal Fermi level separation plus any interfacial potential barriers (denoted ΔV), 𝜌 is the device resistivity (including that of all layers and the Ohmic contacts), and Jf is the forward current density. Equation 7.11 is most useful for the LED under full operation. At low currents, the LED “turn-on” characteristic is governed by the Shockley diode equation including an ideality factor (as discussed in reference [3] of Chapter 1).

7.3

Material Systems

The material systems that dominate the LED world today are all III–V compound semiconductor alloy systems. A chart showing the lowest energy bandgaps vs. growth-plane atomic lattice constant for these materials is shown in Figure 7.5. 7.3.1

Ga(As,P)

The first demonstration of visible light emission from a practical LED was achieved by Nick Holonyak Jr, while at GE Research Labs in Syracuse, New York, in 1962. This was demonstrated by taking an “alloy engineering” approach to mixing GaAs and GaP using

Light Emitting Diode Materials and Devices

281

Figure 7.5 Lowest energy bandgap vs. crystal in-plane atomic lattice spacing for Wurtzite (Ga,In)N and cubic (Al,Ga)As and (Al,Ga)InP material systems. The latter two are latticed matched to GaAs, and provide direct bandgap emission for near-infrared to red – (Al,Ga)As – and red to amber – (Al,Ga)InP. (Ga,In)N is grown lattice-locked to GaN, and undergoes strong compressive stress as InN mole fraction is increased. Practical performance for (Ga,In)N is achieved from the UV-A to the green

VPE in order to blue-shift from the emission of direct-bandgap GaAs into the visible red wavelength regime. The approach, considered very controversial at the time, was ultimately successful and simultaneously demonstrated the first visible-spectrum laser diode [1]. Ga(As,P) laid the foundation for a practical LED manufacturing process and led to commercially successful red-emitting LEDs (∼630–650 nm) in applications such as handheld calculators, watches, and instrument indicators. 7.3.2

Ga(As,P):N

Shifting Ga(As,P) further into the visible became problematic as the indirect–direct transition was approached, at about 45 mol% GaP. A solution to this problem was achieved by introducing nitrogen, an isoelectronic impurity, into the Ga(As,P) lattice. The very small atomic radius of N compared with the surrounding atoms led to very localized carriers in its vicinity. These highly localized carriers were associated with much higher uncertainty in their momentum, in line with Heisenberg’s uncertainty principle of quantum theory. This allowed carriers to achieve conservation of momentum and recombine radiatively with substantially higher probability than would be the case without nitrogen doping [7].

282

Materials for Solid State Lighting and Displays

This approach surpassed a competitive impurity-based approach using GaP:Zn,O [8] and led not only to yellow- but also green-emitting LEDs for the first time [9]. 7.3.3

(Al,Ga)As

The next step in alloy engineering focused on the mixed-column-III, as opposed to column-V, approach. Aluminum atoms have a very similar size to those of gallium and in the ternary (Al,Ga)As system the two elements are virtually interchangeable, without significantly impacting the crystal lattice parameter. Based on low-defect-density GaAs substrates, this approach provided for a lattice-matched alloy system with direct bandgap emission from the infrared (GaAs, 870 nm) to red (Al0.45 Ga0.55 As, 624 nm), providing a platform for much higher crystal quality than what was possible with the previous mixed-column-V approaches. Based first on LPE, (Al,Ga)As red-emitting LEDs were developed for applications such as traffic signals and rear automotive lighting, and with reasonably high efficiency [10]. Even transparent-substrate (TS) (Al,Ga)As devices were made by transitioning from GaAs to (Al,Ga)As during LPE and then eventually removing the starting GaAs substrate. (Al,Ga)As devices were also successfully demonstrated using MOCVD, an epitaxial growth platform with far more precision and control than standard VPE and LPE processes [11]. Unfortunately, the high aluminum content meant these materials were very susceptible to hydrolysis, and their track record in demanding environments such as signaling and automotive left something to be desired. With all the inherent technological advantages of (Al,Ga)As, this limitation, along with the lack of path for shorter wavelength devices, set the stage for a new entry into the LED field. 7.3.4

(Al,Ga)InP

Gallium indium phosphide can be grown lattice-matched to GaAs at a GaP molar fraction of 52%, that is, Ga0.52 In0.48 P. Just as for (Al,Ga)As, aluminum–gallium substitution is allowed, providing a lattice-matched (to GaAs), direct-bandgap material system from the red (Ga0.52 In0.48 P, 650 nm) to yellow ((Al0.53 Ga0.47 )0.52 In0.48 P, 555 nm). While considerably more challenging to engineer than (Al,Ga)As, this quaternary system provided much lower Al content and thus higher reliability devices, as well as practical light emission from 650 to about 580 nm [12, 13]. With all its inherent advantages, and the availability of MOCVD technology to control the complex deposition process, considerable engineering effort was put into advancing this system. The first area of focus was increasing current spreading and light out-coupling by use of a thick, transparent “window” layer [typically GaP or (Al,Ga)As] on top of the thin epi structure [14]. This allowed current injected from an electrode at the top of the device to spread throughout the active region to maximize light generation, and in addition, out-couple more light by utilizing the side surfaces of the window layer in addition to the top surface of the device. Another area of focus was removal of the (absorbing) GaAs substrate. Two basic approaches were adopted. In one, led by Hewlett-Packard Company in San Jose, California, the GaAs substrate was removed and the remaining LED and window layers were bonded to a GaP substrate, which is transparent to light at wavelengths greater than 550 nm. The wafer-bonding process was at elevated temperature and pressure and provided

Light Emitting Diode Materials and Devices

283

for a conductive interface between the GaP substrate and LED layers; effectively, the GaP substrate replaced the GaAs [15]. This TS platform for (Al,Ga)InP provided means for advanced chip architectures, including chip shaping to increase light extraction, which led to above 100 lm/W LED performance for the first time, in 1999 [16]. The other approach to dealing with the absorbing GaAs substrate was to remove it and bond the remaining structure to a reflective, metallic, bond layer applied to another substrate (e.g., GaAs or Ge) [17]. This vertical thin-film (TF) approach was pursued in particular by Osram Opto Semiconductors, of Regensburg, Germany, who used it to compete with the TS approach. Today, both approaches are still used commercially. 7.3.5

(Ga,In)N

Success in GaAs- and GaP-based semiconductor devices provoked the question about gallium nitride (GaN), an even higher bandgap material system for ultraviolet (UV) and visible-spectrum devices. However, GaN was peculiar to its predecessors for at least two reasons: (1) early demonstrations of its synthesis yielded a Wurtzite (i.e., assymetric) crystal, instead of the familiar (and symmetric) cubic structures of Si, Ge, GaAs, and GaP; and (2) because of the very high vapor pressure of nitrogen in molten GaN, there was no straightforward path to synthesize substrates of the material using conventional crystal growth techniques. The first GaN layers were synthesized by vapor deposition on sapphire substrates in 1969 [18], and exhibited very high crystal dislocation densities due to the crystal lattice mismatch between GaN and sapphire. p-Type material (using Mg as an acceptor dopant) was attempted but was impractical, and exploration of metal–insulator–semiconductor (MIS) structures did not result in meaningful performance [19]. It was not until the late 1980s and early 1990s that GaN materials development would make a first impact on the semiconductor industry. Two key developments, both originating in Japan, would usher this era in. The first, demonstrated at Nagoya University by Isamu Akasaki and Hiroshi Amano, was the use of a very low growth temperature buffer layer for synthesizing high quality single-phase GaN on sapphire substrates by MOCVD [20]. The second was a practical method for forming p-type GaN using Mg as an acceptor dopant and properly activating the Mg by driving off hydrogen incorporated into crystal during growth with a thermal annealing treatment [21]. With reasonable quality GaN and p-type capability demonstrated, the remaining issue was the very high dislocation density in the material due to lack of an available high quality, native substrate. Rather surprisingly, it was Mother Nature who came to the rescue here. Contrary to other III–V materials, it was found that the luminescence efficiency of (Ga,In)N was far less sensitive to dislocation density than other material systems [22]. The difference is nothing less than amazing and is illustrated in Figure 7.6 in terms of relative internal quantum efficiency against etch pit density (as a proxy for threading dislocation density, TDD). Leveraging this rather remarkable attribute of (Ga,In)N, it was only a matter of time that the techniques of alloy and heterostructure engineering would be applied and unlock the potential of this material system. Indeed, in rapid succession in the 1990s, GaN-based LEDs were demonstrated emitting over a wide range of wavelengths, from violet to red, and also laser diodes for the shorter emission wavelengths, which became the basis for Blu-ray™ optical storage [23–25].

284

Materials for Solid State Lighting and Displays 1 (Ga,In)N GaP

0.9 0.8

GaP

Normalized efficiency

0.7

GaP

GaAs 0.6

GaP

0.5 Ga(As,P):N

0.4 0.3 (AI,Ga)As 0.2 0.1 0 1. E+03

Ga(As,P):N 1. E+04

1. E+05 Etch pit density

1. E+06

1. E+07

1. E+08

(cm–2)

Figure 7.6 Normalized photoluminescence efficiency vs. etch pit density for several III–V materials. (Ga,In)N is an outlier, maintaining relatively high efficiency for defect densities that would quench emission in other materials

The rapid progress of (Ga,In)N optoelectronic devices was surprising to many since the ranges of TDD were in a regime for which conventional III–V semiconductors could not support device performance. Indeed, (Al,Ga)As luminescence was known to be almost completely quenched at a TDD of 1 × 106 cm−2 or more, whereas good performance for (Ga,In)N was observed for TDDs in the 1 × 108 –1 × 109 cm−2 regime. Many theories have been put forward to justify the difference. One that has some supporting experimental evidence is that incorporation of relatively large In atoms into the lattice provides for localization of carriers which could effectively reduce carrier diffusion lengths [26]. A challenge that arose for LEDs in the (Ga,In)N system was the very low conductivity of p-type layers due to the high activation energy (∼250 meV) for Mg acceptors in GaN. This high resistivity, coupled with the lack of a clear path for growing thick p-type GaN layers, led the industry to a novel solution for current spreading: the use of a semi-transparent Ohmic contact layer. This layer, typically thin NiAu, was deposited on the p+ GaN surface above substantially all the active region area, and electrically coupled to a wirebond pad at the corner of the device. This allowed effective current spreading in a simple structure, although it did have the downside of considerable optical loss and presented an unhappy trade-off in the case of high current density operation, for which thicker NiAu layers would be necessary. Eventually NiAu was replaced by indium–tin–oxide (ITO) contacts, which are substantially more transparent, and commonly used today. For high power operation, the constraints of the “epi-up” configuration led the industry to consider alternatives. The flip-chip configuration, driven mainly by Lumileds Lighting of San Jose, California, a joint venture between Agilent Technogies (formerly

Light Emitting Diode Materials and Devices

285

Hewlett-Packard) and Royal Philips Electronics, allowed the replacement of the semi-transparent contact with a thick, reflective contact, typically Ag-based [27]. Light was out-coupled predominantly through the sapphire substrate, which was now on “top” of the device. This had the advantage of decoupling power density from light extraction efficiency, since the reflective Ohmic contact could be as thick as necessary for current spreading. Furthermore, the thin p-layers typical for GaN provided for the exploitation of optical microcavity effects due to the close proximity of the QW-localized light generation region to the highly reflective p-contact. Indeed, active region design and p-layer thickness tuning became critical for optimized performance in the flip-chip configuration [28]. One remaining issue in the configurations above was light trapping that occurred within the (Ga,In)N epitaxial layers. These layers have a refractive index of about 2.4, much higher than that of sapphire (n ∼ 1.77). The result is that about one-third of the light generated in the active region is trapped in the epitaxial layers, and is very susceptible to reabsorption within the Ohmic contacts, electrodes, or even within the epitaxial layers themselves. For most epi-up configurations, the solution to this problem was the engineering of micrometer-scale etched surfaces into the sapphire growth substrate. This patterned-sapphire-substrate (PSS) approach helped scatter light out of the epitaxial layers and improved the light extraction efficiency of the device [29]. For the flip-chip configuration, it was possible to completely remove the sapphire substrate, by an excimer laser lift off technique, and texture its surface [30]. The same lift off approach was successfully applied to TF structures for (Ga,In)N, an approach adopted by Osram Opto Semiconductors [31]. While improved buffer layer growth techniques could reduce TDD to about 1 × 108 cm−2 for GaN on sapphire, this density was not low enough for reliable operation of laser diodes. Noting that dislocation density reduced as GaN layer thickness increased, several entities in Japan developed high growth rate hydride-vapor-phase-epitaxy (HVPE) of GaN to grow layers several millimeters thick [32]. The resulting “boules” could be used to manufacture bulk GaN substrates, with TDD of about 1 × 106 cm−2 and low enough for sufficient laser diode reliability to support the Blu-ray industry. Even though they were about a hundred times more expensive than sapphire, these substrates were eventually commercialized even for LEDs. Pioneered by Soraa, Inc., in Silicon Valley, California, this technological approach produced very efficient violet-emitters for general lighting by following the TS shaped-chip path previously paved in the development of (Al,Ga)InP LEDs [33]. Advantages provided by the bulk GaN substrate were not only a lower TDD for the active region, but also very high thermal and electrical conductivities, which supported high power density operation (about 10 times higher than GaN-on-sapphire devices) and justified the cost of the very expensive substrates [34]. An overview of GaN-based LED history and technology is provided in a review article co-written by the present author and Nobel Prize winner Professor Shuji Nakamura [35]. 7.3.6

White Light Generation

Before one can discuss the details of white light generation it is important to establish what exactly is “white” light. The Commission Internationale de l’Éclairage (CIE) has developed standard reference illuminants for this purpose that serve as guideposts. For example, illuminant CIE-A represents that of a blackbody radiator at a temperature of 2856 K, and can

286

Materials for Solid State Lighting and Displays

be used as a model for radiation from incandescence of heated tungsten wire, for example. Illuminant D65, on the other hand, is designed to model direct radiation from the sun at the earth near noon in the northern hemisphere. The name refers to the temperature required for a blackbody radiator to achieve the same color point, or chromaticity, in this case 6504 K, also known as the correlated color temperature (CCT) for color points off the blackbody curve. Such “warm” and “cool” white light reference illuminants guide LED manufactures in the types of products they should aim for. The spectra for CIE-A and D65 are shown in Figure 7.7. It is clear from these that they are broadband sources that cannot be mimicked by a singular LED type. Instead, multiple emission sources must be combined. There are basically two ways to do this: (1) a multi-primary approach, which involve multiple LEDs of different emission colors (e.g., red–green–blue) that are configured in balanced operation to maintain a white color point; or (2) a down-conversion approach, wherein light from a single primary LED is shifted to (and, most commonly, mixed with) longer wavelength emission from luminescent materials in order to achieve broadband white light. As will be discussed later in this chapter, today there are considerable performance limitations for direct green-emitting LEDs, and also for red-emitting LEDs at typical operating temperatures. So, while on paper the multi-primary approach should ultimately be the most efficient, since it avoids the fundamental Stokes’ loss of down-conversion processes (i.e., converting a higher-energy photon into a lower-energy one), currently it is unattractive except for special applications. Accordingly, the down-conversion approach

CIE-A (2856 K)

D65 (6504 K)

350

400

450

500

550 600 Wavelength (nm)

650

700

750

Figure 7.7 Equal-lumen emission spectra for two “white” reference illuminants: CIE-A, which approximates tungsten incandescence at a color temperature of 2856 K; and D-65, which approximates typical noon-day solar radiation at the earth, at a color temperature of 6504 K

Light Emitting Diode Materials and Devices

287

is by far the most common method of producing white light, mostly using blue- or violet-emitting primary LEDs in combination with one or more phosphor materials. Various phosphor materials have been described in detail elsewhere in this book (Chapters 3 and 4). Here, we will focus on the most common configurations for use in general lighting. This simplest method to produce white light from LEDs is to mix a blue-emitting primary LED with a broadband, yellow-emitting Y3 Al5 O12 ∶Ce3+ , or “YAG”, phosphor, a well-known garnet structure for efficient luminescence when excited, or “pumped”, with blue light [36]. The mix of leaked blue light plus the yellow phosphor emission is engineered to achieve a white point close to the blackbody loci on the CIE chromaticity diagram. This approach can achieve CCTs between 4000 and 7000 K with color rendering indices in the 65–80 regime, useful for certain lighting application such as outdoor, automotive forward lighting, portable (flashlights), and low-color-gamut liquid crystal display (LCD) screens. Warmer CCTs, and higher color rendering, are achieve by the addition of a red phosphor to the mix, the most common of which is CaAlSiN3 ∶Eu2+ and its variants [37]. Especially in combination with “LuAG”, which is a blue-shifted version of YAG made possible by Lu/Y substitution, very high color rendering (or color gamut, for displays) is achievable, even up to a color rendering index (CRI, or Ra ) of 98. As reference points, incandescent lamps achieve a CRI of ∼100, “deluxe” quality fluorescent lamps a CRI of 90, and standard fluorescent lamps have a CRI of 80. In the USA, the Environmental Protection Agency’s ENERGY STAR® certification requires a minimum CRI of 80 [38]. In addition to powder forms, phosphors have been incorporated into solid-state forms, such as phosphor grains embedded in glass [39] and even pure ceramic forms [40]. These solid-state down-converters have the advantage of higher light density and better thermal properties than powder-based phosphors, but are less flexible in terms of color choice and are generally more expensive [41]. Their high brightness aspects have made them useful in automotive forward lighting. An amber-emitting ceramic, (Ba, Sr)2 Si5 N8 ∶Eu2+ , has been developed and commercialized and now outperforms (Al,Ga)InP LEDs of the same dominant wavelength of emission [42]. One element missing from the common blue-emitting LED approach to white light is that very little to no violet or UV light is emitted. This is meaningful because such short wavelength light can be used to excite optical brightening agents (OBAs) found in many manufactured products (e.g., paper, clothing, cosmetics) and also in nature (e.g., human teeth and eyes). These OBAs absorb short wavelength violet and, if available, UV light and re-emit in the blue, causing a perceived enhanced brightness and “whiteness” in these materials, an effect readily provided by sunlight and incandescence. This effect is critical for the appearance of certain materials—bridal gowns, for instance, without which would appear flat white or even have a “cream” color appearance. Psychophysical experiments employing LED light sources with and without OBA excitation showed that subjects dramatically preferred the light sources that activated OBAs, even ranking their own personal appearance higher when illuminated by such light [43]. While there is yet no common measure or standard for OBA excitation from light sources, there are LED products on the market that provide this effect, either using violet-based primary emission, or a mix of violet- and blue-emitting LEDs, in combination with one or more phosphors. There is considerable activity in the development of new down-converting materials for white-emitting LEDs. The solutions described above are based on very, even too, broad emission spectra, especially when it comes to red-emitting phosphors. Narrower band red

288

Materials for Solid State Lighting and Displays

emission could be exploited to provide reasonable color rendering for warm white LEDs but with significantly higher lumen equivalent of radiation (LER). Recently, General Electric announced a practical version of a Mn4+ -based phosphor, a line-emitter at about 631 nm, that is designed to do just this [44]. Replacing the broadband CaAlSN3 ∶Eu2+ with the narrower emitter increases LER by about 20–30% while maintaining a CRI of 90 and a high R9 (test color for deep red). Improvements have also been made in narrowing Eu2+ emitters, as demonstrated by Lumileds’s development of Sr[LiAl3 N4 ]∶Eu [45]. Such phosphor developments will continue and serve to increase the efficacy of LEDs in general lighting applications, but also for dislpays in which narrower band red-emitters can increase color gamut and are better matched to the color filters employed in LCD panels [46]. In addition, semiconductor nanoparticles, or “quantum dots” (QDs), are a potential solution for narrow band, even tunable down-conversion emitters. QDs are achieving commercial success in LCD backlighting, but not yet in general lighting applications, which generally have more stringent operating conditions in terms of power density and temperature. (QDs are described in much more detail in Chapter 2.)

7.4 7.4.1

Packaging Technologies Low Power

For the first decades after the invention of the LED, its primary function was that of an indicator. As such, it only needed to be able to produce enough light that it could be visible when directly viewed by the human eye. The primary metric was brightness, measured by intensity in lumens pre steradian, or candela (cd, or mcd), and the goal was to produce the required mcd numbers with the lowest cost device. The typical operating current was 20 mA or less. The result of this approach, from a packaging standpoint, was a very low cost lead-frame-based approach. For example, a very common approach was Ni-plated steel lead frames, wherein the LED chip was conductive-epoxy-attached to one lead, usually formed into a “reflector cup” shape. A wirebond connection was made between the LED chip and the other lead. Then, the leads and chips were inserted into a mold filled with epoxy resin which served the dual purpose of mechanically securing the assembly while also focusing light emitted from the LED in the reflector cup to increase on-axis intensity. This was accomplished by shaping the epoxy in a proper lens shape, as shown in Figure 7.8 for the so-called “T-1 3∕4” LED. 7.4.2

Mid Power

Improvements to the efficacies of LEDs by utilizing the (Al,Ga)InP and (Ga,In)N material systems made it possible for LEDs to be used as more than just indicators. Now, they could be used in “power signaling” applications, such as traffic signals and automotive rear lighting. In order to produce enough light for these applications, LEDs had to be run at 50–150 mA, significantly higher than for low power indicator LEDs. In order for these “mid power” devices to operate efficiently and reliably, it was important for the package to be redesigned to remove waste heat from the LED chip, which was also made larger than for low power LEDs. The approach taken was to use a Cu-based lead frame and increase its volume and thickness. This enabled the junction-to-case thermal resistance to reduce from ∼200 K/W, typical for low power LEDs, to about 100 K/W [47].

Light Emitting Diode Materials and Devices

LED Package Type

Historical

Low-Power < ¼ Watt (5–50 mA)

“ T1 ¾ ” circa 1970+

Mid-Power – ½ Watt (50–150 mA)

289

circa 2016

4.0 mm x 1.0 mm = “4010”

~¼

3.0 mm x 3.0 mm = “3030”

circa 1990+

High-Power ~ 1–5 Watts (0.35 -1.5 A) circa 2000+

3.0 mm x 5.0 mm = “3050”

Figure 7.8 Historical (center) and present-day (right, circa 2016) low-power, mid-power, and high-power packages typical for LEDs

Surface mount versions of mid power LEDs were also developed and in fact, by volume, represent the largest population of LED-type sold for general lighting today. The lead frames are designed for eventual solder reflow, and are encased in an epoxy or silicone molding compound, usually loaded with filler to provide a low-optical-loss white body which includes a cavity for containing the LED chip, phosphor (if used), and lens silicone material. The same basic approach is used for low power LEDs today, with dramatically smaller form factor compared with their historical counterparts, as shown in Figure 7.8. 7.4.3

High Power

Continued LED efficacy improvement eventually made Watt-class devices viable. First pioneered by Hewlett-Packard Company in the late 1990s, these “high power” devices required a complete chip and package redesign compared with low and mid power devices [40]. In particular, initially heatsinking was provided by a “slug” of Cu including a reflector cup shape that was electrically isolated from the lead frame. To accommodate thermal expansion mismatch between the larger (typically 0.5–1.0 mm or more wide) LED chip and the copper slug, an intermediary submount (e.g., Si) was provided, and the LED chip was solder-attached to the submount, rather than using epoxy. This configuration dropped the junction-to-case thermal resistance substantially, down to around 10 K∕W [41], and allowed drive currents up to 1.0 A (∼3 W input power) and more.

