Light-Emitting Electrochemical Cells : Concepts, Advances and Challenges 978-3-319-58613-7, 3319586130, 978-3-319-58612-0

Table of contents :
Front Matter ....Pages i-xii
Front Matter ....Pages 1-1
Light-Emitting Electrochemical Cells: Mechanisms and Formal Description (Stephan van Reenen, Martijn Kemerink)....Pages 3-45
Front Matter ....Pages 47-47
Optical-Beam-Induced-Current Imaging of Planar Polymer Light-Emitting Electrochemical Cells (Faleh AlTal, Jun Gao)....Pages 49-75
Optical Engineering of Light-Emitting Electrochemical Cells Including Microcavity Effect and Outcoupling Extraction Technologies (Hai-Ching Su)....Pages 77-92
The Use of Additives in Ionic Transition Metal Complex Light-Emitting Electrochemical Cells (Lyndon D. Bastatas, Jason D. Slinker)....Pages 93-119
Improving Charge Carrier Balance by Incorporating Additives in the Active Layer (Hai-Ching Su)....Pages 121-137
Morphology Engineering and Industrial Relevant Device Processing of Light-Emitting Electrochemical Cells (G. Hernandez-Sosa, A. J. Morfa, N. Jürgensen, S. Tekoglu, J. Zimmermann)....Pages 139-163
Front Matter ....Pages 165-165
Development of Cyclometallated Iridium(III) Complexes for Light-Emitting Electrochemical Cells (Catherine E. Housecroft, Edwin C. Constable)....Pages 167-202
Recent Advances on Blue-Emitting Iridium(III) Complexes for Light-Emitting Electrochemical Cells (Lei He)....Pages 203-235
Thermally Activated Delayed Fluorescence Emitters in Light-Emitting Electrochemical Cells (Michael Yin Wong, Eli Zysman-Colman)....Pages 237-266
White Emission from Exciplex-Based Polymer Light-Emitting Electrochemical Cells (Yoshinori Nishikitani, Suzushi Nishimura, Soichi Uchida)....Pages 267-286
Luminescent Cationic Copper(I) Complexes: Synthesis, Photophysical Properties and Application in Light-Emitting Electrochemical Cells (Margaux Elie, Sylvain Gaillard, Jean-Luc Renaud)....Pages 287-327
Small Molecule-Based Light-Emitting Electrochemical Cells (Youngson Choe, Chozhidakath Damodharan Sunesh, Madayanad Suresh Subeesh, Kanagaraj Shanmugasundaram)....Pages 329-349
Quantum Dot Based Light-Emitting Electrochemical Cells (Meltem F. Aygüler, Pablo Docampo)....Pages 351-371

Citation preview

Rubén D. Costa Editor

Light-Emitting Electrochemical Cells Concepts, Advances and Challenges

Light-Emitting Electrochemical Cells

Rubén D. Costa Editor

Light-Emitting Electrochemical Cells Concepts, Advances and Challenges

123

Editor Rubén D. Costa IMDEA Materiales Parque Científico y Tecnológico-Tecnogetafe Getafe (Madrid) Spain

ISBN 978-3-319-58612-0 DOI 10.1007/978-3-319-58613-7

ISBN 978-3-319-58613-7

(eBook)

Library of Congress Control Number: 2017940608 © Springer International Publishing AG 2017 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Printed on acid-free paper This Springer imprint is published by Springer Nature The registered company is Springer International Publishing AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Foreword

Light-Emitting Electrochemical Cells: organic semiconductor devices augmented by ions Polymer light-emitting electrochemical cells (LEC) was invented in 1994 to facilitate the injections of charge carriers into luminescent conjugated polymers in light emitting diodes. Initially, the benefits of adding a solid electrolyte into the organic semiconductor devices was eminent: electrons could be effectively injected from a stable metal like aluminum into the lowest unoccupied molecular orbitals which are 1 eV higher than aluminum’s work function; and the driving voltage was substantially reduced. Also, the emissive layer could be much thicker than that allowable for tunneling charge injection interfaces. Michael Rubner at MIT obtained similar results with conjugated small molecules such as a soluble  2 þ    derivative of RuðbpyÞ3 PF6 2 , a compound that had been studied for eletrogenerated chemiluminescence long before LEC was first reported. The presence of mobile ions in the LECs and consequently overpotential could induce degradation which has been partially addressed over the years. Karl Leo and Junji Kido, et al. introduced immobile ions or “doped” the organic semiconductor via organic donor-acceptor complex, which have leap frogged the operational stability of OLEDs. Works by Richard Friend, Ian Parker and others in the early 1990s showed that the conjugated polymer/electrode interfaces could be modeled as tunnelling barriers. The work function difference between ITO, a commonly used transparent anode, and calcium is around 1.8 eV. Low electron and hole injection barriers are feasible when using a small band gap polymer such as MEH-PPV to produce orange light emission. For blue light emitting polymers, large barriers are inevitable. The emissive layer has to be made as thin as processing can produce a defect-free layer, typically around 100 nm for spin coating. Even so, the driving voltage was often tens of volts, and the blue polymer OLEDs fabricated at the time

v

vi

Foreword

were very short lived, even though the quantum efficiency was decent. In finding a solution to this challenge, the doping propagation model that Olle Inganäs and me used to simulate the polypyrrole/polyethylene bilayer bending beams inspired me to introduce electrochemical doping into the polymer OLEDs. Thus, a commonly used ionically conductive polymer, polyethylene oxide plus lithium triflate, was selected to supply the mobile dopants. The Wessling precursor of PPV was selected thanks to its compatibility with the PEO-lithium salt. The resulting blend of an ionically conductive and an electronically conductive polymers showed remarkably improved electroluminescent performance compared to control OLEDs based on ITO/PPV/Al. The driving voltage was lower, the quantum efficiency was higher, and the operational stability was also enhanced. Alan Heeger, Jun Gao, Ludvig Edman, and others separately confirmed the formation of p-i-n junction in the polymer LECs by optical beam induced current measurements, direct imaging under microscope, and measurement of electrostatic potential distribution by scanning Kelvin probe microscopy. Light emission and major potential drop were observed to occur at the junction. This p-i-n junction model may not rule out other mechanisms, particularly when the junction is not formed to exhibit the ideal doping profiles at the electrode interfaces. Electrical double layers could dominate at driving voltage well below the band gap of the conjugated polymer, i.e. the onset of simultaneous p- and n-doping. The junction model essentially guides the material selection to fabricate high performance LECs. Three electronic/ionic polymer blend systems were examined in the early years include (MEH-PPV + PEO-lithium salt) and (polyfluorene with ethylene oxide side chains + lithium salt), in addition to the (PPV + PEO-LiTf) system used in the very first LEC device. Morphological control was a critical factor in the device performance: one had to consider the tradeoffs among carrier transport, ionic mobility, luminescence quenching, and accessibility of doping ions into the low-polarity conjugated polymer domain. Furthermore, mobile ions could lead to electrochemical over-reaction or degradation at high driving voltages. Freezing the ions after the formation of the p-i-n junction appears to be effective to slow down such degradation. The added freedom in electrode selection for LECs allows the fabrication of devices without the use of high vacuum: Sue Carter printed silver paste as the cathode; carbon nanotube coating could also be used as the cathode, as well as the anode. LECs formed by sandwiching the emissive polymer layer between a pair of carbon nanotube electrodes were flexible, even stretchable if the nanotubes were coated on elastomeric substrate. The LEC is now intertwined with many other fields. It is exciting to witness the latest progress in LEC performance and exploration of unique applications. Here I merely state my personal view on what occurred in the past that helped shape the field as it is today. More history and exciting developments are covered by the authors who wrote the chapters of this book. My hat’s off to these active researchers

Foreword

vii

who have made critical contributions to the field. I am indebted to Dr. Chi Zhang, Dr. Yang Yang, Dr. Yong Cao for helping fabricate the first generation of polymer LECs, Prof. Alan Heeger for polishing the junction model, and Dr. Gang Yu for suggesting the planar LEC structure to image the junction. Qibing Pei Department of Materials Science and Engineering Henry Samueli School of Engineering and Applied Science University of California, Los Angeles, CA, USA

Preface

The origins of the organic-basedsolid-state lighting (SSL) date back to 1953, but it was only in the 90s when the organic-based light-emitting diode (OLED) and light-emitting electrochemical cell (LEC) technologies started to flourish. Although OLEDs have made all the way from laboratory to commercial products, the LEC technology is considered as the simplest SSL device. The two pillars of LECs are the type of emitter that holds charge injection, charge transport, and emission and the ionic additive that assists charge injection at low applied voltages. As introduced by Prof. Pei in the foreword, the LEC revolution is based on the use of ions to reduce the turn-on voltage. After 15 years of research, we have gained a mature understanding of the device mechanism. This has, in addition, been achieved along with the optimization of the two traditional emitters (luminescent conjugated polymers and ionic transition metal complexes), the ionic additives for each type of emitters, and the type of poling modes. After having fully understood the device limitations, we have achieved several breakthroughs with respect to the efficiency using multilayered architectures (cascade and/or tandem), frozen junctions, color converting layers, etc. and low-cost and up-scalable fabrication protocols using, in addition, unconventional conductive substrates. As the most recent research action, we have focused on investigating different types of emitters like small molecules, nanoparticles, quantum dots, etc. Hence, the last two decades have been a successful test-bed time for LECs, reaching both a high industrial relevance and an always-rising research interests, as LECs are an easy set-up to investigate the electroluminescence features of the emitters and the device physics of ionic-based optoelectronics. Overall, I felt that it was now the right time to bring together all the efforts of the LEC community in this first book devoted to the LEC technology. The intention of this book is to provide to young students a general description of the LEC technology with a focus on the device mechanism and the different techniques to elucidate the role of mobile anions (Part I). After this general view, they will enjoy two sections specialized on the definition and role of the ionic additives (Part II), as well as the last advances in traditional and new electroluminescent materials (Part III). Part II is divided into five chapters that describe in-depth the type of ionic ix

x

Preface

additives and the different techniques to study the effect of the mobile ions on the device mechanism (Chaps. 2 to 5), as well as how the ionic electrolytes are crucial for the fabrication of LECs using deposition tools of industrial relevance (Chap. 6). Part III consists of seven chapters summarizing i) the progress in designing iridium (III) complexes (Chap. 7), in general, and blue-emitting iridium(III) complexes (Chap. 8), in particular, ii) the studies on new materials with thermally activated delayed fluorescence features (Chap. 9) as well as exciplex emission in conjugated polymers (Chap. 10), and iii) the last advances in new electroluminescent materials, such as copper(I) complexes (Chap. 11), small-molecules (Chap. 12), and quantum dots (Chap. 13). My intention is to provide a comprehensive vision of the past and present developments in the LEC technology as insights for future advances covering new device designs, industrial progress, and novel types of emitters. Erlangen, Germany

Rubén D. Costa

Contents

Part I 1

Light-Emitting Electrochemical Cells: Mechanisms and Formal Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stephan van Reenen and Martijn Kemerink

Part II 2

3

4

Introduction to the Light-Emitting Electrochemical Cell Technology 3

Definition and Role of the Ionic Additives

Optical-Beam-Induced-Current Imaging of Planar Polymer Light-Emitting Electrochemical Cells . . . . . . . . . . . . . . . . . . . . . . . . . Faleh AlTal and Jun Gao Optical Engineering of Light-Emitting Electrochemical Cells Including Microcavity Effect and Outcoupling Extraction Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hai-Ching Su The Use of Additives in Ionic Transition Metal Complex Light-Emitting Electrochemical Cells . . . . . . . . . . . . . . . . . . . . . . . . . Lyndon D. Bastatas and Jason D. Slinker

49

77

93

5

Improving Charge Carrier Balance by Incorporating Additives in the Active Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 Hai-Ching Su

6

Morphology Engineering and Industrial Relevant Device Processing of Light-Emitting Electrochemical Cells . . . . . . . . . . . . . 139 G. Hernandez-Sosa, A.J. Morfa, N. Jürgensen, S. Tekoglu and J. Zimmermann

xi

xii

Contents

Part III

Traditional and New Electroluminescent Materials

7

Development of Cyclometallated Iridium(III) Complexes for Light-Emitting Electrochemical Cells . . . . . . . . . . . . . . . . . . . . . . 167 Catherine E. Housecroft and Edwin C. Constable

8

Recent Advances on Blue-Emitting Iridium(III) Complexes for Light-Emitting Electrochemical Cells . . . . . . . . . . . . . . . . . . . . . . 203 Lei He

9

Thermally Activated Delayed Fluorescence Emitters in Light-Emitting Electrochemical Cells . . . . . . . . . . . . . . . . . . . . . . 237 Michael Yin Wong and Eli Zysman-Colman

10 White Emission from Exciplex-Based Polymer Light-Emitting Electrochemical Cells . . . . . . . . . . . . . . . . . . . . . . . . . 267 Yoshinori Nishikitani, Suzushi Nishimura and Soichi Uchida 11 Luminescent Cationic Copper(I) Complexes: Synthesis, Photophysical Properties and Application in Light-Emitting Electrochemical Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287 Margaux Elie, Sylvain Gaillard and Jean-Luc Renaud 12 Small Molecule-Based Light-Emitting Electrochemical Cells . . . . . . 329 Youngson Choe, Chozhidakath Damodharan Sunesh, Madayanad Suresh Subeesh and Kanagaraj Shanmugasundaram 13 Quantum Dot Based Light-Emitting Electrochemical Cells . . . . . . . 351 Meltem F. Aygüler and Pablo Docampo

Part I

Introduction to the Light-Emitting Electrochemical Cell Technology

Chapter 1

Light-Emitting Electrochemical Cells: Mechanisms and Formal Description Stephan van Reenen and Martijn Kemerink

Abstract In 1995 Pei and coworkers reported the first light-emitting electrochemical cell (LEC). LECs are unique because of a simple device layout that does not compromise efficient charge injection and transport. This is achieved by the presence of mobile ions in the active layer. Despite the simple device layout, the device physics has proved to be extremely complicated. In this chapter, an overview is given of the development and methods towards the current understanding of LEC device operation. A large amount of experimental and modeling work in the last 20 years has proved that LECs can operate in different regimes, depending on charge injection and applied bias voltage. Processes related to this device operation range from electric double layer formation at the contacts to electrochemical doping and the formation of a dynamic p-i-n junction in the bulk of the active layer. We discuss these and, where possible, include formal descriptions of transient phenomena relating to the turn-on and stability of LECs on the one hand and on the other hand steady-state phenomena relating to the current density, potential distribution, and recombination properties of LECs.







Keywords Light-emitting electrochemical cell Organic light-emitting diode Electroluminescence Organics Scanning kelvin probe microscopy Electrical impedance spectroscopy










S. van Reenen Department of Physics, University of Oxford, Parks road, Oxford OX1 3PU, UK M. Kemerink (&) Complex Materials and Devices, Department of Physics, Chemistry and Biology (IFM), Linköping University, SE-58 183 Linköping, Sweden e-mail: [email protected] © Springer International Publishing AG 2017 R.D. Costa (ed.), Light-Emitting Electrochemical Cells, DOI 10.1007/978-3-319-58613-7_1

3

4

1.1

S. van Reenen and M. Kemerink

Purpose and Aims

The light-emitting electrochemical cell (LEC) is a subclass of organic optoelectronic devices which generates light from electricity by means of electron–hole recombination in an organic semiconductor. The major characteristic of LECs is the inclusion of mobile ions in the active layer, which leads to complex device physics and, more importantly, efficient and robust operation in a simple device configuration. As such, various material classes can be used to produce LECs. This chapter expands on the general description of LECs with a focus on development towards a common, universal understanding of the LEC device operation. Furthermore, we provide a formal description of this operation that enables numerical and analytical models to qualitatively and quantitatively describe LEC device behavior. Section 1.2 will provide the reader with a general overview of LECs in terms of (i) the discovery and development of LECs, (ii) a classification in three different classes, (iii) the device architecture and figures-of-merit, (iv) the formulation of the operational model. The main intention is to provide the reader with a precise background to assist the understanding of the following chapters. Sections 1.3 and 1.4 will then concentrate on the study of, respectively, transient and steady-state processes in LECs. The final Sect. 1.5 summarizes the current understanding of LECs and provides an outlook on future challenges and work regarding the understanding of LECs.

1.2

Overview

The discovery of LECs in 1995 was preceded by various other discoveries relating to lighting from organic semiconductors. Here, a short overview is given on key events preceding the establishment of LECs as a class of organic lighting devices. Furthermore, we will introduce the classification of LECs into three main classes based on the type of semiconductor, i.e., conjugated polymers (CP), ionic transition metal complexes (iTMC), and the so-called third generation spanning small molecules, quantum dots and luminescent nanoparticles, to be discussed in more detail in Sect. 3. Finally, we discuss in depth the suggested operational mechanisms of LECs and give an overview of the current model of LEC device operation.

1.2.1

Background

Electroluminescence from organic materials was first reported by A. Bernanose and coworkers in 1953, when they applied high alternating voltages to films based on, e.g., acridine orange [1]. Subsequent development of organic electroluminescent

1 Light-Emitting Electrochemical Cells …

5

devices took, however, until 1987 due to the requirement of high drive voltages and consequently poor power efficiencies. In 1987, Tang and VanSlyke reported the first organic light-emitting diode (OLED), which started the current period of global OLED research and development [2]. They produced a diode with a bilayer structure by vapor deposition of thin films of organic semiconductors. Electrodes were used with such work functions that carrier injection in the organics was relatively efficient. Furthermore, the bilayer structure enforced recombination at the interface between the two organic semiconducting films which avoided significant losses related to electroluminescence quenching at the electrodes. A few years later, in 1990, Burroughes et al. introduced an OLED based on CP [3]. One advantage of such polymers is that they enable a larger range of deposition techniques as they can be processed from solution to form uniform, thin films. Furthermore, due to their macrostructure, they can, in principle, lead to higher long-term stability by avoiding, e.g., recrystallization or other structural changes, which can be expected for films based on small molecules. For efficient OLED operation, electrons and holes need to recombine in the bulk of the organic semiconductor to avoid electroluminescence quenching at the electrodes. To achieve this, efficient and balanced carrier injection and transport in the organics is paramount. Electron injection in organic semiconductors occurs in the lowest unoccupied molecular orbital (LUMO), whereas hole injection occurs in the highest occupied molecular orbital (HOMO). In efficient OLEDs, one high- and one low-work function electrode are therefore required for efficient hole and electron injection, respectively. This leads to the typical diode-like behavior of these devices [4]. Injected charge carriers form positively and negatively charged polarons that move through the organic semiconductor with a given mobility that can vary for different types of semiconductor [5]. Carrier transport through the organic occurs through thermally activated tunneling, or ‘hopping’ for short, of the charge carriers between localized sites in a disordered energy landscape [6, 7]. As mobilities are relatively low in organic semiconductors used for OLED applications, the organic films in OLEDs typically need to be extremely thin to achieve efficient operation at still acceptable biases: in the range of hundreds of nanometers. Electroluminescence in OLEDs occurs via recombination of electrons and holes that form singlet and triplet exciton pairs that can decay radiatively, decay non-radiatively, or dissociate. The emitted phonon typically has an energy that is close to the bandgap of the organic semiconductor. One major source of non-radiative decay is exciton quenching near the metallic injecting electrodes [8–10]. This is likely to occur in case of a large imbalance in carrier injection and/or bulk carrier transport properties between electrons and holes. In OLEDs, issues like these are typically avoided by adopting a bilayer structure [2] or by introduction of additional organic layers [11] that facilitate injection and transport of one polarity of charge carrier, while blocking the other. In modern OLEDs, such layers are used on both sides of the emissive layer(s), which leads to a massive increase in device complexity and fabrication cost [12].

6

S. van Reenen and M. Kemerink

Fig. 1.1 Left Schematic device layout of a CP-based LEC including the molecular structures of the active layer contents. Central typical current-voltage-luminescence characteristic of a CP-based LEC, (redrawn after Ref. [13]). Right photograph of a planar CP-based LEC. (reproduced from Ref. [14])

During the development of the OLED, a new type of organic light-emitting device was discovered by Qibing Pei and coworkers in 1995 [13]. By including mobile ions in the organic semiconductor, they were able to produce a device with strongly contrasting properties compared to the OLED. The device had a simple device layout consisting of only one, mixed organic layer sandwiched between metallic electrodes as shown in Fig. 1.1a. Furthermore, the device did not behave like a diode as it did not block current in reverse bias despite the asymmetry created by the use of one high- and one (moderately) low work-function electrode. This behavior is illustrated in the current-voltage characteristic shown in Fig. 1.1b. Remarkably, the light intensity is similar for both bias polarities, showing that the device is as efficient in forward as in reverse bias conditions, despite the presence of large injection barriers for at least one of the biasing conditions. This highlights that despite large injection barriers, this new type of device, which was coined LEC, can operate efficiently at low voltages, as opposed to OLEDs. The turn-on voltage of the cell is observed to be around the bandgap of the semiconductor. Another striking property of LECs is shown in Fig. 1.1c. Efficient LECs can be produced in a planar configuration with an interelectrode spacing of, in this case, 15 lm between gold electrodes. At a bias voltage of 4 V, light emission can be observed, which shows that operation at low voltage is still possible despite a thick or, in this case, wide active layer. Moreover, it indicates that efficient operation is possible while using air stable materials to contact the active layer, despite the consequent large injection barrier for electron injection. The active layer of the first LECs shown in Fig. 1.1 consists of a CP that is mixed with a salt and an ion-solvating polymer, poly(ethylene oxide) (PEO). However, the combined use of ions and organic semiconductors in LECs is not limited to these materials. For instance, iTMCs are luminescent ionic materials that consist of a metal center surrounded by organic ligands. The first electroluminescent device based on iTMCs was published by Lee et al. in 1996 [15]. They used a tris(bathophenanthrolinedisulfonate)ruthenium(II) sodium salt with a bandgap of

1 Light-Emitting Electrochemical Cells …

7

Fig. 1.2 Left Schematic device layout of an iTMC-based LEC including the molecular structure of the active layer contents. Right current (open symbols) and luminance (closed symbols) characteristics of iTMC-based LECs, in which the iTMC layer is formed by spin-coating (circles) and self-assembly (triangles). (reproduced from Ref. [15])

roughly 2.6 eV as shown in Fig. 1.2a. This device was not reported as an LEC, and only a current-voltage-luminance (IVL) characteristic in forward bias conditions was reported, as shown in Fig. 1.2b. It is, however, interesting that the spin-coated device (circles in Fig. 1.2b), despite having Al and ITO electrodes that would give rise to substantial injection barriers, shows a turn-on in current and luminescence for voltages near the bandgap, which suggests LEC-like operation. Later reports in 1999–2001 showed slow time components in the electrical characterization in similar cells which are attributed to slow ionic transport within the active layer [16– 18]. Recently, the scientific community started with the development of the so-called next generation of electroluminescent materials for LECs [19–32]. Small molecules featuring thermally delayed activated fluorescence, sustainable copper(I) complexes, quantum dots and luminescent nanoparticles are leading examples of this new type of LECs, see section III of the book. Noteworthy, these scattered works indicate that the LEC behavior holds independently of the type of materials and additives used. As such, the current chapter will focus on providing a general description of LEC operation that is predominantly based on works focusing on CPand iTMC-based LECs.

1.2.2

Figures of Merit and Device Architectures

Possible applications of LECs are commonly thought to be found in domains where the benefits associated with facile large-area fabrication are maximal, like lighting, signage and possibly displays. Consequently, LEC performance is typically described by the following figures of merit: 1. The turn-on time (ton), which is defined by the time between switch-on of the cell by application of a bias voltage and the time at which the luminance reaches

8

2.

3.

4.

5.

6.

S. van Reenen and M. Kemerink

a certain predefined level, typically equal or close to the quasi-steady-state luminance. The luminance (cd/m2) describes the amount of luminous power per unit area, corrected for the wavelength-dependence of the sensitivity of the human eye. It gives a measure of the brightness of the LEC (Bmax). The electroluminescent efficiency (Effmax), expressed either in lm/W (power efficiency) or cd/A (efficacy) where 1 lm = 1 cd
sr. This efficiency relates to the conversion efficiency of electronic carriers in photons, weighted by the sensitivity curve of the human eye. Note that when expressing the efficiency in lm/W, unlike cd/A, the bias voltage that drives operation is taken into account. External quantum efficiency (EQE): The ratio of photons emerging from the device per injected electrons. EQE is also defined through the equation EQE=b//2n2, where b is the recombination efficiency (equal to unity for two ohmic contacts), / is the fraction of excitons that decay radiatively (photoluminescence quantum yields), and n is the refractive index of the glass substrate and is equal to 1.5 (the factor 1/2n2 accounts for the light outcoupling of the device). The lifetime (t1/2), which can be defined either on-shelf or under operation. For the latter one can take as a measure, e.g., the time for the luminance to decay to half-maximum or to below a certain threshold that can depend on the foreseen application. Another parameter is the total emitted energy (Etot). It is calculated by integrating the radiant flux of the device versus time from t=0 (application of bias) to t=t1/5. If this value is divided by the electrode area, it yields the total emitted energy density Utot, which allows devices having electrodes of different shapes to be compared. Color coordinates are used to compare color in a standardized manner. The human eye has color sensors in three different wavelength regions: short (max. 440 nm), medium (max. 535 nm), and long (max. 567 nm), which allow to differentiate about 10 million colors. The number is actually low compared to the amount of colors defined by the Commission Internationale de l’Eclairage (CIE) in 1931. It states that every color can be described by the three color mapping functions
x(k),
y(k) and
z(k).

For research purposes typically two LEC architectures are used, stacked or diode-type, illustrated in Fig. 1.3a, and planar, illustrated in panel b of the same figure. The major advantage of the latter is the good access to the (various parts in the) active layer by microscopic techniques, see, e.g., Figs 1.1, 1.6 and 1.10; the small relative area of the stripe-like emissive zone in planar LECs (Fig. 1.1c) makes this geometry effectively useless for practical applications. As compared to the multilayer stacks encountered in modern OLEDs the stacked LEC layout is extremely simple. In combination with the inherent tolerance to layer thickness variations, which follows from the operational mechanism that is discussed in the next sections, this rationalizes the main asset of LECs: low cost by virtue of facile solution-based fabrication. It should, however, be kept in mind that the total

1 Light-Emitting Electrochemical Cells …

9

Fig. 1.3 Schematic layout of a stacked and planar LECs

Table 1.1 Comparison of different types of organic light-emitting devices, PLEDs, OLEDs and LECs Parameter

PLEDs

OLEDs

LECs

Benefits of LECs

Active layers

1–3

4 or more

1 or 2

Typical thickness per layer Cathode

60–120 nm

1–40 nm

100–500 nm

Simple device architecture Thicker films promise robust processes

Air sensitive

Air sensitive

Air stable

High (10−6 g/m2day)

High (10−6 g/m2day)

Low (10−6 g/m2day)

Solution-based

Vacuum-based

Solution-based

Aromatic

n.a.

Benign (alcohols, water, etc.)

Encapsulation requirements (permeation rate of H2O) Processing of organic layers Solvent

Air stable metals like Al, Ag, Au can be used Air stable electrodes promise less demanding packaging Cost-efficient R2R processing Environmental friendly

manufacturing costs of an LEC module also include materials costs for the active layer (synthesis), as well as for the substrate, electrodes, and packaging. Regarding future application of LECs, it is important to determine the key features of LECs as opposed to related technologies, i.e., OLEDs and PLEDs. Table 1.1 gives a rough overview and summarizes the benefits of LECs. Apart from advantages on conventional substrates like foil or glass, LECs can also be easily fabricated on a large range of different substrates, like fibers that can possibly be used to produce light-emitting clothing, [33] paper, [34] and even on complex-shaped surfaces like kitchen forks [35]. Current industry interest seems focused on OLEDs because of their energy efficiency, color quality, and high-contrast compared to LCD technology. A key selling point of organics, however, has always been to become a low-cost technology, which to date has not

10

S. van Reenen and M. Kemerink

been achieved. LECs, however, can make a strong case to fill this gap as recently described by Sandström and Edman [36]. If LECs are produced with a reasonable luminance of 1000 cd/m2 in a high-volume roll-to-roll-coating scenario, then the cost per lumen would be roughly 0.0036 €/lm. This cost would be one order of magnitude lower than the projected future costs for LEDs and OLEDs.

1.2.3

Suggested Operational Mechanisms for LECs

From a processing and fabrication perspective, the single-layer architecture of LECs is a major benefit. However, the mixing of ionic and electronic conductors has led to a significant complexity of the device physics. Understanding the device physics furthermore suffered hereof furthermore suffered from the large variety of materials studied in LECs, on top of variations that may be expected to arise between works done in different laboratories. With present knowledge, it is quite understandable that the fact that LECs can exhibit rather different behaviors depending on subtle parameter variations, has led to confusion. Over time, various operational mechanisms have been suggested in the literature, which are summarized below.

1.2.3.1

Electrochemical Doping Model (ECDM)

The electrochemical doping model was first proposed by Pei et al. [13, 37] and later on supported by theoretical work by Smith [38] and Manzanares et al. [39]. According to this model, the LEC operation is mainly determined by electrochemical doping of the active layer that results in the in situ formation of a p-i-n structure. A schematic of this model is shown in Fig. 1.4a. Injection of electrons and holes leads, respectively, to the oxidation and reduction of the semiconductor at the anode and the cathode. The oxidized and reduced semiconductor is electrostatically compensated by anions and cations, respectively, which results in p-type and n-type doped regions in the bulk. The region between the doped regions remains intrinsic and is the place where the injected electronic charges recombine. A relatively large electric field is present in the intrinsic region which compensates its relatively low conductivity. This electric field is due to space charge in the junction that originates from electronic carriers. This is in contrast to conventional p-n junctions in which the space charge in the junction originates from dopant ions [38]. The doped regions at the electrode interfaces lead to low-resistance contacts that allow efficient carrier injection despite large injection barriers.