290

Materials for Solid State Lighting and Displays

In the last decade, a ceramic-based platform for high power LEDs swept the industry and replaced the previous lead-frame approach [48]. In this approach, a single- or multi-layer ceramic, typically alumina, is bonded to or plated with Cu to provide bond pads on the top surface and solder pads on the back surface, using through-hole (Cu-filled) vias. An electrically isolated Cu thermal pad is sometimes provided. The ceramic approach provided most of the heatsinking capability of the “thermal slug” approach but in a much thinner and more cost-effective package. Lens action could be provided by molding a silicone dome lens around the LED chip, rather than employing a reflector cup. For very high power LEDs, the choice of ceramic material moved to aluminum nitride, rather than alumina. This approach has been successfully applied to automotive forward lighting applications. Recently, the ceramic approach has also been applied to certain mid power LEDs. 7.4.4

Chip-On-Board LEDs

While the high power packages are able to handle a few watts of power and thus hundreds of lumens of light output, there are many applications where thousands of lumens are required from a single source. A common solution here is to use a chip-on-board (COB)-style LED. In this configuration, multiple LED chips are mounted onto a package substrate, typically metal-core board or ceramic, with electrical traces to maintain the LEDs in a desired series/parallel configuration to target a certain operating voltage. The ceramic material is typically made white (e.g., alumina) and the traces are often Ag-plated in order to reduce optical loss. A containment ring or “dam”, typically of titania-loaded silicone or similar white material, surrounds the LEDs and provides a containment structure for silicone encapsulation and phosphor material, which is dispensed over the LED chips and results in an “egg yolk” appearance. COB LEDs often have through-holes for convenient bolting onto a heatsink platform. In some products, even electrical connectors are provided directly on the COB to facilitate quick “plug and play” electrical hook up. 7.4.5

Multi-Color LEDs

Both lead-frame- and ceramic-based packages can lend themselves to incorporating multiple LEDs with separate electrical connections, for color tuning capability. A challenge in this configuration is “mixing” of the multi-colored light to provide a color-uniform light source. Strong diffusive elements proximal to the LED chips can increase the optical source size, reducing brightness, and also optical losses by back-scattering light towards the LED chips and package materials. It is more common to mix the light downstream, in secondary optics, or even at the target. In the latter case, application use conditions should be carefully studied to make sure that “color shadows” do not make themselves appear in an unacceptable way. Another challenge with multi-color LEDs is maintaining stable color point over operating conditions (drive current, temperature) and over lifetime. For guaranteed long-term consistency, some form of optical feedback may be necessary. 7.4.6

Electrostatic Discharge Protection

As with many electrical components, high yield manufacturing requires protection against electrostatic discharge (ESD). While proper handling practices are sufficient for many components, more direct protection is warranted especially for (Ga,In)N LEDs, which are

Light Emitting Diode Materials and Devices

291

particularly susceptible to ESD damage, presumably due to higher TDD compared with conventional LED materials. Standard practice is to electrically connect a separate discrete avalanche diode in reverse-parallel with the LED chip. A typical target protection level is 2 kV Human Body Model (HBM) and 1 kV Machine Model (MM). The ESD protection or “transient-violet suppression” (TVS) diode is most often mounted directly into the LED package, which presents a nuisance since it can affect the optical performance of the LED. Embedded protection diodes are available but not as cost-effective today as using discrete devices. In COB LEDs, ESD protection components are typically mounted outside the containment ring for the LEDs, out of the optical pathway of the device.

7.5 7.5.1

Performance Light Extraction Efficiency

All of the LED active region materials have very high optical refractive indices, ranging from 2.4 (GaN) to 3.3 (GaP). The very large difference in optical density between these materials at the target exit medium (i.e., air) presents a problem. Electromagnetic boundary conditions require phase-matching at the semiconductor/air interface which results in a large fraction of light generated within the semiconductor remaining trapped due to total internal reflection (TIR). The critical angle of incidence for TIR is: 𝜃c = sin−1 (na ∕ns )

(7.13)

where ns and na are the refractive indices of the semiconductor and target ambient, respectively. Assuming isotropic light generation within the semiconductor, the amount of light transmitted through to ambient is (1-cos𝜃 c )∕2, which is just the solid angle fraction of light that is not reflected due to TIR. This percentage can be extremely small. For example, for the case of GaP to air in the situation just described, the amount of light reflected by TIR is 97.6%! And, an additional portion of the non-TIR light will not be transmitted due to Fresnel reflection, which at its minimum is (ns − na )2 ∕(ns + na )2 ∼ 29% in the case described. This means less than 2% of isotropic light generated within GaP would transmit through a single GaP/air interface upon first pass. Significant effort has been expended by the LED industry to overcome this challenge, especially within the last two decades. For (Al,Ga)InP, first was the removal of the absorbing GaAs substrate and replacing it with either a transparent [10] or reflective [12] substrate, as described earlier. Surface texturing [24, 25] and/or chip shaping [11, 26] were also found to help improve light extraction significantly. (Ga,In)N devices on sapphire enjoyed a transparent substrate but still had light trapping at the GaN/sapphire interface as described earlier. The solution to this was patterning [23, 49] or removing the sapphire substrate. The employment of the native GaN substrate made TS devices possible for (Ga,In)N which allowed chip shaping, and represent the highest light extraction efficiencies for visible spectrum LEDs claimed to date [26]. Light extraction efficiency is not directly measurable, but can be estimated by combining certain measurements with simulations [50]. A summary of the evolution of light extraction efficiencies for (Al,Ga)InP and (Ga,In)N LEDs is shown in Figure 7.9 [28, 49–51], and some selected LED architectures are illustrated in Figure 7.10 [16, 30, 33, 49]. Best performance is achieved for devices encapsulated with a silicone lens

292

Materials for Solid State Lighting and Displays 100% Shaped TS

90% ITO-PSS TF-FC

Light extraction efficiency

80% TF 70%

Improved TF Shaped TS

60%

TF

50%

FC(Ag) NiAu-PSS

FC(Al) Improved TS

40% NiAu

30%

TS

20%

(Al,Ga)InP

Thick AS

10%

(Ga,In)N

Thin AS 0% 1990 AS

1995

2000

Absorbing Substrate

2005

2010

FC

Flip-Chip

2015

2020

TS

Transparent Substrate

TF

Thin-Film

Al

Aluminum

ITO

Indium-Tin-Oxide

Ag

Silver

PSS

Patterned Sapphire Substrate

Figure 7.9 Evolution of LED chip light extraction efficiency over the last 25 years for both (Al,Ga)InP and (Ga,In)N LEDs. These data are for chips that are emitting into an encapsulating medium, for example, epoxy or silicone

material, since it is easier to out-couple light into a medium of refractive index of 1.4–1.5 than directly to air. Certain applications prevent or discourage the use of such lenses, in which cases the LED performance will suffer a bit due to the loss of “encapsulation gain” which is on the order of +20% for state-of-the-art (Ga,In)N devices [44]. 7.5.2

Monochromatic Performance

The radiometric light output of an LED in optical watts is: Po = If 𝜂ext Ep

(7.14)

where If is the forward current in amperes, Ep is the average photon energy (in eV), and 𝜂ext is the external quantum efficiency, given by: 𝜂ext = 𝜂int Cext Cpkg

(7.15)

where Cext is the light extraction efficiency of photons from the LED chip into the package, Cpkg represents the extraction of light out of the package, and 𝜂int is the internal quantum efficiency, defined earlier. The power-conversion efficiency (sometimes called “wall-plug efficiency”, a misnomer) is then: 𝜂 = Po ∕(If Vf ) = 𝜂ext Ep ∕Vf ∼ 𝜂ext (1240∕𝜆p )∕Vf

(7.16)

The ratio Ep ∕Vf is sometimes referred to as “electrical efficiency”, although technically this quantity can be greater than unity.

Light Emitting Diode Materials and Devices

293

(a)

(b)

(c)

(d)

Figure 7.10 Selected chip architectures that have been successful in achieving high light extraction efficiencies: (a) shaped transparent-substrate (TS) (Al,Ga)InP LED; (b) patterned sapphire substrate (Ga,In)N LED with indium–tin–oxide (ITO) contacts; (c) thin-film flip-chip (Ga,In)N LED; and (d) shaped TS (Ga,In)N LED. Source: Krames et al. 1999 [16], Shchekin et al. 2006 [30], Narukawa et al. 2010 [49] and Hurni et al. 2015 [33]. Reproduced with permission of AIP Publishing LLC

294

Materials for Solid State Lighting and Displays

A summary of the best reported external quantum efficiencies for (Al,Ga)InP and (Ga,In)N LEDs, as a function of peak emission wavelength, is shown in Figure 7.11 [27, 43, 51–55; P. Deb, personal communication, 2015]. These data are for reasonably high current density operation (35 A/cm2 or higher) and at room temperature. For (Al,Ga)InP, decent high external quantum efficiencies (above 50%) are observed for red wavelengths (∼625 nm) when the AlInP mole fraction in the active region is small, or zero. As more Al is incorporated, which is required to raise the bandgap energy and shorten the emission wavelengths, the system moves towards a direct–indirect bandgap transition with a Γ-X crossover at about 45% AlInP. This means that the indirect X valley begins to hold more and more carriers as one tries to shorten the emission wavelength (the problem is even further exacerbated by temperature). Another challenge for (Al,Ga)InP LEDs is that the potential barrier to electrons from available confining layers (e.g., AlInP) are relatively small, so that electron leakage can be a problem. This risk is also increased by increasing AlInP mole fraction in the active region. The sum of these effects is that (Al,Ga)InP LED performance drops precipitously from the red toward the shorter wavelengths, with the limit in commercially viable performance at about 585 nm, in the “amber” color regime. For (Ga,In)N, the efficiency vs. wavelength curve is almost a mirror image of that for (Al,Ga)InP. When then InN mole fraction is relatively low, that is, for violet to blue emission, the performance can be extremely high, nearing 80% external quantum efficiency at

80% ≥ 35 A/cm2

External quantum efficiency

70%

T = 300k

60% 50% 40%

(Al,Ga)InP

PC

(Ga,In)N 30% 20% V(λ) 10% 0% 400

450

500

550

600

650

700

Peak wavelength (nm) Figure 7.11 Best-reported LED external quantum efficiencies as a function of peak emission wavelength, at current densities of 35 A/cm2 or greater, and at room temperature. In addition to primary (Ga,In)N and (Al,Ga)InP LED performance, full-conversion phosphor-converted (PC) LED performance for green and amber emission are indicated

Light Emitting Diode Materials and Devices

295

operating currents and with peak values (at lower current densities) even higher. Estimates for the internal quantum efficiencies for these high performing LEDs is 90% or more [26]. With increasing InN mole fraction in the active region, however, performance deteriorates. In the green, at 535 nm, the best reported value is 35% [49], less than half that of the best violet- or blue-emitting LEDs. At longer wavelengths, the performance continues to drop, and commercial viability is lost at about 540 nm. Since Wurtzite (Ga,In)N is a complete direct bandgap system, with nearest indirect valleys several hundred meV above the Γ point, it is difficult to blame the deteriorating performance on bandstructure as for (Al,Ga)InP. Unlike that material system, however, (Ga,In)N grown on GaN is not a lattice matched system. As InN mole fraction is increased, compressive stress builds up in the high In containing layers, and is known to result in crystallographic break-down if the total InN content exceeds a certain value [56]. In addition, increased strain raises the built-in electric fields present in (Ga,In)N heterostructures which can serve to separate electrons and holes (thus reducing radiative rate) and potentially complicate carrier transport [57]. This effect is strong, and red-shifts of emission wavelength with increasing quantum well thickness are readily observed in (Ga,In)N LEDs, often referred to as the quantum confined Stark effect (QCSE) [58]. In addition, recent studies of high InN mole fraction layers suggest non-uniform distribution of In atoms can lead to “lumpy” transport and are a possible reason that longer wavelength (Ga,In)N LEDs have larger forward voltages, despite having a smaller bandgap [59]. Considering multiple (n) sources and a target spectrum (e.g., white light), the overall efficiency can be shown to be: )−1 (∑n fi ∕𝜂i (7.17) 𝜂= i=0

where fi is the optical fraction of light for the ith source, that is: ) (∑n fi = Po,i ∕ Po,j j=0

and

∑n

f j=0 i

=1

(7.18)

(7.19)

and 𝜂i is the power conversion efficiency of the ith source. From Equation 7.17, it is seen that, for a target spectrum, the lowest efficiency device dominates the performance of the system. This becomes a critical issue in multi-primary LED light sources. In most applications for visible-spectrum LEDs, it is the photometric (or, human-eye response) that matters. The CIE has defined a luminosity function, V(𝜆), representative of the human eye’s responsiveness to light of different wavelengths, according to the sensitivity of cones in the retina and other details. It is a response curve with its peak at 555 nm and is shown in Figure 7.11. The luminosity function is used to define the lumen, for which 1 W of optical power at 555 nm is defined to be 683 lm. The emission spectrum of an LED can be used to determine its lumen-equivalent-of-radiation (LER): LER = 683

∫

S(𝜆) V(𝜆) d𝜆∕ S(𝜆) d𝜆 ∫

(lumens per optical watt)

where S is the relative spectral power density of emission.

(7.20)

296

Materials for Solid State Lighting and Displays

With LER defined, the luminous output of an LED is simply: Φ = Po ⋅ LER

(lm)

(7.21)

and the luminous efficacy: 𝜂l = Po ⋅ LER∕Pin = 𝜂 ⋅ LER

(lumens per electrical watt)

(7.22)

Summary performance for commercially available monochromatic high power LEDs is shown in Table 7.1. These data are gathered or derived from the LUXEON Rebel Colors data sheet from 2015 for 350 mA drive at room temperature [60]. PC refers to “phosphor converted”. The (simulated) emission spectra for these emitters are shown in Figure 7.12 and the chromaticity points in the 1931 CIE chromaticity space are shown in Figure 7.13. Also illustrated in Figure 7.13 are the dominant wavelengths (dashed gray lines), which are a common way to designate color in the LED industry for perception of hue. The dominant wavelength is that of a purely saturated source if the LED emission were translated to the edge of the chromaticity space collinear with the chromaticity point of the CIE illuminant E (“equal energy” spectrum). Both Figure 7.12 and Figure 7.13 illustrate the broadening of emission spectra for (Ga,In)N and (Al,Ga)InP as InN and AlInP mole fractions, respectively, are increased in these alloy systems. The broadening for (Al,Ga)InP is understandable as there is a transition from a ternary (GaInP) to quaternary system. For (Ga,In)N, the broadening is clearly due to other mechanisms, such as were discussed earlier. The performance listed in Table 7.1 is consistent with the wavelength trend of external quantum efficiency in Figure 7.11, only the numbers are lower here for these commercially available devices at typical performance levels, in contrast to the record level performance in the figure. In the blue regime, (Ga,In)N emitters achieve about 50% power-conversion efficiency, falling off dramatically as InN increased, and achieving only 21% in the green. For (Al,Ga)InP, near 50% efficiency is obtained in the deep red, with performance falling off to 23% in the amber. Again, all these data are for room temperature. The lack of commercially viable primary LEDs in the 540–585 nm regime is sometimes referred to as the “green gap”. It is particularly painful because it occurs almost exactly where the human eye response is peaked, as shown in Figure 7.11. This situation prompted LED manufacturers to consider photonic down-conversion approaches to monochromatic LEDs in the green gap regime. The PC Green and PC Amber data in Figure 7.11 and Table 7.1 refer to such LEDs, which employ phosphor down-conversion, and have been recently commercialized. For these devices, care is taken to eliminate or minimize “pump” LED primary emission from leaking into the final emission spectrum, so that the color point is not too much de-saturated. Typical light output vs. current (L–I) characteristics for (Ga,In)N blue-emitting and (Al,Ga)InP deep-red-emitting LEDs are shown in Figure 7.14 [61]. One striking feature is the sublinear performance of the (Ga,In)N LED beginning at relatively low currents, with the resulting L–I characteristic “drooping” away from the customary linear trend (in contrast to what is exhibited by the red-emitting LED). This droop characteristic occurs at low currents and even at low temperatures. Originally believed to be due to carrier transport issues, perhaps exacerbated by built-in polarization fields due to the Wurtzite crystal structure, the cause was later shown to be Auger recombination. First proposed in 2007 with great controversy [2], experimental work eventually detected Auger carriers

(Ga,In)N (Ga,In)N (Ga,In)N (Ga,In)N (Ga,In)N

(Al,Ga)InP (Al,Ga)InP (Al,Ga)InP (Al,Ga)InP

(Ga,In)N (Ga,In)N

Green Cyan Blue Royal Blue Violeta

Deep Red Red Red-Orange Amber

PC Green PC Amber

0.2709 0.3058 0.3164 0.4222

0.6879 0.5529 0.0703 0.0294 0.0153

y

0.4208 0.5472 0.5758 0.4195

0.7290 0.6935 0.6828 0.5766

0.2005 0.1041 0.1335 0.1536 0.1668

x

Chromaticity

538 596

660 633 627 592

522 503 465 448 415

nm

568 591

657 622 617 590

530 505 470 455 440

Nm

100 80

20 20 20 20

30 30 20 20 15

nm

2.90 3.05

2.10 2.10 2.10 2.10

2.90 2.90 2.95 2.90 3.00

V

15 62 72 83

360 358 326 168

1.0 192 419 1.1 112 321

0.7 0.7 0.7 0.7

Luminous efficacy

LER

189 105

21 84 98 113

100 82 40 23 7

458 349

42 173 221 493

489 317 78 44 16

lm mW lm∕Welec lm∕Wopt

Light output

1.0 102 209 1.0 83 262 1.0 41 526 1.0 23 520 1.5 11 675

W

Peak 𝜆 Dominant 𝜆 FWHM Forward Pin Voltage

All data representative of packaged 1 x 1 mm2 LED chips at 350 mA forward current operation at room temperature, unless otherwise specified. LER, Lumen equivalent of radiation. a Violet-emitting LED data given for 500 mA operation.

III-V System

Typical performance characteristics for “colored” LEDs circa 2015

Color

Table 7.1

52% 44%

55% 52% 47% 23%

25% 30% 56% 54% 45%

41% 30%

49% 49% 44% 23%

21% 26% 51% 51% 45%

External Power quantum conversion efficiency efficiency 𝜂ext 𝜂

Light Emitting Diode Materials and Devices 297

298

Materials for Solid State Lighting and Displays (Al,Ga)InP

PC

Spectral power distribution

(Ga,In)N

400

450

500

600

550

650

700

Wavelength (nm) Violet PC Amber

Royal Blue Amber

Figure 7.12

Blue Red-Orange

Cyan Red

Green Deep Red

PC Green

Simulated emission spectra for the LEDs of Table 7.1

in an operating (Ga,In)N LED [61]. The evidence of Auger electrons in such a wide bandgap semiconductor suggests that a simple four-band model may be insufficient to describe carrier–carrier interaction in such materials, and that inter-valley effects might be considered in addition to photon and defect related effects. Since carrier separation due to QCSE increases carrier lifetime and thus carrier density, it was believed that reducing polarization fields in (Ga,In)N heterostructures could reduce carrier density, and thus alleviate droop. The boldest approach to exploring this avenue was moving away from standard (0001) basal plane (i.e., c-axis or “polar”) structures, to highly off-axis “semi-polar” or even “non-polar” (e.g., m-plane or a-plane) platforms for optoelectronic device construction. This was made possible by the availability of HVPE grown bulk GaN boules, which could be cut and polished at large offcut angles, including perpendicular to the boule growth direction, typically (0001) [62]. While in principle reduced built-in fields should improve overlap and reduce carrier density, and thus “push out” the onset of droop due to Auger recombination, leveraging this potentially advantage in commercial devices has remained elusive, and best performing LEDs today are all based on the c-plane growth platform. 7.5.3

White-Emitting Performance

As discussed earlier, due to the weak performance of green- and longer-wavelength (Ga,In)N LEDs compared with shorter wavelength devices, it is most efficient to employ down-conversion to generate white light. It is also relatively simple. For example, a blue-emitting LED may be combined with a single YAG broad-band yellow-emitting phosphor with the combined light mixed to make white [29]. Using the formalism described earlier, the terms driving the efficiency of a down-conversion (e.g., phosphor-based), LED

Light Emitting Diode Materials and Devices

299

0.9 520 Green 530

0.8 510

D om i 540 nan

tW av e 550 len

0.7 n) N

Cyan

(n m 560 )

(G a,I

0.6

PC Green 570

500 y-chromaticity

gt h

0.5

580 PC

Amber 590

0.4 600

(A

CIE Illuminant E

I,G

a)

In

0.3 490

P

RedOrange 610 620 Red 700 Deep Red

0.2 480 0.1

Blue 470 Royal Blue Violet 0.0 0.1 0.0

0.2

0.3

0.4 0.5 x-chromaticity

0.6

0.7

0.8

Figure 7.13 Color points in x–y chromaticity space for the LEDs of Table 7.1, which also determine the LEDs dominant wavelengths which are a projection of their color points from the CIE-E “equal energy” illuminant to the edge of the chromaticity space

may be examined in more detail. Considering a simple one-down-converter scheme based on a primary blue-emitting LED, the power conversion efficiency may be written, from Equation 7.13, as: 𝜂 = [ fb ∕(𝜂b Cpkg,b ) + fph ∕(𝜂b 𝜂ph 𝜂Stokes Cpkg,ph ) ]−1

(7.23)

where fb and fph are the optical fractions of blue primary and down-converted light, respectively, 𝜂ph is the quantum efficiency of the phosphor, Cpkg,b and Cpkg,ph represent losses in the package for blue primary and down-converted light, respectively, and 𝜂b is the power conversion efficiency of the blue-emitting primary LED in a “perfect package”, that is: (7.24) 𝜂b = 𝜂int Cext Ep ∕Vf

300

Materials for Solid State Lighting and Displays 1400

1200

Forward current (mA)

1000 (Ga,In)N λ ∼ 450 nm

800

600 (Al,Ga)InP λ ∼ 650 nm 400

200 T = 300K 0 0

200

400 600 Light output (mW)

800

1000

Figure 7.14 Typical light output vs. current (L–I) characteristics for (Ga,In)N blue-emitting and (Al,Ga)InP deep-red-emitting LEDs. The sublinear behavior for the (Ga,In)N LED, sometimes termed efficiency “droop”, is due primarily to Auger recombination

since the package loss terms have been broken out. Furthermore, 𝜂Stokes represents the Stokes’ loss, or “quantum deficit,” given by: centroid centroid 𝜂Stokes = Ep,ph ∕Ep,b ∼ 𝜆b ∕𝜆ph