1.2.3.2

Electrodynamic Model (EDM)

The electrodynamic model was first proposed by deMello et al. [41–43]. According to this model, the LEC operation is mainly determined by the formation of electric

1 Light-Emitting Electrochemical Cells …

11

Fig. 1.4 Schematics of the ECDM (left) and the EDM (right), with the associated spatial distribution of the electric field shown below. (reproduced from Ref. [40])

double layers at the interfaces as shown in the schematic in Fig. 1.4b. The electric double layers are formed by mobile ions in the active layer that, upon application of a field, drift towards the electrode interfaces and create large interfacial electric fields. The ions continue to move towards the interfaces until the bulk of the active layer is field-free. The electric fields created at the interfaces promote the injection of electronic charge carriers. The injected carriers then move through the bulk by diffusion and recombine when electrons and holes meet.

1.2.3.3

Preferential Electrochemical Doping Model (PECDM)

The preferential electrochemical doping model was first proposed by Leger et al. [44]. This model shows high similarity to the ECDM with the main difference that only one type of doping occurs in the active layer: either n-type or p-type doping [44–46]. As a result, recombination occurs close to one of the electrode interfaces.

12

S. van Reenen and M. Kemerink

Table 1.2 Overview of the universal operational mechanism of light-emitting electrochemical cells Light-emitting electrochemical cells Requirement (depends 0 ohmic contacts on applied bias voltage and injection barriers) Operating mechanism:

Voltage distribution (back line) and recombination zone (orange) Electron (•), Hole(∘), anion (⊕), cation (⊖) distributions:

1.2.4

Electrodynamic model (EDM)

+

+ -

Can be anywhere in active layer

-

+ + +

1 ohmic contact

2 ohmic contacts

preferential Electrochecmical doping model (pECDM)

Electrochecmical doping model (ECDM) -

+

+ + + + - ++ + + + + + -

+

+

-

- - -

+ + + + + + + -

Current Understanding of Operational Mechanism of LECs

The variation in operational behavior of LECs reported experimentally and numerically indicates that LECs operate in different regimes. Systematic experiments [46, 47] and numerical modeling [47, 48] have confirmed this fact and showed that the three models, the EDM, the PECDM, and the ECDM coexist. The applicability of each model was found to depend on the ability to form Ohmic injecting contacts, [47] which depends on a combination of applied bias voltage and the height of the barriers for carrier injection [47, 48]. In case no Ohmic contacts are formed, the LEC follows the EDM. In case one Ohmic contact is formed, the LEC follows the PECDM. Here, p-type doping or n-type doping occurs in case an Ohmic contact is formed at the anode or the cathode, respectively. In case two Ohmic contacts are formed, the LEC follows the ECDM. An overview of this unifying model is shown in Table 1.2. This unifying model has been confirmed in both CP- and iTMC-based LECs by various experimental and numerical studies [13, 14, 38, 41, 45–47, 49–55]. From an application perspective, device operation in the ECDM mode is preferred [47].

1.2.5

Basic Equations to Describe LEC Operation

Here, the basic equations are summarized that describe the various processes related to charge injection, transport, and recombination in the active layer of LECs. Lateral heterogeneity that can arise from, for example, phase separation will be

1 Light-Emitting Electrochemical Cells …

13

ignored. Hence, a 1D model suffices. Although computationally cumbersome, extension to 2D or 3D is straightforward.

1.2.5.1

Drift and Diffusion for Ionic and Electronic Charges

Transport of electrons, holes, anions, and cations can be described by drift and diffusion where the diffusivity D is related to the mobility l by the Einstein relation Di ¼ li kB T: Ji ¼ ni qli

dV dni þ bDi ; dx dx

ð1:1Þ

where b is +1 for electrons (i = n) and anions (i = a), and −1 for holes (i = p) and cations (i = c). Moreover, Ji is the current density of species i, ni the charge density, q the elementary charge, V the electrostatic potential, x the position, kB the Boltzmann constant, and T the temperature.

1.2.5.2

Poisson’s Equation

The electrostatic potential and field, and the density of electrical charge carriers are related through Poisson’s equation: r2 V ¼ rE ¼ 


q np  nn þ nc  na e0 er

ð1:2Þ

where E is the electrostatic field, e0 the vacuum permittivity, and er the relative permittivity of the material.

1.2.5.3

Binding Energy for Anion/Cation and Ion/Electronic Charge Pairs

The mobile ions in the active layer of LECs can recombine with each other to form neutral salt molecules with a binding energy Eac. The anion–cation capture rate Cac, dissociation rate Bac, and net binding rate Uac can be described, respectively, by Cac ¼ cac
na
nc ;

ð1:3Þ

Bac ¼ Kac
bac
nac ;

ð1:4Þ

Uac ¼ Cac  Bac ;

ð1:5Þ

14

S. van Reenen and M. Kemerink

where cac is the anion–cation capture rate coefficient and nac the density of bound anion–cation pairs. Kac is a mass-action law constant, and bac the anion–cation dissociation rate coefficient that can be described by   Eac bac ¼ cac
exp  : kT

ð1:6Þ

Although this has not been explored, a binding energy is also expected for doping complexes, i.e., between anions and holes (Eap) and between cations and electrons (Ecn). In analogy to Eqs. (1.3)–(1.6), the anion-hole capture rate Cap, dissociation rate Bap, and net binding rate Uap can be described, respectively, by Cap ¼ cap
na
np ;

ð1:7Þ

Bap ¼ Kap
bap
nap ;

ð1:8Þ

Uap ¼ Cap  Bap ;

ð1:9Þ

where cap is the anion-hole capture rate coefficient and nap the p-type doping density. Kap is a mass-action law constant, and bap the anion-hole dissociation rate coefficient that can be described by bap

  Eap ¼ cap
exp  : kT

ð1:10Þ

Cation-electron binding can be described by expressions similar to Eq. (1.7)– (1.10).

1.2.5.4

Electron–Hole Recombination

Recombination between electrons and holes in electroluminescent materials leads to the formation of exciton complexes that can decay radiatively. The materials used in LECs are typically low dielectric constant materials, leading to strong Coulombic interaction between the electron and hole in the exciton complex. As a result, the excitons are relatively small and typically reside on a single unit cell. The recombination process can be described by either Langevin recombination [56] or trap-assisted recombination, i.e., Shockley-Read-Hall (SRH) recombination [57, 58]. In case of Langevin recombination, the recombination occurs between free electrons and holes that drift towards each other. In case of SRH recombination, recombination occurs between a trapped and a free carrier. Here, the recombination rate is set by the mobility of the free charge carrier. Hence, both Langevin recombination and SRH recombination are governed by charge carrier mobilities [59]. The Langevin recombination rate RL is described by

1 Light-Emitting Electrochemical Cells …


q ln þ l p RL ¼ np nn : e0 er

1.2.5.5

15

ð1:11Þ

Continuity Equations

The continuity equations for electrons, holes, anions, and cations can be described, respectively, by dnn 1 dJn ¼  RL  Ucn ; q dx dt

ð1:12Þ

dnp 1 dJp ¼  RL  Uap ; q dx dt

ð1:13Þ

dna 1 dJa ¼  Uac  Uap ; q dx dt

ð1:14Þ

dnc 1 dJc ¼  Uac  Ucn : q dx dt

ð1:15Þ

Continuity equations for salt and doping complexes can be described, respectively, by

1.2.5.6

dns ¼ Uac ; dt

ð1:16Þ

dnap ¼ Uap ; dt

ð1:17Þ

dncn ¼ Ucn : dt

ð1:18Þ

Boundary Conditions

The density of the electronic charge carriers at the interfaces, are determined by the model that describes carrier injection. A realistic model is the Emtage-O’Dwyer model [60] that describes injection with an exponential electric field dependence. This model also considers the injection barrier created by the energy level offset between the electrode Fermi-level and the transport level in the semiconductor in which the charge carriers are injected. In addition, it also considers stabilization of injected charge carriers in the semiconductor by the formation of an image charge in the electrode. Implementation of the Emtage-O’Dwyer injection model is numerically challenging as a relatively small grid-point spacing is required, which can lead

16

S. van Reenen and M. Kemerink

to an enormous grid when equal grid-spacings are required [61]. deMello [41] solved this problem by using an adaptive grid with a variable grid-point spacing. Van Reenen et al. [47] used a ‘modified Boltzmann’ model which ignores the grid-point spacing related to carrier injection, but simply assumes thermal activation of charge carriers over an injection barrier that is modified by the local electric field. In LECs this field results typically from the formation of an electric double layer at the interfaces due to pile-up of ions. For ions, the electrodes are typically assumed blocking. Experiments in planar cells have, however, suggested that ions can penetrate and even travel through Au electrodes [14].

1.3 1.3.1

Transient Phenomena Turn-on and the Role of Ion Motion

The presence of ions in LECs makes operation of these cells strongly time-dependent. Processes related to electronic transport in organic semiconductors typically take far less than a second to reach quasi-steady state. Ions, however, are much slower as they physically move through a solid-state material, leading to turn-on times that are relatively long as compared to OLEDs. Turn-on times in LECs can range from several seconds [13] to several hours, [62] dependent on the combination of the constituents and the active layer thickness as well as the applied bias voltage. For the vast amount of LEC device configurations studied and reported in the literature, large quantitative differences are observed in the turn-on transients. Nevertheless, the turn-on transients of CP-based LECs in stacked [63, 64] and planar [47, 65–68] configuration show strong qualitative resemblance with iTMC-based LECs in stacked [18, 54, 62, 69–72] and planar configuration [53]. This resemblance was studied systematically by Van Reenen et al. [55]. They studied the time-dependent current, luminance, and efficacy just after switch-on of freshly prepared CP- and iTMC-based LECs. They found that the timescale at which turn-on occurs is strongly affected by the device temperature. By normalizing to the turn-on time this temperature dependence is scaled out. Furthermore, if the current, luminance and efficacy are normalized as well, the transients of both types of LECs are found to follow a universal shape as shown in Fig. 1.5. Moreover, the activation energy of the turn-on time and the ion conductivity measured in the off-state were found to be the same, which substantiates that the turn-on of LECs is determined by the ion conduction. In the remainder of this paragraph the turn-on behavior of planar and stacked LECs is reviewed separately in more detail. Being not fundamental, the differences in turn-on between CP- and iTMC-based LECs will not be considered.

1 Light-Emitting Electrochemical Cells …

17

Fig. 1.5 Normalized current, luminance and efficacy transients of CP- and iTMC-based LECs at two temperatures each and biased at 3.5 V. (reproduced from Ref. [55])

1.3.1.1

Studies in Planar LECs

The ability to construct functional LECs in planar cell configuration has given researchers access to the active layer for various experimental surface techniques. As shown in Fig. 1.1c, this allowed researchers to study the electroluminescence (EL) position in LECs. Besides EL, also photoluminescence (PL) can be studied in planar LECs, using UV illumination to excite the semiconductor. Electrochemical doping of semiconductors is known to quench PL [73]. Hence, this technique allows one to keep track of the electrochemical doping process in LECs during turn-on [50, 53, 66, 74]. A typical example of such UV-excited PL measurements in planar LECs is shown in Fig. 1.6a [50]. The experiments show that p- and n-type electrochemical doping occurs in the LEC by accelerating doping fronts that move through the active layer, starting from the electrodes, until both meet [75]. The position where the p- and n-type doping fronts meet coincides with the region where light emission takes place as shown in the last photograph in Fig. 1.6a. The corresponding device current grows during front propagation and continues to grow after the fronts connect (Fig. 1.6b).This continuous growth in current is attributed to on-going electrochemical doping of the doped regions. The enhanced quenching of PL observed in Fig. 1.6a after doping front connection further constitutes this view. Other work has also shown that besides this continuation of doping after front connection, the recombination zone can also shift towards the anode or the cathode with time [53, 76]. Van Reenen et al. studied the turn-on in planar cells by means of a numerical drift-diffusion model, based on the equations described in Sect. 1.2.5. It was possible to calculate a current transient (see Fig. 1.6c) that qualitatively reproduces the experiment (Fig. 1.6b) [77]. The potential profile evolution in the active layer was studied experimentally, by use of scanning Kelvin probe microscopy (see Fig. 1.6d), and numerically (see Fig. 1.6e). This combined work gives the following picture of the turn-on of planar LECs. After switch-on of the bias voltage, electric double layers form at the interfaces, see Fig. 1.6d, e, enabling carrier injection and consequently n-type and p-type

18

S. van Reenen and M. Kemerink

Fig. 1.6 Top Photographs under UV illumination of planar CP-based LECs with an electrode gap of 90 lm during operation at V = 8 V and T = 333 K. The positive and negative electrodes are indicated in the photographs and the dashed lines indicate where they contact the active layer. Central left Experimental current measured during turn-on of the planar CP-based LEC shown at the top pictures. Central right Modeled current in a planar CP-based LEC. Bottom left Experimental potential profile evolution during turn-on of a similar planar CP-based LEC as shown at the top pictures. Bottom right Modeled potential profile evolution of a planar CP-based LEC. (adapted from Ref. [77])

electrochemical doping of the active layer. The electrochemically doped regions grow towards the opposite electrodes until they meet. When the fronts connect, the potential is distributed more or less evenly across the active layer as shown in Fig. 1.6d (at t * 10 s) and e (as indicated). Doping continues as there are still mobile ions available that do not yet contribute to electrochemical doping. Hence the current continues to grow. The current reaches a peak value in both the model and experiment. Around this time, the potential is observed to change dramatically: the potential becomes distributed mainly in the region where recombination takes place as shown in Fig. 1.6d (at t * 30 s) and e (as indicated). The model shows that at the same time, the recombination zone becomes depleted of ions so that the

1 Light-Emitting Electrochemical Cells …

19

Fig. 1.7 Top left Normalized front position during switch-on in an LEC as obtained experimentally (blue circles) [65], analytically (red line) [78], and numerically (black line) [79]. Top, right Schematic of the ion mobility criteria with respect to the doping-dependent electron and hole mobility (blue line) that result in the formation of accelerating doping fronts in LECs during switch-on. Bottom Transient electron (closed circles) and hole (open circles) density profiles in LECs for a constant (left) and doping-dependent electron/hole mobility (right) and an ion mobility of 5·10−11 m2V−1s−1. The anode and cathode are positioned at position = 0 and 2000 nm, respectively. Movement in time is expressed by the color change of the graphs from light gray to black to red. (reproduced from Ref. [79])

recombination zone becomes intrinsic, i.e., undoped, and therefore has a lower conductivity. This lower conductivity in the recombination zone, as compared to the high conductivity in the doped regions, necessitates a larger field in the recombination zone to have current conservation across the device. The voltage redistribution towards the recombination zone leads to reduction in current in the later stages of the transients (see Fig. 1.6b, c). The formation of doping fronts in planar LECs to electrochemically dope the active layer has been a major topic of study as it reveals various interesting properties regarding carrier mobilities and local field distributions. A report by Shin et al. showed that the doping fronts typically accelerate during propagation, as shown in Fig. 1.7a (blue circles) [65]. Robinson et al. used an analytical model to describe this behavior. The model is based on matching the current density through the doped regions with the ion current in the intrinsic region. They assumed that the applied potential drops nearly completely over the intrinsic region during the turn-on of the cell because of the mismatch in conductivity between the intrinsic region and the doped regions. A fit of the analytical model shown in Fig. 1.7a (red line) shows that the model indeed predicts accelerating doping fronts. The assumption regarding the potential distribution, however,

20

S. van Reenen and M. Kemerink

was later proven to not necessarily be true, [77] as shown in Fig. 1.6d, e: the potential profile hardly changes during doping front propagation. An attempt to model doping front propagation numerically was made using a drift-diffusion model of a planar LEC [79]. It was found that doping-dependent electron and hole mobilities are required to achieve pronounced doping fronts (Fig. 1.7d). Without such a doping dependence, electrochemical doping of the active layer would occur more gradually as shown in Fig. 1.7c. More specifically, to achieve accelerating doping fronts, the mobility of electrons and holes in the undoped region, lp/n,0, needs to be similar in order of magnitude to the ion mobility (schematic in Fig. 1.7b). The doping density dependence of the mobility in electrochemically doped systems is relatively strong compared to, e.g, field effect mobility enhancements [80]. A low mobility in weakly doped semiconductors is likely due to localization of the electronic carriers at the electrostatically compensating doping sites: [81] doping sites act like charge traps at low doping densities. At higher doping densities, this localization is lifted as there are more neighboring doping sites available to hop to without the need to gain additional energy to escape the energetically favorable trap [80, 81]. Hence, the mobility of the charge carrier is significantly enhanced. Another typical feature of doping fronts in planar LECs is the instability at the doping fronts, which leads to the formation of ‘fingers’. An example of this is shown in Fig. 1.8a. This instability was furthermore found to be enhanced by the use of higher operational voltages [78]. To understand this behavior, Bychkov et al. carried out various numerical modeling studies on doping front propagation [82, 83]. They found that the local electric field at the apexes of the doping front is relatively high (see Fig. 1.8b, c). This increase of the local electric field then accelerates the doping front propagation locally, which leads to an enhancement of the finger shape. This process explains the instability at the doping fronts as shown

Fig. 1.8 Left, top Experimental photo of a p-type doping front in a planar CP-based LEC. The white line is a simulation of the p-type front shape. Left central and bottom show the relative increase in the electric field in the undoped region obtained numerically for b the whole front and c a selected part with equal color scales as indicated below. Right Schematic of the p- and n-type doping fronts in a LEC. (reproduced from Ref. [82])

1 Light-Emitting Electrochemical Cells …

21

in Fig. 1.8a and allowed the researchers to model a similarly shaped doping front (see Fig. 1.8a white line) [82].

1.3.1.2

Studies in Stacked LECs

Like in planar cells, the turn-on time in stacked cells is dominated by ion transport. A logarithmic plot of a current transient, determined from a CP-based LEC, is shown in Fig. 1.9a. Van Reenen et al. carried out a study to model these transients [63]. By experimental determination of the ion conductivity using electrochemical impedance spectroscopy, [63] they tried to use the determined ion mobility and density to model the turn-on transient of the same CP-based LEC, see Fig. 1.9 dashed line. The timescale of the modeled transient was, however, found to be completely off compared to the experiment: e.g, the modeled current reached a quasi-steady-state around t = 100 s, whereas in the experiment it took over 101 s to do so. Only by including a binding energy (see Eq. (1.3)–(1.6)) between the ions, in this case of 0.15 eV, [38] transients could be modeled that took over 100 s to reach steady-state while taking the measured ion conductivity values into account (see Fig. 1.9b solid line). Turn-on times in the order of seconds or even larger are not suitable for various applications of LECs, like displays. Consequently, much effort has been put into improvement of the turn-on time. Addition of the ion-dissolving poly(ethylene oxide) (PEO) has been found to accelerate turn-on significantly in CP-based LECs [13] as well as iTMC-LECs [69]. This addition possibly leads to improved ion conduction and enhanced salt dissociation as PEO has a relatively high dielectric constant of 6 as compared to the dielectric constant of typical CPs around 3 [84]. Improvement of the electrolyte by use of different ion-solvating polymers has already led to significant improvement of the turn-on time in LECs, besides improved lifetime [85–88]. Ion conduction and salt dissociation also depend on the

Fig. 1.9 Left Experimental current transient of a pristine CP-based LEC biased at 3.5 V. (adapted from Ref. [55]) Right Modeled current transient of a LEC biased at 3.5 V with (straight line) and without (dashed line) binding energy between anions and cations. (reproduced from Ref. [63])

22

S. van Reenen and M. Kemerink

type of ions used in LECs. Therefore the use of other ions [65, 89, 90] or ionic liquids [91] in LECs can also lead to improved turn-on times. Turn-on can also be accelerated by driving LECs with a fixed current instead of voltage [62, 92, 93]. This typically translates into a relatively high initial bias voltage applied to the cell when it is in its least conductive state. Use of a pulsed driving scheme was furthermore shown to be able to stabilize the device in an intermediate state during turn-on of the cell, preventing complete electrochemical doping of the active layer and the associated efficiency roll-off that is further discussed in Sect. 1.4.4 [94]. Another method to avoid long turn-on times in LECs is by preparing the LEC in such a way that electrochemical doping is fixed. This can be done chemically by fixing doping complexes using, e.g., polymerization reactions of dopant ions [95, 96]. Alternatively, ions can be frozen into position by lowering the device temperature after electrochemical doping [97–99]. This topic will be discussed further in the next section.

1.3.2

Polarization Reversal and Hysteresis

The dynamic character of mobile ions gives LECs the unique feature that they can be operated efficiently both in forward and reverse bias conditions, as illustrated already in Fig. 1.1b. Polarization reversal was studied in planar cells by scanning Kelvin probe microscopy by Matyba et al. [14]. Steady-state profiles of the potential are shown in Fig. 1.10a and c where a planar CP-based LEC was subsequently biased +5 V and −5 V. The potential profiles are essentially mirror images of each other, which indicates that the processes related to the operation of LECs are highly reversible and therefore enable polarization reversal without significant penalties relating to irreversible side reactions. Figure 1.10b shows the complex evolution of the potential profile during a polarization reversal. Despite the electrochemical doping process being essentially reversible, hysteresis effects are significant in LECs. Li et al. carried out various experiments to study the effects of relaxation of doping in stacked CP-based LECs [100]. Prolonged operation of LECs typically leads to reduction of the electroluminescence by electrochemical doping, similar to PL quenching in planar cells (see, e.g., Figure 1.6a). Li et al. studied the recovery of luminance by relaxing the cell at open-circuit voltage after prolonged operation, as shown in Fig. 1.11. The cells required a relatively long time at elevated temperatures (to speed up the relaxation process) to recover a large part of the luminance. During this time, part of the electrochemical doping is removed by dedoping of the cell. The PL images in Fig. 1.11 (compare top and middle photograph of the device) furthermore confirm that PL quenching is significantly reduced after allowing the cell to relax for 7 h at 60 °C. Although part of the luminance is recovered after a long time of relaxation, another part did not recover which indicates the occurrence of irreversible reactions in the LEC during operation. These will be further discussed in Sect. 1.3.3.

1 Light-Emitting Electrochemical Cells …

23

Fig. 1.10 Experimental data illustrating the polarization reversal in planar CP-based LECs by scanning Kelvin probe microscopy. Steady-state operation at Vbias = +5 V (top). Temporal evolution after subsequent switch to Vbias = −5 V where the arrow indicates the time (central). Steady-state operation at Vbias = −5 V (bottom). The dashed lines indicate the electrode positions as determined from height data from atomic force microscopy measurements. (reproduced from Ref. [14])

Reduction of hysteresis in LECs can be achieved by fixation of the electrochemical doping. One method is to pre-bias the device at elevated temperature at which ions are mobile, followed by reduction of the device temperature, leading to significant reduction of the ion conductivity [97–99]. For practical applications, the latter requires that the freeze-in temperature is significantly above room temperature. Edman substituted PEO in planar CP-based LECs by a crown ether that melts at a temperature of 56 °C [98]. The device was switched on at 85 °C and subsequently cooled down to room temperature. Consequently, the p- and n-type doped regions were frozen-in by crystallization of the crown-ether phase. The same can be achieved when using PEO by cooling the device down to 100 K, which is below the glass transition temperature of PEO [97]. In both instances, the response times of the LECs improved significantly, enabling frozen-junction LECs to match the response times of similar polymer-based LEDs [97]. Chemical binding of doping offers an alternative solution to reduce hysteresis and improve turn-on of LECs [95, 96]. Hoven et al. reported an LEC that consisted of a bilayer structure [95]. This bilayer comprised a film based on a cationic conjugated polyelectrolyte with mobile fluoride counter (an)ions (PFP-F) and a film based on a neutral CPs with functional groups that enabled trapping of fluoride anions (PFP-BMes) as shown in Fig. 1.12. By application of a bias voltage, the mobile fluoride anions move from the PFP-F film into the PFP-BMes film to enable electrochemical doping. The fluoride is then covalently bonded to the PFP-BMes. As a result, these devices have reduced hysteresis and do not relax back to the

24

S. van Reenen and M. Kemerink

Fig. 1.11 Time evolution of luminance of a stacked CP-based LEC at a constant current of 167 mA cm−2. The device was tested multiple times after storage for 17 days since the last run at 167 mA cm−2. Delay and heating applied to the cell are indicated in the graph. The images of the cell to the right show (top) fluorescent image after run 7 before heating; (middle) fluorescent image after run 7 and 7 h of heating at 60 °C; (bottom) electroluminescent image after run 8 and 7 h of heating at 60 °C. (reproduced from Ref. [100])

pristine state by dedoping after initial charging. Furthermore, this device showed rectification unlike typical dynamic LECs.

1.3.3

Degradation, Side Reactions, and Electrochemical Stability

Before discussing degradation and side reactions in LECs, it is important to note that reduction in luminance and efficiency, as well as current over time in LECs is not necessarily related to irreversible side reactions. Luminescence quenching and a reduction in efficiency are generally observed in LECs but can be partially recovered by relaxation of the cell, as illustrated in Sect. 1.3.2. Meier et al. studied reversible and irreversible effects in stacked iTMC-LECs during operation by measurement of the PL intensity as a function of operating time (see Fig. 1.13) [52]. At different moments during the operating time, the cells were allowed to recover by turning off the bias voltage. Recovery of the PL required several hours. Moreover, as shown in Fig. 1.13, only part of the PL could be recovered, dependent

1 Light-Emitting Electrochemical Cells …

25

Fig. 1.12 Design and materials for chemically fixed heterojunctions. Top molecular structures of (left) PFP-BMes and (right) PFP-F. Central pristine device with all ions in the FPF-F layer (left, dark blue), under an applied bias (central), the mobile fluoride anions (yellow) move into the FPF-BMes layer (light blue), where also electronic carriers are injected from the cathode (gray) and anode (white), the ions are compensated by injected charge carriers creating a p-n junction (right). A new immobile borate species (red) is formed. (reproduced from Ref. [95])

26

S. van Reenen and M. Kemerink

Fig. 1.13 Visualization of the reversible recovery and irreversible loss of PL intensity in a stacked iTMC-LEC as a function of operating time. The lines serve as a guide to the eye. (reproduced from Ref. [52])

Fig. 1.14 Time evolution of luminance of a stacked CP-based LEC at a constant current of 167 mA cm−2. The operating time on the horizontal axis indicates accumulated run time under bias. The cell was stored for 30 days in a N2-filled glovebox at room temperature after each run. The first run started right after deposition of the top Al electrode. Photographs of the electroluminescence of the LECs were taken at the end of the runs under which they are shown. (reproduced from Ref. [101])

on the operating time. This further proves that reversible and irreversible processes occur simultaneously and quench the luminescence in these LECs. In the literature, various irreversible processes have been observed and studied that lead to reduced performance of LECs. Here we will shortly mention the main processes.

1 Light-Emitting Electrochemical Cells …

27

Fig. 1.15 Left Schematic electron-energy level diagram for an LEC, with the reduction level for the PEO + KCF3SO3 electrolyte positioned within the bandgap of the CP (MEH-PPV). Right Optical microscopy images of the anodic (left) and cathodic (right) interfaces after 12 h of operation at V = 30 V and T = 360 K of a planar CP-based LEC with a 1 cm interelectrode spacing. The line just left of the cathode appears to be due to cathodic electrochemical side reactions. (reproduced from Ref. [103])

Black spot formation is typically observed in LECs after prolonged operation as shown in Fig. 1.14 [101]. In the results shown here, the irreversible decay of luminance is concomitant with black spot formation in the polymer film. The black spots lead to an effective reduction of the emitting area. In the same paper, and in a follow-up work, [102] the authors show that black spots are formed in unbiased cells only if the cells are stored after deposition of the top Al electrode, indicating that black spots do not necessarily appear during operation of LECs. AlTal et al. [102] found that black spots in EL also appeared in PL after prolonged operation. However, after switch-off of the bias voltage, the black spots in the PL were observed to disappear. Together these results strongly indicate that black spots may be due to heavy doping, promoted by chemical changes occurring at the cathode/polymer interface. The exact origin of black spots formation in LECs has, however, not been identified. Similar black spots have also been observed in LECs comprising iTMCs by Kalyuzhny et al. [71]. The researchers found that near the cathode where light emission takes place, a quencher is formed by side reactions with the iTMC that are assisted by moisture or oxygen. Another cause of irreversible degradation in LECs is a side-reaction with the ion transport material PEO. In case of typical CP-based LECs based on PPV and PEO, the reduction level of the electrolyte lies below the conduction band of the CP (see Fig. 1.15a). Consequently, injection of electrons results in reduction of either the electrolyte or the CP [76, 103]. Although the former is energetically favorable, the latter is kinetically preferred as electronic carrier transport though the electrolyte is low due to the electrically insulating character of PEO. Nevertheless, side reactions with the PEO are likely to occur at the cathode interface, as evidenced in the planar LEC shown in Fig. 1.15b. Such reactions may hamper electron injection, causing an imbalance in carrier transport through the active layer. Such imbalance can ultimately lead to microshorts [104], when one of the doped regions grows completely from one electrode towards the other. Side reactions with PEO may also be

28

S. van Reenen and M. Kemerink

related to black spot formation as discussed above. Lowering of the conduction band of the CP by use of different materials is a route to avoid such side reactions and improve the overall performance of LECs [105]. Use of alternative ion-solvating polymers has led to improved lifetime of LECs by improvement of the electrochemical stability window of the electrolyte [85–88].