(7.25)

where 𝜆b and 𝜆ph are the centroid emission wavelengths of the blue primary LED and phosphor emission spectra, respectively. Strictly speaking, the loss terms Cpkg,b and Cpkg,ph are not relegated only to package effects since in reality it can be considered that the primary LED chip light extraction efficiency may be compromised by surrounding it with phosphor. However, in the LED industry it has been customary to lump all the unaccounted for losses for phosphor-converted LEDs into these “package efficiency” terms, regardless of where and how exactly they occur. It is furthermore customary to make the assumption that package efficiency is similar for both blue primary and phosphor-converted light, so that one may write: 𝜂 = 𝜂b Cpkg [ fb + fph ∕(𝜂ph 𝜂Stokes )]−1

(7.26)

If the leaked blue light fraction is low, or the phosphor quantum efficiency is near unity, or both, Equation 7.22 can be simplified to a convenient expression often used in the LED industry, that is: (7.27) 𝜂 ∼ 𝜂b Cpkg 𝜂ph [ fb + fph ∕𝜂Stokes ]−1

Light Emitting Diode Materials and Devices

301

Expanding the terms for the blue primary LED, Equation 7.23 becomes: 𝜂 ∼ 𝜂int Cext (Ep ∕Vf ) Cpkg 𝜂ph [ fb + (1 − fb )∕𝜂Stokes ]−1

(7.28)

In Equation 7.28 one can observe the simple multiplication of terms that make up the phosphor-based LED efficiency. The photometric version of the expression is given by simply appending the LER, that is: 𝜂l ∼ 𝜂int Cext (Ep ∕Vf )Cpkg 𝜂ph [ fb + (1-fb )∕𝜂Stokes ]−1 LER

(7.29)

White-emitting LED performance has grown steadily over the last two decades since the invention of the blue-emitting LEDs. The performance is tracked by certain entities, such as the US Department of Energy (DOE), whose roadmap for performance in luminous efficacy (past actual and future target) is shown in Figure 7.15 [63]. Also shown are the typical performance levels for conventional light sources such as incandescent tungsten, tungsten–halogen, compact and linear fluorescent, and high intensity discharge white light products. LED performance has surpassed all these conventional light sources and still has room to continue to improve. Cost savings from energy efficiency alone has made the transition to LEDs undeniable. Approximately 20% of site electricity is used for lighting today, and the US DOE estimates US$ 380 B in savings nationwide are possible cumulatively between the years of 2013 and 2030 through aggressive adoption of LED technology [64]. Representative performance for white-emitting LEDs, at the time of this writing, is shown in Table 7.2, for commercially available LEDs from Lumileds circa 2015 [65]. These LEDs are available in a wide range of color temperatures, from warm (5000 K). Within each color temperature, different levels of CRI are provided. “Low” CRI values of the mid-80s are deemed reasonable for general illumination applications wherein precise color accuracy is not so important. This includes general purpose lighting, outdoor applications, and so on. “High” CRI, at values greater than 85, is preferred in applications where color accuracy is deemed important, such as in certain retail and hospitality settings, and certain residential homes. As shown in Figure 7.16, luminous efficacy for LEDs generally increases as CCT is increased, and is generally lower for High-CRI LEDs than for Low-CRI LEDs. This is primarily due to two reasons. The first is that reducing CCT means less primary LED light leakage is allowed, meaning that more light is going through the down-conversion process. This means a more lossy process, due to quantum efficiency of the phosphor and Stokes’ loss, since more photons are being converted, and also due to the fact that more scattering events are occurring for the primary LED light. The second reason is that, as CCT decreases, LER decreases. This is fundamental, since more red light is required to track the blackbody loci in chromaticity space, and there are simply less lumens in red light compared with shorter wavelength light. This second reason is also why High-CRI LEDs under-perform compared with Low-CRI LEDs. More red light is needed to render deep red colors appropriately, and again, this costs lumens. These effects can be appreciated by viewing emission spectra corresponding to the LEDs in Table 7.2, which are shown in Figure 7.17. The LED spectra from Figure 7.17 are shown in the x–y chromaticity diagram in Figure 7.18. Also shown are typical color points for a blue-emitting LED and phosphor emission from those most commonly used for LED in general illumination, that is, LuAG, 160

Luminous efficacy (Im/W)

70

65

140

“Cool white”

120

85

85

100

85

85

“Warm white” 80

95 95

60 40 20 Labels = Color Rendering Indices, Ra 0 2000

2500

3000

3500

4000

4500

5000

5500

6000

Correlated color temperature (K) Figure 7.16 White-emitting LED luminous efficacy as a function of correlated color temperature. Warm white LEDs are generally less efficacious than cool white LEDs, due to their need for more red lumens. The color rendering index for each data point is indicated

304

350

Materials for Solid State Lighting and Displays

400

450

5650K, CRI 70 3000K, CRI 85

500

550 600 Wavelength (nm)

5000K, CRI 85 2700K, CRI 85

650

4100K, CRI 65 3000K, CRI 95

700

750

800

4000K, CRI 85 2700K, CRI 95

Figure 7.17 Emission spectra for the blue-pumped phosphor-converted white-emitting LEDs of Table 7.2, with their varying color temperatures and color rendering indices (CRIs)

YAG and “CASN” phosphors. Visualizing this way, one can get an idea for the breadth of colors that are possible to achieve with phosphor-converted LEDs, which showcases the power of this approach, but also raises the issue of reproducing the correct color over and over again for a give application. Indeed, color targeting is an important issue for white-emitting LEDs, since small changes in chromaticity in the white color regimes can be perceived by the human eye and deemed disagreeable. It is clear from the above discussion that primary LEDs of different peak emission wavelengths will yield different white color points when utilizing the same phosphor materials mix. The LED industry has reacted to this issue by applying fairly sophisticated color binning schemes in which LEDs of different peak emission

Light Emitting Diode Materials and Devices

305

0.9 520 530

0.8

540

510

0.7

550

Lu3AI5O12:Ce 560

Y3AI5O12:Ce

0.6 570

y-chromaticity

500

0.5

580

CaAISiN3:Eu 590

0.4

3000K

1800K

600

ci y Io bod

5000K

610

k

0.3

490

10000K

Blac

620 700

20000K

0.2 480

0.1 Blue-emitting 470 “pump” LED

0.0 0.0

0.1

0.2

0.3

0.4 x-chromaticity

0.5

0.6

0.7

0.8

Figure 7.18 Color points for the LEDs of Table 7.2 in x–y chromaticity space. Also shown are the color points for a primary blue-emitting LED as well as for phosphors most commonly used today for general illumination applications

wavelengths are matched to the appropriate phosphor application process step in order to yield the desired color. Color differences between white-emitting LEDs are typically quantified in terms of Standard Deviations of Color Matching (SDCM), or McAdams Ellipse (MCE). At the time of writing, top tier LED companies can guarantee initial white color point distributions of 3 MCE, which for 3000 K corresponds approximately to a 0.003 (“3 point”) diameter circular region in u’v’ chromaticity space. There is a push to tighten color control even further, with a target of 1 MCE as a worthy goal. For displays, LEDs are most commonly used as a backlight source for LCD panels, replacing the previously favored cold-cathode fluorescent lamps. The very first LED-backlit LCD television used individual red, green, and blue emitters. However, the LED and assembly costs, coupled with the challenge to maintain a fixed color point, prompted the industry to move towards white-emitting LEDs for LCD backlighting. Choice of phosphors can determine the level of color gamut that the display can achieve,

306

Materials for Solid State Lighting and Displays

with the least expensive “single-YAG” phosphor approach achieving only about 67% of NTSC. Replacing the yellow YAG with separate green- and red-emitting phosphors allows to push the gamut up to above 80% NTSC. In the LCD application, it would be desirable to have fairly narrow band green and red emitters, in order to match the color filters used in the LCD panels. This approach is today pursued using line-emitting phosphors or QDs to replace existing LED phosphors, by incorporating them into remote elements (sheets or rods) or even into the LED packages themselves. Using such approaches color gamuts in excess of 100% NTSC have been achieved. There is also today research and development in direct-view “micro” primary LEDs as pixels for very low power consumption wearable displays and/or displays with daylight viewability. 7.5.4

Temperature Effects

Operating temperature has a profound effect on LED performance, and it is worth summarizing that for the reader interested in practical application or characterization of LEDs or LED-based systems. The critical temperature for the LED chip itself is its junction temperature. The junction temperature is given by: (7.30) Tj = Pin (1 – 𝜂)Θja + Ta where Pin is the input electrical power to the LED, 𝜂 is the LED’s power conversion efficiency, Θja is the total thermal resistance from junction to ambient, and Ta is the ambient temperature. Equation 7.30 shows the importance of power conversion efficiency, since power converted to light is radiated away and does not generate heat. It is not always easy to know the thermal resistance from junction to ambient. In system design, a method often used is to specify a case or board temperature (e.g., 85 ∘ C for standard PCBs) and then use the junction-to-case thermal resistance which is also the LED “package thermal resistance” that is specified in LED data sheets. For LED active region radiative efficiency, the effects of elevated temperature are all deleterious. Rising temperatures increase the activity of traps and phonons which serve to increase non-radiative recombination under both high and low carrier density operation. The effects of temperature on current injection can be either negative, positive, or (ideally) negligible. In the case of material systems with proximal indirect bandgap valleys, for example, (Al,Ga)InP, increasing temperature decreases the percentage population of carriers in the direct gamma valley, leading to reduced radiative efficiency. In the case of (Ga,In)N, the very high activation energy for Mg as an acceptor in GaN can result in increased hole concentrations with elevated temperature, which can actually improve current injection efficiency. Forward voltage of LEDs is typically reduced as temperature is increased, as the bandgap reduction due to thermal lattice expansion (Varshni effect) typically overwhelms any increase in device resistivity. Typical shifts are a few millivolts per degree Kelvin. This same effect results in emission peak wavelength red-shift, and slight emission width broadening, with increasing temperature. 7.5.5

Reliability

LEDs, unlike conventional light sources such as incandescence or electrode discharge, employ fundamentally non-destructive processes to generate light. It is not surprising,

Light Emitting Diode Materials and Devices

307

then, that the operating lifetimes of products based on them are several factors, even orders of magnitude, higher than conventional lighting products. Even so, there are degradation mechanisms at play in LEDs, typically related to materials issues within the packaging scheme, including sometimes phosphor materials. An industry-led effort to summarize these effects, for phosphor-based white-emitting LEDs, has resulted in a predictive tool for LED light output preservation, typically referred to as lumen maintenance. This guideline is published by the Illuminating Engineering Society (IES) under their Technical Memorandum (TM) no. 21 [66]. It establishes that up to 25 000 h lumen maintenance may be estimated by using 6000 h of lifetime testing after 1000 h. It has been used, for example, by the Environmental Protection Agency’s Energy Star program which requires 70% lumen maintenance (“L70”) at 25 000 h for most LED-based lighting products for certification [31]. It is worth noting that IES has also published potential improvements to measuring color rendering [67], which is under review by the CIE. Another degradation issue that can arise for LEDs is color shift over operating lifetime. This issue was also present for conventional light sources but was not considered such a big issue since it was understood that they would be replaced on a fairly regular basis, given their short lifetimes. A consequence of the very long operating lifetimes for LEDs, in contrast, has made color maintenance a concern [68]. This is a real issue, as it is easier for a human eye to detect a few-point chromaticity shift between neighboring light sources, than a 20% variation in lumen output. At the time of this writing, a generally accepted guideline for predicting color shift over time has not yet been established. Partly this is due to the fact the specific choices on primary LED emission wavelength, phosphor materials, and packaging architecture can effect color maintenance differently, so it is more difficult to expect homogeneous behavior across different products and different suppliers. The current, and arguably insufficient, Energy Star guideline only requires less than 7 points chromaticity shift in u’v’ space over 6000 h of operating life [31].

References 1. 2. 3. 4. 5. 6.

7. 8. 9.

Holonyak, N. Jr. and Bevacqua, S.F. (1962) Coherent (visible) light emission from Ga(AsP) junctions. Appl. Phys. Lett., 1, 82–83. Shen, Y.C., Mueller, G.O., Watanabe, S. et al. (2007) Auger recombination in InGaN measured by photoluminescence. Appl. Phys. Lett., 91, 141101. Kioupakis, E., Steiauf, D., Rinke, P. et al. (2015) First-principles calculations of indirect Auger recombination in nitride semiconductors. Phys. Rev. B, 92, 035207. Schubert, E.F. (2006) Light-Emitting Diodes, 2nd edn, Cambridge University Press. R. Fitzgerald (2000) Physics Nobel Prize honors roots of Information Age. Physics Today, vol. 53 (12), pp. 17-19. R.D. Dupuis, P.D. Dapkus, N. Holonyak Jr, et al. (1978) Room-temperature laser operation of quantum-well Ga1-xAlxAs-GaAs laser diodes grown by metalorganic chemical vapor deposition. Appl. Phys. Lett., vol. 32 (5), pp. 295–297. D.G. Thomas and J.J. Hopfield (1966) Isoelectronic traps due to nitrogen in gallium phosphide. Phys. Rev., 150 (2), 680–689. Logan, R.A., White, H.G. and Trumbore, F.A. (1967) p-n junctions in compensated solution grown GaP. J. Appl. Phys., 38, 2500–2508. M.G. Craford, Recent developments in light-emitting-diode technology. IEEE Trans. Electron Dev., vol. ED-24 (7), pp. 935–943, 1977.

308

Materials for Solid State Lighting and Displays

10. H. Rupprecht, J.M. Woodall, and G.D. Petit, Efficient visible electroluminescence at 300 K from GaAlAs p–n junctions grown by liquid phase epitaxy. Appl. Phys. Lett., vol. 11 (3), pp. 81–83, 1967. 11. R.D. Dupuis and P.D. Dapkus, Room-temperature operation of Ga(1-x)AlxAs/GaAs double heterostructure lasers grown by metal organic chemical vapor deposition. IEEE Trans. Electron Dev., vol. ED-24 (9), pp. 1195–1196, 1977. 12. M. Ikeda, K. Nakano, Y. Mori et al. (1986) MOCVD growth of AlGaInP at atmospheric pressure using triethylmetals and phosphine. J. Cryst. Growth, vol. 77 (1–3), pp. 380–385. 13. M. Ikeda, Y. Mori, H. Sato et al. (1985) Room-temperature continuous-wave operation of an AlGaInP double heterostructure laser grown by atmospheric pressure metalorganic chemical vapor deposition. Appl. Phys. Lett., vol. 47 (10), pp. 1027–1028. 14. Huang, K.H., Yu, J.G., Kuo, C.P. et al. (1992) Twofold efficiency improvement in high performance AlGaInP light-emitting diodes in the 555–620 nm spectral region using a thick GaP window layer. Appl. Phys. Lett., 61, 1045–1047. 15. F.A. Kish, D.A. DeFevere, D.A. Vanderwater et al. (1994) High luminous flux semiconductor wafer-bonded AlGaInP/GaP large-area emitters. Electron. Lett., vol. 30 (21), pp. 1790–1792. 16. Krames, M.R., Ochiai-Holcomb, M., Hofler, G.E. et al. (1999) High-power truncated-inverted-pyramid (AlGa)InP-GaP light-emitting diodes exhibiting > 50% external quantum efficiency. Appl. Phys. Lett., 75, 2365–2367. 17. K. Streubel, N. Linder, R. Wirth, and A. Jaeger, High brightness AlGaInP lightemitting diodes. IEEE J. Sel. Topics Quantum Electron., vol. 8 (2), pp. 321–332, 2002. 18. Maruska, H.P. and Tietjen, J.J. (1969) The preparation and properties of vapordeposited single-crystal-line GaN. Appl. Phys. Lett., 15, 327–329. 19. H.P. Maruska and D.A. Stevenson, ‘Mechanism of light production in metal-insulatorsemiconductor diodes; GaN:Mg violet light-emitting diodes,’ Solid-State Electron., vol. 17, pp. 1171–1179, 1974. 20. H. Amano, N. Sawaki, I. Akasaki, and Y. Toyoda, ‘Metalorganic vapor phase epitaxial growth of a high quality GaN film using an AlN buffer layer,’ Appl. Phys. Lett., vol. 48, pp. 353–355, 1986 21. Nakamura, S., Mukai, T., Senoh, M. and Iwasa, N. (1992) Thermal annealing effects on p-type Mg-doped GaN films. Jpn. J. Appl. Phys., 31, L139–L142. 22. Lester, S.D., Ponce, F.A., Craford, M.G. and Steigerwald, D.A. (1995) High dislocations densities in high efficiency GaN-based light-emitting diodes. Appl. Phys. Lett., 66, 1249–1251. 23. Nakamura, S., Senoh, M., Iwasa, N. and Nagahama, S. (1995) High-brightness InGaN blue, green and yellow light-emitting diodes with quantum well structures. Jpn. J. Appl. Phys., 34, L797–L799. 24. Nakamura, S., Senoh, M., Iwasa, N. et al. (1995) Superbright green InGaN singlequantum-well-structure light-emitting diodes. Jpn. J. Appl. Phys., 34, L1332–L1335. 25. Nakamura, S. and Fasol, G. (1997) The Blue Laser Diode, Springer-Verlag, pp. 216–219. 26. Chichibu, S.F., Abare, A.C., Minsky, M.S. et al. (1998) Effective band gap inhomogeneity and piezoelectric field in InGaN/GaN multiquantum well structures. Appl. Phys. Lett., 73, 2006.

Light Emitting Diode Materials and Devices

309

27. Steigerwald, D.A., Bhat, J.C., Collins, D. et al. (2002) Illumination with solid state lighting technology. IEEE J. Sel. Topics Quant. Electron., 8, 310–320. 28. M.R. Krames, O.B. Shchekin, R. Mueller-Mach et al. (2007) Status and future of high-power light-emitting diodes for solid-state lighting. IEEE/OSA J. Display Technol., vol. 3 (2), pp. 160-175. 29. Yamada, M., Mitani, T., Narukawa, Y. et al. (2002) InGaN-based near-ultraviolet and blue-light-emitting diodes with high external quantum efficiency using a patterned sapphire substrate and a mesh electrode. Jpn. J. Appl. Phys., 41, L1431–L1433. 30. Shchekin, O.B., Epler, J.E., Trottier, T.A. et al. (2006) High performance thin-film flip-chip InGaN-GaN light-emitting diodes. Appl. Phys. Lett., 89, 071109. 31. Haerle, V., Hahn, B., Kaiser, S. and Weimar, A. (2004) S. bader, F. Eberhard, A. Plössl, and D. Eisert, High brightness LEDs for general lighting applications using the new ThinGaN technology. Phys. Status Solidi A, 201, 2736–2739. 32. Motoki, K., Okahisa, T., Nakahata, S. et al. (2002) Growth and characterization of freestanding GaN substrates. J. Crystal Growth, 237–239, 912–921. 33. Hurni, C.A., David, A., Cich, M.J. et al. (2015) Bulk GaN flip-chip violet light-emitting diodes with optimized efficiency for high-power operation. Appl. Phys. Lett., 106, 031101. 34. M. Krames, Light emitting diodes: GaN-on-GaN platform removes cost/performance tradeoffs in LED lighting. Laser Focus World, vol. 49 (9), p. 37(4), 2013. 35. S. Nakamura and M.R. Krames, History of gallium–nitride-based light-emitting diodes for illumination. Proc. IEEE, vol. 101 (10), pp. 2211-2220, 2012. 36. Schlotter, P., Schmidt, R. and Schneider, J. (1997) Luminescence conversion of blue light emitting diodes. Appl. Phys. A, 64, 417–418. 37. Xie, R.-J., Hirosaki, N., Li, Y. and Takeda, T. (2010) Rare-earth activated nitride phosphors: Synthesis, luminescence and applications. Materials, 3, 3777–3793. 38. ENERGY STAR Program Requirements for Lamps - Eligibility Criteria (Rev. August 2014). 39. Fujita, S., Yoshihara, S., Sakamoto, A. et al. (2005) YAG glass-ceramic phosphor for white LED (I): Background and development. Proc. SPIE, 5941, 594111. 40. Bechtel, H., Schmidt, P., Busselt, W. and Schreinemacher, B.S. (2008) Lumiramic - A new phosphor technology for high performance solid state light sources. Proc. SPIE, 7058, 7058E. 41. Mueller-Mach, R., Muller, G.O. and Krames, M.R. (2007) Phosphor-converted high power LEDs. Proc. SPIE, 6797, 6797G. 42. Mueller-Mach, R., Mueller, G.O., Krames, M.R. et al. (2009) All-nitride monochromatic amber-emitting phosphor-converted light-emitting diodes. Phys. Status Solidi RRL, 3, 215–217. doi: 10.1002/pssr.200903188 43. Wei, M., Houser, K.W., David, A. and Krames, M.R. (2014) Perceptual responses to LED illumination with colour rendering indices of 85 and 97. Lighting Res. Technol., 47, 810–827. 44. A.A. Setlur, E.V. Radkov, C.S. Henderson et al. (2010) Energy-efficient, highcolor-rendering LED lamps using oxyfluoride and fluoride phosphors. Chem. Mater., vol. 22 (13), pp. 4076-4082, DOI: 10.1021/cm100960g. 45. Pust, P., Weiler, V., Hecht, C. et al. (2014) Narrow-band red-emitting Sr[LiAl3 N4 ]:Eu as a next-generation LED-phosphor material. Nat. Mater., 13, 891–896.