1.4

Steady-State Phenomena

1.4.1

Potential and Ion Distribution

The potential profile in an LEC can be regarded as the fingerprint of the operating mechanism that the LEC follows. The ion distribution furthermore determines the potential profile, dependent on the formation of electric double layers and electrochemically doped regions. Characterization of either of the two therefore gives crucial information on the operation of the device, see also Table 1.2.

1.4.1.1

EDM

Stacked LECs typically have a too small interelectrode distance to enable surface techniques like electrostatic force microscopy (EFM) or scanning Kelvin probe microscopy (SKPM) to directly measure the potential distribution in them. Planar cells are however sufficiently large to allow such techniques to extract useful information on the potential distribution. During EFM, a conductive tip is scanned in non-contact mode over an area without a feedback mechanism on the potential. This is opposite to SKPM, where a DC bias feedback is in place which keeps the electrostatic potential difference between the conductive tip and the sample at zero. For LECs following the EDM, the potential distribution is dominated by large potential drops at the electrode interfaces and a field-free bulk [47, 51]. Figure 1.16 shows the device layout and the typical potential distribution in an LEC following the EDM from EFM measurements on a planar Ru-based iTMC-LEC [51]. In CP-based LECs, similar potential profiles were obtained by SKPM in an N2-filled glovebox only after allowing the contacting Al electrodes to oxidize by contact to air [47]. Consequently, the appearance of the EDM model is assigned to devices with relatively poor carrier injection properties. The experiments show that the potential drop in LECs following the EDM is mainly located at the electrode interfaces due to formation of ionic space charge. As a result, the bulk of the active layer becomes field-free. Numerical modeling by deMello et al. [41–43] and van Reenen et al. [47] confirm this picture and confirm the requirement of relatively weak carrier injection to get into the EDM operating regime. Due to the absence of significant electrochemical doping, the majority of mobile ions remains distributed throughout the bulk of the active layer resulting in the absence of net ionic charge.

1 Light-Emitting Electrochemical Cells …

Au

29

Au

Substrate

Fig. 1.16 Top Schematic diagram of an unpatterned planar iTMC-LEC. Bottom Experimental characterization of the time dependence (green to red; spaced by equal increments of 15 min) of the potential profiles for an Au/[Ru(bpy)3][PF6]2/Au device at Vbias = 5 V by EFM. The electrode positions are indicated by the dashed lines. (reproduced from Ref. [51])

1.4.1.2

ECDM

Matyba et al. first reported on the potential distribution in planar CP-based LECs following the ECDM as shown in Fig. 1.17 [14]. Figure 1.17b clearly demonstrates a significant potential drop in the bulk of the device that indicates a (dynamic) p-i-n junction. Immediately after release of the bias voltage in this cell, i.e., at open-circuit conditions, a built-in potential of 1.5 V can be observed at the same position (see Fig. 1.17c), which is slightly smaller than the bandgap of the semiconductor used in this cell. This built-in voltage indicates the presence of a connection between an n-type doped semiconductor (on the left) and a p-type doped semiconductor (on the right). In Fig. 1.17b, no potential drops are observed at the interfaces. In other experiments, however, such potential drops have been observed upon use of Al injecting contacts [47]. This discrepancy is suggested to be due to the formation of a thin layer of ion-containing material on top of the Au electrodes that screens some of the potential [14]. The formation of this thin layer may be due to ions diffusing through the Au electrodes, which is not possible in Al electrodes; the potential drops observed at Al electrode interfaces were found to be similar to the expected energetic injection barriers, determined from the difference in electrode Fermi-level and the positions of the HOMO and LUMO levels of semiconductor.

30

S. van Reenen and M. Kemerink

Fig. 1.17 Experimental SKPM data illustrating the potential distribution in a planar CP-based LEC following the ECDM, showing 2D topography image by atomic force microscopy (top, left), electrostatic potential profile during steady-state operation at Vbias = +5 V (central, left), transient potential profile measured with the device disconnected (bottom, left), directly after operation at Vbias = +5 V, micrograph showing light emission from the same device during steady-state operation at Vbias = +5 V.(right) (reproduced from Ref. [14])

Similar potential profiles relating to the ECDM have also been observed in planar iTMC-LECs [53]. A technique that can be used to experimentally determine the ion distribution in stacked LECs is time-of-flight secondary ion mass spectroscopy (ToF-SIMS) [106, 107]. ToF-SIMS basically consists of sputtering the sample with a focused ion beam while monitoring the secondary ions. This way a depth profile of system components can be determined, which can give information on the ion distribution in LECs. One of the difficulties of this technique is that the ions used for sputtering induce positive charges on the thin film surface [108]. The induced charges give rise to large electric fields, which disturb the ionic charge distribution in LECs studied by ToF-SIMS. Nevertheless, side-by-side comparison between devices can still give information on ionic charge distributions in LECs [107]. ToF-SIMS was carried out on stacked CP-based LECs as shown in Fig. 1.18 [107]. The results in Fig. 1.18b show first of all that an unbiased device (red line/symbols) suffers from the charge redistribution caused by the sputtering process. Comparison with devices biased at 7 V for 2 min (blue line/symbols) and for 3 min (green line/symbols), however, shows that because of prior device operation, Li+ and CF3SO3−ions have

1 Light-Emitting Electrochemical Cells …

31

Fig. 1.18 Left Schematics illustrating a standard ToF-SIMS measurement on a LEC device comprising the salt Li+CF3SO3−. Right ToF-SIMS profiles of Li+ (lines) and F− (circles; F− originates from CF3SO3−) for dynamic junction LECs with a 100 nm thick active layer charged at 7 V for 3 min (blue), 2 min (green), and uncharged (red). (reproduced from Ref. [107])

Fig. 1.19 Experimental SKPM data illustrating the potential distribution in a planar CP-based LEC with Ca electrodes operated at Vbias = 5 V following the PECDM. The dashed lines indicate the electrode positions. (reproduced from Ref. [46])

moved towards the cathode and anode, respectively. This behavior is in line with the ECDM. The combined results on potential and ion density profile distributions show that LEC operating according to the ECDM, separate anions from cations to form electric double layers at the interfaces and doped regions in the bulk. The doped regions are separated by a low-conductive intrinsic region. Consequently, sharp potential drops are observed at the interfaces and in a small region in the bulk. Numerical calculations confirm these experiments [38, 47].

1.4.1.3

PECDM

Potential profiles of LECs following the PECDM were first reported by Pingree et al. [45] and later on by Rodovsky et al. [46] as shown in Fig. 1.19. The majority of the potential drops at the injecting contact that has the largest barrier for carrier injection. For the device shown in Fig. 1.19, low-work function Ca electrodes were used to contact the active layer to inject electrons and holes in the polymer MDMO-PPV. Use of Ca leads to a large injection barrier for hole injection and a

32

S. van Reenen and M. Kemerink

small injection barrier for electron injection. As a result, the active layer is predominantly n-type doped and the light-emitting junction is positioned at the contact with the largest injection barrier, being the anode. Therefore, this device shows preferred n-type doping. Devices with preferred p-type doping have also been reported [45]. No numerical modeling has been reported in which potential profiles are calculated in devices following the PECDM.

1.4.2

Position and Width of the Recombination Zone

1.4.2.1

Studies in Planar LECs

Properties of the recombination zone depend strongly on the operation mechanism in LECs. These have been studied extensively in planar LECs by photographs of the device during operation. For planar cells operating in the EDM or the PECDM, the recombination zone is observed to sit close to one or even both of the electrode interfaces [46, 51]. For planar cells operating in the ECDM, however, the position of the light-emitting junction can be found anywhere in the bulk of the device, dependent on properties related to carrier injection [45, 46] and doping-dependent transport [78, 79]. Moreover, as these properties change during the whole electrochemical doping process of the active layer, the junction position is also found to move during operation [53]. For device performance, control of the junction position is important as to avoid the junction to lie close to the electrodes where strong EL quenching takes place. The initial position of the light-emitting junction is determined by the time required to form Ohmic contacts at both electrodes and the doping front propagation. Ohmic contact formation takes longer in case a larger injection barrier needs to be overcome by EDL formation [45, 46]. Immediately when an Ohmic contact is formed, carriers are injected, which initiates the doping front propagation towards the opposite electrode. Doping front propagation itself depends strongly on the mobility of the electronic carriers that must be supplied to the front for electrochemical doping [79]. PPV-based planar CP-based LECs with Au electrodes have a relatively large injection barrier for electrons compared to holes, as well as a hole mobility that is one order of magnitude above the electron mobility [109]. Hence, the initial position of the light-emitting junction in these devices is typically found near the cathode [46]. Under such conditions, distinguishing ECD and PECDM from ECDM behavior can be tricky. The position of the junction region can, however, shift after initial formation because of continuous electrochemical doping [53]. The final position of the junction region is then strongly determined by the respective electronic conductivity of the n-type and p-type doped regions, which must lead to balanced transport of electrons and holes towards the recombination zone [47, 53, 77, 78].

1 Light-Emitting Electrochemical Cells …

33

The width of the junction in planar cells is found to range from 1–3 lm, independent of the interelectrode distance [13, 46].

1.4.2.2

Electrical Impedance Spectroscopy

Studies on stacked LECs often employ electrical impedance spectroscopy (EIS) in which the large differences in time scales associated with the various physical and chemical processes are used to separate these processes. In particular, EIS measures the dielectric properties of a medium as a function of frequency. It is based on the perturbation of a quasi-equilibrium state by an AC bias voltage Vac. This perturbation must be sufficiently small so that a linear response can be assumed. The ~ perturbation results in a change in current that follows the oscillating bias voltage V ~ with a phase difference. Measurement of this oscillating current I allows determination of the complex admittance Y: Y¼

~I ¼ G þ jC; ~ V

ð1:19Þ

with G the (real) conductance and C the (real) capacitance. Dependent of the device under test, the overall conductance and capacitance are, however, the result of a complex combination of various processes in the device. Hence interpretation of EIS is generally not straightforward and should therefore preferably be accompanied by modeling. This modeling can be either by the use of equivalent circuits [84] or a numerical device model [63] based on the transport equations described in Sect. 1.2.5. In particular, interpretation of EIS in LECs remains complicated due to the presence of four types of charge carriers, each with different, position-dependent, densities and mobilities. Furthermore, effects from binding and recombination between these carriers as described in Sect. 1.2.5.3 will further complicate matters. Studies reported on impedance spectroscopy in LECs have shown that this technique allows determination of ion conductivity and electric double layer formation in unbiased cells, as well as determination of the recombination zone width in biased cells that show light emission [63, 84, 93]. It is anticipated that more information can be obtained from biased LECs, however, this has yet to be conclusively explored and reported.

1.4.2.3

Studies in Stacked LECs

In LECs, the recombination zone is intrinsic and is therefore a relatively low-conductive region, sandwiched by relatively high-conductive doped regions. The light-emitting p-i-n junction can thus, in first-order approximation, be considered as a parallel plate capacitor. The junction width is then

34

S. van Reenen and M. Kemerink

Fig. 1.20 Calculated potential drop across the junction (solid line) and junction width (dashed line) as a function of applied bias. (reproduced from Ref. [111])

Fig. 1.21 I-V characteristics (left), as well as capacitance (filled symbols) and junction thickness (empty symbols) measurements (right) at different times during Vbias = 3.5 V operation of a stacked Ir-based iTMC-LEC. The black lines in a indicate a quadratic dependence. (reproduced from Ref. [54])

Lj ¼ e0 er =C;

ð1:20Þ

where C is the areal capacitance in F m−2. In EIS measurements, the relatively slow ions typically dominate at low frequencies below 10 kHz, whereas electronic processes dominate above this frequency. Therefore, the capacitance for frequencies above 10 kHz is related to the junction, hence the junction width can be estimated by this technique [54, 63, 84, 110, 111]. Campbell et al. modeled impedance spectra extracted from stacked CP-based LECs and found that the junction width depends on the bias voltage as shown in Fig. 1.20 [111]. At the turn-on of the cell above *1.7 V, the junction width is observed to decrease whereas the junction potential is observed to increase. The width of the junction in this stacked CP-based LEC decreases to values as low as 15 nm, which is much lower than the micron-sized widths found in planar CP-based LECs [13, 46]. The low conductivity and the absence of doping in the junction region lead to the formation of space charge during operation. Lenes et al. [54] found that the current

1 Light-Emitting Electrochemical Cells …

35

through this region is space charge limited as shown by the quadratic dependence of the current in Fig. 1.21a. Space charged limited current can be described by JSCLC ¼ a

V2 ; L3j

ð1:21Þ

where a is a coefficient that depends on the dielectric constant, carrier mobilities, and bimolecular recombination rate [112]. The increase in current during fixed voltage operation in Fig. 1.21a can therefore be explained by a continuous reduction in junction width Lj. This is confirmed by determination of the junction width in similar cells under similar conditions by characterization of the high frequency capacitance as shown in Fig. 1.21b. This continuous reduction of the junction width is due to continuous doping of the active layer.

1.4.3

Current-Voltage Characteristic

Measurement of a general I-V characteristic in LECs is complicated because of the typical hysteresis due to on-going reversible and irreversible reactions, [100, 101] combined with the large variation in turn-on times [55]. Hence, a difference in voltage sweep rate can have dramatic impact on the resultant I-V characteristic. This prompts the need to always report time-dependent characteristics of current, luminance, and efficiency next to voltage-dependent characteristics [63, 64]. Nevertheless, experimental data in the literature still show resemblance between IV characteristics for various materials used in LECs. A few examples of experimentally determined I-V characteristics are shown in Fig. 1.22. Understanding of the I-V characteristics in LECs necessitate numerical modeling due to the large amount of processes happening simultaneously in LECs. Such numerical modeling efforts were initiated by D. L. Smith who used a steady-state model to describe carrier transport in LECs [38]. The model predicted that the

Fig. 1.22 I-V characteristics of stacked LECs based on a multifluorophoric conjugated copolymer mixed with a LiCF3SO3 + trimethylolpropane ethoxylate electrolyte (left, reproduced from Ref. [113]), a MDMO-PPV CP mixed with a KCF3SO3 + poly(ethylene oxide) electrolyte (central, reproduced from Ref. [63]), and a Tris(2,2’-bipyridine)ruthenium(II) complex (straight lines: tested in drybox; dashed lines: tested in air) (right, reproduced from Ref. [71])

36

S. van Reenen and M. Kemerink

Fig. 1.23 Calculated I-V characteristics in LECs by Smith (Left, reproduced from Ref. [38]), Manzanares et al. (central, reproduced from Ref. [39]), and Mills and Lonergan. (right, reproduced from Ref. [48])

electrochemical doping density depends exponentially on the applied bias voltage for voltages near the band gap of the semiconductor. As a result, the current grows exponentially for increasing bias voltage as shown in Fig. 1.23a. Such exponential growth is also observed in experiments as shown in Fig. 1.22b and c for bias voltages around the bandgap of the semiconductor, i.e., roughly 2 eV. The experiments displayed in Fig. 1.22b show that at relatively high bias voltages, the current levels off. Modeling by Manzanares et al. showed that the IV curve levels off at high doping densities as shown in Fig. 1.23b [39]. This is due to all ions being used for electrochemical doping, preventing further doping at even higher bias voltages. According to their model, any additional voltage is added to the interfaces. Other work including modeling and experiments [47] have shown this not to be the case: the additionally applied bias voltage drops in the junction region. The enhanced field in the junction can furthermore lead to a broadening of the recombination zone due to the requirement for current conservation [63]. Mills and Lonergan developed a numerical and analytical model that describes the complete I-V characteristic across the whole possible bias voltage range, as shown in Fig. 1.23c [48]. They discern four regimes. In the first, low injection regime, the current grows exponentially due to increased carrier injection by EDL formation. Here, the bulk is field-free and the operation of the LEC follows the EDM. Increasing the bias voltage gets the LEC in the second, space charge regime, where one type of electronic carrier is injected efficiently. In this regime the device follows the PECDM. The space charge of this majority carrier in the device then affects the device potential in such a way that further injection is retarded, leading to only a weak increase of the current density. Simultaneously, increase of the bias voltage within this regime leads to enhancement of the injection of the minority carriers. Increasing the bias even further gets the cell in the third, bipolar injection regime where carrier injection for both electron and hole injection is efficient, allowing anions and cations to separate to form electrochemically doped regions, while the current again increases exponentially with bias. At these bias voltages, the device follows the ECDM. Further enhancement of the bias voltage will ultimately lead to complete exhaustion of the available mobile ions so the device enters the fourth, high injection regime. At this point, enhancement of the bias voltage does

1 Light-Emitting Electrochemical Cells …

37

not enhance the doping density in the device anymore, resulting in a much weaker increase in current density. Any additional voltage is then dropped predominantly over the junction region and, to a minor extent, at the interfaces to enhance carrier injection to accommodate the increase in current density. The bias voltage at which each regime occurs depends strongly on the available mobile ion densities and the injection barriers for electrons and holes. Hence, large variations are present between I-V characteristics in different LECs.

1.4.4

Luminescence Quenching and Reabsorption

Mobile ions in LECs enable efficient carrier injection and transport by EDL formation and electrochemical doping at relatively low bias voltages. However, to achieve an efficient electroluminescent device, the resultant large density of current must be converted into excitons that subsequently decay radiatively, generating photons that leave the device without being reabsorbed. To achieve a maximal recombination efficiency of electrons and holes into excitons, it is required that injected electrons and holes do not leave the active layer at the opposite electrode, i.e., the anode and the cathode, respectively. This is the case if LECs operate in the ECDM and both n- and p-type electrochemical doping has developed in such a way that the recombination zone is sufficiently far away from the electrodes. High doping densities should then make it impossible for electrons or holes to move through oppositely doped regions without recombining [47]. The fraction of excitons that decays radiatively depends first of all on the type of emitter that is used and the distribution in exciton spin-states. Conjugated polymers like PPV are singlet emitters, whereas, e.g., Ir-based iTMCs are triplet emitters. Second, it depends on the rates of non-radiative decay processes that compete with radiative decay. Excitons can, e.g, be quenched by the presence of large electric fields, [114] other excitons, [115] or polarons [116]. Exciton quenching by polarons has been found to contribute significantly to loss in efficiency in LECs [64, 117]. The reason is the relatively large density of electrochemical doping next to a relatively thin recombination zone, e.g, *15 or *22 nm thick as shown in Figs. 1.20 and 1.21b. Exciton quenching by polarons occurs through diffusion of excitons through the semiconductor during their lifetime, followed by either of two possible processes: Förster resonance energy transfer (FRET) or charge transfer (CT) [117]. Both processes are depicted in Fig. 1.24. In case of FRET, the exciton transfers its energy to a polaron leading to the loss of the exciton and an excited polaron that will relax and generate heat. In case of CT, the polaron recombines with the exciton, leading to the loss of the exciton and the generation of a new polaron. The effect of these quenching mechanisms becomes stronger for higher doping densities. In previous paragraphs we saw that doping densities increase in time (Sect. 1.3.1) and for enhanced voltages (Sect. 1.4.3). Indeed, in operational LECs

38

S. van Reenen and M. Kemerink

Fig. 1.24 Schematic representation of exciton quenching mechanisms in LECs. Left exciton diffusion follow by Förster resonance energy transfer. Right exciton diffusion followed by charge transfer. (reproduced from Ref. [117])

Fig. 1.25 Absorption and EL intensity in a stacked CP-based LEC. The absorption spectra are shown for a doped and undoped device as indicated. (reproduced from Ref. [118])

the efficiency is found to typically roll-off in time (see also Fig. 1.5 in Sect. 1.3.1) and for enhanced voltages [55]. Integrating the above-mentioned quenching mechanisms into a device model allows to reproduce this roll-off, see also Sect. 1.4.6 [64]. Following radiative decay, the generated photons need to leave the device. However, Kaihovirta et al. found that a significant fraction of the generated photons can be reabsorbed by electrochemical doping in CP- and iTMC-based LECs [118, 119]. Electrochemical doping namely leads to an additional absorption band at similar wavelengths as the emitting transition, shown for example in Fig. 1.25. For a CP-based LEC based on Super Yellow PPV, Kaihovirta found that 100 nm of active layer leads to *10% reabsorption, whereas 1.0 lm of active layer leads to over 70% reabsorption. The effect depends on the overlap between the emission spectrum and the self-absorption spectrum induced by doping. Hence this effect can vary strongly for different materials. The reabsorption in a yellow-emitting Ir-based iTMC-LEC was found to be significantly less: 4% in a 95 nm thick device and 40%

1 Light-Emitting Electrochemical Cells …

39

in a 1.0 lm thick device. This shows that thin devices are likely preferred to achieve efficient LECs. It also suggests a materials optimization target.

1.4.5

Color Tuning and Cavity Effects

Color tuning in electroluminescent devices based on organics is relatively easy compared to those based on inorganics due to the large number of materials with different bandgaps available through synthesis. In polymers and small molecules, the bandgap can be tuned by using different repeat units or by adding different electron donating/accepting moieties in the periphery of the compound, respectively [113, 120]. In iTMCs, variation of the organic ligands allows coverage of the complete visible light emission spectrum, despite the limited possibilities in variation of the metal complex [121]. Finally, the color of nanoparticles and quantum dots are mainly related to their size as a consequence of the quantum confinement, facilitating color tuning in LECs based on these emitters. These aspects are further described in the section III of the book. Besides color tuning through materials design, the emission spectrum of LECs can also be modified through photonic effects since the thickness of the active layer in stacked LECs is comparable to the wavelengths of the emitted photons. Moreover, typically one reflective electrode is used to contact the device. This combination gives stacked LECs a microcavity structure that through interference effects can have a significant influence on the emission spectrum and must therefore

Fig. 1.26 Simulated and measured EL spectra of a stacked 450 nm thick iTMC-LEC at 8 (top, left), 12 (top, right), 18 (bottom, left), and 58 (bottom, right) min after a bias of 2.5 V was applied. The recombination zone position (zi) with respect to the cathode position was determined by fitting the simulated and measured EL spectra and is shown in each figure. (reproduced from Ref. [127])

40

S. van Reenen and M. Kemerink

be taken into account when optimizing LECs. These cavity effects are known to be important in OLEDs [122–124] and in organic photovoltaic cells [125]. Although Chap. 3 is devoted to color tuning by cavity effects, we will briefly comment on it here. The position where recombination takes place in electroluminescent devices has a strong effect on the emission spectrum [124] Wang and Su took advantage of cavity effects in 450 nm thick LECs to study the migration of the light-emitting junction in time as shown in Fig. 1.26 [126]. They found that the junction first formed closest to the electrode with the largest injection barrier, which is the cathode in this case (see Fig. 1.26a). This is due to more ions being required to turn the cathode into an Ohmic contact. In time, however, this Ohmic contact is formed, resulting in balanced carrier injection that leads to a junction shift towards the center of the device (see Fig. 1.26d). These results show that effects of carrier injection and transport on the junction position during operation can be studied using microcavity effects in stacked LECs. We are not aware of non-diagnostic application of these effects in color tuning of LECs.

1.4.6

Efficiency: Values and Limits

Comparison of performance between LEC devices is extremely complicated. Efficiencies as high as 39.8 lm/W, [128] luminances above 10000 cd/m2, [129] and lifetimes of 3000 h [130] (defined here as the time for the brightness to decay to half-maximum) have been reported for LECs. However, these values do not apply to a single device and are all measured under different experimental conditions. For example, the efficiency of LECs depends strongly on the emission wavelength as well as the brightness and time at which the efficiency was measured [131]. The time dependence of operation of LECs has already been discussed in previous paragraphs. As a virtually unavoidable result of the device physics, variations in efficiency over time are expected due to, e.g., continuous electrochemical doping and changes in the recombination zone width and position. It has, however, been shown by Tordera et al. that the characteristic efficiency roll-off of LECs (see Sect. 1.3) can be significantly suppressed by a pulsed instead of a DC driving scheme [94]. The tradeoff between efficiency and brightness has been discussed quantitatively in Ref. [64]. Enhancement of the brightness in an LEC is achieved by improving the current density, which in turn is achieved by enhancing the doping density. However, this enhanced doping density simultaneously boosts the rate of exciton-polaron quenching by Förster resonant energy transfer (FRET) and/or charge transfer (CT) as discussed in Sect. 1.4.4. This quenching results in a lowering of the electroluminescent efficiency [64]. To avoid this tradeoff it is required to (1) reduce FRET by reduction of the overlap between exciton emission spectra and polaron absorption spectra and (2) reduce CT by suppression of exciton diffusion by, e.g., introduction of structural disorder. Unfortunately, the latter would

1 Light-Emitting Electrochemical Cells …

41

lead to (even) lower charge carrier mobilities. For iTMC complexes the structural disorder can possibly be tuned through the addition of large side-chains [132].

1.5

Conclusion and Outlook

Work carried out by various groups around the world has led to a coherent understanding of the complicated device physics of LECs. The operation of LECs strongly depends on carrier injection in the active layer, which is predominantly determined by the injection barrier and the applied bias voltage. The preferred operational mechanism of LECs is the electrochemical doping model that occurs in case carrier injection does not limit device operation, leading to electrochemical doping of the active layer. The process of electrochemical doping in LECs governs the transient properties relating to turn-on, hysteresis and lifetime, as well as the steady-state properties relating to potential and carrier distribution, carrier recombination, and electroluminescence quenching. As the majority of the device physics is now relatively well understood, it seems logical to direct further efforts towards materials development to limit electroluminescence quenching, self-absorption, and degradation, which lead to significant losses in performance. Electroluminescence quenching can be reduced by use of dyes that limit exciton diffusion towards polarons or by enhancement of the radiative decay rate. Self-absorption can be reduced by use of relatively thin active layer films and improved control of the position of the recombination zone. The latter may also be used to tune the emission spectrum and minimize its overlap with quenching or absorbing transitions, see Sects. 1.4.4 and 1.4.5. Regarding degradation, its origins are yet to be fully understood. Nevertheless, it seems likely that ion-solvating polymers with a larger electrochemical stability window are desired. To accommodate materials research and device development, the field would greatly benefit from standardized tests on reporting performance characteristics of LECs. The extensively discussed transient behavior of LECs makes their performance strongly dependent on measurement time and driving conditions, often making it nearly impossible to compare reported performance indicators. As a first starting point, peak light outputs and efficiencies should always be accompanied by information about their transients, such as the turn-on time and the time to half-peak value. Setup of a standardized characterization protocol is, however, far from straightforward because of the multidimensional parameter space of light-emitting electrochemical cells.