310

Materials for Solid State Lighting and Displays

46. L. Wang, X. Wang, T. Kohsei et al. (2015) Highly efficient narrow-band green and red phosphors enabling wider color-gamut LED backlight for more brilliant displays. Opt. Express, vol. 23 (22), pp. 28707–28717. DOI: 10.1364/OE.23.028707. 47. G.E. Hoefler, C. Carter-Coman, M.R. Krames et al. (1998) High-flux high-efficiency transparent-substrate AlGaInP/GaP light-emitting diodes. Electron. Lett., vol. 34 (18), pp. 1781-1782. 48. G. Chen, M. Craven, A. Kim et al. (2008) Performance of high-power III-nitride light emitting diodes. Phys. Status Solidi A, vol. 205 (5), pp. 1086–1092. 49. Narukawa, Y., Ichikawa, M., Sanga, D. et al. (2010) White light emitting diodes with super-high luminous efficacy. J. Phys. D: Appl. Phys., 43, 354002. 50. David, A., Hurni, C.A., Aldaz, R.I. et al. (2014) High light extraction efficiency in bulk-GaN based volumetric violet light-emitting diodes. Appl. Phys. Lett., 105, 231111. 51. Broell, M., Sundgren, P., Rudolph, A. et al. (2014) New developments on highefficiency infrared and InGaAlP light-emitting diodes at OSRAM Opto Semiconductors. Proc. SPIE, 9003, 90030L. 52. R. Hashimoto, J.-I. Hwang, S. Saito, and S. Nunoue, High-efficiency green-yellow light-emitting diodes grown on sapphire (0001) substrates. Phys. Status Solidi C, vol. 10 (11), pp. 1529– 1532 (2013). 53. R. Hashimoto, J.-I. Hwang, S. Saito, and S. Nunoue, High-efficiency yellow light-emitting diodes grown on sapphire (0001) substrates. Phys. Status Solidi C, vol. 11 (3–4), pp. 628– 631, 2014. 54. Saito, S., Hashimoto, R., Hwang, J.-I. and Nunoue, S. (2013) InGaN light-emitting diodes on c-face sapphire substrates in green gap spectral range. Appl. Phys. Express, 6, 111004. 55. Hwang, J.-I., Hashimoto, R., Saito, S. and Nunoue, S. (2014) Development of InGaNbased red LED grown on (0001) polar surface. Appl. Phys. Express, 7, 071003. 56. F.A. Ponce, S. Srinivasan, A. Bell et al. (2003) Microstructure and electronic properties of InGaN alloys. Phys. Status Solidi B, vol. 240 (2), pp. 273–284. 57. Della Sala, F., Di Carlo, A., Lugli, P. et al. (1999) Free-carrier screening of polarization fields in wurtzite GaN/InGaN laser structures. Appl. Phys. Lett, 74, 2002. 58. Chichibu, S.F., Abare, A.C., Mack, M.P. et al. (1999) Optical properties of InGaN quantum wells. Mater. Sci. Eng., B59, 298–306. 59. Yang, T.-J., Shivaraman, R., Speck, J.S. and Wu, Y.-R. (2014) he influence of random indium alloy fluctuations in indium gallium nitride quantum wells on the device behaviour. J. Appl. Phys, 116, 113104. 60. http://www.lumileds.com/products/color-leds/luxeon-rebel-color. Accessed December 26, 2015. 61. J. Iveland, L. Martinelli, J. Peretti et al. (2013) Direct measurement of Auger Electrons emitted from a semiconductor light-emitting diode under electrical injection: Identification of the dominant mechanism for efficiency droop. Phys. Rev. Lett., vol. 110 (17), pp. 177406-177410. 62. M.C. Schmidt, K.C. Kim, H. Sato et al. (2007) ‘High power and high external efficiency m-plane InGaN light emitting diodes,’ Jpn. J. Appl. Phys., vol. 46, pp. L126–L128.

Light Emitting Diode Materials and Devices

311

63. Solid state lighting R&D plan, U.S. Department of Energy, 2015. 64. Energy savings forecast of solid-state lighting in general illumination applications, Prepared for the U.S. Department of Energy by Navigant Consulting, Inc., 2014. 65. http://www.lumileds.com/products/high-power-leds/luxeon-rebel. Accessed December 26, 2015. 66. Projecting long term lumen maintenance of LED light sources, Illuminating Engineering Society TM-21-11, 2011. 67. IES method for evaluating light source color rendition, Illuminating Engineering Society TM-30-15, 2015. 68. Royer, M., Tuttle, R., Rosenfeld, S. and Miller, N. (2013) Color maintenance of LEDs in laboratory and field applications, Pacific Northwest National Laboratory, U.S. Department of Energy.

8 Alternating Current Thin Film and Powder Electroluminescence Adrian Kitai McMaster University, Hamilton, Ontario, Canada

Alternating current (AC) electroluminescent light emitting materials and devices play a small but continuing role in today’s solid state lighting and display markets. Coverage in this book has both academic and industrial value: The intrinsic light emission mechanisms and device features are unique and the materials used are low in cost compared with materials used in either organic or inorganic light emitting diode (LED) devices.

8.1

Introduction

The semiconductor-based solid state conversion of electrical energy to visible light has become dominated by both inorganic and organic p-n junction LED devices, however there are other mechanisms of effecting luminescence from electron–hole pair recombination in devices that do not necessarily have both p-type and n-type regions. Powder electroluminescence (EL) was discovered in the 1930s and continues to find niche applications in both small and large area electronics. Thin film electroluminescence (TFEL) which was developed in the 1970s has niche applications in small but very rugged displays. Here, an electron is excited to a higher energy level within an ion and when it returns to its ground state a photon may be emitted. The purpose of this chapter is to maintain awareness of these rather interesting alternative mechanisms of light generation in spite of their small market penetration. It is useful to start by examining features that distinguish both powder EL and TFEL materials and devices from diode-type devices. The materials used in powder EL and TFEL devices are generally inorganic polycrystalline solids rather than the single crystal materials used in LEDs.

Materials for Solid State Lighting and Displays, First Edition. Edited by Adrian Kitai. © 2017 John Wiley & Sons Ltd. Published 2017 by John Wiley & Sons Ltd.

314

Materials for Solid State Lighting and Displays

This permits large area devices to be manufactured at a reasonable cost. Highly uniform light emission over large areas is available. This is difficult to achieve with LEDs that are small area, albeit intense, sources of light. The polycrystalline nature of these EL powder and thin film materials is the key reason for the uniformity of the light emission. The light emitted from each grain of material is normally not seen as a distinct source by the observer, as the light per unit area emitted is the result of light from numerous grains or crystals. For this reason, a less perfect material than single crystal LED materials is able to create highly reproducible lighting without requiring the binning associated with LED devices. Another key distinction between LEDs and powder EL and TFEL materials lies in the light emission process. In LEDs, band-to-band recombination generally occurs. This means that traps and defects that lead to unwanted, non-emissive recombination events between electrons and holes must be minimized by using high purity, low defect density materials. This is fundamentally due to the delocalization of electrons and holes in energy bands. Since electron and hole conduction is a necessary condition for efficient diode-type devices, conduction and valence band transport is necessary. This carrier delocalization is therefore a mixed blessing, as it allows for high current density carrier transport and the associated availability of electrons and holes to supply recombination, however, it also allows carriers relatively free access to any traps located within a diffusion length of the carriers. In powder EL and TFEL materials, luminescence is derived from non-mobile charges. They are trapped in donor/acceptor-type traps (powder EL) or within atomic orbitals in the case of TFEL. Powder EL and TFEL materials are based on electrically insulating semiconductors. Charge transport occurs by a mechanism other than ohmic transport: upon application of a high electric field, high field breakdown (avalanche breakdown) occurs and charges flow, even though the EL materials are normally insulators. In contrast to organic light emitting diode (OLED) devices, powder EL and TFEL materials are relatively air and moisture insensitive due to their use of inorganic crystalline materials. OLEDs have advantages in several other aspects, however, including higher efficiencies, wider choices of available colors and low voltage operation.

8.2

Background of TFEL

In 1967, Russ and Kennedy [1] demonstrated a double-insulating layer EL structure. Figure 8.1 shows the structure of a double-insulating layer TFEL device which is the commercially available structure for TFEL devices. It consists of a substrate, usually glass, on which a series of thin film layers is grown. There are two electrodes, two dielectric layers, and a phosphor layer. More recently, a number of variants on this structure have been developed in which thick films or even ceramic sheets are used in place of the thin film insulating layers. In all cases, we will classify these structures as TFEL devices because the phosphor layer, which emits the light, is always a thin film. Note that sometimes only one dielectric layer is shown although two insulating layers provide more symmetrical, stable and reliable operation. Two thin (e.g., 10 nm) interface layers on the two sides of the phosphor layer may be formed as part of the EL structures even if there is only one thicker insulating layer. A number of alternative device structures that can produce TFEL have been developed. These structures offer specific advantages but have not been commercialized, and are now only briefly reviewed.

Alternating Current Thin Film and Powder Electroluminescence

315

Rear electrode Dielectric layer 2 Phosphor layer Dielectric layer 1 Transparent electrode Glass substrate

(a)

Rear electrodes

Phosphor layer

Transparent front electrodes Glass substrate (b)

Figure 8.1 (a) Thin film double-insulating-layer TFEL device structure showing sequence of layers deposited as thin films on a glass substrate. Typical materials used are:

• • • • •

glass substrate: Corning 1737 glass; transparent electrode: Indium tin oxide (ITO), 150 nm; dielectric layers: Aluminium titanate, 200 nm each; phosphor layer: Manganese-doped zinc sulfide (ZnS:Mn), 500 nm; rear electrode: Aluminum, 100 nm.

(b) TFEL display with rows and columns. The sheet resistance of the transparent (ITO) electrodes is important for matrix addressing. ITO can be grown with under 10 ohms per square and over 80% transmission of light

8.2.1

Thick Film Dielectric EL Structure

Developed by iFire Inc., Thick Dielectric EL (TDEL) devices [2] offer high EL brightness and efficiency. Here, the thin film dielectric layers are replaced with a single thick film dielectric made from high dielectric constant (relative dielectric constant over 1000) ceramics such as barium titanates or lead zirconium titanates. The thick film layer may be deposited by screen printing. A subsequent planarization layer provides a sufficiently smooth dielectric surface to permit the deposition of good-quality phosphor thin films. See Figure 8.2.

316

Materials for Solid State Lighting and Displays ITO front contact Dielectric layer Phosphor layer Thick film dielectric layer Thick film metalization Ceramic substrate

Figure 8.2 Thick film ceramic dielectric EL device. The TFEL phosphor and front electrode are grown on a thick film laminate to form a robust EL device. This structure has the advantage of a thick dielectric layer that is far less sensitive to defects that thin film dielectric layers. As a result of the high dielectric constant of the dielectric layer, high charge injection levels may be reached that permit higher phosphor brightness. In addition, more efficient light out-coupling than for the structure of Figure 7.1 may be obtained

ITO Dielectric Phosphor BaTiO3 substrate Metal

Figure 8.3 Ceramic sheet dielectric EL device showing self-supporting substrate design. Rear electrode may be a thin film metalization. Due to the brittle nature of the thin ceramic sheet, only small-area devices are practical. The barium titanate substrate is typically 200 𝜇m thick

8.2.2

Ceramic Sheet Dielectric EL

The use of free-standing and self-supporting ceramic sheets for EL substrates has been employed to develop a number of new EL phosphors [3]. Typically a barium titanate ceramic is prepared by sintering barium titanate powder to form a dense ceramic and then polishing at least one side of the resulting ceramic sheet to prepare for the phosphor layer. The sheet is typically about 200 μm thick. See Figure 8.3. 8.2.3

Sphere-Supported TFEL

The device structure [4] is effectively composed of numerous spherical, small-area TFEL devices embedded within a polymer (polypropylene) film. Each spherical TFEL device consists of a barium titanate sphere coated with a series of thin films, including the EL phosphor layer on the upper surface and an electrode on the lower surface. See Figure 8.4.

Alternating Current Thin Film and Powder Electroluminescence

317

Gold on polypropylene (~30 nm) ITO (~100 nm) Al2O3 (~50 nm)

polypropylene (~40 μm)

Phosphor (~700 nm)

Gold (~60 nm) BaTiO3 (~58 μm)

Figure 8.4 High dielectric constant BaTiO3 spheres are sandwiched between upper and lower electrodes to form many individual TFEL devices that form a flexible sheet of light. BaTiO3 spheres are coated with thin film interface layers and phosphor layer on the upper side and conductive electrode layers on both upper and lower sides

8.3

Theory of Operation

The most important electronic processes in TFEL devices occur both within the phosphor layer as well as at the phosphor interfaces. The phosphor layer must satisfy a number of criteria to enable efficient light emission: • • • •

It must be transparent to the wavelength of light being emitted. It must contain impurities having localized quantum states. It must be an electrical insulator. It must exhibit an avalanching-type breakdown process once a critical electric field is reached. The critical field is on the order of 108 V/m. In a simplified device structure in which a phosphor thickness of 1 μm is subjected to a potential difference of 100 V, this critical field is reached in the phosphor. • The electrons that generate light must be able to fall into a localized ground state to cause light emission even in the presence of a high electric field in the phosphor layer. The process by which light is emitted is shown in Figure 8.5 [5]. The light emission process begins upon application of a voltage across the electrodes of the TFEL device, causing the phosphor layer to sustain a high electric field. This electric field allows electrons trapped in interface states at the interface layer, on the left-hand side of Figure 8.5, to tunnel into the conduction band of the phosphor layer. Once traveling in the conduction band, these electrons become “hot”, possessing a few electron volts of kinetic energy, and may impact-excite impurity centers causing the electrons in the ground state of the impurity center to become excited. When these excited electrons return to their ground state, light is emitted. Conduction band electrons may also excite other valence band electrons into the conduction band by an avalanche process. These further electrons may also excite impurity centers. Eventually, electrons in the conduction band reach the opposite interface layer on the right-hand side

318

Materials for Solid State Lighting and Displays

hν Phosphor layer in avalanching electric field

Dielectric layer Energy

Position

Figure 8.5 Energy band diagram of TFEL device showing electron trapping at phosphor/ dielectric interfaces (solid circles are occupied traps, hollow circles are empty traps), tunneling, impact excitation of activator, and re-trapping

of Figure 8.5 and get trapped at interface states there. No further excitation occurs until the electric field is reversed in polarity, allowing these trapped electrons to return to the left interface layer. Light emission occurs again during this process. A certain proportion of the conduction band electrons will return to the valence band, allowing avalanche processes to repeat every time electrons cross the phosphor layer. Light emission is pulsed, and an AC voltage is necessary for sustained operation of this light emission process. This voltage generally consists of a series of voltage pulses of peak voltage Vp , with alternating polarity, however any AC voltage such as a sine wave will cause light emission, provided that the critical electric field allowing electron release from traps and the subsequent avalanche process in the phosphor layer is attained. If a direct current (DC) voltage is applied to the TFEL device, only one flash of light can occur. The excess DC voltage will thereafter drop across an insulation layer preventing a sufficient electric field for electron transport from developing in the active semiconductor layer. The light emission from a typical device is shown in Figure 8.6 [5]. Here, the phosphor material is ZnS:Mn. ZnS is a well-known semiconductor having a valence band as well as a conduction band, with an energy gap of about 3.6 eV. The ZnS is doped with manganese which is well known to possess a localized electronic ground state and a localized excited state within the Mn2+ ion. The 3d shell of the Mn2+ ion has orbitals to accommodate 10 electrons but contains only 5 electrons and is therefore only half occupied. The spins of these 5 electrons are aligned in the ground state of the ion. A spin flip transition within the 3d shell forms an excited state which can decay to the ground state with the production of a photon. The spontaneous decay time is relatively long (in the millisecond range) due to the spin flip (phosphorescence) transition. (See Chapter 1.) In addition, the emission spectrum

Alternating Current Thin Film and Powder Electroluminescence 104

319

103

ZnS:Mn (0.65 μm) 1 kHz Sinusoidal wave drive

Luminous L (cd/m2)

L30 103

102

102

101 η

100

101

Threshold voltage 100 100

150

200 Voltage V (V)

Luminous efficiency η (Im/W)

L

Vth 250

10–1 300

Figure 8.6 Brightness–voltage characteristic of ZnS:Mn EL phosphor measured in a thin film device similar to that shown in Figure 8.1. Note that the drive frequency is 1000 Hz. The device exhibits a characteristic sharp threshold voltage Vt near 160 V due to the critical electric field required for electron transport through the phosphor layer. The luminous efficiency is also shown. Source: Ono 1995 [5]. Reproduced with permission of World Scientific Press

is sensitive to the host crystal because the crystal field perturbs the 3d orbitals. In ZnS, the Mn2+ ion is a yellow emitter, whereas in oxide hosts the emission color is typically green. The formal notation for the Mn2+ spin flip transition in a ZnS zinc blende host is 4 T (4 G) → 6 A (6 S). 1 1 Note the sudden onset of luminance at a specific threshold voltage Vt . This is inherent in an avalanche process, and is useful for display applications of TFEL. In a passive matrix flat panel TFEL display, partial voltages are present across “off” pixels. If they lie below the threshold voltage, these partial voltages do not cause light emission, leading to high contrast flat panel displays with passive matrix addressing. The ability of TFEL displays to function well without active matrix addressing was critical due to the lack of active matrix technology in the 1970s. Details of multiplexing drive methods of a TFEL display are well described in the literature [5] and will not be discussed further in this chapter. It is now appropriate to describe the purpose of the dielectric insulator layer(s) in the EL device. For drive voltages below the threshold voltage, the phosphor layer is an insulator. However, for higher drive voltages, the phosphor layer is an avalanching state, and therefore the voltage drop across this layer is clamped. Without the insulating layer, high currents could flow and damage the device. The dielectric layer limits the maximum current flow to

320

Materials for Solid State Lighting and Displays

the phosphor layer according to the general relationship: I=C

dV dt

(8.1)

where C is the effective dielectric layer capacitance per unit area and V is applied voltage (assumed to be larger than the threshold voltage). An equivalent circuit of a TFEL device is shown in Figure 8.7. A capacitor Ci is connected in series with two back-to-back zener diodes. Ci represents the effective dielectric capacitance per unit area of the TFEL device, and the back-to-back zener diodes represent the phosphor layer. The zener diode voltage represents the phosphor threshold voltage Vt(phosphor) and is determined by the critical electric field across the phosphor layer for avalanching to occur as well as the phosphor layer thickness. It is experimentally determined that the brightness of an EL device is substantially proportional to the amount of charge per unit area that flows across the phosphor layer during each voltage pulse, as well as to the frequency of these pulses. The charge flowing through the phosphor layer flows only for applied voltages that exceed Vt and is determined from: Q = Ci (V − Vt )

(8.2)

where Vt is the threshold voltage and V is the applied voltage. Clearly, a high capacitance dielectric is desirable to maximize the transferred charge through the phosphor of the EL device and hence its brightness. We know that Ci = (𝜀0 𝜀d )∕d

(8.3)

where d is the dielectric thickness and 𝜀d is the relative dielectric constant of the dielectric material. For this reason, high dielectric constant dielectric materials play an important role in high performance EL devices. In addition, the thickness of the dielectric layer needs to be minimized, but this thickness may be much higher for materials with high dielectric constants. For example, a 20 μm barium titanate-based layer having relative dielectric constant

Ci

Cp

}Zener diodes: breakdown voltage Vt (Phosphor)

Figure 8.7 Equivalent circuit of TFEL device showing insulator capacitance Ci (note that this represents the series capacitance of both dielectric layers if two layers are present), Cp the phosphor layer capacitance, and phosphor breakdown voltage Vt (phosphor) represented as the zener diode turn-on voltage in reverse bias

Alternating Current Thin Film and Powder Electroluminescence

321

of 2000 will actually provide a value of Ci that is almost 4 times higher than a 0.4 μm alumina layer having a dielectric constant of 11. Now, Q = Ci (V − Vt ) will be higher, and the maximum EL brightness will increase. Below the threshold voltage, the TFEL device, having capacitance CEL per unit area, consists of two capacitors in series, represented by Cp (the phosphor capacitance per unit area) as well as by Ci (the insulator capacitance per unit area). As a result, the applied voltage Va is divided as Vp across Cp and Vi across Ci according to the following equations: ( [ ]) VP = Va Ci ∕ Ci + Cp (8.4a) [ ( ]) (8.4b) Vi = Va Cp ∕ Ci + Cp with CEL = Ci Cp ∕(Ci + Cp )

(8.4c)

Once Va reaches Vt , the EL device capacitance CEL increases to Ci and current flows through the phosphor layer. This results in power dissipation and transferred charge. Figure 8.8 shows the relationship between transferred charge and applied voltage for an EL device operating below and above the threshold voltage. Note that below the threshold voltage, the relationship follows the simple equation Q = CV (Figure 8.8a). The Q − V plot is a straight line having slope CEL , since CEL represents the series combination of Ci and Cp , and there is no power consumption. Above threshold, power consumption will occur as charge flows across the phosphor layer. This results in an area inside the Q–V plot (Figure 8.8b). This area precisely determines the electrical energy dissipated in the TFEL device per cycle. There are now two different slopes in the Q–V plot. The smaller slope is the same slope as observed in Figure 8.8a. The larger slope is obtained when the electric field across the phosphor layer has reached the critical field, and therefore the voltage across the phosphor layer is clamped. Now the larger slope is equal to Ci or the insulator capacitance. As the voltage is increased further, the loop becomes larger as more charge is driven through the phosphor layer in each cycle. The additional applied voltage falls across the dielectric layer(s), and this dielectric must be capable of supporting the desired operating voltage. The total charge that flows through the EL device may now be evaluated as the change in Q = 2Ci (Va − Vt ), where Va is the applied voltage, as the voltage goes from −Va to +Va . Hence during one full cycle, the total charge that flows across the phosphor is: Qtotal = 4Ci (Va − Vt ) The energy dissipated in one cycle is the product: (Qtotal ) × (phosphor ) × (dphosphor ) where phosphor is the critical field in the phosphor layer and dphosphor is the phosphor layer thickness. Hence the power dissipated is: P = 4Ci (Va − Vt )(phosphor )(dphosphor )f

(8.5)

where f is the frequency of the applied voltage. A convenient way to measure these parameters is to use the Sawyer–Tower circuit method [5] to plot the charge–voltage curve. The

322

Materials for Solid State Lighting and Displays (a)

Q

V

(b)

Q

V

Figure 8.8 Q–V relation for EL devices operating (a) below and (b) above the threshold voltage. Note the abrupt change in slope above the threshold voltage when avalanching in the phosphor layer occurs

charge is derived from a small voltage across a sense capacitor Cs inserted in series with the EL device such that the voltage drop across the sense capacitor Vs is always much smaller (at least 100 times smaller) than the voltage drop across the EL device. See Figure 8.9. The x-axis of the oscilloscope is connected to the applied voltage, which is almost identical to the sample voltage, and the y-axis measures Vs . The oscilloscope is used in x − y mode. The shape of the oscillogram is therefore that of the QV loops (Figure 8.10). For a set of different applied voltages, the QV loops as well as the sample brightnesses in cd∕m2 are recorded. The area A inside the loop on the oscillogram multiplied by the frequency f of the applied voltage is the electrical power dissipated. Therefore, from Equation 8.5, the following information can now be obtained: • capacitance of EL device below threshold; • capacitance of EL device above threshold;

Alternating Current Thin Film and Powder Electroluminescence

323

x-axis of scope Applied voltage

CEL y-axis of scope Cs

Figure 8.9 Sawyer–Tower circuit showing the sense capacitor in series with the EL device. The sense capacitor is often chosen to be approximately 100 times larger than the EL device capacitance

VCs = Q/Cs ( y-axis of scope)

Va (x-axis of scope)

Figure 8.10 Sawyer–Tower output trace showing the charge versus applied voltage relationship for an EL device. The charge is derived from the voltage across a sense capacitor according to Q = Cs Vs and is plotted versus applied voltage Va

• • • • •

degree of symmetry of turn-on characteristic for positive- and negative-going voltages; threshold voltage; sharpness of threshold voltage; electrical power dissipation using loop area of QV loop; device efficiency in lumens per watt which may be obtained by dividing the optical power emitted by the EL device (in lumens) by the electrical power P dissipated (in watts).

The efficiency plot in Figure 8.6 may be obtained using the Sawyer–Tower method. Therefore, this is a powerful tool in analyzing and understanding EL device performance.