References 1. A. Bernanose, M. Comte, P. Vouaux, J. Chim. Phys. 50, 64 (1953) 2. C.W. Tang, S.A. VanSlyke, Appl. Phys. Lett. 51, 913 (1987)

42

S. van Reenen and M. Kemerink

3. J.H. Burroughes, D.D.C. Bradley, A.R. Brown, R.N. Marks, K. Mackay, R.H. Friend, P.L. Burns, A.B. Holmes, Nature 347, 539 (1990) 4. D.R. Baigent, N.C. Greenham, J. Grüner, R.N. Marks, R.H. Friend, S.C. Moratti, A.B. Holmes, Synth. Met. 67, 3 (1994) 5. K. Fesser, A.R. Bishop, D.K. Campbell, Phys. Rev. B 27, 4804 (1983) 6. A. Köhler, H. Bässler, Electronic Processes in Organic Semiconductors: An Introduction (Wiley, 2015) 7. P.W.M. Blom, M.C.J.M. Vissenberg, Mater. Sci. Eng. R Rep. 27, 53 (2000) 8. H. Imahori, H. Norieda, Y. Nishimura, I. Yamazaki, K. Higuchi, N. Kato, T. Motohiro, H. Yamada, K. Tamaki, M. Arimura, Y. Sakata, J. Phys. Chem. B 104, 1253 (2000) 9. P. Andrew, Science 290, 785 (2000) 10. E. Dulkeith, A.C. Morteani, T. Niedereichholz, T.A. Klar, J. Feldmann, S.A. Levi, F.C.J.M. Van Veggel, D.N. Reinhoudt, M. Möller, D.I. Gittins, Phys. Rev. Lett. 89, (2002) 11. S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Lüssem, K. Leo, Nature 459, 234 (2009) 12. M. Mesta, M. Carvelli, R.J. de Vries, H. van Eersel, J.J.M. van der Holst, M. Schober, M. Furno, B. Lüssem, K. Leo, P. Loebl, R. Coehoorn, P.A. Bobbert, Nat. Mater. 12, 652 (2013) 13. Q. Pei, G. Yu, C. Zhang, Y. Yang, A.J. Heeger, Science 269, 1086 (1995) 14. J.-K. Lee, D.S. Yoo, E.S. Handy, M.F. Rubner, Appl. Phys. Lett. 69, 1686 (1996) 15. E.S. Handy, A.J. Pal, M.F. Rubner, J. Am. Chem. Soc. 121, 3525 (1999) 16. F.G. Gao, A.J. Bard, J. Am. Chem. Soc. 122, 7426 (2000) 17. H. Rudmann, M.F. Rubner, J. Appl. Phys. 90, 4338 (2001) 18. A.J. Norell Bader, A.A. Ilkevich, I.V. Kosilkin, J.M. Leger, Nano Lett. 11, 461 (2011) 19. S. Tang, W.-Y. Tan, X.-H. Zhu, L. Edman, Chem. Commun. 49, 4926 (2013) 20. A. Pertegás, D. Tordera, J.J. Serrano-Pérez, E. Ortí, H.J. Bolink, J. Am. Chem. Soc. 135, 18008 (2013) 21. C. Bizzarri, C. Strabler, J. Prock, B. Trettenbrein, M. Ruggenthaler, C.-H. Yang, F. Polo, A. Iordache, P. Brüggeller, L.D. Cola, Inorg. Chem. 53, 10944 (2014) 22. G. Qian, Y. Lin, G. Wantz, A.R. Davis, K.R. Carter, J.J. Watkins, Adv. Funct. Mater. 24, 4484 (2014) 23. D. Asil, J.A. Foster, A. Patra, X. de Hatten, J. del Barrio, O.A. Scherman, J.R. Nitschke, R. H. Friend, Angew. Chem. Int. Ed. 53, 8388 (2014) 24. S. Keller, E.C. Constable, C.E. Housecroft, M. Neuburger, A. Prescimone, G. Longo, A. Pertegás, M. Sessolo, H.J. Bolink, Dalton Trans. 43, 16593 (2014) 25. M.Y. Wong, G.J. Hedley, G. Xie, L.S. Kölln, I.D.W. Samuel, A. Pertegás, H.J. Bolink, E. Zysman-Colman, Chem. Mater. 27, 6535 (2015) 26. M.F. Aygüler, M.D. Weber, B.M.D. Puscher, D.D. Medina, P. Docampo, R.D. Costa, J. Phys. Chem. C 119, 12047 (2015) 27. H. Zhang, H. Lin, C. Liang, H. Liu, J. Liang, Y. Zhao, W. Zhang, M. Sun, W. Xiao, H. Li, S. Polizzi, D. Li, F. Zhang, Z. He, W.C.H. Choy, Adv. Funct. Mater. 25, 7226 (2015) 28. M.S. Subeesh, K. Shanmugasundaram, C.D. Sunesh, T.P. Nguyen, Y. Choe, J. Phys. Chem. C 119, 23676 (2015) 29. M.D. Weber, V. Nikolaou, J.E. Wittmann, A. Nikolaou, P.A. Angaridis, G. Charalambidis, C. Stangel, A. Kahnt, A.G. Coutsolelos, R.D. Costa, Chem. Commun. 52, 1602 (2016) 30. M.D. Weber, M. Adam, R.R. Tykwinski, R.D. Costa, Adv. Funct. Mater. 25, 5066 (2015) 31. S. Jenatsch, L. Wang, M. Bulloni, A.C. Véron, B. Ruhstaller, S. Altazin, F. Nüesch, R. Hany, A.C.S. Appl, Mater. Interfaces 8, 6554 (2016) 32. Z. Zhang, K. Guo, Y. Li, X. Li, G. Guan, H. Li, Y. Luo, F. Zhao, Q. Zhang, B. Wei, Q. Pei, H. Peng, Nat. Photonics (2015) 33. A. Asadpoordarvish, A. Sandström, C. Larsen, R. Bollström, M. Toivakka, R. Österbacka, L. Edman, Adv. Funct. Mater. 25, 3238 (2015) 34. A. Sandström, A. Asadpoordarvish, J. Enevold, L. Edman, Adv. Mater. 26, 4975 (2014) 35. A. Sandström, L. Edman, Energy Technol. 3, 329 (2015) 36. Q. Pei, Yang, G. Yu, C. Zhang, A.J. Heeger, J. Am. Chem. Soc. 118, 3922 (1996)

1 Light-Emitting Electrochemical Cells … 37. 38. 39. 40. 41. 42. 43. 44. 45. 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.

43

D.L. Smith, J. Appl. Phys. 81, 2869 (1997) J.A. Manzanares, H. Reiss, A.J. Heeger, J. Phys. Chem. B 102, 4327 (1998) J.C. deMello, Phys. Rev. B 66, (2002) J.C. deMello, N. Tessler, S.C. Graham, R.H. Friend, Phys. Rev. B 57, 12951 (1998) J.C. deMello, J.J.M. Halls, S.C. Graham, N. Tessler, R.H. Friend, Phys. Rev. Lett. 85, 421 (2000) J.M. Leger, S.A. Carter, B. Ruhstaller, J. Appl. Phys. 98, 124907 (2005) L.S.C. Pingree, D.B. Rodovsky, D.C. Coffey, G.P. Bartholomew, D.S. Ginger, J. Am. Chem. Soc. 129, 15903 (2007) D.B. Rodovsky, O.G. Reid, L.S.C. Pingree, D.S. Ginger, ACS Nano 4, 2673 (2010) S. van Reenen, P. Matyba, A. Dzwilewski, R.A.J. Janssen, L. Edman, M. Kemerink, J. Am. Chem. Soc. 132, 13776 (2010) T.J. Mills, M.C. Lonergan, Phys. Rev. B 85, (2012) Q. Pei, A.J. Heeger, Nat. Mater. 7, 167 (2008) J. Gao, J. Dane, Appl. Phys. Lett. 84, 2778 (2004) J.D. Slinker, J.A. DeFranco, M.J. Jaquith, W.R. Silveira, Y.-W. Zhong, J.M. Moran-Mirabal, H.G. Craighead, H.D. Abruña, J.A. Marohn, G.G. Malliaras, Nat. Mater. 6, 894 (2007) P. Matyba, K. Maturova, M. Kemerink, N.D. Robinson, L. Edman, Nat. Mater. 8, 672 (2009) S.B. Meier, D. Hartmann, D. Tordera, H.J. Bolink, A. Winnacker, W. Sarfert, Phys. Chem. Chem. Phys. 14, 10886 (2012) S.B. Meier, S. van Reenen, B. Lefevre, D. Hartmann, H.J. Bolink, A. Winnacker, W. Sarfert, M. Kemerink, Adv. Funct. Mater. 23, 3531 (2013) M. Lenes, G. Garcia-Belmonte, D. Tordera, A. Pertegás, J. Bisquert, H.J. Bolink, Adv. Funct. Mater. 21, 1581 (2011) S. van Reenen, T. Akatsuka, D. Tordera, M. Kemerink, H.J. Bolink, J. Am. Chem. Soc. 135, 886 (2013) P. Langevin, in Recomb. Mobilites Ions Dans Gaz Ann. Chim Phys (1903), p. 433 W. Shockley, W.T. Read, Phys. Rev. 87, 835 (1952) R.N. Hall, Phys. Rev. 87, 387 (1952) M. Kuik, L.J.A. Koster, G.A.H. Wetzelaer, P.W.M. Blom, Phys. Rev. Lett. 107, (2011) P.R. Emtage, J.J. O’Dwyer, Phys. Rev. Lett. 16, 356 (1966) J. C. deMello, J. Comput. Phys. 181, 564 (2002) D. Tordera, S. Meier, M. Lenes, R.D. Costa, E. Ortí, W. Sarfert, H.J. Bolink, Adv. Mater. 24, 897 (2012) S. van Reenen, R.A.J. Janssen, M. Kemerink, Adv. Funct. Mater. 22, 4547 (2012) S. van Reenen, R.A.J. Janssen, M. Kemerink, Adv. Funct. Mater. 25, 3066 (2015) J.-H. Shin, N.D. Robinson, S. Xiao, L. Edman, Adv. Funct. Mater. 17, 1807 (2007) J. Fang, Y. Yang, L. Edman, Appl. Phys. Lett. 93, 63503 (2008) Y. Hu, J. Gao, J. Am. Chem. Soc. 133, 2227 (2011) Y. Zhang, J. Gao, J. Appl. Phys. 100, 84501 (2006) C.H. Lyons, E.D. Abbas, J.-K. Lee, M.F. Rubner, J. Am. Chem. Soc. 120, 12100 (1998) H.J. Bolink, L. Cappelli, E. Coronado, M. Grätzel, E. Ortí, R.D. Costa, P.M. Viruela, M.K. Nazeeruddin, J. Am. Chem. Soc. 128, 14786 (2006) G. Kalyuzhny, M. Buda, J. McNeill, P. Barbara, A.J. Bard, J. Am. Chem. Soc. 125, 6272 (2003) J. Slinker, D. Bernards, P.L. Houston, H.D. Abruña, S. Bernhard, G.G. Malliaras, Chem Commun 2392 (2003) A.L. Holt, J.M. Leger, S.A. Carter, J. Chem. Phys. 123, 44704 (2005) J. Gao, J. Dane, J. Appl. Phys. 98, 63513 (2005) J.-H. Shin, N.D. Robinson, S. Xiao, L. Edman, Adv. Funct. Mater. 17, 1807 (2007) T. Wågberg, P.R. Hania, N.D. Robinson, J.-H. Shin, P. Matyba, L. Edman, Adv. Mater. 20, 1744 (2008)

44

S. van Reenen and M. Kemerink

76. S. van Reenen, P. Matyba, A. Dzwilewski, R.A.J. Janssen, L. Edman, M. Kemerink, Adv. Funct. Mater. 21, 1795 (2011) 77. S. van Reenen, R.A.J. Janssen, M. Kemerink, Org. Electron. 12, 1746 (2011) 78. H. Shimotani, G. Diguet, Y. Iwasa, Appl. Phys. Lett. 86, 22104 (2005) 79. V.I. Arkhipov, E.V. Emelianova, P. Heremans, H. Bässler, Phys. Rev. B 72, 235202 (2005) 80. N.D. Robinson, J.-H. Shin, M. Berggren, L. Edman, Phys. Rev. B 74, (2006) 81. V. Bychkov, P. Matyba, V. Akkerman, M. Modestov, D. Valiev, G. Brodin, C.K. Law, M. Marklund, L. Edman, Phys. Rev. Lett. 107, (2011) 82. V. Bychkov, O. Jukimenko, M. Modestov, M. Marklund, Phys. Rev. B 85, (2012) 83. A. Munar, A. Sandström, S. Tang, L. Edman, Adv. Funct. Mater. 22, 1511 (2012) 84. S. Tang, L. Edman, J. Phys. Chem. Lett. 1, 2727 (2010) 85. S. Tang, J. Mindemark, C.M.G. Araujo, D. Brandell, L. Edman, Chem. Mater. 26, 5083 (2014) 86. J. Mindemark, S. Tang, J. Wang, N. Kaihovirta, D. Brandell, L. Edman, Chem. Mater. (2016) 87. J. Mindemark, L. Edman, J. Mater. Chem. C 4, 420 (2016) 88. Y. Shen, D.D. Kuddes, C.A. Naquin, T.W. Hesterberg, C. Kusmierz, B.J. Holliday, J.D. Slinker, Appl. Phys. Lett. 102, 203305 (2013) 89. K.J. Suhr, L.D. Bastatas, Y. Shen, L.A. Mitchell, B.J. Holliday, J.D. Slinker, ACS Appl. Mater. Int. 8, 8888 (2016) 90. R.D. Costa, A. Pertegás, E. Ortí, H.J. Bolink, Chem. Mater. 22, 1288 (2010) 91. J. Fang, P. Matyba, L. Edman, Adv. Funct. Mater. 19, 2671 (2009) 92. G. Yu, Y. Cao, C. Zhang, Y. Li, J. Gao, A.J. Heeger, Appl. Phys. Lett. 73, 111 (1998) 93. D. Tordera, S. Meier, M. Lenes, R.D. Costa, E. Ortí, W. Sarfert, H.J. Bolink, Adv. Mater. 24, 897 (2012) 94. C.V. Hoven, H. Wang, M. Elbing, L. Garner, D. Winkelhaus, G.C. Bazan, Nat. Mater. (2010) 95. I.V. Kosilkin, M.S. Martens, M.P. Murphy, J.M. Leger, Chem. Mater. 22, 4838 (2010) 96. J. Gao, G. Yu, A.J. Heeger, Appl. Phys. Lett. 71, 1293 (1997) 97. L. Edman, J. Appl. Phys. 95, 4357 (2004) 98. Y. Shao, G.C. Bazan, A.J. Heeger, Adv. Mater. 19, 365 (2007) 99. X. Li, J. Gao, G. Liu, Appl. Phys. Lett. 102, 223303 (2013) 100. X. Li, F. AlTal, G. Liu, J. Gao, Appl. Phys. Lett. 103, 243304 (2013) 101. F. AlTal, J. Gao, Org. Electron. 18, 1 (2015) 102. J. Fang, P. Matyba, N.D. Robinson, L. Edman, J. Am. Chem. Soc. 130, 4562 (2008) 103. J.-H. Shin, S. Xiao, L. Edman, Adv. Funct. Mater. 16, 949 (2006) 104. P. Matyba, M.R. Andersson, L. Edman, Org. Electron. 9, 699 (2008) 105. S.B. Toshner, Z. Zhu, I.V. Kosilkin, J.M. Leger, A.C.S. Appl, Mater. Interfaces 4, 1149 (2012) 106. T.D. Shoji, Z. Zhu, J.M. Leger, A.C.S. Appl, Mater. Interfaces 5, 11509 (2013) 107. S. Krivec, T. Detzel, M. Buchmayr, H. Hutter, Appl. Surf. Sci. 257, 25 (2010) 108. L. Bozano, S.A. Carter, J.C. Scott, G.G. Malliaras, P.J. Brock, Appl. Phys. Lett. 74, 1132 (1999) 109. Y. Li, J. Gao, G. Yu, Y. Cao, A.J. Heeger, Chem. Phys. Lett. 287, 83 (1998) 110. I.H. Campbell, D.L. Smith, C.J. Neef, J.P. Ferraris, Appl. Phys. Lett. 72, 2565 (1998) 111. M.A. Lampert, P. Mark, Current Injection in Solids (Academic Press, 1970) 112. J. Kalinowski, W. Stampor, J. Szmytkowski, D. Virgili, M. Cocchi, V. Fattori, C. Sabatini, Phys. Rev. B 74, (2006) 113. S.M. King, D. Dai, C. Rothe, A.P. Monkman, Phys. Rev. B 76, (2007) 114. J.M. Hodgkiss, S. Albert-Seifried, A. Rao, A.J. Barker, A.R. Campbell, R.A. Marsh, R.H. Friend, Adv. Funct. Mater. 22, 1567 (2012) 115. S. van Reenen, M.V. Vitorino, S.C.J. Meskers, R.A.J. Janssen, M. Kemerink, Phys. Rev. B 89, (2014) 116. N. Kaihovirta, A. Asadpoordarvish, A. Sandström, L. Edman, ACS Photonics 1, 182 (2014)

1 Light-Emitting Electrochemical Cells …

45

117. N. Kaihovirta, G. Longo, L. Gil-Escrig, H.J. Bolink, L. Edman, Appl. Phys. Lett. 106, 103502 (2015) 118. S. Tang, J. Pan, H.A. Buchholz, L. Edman, J. Am. Chem. Soc. 135, 3647 (2013) 119. S. Tang, J. Pan, H. Buchholz, L. Edman, A.C.S. Appl, Mater. Interfaces 3, 3384 (2011) 120. M.S. Lowry, S. Bernhard, Chem. - Eur. J. 12, 7970 (2006) 121. M.-H. Lu, J.C. Sturm, J. Appl. Phys. 91, 595 (2002) 122. J.M. Leger, S.A. Carter, B. Ruhstaller, H.-G. Nothofer, U. Scherf, H. Tillman, H.-H. Hörhold, Phys. Rev. B 68, (2003) 123. S.L.M. van Mensfoort, M. Carvelli, M. Megens, D. Wehenkel, M. Bartyzel, H. Greiner, R.A.J. Janssen, R. Coehoorn, Nat. Photonics 4, 329 (2010) 124. J. Gilot, I. Barbu, M.M. Wienk, R.A.J. Janssen, Appl. Phys. Lett. 91, 113520 (2007) 125. T.-W. Wang, H.-C. Su, Org. Electron. 14, 2269 (2013) 126. H.J. Bolink, E. Coronado, R.D. Costa, N. Lardiés, E. Ortí, Inorg. Chem. 47, 9149 (2008) 127. Y. Xiong, L. Li, J. Liang, H. Gao, S. Chou, Q. Pei, Mater. Horiz. 2, 338 (2015) 128. H.J. Bolink, E. Coronado, R.D. Costa, E. Ortí, M. Sessolo, S. Graber, K. Doyle, M. Neuburger, C.E. Housecroft, E.C. Constable, Adv. Mater. 20, 3910 (2008) 129. T. Hu, L. He, L. Duan, Y. Qiu, J. Mater. Chem. 22, 4206 (2012) 130. R.D. Costa, E. Ortí, H.J. Bolink, S. Graber, C.E. Housecroft, E.C. Constable, Adv. Funct. Mater. 20, 1511 (2010) 131. J.C. deMello, Nat. Mater. 6, 796 (2007) 132. H.-C. Su, J.-H. Hsu, Dalton Trans. 44, 8330 (2015)

Part II

Definition and Role of the Ionic Additives

Chapter 2

Optical-Beam-Induced-Current Imaging of Planar Polymer Light-Emitting Electrochemical Cells Faleh AlTal and Jun Gao

Abstract In this chapter, we describe optical-beam-induced-current (OBIC) and scanning photoluminescence (PL) imaging of extremely large planar LECs that have been frozen to preserve the doping profile. This complements the basics described in Chap. 1 with respect to device mechanism and characterization. We succeeded in resolving the depletion width, for the first time, of a frozen LEC pn junction and a frozen LEC p-i-n junction. These optical scanning results reveal a surprisingly strong built-in potential that is independent of the electrode work function and an extremely narrow junction depletion region that is about 0.2% of the interelectrode spacing. These findings provide new insight into the electronic structure of the LEC junction. Since only about 0.2% of the entire device area is photoactive in response to an incident optical beam, the effective junction width (or volume) of polymer-based LECs must be dramatically increased to realize a more efficient device. Keywords P-i-n junction Optical-beam-induced current

2.1 2.1.1


P-n junction
Electrochemical
Light-emitting electrochemical cell

doping




Polymer Light-Emitting Electrochemical Cells Background

A light-emitting electrochemical cell (LEC) is a solid-state, two-terminal device that employs a mixed ionic/electronic conductor as the active layer—see Chap. 1 for more details [1–7]. The first LECs were demonstrated by Pei et al. and consisted of a luminescent conjugated polymer as the emitter, and a polymer electrolyte as the ion conductor sandwiched between two electrodes [8]. A direct current (DC) bias, F. AlTal
J. Gao (&) Department of Physics, Engineering Physics and Astronomy, Queen’s University, 64 Bader Lane, Kingston, ON K7L 3N6, Canada e-mail: [email protected] © Springer International Publishing AG 2017 R.D. Costa (ed.), Light-Emitting Electrochemical Cells, DOI 10.1007/978-3-319-58613-7_2

49

50

F. AlTal and J. Gao

comparable to the luminescent polymer energy gap, is required to activate a polymer LEC. Charge injection at the positive electrode (anode) and the negative electrode (cathode) leads to the oxidation and reduction of the luminescent polymer. Meanwhile, the mobile ions from the polymer electrolyte redistribute and compensate the injected electronic charges, causing in situ electrochemical doping of the luminescent polymer. The doped polymer is electrically neutral but has elevated electronic conductivity due to the extra, injected electronic charges. With remarkable insight, it was hypothesized by the inventors of LECs that the in situ doping was a dynamic process: doping initially occurred at the electrode-polymer interfaces, but the doped regions would expand until they meet to form a p-n junction. The formation of the p-n junction opens up a continuous pathway for the electronic charges, namely electrons and holes, which recombine in the junction region to give off light emission [9]. The LEC operation mechanism, as described above, is depicted schematically in Fig. 2.1. Light emission in an LEC is the result of radiative recombination of the injected electrons and holes in the vicinity of the junction. In this regard, the polymer LEC is analogous to a conventional pn junction light-emitting diode (LED) in that both contain a semiconductor homojunction. A polymer LEC is therefore fundamentally different from a polymer LED made with the same luminescent polymer [10]. The polymer-based LEC was developed to address two main drawbacks of conventional polymer LEDs. The active layer of a prototypical polymer LED is a pristine light-emitting polymer film. For visible light-emitting applications, the energy gap of the polymer needs to be between 1.6 and 3.1 eV. Due to the large energy gap, undoped polymer films have high resistivity and need to be very thin in order to inject a sufficient amount of current. The ultrathin (ca. 100 nm) polymer film, however, is prone to pinholes and the effect of exciton quenching [11]. Moreover, the injection of electrons and holes needs to be balanced for maximum electroluminescence (EL) efficiency. To facilitate electrons injection, a low work function reactive metal cathode is commonly used. The highly reactive cathode material (such as calcium) increases the chance of device failure unless the polymer LED is carefully encapsulated. The polymer LEC, by contrast, operates on in situ electrochemical doping of the luminescent polymer due to the presence of mobile ions in the composite material. The doped polymer is much more conductive than a pristine one so that a thicker active layer could be used. The high conductivity of the doped LEC film also ensures strong and balanced charge injection at the electrode interfaces. Regardless of the electrode work function, efficient charge

Fig. 2.1 Illustration of doping propagation and junction formation process in polymer-based LECs. h stands for holes, e stands for electrons, A stands for anions and C stands for cations

2 Optical-Beam-Induced-Current Imaging of Planar Polymer …

51

injection can occur via quantum mechanical tunnelling between a metal electrode and a heavily doped semiconductor. Heavy doping is in fact how an ohmic contact is made in conventional inorganic semiconductor devices. Indeed, highly efficient sandwich LECs had been demonstrated very early on with air-stable cathode materials and without express optimization of the active layer thickness [12–14]. The unique advantages of LECs have led to a renaissance of interest in these devices in recent years [2, 15–23]. LECs can be made with not only luminescent polymers and polymer electrolytes, but also ionic metal complexes or low-cost small molecules [24–32]. The LEC operation mechanism is inherently complex due to the presence of both ionic and electronic charges in a mixed ionic/electronic conductor. The fundamental operating mechanism of polymer-based LECs, for example, had been a subject of intense debate [33, 34]. The aim of many studies, both theoretical and experimental, had been to elucidate the basic processes of LECs and the properties of the LEC junctions [35–40]. For polymer-based LECs, it is established that in situ electrochemical doping and junction formation are the fundamental processes that dictate the dynamic activation/turn-on behaviour of LECs, as well as the static junction properties. However, we still lack knowledge about the basic properties of an LEC pn or p-i-n junction. In this chapter, we describe our recent experimental work on the scanning optical imaging of frozen planar LECs to probe the electronic structures of an LEC junction. This is meant to complement the basics described in Chap. 1 with respect to device mechanism and characterization. In brief, we performed four consecutive optical-beam-induced-current (OBIC) and scanning photoluminescence (PL) imaging of LECs with a planar (versus sandwich) configuration. We devised four different scanning setups, each with an increasing scanning resolution than the previous one [41–44]. We succeeded in resolving the depletion width, for the first time, of a frozen polymer p-n junction and a polymer p-i-n junction. These optical scanning results reveal a surprisingly strong built-in potential that is independent of the electrode work functions and an extremely narrow junction depletion region that is less than 0.2% of the interelectrode spacing. These findings have profound implications on the development of more practical and efficient LECs, as well as the fundamental science of mixed conductors—e.g. see Chaps. 4, 5, 6, and 10. The scanning measurements were performed on extremely large, frozen-junction planar LECs. Gao and his colleagues were the inventors of both extremely large planar LECs and frozen-junction LECs, two key LEC concepts that will be briefly introduced below before the optical scanning experiments are presented in Sects. 2.2, 2.3, and 2.4.

2.1.2

Frozen-Junction LECs

Despite possessing some very attractive device characteristics, polymer LECs are not without drawbacks. The very operation mechanism responsible for LEC’s insensitivity to the active layer thickness and the electrode work function also

52

F. AlTal and J. Gao

brings serious comprises. The turn-on or activation of an LEC typically takes seconds or even minutes as the junction is slowly established by the slow moving ions [12, 14, 45]. Unlike the doping of silicon by chemical diffusion, the in situ electrochemical doping of an LEC is a room temperature process. The mobile ions, which serve as counter ions to compensate the injected electrons and holes, do not become part of the polymer chain and remain mobile. Once the applied voltage bias is removed, the LEC junction will eventually disappear as the doped regions relax back to the undoped state. Therefore, LECs are slow to turn-on and exhibit strong hysteresis if they are turned on before reaching a fully relaxed state from a previous operation [46]. LECs also suffer from burn-out if a voltage bias much higher than 4 V (for sandwich cells) is applied due to the limited electrochemical stability window of the electrolyte materials used. This means that LECs are not suitable for high intensity applications. It is apparent that a fixed LEC junction is desirable, since it will retain all the advantages of LECs, while also be fast and stable. In an LEC, fixing the junction means fixing the counter ion placement. Realizing that the ion transport/mobility in a polymer electrolyte is strongly temperature dependent, Gao et al. devised a simple method to fix the LEC junction by cooling the cell after the junction formation [46, 47]. When the LEC temperature is below the glass transition temperature (Tg) of the polymer electrolyte, the ions, and therefore the LEC junction, are immobilized. The first demonstration of a frozen LEC junction was on a sandwich cell. The cell had an active layer of poly[2-methoxy-5-(2′-ethylhexyloxy)-p-phenylene vinylene] (MEH-PPV):poly[ethylene oxide] (PEO):lithium triflate (LiTf) blend film sandwiched between an Indium-Tin-oxide (ITO) electrode and an aluminium electrode. The cell was activated at room temperature with a fixed positive (ITO biased positive) or negative voltage bias to emit strongly. Subsequently, the cell was cooled to 100 K. Two key factors contributed to the success of the frozen junction. First, the voltage bias was maintained until the cell temperature reached the target temperature. Second, the target temperature of 100 K was well below the Tg of PEO (about 208 K). This ensured that the LEC junction was fully stabilized. The resulting “frozen-junction” LEC exhibit much faster response time (ls) than the same cell operated at room temperature. In a frozen-junction LEC, the ions are immobilized and the device response time is no longer limited by the slow doping process. The frozen-junction LECs also exhibit diode-like current versus voltage versus light intensity (I-V-L) characteristics, as shown in Fig. 2.2. Significant current and EL had only been observed under forward bias, here defined as a bias with the same polarity as the applied activation bias. Varying the polarity of the activation bias can, therefore, change the polarity of the frozen junction. This behaviour is completely different from both regular LECs, which can conduct and emit under both forward and reverse bias at the same time, or the polymer LEDs, whose polarity is fixed. The frozen junction also brings a new functionality to the LEC. For the first time, an activated LEC can operate as a photovoltaic (PV) cell for power generation. Figure 2.3 shows the I-V traces of a frozen-junction cell in dark and under

2 Optical-Beam-Induced-Current Imaging of Planar Polymer …

53

Fig. 2.2 Current and light vs voltage (I-V-L) data measured at 100 K (upper curves) after cooling under +4 V bias. The LEC was subsequently heated to 300 K without external bias, then biased at −3 V and cooled (after reaching steady state) to 100 K. The I-V-L characteristics (lower curves) were reversed after prebiasing at −3 V, with roughly mirror symmetry relative to 0 V. Reprinted with permission from reference [47]. Copyright (1997) American Institute of Physics

Fig. 2.3 The photovoltaic response of a frozen sandwich LEC at 100 K. the upper panel shows the response when the cell was activated using +4 V. The lower panel shows the response after the cell was reheated and activated using −3 V then was frozen again. Reprinted with permission from reference [47]. Copyright (1997) American Institute of Physics

illumination. The low rectification ratio of the dark I-V curves can be attributed to the high resistance of the film at low temperature. The cell, however, showed a pronounced photovoltaic response in either polarity, as shown by the dashed curves. It is remarkable that the same cell can exhibit either a positive or a negative open-circuit voltage (VOC) (or short-circuit current, ISC) depending on the polarity of the activation bias. The photovoltaic response of a frozen-junction LEC, just like its EL, is no longer dependent on the electrode work functions. Rather, the LEC pn or p-i-n junction determines both the electrical and optical properties of the frozen cell. The large VOC of −1 V or +1.3 V suggests a large built-in potential in the LEC junction. Adding an electron-accepting polymer to the LEC blend created a more efficient frozen-junction polymer photovoltaic cell [48].

54

F. AlTal and J. Gao

While it is fairly straightforward to cool an activated LEC to stabilize its junction in a laboratory, eventual application of frozen-junction LECs requires the junctions, once formed, to be frozen at room temperature. Progress toward this goal has been made by using electrolytes with a high Tg [49, 50]. The junction is formed at elevated temperatures and the cell is subsequently cooled to room temperature. The LEC junctions can also be fixed chemically [51–56]. One method used ion pair monomers in the LEC blend [51]. Upon activation, these ions dope the polymer and cause radical-induced polymerization that would significantly reduce the ions mobility and fix the junction. Pei et al. used PEO oligomer capped with methacrylate as the ions conducting component. It was found that polymerizing the methacrylate group during junction formation results in a stable junction with lifetime and efficiency comparable to polymer LEDs [57]. Another approach utilized ionic trapping polymers to establish a permanent junction after junction formation [54]. Also, fixed junctions were formed by incorporating cross-linkable materials that were cured after the junction formation [58, 59]. Finally, our group recently showed that when the ion solving/transporting material such as PEO was removed altogether, the resulting cell, now only contains a luminescent polymer and a lithium salt, could still be activated by applying a much higher bias voltage. The activated cells exhibit characteristics of LECs with strong evidence of doping. More important, the activated state was stable for more than 100 hours without an applied bias. This is the longest reported shelf-life of a frozen junction at room temperature [60, 61].

2.1.3

Extremely Large Planar LECs

Unlike organic or polymer LEDs, LECs can operate in both sandwich and planar configurations. In a planar configuration device, the overall device resistance is dominated by the bulk resistance of the active layer, which can be enormous for an undoped semiconducting material if the interelectrode spacing is large. For example, a planar polymer LED with an interelectrode spacing of 30 lm exhibited an EL turn-on voltage of 500 V and was only operational at liquid nitrogen temperatures [62]. A planar LEC of similar dimensions, however, can be turned on to emitted light with a mere 4 V bias [8]. In an LEC, the presence of mobile ions and the subsequent in situ electrochemical doping render the active film highly conductive. The images of these planar LECs offered the first visualization of an LEC junction [8, 9, 63]. The relatively small size (interelectrode gap size) of these early planar LECs means that they were difficult to fabricate and to study. There had been almost no follow-up studies of planar LECs in the late 90s and early 2000s. In 2003, Gao and Dane demonstrated planar LECs with an interelectrode gap size of 1.5 mm [64]. An 800 V bias was applied to turn on two 1.5 mm cells in series at room temperature, as shown in Fig. 2.4. The millimetre-sized, extremely large planar LECs are easy to fabricate via shadow masking compared to photolithographic patterning. More important, the slow turn-on process of these planar LECs

2 Optical-Beam-Induced-Current Imaging of Planar Polymer …

55

Fig. 2.4 The photograph of two working 1.5 mm polymer-based LECs in series under 800 V bias. Also shown is the device configuration and biasing condition. The green-emitting device is made with a green emitter; and the orange-emitting device is made with MEH-PPV. The shutter speed is 5 s and the aperture is f/10. Reprinted with permission from reference [64]. Copyright (2003) American Institute of Physics

is highly advantageous for time-resolved studies of the dynamic doping process and the effects of various operational and material parameters. In situ electrochemical doping of the LEC film affects not only its electrical conductivity but also its optical properties. Doping introduces mid-gap impurity states that quench the PL of the luminescent polymer [65, 66]. The optical effect of doping had been exploited by Gao and Dane to elucidate the very doping process of LECs [67]. Extremely large planar LECs were imaged under UV light illumination, and for the first time, the dynamic LEC doping process had been visualized. Figure 2.5 displays the time-lapse PL images of a 1.5 mm planar LEC under a voltage bias of 140 V. The LEC film exhibits the characteristic orange-red PL of MEH-PPV. Also visible are finger-like, darkened regions expanding from the anode toward the cathode. On the cathode side, faint but discernible darkening of the polymer film could be observed. These darkened regions are, in fact, p- and ndoped regions whose PL had been partially quenched. The darker p-doped region expanded at a faster speed than the n-doped region. The expansion stopped once the propagating doping fronts had met to form a p-n junction. Moreover, strong EL could be observed in the last image from the forward-biased p-n junction. These visualizations provide indisputable proof that doping did occur in a polymer LEC, and the formation of a p-n junction was necessary for EL to occur. From the time-lapse images of planar LECs, the average doping propagation speed was extracted and shown to be highly sensitive to the operating temperature [68]. By moderately increasing the operating temperature, Gao et al. successfully demonstrated the largest planar LEC ever with a gap size of over 10 mm [69]. Edman et al., on the other hand, showed that planar LECs with a gap size of 1 mm could be turned on with only a 5 V bias when heated to 360 K [70]. Figure 2.6 shows an example of the largest planar LEC under UV illumination during the activation process [71]. In this cell, both p- and n-doping are clearly visible. Also,

56

F. AlTal and J. Gao

Fig. 2.5 Photographs of a working 1.5 mm MEH-PPV polymer-based LEC under 365 nm UV illumination. The device was tested at 310 K under a voltage bias of 140 V. The electrode to the left is positively biased (anode, denoted as “+”) relative to the electrode to the right (cathode, denoted as “−”. The photographs were taken at different times after the application of the voltage bias. a 8 min; b 13 min; c 18 min; d 43 min. The exposure time is 20 s. The aperture is f/10. Reprinted with permission from reference [67]. Copyright (2004) American Institute of Physics

once again EL was only observed when the p- and n-doping fronts had made contact to form a p-n junction. The cell current had increased by several orders of magnitude during the activation process. Subsequently, the cell was cooled to freeze the junction. The large surface area of this cell allowed for contact probing the cell surface in a micromanipulated cryogenic probe station. The time-lapse fluorescence imaging of extremely large planar LECs has proven to be a powerful and versatile technique in the elucidation of LEC processes. The effect of thermal annealing, [72] electrode work function, [73, 74] electrolyte salt, [75–77] and operating voltage [78] had all been studied. The static doping profile of

2 Optical-Beam-Induced-Current Imaging of Planar Polymer …

57

Fig. 2.6 Time-lapse fluorescence imaging of a 10.4 mm MEH-PPV: PEO: CsClO4 planar LEC during turn-on and cooling. A fixed DC bias of 400 V was applied to turn on the cell, which was at 335 K and under UV illumination. Time since the DC bias was applied to the cell: a no bias b 2 min c 5 min d 8 min e 19 min f 37 min g 54 min. Panel H shows the cell current and temperature as a function of time after the DC bias was applied. Uniform enhancement (Level adjustment in Photoshop) has been applied to images A-G. Reprinted with permission from reference [71]. Copyright (2011) American Chemical Society

a frozen-junction LEC had been directly observed in a frozen planar LEC [68]. Heating the frozen cell briefly, however, led to partial relaxation of doping and the formation of a p-i-n junction [79]. Under the right conditions, the frozen p-in junction is a much more efficient emitter than an as-formed p-n junction due to the former’s less quenched emission zone. A frozen p-i-n junction also exhibits a record open-circuit voltage [80].

2.2 2.2.1

Scanning Optical Imaging of Planar LECs The Optical-Beam-Induced Current (OBIC) Technique

Passive, time-lapse fluorescence imaging of the entire planar LECs has led to many discoveries described above. The fully exposed surface of a planar LEC also offers

58

F. AlTal and J. Gao

a unique platform to perform spatially resolved, electrical probing of local electrical properties of the cell. Recently, scanning probe microscopic techniques have been applied to planar LECs to determine the electric potential distribution across biased planar LECs operated at room temperature [81–84]. With the extremely large 10.4 mm planar cell shown in Fig. 2.6, Gao and Hu used a micromanipulated cryogenic probe station to map both the electrical potential and the conductivity profiles of the frozen cell [71]. These studies establish that the planar LEC is doped, and a p-n junction is formed. The p-doped polymer (MEH-PPV) is much more conductive than the n-doped polymer, and the level of doping is not constant in each of the doped regions. While the above scanning probe studies aim to map the spatial distribution of an externally applied potential, the LEC junction also possesses a built-in potential/field just like a conventional p-n junction. The presence of the junction built-in potential/field is evidenced by the strong PV response of a sandwich frozen-junction LEC described in Sect. 2.1.2. In planar frozen-junction cells, VOC approaching the magnitude of the band gap energy has been observed despite the use of identical electrodes [80]. For any semiconductor homojunction, it is important to know the junction depletion width in order to design a more efficient device structure. The depletion width plays a major role in determining the response time, carrier recombination and photogeneration of the device. The optical-beam-induced current (OBIC) technique is especially well suited to probe the electronic structure of a semiconductor junction [85–89]. In OBIC measurement, a focused light beam is scanned across the device; a photocurrent is generated when the focused beam illuminates the depletion region. In neutral p- or n-doped regions, by contrast, a null OBIC signal is expected due to the absence of a built-in field that can sweep the photogenerated charge carriers before they recombine. Figure 2.7 illustrates a p-n junction under illumination when connected to a load resistor. Absorption of photons with energy larger than the band gap energy generates electron and hole pairs. The electrons and holes generated in the depletion region are subsequently swept to opposite directions shown. This creates a photocurrent and a voltage drop across the junction. The flow of a net current is reflected by the gradient in the Fermi level of the junction. It should be mentioned that the OBIC technique is typically carried out under the short-circuit condition where the detected OBIC is a short-circuit current. In addition, photogeneration just outside of the depletion region (within one diffusion length) can also give rise to an OBIC signal when the photogenerated charge carriers enter the depletion region by diffusion. In the absence of an externally applied electric field, the drift-diffusion equations that govern the current generation in a semiconductor are given by

2 Optical-Beam-Induced-Current Imaging of Planar Polymer … Fig. 2.7 A schematic of a pn junction under light illumination and the associated energy band diagram

59

e

EC EF

h

EV

-

+

-

+

R

kB T l rn  ln nru q n kB T l rp  lP pru Jp =q ¼  q p kB T ðln rn  lp rpÞ  ruðln n þ lp pÞ; ) JT =q ¼ q

Jn =q ¼

ð2:1Þ

where J is the current density, q is the elementary charge, kB is the Boltzmann constant, T is the temperature, n is the free electrons concentration, p is the holes concentration, l is the mobility coefficient and u is the electrostatic potential. The subscripts n, p and T stand for electrons, holes and total respectively. The first part of the last line in the equation is the diffusion part of the current density, while the second part is the drift part. The Einstein relation was assumed to be valid. The electron and hole components of the diffusion current counter each other and vanish when the diffusion coefficients of electrons and holes are equal, assuming balanced electron–hole generation/recombination. Also, if electrons and holes have different mobilities, the higher mobility component will be more depleted from the generation zone. Hence it is expected that the electron and hole diffusion currents tend to cancel each other and minimize the net diffusion current. On the other hand,

60

F. AlTal and J. Gao

the electron/hole drift currents add up to maximize the net drift current and proportional to the electrostatic potential gradient. Moreover, according to Onsager– Braun model, free carriers generation rate in organic semiconductors is strongly enhanced by an electric field [90, 91]. Finally, if a symmetric beam is used for excitation, the electron and hole concentration gradients will be equal around the centre of the beam and nulls the net diffusion current. Therefore, it is expected that the OBIC scan would generate a significant signal only in regions with an electrostatic potential gradient, i.e. a built-in electric field. In a p-n junction, the electrostatic potential varies along the depletion region, and it causes the OBIC signal to be significant there and to null in the neural doped regions. It is possible to extract the electrostatic potential profiles from the OBIC profiles. However, this is not a trivial problem since the relationship is highly nonlinear. The details are beyond this text [92]. Compared to scanning electron beam or scanning Kelvin probe techniques, the OBIC method is simple to implement and widely used to characterize semiconductor junction structures and to map the minority carrier lifetime, defects, local resistance and cell uniformity of thin film solar cells. Application of the OBIC technique to an LEC was first reported by Dick et al [63]. A focused Argon laser beam was scanned across an encapsulated planar LEC approximately 20 lm wide mounted on a cooled scanning stage. VOC, rather than ISC, was measured to prevent rapid dedoping of the activated cell. The detection of a peak VOC where the PL intensity showed a large step is consistent with the presence of a p-n junction. The width of the junction was estimated to be about 2 lm or 10% of the interelectrode gap. Since the activated device was only cooled to 250 K, well above the Tg of the electrolyte, the planar LEC was still prone to dedoping. The magnitude of the VOC, at only a few tens of lV, was minuscule compared to the energy gap of the luminescent polymer. Since the I-V characteristics of a p-n junction are not linear, the VOC profile would be different from the OBIC profile in width and shape. The challenges of optically scanning a fully frozen, small planar LEC in a cryogenic vacuum chamber (to avoid condensation) meant it was the only study of its kind in 15 years since its publication. With the advent of extremely large planar LECs, we carried out several OBIC studies of planar LECs that are frozen in a vacuum cryostat. The frozen cells had an interelectrode spacing ranging from 700 lm to 4.6 mm. The extremely large gap size made it possible to use a variety of optical/cryogenic setups for added functionalities. The large planar cells were also easier to fabricate using shadow masking (versus photolithographic) techniques. In all these studies, the planar cells were activated at elevated temperatures, and cooled to 200 K or below to freeze the junction. In the remainder of this section, we briefly introduce the first two OBIC studies carried out in our lab. In Sects. 1.3 and 1.4, we describe in detail the latest OBIC/PL scans with a focused laser beam.

2 Optical-Beam-Induced-Current Imaging of Planar Polymer …

2.2.2

61

OBIC Scanning of Planar LECs with a Micromanipulated Cryogenic Probe Station

Hu and Gao turned on a 3.1 mm planar LEC in a micromanipulated cryogenic probe station (Janis ST-500) under vacuum [41]. The planar LEC had a composition of MEH-PPV (10):PEO(10):CsClO4(3) in weight ratio and a pair of Au/Al electrodes deposited on top of the polymer film. The micromanipulated probes were used to make electrical contacts to the electrodes. The planar LEC was activated with a voltage bias of 300 V at 335 K. The cell current reached 3 mA and decreased when the cell was gradually cooled to 200 K to freeze the junction. The probe station was equipped with a fibre optical arm. The optical fibre was coupled to a 442 nm He-Cd laser beam and manually scanned across the entire interelectrode gap of the frozen planar LEC. OBIC scans were performed in both short-circuit and open-circuit conditions along the same path. Figure 2.8 shows the spatial OBIC and VOC profiles of the scans along the path shown. Both profiles were very broad due to the large core diameter of the optical fibre used (200 lm) and the jaggedness of the junction. The depletion width of the junction, therefore, was not resolved. The OBIC and VOC peaks, however, coincided precisely with the position of the p-n junction, shown above the scan profiles. The peak VOC was over 0.6 V (versus a few tens of lV of the initial OBIC study), indicating a significant junction built-in potential.

Fig. 2.8 OBIC photocurrent and photovoltage profiles of a 3.1 mm frozen planar LEC measured in a Janis micromanipulated cryogenic probe station. Top the activated cell under UV illumination. The vertical white lines indicate the electrode/polymer film interfaces. The yellow line and arrow depict the scan path and direction. Bottom Photocurrent and photovoltage profiles for the scan path shown at the top. Reprinted from reference [41], Copyright (2011), with permission from Elsevier

62

2.2.3

F. AlTal and J. Gao

Concerted OBIC and Scanning PL Imaging of Planar LECs with a Fluorescence Microscope

Subsequently, Inayeh et al. utilized a low profile microscopy cryostat and a fluorescence microscope to perform concerted PL and OBIC scans of planar frozen-junction LECs [42]. A schematic of this setup is shown in Fig. 2.9. The scanning optical beam was a focused beam of the mercury lamp attached to the fluorescence microscope. An octagon-shaped aperture was placed in the optical path. Moreover, the beam was focused to the surface of the planar LEC with a 40 objective to a size of about 35 µm in diameter. This represented a significant improvement in scanning resolution compared to the probe station. Moreover, the setup allows for a simultaneous recording of the PL intensity of the film with the photodiode positioned below the cryostat. This configuration was made possible by

Fig. 2.9 Top Image of the undoped MEH-PPV:PEO:KTf planar LEC with a 1.0 mm interelectrode gap. Also shown are the illumination spots created using a 10 objective lens and a 40 objective lens. Bottom Schematic illustrating the experimental setup used to perform the OBIC and photoluminescence scans. Blue light (from 448 nm to 497 nm) originating from the mercury lamp is focused through a 40 objective lens. The light travels through the cryostat window and excites the surface of the device. Unabsorbed blue light and photoluminescence from the LEC travel through the bottom window of the cryostat. A 550 nm longpass filter removes the unabsorbed blue light. The photodiode detects the photoluminescence intensity of the LEC film. The LEC is mounted in a microscopic cryostat and kept under vacuum. Reprinted with permission from reference [42]. Copyright (2012) American Institute of Physics

2 Optical-Beam-Induced-Current Imaging of Planar Polymer …

63

Fig. 2.10 OBIC photocurrent and photoluminescence intensity profiles of the frozen-junction cell shown in Fig. 2.2, an MEH-PPV:PEO:KTf planar LEC with a 1.0 mm interelectrode gap turned on with 25 V and cooled to 200 K. Top portion of the cell illuminated under blue light (the full cell is not illuminated during the OBIC scan). The blue lines depict the area of the cell exposed to light during the scan and the blue arrow indicates the direction of the scanning optical beam. Middle photocurrent and fluorescence intensity profiles of the scan path shown at top. Bottom photocurrent and differential change in fluorescence intensity profiles of the same scan path. Reprinted with permission from reference [42]. Copyright (2012) American Institute of Physics

the fact that the microscopy cryostat had both a top and a bottom optical windows, and the copper cold finger was partially hollowed out to let the light beam through. The beam was scanned across the cell by moving the cryostat which was mounted on a motorized scan stage with steps of 10 µm. Figure 2.10 shows the acquired OBIC and PL profiles from a 1 mm frozen-junction planar LEC. Even without the photograph shown at the top, we can easily observe that the OBIC peak is located at the junction region, where the PL intensity underwent a sharp transition. Detailed analysis, however, revealed that both the PL transition region and the OBIC peak had a width comparable to the beam width. This indicates that the beam was still not narrow enough to resolve the depletion width of the frozen junction. Further improvement to the scanning resolution was needed.

64

2.3 2.3.1

F. AlTal and J. Gao

OBIC and Scanning PL Probing of a Frozen Planar pi-n Junction Introduction

In Sect. 2.2, we introduced the OBIC technique as a valuable tool to probe the electronic structure of planar LECs. Dick et al. first demonstrated the feasibility of measuring the spatially resolved VOC and PL intensity across a planar LEC, although the LEC studied was not fully frozen, nor was the photocurrent measured [63]. The invention of extremely large planar LECs makes the OBIC technique extremely attractive due to their ease of fabrication and compatibility with various scanning apparatus. Our group’s first two attempts at scanning extremely large planar LECs were successful in that (1) the cells were sufficiently cooled to freeze the LEC junction. (2) the OBIC was measured for the first time as well as the VOC. The latter was on the order of the built-in potential expected, and (3) the OBIC and PL signals were obtained concurrently. The PL trace provides a reliable reference for determining the position of the OBIC peak relative to the junction position. The results showed that the peak OBIC and VOC occurred at the junction. The polymer/electrode interfaces and much of the neutral doped regions were not photovoltaic-active. These studies described briefly in Sects. 2.2.2 and 2.2.3, however, did not resolve the depletion width of the LEC junction due to the limitation of the scanning apparatus. The beam size in the second study was about 35 lm, so it was interesting to note that the increase in the gap size of the planar cell did not lead to a wider depletion width that scales with the size of the planar cell. To improve the scanning resolution, the second scanning setup was modified and employed to scan a unique planar polymer p-i-n junction [43]. For the first time, the depletion width of a planar LEC junction has been resolved from a scanning optical measurement. This section describes in detail the experimental setup and the obtained results.

2.3.2

Experimental Details

A planar LEC with a 700 lm gap size was fabricated on a 15 mm  15 mm 1 mm sapphire substrate inside a nitrogen-filled glovebox/evaporator system. The LEC film was spin cast from a cyclohexanone solution of MEH-PPV, PEO and potassium triflate (KTf) with a weight ratio of 10:5:1.2 and subsequently dried for 5 h at 50 °C. After that, aluminium electrodes with a thickness of 100 nm were deposited on top of the LEC film under a vacuum of *1.5  10−6 torr. The finished planar LEC had an active area of 8 mm by 700 lm. The planar LEC was loaded into a microscopy cryostat, sealed and transferred out of the glovebox for testing.

2 Optical-Beam-Induced-Current Imaging of Planar Polymer …

65

The scanning OBIC/PL apparatus again consists of a Nikon fluorescence microscope and a motorized scanning stage onto which the cryostat is mounted. In addition, a Keithley-237 source measurement unit was used to supply the voltage bias and simultaneously measure the device current during both the activation and the scanning processes. A photodiode recorded the PL intensity of the cell during the scan, and the cell temperature was controlled via a Cryocon 32 controller. The measurement was controlled with a custom LabView program. These arrangements are similar to the last OBIC/PL scan. The optics of the scanning setup, however, had been changed. Instead of focusing down the beam of the attached mercury lamp, the lamp assembly was removed from the microscope. A single mode He-Cd laser (442 nm) was used as the light source. Mirrors and a 6.5/1.1 Galilean telescope were used to steer and expand and elevated the laser beam. A plano-convex lens with a focal length of 15 cm was used to couple the beam into the microscope’s rear aperture. The dichroic mirror inside the microscope directed the laser beam into a 10 objective and focused it onto the device under test through the cryostat optical window. The 2D beam profile was captured using the CCD camera mounted on top of the microscope. The 1D intensity profile of the scanning beam was obtained by integrating the 2D profile perpendicular to the scanning direction. The resulting profile fits well to a Gaussian with a 1/e2 waist diameter of *13 lm, about 1/10 of the size of the focused lamp beam using the same 10 objective.

2.3.3

Resolving the Depletion Width of a Planar p-i-n Junction

The 700 lm planar LEC was activated by applying a voltage bias of 150 V. The cell temperature was kept at 325 K for about 250 s and subsequently increased to 330 K to speed up the activation process. When the device current had reached about 1.3 mA (t = 850 s) the flow of liquid nitrogen was turned on to cool the device at a rate of −0.365 K/s until the device reached 200 K. The device was illuminated with a 365 nm wavelength UV lamp during the activation process. The time-lapse fluorescence images of the cell are shown in the top part of Fig. 2.11. Like the large planar cells shown in Figs. 2.5 and 2.6, in situ electrochemical doping, manifested as PL quenching, was again highly visible under UV illumination. The doped regions expanded until the doping fronts met and the EL of the cell started to grow stronger. Cooling led to a dimmer and red-shifted EL at t = 1300 s. The last image in the second row was taken at 200 K with the 150 V bias removed. A significant observation was a bright line between the p- and n-doped regions. This is different from the frozen cells shown in Figs. 2.8 and 2.10. The bright line has stronger PL than the neighbouring p- and n-doped regions and is a less quenched quasi-intrinsic region. Therefore, the doping profile of this planar cell is that of a p-i-n junction. The focused laser beam was scanned across the frozen planar LEC with a step size of 1 lm, in a direction that was perpendicular to the p-i-n junction. Each scan

66

F. AlTal and J. Gao

Fig. 2.11 Top device snapshots during the turn-on and freezing process. The top electrode (anode) was positively biased relative to the lower electrode (cathode). The p-doped region at the top is visibly darker than the n-doped region in the lower part of the film. Only 1,100 lm by 700 lm of the entire device active area is shown. Bottom Deconvoluted photocurrent and PL profiles. The background in the lower panel shows the frozen LEC pictured at zero bias. The scan was done along the central horizontal line in the figure. The UV lamp was turned off. Reprinted with permission from reference [43]. Copyright (2015) WILEY-VCH Verlag GmbH & Co

generated a set of raw OBIC and PL profiles. Since the scanning laser beam had a significant size of about 13 lm in diameter, deconvolution was necessary to extract the true OBIC/PL profiles. This was done by fitting the measured PL and OBIC peaks to separate Gaussian functions. The Gaussian laser beam was then deconvoluted from these fitting functions. The deconvoluted OBIC and PL curves are shown in the lower part of Fig. 2.11. The photocurrent peak has a 1/e2 width of *18 lm. The PL peak, on the other hand, extends over a wider area. Since the exciton diffusion length in PPVs is on the order of nm, the width of the deconvoluted OBIC peak is a good approximation of the width of the depletion region. We note that the PL profile indeed includes a prominent PL peak in the junction region. Also, the OBIC peak is displaced by about 12 lm to the right, close to the i/p interface. It is reasonable to assume that the peak OBIC position is

2 Optical-Beam-Induced-Current Imaging of Planar Polymer …

67

also the position of the peak built-in field. The most significant discovery of this study is the determination of the depletion width of a planar, frozen polymer junction. Moreover, the LEC junction has the doping profile of a p-i-n junction. Manzanares and Heeger first discussed the possibility of a p-i-n junction in a polymer LEC. A majority of the doping profiles observed in extremely large planar LECs, however, have been those of a p-n junction. These as-formed p-n junctions can relax into a p-i-n junction. The OBIC profile of a p-n junction, as well as a partially relaxed p-n junction, is described in the next section.

2.4 2.4.1

High-Resolution OBIC and Scanning PL Imaging of a Frozen Planar Polymer p-n Junction Introduction

The ultimate goal of a scanning optical measurement, when applied to planar LECs, is to uncover the inner electronic structure of the LEC junction that is responsible for all of LEC’s functionalities as either a light-emitting device or a photovoltaic cell. To achieve this goal, the scanning resolution must be sufficiently high compared to the features to be resolved. The scanning setup described in Sect. 2.3 produced a scanning beam of about 13 lm in diameter. While it was sufficient to resolve the depletion width of a planar p-i-n junction, the beam diameter is still too large compared to what is obtainable within the diffraction limit. We note that in the first OBIC study by Dick et al. the laser beam was focused to have a diameter of about 1 lm with the use of a 40 microscope objective [63]. A high-powered objective, however, could not be used in the previous setup as it led to a second focused spot on the device surface. This was likely caused by some internal reflections of the laser beam which entered the microscope from the lamp port. In this section, we describe a revised optical scanning setup that has overcome this problem. The focused beam had a Gaussian shape with a 1/e2 width of only 1.9 µm. The planar LEC itself had also been optimized by fine-tuning the LEC blend, the electrode material and the operating conditions. The planar LEC was activated to exhibit a straight, long and highly emissive p-n junction that had been successfully frozen. The concerted, high-resolution OBIC and scanning PL imaging the frozen planar LEC exposed the narrowest p-n junction in a frozen LEC ever reported [44].

2.4.2

Experimental Details

The planar LECs in this study had a composition of MEH-PPV (10): PEO (5): KTf (1.2) by weight and a thickness of about 0.5 lm. A pair of thermally evaporated gold electrodes defined a 700 µm interelectrode spacing and a cell length of 7 mm.

68

F. AlTal and J. Gao

Fig. 2.12 Left Experimental setup for scanning OBIC and PL imaging of frozen planar LECs. Right 2D beam profile and Gaussian fit of the focused excitation beam. Reprinted with permission from reference [44]. Copyright (2016) American Institute of Physics

The same microscopy cryostat was used to house the cell, and the device fabrication and testing procedures are identical to those described in Sect. 2.3.2. The light source and optics have been modified. The experimental setup is shown in Fig. 2.12. A blue single mode diode laser (473 nm) was used as the excitation source. The laser beam was steered using mirrors (M1, M2) and coupled into a single mode fibre (SMF) via a coupler lens. Another lens was attached at the other terminal of the fibre to collimate the output beam which was then expanded in size. A 50/50 beam splitter cube then redirected the beam into a 50 objective and at the same time allowed light collected by the objective to pass through. This arrangement allowed the excitation beam to by-pass the microscope optics. In addition, the mercury lamp was reattached which was used to take fluorescence images of the cell during activation. The focused Gaussian beam had a 1/e2 width of 1.9 µm at the device surface. A silicon photodiode positioned under the bottom optical window collected the PL from the LEC film. The excitation beam was rejected by a low-pass filter (LPF) with a cut-off wavelength of 500 nm, between the optical window and the photodiode. The device PL pictures were captured via a 10 objective using a CCD camera that is mounted on top of the microscope. This provided a field of view that covers an area of 0.9 mm  1.8 mm.