324

Materials for Solid State Lighting and Displays

8.4

Electroluminescent Phosphors

The first high performance phosphor material and still the material that has been studied the most is ZnS:Mn. ZnS is a II–VI semiconductor with an energy gap of 3.6 eV [5]. There are two structural variants, namely a cubic phase and a hexagonal phase. However, these two phases differ only in terms of second-nearest neighbors, and both phases are excellent EL phosphors. In either phase, each Zn2+ ion is coordinated by four S2− ions in a tetrahedral configuration. ZnS is a relatively stable sulfide host that may be stored in ambient atmospheric conditions without degradation. However, it does require moisture protection when it is subjected to high electric fields in an EL device. This may be due to the field-assisted dissociation of water to form OH- ions that hydrate the ZnS material. The EL phosphor ZnS:Mn is more correctly described as Zn(1−x) Mnx S with 0.005 < x < 0.02. The dopant Mn provides the observed yellow luminescence, and Mn as a 2+ ion substitutes effectively for Zn2+ ions. This is achieved since both Zn and Mn are 2+ ions and have similar ionic radii (ionic radius of Zn2+ = 0.74 nm, ionic radius of Mn2+ = 0.8 nm) and both bond in tetrahedral configurations. The concentration of the Mn2+ ion in the ZnS host is optimized by incorporating as much Mn as possible while maintaining isolated Mn2+ behavior. Too high a Mn2+ concentration results in concentrating quenching of the desired luminescence due to energy transfer between Mn2+ ions. Mn2+ , being a transition metal ion, has an unfilled inner shell or 3d shell, which gives rise to a ground state as well as an excited state. It is important to note that the electronic transition from excited state to ground state generates a photon, but it does not affect the charge on the Mn2+ ion, which remains at 2+ throughout this transition. The luminescent process is therefore not a charge transfer process in contrast to other well-known ZnS phosphors, such as cathode ray tube (CRT) phosphors ZnS:Cu and ZnS:Ag which rely on charge transfer processes in which an electron falls from a higher energy state elsewhere in the phosphor and is accepted by the Cu2+ or Ag2+ dopant. This results in a Cu1+ or Ag1+ state, and the luminescence results from the transfer of the electron to the metal ion, which is a trap-assisted electron–hole recombination process. This distinction in mechanism is important, as only luminescent centers that function without charge transfer generally exhibit good EL properties, including high efficiency and long life. When a high electric field is applied to the EL phosphor, the avalanche process takes place, and the hot electrons generated impact-excite the Mn2+ centers. This impact-excitation process is characterized by an impact excitation cross-section, which describes the success rate of hot electrons exciting the Mn2+ ion. Mn2+ has a large impact excitation cross-section in ZnS, and the applied electric field does not change the ionization state and therefore affect the efficient recombination process that takes place within the Mn2+ ion. Trap-assisted electron–hole luminescence processes are not useful for high field EL, since the electric fields established to cause avalanching, as well as the acceleration of electrons to become “hot”, sweep free carriers away and prevent their recombination. Therefore, electrons that normally recombine and generate photons in CRT phosphors, such as ZnS:Cu and ZnS:Ag, are driven back and forth between the phosphor/dielectric interfaces under EL drive conditions, and eventually may find their way back to the valence band via non-radiative trapping. For this reason, ZnS:Cu and ZnS:Ag are not useful as TFEL phosphors. Mn2+ ions may also generate luminescence in other non-sulfide EL phosphor hosts. An effective oxide phosphor Zn2−x Mnx SiO4 [3] relies on the same inner-shell radiative

Alternating Current Thin Film and Powder Electroluminescence

325

Table 8.1 Important sulfide and oxide EL phosphors showing their composition color, CIE coordinates and reported luminous efficiency. Note that in many cases a range of efficiency values is given, since various values are reported depending on the device structure and preparation conditions Phosphor ZnS:Mn ZnS:Tb SrS:Ce SrS:Ce, Eu BaAl2 S4 ∶Eu SrGa2 S4 ∶Eu Zn2 SiO4 ∶Mn Zn2 Six Ge1−x O4 ∶Mn ZnGa2 O4 ∶Mn Ga2 O3 ∶Eu Y2 O3 ∶Mn Yx Gay O3 ∶Mn Yx Gey O3 ∶Mn

Color

CIE coordinates

Efficiency (lm/W)

Reference

Yellow Green Blue-Green White Blue Green Green Green Green Red Yellow Yellow Yellow

0.5, 0.5 0.32, 0.6 0.19, 0.38 0.41, 0.39 0.135, 0.1 0.226, 0.701 0.2, 0.7 0.2, 0.7 0.08, 0.68 0.64, 0.36 0.51, 0.44 0.54, 0.46 0.43, 0.44

3–10 0.5–2 0.5–1.5 0.4 0.5–1.5 1–2 0.5–2 1–3 1–2 0.5–1 10 10 10

[5] [5] [5] [5] [10] [7] [11] [11] [11] [9] [12] [12] [12]

transition as in ZnS:Mn. The color of the luminescence is green instead of yellow, due to a change in the crystal field surrounding the Mn2+ ion. Nitride and fluoride host materials have also been identified, but EL performance has not been as successful to date. Other efficient EL luminescent centers exist in place of Mn. Eu2+ exhibits excellent blue luminescence in BaAl2 S4 phosphors. Eu2+ has a 4f–4d transition. Ce3+ exhibits bright green-blue luminescence in SrS phosphors, and Ce3+ has a 4f–4d transition. A number of other ternary host materials are available. For example, Zn1−x Mgx S:Mn phosphors exhibit a blue-shift compared with ZnS:Mn material, which gives this phosphor a peak wavelength below 580 nm [6]. SrGa2 S4 :Eu is a green EL phosphor [7] not the same as, but related to, BaAl2 S4 :Eu. Zn2 Si1−x Gex O4 :Mn represents a family of green oxide phosphors with higher brightness and efficiency and relatively lower processing temperature requirements compared with other oxide phosphors such as Zn2 SiO4 :Mn [8]. ZnS:Tb is also a bright green phosphor [5] that is unusual in that Tb is not readily soluble in the ZnS host due to the incompatible size and charge of the Tb3+ ion. A charge compensation coactivator, such as F or O, can play a role in improving the performance of these phosphors. Eu3+ exhibits bright orange-red luminescence in Ga2 O3 host material. Eu3+ has a 4f–f transition. This phosphor is unusual in that it works in an amorphous phase, demonstrating that crystallization is not a requirement for all EL phosphors. Eu3+ is not size-compatible with the Ga cations in the host material [9]. A list of some important EL phosphors together with their properties is shown in Table 8.1 [5, 7, 9–12].

8.5

Thin Film Double-Insulating EL Devices

Due to the commercial success of the device structure in Figure 8.1, it will now be discussed in more detail. It relies on thin film processing for all the layers, as well as the properties

326

Materials for Solid State Lighting and Displays

of glass substrates in terms of their smoothness, thermal stability, and transparency. Aluminosilicate glass used in active matrix liquid crystal display (LCD) manufacturing is an excellent substrate material. One key feature of this TFEL device structure is the reliance on high-quality, thin film dielectrics. These may be grown by RF sputtering or by ion-assisted deposition. From Equation 8.2 and Equation 8.3 it is clear that we wish to use dielectric materials with high dielectric constants, which also exhibit high breakdown voltages. Binary oxide dielectric materials include silica, alumina, silicon nitride, and yttrium oxide. These may be grown by RF sputtering or by electron-beam evaporation. See Table 8.2. With the exception of SiO2 , whose dielectric constant is too low, these materials are all suitable for EL dielectrics. Nevertheless, the desire for higher dielectric constant materials has motivated the development of more complex oxide and oxynitride dielectric materials. See Table 8.3. In the thin film double-insulating-layer TFEL device structure of Figure 8.1, the choice of dielectrics is limited. This comes about owing to the desire for self-healing behavior of the completed EL device. Micrometer- or submicrometer-sized weak spots in the dielectric or phosphor layers generally exist. These may occur due to imperfections in the layers such as pin holes, impurities, particulates, or substrate non-uniformities. The result is that these weak spots are susceptible to catastrophic dielectric breakdown at voltages that are required for the normal operation of the device. Self-healing behavior allows this highly localized breakdown to occur without causing overall failure of the EL device. The key is to allow at least one of the electrodes to fail locally around the defect, thereby isolating it from the remainder of the device (Figure 8.8). Note that heat is created at the defect as leakage current flows, and the weakest electrode opens like a protective fuse around the defect. Since the size of this open region is in the micrometer scale, it is not visible to the naked eye in a working EL display. Self-healing requires a suitable combination of dielectric layers Table 8.2 Dielectric properties of binary oxide dielectrics Oxide

Dielectric constant

Breakdown field (×108 V∕m)

SiO2 Al2 O3 Si3 N4 Y 2 O3

4.0 8 8 12

6 5 6–8 3–5

Source: Ono 1995 [5]. Reproduced with permission of World Scientific Press.

Table 8.3 Dielectric properties of complex oxide dielectrics SiAlON BaTa2 O5 SrTiO3 BaTa2O6 PbTiO3

Dielectric constant

Breakdown field (×108 V∕m)

6 22 140 22 150

7 3.5 1.5–2 3.5 0.5

Source: Ono 1995 [5]. Reproduced with permission of World Scientific Press.

Alternating Current Thin Film and Powder Electroluminescence

Aluminum

327

Damaged conductive region

Phosphor ITO Glass substrate

Figure 8.11

Self-healing due to aluminum evaporation around short circuit

and electrode materials in suitable thickness ranges. For example, if the dielectrics are damaged by heat before the electrode has a chance to open, then propagating breakdown will occur. Here, the heat causes a hot spot to travel along a random path within the EL device and the entire EL device will be destroyed. If the aluminum layer of Figure 8.11 were too thick, this could occur. If the aluminium were replaced by a high melting point metal such as tungsten, this might also occur. Note that the transparent ITO electrode in Figure 8.11 may no longer self-heal because ITO is a high melting point oxide material. In addition, the ITO is sandwiched between other materials that conduct heat away from it. Commercially successful EL devices to date have taken advantage of the self-healing properties of thin film electrodes. In addition, the choice of dielectric also determines the ability of self-healing to function. Unfortunately, the high dielectric constant dielectric materials such as BaTiO3 are chemically less stable than simpler dielectrics such as Al2 O3 , and catastrophic dielectric degradation may occur before the rear electrode is able to become an open circuit around the defect. A transparent variation of the TFEL device structure of Figure 8.1 has been commercially successful and continues to be manufactured. Both front and rear electrodes are transparent thin films (such as ITO). The achievement of virtually defect-free thin film growth enables these transparent TFEL structures to be reliable and rugged.

8.6

Current Status of TFEL

TFEL has a very small market share compared with more mainstream display technologies such as LCD and OLED. There are several factors behind this lack of TFEL adoption: 1. TFEL technology achieves lower efficiencies (0.5–10 lm/W) than current LEDs or OLEDs (10–200 lm/W). 2. TFEL does not offer full colour quality sufficient to meet current colour standards. From Table 8.1 it is clear than the most efficient TFEL materials are yellow emitters and a set of high efficiency red/green/blue TFEL materials is not available. Full color TFEL display prototypes using optical filters as well as fluorescent conversion materials to produce red and green light from a blue TFEL phosphor have been developed but have not been successfully commercialized. 3. TFEL does not use active matrix addressing. While this is a benefit it also puts an upper limit on resolution. This upper limit is in the range of hundreds of rows, but mainstream

328

Materials for Solid State Lighting and Displays

full HD displays now require 1080 rows and emerging standards such as 4K require 2160 rows. 4. TFEL displays require high drive voltages (over 100 V) which adds to driver cost. These limitations have relegated TFEL to display applications requiring very wide temperature ranges where LCD or OLED displays are unable to function reliably. A small market for transparent TFEL displays also exists.

8.7

Background of AC Powder EL

For the remainder of this chapter we will turn our attention to AC powder EL. AC powder EL devices were first discovered by Destriau [13] in 1936 using ZnS crystals activated with copper. Following further development by Sylvania in the 1950s this resulted in the typical structure of an AC powder ZnS EL device shown in Figure 8.12. The EL active phosphor layer consists of suitably doped ZnS powder with particle size of approximately 20 μm embedded in a dielectric, which also acts as a binder. This phosphor layer is sandwiched between two electrodes, at least one of which is transparent, and is supported by a substrate usually made from a flexible plastic sheet. The EL color of the film depends on the activator and coactivator in the ZnS phosphor. The most common ZnS phosphor used is the green-emitting ZnS:Cu which is co-activated with an additional impurity consisting of either Cl or Al. In this material, the Cu activator acts as an acceptor while Cl (or Al) acts as a donor. The emission of light occurs due to an electron dropping from the donor energy level and into the acceptor energy level. This is called a donor–acceptor pair emission process. The concentration of Cu added in the preparation process of these phosphors is 10−3 –10−4 gram per gram of ZnS, and is about one order of magnitude higher than that

Al rear electrode

Dielectric Zns:Cu, Cl phosphor Transparent electrode (ITO) Glass or plastic substrate

Figure 8.12 Typical structure of AC powder EL device. The EL active phosphor layer consists of suitably doped ZnS powder with particle size of approximately 20 𝜇m embedded in a polymer binder with thickness not much more than the phosphor particle diameter. This phosphor layer is sandwiched between two electrodes, at least one of which is transparent, and is supported by a substrate usually made from a flexible plastic sheet. The dielectric layer is a composite material with a high average dielectric constant and is generally made by embedding small grain barium titanate powder within a polymer binder

Alternating Current Thin Film and Powder Electroluminescence

329

Table 8.4 A set of powder phosphors known to exhibit AC powder EL Phosphor

Color

Reference

ZnS:Cu, Cl(Br, I) ZnS:Cu, Cl(Br, I) ZnS:Mn, Cl ZnS:Mn, Cu, Cl ZnSe:Cu, Cl ZnSSe:Cu, Cl ZnCdS:Mn, Cl (Cu) ZnCdS:Ag, Cl (An) ZnS:Cu, Al

Blue Green Yellow Yellow Yellow Yellow Yellow Blue Blue

14, 15 14, 15 16 17 17 17 14 14 18

added to ZnS phosphors historically used in CRTs. As discussed below, the excess Cu plays an import role in high-field AC powder EL in addition to acting as the activator. See Table 8.4 [14–18]. The embedding dielectric in the device of Figure 8.12 is an organic material with good dielectric properties such as cyanoethylcellulose, or low melting temperature glass [5]. In order to increase the stability and protect the EL device against catastrophic dielectric breakdown, a high dielectric constant insulating layer consisting of BaTiO3 powders dispersed in a polymer matrix is inserted between the EL active layer and the Al rear electrode.

8.8

Mechanism of Light Emission in AC Powder EL

The bipolar field-emission model, proposed by Fischer [19] is the accepted explanation for the light emission mechanism in AC powder EL materials. A careful study of the interior of ZnS:Cu, Cl particles using an optical microscope was performed. It was observed that the shape of the lighting-emitting region within a single EL particle takes the form of double tails joined by a line with shapes similar to tails of a comet as shown in Figure 8.13. On further observing the ZnS phosphor particles under the microscope, Fischer found that there were many dark segregations and precipitates inside the phosphor particles as shown in Figure 8.14 [20]. According to these observations, Fischer proposed the following model for the EL mechanism. ZnS EL powders are typically prepared by firing at high temperatures (1100–1200 ∘ C), where the hexagonal wurtzite phase of ZnS predominates. When the powders are cooled, there is a phase transition to the cubic zinc-blende structure. Excess copper that is beyond the solubility limit of copper in ZnS preferentially precipitates on defects formed in the hexagonal-to-cubic transformation with the reduction of copper solubility during cooling. The Cu forms thin embedded Cu2-x S needles in the crystal matrix (Figure 8.15a)[21]. Cu2-x S is known to be a p-type semiconductor with high conductivity. Between these Cu2-x S precipitates and ZnS powder, hetero-junctions are formed (Figure 8.15b) [19, 22]. When an electric field is applied to the phosphor particles, relatively high electric fields will be concentrated on the tips of Cu2-x S conducting needles (the effective tip radius is under 100 nm) compared with the average electric field. See Figure 8.16.

330

Materials for Solid State Lighting and Displays

E > Et

E = Et

Figure 8.13 Typical microscopic view of EL from ZnS:Cu, Cl particles. Double tails at threshold voltage and above the threshold voltage are illustrated [19] ZnS particle

Segregation

Field Direction

Figure 8.14 Phosphor particles containing dark segregations and emitting spots. Source: Chen and Xiang 2008 [20]. Reproduced with permission of John Wiley and Sons

Therefore, an applied field of 106 –107 V/m can induce a local field of 108 V/m or more. This electric field is strong enough to induce tunneling of holes from one end of the needle and electrons from the other to the ZnS:Cu, Cl lattice. The electrons are captured in shallow traps at Cl donor sites, while the holes are trapped by the Cu recombination centers (acceptor sites). When the field is reversed, the emitted electrons recombine with the trapped holes to produce EL. Figure 8.17 [22] shows the illustration of the basic principle of the bipolar field-emission model. Details of the trapping centers for ZnS:Cu, Al which are similar to those for ZnS:Cu, Cl are shown in Figure 8.18 [23]. Before excitation, the Cu (acceptor) is monovalent (1+), while the Al (donor) is trivalent (3+), so that charge compensation is realized in the lattice. Absorption A of the figure located at about 400 nm gives the characteristic excitation band of the center. When excited, Cu and Al become divalent (2+). The levels of Cu2+ (3d9 configuration) are split by the crystal field into 2 T2 and 2 E states, with 2 T2 lying higher

Alternating Current Thin Film and Powder Electroluminescence

331

ZnS powder grain Conductive Cu2–xS precipitates

(a) ZnS:CU, Cl

CUxS

ZnS:CU, Cl

Cl e e e e e e e e e e Cu Cl h h h h h h hh h

hv Cu

(b)

Figure 8.15 EL emission mechanism and schematic energy-band diagram of AC powder EL devices: (a) Cu2-x S needles [21]; and (b) energy-band diagram [19, 22]

Conductor

Insulator

Field lines

Figure 8.16 Conducting needle embedded in insulator. A uniform electric field is applied parallel to the needle. Geometrical field intensification occurs at the ends. Source: Fischer 1962 [19]. Reproduced with permission of The Electrochemical Society

332

Materials for Solid State Lighting and Displays ZnS particle

Conducting line Hole traps −

+

Electron traps

(a)

−

+

Recombination (b)

Figure 8.17 Illustration of the basic principle of the field-emission model. (a) upon field application, electrons and holes are ejected from the opposite ends of the conducting inclusion, where the field is intensified, into the ZnS lattice and to the relevant traps. Holes are trapped after a short path. Electrons can travel further and are also trapped. (b) Upon field reversal, trapped electrons recombine with trapped holes causing light emission. Simultaneously, other electrons and holes are field-emitted on the opposite two sides of the conducting inclusion. Source: Fischer 1963 [22]. Reproduced with permission of The Electrochemical Society. 4.0

B Al3+

Al2+

Photon energy (eV)

3.0 A 2.0

2T

1.0

Cu+

Cu2+

C

2

D 2E

0 Valence band (a)

(b)

Figure 8.18 Energy levels and absorption transitions of ZnS:Cu, Al phosphor before excitation (a) and during excitation (b). Source: Suzuki and Shionoya 1971 [23]. Reproduced with permission of Journal of the Physical Society of Japan

Alternating Current Thin Film and Powder Electroluminescence

333

in the zinc-blende structure. In the process of the relaxation due to the reverse of the electric field, the electron trapped by Al2+ would recombine with the hole trapped by Cu2+ and luminescence is generated. As a result, Al and Cu become monovalent and trivalent, respectively. In summary, according to the bipolar field emission model, EL emission from a ZnS-powder-EL device is caused by the radiative recombination of electron–hole pairs via donor acceptor pairs. Electrons and holes are injected from CuS precipitates to the donors and acceptors, respectively. The injection process is caused by tunneling of the relevant carriers and is enabled by the high electric fields generated at the tips of conductive CuS precipitates. The instantaneous field-emission current I through the Cux S-ZnS contact follows the Fowler–Nordheim equation: ( ) 3 AE2 BW 2 I = 3∕2 exp − (8.6) E W where A and B are constants, E is the field strength, and W is the work function. In this case, W corresponds to the energy difference between the electron affinity of ZnS and that of Cux S. Experimental evidence has been provided by Ono et al. [24] to support the validity of Fischer’s model, by carefully examining ZnS:Cu phosphor particles showing EL using transmission electron microscopy (TEM). Black specks in the shape of narrow needles with diameters of 20–40 nm along the boundaries of micro-twin crystals inside a ZnS particle were observed using TEM. The specks were believed to be the Cu2 S, further confirmed by measuring the wavelength of characteristic X-rays emitted from the precipitates. Cu2 S is well known to be a p-type semiconductor with high metallic conductivity. Therefore, these observations are consistent with Fischer’s predictions.

8.9

Electroluminescence Characteristics of AC Powder EL Materials

EL is observed when an AC voltage of about 100–200 V corresponding to an electric field of order of 104 V∕cm is applied across the electrodes of the device. Luminance–voltage characteristics of a typical EL device are shown in Figure 8.19 [25]. The observed dependence of the luminance L on the applied voltage V is expressed [25] by: [ ( ) ] V 1∕2 (8.7) L = L0 exp − 0 V The parameters L0 and V0 depend on the particle size of the phosphor, the concentration of the powder in the dielectric, the dielectric constant of the embedding medium and the device thickness. It has been established empirically that one of the key parameters affecting EL characteristics is the particle size, and a critical trade-off between the EL efficiency and the operational lifetimes, defined by the time when the luminance becomes one half of the initial value, exists as follows. The efficiency increases approximately in proportion to d−0.5 where d is the particle size. However, the operational lifetime, decreases approximately

334

Materials for Solid State Lighting and Displays

f = 400 Hz 102

Luminance L (cd/m2)

η

101

100

Luminous efficiency η (lm/W)

101

L

100

0

100

200 300 Voltage V (V)

400

Figure 8.19 Typical luminance–voltage and efficiency–voltage characteristics of an AC powder EL device. Source: Ono 1995 [5]. Reproduced with permission of World Scientific Press

in proportion to d. In addition, the luminance increases with frequency in the frequency range of ∼100–10 kHz. Luminance of 100cd∕m2 has been achieved for devices driven at a frequency of 400 Hz and a voltage of 200 V [25]. A typical voltage dependence of EL efficiency 𝜂 is shown in Figure 8.19. Typical values of the efficiency are 1–10 lm∕W. The efficiency increases initially with increased applied voltage up to a saturation value, but then decreases gradually with further increases in voltage. The EL efficiency dependence on the voltage V is expressed by 𝜂 = L0.5 V −2 . The maximum efficiency is obtained at a voltage well below the highest luminance level.

8.10

Emission Spectra of AC Powder EL

Typical emission spectra of AC powder EL devices are shown in Figure 8.20 [25]. Emission colors depend on the luminescent centers incorporated in the phosphors. When the ZnS lattice is activated with Cu (activator) then various emission bands form depending on the energy level of the coactivator. The combination of Cu and Al in ZnS:Cu, Al produces a green emission color. The combination of Cu and Cl in ZnS:Cu, Cl gives blue and green emission bands, their relative intensity depending on the relative amounts of Cu and Cl. ZnS:Cu, I yields blue emission. It should be noted that ZnS:Cu, in which no coactivators are incorporated, shows a red emission. By further incorporating Mn2+ ions into ZnS:Cu, Cl phosphors, the resultant ZnS:Cu, Mn, Cl shows yellow emission due to energy transfer to Mn2+ ions.

Alternating Current Thin Film and Powder Electroluminescence

EL intensity (a.u.)

ZnS:Cu, Cl

ZnS:Cu, Al

ZnS:Cu, Mn, Cl ZnS:Cu

ZnS:Cu, l

400

335

500

600 Wavelength (nm)

700

Figure 8.20 AC Powder EL spectra of several ZnS phosphor materials. Source: Ono 1995 [5]. Reproduced with permission of World Scientific Press

8.11

Luminance Degradation

Limited lifetime due to luminance degradation as a function of operating time is a key issue for the application of AC powder phosphor EL. Lifetime is defined as the operating time over which the luminance decreases to one half of the initial value. Figure 8.21 shows a typical example of EL light output versus time [26]. The degradation rate depends on driving conditions (such as frequency and luminance levels) and on environmental conditions, especially temperature and humidity. The luminance decay with time is usually expressed by: L = (1 + 𝛼t)−1 (8.8) L0 where 𝛼 is a constant roughly proportional to the driving frequency [25]. It is therefore not possible to realize long lifetimes at high luminance.

Ratio L/L0 (%)

100 condition: 20°C, 70% power suppy: inverter 200 V, 400 Hz 50 L/L0 = (1 + αt)−1 α = 4.5 × 10−4 h−1

0

500

1000

1500

2000

2500

Time t (h)

Figure 8.21 Typical luminance maintenance curve of AC powder EL device. Source: Howard 1981 [26]. Reproduced with permission of Elsevier

336

Materials for Solid State Lighting and Displays

According to Fischer’s bipolar field-emission model, the degradation may be related to the diffusion of copper in the ZnS lattice. As mentioned above, CuS precipitate needles (dislocation lines) play a crucial role in EL emission. However, the tips of the CuS-decorated imperfection lines can be compromised by drift or diffusion of copper ions under the influence of the high AC field applied to the needle tips [21]. This is believed to be the leading cause for luminance degradation as a function of operating time.

8.12

Moisture and Operating Environment

Another challenge for powder EL is that it is sensitive to the moisture and the operating environment. The following reaction occurs between ZnS and water to produce SiO2 : ZnS + 2H2 O → SO2 + Zn + 2H2 Through this reaction, sulfur escapes from the ZnS phosphor generating sulfur and zinc vacancies in the phosphor. It has been reported that the luminance of the ZnS phosphor deteriorates when the number of sulfur vacancies increases [27]. For this reason, two methods to prevent degradation have been developed. One method uses moisture blocking polymer film such as fluorocarbon film [28] to package the entire device. Another very successful method is to coat each phosphor particle individually with moisture impervious oxide coatings [29–32].

8.13

Current Status and Limitations of Powder EL

Powder EL lamps and information displays historically suffer from shortcomings including low luminance, short useful operating life, poor visibility in normal room light, and no visibility under high ambient light. The latter is a result of the powder nature of the EL devices which causes high ambient light scattering and therefore limited contrast. In addition, the high voltage requirement may add cost, although in certain applications AC line voltage can directly drive AC powder EL devices without the need for any power supply. Operation at temperatures much above room temperature accelerates luminance deterioration presumably due to the enhanced mobility of copper ions at higher temperatures. Nevertheless, the ability of powder EL to form rugged flexible plastic substrate light sheets is unique among light emission materials and devices. This leads to a number of niche markets including advertising and safety lighting products.

8.14

Research Directions in AC Powder EL and TFEL

Some more recent research in the field of powder EL is summarized in review papers [33, 34]. Investigations of powder phosphor particle alignment and surface properties have been reported [35, 36]. Work on binder optimization for flexibility and thinness has been reviewed also. Alternative nanomaterials proposed to act in place of CuS as electric field enhancement in donor–acceptor pair luminescent materials [37] could potentially allow for longer life

Alternating Current Thin Film and Powder Electroluminescence

337

performance although much remains to be done in order to achieve commercially meaningful results, and significant industrial interest in powder EL research is not evident. Materials under investigation include carbon nanotubes [38] and silicon carbide whiskers [39]. In the field of TFEL carbon nanotubes have also been used to demonstrate the achievement of locally enhanced electric fields [40]. Also, a red-emitting perovskite ((Ca0.6 Sr0.4 )0.997 Pr.002 )TiO3 has been shown to produce red EL at low voltages [41]. LED and OLED devices have extensively displaced AC EL as a light source in both lighting and display applications. Nevertheless, the ability for AC powder EL devices to produce light in low cost semiconductor materials without direct electrical contact to the semiconductor is unique and could lead to a new era in this field if materials and manufacturing methods can improve.

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

Russ, M.J. and Kennedy, D.I. (1967) J. Electrochem. Soc., 114, 1066. Wu, X., Carkner, D., Hamada, H. et al. (2004) SID Digest, 35, 1146. Minami, T., Kobayashi, Y., Shirai, T. et al. (1991) Jpn. J. Appl. Phys., 30 (T), L117. Xiang, Y., Kitai, A.H. and Cox, B. (2005) J. Soc. Inform. Display, 13, 493. Ono, Y.A. (1995) Electroluminescent Displays, World Scientific Press, Singapore. Mikami, A. (1997) Proc SID Int. Symp. Dig, vol. 28, p. 851. Yano, Y., Oike, T. and Nagano, K. (2002) Proc. EL2002, Gent, Belgium, p. 225. Xiao, T., Kitai, A.H., Liu, G. and Nakua, A. (1997) SID Digest, 28, 415. Stodilka, D., Kitai, A.H., Huang, Z. and Cook, K. (2000) SID Int. Symp. Dig., 31, 11. Xin, Y., Hunt, T. and Acchione, J. (2004) SID Digest, 35, 1138. Minami, T. (1998) Ex. Abst., The 4th International Conference on the Science and Technology. of Display Phosphors, Bend, OR, USA, September 14–17, 1998, p. 195. Minami, T. (2002) Proc. EL2002, Gent, Belgium, p. 219. Destriau, G. (1936) J. Chem. Phys., 33, 587. Destriau, G.J. (1937) Chim. Phys., 34, 327. Jaffe, P.M. (1961) J. Electrochem. Soc., 108, 711. Gobrecht, H. and Gumlich, H.E. (1956) J. Phys. Radium, 17, 754. Thornton, W.A. (1958) Bull. Am. Phys. Soc., 3, 233. Gobrecht, H., Gumlich, H.E., Nelkowaki, H. and Langar, D.Z. (1957) Z. Phys., 149, 504. Fischer, A.G. (1962) J. Electrochem. Soc., 109 (11), 1043. Chen, F. and Xiang, Y. (2008) Luminescent Materials and Applications, John Wiley & Sons, Ltd, Chichester. Kitai, A.H. (1993) Solid State Luminescence, Chapman & Hall, London, p. 209. Fischer, A.G.J. (1963) Electrochem. Soc., 110, 733. Suzuki, A. and Shionoya, S. (1971) J. Phys. Soc. Jpn., 31, 1455. Ono., Y., Shiraga, N., Kadokura, H., and Yamada, K. (1990) Electron. Inform. Commun. Eng., Tech. Rep. 89(378). Shionoya, S. and Yen, W.M. (1998) Phosphor Handbook, CRC Press LLC., New York. Howard, W.E. (1981) Proc. Soc. Display, 22, 47. Hirabayashi, K., Kozawaguchi, H. and Tsujiyama, B. (1983) J. Electrochem. Soc., 130, 2259.

338

Materials for Solid State Lighting and Displays

28. Hirabayashi, K., Kozawaguchi, H., and Tsujiyama, B. (1985) Kenkyu Jitsuyoka, Hokoku, 34: 525 (in Japanese). 29. Sigai, A.G. (1991) US Pat. 5,051,277. 30. Yan, S., Maeda, H., Hayashi, J.-I. et al. (1993) J. Mater. Sci., 28, 1829. 31. Budd, K.D. (1995) US Pat. 5,418,062. 32. Budd, K.D. (1999) US Pat. 5,958,591. 33. Withnall, R., Silver, J., Harris, P.G. et al. (2011) J. SID, 19 (11), 798. 34. Bredol, M. and Dieckhoff, H.S. (2010) Materials, 3, 1353. 35. Silver, J., Ireland, T.G., Harris, P. et al. (2013) SID Digest, 44, 224. 36. Sychov, M.M., Bakhmet’ev, V.V., Nakanishi, Y. et al. (2003) J. Soc. Inform. Display, 11, 33. 37. Chen, F. and Kitai, A.H. (2008) J. Luminescence, 128, 1856. 38. Bae, M.J., Park, S.H., Jeong, T.W. et al. (2009) Appl. Phys. Lett., 95, 071901. 39. Wagstaff, B. and Kitai, A.H. (2015) J. Luminescence, 167, 310. 40. Chen, K.-F., Wang, F.-H., Chien, Y.-H. et al. (2011) IEEE Electron Device Lett., 32, 5. 41. Takashima, H., Shimada, K., Miura, N. et al. (2009) Adv. Mater., 21, 3699.

Index

A 3M, 62 absorption edge, 22, 23 accelerating charges, 1–5 acoustic cavitation, 52 acridine, 189, 190 ACTFEL devices, 33 activator emission, 44–5 activators Ce3+ activated phosphors, 97, 99–101 co-activation with praseodymium, 101–6 Yx Gd1−x AG:Ce, 106–10 Eu activated phosphors, 97, 112 luminescence spectra of, 142–64 in nitride and oxynitride phosphors, 136–7 active-matrix addressing, 204, 205, 319, 327–8 Active Matrix-QLED displays, 68 AC voltage, 33, 318 AC powder EL, 328–37 aggregation, 192–3, 199, 201 aggregation-induced emission (AIE), 201–2, 250 aggregation quenching, 250 Agilent Technologies, 284–5 Al2 O3 –AlN binary system, 162 (Al,Ga)As material system, 282, 284 (Al,Ga)InP material system, 282–3, 288 broadening of emission spectra, 296 LEDs, 297, 298, 299 efficiencies, 294, 295

external quantum efficiencies, 294, 297 light extraction efficiency, 291, 292, 293 light output v current characteristics, 296, 300 Alivisatos group, 66 alkaline earth metal oxo-nitride phosphors, 120–1 alkaline earth metal silicate phosphors, 111–12 alkaline earth metal sulfide based phosphors (MS:Eu2+ ), 113–14 alkaline earth metal thiogallate phosphors (MGa2 S4 :Eu2+ ), 114–17 alloy nitridation, 140–1 AlN, 143, 156 preparation of, 140, 141 AlN:Eu,Si, 145–6 [gamma]-Alon, 162–4 Alq3 (tris(8-hydroxyquinolinato) aluminium), 187, 188, 193, 243 doped with aromatic dyes, 194 in QD-LEDs, 216 alumina dielectrics, 326, 327 Amazon Kindle Fire HDX, 62 ammonothermal synthesis, 141–2 AMOLED displays, 73, 205–7 compared to other devices, 68, 208, 209 fabrication of, 207–8 in smartphones, 213 ANSI standard for solid state lighting, 241–2

Materials for Solid State Lighting and Displays, First Edition. Edited by Adrian Kitai. © 2017 John Wiley & Sons Ltd. Published 2017 by John Wiley & Sons Ltd.

340

Index

anthracene, 189, 190, 191 anti-Stokes shift/upconversion, 43 antisymmetric wavefunctions, 13, 14–15, 16 aromatic dyes, fluorescent, 194, 195 aromatic rings, rotation of, 201 atomic orbitals dipole radiation from, 5 in TFEL, 314 Auger recombination, 41, 45–6, 47 and efficiency droop, 124 in LEDs, 275–6 on light output v current curve, 296, 298, 300 in QD-LEDs, 218 in QLEDs, 72 avalanching, 314 in TFEL devices, 317–18, 324 on Q-V plot, 321, 322 threshold voltage, 319, 320 B BaAl2 S4 :Eu phosphors, 325 Ba2 AlSi5 N9 , 143, 150–1 backlight, 172–3, 305–6 using QDs for, 54–5, 60–1 see also BLU (backlight units) Ba2 LiAlSi7 N12 :Eu2+ , 167, 168 Ba[Li2 (Al2 Si2 )N6 ]:Eu2+ phosphor, 121, 167–8 band edge recombination/emission, 41, 42–3 band gap, 31–2, 48, 275 and changes in QD size, 36–7, 38 and HOMO-LUMO difference, 40, 184 and organic materials in OLEDs, 185–6, 188 and quantum confinement effects, 36–7, 38–41 values for selected semiconductors, 39 band model, p-n junction diode, 28–9 band-to-band transitions, 19–23 barium sulfide, 31 barium titanate, 315, 316, 317, 320–1, 327 Ba5 Si11 Al7 N25 , 143, 151–2

Ba2 Si5 N8 , 141, 156 BaSi7 N10 , 140, 143 BaSi2 O2 N2 , 144, 148, 150, 155, 171 (Ba1−x−y Srx Cay )SiO4 :Eu2+ based phosphors, 111 (Ba,Sr)2 Si5 N8 :Eu2+ phosphor, 118, 287 Be(bq)2 , 188 benzene, 189, 190, 193 bimolecular radiative recombination, 275 biolabelling, QDs in, 79 biological applications of QDs, 35, 49, 78–81 birefringence, 243 1,3-bis(N-carbazolyl)benzene (mCP), 193, 194 2,6-bis(N-carbazolyl)pyridine (26mCPy), 193, 194 bis(10-hydroxybenzo[h]quinolinato) beryllium (Be(bq)2 ), 188 blackbody radiators, 26–7, 240, 285–6 blinking effect, 46–7 Bloch functions, 7 blocking layers, 233, 234 BLU (backlight units), 54 blue InGaN LED BLU, 54, 55, 56, 58 concerns about, 172 direct-lit BLU, 60 and edge optics, 61 see also backlight blue color and Ce3+ , 155 and coumarin moieties, 196 phosphorescence problems with, 257 blue emitters, 287 commercial fluorescent, 212 Eu2+ doped phosphors for, 145–7 in host-guest systems, 193 LEDs, 92, 99, 136, 278 OLEDs, 212, 251, 261 aggregate-induced emission materials, 202 deep blue, 193 for white light, 235, 236 P-OLEDs, 192 QD-LEDs, lifetime of, 220

Index

blueshift, 194, 196 in metal chelates, 188 and QDs, 36, 38 Blu-rayTM optical storage, 283 Bohr magneton, 16 bottom-emitting OLEDs, 235, 236, 246–7 bottom-up synthesis for QDs, 50–3 bound excitons, 9, 17, 217 brightness, 239 of LEDs, 92, 93 of OLEDs, for displays, 212 perception of, 23–4, 238 brightness-voltage characteristic, 31 BSSNE,2–5–8 phosphor, 118 5-BTMPS, 202 BTPE-PI, 202 C C47, 195, 196 C540 (aromatic dye), 194, 195 C545P, 196 CaAlSiN3 , 119–20, 156, 159, 169 in backlights, 172, 173 preparation of, 140, 141–2 product success, 155 structure and properties, 143, 152–3, 154 in white LEDs, 170, 171, 287 cadmium-free QLEDs, 70–1 CAMFR (software tool), 248 candela, 23–4 capping for OLEDs, 247–8 for QDs, 47–8 carbazole, 189, 190 carbon-carbon double bonds, 186 carbothermal reduction and nitridation, 140, 141 carrier transport properties, 193 CaS:Eu, 113, 114, 122 Ca-[alpha]-sialon, 140, 141, 144, 150, 156 Ca-[alpha]-sialon:Eu2+ , 150, 151, 155 in white LEDs, 169, 170, 171 Ca-[alpha]-sialon:Yb2+ , 110–11, 171

341

CaSiN2 , 156, 159–60, 171 CaSi2 O2 N2 , 144, 148, 155 (Ca1−x−y Srx Euy )AlSiN3 , 141 CaSrSi5 N8 , preparation of, 141 cathodoluminescence (CL), 32, 33 by alkaline earth metal thiogallate phosphors, 114 by YAG:Ce3+ phosphor, 99 cavity effect, 244 CBP (4,4’-bis(N-carbazolyl)-1, 1’-biphenyl), 67, 193, 194, 219 CC2TA, 200, 201 CCT (correlated color temperature), 27, 240–2, 286 of white LEDs, 170, 171, 172, 302 Cd-based QDs, 36, 37, 214, 215 applications and restrictions, 54, 62 in QD-LEDs, 215, 216, 218, 219 in QLEDS, 67–70 regulation and environmental issues, 75 CdSe, 37, 215, 216 CdSe/CdS, 216, 217 CdSe/CdS-PPV bilayer, 216 CdSe/CdS QD-LEDs, 216, 219 CdSe/CdS/ZnS graded shell QDs, 215 CdSe/CdS/ZnS QD-LEDs, 218 CdSe QDs, 36, 37, 214 in QD-LEDs, 215 regulation and environmental issues, 75 CdSe/ZnS, 214, 217 CdS QDs, 36, 51, 215 CdS/ZnS semiconductor, 217 Ce3+ in phosphors, 97, 99–101, 136–7, 169 co-activation with praseodymium, 101–6 nitride and oxynitride, 155–60 in TFEL devices, 325 cell/mobile phones, 63, 65, 205 see also smartphones cellular imaging, 80 ceramic sheet dielectric EL, 316 cerium (Ce) see Ce3+ Changhong, 63

342

Index

charge Q-V plot for TFEL devices, 321, 322 radiation from, 1–5, 6–7 chemical etching, 49, 207, 208 chemical methods of self-assembly, 50–2 chemical vapor deposition (CVD), 53 chemiluminescence, 32 chip-on-board (COB LEDs), 290, 291 chromaticity of LEDs, 297, 302 see also CIE diagram CIE, 237 standards for white light, 286, 295–6, 303–4, 305 CIE diagram, 25–6, 55, 56, 240, 241 for LEDs, 94, 100, 101–2, 299 for OLEDs with PtL2 Cl single emitters, 258 SrGa2 S4 :Eu2+ phosphor, 116 TFEL devices, 325 white light on, 242 CL (cathodoluminescence), 32, 33, 99, 114 COB LEDs, 290, 291 cold cathode fluorescent lamps (CCFLs), 172 cold white see cool/cold white colloidal QDs, 213–20, 235 see also QDs color and nitride/oxynitride phosphors, 137, 142 perception of, 25, 238 and QD sizes, 213 RGB color patterning in OLEDs, 203–4, 207 in QD-LEDs, 217 shift over lifetime, 307 and TFEL devices, 325, 327 vivid, with OLEDs, 63 color conversion phosphors, 91–125 metal fluoride, 121–2 metal nitrides, 117–20 metal oxide, 99–113 metal oxo-nitrides, alkaline earth, 120–1

metal sulfide, 113–17 multi-phosphor pcLEDs, 122–3 outstanding problems, 98 requirements, 95–7 synthesis, 99 using LEDs without, 93–5 white light generation, 97–8 color gamut for an average person, 240 with OLEDs, 63 of phosphor-converted white LEDs, 172, 173, 305–6 of QDs, 54, 55, 56 of QLEDS, 64, 68, 69 Color IQTM edge optic component, 62 color matching functions, 240 color rendering see CRI color space, 26, 27, 54 color space diagram see CIE diagram color temperature see CCT computer monitors, QDs in, 53, 62 see also displays COMSOL (software tool), 248 conduction band, 7, 19–23, 31–2 in LEDs, 275 p-n junction diode, 28 conduction band-level-to-acceptor-level recombination, 275 conjugated polymers, 191–2 conjugation and aggregate-induced emission, 201 in conjugated polymers, 192 in phosphorescent emitters, 197 contrast perception, 238 cool/cold white, 240, 241, 286 LEDs for, 169, 301, 303 core-shell systems, 48–9, 214–15, 220 coumarin moiety, dopants from, 196 CRI (color rendering index), 27, 74, 96, 241 and QLEDs, 74–6 of white LEDs, 97, 98, 169–71, 302, 303 high CRI, 287, 303 crystal field splitting, 136, 137, 155

Index

crystal structure see structure CuInS2 , 75 current-voltage characteristics of p-n junction diodes, 29–30 staircase, for quantum dots, 35 cyanostilbene, 201, 202 4CzIPN, 200, 201, 259, 260

D D-65, 286 dark states, 43–4 DCM2 aromatic dye, 194, 195 DCM aromatic dye, 194, 195 DCTJB, 194, 195 deep charge traps, 185 defect recombination/emission, 41, 43–4 defect states, 43–4 dendrimers, 267 density of states (DOS), 34–5, 36 Dexter energy transfer, 19, 185, 193, 253, 255 N,N’-dicarbazolyl-4,4’-biphenyl (CBP), 67, 193, 194, 219 dielectrics OLED capping layer, 247–8 in TFEL, 314, 315–16, 317, 318 double-insulating device, 326–7 purpose of, 319–21 7-(diethylamino)-4-(methyl)-coumarin (C47), 195, 196 diffusion current, p-n junction diode, 28–9, 30 diffusion of luminescent materials, 210 diodes for ESD protection, 291 LEDs see LEDs OLEDs see OLEDs p-n junction, 28–30 polymer-based PLEDs, 263–9 P-OLEDs, 191–2 quantum dot light emitting see QD-LEDs; QLEDs dipole radiation, 4–5, 8, 9, 15–16

343

dipoles, radiation from, 4–7, 19 direct band gap semiconductors, 8, 9, 276, 277 absorption edge, 22, 23 band gap energy, 31–2 photon emission rate, 19 direct-lit LCD module, 60 display addressing approaches, 204–7 displays, 305–6 OLEDs in, 203–13 top-emitting structures, 232 QDs in, 53–73 integration of quantum dots, 57–61 market and product overview, 53–4 mass produced products available now, 62, 63 performance and energy advantages, 54–7 to replace LED phosphors, 306 TFELs in, 319 see also flat panel displays DMAC-DPS, 200, 201 Dolby Maui PRM, 62 dominant wavelengths of LEDs, 296, 297, 299 donor-acceptor emitters, 200 donor/acceptor-type traps, 314 donor-level-to-valence-band recombination, 275 DOS (density of states), 34–5, 36 double bonds, carbon-carbon, 186 double-heterojunction (DH) LEDs, 278, 279 double-insulating TFEL devices, 314, 315, 325–7 down-conversion, 73–4 fills green gap in LEDs, 296 nitride phosphors for, 169 quantum dots for, 76–8 in white LEDs, 286–8, 298–306 drift current, p-n junction diode, 28, 29 drug delivery systems, QDs in, 78 dry processes in fabrication, 208, 210 DTC-DPS, 200 (dtpb)Cu(I)Br, 200, 201