2.4.3

Results and Discussion

The planar LEC was activated by applying a 20 V DC bias at a temperature of 360 K. Subsequently; it was cooled at a rate of 0.19 K/s to 170 K. The lower freezing temperature of 170 K compared to 200 K used in previous scans ensured a fixed junction for multiple optical scans. The frozen junction was verified by repeated I-V scans, which did not show change over the course of the experiment. The top panel of Fig. 2.13 depicts the cell geometry and the time evolution of cell current and cell temperature during the activation process. We note that a very high

2 Optical-Beam-Induced-Current Imaging of Planar Polymer …

69

peak current of over 5 mA had been reached, which is an indication of strong doping and good film quality. The lower panel displays the time-lapse fluorescence images of the cell. The LEC film is uniformly fluorescent before the voltage bias was applied (t = 0 s). The doping fronts of this cell propagated much faster than the earlier cells shown in Sects. 2.2 and 2.3. At t = 3 s, a large part of the LEC film was already doped to either p-type on the anode (+) side or n-type on the cathode (−) side. At t = 9 s, a continuous light-emitting p-n junction was visible between the doped regions. The p-n junction was initially very uneven but became much straighter with time. Meanwhile, the doped regions became darker, indicating an increase in doping level. The darkening of the LEC film was accompanied by the rapid increase in cell current. The applied voltage bias was removed once the target temperature of 170 K had been reached. The PL image of the frozen cell revealed a p-n junction doping profile. A thin bright line was observed in the junction region. However, the intensity of the bright line was not high enough to suggest a p-in junction structure. Moreover, it will be shown that the as-activated p-n junction could be relaxed into a p-i-n junction when subjected to warming/re-cooling cycles. A total of 18 OBIC/PL scans were performed across the frozen p-n junction as seven locations. Figure 2.14 shows the OBIC and PL profiles of a full

Fig. 2.13 Device activation process. Top Time evolution of the cell current and temperature during the activation process. A DC voltage bias of 20 V was applied. The inset shows a schematic of the planar LEC. Bottom Time-lapse fluorescence images of the planar LEC during the activation process. Only a section of the entire cell is shown. The bright line formed after t = 9 s is due to electroluminescence of the cell. Reprinted with permission from reference [44]. Copyright (2016) American Institute of Physics

70

F. AlTal and J. Gao

Fig. 2.14 OBIC and PL profiles of the frozen planar LEC across the same junction region. Top Full OBIC and PL scans of the planar LEC across the entire planar LEC. The cathode is located at −230 µm and the anode is located at 470 µm, as indicated by the vertical dashed lines. Bottom OBIC data of four scans and Gaussian fit to the data near the OBIC peak. Reprinted with permission from reference [44]. Copyright (2016) American Institute of Physics

(electrode-to-electrode) scan. The OBIC peak was assigned a beam position of zero lm as a reference. The PL profile indicates an n-doped region with the highest overall PL intensity and a darker, p-doped region to the right of the OBIC peak. A small local PL peak just to the left of the OBIC peak can be attributed to the thin bright line observed in the PL image. The sensitivity of the PL scan is evident in the detection of a PL “transition zone” on the p side of the junction, between 0 and +30 lm. In this zone, the PL intensity decreased continuously from the high level of the n side to the much lower level of the p side. The large variation in PL intensity is also visible near both electrode interfaces. This variation indicates a large doping gradient, and the level of doping is the highest just inside the electrode edges, for both p- and n-doped films. This observation is consistent with the results of the contact probing measurement. The most significant observation is the extremely narrow and prominent OBIC peak at x = 0 lm. The lower panel of the figure provides an expanded view of the OBIC peak that incorporated the data points from a total of four scans across the same junction region. The data points of the individual scan were adjusted/shifted so that their peak positions coincide. This procedure was necessary to eliminate any offset due to the mechanical hysteresis of the scanning stage. The data are fitted to a Gaussian function, from which a 1/e2 width of 3.1 lm was determined. Since both the scanning beam and the measured OBIC profile are Gaussian functions, deconvolution also results in a Gaussian whose width is given by pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 2  WBeam WOBIC , where W is the 1/e2 Gaussian width. The deconvoluted OBIC peak width is 2.45 lm. We note that all 18 scans yielded an OBIC peak width that was larger than the excitation beam diameter of 1.9 lm. An average of all 18 scans gives an OBIC width of 2.53 lm with a standard deviation of 0.45 lm. The average OBIC peak width after deconvolution is 1.5 lm. The average OBIC signal is 292 ±40 pA for an excitation beam intensity of 2 µW.

2 Optical-Beam-Induced-Current Imaging of Planar Polymer …

71

Since the exciton diffusion length in MEH-PPV is on the order of nm, the OBIC peak width should represent the junction depletion width. The average junction width of 1.5 lm is the smallest ever measured in a planar LEC, regardless of the interelectrode spacing. The junction width accounts for a mere 0.21% of the interelectrode gap, which is the lowest value reported for LECs. The OBIC junction width is also much smaller than the width of the EL zone (*10 µm) and the PL transition zone (*30 µm). The small junction width obtained attests to the high resolution of the scans. The drop-off in PL intensity to the right of the OBIC peak indicates a doping gradient at the edge of the p-n junction. There is a strong possibility that the doping gradient is caused by the presence of rough features on the sub-micrometre scale. Sharp protrusions on a larger scale had been observed along a planar p-n junction [80]. They contributed to a tunnelling leakage current that degraded the rectification ratio of the as-formed frozen p-n junction. The sharp features, however, could be removed by subjecting the frozen p-n junction to a warming/cooling cycle that caused partial dedoping and smoothing the p-n junction. The “relaxed” p-n junction exhibit improved rectification and a much larger VOC when illuminated. In an attempt to smooth out the junction and eliminate any possible submicron protrusions the frozen cell was partially relaxed after the aforementioned scans. The relaxation/dedoping cycle was carried out by briefly (for a few minutes at a time) warming the frozen cell to 260 K and cooling it back to 170 K. During these thermal cycles, the cell was kept at an open-circuit condition to avoid fast dedoping and the loss of the active junction. In total, four relaxation cycles were carried out. The cell current decreased after each cycle, and the I-V curves became more nonlinear and less symmetric. This is a strong indication that dedoping had occurred. OBIC scans were performed after each dedoping cycle along the same junction location. As the cell became more resistive, the input optical power was increased to obtain a measurable OBIC signal. Figure 2.15 compare the normalized OBIC profiles without dedoping and after the final dedoping cycle. The measured OBIC profile narrowed to *2 lm after dedoping. After subtracting the beam width, the junction width is only 0.6 lm compared to 2.45 lm without dedoping. A very prominent PL peak appeared to the left of the OBIC peak. The PL profile became similar to that of a pi-n junction shown in Fig. 2.11. The p-i-n junction formed by controlled dedoping, however, is quite different from an as-formed p-i-n junction discussed in Sect. 1.3. Here, the removal of the fine features along the edge of the junction resulted in the narrowest OBIC peak ever reported for a planar LEC. The p-i-n junction had sharp boundaries between the differently doped regions. Indeed, the PL transition region between the “i” region and the p region has narrowed to about 3 lm from nearly 30 lm in the as-formed p-n junction. The more abrupt PL transition is consistent with a narrowed OBIC profile and a reduced junction width.

72

F. AlTal and J. Gao

Fig. 2.15 Normalized OBIC scans at the same location before and after four dedoping cycles. The excitation laser power used was 2 µW (red) before dedoping and 50 µW (blue) after four dedoping cycles. Reprinted with permission from reference [44]. Copyright (2016) American Institute of Physics

2.5

Conclusion and Outlook

Polymer-based LECs are intriguing and promising devices that offer attractive device characteristics not easily attained by other organic electronic devices. Doping plays an essential role in the operation of polymer-based LECs. On one hand, doping leads to conductivity increase that gives LECs their desirable electrical properties. On the other hand, the optical effect of doping allows for the visualization of the dynamic doping process via time-lapse fluorescence imaging. In this chapter, we summarized our recent experimental work on the scanning optical imaging of planar LECs. The experiments exploited both the electrical and optical effects of doping, as well as the temperature dependence of ionic conductivity of the LEC film. For the first time, we had resolved the junction depletion width of a planar p-i-n junction and a planar p-n junction. The narrowing of the junction upon thermal cycling strongly suggests the presence of fine structures on the edge of the junction. For an as-formed, frozen p-n junction, the depletion width accounted for only 0.2% of the entire cell area enclosed by the electrodes. Since only the junction region contributes to EL and PV response in an LEC, the narrow junction width is not ideal for efficient operation of LECs as either a light-emitting device or a PV cell. An igneous solution to this problem is to form multiple junctions that are simultaneously emitting without increasing the total cell area. By introducing dispersed, metallic particles to the LEC film, we successfully demonstrated a new device structure called a bulk homojunction (BHoJ) LEC [93–96]. A BHoJ planar LEC exhibits vastly improved effective emitting area as well as a giant VOC when operated as a PV cell. BHoJ LECs in a sandwich configuration and frozen at room temperature represent a major challenge and opportunity in LEC research. Acknowledgements The optical scanning studies were supported by the Natural Sciences and Engineering Research Council of Canada (NSERC). Faleh AlTal is supported by an Ontario Trillium Scholarship.

2 Optical-Beam-Induced-Current Imaging of Planar Polymer …

73

References 1. S.B. Meier, D. Tordera, A. Pertegas, C. Roldan-Carmona, E. Orti, H.J. Bolink, Mater. Today 17, 217–223 (2014) 2. Y. Xiong, L. Li, J.J. Liang, H. Gao, S.Y. Chou, Q.B. Pei, Mater. Horiz. 2, 338–343 (2015) 3. A. Sandstrom, L. Edman, Energy Technol. 3, 329–339 (2015) 4. Q. Pei, Y. Yang, G. Yu, Y. Cao, A.J. Heeger, Synth. Met. 85, 1229–1232 (1997) 5. Q. Sun, Y. Li, Q. Pei, J. Disp. Technol. 3, 211–224 (2007) 6. J.F. Fang, Y.L. Yang, L. Edman, Appl. Phys. Lett. 93 (2008) 7. J.M. Leger, Adv. Mater. 20, 837–841 (2008) 8. Q.B. Pei, G. Yu, C. Zhang, Y. Yang, A.J. Heeger, Science 269, 1086–1088 (1995) 9. Q.B. Pei, Y. Yang, G. Yu, C. Zhang, A.J. Heeger, J. Am. Chem. Soc. 118, 3922–3929 (1996) 10. R.H. Friend, R.W. Gymer, A.B. Holmes, J.H. Burroughes, R.N. Marks, C. Taliani, D.D.C. Bradley, D.A. Dos Santos, J.L. Bredas, M. Logdlund, W.R. Salaneck, Nature 397, 121–128 (1999) 11. Y. Cao, I.D. Parker, G. Yu, C. Zhang, A.J. Heeger, Nature 397, 414–417 (1999) 12. Y. Cao, Q.B. Pei, M.R. Andersson, G. Yu, A.J. Heeger, J. Electrochem. Soc. 144, L317–L320 (1997) 13. Y. Cao, G. Yu, A.J. Heeger, C.Y. Yang, Appl. Phys. Lett. 68, 3218–3220 (1996) 14. Y. Yang, Q.B. Pei, J. Appl. Phys. 81, 3294–3298 (1997) 15. Z.T. Zhang, K.P. Guo, Y.M. Li, X.Y. Li, G.Z. Guan, H.P. Li, Y.F. Luo, F.Y. Zhao, Q. Zhang, B. Wei, Q.B. Pei, H.S. Peng, Nat. Photonics 9, 233–238 (2015) 16. Y. Nishikitani, D. Takizawa, H. Nishide, S. Uchida, S. Nishimura, J. Phys. Chem. C 119, 28701–28710 (2015) 17. K.O. Burnett, P.P. Crooker, N.M. Haegel, Y. Yoshioka, D. MacKenzie, Synth. Met. 161, 1496–1499 (2011) 18. Z.Z. You, G.J. Hua, Int. J. Electron. 97, 99–104 (2010) 19. J. Mindemark, S. Tang, J. Wang, N. Kaihovirta, D. Brandell, L. Edman, Chem. Mater. 28, 2618–2623 (2016) 20. J. Mindemark, L. Edman, J. Mater. Chem. C 4, 420–432 (2016) 21. Z. Shu, O. Pabst, E. Beckert, R. Eberhardt, A. Tunnermann, Mater. Today Proc. 3, 733–738 (2016) 22. J.H. Jang, L.H. Kim, Y.J. Jeong, K. Kim, T.K. An, S.H. Kim, C.E. Park, Org. Electron. 34, 50–56 (2016) 23. A. Asadpoordarvish, A. Sandstrom, L. Edman, Adv. Eng. Mater. 18, 105–110 (2016) 24. K. Shanmugasundaram, M.S. Subeesh, C.D. Sunesh, Y. Choe, RSC Adv. 6, 28912–28918 (2016) 25. A. Pertegas, M.Y. Wong, M. Sessolo, E. Zysman-Colman, H.J. Bolink, Ecs J. Sol. State Sci. Techn. 5, R3160–R3163 (2016) 26. Q.Y. Zeng, F.S. Li, T.L. Guo, G.G. Shan, Z.M. Su, Sci. Rep. 6, 9 (2016) 27. J.Q. Wu, F.S. Li, Q.Y. Zeng, C. Nie, P.C. Ooi, T.L. Guo, G.G. Shan, Z.M. Su, Org. Electron. 28, 314–318 (2016) 28. M.S. Subeesh, K. Shanmugasundaram, C.D. Sunesh, R.K. Chitumalla, J. Fang, Y. Choe, J. Phys. Chem. C 120, 12207–12217 (2016) 29. D.A.W. Ross, P.A. Scattergood, A. Babaei, A. Pertegas, H.J. Bolink, P.I.P. Elliott, Dalton Transact. 45, 7748–7757 (2016) 30. L.D. Bastatas, K.Y. Lin, M.D. Moore, K.J. Suhr, M.H. Bowler, Y.L. Shen, B.J. Holliday, J.D. Slinker, Langmuir 32, 9468–9474 (2016) 31. M.D. Weber, M. Adam, R.R. Tykwinski, R.D. Costa, Adv. Funct. Mater. 25, 5066–5074 (2015) 32. M.F. Ayguler, M.D. Weber, B.M.D. Puscher, D.D. Medina, P. Docampo, R.D. Costa, J. Phys. Chem. C 119, 12047–12054 (2015) 33. Q. Pei, A.J. Heeger, Nat. Mater. 7, 167 (2008)

74

F. AlTal and J. Gao

34. G.G. Malliaras, J.D. Slinker, J.A. DeFranco, M.J. Jaquith, W.R. Silveira, Y.-W. Zhong, J.M. Moran-Mirabal, H.G. Craighead, H.D. Abruna, J.A. Marohn, Nat. Mater. 7, 168 (2008) 35. S. Jenatsch, L. Wang, M. Bulloni, A.C. Veron, B. Ruhstaller, S. Altazin, F. Nuesch, R. Hany, A.C.S. Appl, Mater. Interf. 8, 6554–6562 (2016) 36. S. van Reenen, R.A.J. Janssen, M. Kemerink, Adv. Funct. Mater. 25, 3066–3073 (2015) 37. S. van Reenen, P. Matyba, A. Dzwilewski, R.A.J. Janssen, L. Edman, M. Kemerink, J. Am. Chem. Soc. 132, 13776–13781 (2010) 38. S. van Reenen, R.A.J. Janssen, M. Kemerink, Org. Electron. 12, 1746–1753 (2011) 39. G. Gozzi, L.F. Santos, R.M. Faria, Epl. 100 (2012) 40. A. Munar, A. Sandstrom, S. Tang, L. Edman, Adv. Funct. Mater. 22, 1511–1517 (2012) 41. Y.F. Hu, B. Dorin, F. Teng, J. Gao, Org. Electron. 13, 361–365 (2012) 42. A. Inayeh, B. Dorin, J. Gao, Appl. Phys. Lett. 101, 253305 (2012) 43. F. AlTal, J. Gao, Phys. Status Solidi Rapid Res. Lett. 9, 77–81 (2015) 44. F. AlTal, J. Gao, J. Appl. Phys. 120, 8 (2016) 45. Y. Shao, G.C. Bazan, A.J. Heeger, Adv. Mater. 19, 365–370 (2007) 46. J. Gao, Y.F. Li, G. Yu, A.J. Heeger, J. Appl. Phys. 86, 4594–4599 (1999) 47. J. Gao, G. Yu, A.J. Heeger, Appl. Phys. Lett. 71, 1293–1295 (1997) 48. J. Gao, G. Yu, A.J. Heeger, Adv. Mater. 10, 692–695 (1998) 49. G. Wantz, B. Gautier, F. Dumur, T.N.T. Phan, D. Gigmes, L. Hirsch, J. Gao, Org. Electron. 13, 1859–1864 (2012) 50. G. Yu, Y. Cao, M. Andersson, J. Gao, A.J. Heeger, Adv. Mater. 10, 385–388 (1998) 51. J.M. Leger, D.B. Rodovsky, G.R. Bartholomew, Adv. Mater. 18, 3130 (2006) 52. D.T. Simon, D.B. Stanislowski, S.A. Carter, Appl. Phys. Lett. 90, 103508 (2007) 53. J.M. Leger, D.G. Patel, D.B. Rodovsky, G.P. Bartholomew, Adv. Funct. Mater. 18, 1212– 1219 (2008) 54. C.V. Hoven, H.P. Wang, M. Elbing, L. Garner, D. Winkelhaus, G.C. Bazan, Nat. Mater. 9, 249–252 (2010) 55. I.V. Kosilkin, M.S. Martens, M.P. Murphy, J.M. Leger, Chem. Mater. 22, 4838–4840 (2010) 56. S. Tang, K. Irgum, L. Edman, Org. Electron. 11, 1079–1087 (2010) 57. Z.B. Yu, L. Li, H.E. Gao, Q.B. Pei, Sci. China Chem. 56, 1075–1086 (2013) 58. Z.B. Yu, M.L. Sun, Q.B. Pei, J. Phys. Chem. B 113, 8481–8486 (2009) 59. S. Tang, L. Edman, Electrochim. Acta 56, 10473–10478 (2011) 60. B. Gautier, J. Gao, Appl. Phys. Lett. 101, 093302 (2012) 61. B. Gautier, X.M. Wu, F. Altal, S.L. Chen, J. Gao, Org. Electron. 28, 47–52 (2016) 62. U. Lemmer, D. Vacar, D. Moses, A.J. Heeger, T. Ohnishi, T. Noguchi, Appl. Phys. Lett. 68, 3007–3009 (1996) 63. D.J. Dick, A.J. Heeger, Y. Yang, Q.B. Pei, Adv. Mater. 8, 985–987 (1996) 64. J. Gao, J. Dane, Appl. Phys. Lett. 83, 3027–3029 (2003) 65. A.L. Holt, J.M. Leger, S.A. Carter, J. Chem. Phys. 123, 044704 (2005) 66. N. Kaihovirta, A. Asadpoordarvish, A. Sandstrom, L. Edman, ACS Photonics 1, 182–189 (2014) 67. J. Gao, J. Dane, Appl. Phys. Lett. 84, 2778–2780 (2004) 68. J. Dane, C. Tracy, J. Gao, Appl. Phys. Lett. 86, 153509 (2005) 69. Y.F. Hu, C. Tracy, J. Gao, Appl. Phys. Lett. 88, 123507 (2006) 70. J.H. Shin, A. Dzwilewski, A. Iwasiewicz, S. Xiao, A. Fransson, G.N. Ankah, L. Edman, App. Phys. Let. 89, 013509 (2006) 71. Y.F. Hu, J. Gao, J. Am. Chem. Soc. 133, 2227–2231 (2011) 72. S. Alem, J. Gao, Org. Electron. 9, 347–354 (2008) 73. D. Hohertz, J. Gao, Adv. Mater. 20, 3298–3302 (2008) 74. J.H. Shin, P. Matyba, N.D. Robinson, L. Edman, Electrochim. Acta 52, 6456–6462 (2007) 75. Y.F. Hu, J. Gao, Appl. Phys. Lett. 89, 253514 (2006) 76. Y. Hu, Y. Zhang, J. Gao, Adv. Mater. 18, 2880 (2006) 77. J.-H. Shin, N.D. Robinson, S. Xiao, L. Edman, Adv. Funct. Mater. 17, 1807–1813 (2007) 78. N.D. Robinson, J.-H. Shin, M. Berggren, L. Edman, Phys. Rev. B 74, 155210 (2006)

2 Optical-Beam-Induced-Current Imaging of Planar Polymer …

75

79. Y.G. Zhang, Y.F. Hu, J. Gao, Appl. Phys. Lett. 88, 163507 (2006) 80. Y. Zhang, Y. Hu, J. Gao, Appl. Phys. Lett. 91, 233509 (2007) 81. P. Matyba, K. Maturova, M. Kemerink, N.D. Robinson, L. Edman, Nat. Mater. 8, 672–676 (2009) 82. L.S.C. Pingree, D.B. Rodovsky, D.C. Coffey, G.P. Bartholomew, D.S. Ginger, J. Am. Chem. Soc. 129, 15903–15910 (2007) 83. D.B. Rodovsky, O.G. Reid, L.S.C. Pingree, D.S. Ginger, ACS Nano 4, 2673–2680 (2010) 84. J.D. Slinker, J.A. DeFranco, M.J. Jaquith, W.R. Silveira, Y.W. Zhong, J.M. Moran-Mirabal, H.G. Craighead, H.D. Abruna, J.A. Marohn, G.G. Malliaras, Nat. Mater. 6, 894–899 (2007) 85. T. Wilson, C. Sheppard, Theory and Practice of Scanning Optical Microscopy (Academic Press, London; Orlando, 1984) 86. M. Ogura, K. Sakaue, Y. Tokumaru, Jap. J. Appl. Phys. Part 2 Lett. 24, L617–L619 (1985) 87. C. Werner, D. Reuter, A.D. Wieck, Phys. E Low Dimens. Syst. Nanostruct. 32, 508–511 (2006) 88. M.G. Liu, S.L. Wang, J.C. Campbell, J.D. Beck, C.F. Wan, M.A. Kinch, J. Appl. Phys. 98 (2005) 89. C. Raynaud, D.M. Nguyen, N. Dheilly, D. Tournier, P. Brosselard, M. Lazar, D. Planson, Phys. Status Solidi A Appl. Mat. 206, 2273–2283 (2009) 90. L. Onsager, Phys. Rev. 54, 554–557 (1938) 91. C.L. Braun, J. Chem. Phys. 80, 4157–4161 (1984) 92. S. Busenberg, W.F. Fang, K. Ito, Siam J. Appl. Math. 53, 187–204 (1993) 93. C. Tracy, J. Gao, Appl. Phys. Lett. 87, 143502 (2005) 94. C. Tracy, J. Gao, J. Appl. Phys. 100, 104503 (2006) 95. W. Bonnet, C. Tracy, G. Wantz, G.J. Liu, J. Gao, Small 4, 1707–1710 (2008) 96. S.L. Chen, G. Wantz, L. Bouffier, J. Gao, Chem. Electron. Chem. 3, 392–398 (2016)

Chapter 3

Optical Engineering of Light-Emitting Electrochemical Cells Including Microcavity Effect and Outcoupling Extraction Technologies Hai-Ching Su

Abstract Light-emitting electrochemical cells (LEC) have received much scientific interest after the first demonstration in 1995. In addition to enormous progresses in development of novel emissive materials for LECs, charge carrier balance has been improved to significantly enhance device efficiency of LECs. However, further improvement in device performance is still required to meet industrial applications. Optical engineering techniques involving interference, scattering, and waveguiding effects in device optical structures would be feasible approaches to modify the electroluminescence (EL) spectrum and to extract more light output from LECs. In this chapter, the microcavity effect and outcoupling extraction technologies are reviewed in detail. This complements the introduction provided in Chap. 1. Microcavity effect can be employed to tailor the EL spectrum of LECs by adjusting the interference effect. Scattering from microlens array increases light outcoupling from substrates of LECs. Furthermore, color conversion and waveguide coupling result in more light extraction, since both EL trapped in waveguide and substrate mode would be harvested. With these promising techniques, doubled device efficiency can be realized. These results confirm that optical engineering techniques are necessary to achieve highly efficient LECs. Keywords Light outcoupling layer


Waveguiding
Microcavity
Color conversion

H.-C. Su (&) Institute of Lighting and Energy Photonics, National Chiao Tung University, Tainan 71150, Taiwan e-mail: [email protected] © Springer International Publishing AG 2017 R.D. Costa (ed.), Light-Emitting Electrochemical Cells, DOI 10.1007/978-3-319-58613-7_3

77

78

3.1 3.1.1

H.-C. Su

Introduction Microcavity Effect in Organic Thin-Film Devices

As explained in Chap. 1, LEC architecture involves a single active layer of around 100–200 nm sandwiched in between two electrodes. Since the active layer thickness is generally similar to the visible light wavelength and a high-reflectance metal is commonly employed as the cathode (e.g., Al, Ag, etc.), the wavelength dependent emission characteristics of the active layer are adjusted in a microcavity structure due to optical interference. This modifies the density of optical mode and changes the electroluminescence (EL) spectrum spectrally. Output spectrum from a bottom emitting thin-film light-emitting device would be simulated by employing the equation shown in the following [1]:  pffiffiffiffiffi  i 1 þ R1 þ 2 R1 cos 4pz k þ u1 i¼1   jEint ðkÞj2 ; pffiffiffiffiffiffiffiffiffiffi  jEext ðkÞj2 ¼ þ u þ u 1 þ R1 R2  2 R1 R2 cos 4pL 1 2 k T2 N1

N  P

ð3:1Þ

where R1 and R2 are the reflectances from cathode and substrate, respectively, u1 and u2 are the phase changes of reflection from cathode and substrate, respectively, T2 is the transmittance from the substrate, L is the total optical thickness of the active layers, jEint ðkÞj2 is the emission spectrum without microcavity effect, jEext ðkÞj2 is the emission spectrum coming out from the substrate, zi is the optical distance between cathode and the emitting sublayer i. The emitting zone is partitioned into N sublayers and all contributions are integrated. The photoluminescence (PL) spectrum of the emissive material on a quartz substrate is used as the emission spectrum without microcavity effect because no high-reflectance metal is present. It is noted that the optical structure of the device, e.g, active layer thickness and electrode structure, would affect the output EL significantly. Furthermore, the recombination zone of LECs is moving while the doped layers are extending, resulting in temporal evolution in EL spectrum [2]. Hence, the EL spectrum of LECs is modified by tuning the active layer thickness and/or the recombination zone position.

3.1.2

Optical Modes in Organic Thin-Film Devices

In a typical thin-film organic-based device, most of the EL in the active layer is trapped in indium tin oxide (ITO) and organic layer because of a large discrepancy in the refractive indices between the high-index layers (norg.  1.7 and nITO  1.9) and the low-index glass substrate (nglass  1.5). The percentage of EL in waveguide mode is 40–60% of the total intensity [3]. Another emission part is trapped in the

3 Optical Engineering of Light-Emitting Electrochemical Cells … Fig. 3.1 Schematic diagram of the optical modes in an organic thin-film device

79

External mode

Substrate mode

Substrate ITO anode

Waveguide mode

Organic layers Metal cathode

glass substrate because of total internal reflection at the output end. Eventually, only 20% of total intensity can be harvested in the forward direction (external mode). Schematic diagram of the optical modes in an organic thin-film device is shown in Fig. 3.1. Hence, recycling the EL in waveguide and substrate mode is a possible way to enhance device performance of LECs.

3.1.3

Organization of This Chapter

In the following, the state of the art on optical engineering of LECs including microcavity effect and outcoupling extraction technologies are reviewed. The microcavity effect can be used to tailor the output EL of LECs. Blue-green emission form white LECs is suppressed to improve the color purity. White EL can be generated by constructive interference from LECs based on a single blue-emitting material. In addition, near-infrared EL can also be obtained by constructive interference from LECs based on deep-red-emitting material. Such technique enhances obtainable EL spectral range from a limited number of available emissive materials for LECs. Outcoupling extraction technologies are essential to increase device efficiency of LECs. Recycling the confined EL by employing red color conversion layers and waveguide coupling are mentioned. Finally, conclusion and outlook are discussed.

3.2 3.2.1

Tailoring Output EL Spectrum of LECs by Employing Microcavity Effect Suppression of Blue-Green Emission to Achieve Purer White EL

Modifying EL spectra by adjusting microcavity effect is useful for white LECs based on ionic transition-metal complexes (iTMCs) to achieve white EL with

80

H.-C. Su

Commission Internationale de l′Eclairage (CIE) coordinates near (0.33, 0.33), because efficient saturated blue-emitting iTMCs were scarce. Published white LECs containing sky-blue-emitting iTMCs commonly show greenish white emission even mixed with deep-red-emitting iTMCs [4–7]. As shown in the work published by Su et al., [8], reducing green emission part of sky-blue-emitting iTMCs by destructive interference is a possible approach to reach saturated blue EL and purer white EL emission was obtained consequently. The blue-green-emitting iTMC [Ir(dfppz)2(dtb-bpy)][PF6] (1) (where dfppz is 1-(2,4-difluorophenyl)pyrazole and dtb-bpy is [4,4′-di(tert-butyl)-2,2′-bipyridine]) proposed by Tamayo et al. was employed as the host complex [9]. Efficient red laser dye Sulforhodamine 101 (2), which was shown to have a high photoluminescence quantum yield (PLQY) of 0.95 ± 0.02, [10] was used as the guest dopant. The calculated and measured output EL spectra of the white LEC based on an active layer (320 nm) doped with 0.3wt% 2 are shown in the left part of Fig. 3.2. When the emission zone of the white LECs was placed at 195 nm away from cathode, the calculated and measured output EL spectra were well fitted. This emitting zone position was rational because a larger energy level offset in the highest occupied molecular orbital (HOMO) level between host and guest (Fig. 3.3) lead to more pronounced hole trapping and the recombination zone was thus nearer the anode. Reduced bandwidth of the blue EL was attributed to destructive interference of the green emission from complex 1. Saturated blue EL lead to purer white emission with CIE coordinates of (0.32, 0.30), which was close to equal energy point (0.33, 0.33). When the active layer thickness was changed, the spectral region, at which destructive interference occurred was different and the output EL emission was modified. As shown in the right part of Fig. 3.2, destructive interference took place at the red spectral region when the emissive layer thickness was 270 nm. Calculation also predicted well-fitted spectrum with the experimental data (right part of Fig. 3.2). These results showed that microcavity effect can be employed to modify the output EL emission and purer white EL can be obtained without using deep-blue emitting materials.

1.0

Measured Simulated

EL Intensity (a.u.)

EL Intensity (a.u.)