344

Index

E edge-lit displays, 57, 60 edge optic integration of quantum dots, 57, 59, 61 in mass produced products, 62, 63 effective mass approximation model (EMA), 38–9 efficiency external see EQE internal see IQE of laser diodes, 124 of laser diode SSL systems, 124, 125 of LEDs, 93, 295 pcLEDs, 97 and white light generation, 93, 135–6, 169 of light sources, 25 luminous see luminous efficiency of OLEDs, 185, 196, 203, 204 outcoupling, 245–6, 247, 249–50, 252–3 power conversion see PCE of QD-LEDs, 216, 218 of QDs, 123–4 colloidal QDs, 215 in displays, 56–7 of QLEDs, 67, 68 early devices, 66 InP-based QLEDs, 70 and sub-band gap emission, 65 ZnSe-based QLEDs, 70, 71 quantum see quantum efficiency radiative see radiative efficiency of TFEL devices, 325, 327 efficiency droop, 124 efficiency roll off, 253 EIL (electron injection layers), 233 EL (electroluminescence), 32, 33, 248–9 in anthracene, 189 devices, advantages of, 33 and display applications of quantum dots, 62–73 first reported, 91 with OLEDs, 184, 185–6 white OLEDs, 231 in QDs, 75–6

electrical excitation, 18 electric discharge machining, 207 electric field around point charges, 1–4 in electroluminescence, 33 and light emission in TFEL, 317–18, 321, 324 in p-n junction diodes, 28 electrodeposition in mask fabrication, 207, 208 electroluminescence see EL electroluminescent materials organic, 184, 185–6 phosphors, 324–5 electromagnetic radiation see light; photons; radiation electron beam lithography, 49, 50, 208 electron diffusion current, p-n junction diode, 28 electron drift current, p-n junction diode, 28 electron-hole recombination, 19–23, 41–6, 275 in band edge emission, 41, 42 dipole radiation from, 5 in electroluminescence, 33 in LDs, 275–7 in OLEDs, 184, 185, 186, 233 p-n junction diodes, 28–30 see also radiative recombination electronic transitions in nitride/oxynitride phosphors, 137, 142 Eu2+ doped, 142–3 in [beta]-sialon:Pr3+ , 162 electron injection layers (EIL), 233 electron mobility, 187, 191, 193 electrons Auger, emission of, 41, 45–6 in cathodoluminescence, 33 conduction band, 20 and dipole radiation, 4 and excitons, 7–10 molecular excitons, 16–17, 19 in OLEDs, 184 and photon emission, 5–7

Index

in semiconductors, 31–2 spin see spin in thin film electroluminescence (TFEL), 313, 317–18 two-electron atoms, 10–16 see also electron-hole recombination electron transport layer see ETL electrostatic discharge protection (ESD), 290–1 EMA (effective mass approximation) model, 38–9 emission layer see EML (emission layer) EML (emission layer) in OLEDs, 184, 185, 186, 233–4, 250 phosphorescence, 251, 255–6 in QD-LEDs, 213 in QLEDs, 64, 65, 66 energy band gap, 31–2, 275 and band transitions, 19–23 and HOMO-LUMO difference, 184 of QDs, 36, 37, 38 electric field, 1 of emitted photons, 184 exciton, of host and guest, 193 from an accelerated charge, 4, 5 loss processes, for molecular excitons, 18–19 separating singlet and triplet states, 18, 199–200 energy band diagram of OLEDs, 184, 185, 186 multi-layer OLED, 233 of TFEL device, 318 energy density due to fields, 2 energy flow per unit area, 4 energy levels of CdSe/CdS-PPV bilayer QD-LED, 216 and excitons, 7–10 molecular excitons, 16–17 Energy Star program, 241–2, 287, 307 energy transfer enhanced QD-LEDs, 219 Environmental Protection Agency, US, 241, 287, 307 epi-up configuration, 284, 285

345

EQE (external quantum efficiency), 64, 239 of LEDs, 292, 297 (Ga,In)N material system, 294–5 vs wavelength, 93, 294, 296 of OLEDs, 196, 197, 199, 249–50 aggregate-induced emission materials, 202 and phosphorescence, 252–3 TADFs, 201 of QD-LEDs, red, 215–16 of QDs, colloidal QDs, 215–16 of QLEDs, 67–8, 69, 70 equivalent circuit for TFEL devices, 320–1 ESD (electrostatic discharge protection), 290–1 etching in QD synthesis, 49–50 ETL (electron transport layer) metal chelates for, 187 in OLEDs, 184, 185, 186, 233 in QD-LEDs, 213, 216, 217–18 in QLEDs, 64, 65, 66, 67 E-type delayed fluorescence, 258 Eu (europium) phosphors with, 97, 136–7, 169 metal oxide, 111, 112–13 metal silicates, alkaline earth, 111, 112 nitrides, 118, 142–55, 166–7 oxynitride, 142–55 supply risk and importance of, 73–4 in TFEL devices, 325 europium see Eu (europium) excimeric emission bands, 236, 256–7 excited electrons and light emission, in TFEL, 317–18 and luminescence, from phosphors, 31, 32 excited states in Mn, 324 and molecular excitons, 16–19 radiation from, 15, 16 relaxation processes, 41, 42–5 relaxation from, 41–6

346

Index

excitons, 7–10, 38 bound, 9, 17, 217 molecular, 9–10, 11, 16–19 non-radiative quenching of, 187, 188 aggregation leads to, 201 in OLEDs, 184, 185, 233, 234, 248 fluorescent dyes, 196 host-guest materials, 193 neat emitters, 187 and phosphorescence, 252–6 in QLEDs, 185 relaxation process, 41–6 short lifetime in Ir complex devices, 197 exiplex-type materials, 261 external quantum efficiency see EQE extinction coefficients, 242–3 extrinsic luminescence, 44–5 eyes see human eye and visual system eye sensitivity function, 24–5 F fabrication of LEDs, 273–5 of OLED displays, 203–4 alternative techniques, 208–12 fine metal masks for, 207–8 limits commercialization, 213 transistors and capacitors included, 205, 206 of OLEDs, 231 of QD-LEDs displays, 217 solution processing of P-OLEDs, 192 of SM-OLEDs, 192 see also synthesis; wet processing fill factor, 208, 209, 210 films for integration of quantum dots, 57, 59, 60–1 in mass produced products, 62, 63 TFEL, 313–28 thin, optical properties of, 242–5 fine metal masks (FMM), 207–8 see also shadow masks FIrpic, 197, 198, 214, 219

as OLED emitter material, 234, 235, 253–5 flat panel displays (FPDs) OLEDs for, 183, 203 TFELs in, 319 see also displays flip-chip configuration, 284–5, 293 fluorene, 189, 190 fluorescence, 18, 42 by metal chelates, 187 by OLEDs, 249–51 aromatic dyes, 194–6 commercial use of, 212 PAHs, 189, 191 TADF emitters, 199–201, 257–63 fluorescent lamps, phosphors in, 135 fluoride phosphors, metal, 121–2 FMM (fine metal masks), 207–8 see also shadow masks focused ion beam (FIB) techniques, 50 Förster resonance energy transfer see FRET forward bias on p-n junction diodes, 28, 29, 30 forward voltage, 280 of LEDs, 297, 302, 306 FPDs (flat panel displays) OLEDs for, 183, 203 TFELs in, 319 Frank-van der Merwe mode (FvdM), 52, 53 Fresnel coefficients, 243 Fresnel reflection, 291 FRET (Förster resonance energy transfer), 19, 65, 185 from host to guest, 193 in QD-LEDs, 217, 219 in QLEDs, 251, 255, 256 in self-quenching of QDs, 215 singlet excitons, 193 G GaAs, 8 GaAs/AlGaAs quantum structures, 50 Ga(As,P):N material system, 281–2 Ga(As,P) material system, 280–1

Index

(Ga,In)N, 283–5, 288 broadening of emission spectra, 296 ESD protection, 290–1 LEDs data, 297, 298, 299 external quantum efficiency, 294–5, 297 light extraction efficiency, 291, 292, 293 light output v current characteristics, 296, 298, 300 Ga2 O3 :Eu phosphor in TFEL devices, 325 gas reduction and nitridation, 139–40 GE/General Electric, 77, 280, 288 diodes by, 91 TriGainTM , 122 green color, coumarin moieties for, 196 green emitters LEDs, early, 92, 282 OLEDs, 202, 251 phosphors for, 147–50, 167–8 QD-LEDs, lifetime of, 220 green gap, 296 green window, 93, 95, 97 ground state of helium, 15, 16 and luminescence, from phosphors, 31 guest emitters, 192–201 guest molecules, 185, 186 H hazardous substances, directives on, 54, 67, 75 HDR (high dynamic range) applications, 53, 57, 60, 61, 63 heat removal from LED chips, 288, 289–90 heatsinking in packaging, 289–90 Heizenberg Uncertainty Principle, 7 helium atom, 11, 15–16 Helmholtz-Kolrausch effect, 57 heterojunctions, 277–8, 279 Hewlett Packard, 282, 285, 289 high dynamic range (HDR) applications, 53, 57, 60, 61, 63

347

highest unoccupied molecular orbital see HOMO HIL (hole injection layers) in OLEDs, 233 in QLEDs, 64, 65, 66–7 Hisense, 63 hole current, p-n junction diode, 28 hole diffusion current, p-n junction diode, 28 hole drift current, p-n junction diode, 28 hole injection layers see HIL hole mobility in host emitters, 193 in PAHs, 189 in polymer OLED (P-OLED), 191 holes energy of, 20 and excitons, 7–10 and molecular excitons, 16–17 in OLEDs, 184, 233 in semiconductors, 31–2 see also electron-hole recombination hole transport layer see HTL HOMO (highest unoccupied molecular orbital), 17, 32, 40 in OLEDs, 184, 233, 234 host molecules, 193 in organic semiconductors, 186 in TADF emitters, 200, 259–61 host-guest system, 185, 193 guest emitters, 192–201 and OLEDs, 186, 193, 233–4, 250 hot-solution decomposition process, 51–2 HTL (hole transport layer) in OLEDs, 184, 185, 186, 233 in QD-LEDs, 213, 216, 217 in QLEDs, 64, 65, 66–7 human eye and visual system and CIE, luminosity function, 295 and color shift, over lifetime, 307 eye sensitivity function, 24–5 and luminous efficiency of radiation, 239 and perception, 25–6, 238

348

Index

human eye and visual system (continued) of hue, and LED dominant wavelengths, 296, 297, 299 of white light, 241–2 and photometry, 237 photopic curve/function, 238 and sub-pixel rendering, 208 units relating to, 23–4 Hund’s rule, 199 HVPE (hydride-vapor-phase-epitaxy), 285 hybrid organic-inorganic charge transport QD-LEDs, 7–18 hydride-vapor-phase-epitaxy (HVPE), 285 hydrocarbons, polycyclic aromatic (PAHs), 189–92, 216 hydrothermal synthesis, 52 I IBM, diode use by, 91–2 iFire Inc, 315 I-III-VI materials for QDs, 75 III-V QDs, 50 hazardous nature of, 75 II-VI QDs, 50, 51 regulation and environmental issues, 75 Illuminating Engineering Society (IES), 307 incandescent lamps/bulbs CRI of, 241, 287 luminous efficacy of, 301 as non-Lambertian light sources, 237 indirect band gap semiconductors, 276–7 absorption in, 22 band gap energy, 32 and excitons, 9 higher rates of Auger recombination, 46 indium-tin-oxide (ITO) contacts, 284 InGaN LEDs, 57, 61, 92, 94 blue/UV, for white light, 112, 122 in-home lighting, warm whites for, 169 injection efficiency, 278–80 inorganically passivated QDs, 48–9

inorganic electron transport layers, 217–18 inorganic nanophosphors, 33 inorganic semiconductors, 8–9, 16, 19 InP QDs, 54, 62, 75 QLEDs, 70 integration of quantum dots, 57–61 internal quantum efficiency see IQE intersystem crossing see ISC intrinsic semiconductors, 42 in vivo bioanalytical applications, 49 ion beam nanofabrication, 50 IQE (internal quantum efficiency) and defect densities, 283, 284 of LEDs, 94, 95, 280, 292 high performing LEDs, 295 of OLEDs, 196, 249–50 of PLEDs, 267 Ir(MDQ)2 (acac), OLEDs from, 234, 235 Ir(3’,5’,4-mppy)2 (tmd) (phosphorescent dye), 198 Ir(piq)3 (phosphorescent dye), 197, 198, 214 Ir(ppy)2 (acac), 197, 198 as OLED emitter material, 234, 235 Ir(ppy)3 , 196–7, 198, 214 as OLED emitter material, 234, 235 ISC (intersystem crossing), 42, 43, 199 in metal chelates, 187–8 RISC (reverse intersystem crossing), 199–200, 201, 258–9 ITO (indium-tin-oxide) contacts, 284 ITU-R Recommendation BT.200, 54 I-V curves of p-n junction diodes, 29–30 staircase, for quantum dots, 35 J joint density of states function, 21–2 K K2 SiF6 :Mn4+ , 121–2, 172–3 Konka, 63 KSF phosphors, 77 see also K2 SiF6 :Mn4+

Index

L LaAl(Si6−z Alz )(N10−z Oz ):Ce3+ , 155–6 labelling, QDs in, 78, 79 (La,Ca)3 Si6 N11 , preparation of, 140 Lambertian emitters, 237, 238 Laporte selection rule, 45 laser diodes (LDs), 92, 124–5 lasers, fabrication with, 207, 208, 210, 211 La3 Si6 N11 :Ce3+ , 169, 170 LaSi3 N5 , 140, 156–7, 158 LCAO-MO (Linear Combination of Atomic Orbital Theory and Molecular Orbital Theory), 39–41 LCD backlight, 172–3, 305–6 see also backlight; BLU (backlight units) LCDs (liquid crystal displays) low temperature polysilicon (LTPS), 213 problems with, 62–3 QDs in, 54 in smartphones, 208, 209 LDs (laser diodes), 92, 124–5 lead frame packaging, 288, 289 LED-LCD, 53 LEDs (light emitting diodes), 273–80 CIE chromaticity diagram for, 94, 299 color conversion phosphors for see color conversion phosphors compared to TFEL and powder EL, 313–14 historical overview, 91–3 lighting applications, 92 material systems for, 280–8 nitride phosphors for, 135–73 as non-Lambertian light sources, 237 oxynitride phosphors for, 135–73 packaging technologies for, 288–91 performance of, 291–306 p-n junction diodes, 28–30 QDs in, 35, 77–8, 213 red color rendering problems, 76–7 without color conversion phosphors, 93–5

349

LER (lumen-equivalent-of-radiation), 295–6, 301, 302, 303 LG, 63, 212–13 lifetime of LEDs, 93, 306–7 of OLEDs, 73, 76 and color, 251 in displays, 204, 205, 206, 207, 208, 212 white OLEDs, 204 of QD-LEDs, 220 of QLEDs, 72–3, 76 SSL applications, 76 T95 lifetime, 73 ligands between QDs, reduces FRET, 215 for capping QDs, 47–8 in metal chelates, 187, 188 in phosphorescent emitters, 197, 199 light emission OLED materials, 186–202 in powder EL, 329–33 in TFEL, 313–14, 317–18 as radiation, 237 from accelerating charges, 1–4 see also photons; radiation light emitting diodes see LEDs light extraction efficiency, LED, 291–2, 293 light outcoupling efficiency, 249–50 light output v current characteristics, LED, 296, 300 light sources CCT (correlated color temperature), 27, 240–2 luminance requirements of, 239 luminous efficacy of, 301 Linear Combination of Atomic Orbital Theory and Molecular Orbital Theory (LCAO-MO), 39–41 liquid crystal displays see LCDs liquid phase epitaxy (LPE), 274 Li-[alpha]-sialon:Eu2+ , 169, 170 LITI, 210, 211 local dimming, 60, 61, 63

350

Index

loss mechanisms bottom-emitting OLEDs, 246–7 in LEDs, 275–6, 278–80 lowest unoccupied molecular orbital see LUMO low temperature polysilicon (LTPS) LCDs, 213 LPE (liquid phase epitaxy), 274 LTPS (low temperature polysilicon) LCDs, 213 LuAG (blue-shifted YAG), 287, 303, 305 lumen, 24, 239 lumen-equivalent-of-radiation see LER lumen maintenance (L70), 307 Lumerical (software tool), 248 Lumileds, 118, 121, 122, 284 white LEDs from, 288, 301 luminance, 24, 239 luminescence, 1–30, 31–2 blinking effect, 46–7 extrinsic luminescence, 44–5 processes that can occur during, 41 and relaxation processes of excitons, 42–5 luminescence spectra of activators, 142–64 luminosity function, 295 luminous efficacy, 25 of LEDs, 296, 297 white LEDs, 169, 170, 172, 301, 302, 303 of OLEDs, 239 of organic electroluminescent materials, 185 of SrGa2 S4 :Eu2+ phosphor, 116, 117 of TFEL devices, 319 of YAG:Ce phosphors, 103–6, 108–10 of Yx Gd1−x AG:Ce phosphors, 108 luminous efficiency, 25, 239 of LEDs for white light, 97, 98 of man-made Plankian emitters, 74–5 luminous energy, 238 luminous flux, 24, 239 luminous intensity, 23–4, 25, 239 LUMO (lowest unoccupied molecular orbital), 17, 32, 40

in OLEDs, 184, 233, 234 host molecules, 193 in organic semiconductors, 186, 200 in TADF emitters, 200, 259–61 M MacAdam ellipses, 242, 305 magnetic fields around point charges, 1–4 magnetic moment, 16, 18 magnetic resonance imaging (MRI), 79 manganese see Mn2+ masks in LED fabrication, 274 see also shadow masks Maxwell equations, software for, 248 MBE (molecular beam epitaxy), 53, 274 mCP (1,3-bis(N-carbazolyl)benzene), 193, 194 mechanoluminescence, 32 metal chelates, 187–8 metal fluoride phosphors, 121–2 metal nitride phosphors, 117–20 (Ba,Sr)2 Si5 N8 :Eu2+ (BSSNE, 2–5–8) 118 CaAlSiN3 :0.8 Eu2+ , 119–20 M2 Si5 N8 , 118–19 Sr[LiAl3 N4 ]Eu2+ , 119 Sr2 Si5 N8 :Eu2+ , 118 see also nitride phosphors metal organic chemical vapor deposition (MOCVD), 274 metal oxide based phosphors, 99–113, 325 alkaline earth metal silicates, 111–12 doped with Eu3+ , 112–13 Na2 Gd2 B2 O7 :Ce3+ ,Tb3+ , 111 silicate garnets and related phosphors, 110–11 (Y2−x−y Eux Biy )O3 , 111 yttrium aluminum garnet, 110 YAG:Ce3+ , 99–101 YAG:Ce3+ , Pr3+ , 101–6 Yx Gd1−x AG:Ce, 106–10 see also YAG metal oxide films, 217–18

Index

metal oxo-nitride phosphors, alkaline earth, 120–1 metal silicate phosphors, alkaline earth, 111–12 metal sulfide based phosphors, 113–17, 122 alkaline earth metal thiogallates MGa2 S4 :Eu2+ , 114–17 MS:Eu2+ , 113–14 metal thiogallate phosphors, alkaline earth, 114–17 MGa2 S4 :Eu2+ phosphors, 114–17 microemulsion processes, 51 microwaves in QD synthesis, 52 Mn2+ , 318–19 in K2 SiF6 :Mn4+ , 121–2, 172–3 in Mn-doped [gamma]-alon, 162–4 in TFEL devices, 314, 315, 318–19, 324–5 mobile/cell phones OLEDs in, 63, 65, 205 see also smartphones mobility see electron mobility; hole mobility MOCVD (metal organic chemical vapor deposition), 274 molecular beam epitaxy (MBE), 53, 274 molecular excitons, 9–10, 16–19 molecular orbitals, 16–17, 185–6 see also HOMO; LUMO momentum, 18, 46, 199, 276–7 monochromatic performance, LED, 292–8, 299, 300 monomeric and excimeric emission bands, 236, 256–7 MRI (magnetic resonance imaging), 79 MS:Eu2+ (metal sulfide phosphors), 113–14, 122 MSi2 O2 N2 :Eu, 148, 150, 155 M2 Si5 N8 , 118–19, 141 in pcLEDs, 122 multi-colour LEDs, packaging for, 290 multi-layer OLEDs, 233 multi-phosphor pcLEDs, 122–3 multiple-quantum-well (MQW) structure, 278, 279

351

multi-primary LEDs, 286, 295 MYSi4 N7 , preparation of, 141 N Na2 Gd2 B2 O7 :Ce3+ ,Tb3+ phosphor, 111 Nagoya University, 283 nanobiotechnology, 81 see also biological applications Nanoco, 78 nanomaterials, 34 nanoparticles, 35, 50 QDs as, 123 nanophosphors, 33 NanoPhotonica, 68, 69 nanostructured materials, 34–5 Nanosys, mass production of QD materials by, 62 napthalene, 189, 190 narrow-band green nitride phosphors, emerging, 167–8 narrow-band red nitride phosphors, emerging, 165–7 National Television Standard Committee (NTSC), 172, 306 near-infrared emission, 49 neat emitters, 187–92, 199 nephelauxetic effect, 136, 137, 155 neutral white, 241 NiAu contact layer, 284 Nichia Chemical Company, 96, 122 nitridation, 140–1 nitride phosphors, 117–20 applications of, 169–73 emerging, 165–8 for LEDs, 135–73 Lumileds uses, 122 photoluminescence of, 142–64 synthesis of, 138–42 nitrides for LEDs, 92 see also metal nitride phosphors nitridoaluminosilicate phosphors, 138 nitridosilicate phosphors, 138 non-Lambertian light sources, 237 non-radiative recombination, 41, 45–6, 187, 188 and Cu devices, 201