1.0 0.8 0.6 0.4 0.2 0.0 400

500

600

Wavelength (nm)

700

Measured Simulated

0.8 0.6 0.4 0.2 0.0 400

500

600

700

Wavelength (nm)

Fig. 3.2 Calculated and measured EL spectra from the white LECs based on emissive layer thicknesses of 320 nm (left) and 270 nm (right)

3 Optical Engineering of Light-Emitting Electrochemical Cells …

81

Fig. 3.3 Energy level diagram of the host (1) and the guest (2) molecules

3.2.2

Non-doped White LECs Based on a Single Emissive Material

White LECs have recently attracted much attention, since they are potential in solid-state lighting applications. White iTMC LECs were commonly composed of host-guest active layers to reach better device performance. Nevertheless, low doping concentrations (450 nm) emissive layers of ruthenium(II) complexes [2, 14, 15] and the time-dependent EL spectrum resulted from modified microcavity effect due to recombination zone moving [2]. To clarify the mechanism of the time-dependent EL spectrum from the LECs containing complex 1, device physics responsible for spectral changing are explained in the following. The recombination zone was initially at the central emissive layer (zi = 300–230 nm at 15–57 min, the left part of Fig. 3.4). Blue-green constructive interference occurred and thus the EL showed blue shift and narrowed FWHM. Then the recombination zone shifted towards the cathode

H.-C. Su

0.8

t (min) Measured z (nm) 15 300 30 280 57 230

0.6 0.4 0.2 0.0 400

500

600

700

Wavelength (nm)

800

1.0 0.8

t (min) Measured z (nm) 66 206 75 196 78 193 81 190

EL Intensity (a.u.)

1.0

EL Intensity (a.u.)

EL Intensity (a.u.)

82

0.6 0.4 0.2 0.0 400

500

600

700

Wavelength (nm)

800

t (min) Measured z (nm) 90 170 150 105 174 120

1.0 0.8 0.6 0.4 0.2 0.0 400

500

600

700

800

Wavelength (nm)

Fig. 3.4 Measured and simulated temporal EL spectra from the white LECs containing 1 at 15–57 (left), 66–81 (central) and 90–174 min (right) under 3.7 V. The temporal recombination zone position measured from cathode was extracted by calculations and is labeled for comparison

Fig. 3.5 The energy level diagram of 1 and the work functions of electrodes

with time (zi = 206–190 nm at 66–81 min, the central part of Fig. 3.4). The blue-green emission showed more blue-shifted and even narrower FWHM. In addition, another constructive interference took place at red and dual-emission white emission was obtained. When the recombination zone was further moving to the cathode (zi = 170–120 nm at 90–174 min, the right part of Fig. 3.4), the red emission reduced and finally disappeared. After ca. 3 h, the EL was reaching a steady state, which revealed that stabilized recombination zone was present because of well formed doped layers. Since the time-dependent EL spectrum was matched by the simulation from microcavity effect when the recombination zone was moving, EL spectral migration resulted from material degradation may be ruled out. Time-dependent recombination zone was rationalized by the energy level for the LECs containing complex 1 (Fig. 3.5). When a bias voltage was applied, the recombination zone was at the central active layer in spite of a lower hole injection barrier (0.85 eV) than electron injection barrier (1.29 eV). It may be related to higher electron mobility in a thicker film (490 nm) of complex 1 at 3.7 V. Due to the difference in hole and electron injection barrier, the necessary number of ions at the anode to reach ohmic contact for hole is much lesser than that for electron at the cathode. The temporal increasing rate for hole injection efficiency is, thus, higher than that for electrons, leading to a recombination zone moving from the center of the emissive layer to the cathode. After the doped layers were completely established, the p-i-n layer was steady and the recombination zone froze. These results showed that recombination zone moving lead to red constructive interference, which combined with blue-green emission from complex 1 to reach white emission. It demonstrated a simple approach to realize white EL from non-doped LECs.

3 Optical Engineering of Light-Emitting Electrochemical Cells …

3.2.3

83

Non-doped Near-Infrared LECs Based on Interferometric Spectral Tailoring

Near-infrared (NIR) LECs show potential in NIR light sources because of simple device architecture, compatibility with large-area solution processes, low power consumption, and high device efficiency. Nevertheless, host-guest NIR LECs generally exhibit enormously increased residual host emission when enhancing NIR light output by raising bias [16]. More NIR power was only harvested at the cost of spectral purity. To enhance light output from NIR LECs without losing spectral purity, a new way to obtain NIR emission from non-doped deep-red-emitting LECs by tuning active layer thickness to adjust microcavity effect has been reported [17]. NIR emission from non-doped deep-red-emitting LECs was achieved by tuning emissive layer thickness to move the constructive interference peak to the NIR spectrum. NIR emission coming from enhancement of microcavity effect was insensitive to bias. Hence, without sacrificing spectral purity, 20X NIR output was achieved when compared to reported value from host-guest NIR LECs [16]. NIR emission from non-doped deep-red-emitting LECs was achieved by employing the deep-red-emitting 3 as the emissive material. Complex 3 was [Ru (dtb-bpy)3][PF6]2 (where dtb-bpy is 4,4′-ditertbutyl-2,2′-bipyridine) [18]. The thickness of emissive layer was 605 nm. As shown on Fig. 3.6, LECs based on thick emissive layer finally showed a main NIR emission and a weaker red emission. However, the EL spectrum of relatively thinner device (200 nm) showed a stable peak at ca. 660 nm. Spectral shift in constructive interference because of recombination zone moving showed little effect on spectral changing. Spacing between two constructive interference peaks in thinner reference LECs was much larger. Constructive interference may take place outside the emission spectrum of 3, rendering stable EL. It showed that the output EL emission from non-doped deep-red-emitting LECs can be modified by tuning the emissive layer thickness to adjust microcavity effect.

t (min) 15 42 75 96

1.0

EL Intensity (a.u.)

Fig. 3.6 Temporal EL spectra of LECs with thick emissive layer (605 nm) of 3

0.8 0.6 0.4 0.2 0.0 500

600

700

800

Wavelength (nm)

900

84

H.-C. Su

To clarify the mechanism of the temporal EL from thicker LECs, the proposed technique shown in Sect. 3.1.1 was employed to probe moving of recombination zone. Time-dependent recombination zone for thick LECs based on 3 under 3 V is shown in Fig. 3.7. The ruthenium(II) complexes with bipyridyl ligands were reported to exhibit higher electron mobility than hole mobility [19]. Hence, when the doped layers have not yet well established to facilitate carrier injection, the initial recombination zone was nearer anode because of imbalanced hole and electron carrier mobilities. Subsequent recombination zone moving was rationalized by the energy levels depicted in Fig. 3.8. Since injection barrier for hole (0.42 eV) was lower than that for electron (1.03 eV), the necessary number of ions near anode to reach ohmic contact for hole was lesser than that for electron at the cathode. Enhancing rate for hole injection was, thus, higher than that for electron injection, leading to recombination zone shifting from anode toward the central active layer (Fig. 3.7). After the doped layers were well established, carrier injection efficiency reached a steady state and the recombination zone froze. It further revealed that temporal EL was attributed to modified microcavity effect due to recombination zone moving. Voltage dependent EL spectrum was the main disadvantage of NIR host-guest LECs. As bias voltage enhanced to obtain higher light output, residual host emission was obvious, deteriorating the spectral purity. Low NIR power 490 nm, the actual DnITOSpacing layer was even smaller (Fig. 3.14) and thus almost whole EL was under cutoff condition. Because waveguide coupling stopped, some emission was still confined in ITO layer and spacing layer, resulting in deteriorated device efficiency. It indicated that the refractive index of the spacing layer have to be carefully adjusted to ensure waveguiding of ITO layer and thus to optimize the outcoupling efficiency.

3.4

Conclusion and Outlook

Tremendous progresses have been made in device performance of LECs by developing novel emissive materials and by improving carrier balance. To further enhance device efficiency and to adjust optical properties of LECs, optical engineering was a feasible approach to extract more confined EL in layered device structure. In this chapter, several reported optical engineering techniques have been reviewed. The microcavity effect can be employed to tailor the EL spectrum of LECs due to altered interference effect in different device optical structures. Reduced bandwidth of the sky-blue EL resulted from destructive interference at green can be achieved and hence purer white EL can be obtained without using deep-blue-emitting materials. In addition, recombination zone moving in LECs leads to constructive interference at red. When combined with blue-green EL from LECs, white emission was realized in non-doped LECs based on a single emissive material. Based on a similar concept, non-doped NIR LECs containing a

3 Optical Engineering of Light-Emitting Electrochemical Cells …

91

red-emitting complex were reported to exhibit 20-fold enhancement in NIR light output as compared to conventional host-guest NIR LECs. To further improve light outcoupling efficiency, microlens array was utilized on the output end of glass substrate in LECs and 1.6-fold enhancement in outcoupled EL was achieved. Recycling the trapped EL by employing red CCLs placed under ITO layer would be another effective approach. The blue emission confined in the substrate mode and in the waveguide mode can be transferred into red PL in the CCL and was partially redirected into the external mode. With large NPs (250 nm) to scatter the emission in the CCL, peak EQE (power efficiency) of white LECs reached 20% (40 lmW−1). Such results were almost doubled values when compared to white LECs without scattering red CCLs. Waveguide coupling would also be an effective approach to recycle trapped light in the ITO layer. Two TPR layers were placed under ITO layer. By tuning the effective refractive index of the upper TPR layer, efficient waveguide coupling between ITO layer and the lower TPR layer was realized. Because the lower TPR layer was doped with scattering particles, emission coupled from ITO layer was scattered and transferred into the external mode. Doubled device efficiency can also be achieved by utilizing waveguide coupling. In spite of enormous enhancement in device performance achieved by using optical engineering techniques mentioned above, further improvement will still be required to meet practical applications. Fabricating protrusion structures on and/or below ITO layer would destroy waveguiding of light at the ITO layer, rendering increased light outcoupling efficiency. Incorporating metal NPs in LECs would be another way to enhance EL efficiency of LECs. Emission of excitons can be significantly enhanced by localized surface plasmon resonance on metal NPs. In addition, overall consideration of interference effect in optical structure of LECs, e.g., optimizing the thickness of ITO, poly(3,4-ethylenedioxythiophene):poly (styrene sulfonate) (PEDOT:PSS), and active layer, would be beneficial in maximizing light outcoupling efficiency. With these feasible techniques, further improvements in device efficiency will be expected. Acknowledgements The author acknowledges the financial support from Ministry of Science and Technology (MOST 105-2221-E-009-097-MY2).

References 1. 2. 3. 4.

X. Liu, D. Poitras, Y. Tao, C. Py, J. Vac. Sci. Technol. 22, 764 (2004) T.W. Wang, H.C. Su, Org. Electron. 14, 2269 (2013) A. Chutinan, K. Ishihara, T. Asano, M. Fujita, S. Noda, Org. Electron. 6, 3 (2005) H.C. Su, H.F. Chen, F.C. Fang, C.C. Liu, C.C. Wu, K.T. Wong, Y.H. Liu, S.M. Peng, J. Am. Chem. Soc. 130, 3413 (2008) 5. L. He, J. Qiao, L. Duan, G. Dong, D. Zhang, L. Wang, Y. Qiu, Adv. Funct. Mater. 19, 2950 (2009) 6. L. He, L. Duan, J. Qiao, G. Dong, L. Wang, Y. Qui, Chem. Mater. 22, 3535 (2010) 7. H.C. Su, H.F. Chen, Y.C. Shen, C.T. Liao, K.T. Wong, J. Mater. Chem. 21, 9653 (2011)

92

H.-C. Su

8. H.C. Su, H.F. Chen, P.H. Chen, S.W. Lin, C.T. Liao, K.T. Wong, J. Mater. Chem. 22, 22998 (2012) 9. A.B. Tamayo, S. Garon, T. Sajoto, P.I. Djurovich, I.M. Tsyba, R. Bau, M.E. Thompson, Inorg. Chem. 44, 8723 (2005) 10. R.A. Velapoldi, H.H. Tønnesen, J. Fluoresc. 14, 465 (2004) 11. G.R. Lin, H.F. Chen, H.C. Shih, J.H. Hsu, Y. Chang, C.H. Chiu, C.Y. Cheng, Y.S. Yeh, H.C. Su, K.T. Wong, Phys. Chem. Chem. Phys. 17, 6956 (2015) 12. C.T. Liao, H.F. Chen, H.C. Su, K.T. Wong, J. Mater. Chem. 21, 17855 (2011) 13. C.T. Liao, H.F. Chen, H.C. Su, K.T. Wong, Phys. Chem. Chem. Phys. 14, 9774 (2012) 14. J.S. Lu, J.C. Kuo, H.C. Su, Org. Electron. 14, 3379 (2013) 15. C.L. Lee, C.Y. Cheng, H.C. Su, Org. Electron. 15, 711 (2014) 16. C.C. Ho, H.F. Chen, Y.C. Ho, C.T. Liao, H.C. Su, K.T. Wong, Phys. Chem. Chem. Phys. 13, 17729 (2011) 17. J.H. Hsu, H.C. Su, Phys. Chem. Chem. Phys. 18, 5034 (2016) 18. S. Bernhard, J.A. Barron, P.L. Houston, J.L. Ruglovksy, X. Gao, G.G. Malliaras, J. Am. Chem. Soc. 124, 13624 (2002) 19. W.K. Chan, P.K. Ng, X. Gong, S. Hou, Appl. Phys. Lett. 75, 3920 (1999) 20. N. Kaihovirta, C. Larsen, L. Edman, A.C.S. Appl, Mater. Interfaces. 6, 2940 (2014) 21. J.S. Lu, H.F. Chen, J.C. Kuo, R. Sun, C.Y. Cheng, Y.S. Yeh, H.C. Su, K.T. Wong, J. Mater. Chem. C 3, 2802 (2015) 22. C.C. Liu, S.H. Liu, K.C. Tien, M.H. Hsu, H.W. Chang, C.K. Chang, C.J. Yang, C.C. Wu, Appl. Phys. Lett. 94, 103302 (2009) 23. C.Y. Cheng, C.W. Wang, J.R. Cheng, H.F. Chen, Y.S. Yeh, H.C. Su, C.H. Chang, K.T. Wong, J. Mater. Chem. C 3, 5665 (2015) 24. R.G. Hunsperger, Integrated Optics: Theory and Technology (Springer, New York, 2009)

Chapter 4

The Use of Additives in Ionic Transition Metal Complex Light-Emitting Electrochemical Cells Lyndon D. Bastatas and Jason D. Slinker

Abstract In this chapter, we will describe the use of additives within films of ionic transition metal complexes (iTMC) applied in light-emitting electrochemical cells (LEC). Here, iTMCs generally act as charge transporters for electrons and holes, electrolytes for ion motion and emissive materials. The additives primarily fall into three categories: polymer additives, small molecule host/guest systems, and salt/electrolyte additives. Each class has emerged to optimize specific elements of the device operation. Keywords Ionic transition metal complex Ionic liquids Electrical double layer




4.1


Lithium salts
Host guest systems


Polymer Additives to Decrease Self-quenching for Improved Efficiency

In light-emitting electrochemical cells (LEC), electrons and holes injected from the contacts traverse through the bulk and recombine to form excitons—i.e. bound electron-hole pairs. Ideally, these excitons will decay radiatively, and a chief benefit of LECs based on ionic transition metal complexes (iTMC) is their high luminance efficiency, which can approach 100% (see Chaps. 7, 8 and 11 for details about this type of compounds) [1]. One fundamental issue is the nonradiative self-quenching of excitons due to the presence of excimers and/or exciplexes in solid state [2, 3]. That is, when two or more excited-state molecules are in close proximity to each other, luminescence can be lost as the energy is dissipated between the molecules themselves through a variety of nonradiative processes. This is an inherent problem in pristine iTMC devices, where the metal complexes function as mixed conductors— i.e. facilitating transport of electrons and holes, as well as serving as an electrolyte L.D. Bastatas
J.D. Slinker (&) The University of Texas at Dallas, 800 W. Campbell Rd., Richardson, TX 75080, USA e-mail: [email protected] © Springer International Publishing AG 2017 R.D. Costa (ed.), Light-Emitting Electrochemical Cells, DOI 10.1007/978-3-319-58613-7_4

93

94

L.D. Bastatas and J.D. Slinker

for counterions in the film [4, 5]. Compounding this fact, the operational mechanism of these materials dictates that recombination of electrons and holes occurs within a thin region between the electrodes, and thus excitons are confined to a small space [6, 7]. These combined effects can lead to electroluminescence efficiencies in devices that are lower than anticipated from the photoluminescence quantum yields (/) of the materials found in solution. To counteract these effects, efforts have been focused on spacing the metal complexes apart in various ways. Intrinsically, this has been accomplished through the addition of bulky ligands on the metal complexes themselves [8–10]. Such approaches have led to some of the most efficient LECs produced to date. However, there is a limit to this approach, as very large ligands can decrease both the electronic and ionic conductivity of the films [11]. Alternatively, iTMC spacing can be increased through additives, and early efforts in iTMC films utilized inert polymers to accomplish this. Prior photoluminescence studies had shown that inert polymers, such as poly (methyl methacrylate) (PMMA), can enhance the / of solid state films of ITMCs [12]. This idea was then incorporated into some of the first iTMC devices with important implications for device efficiency and lifetime.

4.1.1

Layer-by-Layer Techniques

To control complex–complex spacing with inert polymer films, the Rubner group first employed layer-by-layer polymer assembly techniques to controllably deposit the electroluminescent layer of ruthenium(II) iTMC-based LECs [13–15]. This involved the alternating deposition of acidic, basic, or neutral layers (Fig. 4.1) which was accomplished either by hand or by automated dipping machines. Such assembly approaches are similar to Langmuir–Blodgett film deposition, and the approach enabled precise control over the composition and thickness of the materials. In the first embodiments of this approach, ruthenium(II) complexes were introduced into devices as pendant groups on polyester polymers, and electrically inert poly(acrylic acid) (PAA) layers were incorporated. By changing the pH of the ruthenium(II) polyester and PAA solutions during the deposition process, both the overall thickness of the devices and the relative ratio of the ruthenium(II) unit to the PAA spacer was varied. The variation of the composition allowed the systematic modification of the site-to-site distance of the ruthenium(II) complexes, which influences self-quenching effects and electronic mobility in the devices. By optimizing thickness and composition, external quantum efficiencies (EQE)—i.e. the emitted photons per injected electron; please refer to Chap. 1 for more details—as high as 3.2% were realized, [15] a breakthrough at the time that remains highly respectable for ruthenium(II) iTMCs. By simply changing the thickness of these films at the optimal composition, the EQE varied from 0.2% at 90 nm to 2% at 200 nm active layer thickness. The turn-on voltage was found to be a function of the number of polyester/PAA bilayers present, with 3 and 5 V turn-on voltages for 15 and 35 bilayer devices, respectively. This was attributed to the high bulk

4 The Use of Additives in Ionic Transition Metal Complex …

95

Fig. 4.1 Structure of an emissive polymer use to make layer-by-layer polymer blended LECs

resistance of the multilayer films. It was shown that the order of voltage sweeping, whether from forward bias to reverse or reverse bias to forward, influenced the symmetry of the current versus voltage and radiance versus voltage characteristics. Furthermore, by changing the composition of the multilayers, devices that exhibited light emission preferentially in either the forward or reverse bias operation were fabricated. These first efforts focused primarily on polymerizable ruthenium(II) compounds, but the Rubner group subsequently showed that a more simplified approach could be followed.

4.1.2

Blended Inert Polymers

iTMCs may be directly mixed with electrically inert polymers to achieve the benefits afforded by a decrease of self-quenching. Rudmann and Rubner first explored this idea with ruthenium(II) complexes mixed into inert polymers such as poly(methylmethacrylate) (PMMA), poly(carbonate) (PC), and poly(styrene) (PS), discovering great improvements in lifetime [16]. Inert polymers were introduced into devices by combining stock solutions of the ruthenium(II) complex and the inert polymer and spin coating. So this approach maintains the ease of solution processing. Polymer blending, when coupled with pulsed voltage driving, increased the lifetimes of the LECs considerably. As seen in Fig. 4.2, the extrapolated lifetimes of single layer devices were extended for some of the polymer blends. The device half-life, the time to decay to half of the radiant flux maximum—see Chap. 1 for more details, reached an excess of 1,000 h for polycarbonate blends. Likewise, the EQE was improved to as high as 2.5% for polymer blended devices, postulated to possibly come from a reduction of leakage current with the incorporation of the inert polymer. Additional efforts by Rudmann et al. with polymer blending led to even higher efficiencies [17]. Ruthenium(II) complexes were modified with ligands that promoted additional spacing between complexes, as shown in Fig. 4.3. This led to EQEs reaching as high as 5.5%, the best among iTMC devices to that time

96

L.D. Bastatas and J.D. Slinker

Fig. 4.2 Light emission versus time of pristine ruthenium(II) tris-bipyridine LECs and those blended with polystyrene (PS), polycarbonate (PC), and polymethylmethacrylate (PMMA). Solid lines show extrapolations to the half-life of the emission. Reprinted from [16] with the permission of AIP Publishing

Fig. 4.3 Structures of the small molecule complexes and side profile illustration of a blended light emitted electrochemical cell. Reprinted with permission from [17]. Copyright 2016 American Chemical Society

(Fig. 4.4). Furthermore, through the use of small counterions coupled with the complex, turn-on times of less than one second could be achieved. However, these devices showed relatively low luminance, typically 10–30 candelas per metre squared (cd/m2), below the benchmarks typically desired for display (>100 cd/m2) and lighting applications (>3,000 cd/m2). Later, the group of Bolink utilized the polymer blending strategy to achieve devices of high brightness and relatively long-lasting emission. In particular, a tris (4,7-diphyenyl-1,10-phenanthroline) ruthenium(II) complex was blended with PMMA to yield a device with appreciable luminance of nearly 400 cd/m2 at a power efficiency of 1.9 lumens per watt (lm/W) and a lifetime of approximately 100 h [18]. The increased performance over more simplified ruthenium(II) complexes was attributed to the hydrophobic nature of the ligands, which would tend to

4 The Use of Additives in Ionic Transition Metal Complex …

97

Fig. 4.4 Luminance and EQE versus time of a LEC from a ruthenium(II) tris-bipyridine complex blended with polymethylmethacrylate (PMMA). Reprinted with permission from [17]. Copyright 2016 American Chemical Society

suppress water-based degradation reactions that had been shown to plague these complexes (see Chaps. 7 and 8 for more details) [19–21]. In conclusion, the polymer blending strategy is one way to circumvent self-quenching in iTMCs-based LECs for improved EQE and lifetime. However, the increased resistance introduced by the inert polymer limits the overall brightness of these devices and also negatively impacts the power efficiency. Thus, there is benefit to retaining small molecules throughout the LEC active layer to maintain high conductivity and brightness.

4.2

Host/Guest LECs to Control Color and Improve Efficiency

As noted above, self-quenching of iTMC-based LECs is an inherent concern, particularly for pristine films of unblended iTMCs [22]. Blending the emitter with a polymer decreases this quenching and improves efficiency, but at a cost to the overall device resistance, such that absolute luminance and power efficiency both suffer. This is due to the fact that these emitters are multifunctional chromophores that also transport electrons and holes through hopping and diffusion processes, and increasing the spacing between these molecules greatly frustrates this transport. Conducting polymers can be used to circumvent this problem, but this ultimately changes the class of devices and falls out of the scope of this chapter (see Chaps. 1 and 10 for details about this aspect). Also, this approach often changes the operational mechanism from that of a LEC to a more conventional light-emitting diode. Notably, the unique operational mechanism of LECs has enabled their use in beneficial architectures such as functional cascaded panels [23, 24] and electroluminescent nanofibers [25, 26].

98

L.D. Bastatas and J.D. Slinker

Alternatively, to overcome the challenges of self-quenching without a cost to conductivity, while maintaining a LEC operational mechanism, two or more iTMCs may be combined in a host–guest approach. In the case of two materials, one may primarily function for charge transport, and the other, typically of lower bandgap, is utilized as an emitter. Alternatively, separate conduction of electrons and holes may be facilitated by the distinctive materials given the details of the energy levels. This has been particularly utilized in blends of iTMC-based LECs. The fabrication of host–guest iTMC-based LECs is facilitated by the solution processability of these materials—i.e. both components can be co-dissolved and deposited from the same solvent. Separating charge transport and emission processes in LECs circumvents the self-quenching that decreases the emission yield in organics. The first instance of the host–guest approach in an iTMC device was from Hoisseini et al., who doped an osmium(II) complex into a ruthenium(II) complex to achieve an improved efficiency of a technologically beneficial color [27]. [Os (phen)3][PF6]2, where phen is 1,10-phenanthroline, emits a deep red color but exhibits a low /. It was doped within a [Ru(bpy)3][PF6]2 matrix, an orange emitter with a higher /. In Fig. 4.5, the photoluminescence spectra of pristine [Os(phen)3] [PF6]2 and [Ru(bpy)3][PF6]2 films are shown along with those from small percentages of [Os(phen)3][PF6]2 within [Ru(bpy)3][PF6]2. Energy transfer from the ruthenium(II) complex to the osmium(II) complex was facilitated by the overlap of the ruthenium(II) complex emission with the osmium(II) complex absorption. The normalized electroluminescence of various blends of these materials at 3 V operation is also shown in Fig. 4.5, and clearly the electroluminescence spectra are Fig. 4.5 Top Normalized photoluminescence spectra of doped and pristine films under 470 nm excitation. Bottom Normalized electroluminescence spectra from devices made with doped and pristine films at 3 V. (Inset) Bias dependence of the emission spectrum for the 1% doped film. Reprinted with permission from [27]. Copyright 2016 American Chemical Society

4 The Use of Additives in Ionic Transition Metal Complex …

99

nearly identical to the photoluminescence spectra. By doping 5% of guest to emitter weight (w/w), nearly all of the emission originates with the osmium(II) complex guest. In the inset, it is shown that increasing the bias increases the emission component from the ruthenium(II) complex, likely due to saturation of the osmium (II) dopant states from exponentially increased carrier injection. Overall, this work showed that it is possible to tune emission color with small amounts of one iTMC within another. In Fig. 4.6, the EQE of the ruthenium(II) complex host, osmium(II) complex guest, and guest 5% w/w blended device is shown. The efficiency of the osmium(II) complex device is essentially an order of magnitude below that of the ruthenium(II) complex device. However, blending the small amount of the guest in the host produced a device that is more efficient than either individual film, at a respectable luminance of 220 cd/m2. The improvement apparently arises from the decreased self-quenching of the osmium(II) complex. Thus, in addition to color tuning, this host/guest approach can improve device efficiency. The host/guest strategy was later applied to iridium(III) complexes, with impressive results. Su and coworkers extended this by doping the green-emitting [Ir (dFppy)2(SB)][PF6] host complex with the orange-emitting [Ir(ppy)2(SB)][PF6] guest complex, where ppy is 2-phenylpyridyl, dFppy is a fluorine substituted ppy, and SB is a spirobifluorene ligand [28]. Previously, these complexes were studied for their high efficiency and color tuning [10]. The highest / was obtained for films with guest concentrations of 25% w/w, at which concentration emission was found to almost completely occur from the guest. Before blending these complexes, devices based separately on the host or guest each showed a respectable EQE of 7.1% [10]. Upon blending, an impressive peak EQE of 10.4% and power efficiency of 36.8 lm/W were obtained [28]. As in the previous case, the improvement in

Fig. 4.6 Temporal evolution of the EQE of pristine osmium(II) complexes and ruthenium(II) complexes devices, as well as a guest 5% w/w blended device. (Inset) dependence of peak efficiency on dopant concentration. Reprinted with permission from [27]. Copyright 2016 American Chemical Society

100

L.D. Bastatas and J.D. Slinker

efficiency over either pristine film signified a decrease in self-quenching of emission upon doping in a matrix. This group also utilized the doping strategy to achieve white light emission between two iridium(III) complexes, as seen in Fig. 4.7 [29]. In this instance, the group combined a blue-green-emitting host with a red-emitting guest. Photoluminescence studies of blended films revealed that emission was roughly balanced between host and guest at a guest concentration of 0.4wt%. Blends of these emitters in single layer LECs showed that the relative emission could be tuned based on the choice of voltage, as previously noted by Hoisseini et al. At low voltages (2.7 V and lower), emission primarily occurred from the red guest, whereas at higher voltages emission became more balanced as presumably the guest emission states became saturated. Power efficiencies as high as 7.8 lm/W were demonstrated at a luminance of approximately 10 cd/m2. Color rendering indices—CRIs; see Chap. 1 for more details—of 80–81 were obtained over the range of operating voltages of 2.7–3.3 V. This group went on to use the host guest strategy to produce devices with

Fig. 4.7 Top White light electroluminescence spectra from a host/guest LEC operating at various biases. Bottom Current density and voltage characteristics from a host/guest white light-emitting LEC. Reprinted with permission from [29]

4 The Use of Additives in Ionic Transition Metal Complex …

101

white electroluminescence having Commission Internationale de l’Eclairage—CIE; see Chap. 1 for more details—coordinates approaching (0.33, 0.33). Overall a promising strategy, the host/guest approach can produce high-efficient devices with color control. It remains to be seen if high luminance (>300 cd/m2) at high efficiency can be achieved with this approach.

4.3 4.3.1

Ionic Additives for Improved Ion Redistribution and LEC Performance Electric Double Layer Formation and Charge Injection

As carefully explained in Chap. 1, LEC operation is facilitated by the redistribution of ions in the film to form electrical double layers at the contacts [6, 30–38]. In particular, ions rearrange into layers, forming a dense (Helmholtz) and diffuse (Gouy-Chapman) layer at each electrode referred to as electric double layers (EDLs). This ionic rearrangement controls carrier injection, and as a result the device response ranges from seconds to as long as days, particularly for pristine small molecule LECs [4, 5, 39]. It follows that controlling the details of ionic redistribution and double layer formation is key to idealize LEC performance.