352

Index

non-radiative recombination (continued) in LEDs, 275, 278 in OLEDs, 185 in QD-LEDs, 218, 220 in QDs, 215 in shell materials, 48 [alpha]-NPD, 243 NTSC standard, 172, 306 O OBAs (optical brightening agents), 287 OLEDs (organic light emitting diodes), 183, 184–6 and Dexter energy transfer, 19, 185 in displays, 63, 203–13 excitation in, 17, 18 external quantum efficiency (EQE) of, 64 host-guest three layer, 186 as Lambertian emitters, 237 lifetimes, 73, 76 materials for, 186–202 molecular excitons in, 10 multi-layer OLEDs, 233 QDs as emitting layer in, 213–20 structure, 231–2 TADF-based, 257–61 three-color polymer, 212 see also PLEDs; P-OLEDs on-chip integration of quantum dots, 57, 59, 61, 62 in LEDs, 77–8 optical brightening agents (OBAs), 287 optical properties and applications, of QDs, 53–81 of thin films, 242–5 optics of white OLED devices, 242–8 orange phosphors, warm whites from, 169 orbitals atomic, 5, 314 molecular, 16–17, 185–6 see also HOMO; LUMO organically capped QDs, 47–8 organic charge transport QD-LEDs, 215–17

organic electroluminescent materials, 184–6 organic light emitting diodes see OLEDs organic semiconductors electrons in, 10–11 excitons in, 9–10, 16, 17 in neat emitters, 187 in OLEDs, 184, 248 PAHs, 189–92, 216 pi and pi* molecular orbitals in, 186 PPV, in QD-LEDs, 215, 216 organometallic compounds, pyrolysis of, 51–2 Orion QD Quantum Dot Linear Lighting, 78 oscillating dipoles, 4–5, 6–7 Osram, 118, 122, 283, 285 use of QDs, 123–4 Ostwald ripening, 52 outcoupling efficiency of OLEDs, 245–6, 249–50 and phosphorescence, 252–3 top-emitting OLEDs, 247 oxide dielectrics, 326 oxides metal oxide films, 217–18 phosphors see metal oxide based phosphors oxo-nitride phosphors, alkaline earth metal, 120–1 oxynitride phosphors, 135–73 applications of, 169–73 emerging, 165 photoluminescence of, 142–64 synthesis of, 138–42 P Pacific Light Technologies, 62 packaging, 288–91, 300, 306 and lifetime, 307 PAHs (polycyclic aromatic hydrocarbons), 189–91 as blocking layer in QD-LEDs, 216 conjugated polymers as, 191–2 paramagnetic ions, 187 particle-in-a-box model, 38

Index

passive-matrix addressing, 204–5, 319 patterned-sapphire-substrate (PSS), 285 see also PEDOT:PSS Pauli exclusion principle, 11, 15, 252 PbS QDs, 36 PCE (power conversion efficiency), 124, 306 of LEDs, 292, 294, 295, 297 white-emitting, 299–301, 302 pcLEDs, 92, 97–8, 122–3 PEDOT:PSS, 66–7 Peng group, 67 PenTile layout, 208, 209 performance of LEDs, 291–306 light extraction efficiency, 291–2, 293 monochromatic performance, 292–8, 299, 300 reliability, 306–7 white-emitting performance, 298–306 PFNBr-DBT5, 191, 192 PFO (poly(9,9’-dioctylfluorene)), 191, 192 PFS (K2 SiF6 :Mn4+ (KSF/PFS)), 121–2 phase transitions of QDs, 37 Philips, 62, 63, 119 see also Royal Philips Electronics phOLED, 196–7, 199 phonons, emission of, 41, 45 phosphole oxide, 201 phosphor-converted LEDs (pcLEDs), 92, 97–8, 122–3 phosphorescence, 18, 42, 196–9 and intersystem crossing, 42, 43 issues with, 257 of OLED materials, 234–5 of OLEDs, 212, 251–7 in polymeric emitters, 267–8 and TADF emitters compared, 261–3 phosphorescent dopants in QD-LEDs, 219 phosphorescent dyes, 196–9 phosphor layers in TFEL devices, 314, 315, 316, 324–5 criteria for, 317 device operation, 317–20, 321 phosphors, 21, 31, 76–7

353

color conversion, 91–125 down-conversion, 73–4, 76–8, 169, 286–306, 296 for laser diode lighting, 124–5 as long established technology, 95 nitride see nitride phosphors oxynitride see oxynitride phosphors synthesis, 99, 112, 113–16 nitride and oxynitride, 138–42, 150 wide use of, 135 photolithographic methods, 205, 212, 236 photoluminescence see PL photoluminescence quantum yield see PL QY photometry, 23–7, 236–42 photon recycling, 278 photons absorption, 32 emission, 32–3, 41, 42–5, 275 by OLEDs, 184 rate of, 7, 19–23 in TFEL, 313, 324 see also light; radiation photopic curve/function, 238 physical methods for QD synthesis, 52–3 physical vapor deposition (PVD), 53, 263 pi and pi∗ orbitals, 186, 188 pixels, 203, 206 in AMOLED displays, 207–8, 209 PL (photoluminescence), 32–3, 41, 42–5, 95 of nitride and oxynitride phosphors, 142–64, 169 Ce3+ doped, 155–60 emerging phosphors, 165–8 Eu2+ doped, 142–55 Mn2+ doped, 162–4 Yb2+ doped, 160–2 of OLEDs, 234–5, 249–50 of QDs, 53–62 in LEDs, 77–8 mass produced products, 62, 63 Planck curve, 26–7, 240–2 Planckian emitters, 74–5, 240 plasma display panels, 135 plasmonic absorption, 246, 247

354

Index

PLEDs white light from, 263–9 see also P-OLEDs PLE spectrum, 32–3 PL QY (photoluminescence quantum yield), 185 of QDs, 53, 54, 58 plug-and-play combinations, 236 PMOLED (passive-matrix driving OLED), 204–5 p-n junction diodes, 28–30, 91 in LEDs, 273 P-OLEDs, 191–2 see also PLEDs polycrystalline materials, emission from, 313–14 polycyclic aromatic hydrocarbon oligomers see PAHs poly(9,9’-dioctylfluorene) (PFO), 191, 192 polymer-based OLEDs (PLEDs), 263–9 polymer OLED (P-OLED), 191–2 polymers conjugated, 191–2 as host-guest materials, 193–4 with photoresist-like properties, 212 PPV, 66, 215, 216 in P-OLEDs, 191, 192 and white light emission, 263–9 polymer TADF, 267–8 poly(p-phenylene) (PPP), 191, 192 poly(p-phenylene vinylene) see PPV poly(3,6-silafluorene-co-2,7-fluorene), 191, 192 polysilicon, low temperature, (LTPS) LCDs, 213 porphine dopants from, 194 see also PtOEP powder EL, 313–14, 328–37 power, 5, 24 power conversion efficiency see PCE power signalling applications, 288 Poynting vector, 4 PPP (poly(p-phenylene)), 191, 192

PPV (poly(p-phenylene vinylene)), 66, 215, 216 in P-OLEDs, 191, 192 Pr3+ , 101–6, 162, 163 praseodymium (Pr), 101–6, 162, 163 precipitation methods of self-assembly, 50 printing in OLED fabrication, 231 displays, 203, 208, 210, 211 in QD-LEDs display fabrication, 217 PSiFF (poly(3,6-silafluoreneco-2,7-fluorene)), 191, 192 PSS (patterned-sapphire-substrate), 285 see also PEDOT:PSS Pt-A (phosphorescent dye), 198, 199 Pt complexes with excimeric emission bands, 256–7, 258 as phosphorescent dyes, 198, 199 PtOEP, 196, 198, 199, 252 PtON7-dtb (phosphorescent dye), 198, 199 Purcell effect, 244–5 PVK (poly(9-vinylcarbazole)), 67, 68, 71, 193, 194 pyran family, dopants from, 194 pyrene, 189 pyrolysis of organometallic compounds, QDs from, 51–2 Q QCSE (quantum confined Stark effect), 295, 298 QD-LEDs (quantum dot light emitting devices), 213–20 CdSe/CdS-PPV bilayer QD-LED, 216 charge transport, 7–18, 215–17 energy transfer enhanced, 219 lifetime of, 220 spectra, 214 structure, 213 see also QLEDs QDs (quantum dots), 35–41, 42–5 applications biological, 35, 78–81

Index

displays, 53–73, 213–20, 306 LCD backlighting, 288, 306 solid state lighting, 73–8 televisions, 62 band edge emission peak, 43 blinking effect, 46–7 CdSe QDs, 36, 37 CdS QDs, 36, 51 colloidal QDs, 213–20 efficiency of, 56–7, 123–4 as electroluminescent light sources, 75–6 InP, 54, 62 optical properties and applications, 53–81 organically capped, 47–8 QD-LEDs, 213–20 QLEDs, 63–73 quantum confinement, 36–7, 38–41, 213 size and properties, 34–5, 213, 214 surface passivation, 47–9 surface states, 36, 45, 47–8 surface-to-volume ratio, 36, 44 synthesis, 49–53, 80, 81, 213 ZnO QDs, 44, 51 QD vision, 62, 67, 77 QLEDs (quantum dot light emitting diodes), 63–73 cadmium-free, 70–1 device lifetime, 72–3, 76 white, 71–2 see also QD-LEDs quantum confined Stark effect (QCSE), 295, 298 quantum confinement, 36–7, 38–41, 213 quantum deficit, 95, 97, 122, 300 Quantum Dot Enhancement Film (QDEF), 62 Quantum Dot Forum, 2016 62 quantum dot light emitting devices see QD-LEDs quantum dot light emitting diodes see QLEDs quantum efficiency, 137, 144, 155, 169 Quantum LightTM , 77

355

quantum mechanics, 5–7, 10–16 and defect states, 43–4 forbidden processes, 193, 249 local states, activators create, 44–5 and surface states, 36, 45, 47 QuantumRail, 62 quantum well (QW) devices (LEDs), 278, 279 quantum yield, 185 of OLEDs aggregate-induced emission materials, 202 conjugated polymers, 192 phosphorescent emitters, 196, 197 and polycyclic aromatic hydrocarbon oligomers, 189, 191 TADFs, 201 see also PL QY Q-V plot for TFEL devices, 321, 322

R radiance, 237–8, 239 radiant energy, 238 radiant flux, 239 radiant intensity, 237, 239 radiation, 1–4, 32–3, 237 dipole, 4–5, 19 see also light; photons; radiative recombination radiative efficiency, 245, 277, 278 radiative recombination, 41, 42–5, 275 in OLEDs, 252–6, 258–61 see also electron-hole recombination radiometric light output, 292 radiometry, 23–7, 236–42 rare earths cerium see Ce3+ europium see Eu lanthanum, 140, 155–7, 158, 169, 170 praseodymium (Pr), 101–6, 162, 163 supply problems, 73–4, 92 traditional phosphors, 76–7 Yb2+ , 160–2 yttrium see yttrium reactive-ion etching (RIE), 49, 50, 208

356

Index

recombination see electron-hole recombination recombination rate, 277 RED2 (aromatic dye), 194, 195 red emitters LEDs, 76–7, 281 OLEDs, 193, 194, 202, 251 phosphors for, 152–5, 165–6 QD-LEDs, 213, 215, 216, 217, 218 lifetime of, 220 redshift in conjugated polymers, 191–2 in fluorescent emitters, 194 in metal chelates, 188 in PAHs, 189, 191 in phosphorescent emitters, 197 in phosphors, 136 in QD emission, 48, 49 refractive index, 242–3 and light extraction efficiency, 291, 292 and loss mechanisms, 246 relaxation processes of excitons, 41–6 see also electron-hole recombination reliability of LEDs, 306–7 see also lifetime resolution and OLED display fabrication, 207, 208, 209 and TFEL devices, 327–8 Restriction of Hazardous Substances (RoHS) Directive, 54, 67, 75 reverse bias, 28, 29, 30 reverse micelle process, 51 RGB color patterning, 203–4, 207, 217 RGB emitting sub-pixels, 203 RGB OLEDs, 212, 213 side-by-side, 203, 204, 207–8 novel shadow masks, 210, 211 RGB units in OLED architecture, 235, 236 RISC (reverse intersystem crossing), 199–200, 201, 258–9 Round, Henry Joseph, 33, 91 Royal Philips Electronics, 285 see also Philips

S Samsung, 62, 63, 123, 212–13 Sawyer-Tower circuit, 321–3 Schrödinger’s equation, 5–6, 11 selective diffusion of luminescent materials, 210 selective transfer of organic layers, 208 self-assembly techniques, in synthesis of QDs, 50–3 self-healing behaviour, 326–7 semiconductor nanoparticles, QDs as, 123 semiconductors, 8–10, 31 band gap and other data for, 39 and QD size, 36–7 in LEDs, 273–4 in OLEDs, 231–2 organic see organic semiconductors PPV, in QD-LEDs, 215, 216 surface states, 47 SETFOS software simulation tool, 248 shadow masks, 207–8, 210, 211 in fabrication of OLED devices, 203, 236 Shanghai Tianma Microelectronics Group, 68 Sharp, 63 shells see core-shell systems [beta]-Sialon, 144, 156 Eu doped, 147–8, 155, 170, 171 in LCD backlights, 167, 172, 173 Pr3+ doped, 162, 163 Sialon-type phosphors, 120 see also Ca-[alpha]-sialon; Li-[alpha]-sialon:Eu2+ ; [beta]-Sialon; Y-[alpha]-sialon side-by-side RGB OLEDs, 203, 204, 207–8 fabrication costs, 213 novel shadow masks, 210, 211 silica dielectrics, 326 silicate garnets, 110–11 silicate phosphors, 110–12 silicon, low temperature poly-, (LTPS) LCDs, 213

Index

silicon carbide/SiC, electroluminescence of, 33, 91 siloles, 201, 202 Silvaco Atlas, 248 SimOLED software simulation tool, 248 simulation tools, 248 singlet states, 14, 15–16, 17–18 in intersystem crossing (ISC), 199 in OLEDs, 193, 196, 248, 249, 250, 252 TADF-based, 258, 259, 260–1 Skyworth, 63 small molecule OLEDs (SM-OLEDs), 191, 192, 212 smartphones LCD, 208, 209 OLEDs in, 205, 208, 212, 213 see also cell/mobile phones SM-OLEDs (small molecule OLEDs), 191, 192, 212 Sn4+ -porphyrin complexes, 200 software simulation tools, 248 solar cells, QDs in, 35 sol-gel techniques, 51 solid state lighting see SSL solid state luminescence see luminescence solid state reaction method, 138–9 solution processing, 192, 203 see also wet processing sonic waves in QD synthesis, 52 Sony, 62, 123, 212 Soraa, Inc., 285 sphere-supported TFEL, 316–17 spin, 12–16, 18, 45, 318–19 and OLED emission, 248–9, 252, 258 SrAlSi4 N7 :Eu2+ phosphor, 153, 154–5 SrAlSiN3 , 141, 143 (Sr1−x−y Bax Cay )2 Si5 N8 :Eu2+ , 122 (Sr0.5 Ba0.5 )Si2 O2 N2 :Eu2+ , 169, 170 Srx Ca1−x S:Eu phosphors, 113–14 SrGa2 S4 :Eu phosphors, 115–17, 122, 325 Sr[LiAl3 N4 ]:Eu2+ , 119, 166, 171, 288 Sr[Mg3 SiN4 ]:Eu2+ , 166–7 SrSi9 Al19 ON31 , 144, 146–7

357

Sr2 Si5 N8 , 140, 141, 143, 156 Eu doped, 118, 152, 153, 155, 171 Sr3 Si2 O4 N2 , 156, 157–8 SrSi2 O2 N2 , 144 Eu doped, 120–1, 122, 148, 150, 155, 161–2, 171 Yb2+ doped, 161–2 SSL (solid-state lighting), 92 ANSI standard for, 241–2 CRI of, 241 customer requirements, 96 laser diodes (LDs) in, 124–5 nitride phosphors in, 136 quantum dots in, 73–7 Standard Deviations of Color Matching (SDCM), 305 stationary states, 5–7 Stokes’ loss, 300, 303 Stokes shift, 42–3, 45, 56 in phosphors, 95, 137 Stranski-Krastonow mode (SK), 52, 53 structure of [gamma]-alon, 162 of emerging nitrides, 166–7 structure-property relationships, phosphor, 142 Ce3+ doped phosphors, 155, 156, 157–8 CaSiN2 :Ce3+ , 159–60 Eu2+ doped phosphors, 143–4 blue-emitting phosphors, 145–6 green-emitting phosphors, 147, 148 red-emitting phosphors, 152, 153, 155 yellow-emitting phosphors, 150, 151–2 sub-band gap emission, 65 sub-pixel rendering, 208 substrate modes, 246–7 SUHD televisions, 62 sulfide phosphors, 113–17, 122 superposition states, 6–7, 19 surface mount mid power LEDs, 289 surface passivation, 47–9, 72 surface recombination in LEDs, 276 surface states and QDs, 36, 45, 47–8

358

Index

surface-to-volume ratio, 36, 44 surface traps, 215, 276 switches, QDs in, 79 symmetric wavefunctions, 13, 14, 15, 16 Synopsys RSoft, 248 synthesis of phosphors, 99, 112, 113–14 alkaline earth metal thiogallate, 114–15 Ce3+ doped, 155–6, 159–60 emerging nitrides, 166, 167, 168 nitride and oxynitride, 138–42, 150 SrGa2 S4 :Eu2+ , 115–16 of QDs, 49–53, 80, 213 see also fabrication systems trackers, 78, 80–1 T T95 lifetime, 73 TADF emitters, 199–201, 257–63, 264 polymer TADF, 267–8 TCL, 63 TCTA, 67 television, standards for, 172, 306 televisions LED-backlit LCD, 305 OLEDs in, 63, 212–13 QDs in, 53, 61, 62, 63, 123 see also displays tetraphenylethene (TPE), 201, 202 5,10,15,20-tetraphenyl-21H, 23H-porphine (TPP), 194, 195 Texas Instruments, 91 TFB, 67 TFEL (thin film electroluminescence), 313–28 background and structure, 314–17 devices, 33, 314, 315, 324–7 materials in, 313–14 theory of operation, 317–23 thermal deposition, 203 thermal evaporation, 231 thermalization, energy losses through, 8 thermally activated delayed fluorescence see TADF

thermal quenching, 137, 169 thermal stability, and OLED devices, 187 thermal transfer techniques, 208 thermoluminescence, 32 thick film dielectric EL structure, 315–16 thin film electroluminescence see TFEL thin films optical properties of, 242–5 see also films thin film transistors (TFTs), 205–6 thiogallate phosphors, alkaline earth metal, 114–17 thiogallates, 114–17 3M, 62 threshold voltage, 319, 320, 321, 322 TNB, 202 top-emitting OLEDs, 236, 247–8 total internal reflection, and loss mechanisms, 246–7 TPD, 66, 67, 216 TPE (tetraphenylethene), 201, 202 TPP(5,10,15,20-tetraphenyl-21H, 23H-porphine), 194, 195 transfer matrix methods, 243, 244, 248 transitions see electronic transitions transparent substrate (TS) devices, 282–3, 285, 291 transport layers in OLEDs, 233, 234 in organic charge transport QD-LEDs, 215, 216 see also ETL; HTL triarylamine, 202 trichromatic illumination, 26 TriGainTM , 77, 122 Triluminos television, 62 triplet harvesting, 251 triplet states, 14, 15–16, 17–18 in intersystem crossing (ISC), 199 in OLEDs, 196, 248, 249, 250, 251–6 triplet–triplet annihilation (TTA), 188, 196, 250 tri-PXZ-TRZ, 200, 201 tris(8-hydroxyquinolinato)aluminium see Alq3

Index

TS (transparent substrate), 282–3, 285, 291 TTA (triplet–triplet annihilation), 188, 196, 250 tumor detection by QDs, 80–1 TVs see televisions two-electron atoms, 10–16 U ultrasound in QD synthesis, 52 ultraviolet LEDs, 110–11, 136, 170 units, 23–7, 236–9 Universal Display Corporation, 73 US Department of Energy, 241, 301 UV-LEDs, 110–11, 136, 170 UV light and Ce3+ , 155 and optical brightening agents, 287 V valence band, 7, 19–23, 31, 275 in indirect band gap, 32 p-n junction diode, 28 vanadate garnet UV-LED phosphor, 110–11 vapor phase epitaxy (VPE), 274 video and photography, 54, 62 Vizio, 63 Volmer-Weber mode (VM), 52, 53 VPE (vapor phase epitaxy), 274 W warm white, 240, 241, 286 LEDs, 196, 287, 301, 303 water-compatible coatings for QDs, 49 water solubility in biological applications, 80 watts, 23, 24 wavefunctions, 5–6, 11–16 waveguided modes, 246–7 wet processing in mask fabrication, 207, 208 in OLED fabrication, 231 and polymer-based OLEDs (PLEDs), 264 see also solution processing

359

white LCDs, performance of, 298–306 white LEDs, 92, 93, 95, 97–8, 124, 285–8 for LCD backlight, 172, 173, 305–6 luminous efficacy, 301, 302, 303 outdoor lighting, 287 pcLEDs for, 122–3 phosphors, 135–6, 287–8, 298–306 alkaline earth metal silicate, 112 fluoride K2 SiF6 :Mn4+ , 121–2 multi-phosphors, 169–71 nitride, 118, 120, 169 powder characteristics, 138 Srx Ca1−x S:Eu, 113–14 vanadate garnet for UV, 110–11 YAG:Ce3+ , 99–101 YAG:Ce3+ , Pr3+ , 101–2 white light artificial source requirements, 95–6 CIE standards, 286, 295–6 LEDs for see white LEDs narrowband QDs for, 74–5 OLEDs for see white OLEDs perception of, 241–2 polymer-based OLEDs (PLEDs) for, 263–9 QLEDs for, 71–2, 75, 204 white OLEDs, 233–6 with CFs, 203–4 commercial use of, 212 fabrication costs, 213 device optics, 242–8 materials for, 248–57 structure, 231–2, 235–6 white QLEDs, 71–2, 75 efficiency of, 204 X X-rays, cathodoluminescence produces, 33 Y YAG phosphors, 110, 122, 169 easily substituted lattice, 102, 110 limit of luminous efficacies of, 103–6

360

Index

YAG phosphors (continued) in pcLEDs, 98, 122 in white LEDs, 172, 287, 298 YAG:Ce3+ , 99–101, 169, 172 YAG:Ce3+ , Pr3+ , 101–6 Yx Gd1−x AG:Ce, 106–10 Y(Al,Ga)G:Ce3+ , 122 Y3 Al5 O12 :Ce3+ phosphor, 287 Yb2+ doped phosphors, 160–2 YCa3 M3 B4 O15 :Eu3+ phosphor, 111 yellow emitters, 282 phosphors for, 150–2, 169, 170 yellow light, human sensitivity to, 238 (Y2−x−y Eux Biy )O3 phosphor, 111 Yx Gd1−x AG:Ce phosphors, 106–10 Y-[alpha]-sialon, preparation of, 140 Y3 Si6 N11 , 141 ytterbium (Yb), 160–2 yttrium aluminum garnet see YAG phosphors

yttrium (Y) phosphors, 111, 140, 141, 287 YAG see YAG phosphors supply risk and importance of, 73–4 Z zero-dimensional QDs, 35 Zn1−x Mgx S:Mn phosphors, 325 Zn1−x Mnx S, 324–5 Zn2−x Mnx SiO4 , 324–5 ZnO, 44, 217, 219 ZnO nanoparticles, 67, 218 ZnO QDs, 44, 51 ZnS:Mn, 314, 315, 318–19, 324–5 ZnS:Tb phosphor layer, 325 ZnS (layers), 215 ZnSe for QLEDs, 70–1 Zn2 Si1−x Gex O4 :Mn, 324–5 Zn2 SiO4 :Mn phosphors, 324–5 ZnTe, close packed arrays of, 50 Zylight F8–100 LED Fresnel Light, 78