4.3.2

Electrolyte—Salt Combinations

Initial efforts to control ionic redistribution in iTMC-based LECs involved using traditional electrolyte systems in combination with salts to control performance. The first approach was carried out by adding poly(ethylene oxide) (PEO) and a salt, lithium triflate, to ruthenium(II) complexes, [40] similar to polymer-based LEC systems which utilize these components along with a conducting, electroluminescent polymer. This reduced the turn-on time (ton) of these LECs from 2 min to 30 s. However, these devices showed limited EQE, well below 1%. Potential reasons for this limitation could be increased electrical resistance due to the spacing of the complexes, as well as possible phase separation of the complexes and electrolyte. A second approach utilizing an explicit electrolyte additive was accomplished by blending a crown ether and a salt along with an emissive binuclear ruthenium(II) complex [41]. In this case, the pristine device did not show emission on the time scale of the experiment, but emitted light with ton as small as 2 min with an optimal electrolyte/salt combination. Again, these devices showed limited external quantum efficiencies of 0.02%. This approach, though potentially offering better phase compatibility than PEO between the emitter and electrolyte, could result in plasticization of the films at high ether contents, thereby limiting optimal conductivity

102

L.D. Bastatas and J.D. Slinker

of the film. As seen, at appropriate concentrations, both of these methods elicited faster response from the devices. However, the EQE remained low, though alternative emitters could have improved this figure-of-merit.

4.3.3

Ionic Liquids

Following efforts with electrolyte-salt blends, it was shown that salts could be beneficially added to iTMC-based LECs without the addition of an explicit electrolyte. In this arrangement, the ionic emitters also act as electrolytes to distribute the salt. One of the widely used classes of materials to increase the conductivity of the active layer of LECs is ionic liquids (ILs) [42]. ILs behave as ionically conductive molten salts at room temperature, primarily because the component ions are structurally uncoordinated. They have high conductivity appropriate for charge transport as well as a wide potential window for stability, both favourable for improving LECs. To understand how these materials improve LEC performance, consider that the ionic conductivity of fully dissociated ions is X rion ¼ ni l i qi ; ð4:1Þ i

where ni is the concentration, µi the mobility, and qi the charge of the ith ionic species. Ionic liquids can improve the ionic conductivity by increasing the concentration of mobile ions and by increasing the mobility of ions in the film. In turn, this can facilitate the redistribution of ions that contribute to beneficial double layer formation as noted above. Consequently, ILs tend to decrease the turn-on time, lower the applied voltage needed to achieve a given luminance, and increase the efficiencies of LECs.

4.3.3.1

Ionic Liquids Shown to Decrease Turn-on Time

Parker et al. first investigated the effect of ionic liquids on the radiant flux, turn-on time, and lifetime of LECs based on an iridium(III) complex [43]. In particular, they prepared devices based on [Ir(ppy)2(dtb-bpy)][PF6] (where ppy is 2-phenylpyridine and dtb-bpy is 4,4′-di-tert-butyl-2,2′-dipyridine) and the ionic liquid 1-butyl-3-methylimidazolium hexafluorophosphate [BMIM][PF6]. They found that as the concentration of ionic liquid was increased from 0–0.25% volume fraction (v/v), the time for the onset of light emission and the time to achieve maximum emission decreased significantly (Fig. 4.8b). This followed the trend anticipated, since the conductivity of the film increases with increasing IL concentration, as given by Eq. 4.1. Simultaneously, the intensity of light emission increased only slightly with concentration, indicating that these ions had little effect on improving

4 The Use of Additives in Ionic Transition Metal Complex …

103

Fig. 4.8 Left Radiant flux versus time of Iridium(III) iTMC-based LECs with various amounts of ionic liquid. Middle Time for onset of emission (ton) and time to radiant flux maximum (tmax) versus ionic liquid concentration in iridium(III) iTMC-based LECs. Right Lifetime and total emitted energy versus ionic liquid concentration in LECs with [Ir(ppy)2(dtb-bpy)][PF6]. Reprinted with permission from [43]. Copyright 2016 American Chemical Society

the balance of injected carriers. Unfortunately, device lifetime was also marginally lowered with increasing ionic liquid concentration (Fig. 4.8c). Nonetheless, this established a convenient, generalizable way to improve the response time of iTMC-based LECs that has been utilized in many works. Similar effects were subsequently observed by Slinker et al. [44] and Lowry et al. [45] in iridium(III) iTMC-based LECs of different colours.

4.3.3.2

High Ionic Conductivity Ionic Liquids Yield High Peak Luminance

Costa et al. evaluated the effect of additional imidazolium-based ionic liquids with differing intrinsic ionic conductivities in LECs from [Ir(ppy)2bpy][PF6] [46].

104

L.D. Bastatas and J.D. Slinker

Specifically, the ILs investigated were [BMIM][PF6], 1-ethyl-3-methylimidazolium hexafluorophosphate [EMIM][PF6], and 1-hexyl-3-methylimidazolium hexafluorophosphate [HMIM][PF6]. The intrinsic conductivity of each IL and the resulting device properties are provided in Table 4.1. Devices containing IL responded more than one order of magnitude faster than the device without IL. Moreover, studies also confirmed that the turn-on time of the devices decreased with the concentration of the IL added. In addition, they observed that there is a correlation between the luminance of the devices and their intrinsic conductivity, with the peak luminance of more conductive devices higher than less conductive devices. The effect of ILs on turn-on times and luminescence were also observed by Park et al. studying similar ILs. They compared ILs with different counterions, namely [BMIM][PF6] and [BMIM][BF4] in LECs with 1:1 emitter-to-IL molar ratio. [47] The conductivity of the luminous layer was 1.22, 1.58, and 1.66 nanosiemens per centimeter (nS/cm) for the pristine film, the film with [BMIM][PF6] and a [BMIM] [BF4] film, respectively. They observed that the turn-on time of the devices was dictated by the conductivity of the film. The peak luminance of the devices also increased with the addition of the ILs, and the device containing [BMIM][BF4] showed the highest peak. They also confirmed the trends of the effect of [EMIM] [PF6] and [BMIM][PF6] as noted above by Costa et al. with a different iridium(III) complex in a separate work [48]. Recently, Sun et al. observed decreased currents and marginal efficiency changes in IL-doped LECs based on iridium(III) complex, but these differences could arise from the high thickness of the devices (*600 nm) and relatively large IL concentrations used (see Chap. 5 for more details) [49]. Table 4.1 Conductivities of various ionic liquids and their effects on LECs based on iridium(III) complexes. Reprinted with permission from [46] IL

Intrinsic conductivity (mS/cm)

iTMC: IL

tdon (min)

te1/2 (h)

Luminance (cd/m2)

[EMIM+] [PF6−]

5.2a

4:1

24

81

497

4.6

6.7

[BMIM+] [PF6−]

1.5b

1:1 4:1

2.9 82

4.1 103

615 269

4.9 8.6

0.6 2.3

1:1 4:1

3.5 132

4.3 134

586 153

16.7 5.6

0.2 1.1

+

[HMIM ] [PF6−]

1.1

c

Power efficiency lm/W

Etotf (J)

1:1 22 5.5 302 14.1 0.1 without 1:0 690 668 219 6.1 6.6 a Value at 298.15 K from ref 23 b Value at 298.15 K from Ref 25 c Value at 295.1 K from Ref 24 d Time to reach 100 cd/m2 e Time to reach half of the maximum luminance f Total emitted energy, calculated by integrating the radiant flux curve from the application of current to the time for the radiant flux to decay to 1/5 of maximum. Measured from a 3 mm2 device

4 The Use of Additives in Ionic Transition Metal Complex …

4.3.3.3

105

Ionic Liquids in LECs Under AC and Pulsed Operation

Slinker et al. demonstrated that additives in iTMC-based LECs can enable devices with direct operation from a standard US outlet at 120 V and 60 Hz, [50] such as one seen in Fig. 4.9. To enable the millisecond response required by 60 Hz, they incorporated [BMIM][PF6] into a ruthenium(II) iTMC-based LEC to enhance the conductivity of the light-emitting layer. To accommodate the large applied bias, they utilized a cascaded device geometry, in which the cathode of one device served as the anode of an adjacent device in series. The rapid response afforded by the ionic liquid supported sufficient ionic redistribution for light emission to consistently appear on forward and reverse sweeps of the applied field. Tordera et al. showed that the use of ionic liquids with the application of pulsed current can produce an almost instantaneous (1 micron) interelectrode spacing works well. The active film in these devices should be meticulously patterned to capture the ion-blocking feature of electrodes [35, 38]. Common scanning techniques to probe the potential distribution in the active film of the device include electric force microscopy (EFM) and scanning Kelvin probe microscopy (SKPM). Along these lines, Lin et al. performed SKPM to investigate the effect of the Li [PF6] ionic additive on the dynamics of planar devices with channel lengths of *10 µm [38]. In particular, the impact of lithium salt additives on the double layer formation of these LECs was investigated in detail. Devices were prebiased at

4 The Use of Additives in Ionic Transition Metal Complex …

111

50 V to provoke a significant ionic redistribution, then tested by SKPM under 9 V applied bias to reveal differences in voltage profiles (Fig. 4.14). For the pristine device, the potential between the electrodes resembles a simple resistor. Upon closer inspection, the potential proceeding from the cathode follows e−jx dependence that extends throughout the bulk, consistent with a diffuse layer of cations. This demonstrated that the [Ir(ppy)2(bpy)]+ do not accumulate with high density at the cathode. Alternatively, devices with Li[PF6] showed a sharp linear drop at the cathode, consistent with a dense accumulation of Li+ and reflecting more idealized double layer formation. Notably, both devices showed similar dynamics at the anode, as both had strictly [PF6]− anions. The spatial electric field profiles for the same planar devices are shown in Fig. 4.15. Both devices show similar electric fields near the anode due to accumulation of [PF6]−. At the cathode there is a much stronger peak electric field in the ionic additive device due to the influence of densely packed Li+. Also indicative of ideal ionic redistribution, the electric field is repressed in the bulk region (12 to 16 lm) in the ionic additive device relative to the pristine device. In ideal LECs, ions accumulate at the contacts until the electric field is effectively cancelled out by

Fig. 4.14 Steady-state potential distribution of planar iTMC-based LECs at 9 V operation. SKPM data (voltage versus position) of pristine (upper graph, green) Au/[Ir(ppy)2(bpy)][PF6]/Au devices and devices with lithium salt additive (lower graph, red), Au/[Ir (ppy)2(bpy)][PF6] + 0.5wt% Li[PF6]/Au under a 9 V bias. Devices were also subjected to a 50 V prebias for 30 min. Dashed lines represent the approximate location of the gold electrode edges. The inset is the overlap range from 7 to 21 lm positions on the graphs. Reprinted with permission from [38]. Copyright 2016 American Chemical Society

112

L.D. Bastatas and J.D. Slinker

Fig. 4.15 Profile of electric field in Au/[Ir(ppy)2(bpy)] [PF6]/Au (green) and Au//[Ir (ppy)2(bpy)][PF6] + 0.5 wt% LiPF6/Au (red) planar devices applied with 9 V. Reprinted with permission from [38]. Copyright 2016 American Chemical Society

the ionic space charge. In real devices, low mobility ions, such as [Ir(ppy)2(bpy)]+, may persist in the bulk away from the contacts even with a substantial electric field in the bulk. Addition of Li[PF6] contributes small cations that redistribute efficiently to cancel out this electric field. It is informative to consider differences in the voltage profiles observed in ruthenium(II) complex based devices, which were extensively studied among the first iTMC-based LECs, with those observed in LECs with iridium(III) complexes with and without the presence of lithium salts. We performed EFM measurements of Au/[Ru(bpy)3][PF6]2/Au planar devices [35]. However, unlike the pristine device, which shows a gradual potential drop from the cathode, significant linear potential drops at both electrodes were observed in pristine ruthenium(II) complex devices. This demonstrated that ion redistribution is more efficient and idealized in devices with ruthenium(II) complexes and clarifies why the turn-on times of these pristine LECs are shorter than those from pristine iridium(III) complex devices. Ion conduction in [Ru(bpy)3][PF6]2 devices benefit from two mobile counterions as opposed to one for [Ir(ppy)2(bpy)][PF6], and the ruthenium(II) complex is, in addition, a smaller iTMC. This supports the postulate that a difference in ion sizes can impact ion transport and packing densities of ions near the contacts, resulting in differences in interfacial fields that impact device efficiency. It also shows that additives are most needed in low mobility films in LECs, such as those from highly efficient iridium(III) complexes. Overall, this work established that Li[PF6] additives enhance double layer formation for improved iTMC-based LEC performance.

4.3.4.3

Optimal Lithium Salt Concentration

Modelling study suggests that a high mobile ion concentration is needed to optimize the device performance [58]. Such device films have an enhanced ionic

4 The Use of Additives in Ionic Transition Metal Complex …

113

conductivity of the active layer that results in an increased steady state current density and recombination rate, yielding enhanced charge injection and a narrow recombination zone. On the other hand, devices with low concentrations of mobile ions can exhibit injection limited current and extremely long turn-on times due to incomplete formation of one or both EDLs near the contacts. However, doping-induced quenching finite salt solubility, and side reactions can occur which can set the threshold amount of salt that can be added to beneficially affect the devices. The concentration dependence of Li[PF6] sandwich-structure LECs with iridium (III) complexes has been explicitly reported [54]. Figure 4.16 shows the absolute luminance of the devices with different weight fractions of Li[PF6], as well as relative luminance and relative lifetimes as a ratio against the performance of salt-free device. As shown, for lower concentrations, addition of increasing amounts of Li[PF6] promotes faster response kinetics, higher peak luminance, and corresponding better maximum efficiency metrics than pristine devices. Above a threshold concentration, this trend reverses and performance suffers with increasing Li[PF6]. Among the concentrations considered, LECs with 0.5 wt% Li[PF6] emerged as the optimal device composition. At this concentration, the devices displayed almost three times the maximum brightness at only half the turn-on time required of pristine devices. Although the absolute half-life of the devices with 0.5 wt% of Li[PF6] appears to be lower than the pristine device, these are difficult to directly compare since the devices were operating at different luminance. Alternatively, for a more balanced comparison, the extrapolated half-life at 100 cd/m2 is 3.5 times longer for the 0.5 wt% lithium salt-additive device than the

Fig. 4.16 The performance of iridium(III) iTMC-based LEC devices with various amounts of Li [PF6] under constant current driving (0.050 A/cm2). Left Luminance versus time for iridium(III) iTMC-based LECs with various Li[PF6] concentrations under constant current driving. Right Relative maximum luminance and relative lifetime extrapolated at 100 cd/m2 versus Li[PF6] concentration for LECs with iridium(III) complexes, each normalized against a pristine LEC. Error bars represent the standard deviations over 4 experimental trials. Reprinted with permission from [54]. Copyright 2016 American Chemical Society

114

L.D. Bastatas and J.D. Slinker

pristine device (Fig. 4.16, right). Surface probe measurements of planar devices suggested that this enhancement is linked to details of double layer formation. Further measurements of sandwich-structure devices allowed for quantitative estimates of this effect.

4.3.4.4

Electrochemical Impedance Spectroscopy of LECs with Lithium Salt Additive

To understand the effect of different lithium salt concentrations, electrochemical impedance spectroscopy was conducted at 0 V to follow ion-dominated charge transport. As explained in detail in Chap. 1, electrochemical impedance spectroscopy (EIS) enables the direct study of thin film sandwich-structure devices and extraction of key metrics of electronic and ionic motion. EIS analysis showed that the conductance of the device is boosted by small additions of salt. Initially, conductance increases as more salt is added in small fraction, peaking at the optimal concentration—i.e. 0.5 wt% Li[PF6], then decreases as more salt is added above the threshold concentration. This is corroborated by the opposite trend of resistance, displayed in the plateau of the Bode plot shown in Fig. 4.17. To further understand this trend, devices were modelled with an equivalent circuit that allows extraction of the thickness of EDL (Fig. 4.17). Apparently, there is a correlation of the device efficiency and initial packing of cations in the cathode. A plot of the normalized efficiency versus inverse double layer thickness shown in Fig. 4.18 exhibits linear behaviour with an R-squared correlation of 0.946. To understand this correlation, consider that EDLs need to form to enhance electronic injection. In particular, to optimize the device performance, the potential barrier, as influenced by the EDL thickness, must be thin enough for easy tunnelling of injected electronic charges (Eq. 4.1) to maintain steady charge balance between the contacts and efficient light emission in the bulk of the device. Thus, it appears that with low Li[PF6] concentration Li+ salt cations compensate the imbalance of space charge near the cathode in the pristine device to balance the effect of accumulated anions at the anode. Alternatively, at high Li[PF6] concentrations, efficiency drops as the cathodic double layer broadens [38]. This broadening could arise from aggregation of Li[PF6], which would plausibly lower ionic conductivity and limit cation accumulation at the cathode. Simulation studies showed that at high concentrations in an ionic liquid environment, Li[PF6] becomes more viscous and forms aggregates that decreases its conductivity and slows down ionic transport [59]. Additionally, high concentrations of Li[PF6] could lead to a lowered efficiency of dissociation of the salt into mobile ions. As pointed out earlier, although addition of salts can improve the complex spacing and minimize self-quenching, excessive

4 The Use of Additives in Ionic Transition Metal Complex …

115

Fig. 4.17 Electrochemical impedance spectroscopy of thin film iridium(III) iTMC-based based LECs with various fractions of lithium salt additives. Left Equivalent circuit used for EIS fitting. Lcab is the inductance of the external cables, Rext is the external resistance, RE is the total electrical resistance of the active layer, CPEGEO is a constant phase element from which is derived the geometric capacitance, Rion is the bulk resistance, CEDLA and CEDLC are capacitors representing the electrical double layers, and REDLA and REDLC are the resistances of the double layers. Right Impedance versus frequency data for iridium(III) iTMC-based LEC devices with various concentrations of Li[PF6]. Solid lines are fits to the data based upon the equivalent circuit in part A. (These are hard to see given the excellent agreement with the data.) Reprinted with permission from [54]. Copyright 2016 American Chemical Society

Fig. 4.18 Normalized EQE and inverse cathodic double layer thickness versus Li[PF6] concentration in iridium(III) iTMC-based LECs. Reprinted with permission from [54]. Copyright 2016 American Chemical Society

116

L.D. Bastatas and J.D. Slinker

separation can compromise electronic charge hopping between complexes and negatively impact ionic charge transport. Thus, an appropriate concentration of added ions must be used in order to observe beneficial effects.

4.3.4.5

Counterion Dependence of Lithium Salts

Most studies of lithium salt additives to date have utilized Li[PF6], since the [PF6]− counterion is commonly used to balance the positive charge of iTMCs. Recently, Bandiello et al. investigated LECs with lithium salts of various counterions and observed distinct performance patterns from each [60]. In addition to Li[PF6], perchlorate ([ClO4]−), tetrafluoroborate ([BF4]−), and triflate ([CF3SO3]−) lithium salts were each studied and blended with iridium(III) complexes. Interestingly, the turn-on times from the LECs modified with these salts were inversely proportional to the ionic conductivity of the salts found in solution. This clearly demonstrates the important role that these salts play in ionic distribution and LEC turn-on time. Luminance and efficiency metrics were all improved approximately 30% by the lithium salts over the pristine device, with the salt-enhanced luminance values reaching approximately 1,000 cd/m2. Interestingly, three salts produced extensions of lifetimes, and two showed over tenfold enhancement, namely Li[PF6] (1,263 h) and Li[BF4] (1,973 h, extrapolated). This work represents further evidence of the ability of lithium salts to produce bright, long-lasting devices.

4.4

Outlook

To date, additives have played a key role in optimizing iTMC-based LEC performance. The impact has been far reaching, from improving key performance metrics, such as brightness, efficiency, and lifetime, but also to enable interface with practical technologies, such as direct interface with ac line power [50]. The future poses intriguing possibilities for additives in iTMC-based LECs. As an example of a potential future application, Fig. 4.19 highlights the electroluminescence from electrospun fibres of microscale dimensions. These fibres are based on a [Ru(bpy)3] [PF6]2 emitter in a PMMA host, similar to those described in a seminal publication [25]. Such devices can be implemented for high resolution spectroscopy or integrated into lab-on-a-chip applications. Clearly, the landscape for these low-cost devices is broad, and the focus for applications will sharpen as fundamental understanding yields superior device performance.

4 The Use of Additives in Ionic Transition Metal Complex …

117

Fig. 4.19 Electroluminescence from two electrospun nanofibres utilizing a [Ru(bpy)3][PF6]2 emitter in a PMMA host on interdigitated electrodes with a one micron pitch. Electroluminescence is seen in the area between the electrodes. Photo courtesy of Prof. Jose Moran-Mirabal

Acknowledgements The authors would like to thank The University of Texas at Dallas for salary support for this effort. The authors would also like to thank Dr. Kristin Suhr and Prof. José Moran-Mirabal for assistance with the figures.

References 1. H. Yersin, Top. Curr. Chem. 241, 1 (2004) 2. E. Margapoti, V. Shukla, A. Valore, A. Sharma, C. Dragonetti, C.C. Kitts, D. Roberto, M. Murgia, R. Ugo, M. Muccini, J. Phys. Chem. C 113, 12517 (2009) 3. J.M. Fernandez-Hernandez, S. Ladouceur, Y. Shen, A. Iordache, X. Wang, L. Donato, S. Gallagher-Duval, M. de Anda Villa, J.D. Slinker, L. De Cola, E. Zysman-Colman, J. Mater. Chem. C, 1, 7440 (2013) 4. S.B. Meier, D. Tordera, A. Pertegás, C. Roldán-Carmona, E. Ortí, H.J. Bolink, Mater. Today 17, 217 (2014) 5. J.D. Slinker, J. Rivnay, J.S. Moskowitz, J.B. Parker, S. Bernhard, H.D. Abruña, G.G. Malliaras, J. Mater. Chem. 17, 2976 (2007) 6. J.C. de Mello, Phys. Rev. B 66, 235210 (2002) 7. J. Gao, J. Dane, Appl. Phys. Lett. 84, 2778 (2004) 8. J.D. Slinker, A.A. Gorodetsky, M.S. Lowry, J. Wang, S. Parker, R. Rohl, S. Bernhard, G.G. Malliaras, J. Am. Chem. Soc. 126, 2763 (2004)

118

L.D. Bastatas and J.D. Slinker

9. H.J. Bolink, E. Coronado, R.D. Costa, E. Ortí, M. Sessolo, S. Graber, K. Doyle, M. Neuburger, C.E. Housecroft, E.C. Constable, Adv. Mater. 20, 3910 (2008) 10. H.-C. Su, F.-C. Fang, T.-Y. Hwu, H.-H. Hsieh, H.-F. Chen, G.-H. Lee, S.-M. Peng, K.-T. Wong, C.-C. Wu, Adv. Funct. Mater. 17, 1019 (2007) 11. J.A. Barron, S. Bernhard, P.L. Houston, H.D. Abruna, J.L. Ruglovsky, G.G. Malliaras, J. Phys. Chem. A 107, 8130 (2003) 12. K. Nagai, N. Takamiya, M. Kaneko, J. Photochem. Photobiol., A 84, 271 (1994) 13. J.K. Lee, D.S. Yoo, M.F. Rubner, Chem. Mater. 9, 1710 (1997) 14. A.P. Wu, J.K. Lee, M.F. Rubner, Thin Solid Films 329, 663 (1998) 15. A.P. Wu, D.S. Yoo, J.K. Lee, M.F. Rubner, J. Am. Chem. Soc. 121, 4883 (1999) 16. H. Rudmann, M.F. Rubner, J. Appl. Phys. 90, 4338 (2001) 17. H. Rudmann, S. Shimada, M.F. Rubner, J. Am. Chem. Soc. 124, 4918 (2002) 18. H.J. Bolink, L. Cappelli, E. Coronado, M. Grätzel, M.K. Nazeeruddin, J. Am. Chem. Soc. 128, 46 (2006) 19. G. Kalyuzhny, M. Buda, J. McNeill, P. Barbara, A.J. Bard, J. Am. Chem. Soc. 125, 6272 (2003) 20. L.J. Soltzberg, J.D. Slinker, S. Flores-Torres, D.A. Bernards, G.G. Malliaras, H.D. Abruña, J.S. Kim, R.H. Friend, M.D. Kaplan, V. Goldberg, J. Am. Chem. Soc. 128, 7761–7764 (2006) 21. J.D. Slinker, J.-S. Kim, S. Flores-Torres, J.H. Delcamp, H.D. Abruña, R.H. Friend, G.G. Malliaras, J. Mater. Chem. 17, 76–81 (2007) 22. S. Bernhard, J.A. Barron, P.L. Houston, H.D. Abruña, J.L. Ruglovksy, X.C. Gao, G.G. Malliaras, J. Am. Chem. Soc. 124, 13624 (2002) 23. D.A. Bernards, J.D. Slinker, G.G. Malliaras, S. Flores-Torres, H.D. Abruña, Appl. Phys. Lett. 84, 4980 (2004) 24. J.D. Slinker, J. Rivnay, J.A. DeFranco, D.A. Bernards, A.A. Gorodetsky, S.T. Parker, M. P. Cox, R. Rohl, G.G. Malliaras, S. Flores-Torres, H.D. Abruña, J. Appl. Phys. 99, 074502 (2006) 25. J.M. Moran-Mirabal, J.D. Slinker, J.A. DeFranco, S.S. Verbridge, G.G. Malliaras, H.G. Craighead, Nano Lett. 7, 458 (2007) 26. Z. Zhang, K. Guo, Y. Li, X. Li, G. Guan, H. Li, Y. Luo, F. Zhao, Q. Zhang, B. Wei, Q. Pei, H. Peng, Nat. Photonics 9, 233 (2015) 27. A.R. Hosseini, C.Y. Koh, J.D. Slinker, S. Flores-Torres, H.D. Abruña, G.G. Malliaras, Chem. Mater. 17, 6114 (2005) 28. H.-C. Su, C.-C. Wu, F.-C. Fang, K.-T. Wong, Appl. Phys. Lett. 89, 261118 (2006) 29. H.-C. Su, H.-F. Chen, F.-C. Fang, C.-C. Liu, C.-C. Wu, K.-T. Wong, Y.-H. Liu, S.-M. Peng, J. Am. Chem. Soc. 130, 3413 (2008) 30. J.C. de Mello, N. Tessler, S.C. Graham, R.H. Friend, Phys. Rev. B 57, 12951 (1998) 31. J.C. de Mello, J.J.M. Halls, S.C. Graham, N. Tessler, R.H. Friend, Phys. Rev. Lett. 85, 421 (2000) 32. Q. Pei, G. Yu, C. Zhang, Y. Yang, A.J. Heeger, Science 269, 1086 (1995) 33. Q. Pei, Y. Yang, G. Yu, C. Zhang, A.J. Heeger, J. Am. Chem. Soc. 118, 3922 (1996) 34. D.L. Smith, J. Appl. Phys. 81, 2869 (1997) 35. J.D. Slinker, J.A. DeFranco, M.J. Jaquith, W.R. Silveira, Y. Zhong, J.M. Moran-Mirabal, H. G. Craighead, H.D. Abruña, J.A. Marohn, G.G. Malliaras, Nat. Mater. 6, 894 (2007) 36. S.V. Reenen, P. Matyba, A. Dzwilewski, R.A. J. Janssen, L. Edman, M. Kemerink, J. Am. Chem. Soc. 132, 13776 (2010) 37. M. Lenes, G. Garcia-Belmonte, D. Tordera, A. Pertegás, J. Bisquert, H.J. Bolink, Adv. Funct. Mater. 21, 1581 (2011) 38. K. Lin, L.D. Bastatas, K.J. Suhr, M.D. Moore, B.J. Holliday, J.D. Slinker, A.C.S. Appl, Mater. Interfaces 8, 16776 (2016) 39. R.D. Costa, E. Ortí, H.J. Bolink, F. Monti, G. Accorsi, N. Armaroli, Angew. Chem. Int. Ed. 51, 8178 (2012) 40. C.H. Lyons, E.D. Abbas, J.K. Lee, M.F. Rubner, J. Am. Chem. Soc. 120, 12100 (1998) 41. J.-C. Lepretre, A. Deronzier, O. Stephan, Synth. Met. 131, 175 (2002)

4 The Use of Additives in Ionic Transition Metal Complex …

119

42. M. Armand, F. Endres, D.R. MacFarlane, H. Ohno, B. Scrosati, Nat. Mater. 8, 621 (2009) 43. S.T. Parker, J.D. Slinker, M.S. Lowry, M.P. Cox, S. Bernhard, G.G. Malliaras, Chem. Mater. 17, 3187 (2005) 44. J.D. Slinker, C.Y. Koh, G.G. Malliaras, M.S. Lowry, S. Bernhard, Appl. Phys. Lett. 86, 173506 (2005) 45. M.S. Lowry, J.I. Goldsmith, J.D. Slinker, R. Rohl, R.A. Pascal Jr., G.G. Malliaras, S. Bernhard, Chem. Mater. 17, 5712 (2005) 46. R.D. Costa, A. Pertegas, E. Ortí, H.J. Bolink, Chem. Mater. 22, 1288 (2010) 47. S. Park, D. Moon, S. Damodharan, M. Chandran, Y. Choe, Mater. Res. Bull. 47, 2807 (2012) 48. C.D. Sunesh, O. Sunseong, M. Chandran, D. Moon, Y. Choe, Mater. Chem. Phys. 136, 173 (2012) 49. R. Sun, C.-T. Liao, H.-C. Su, Org. Electron. 15, 2885 (2014) 50. J.D. Slinker, J. Rivnay, J.A. DeFranco, D.A. Bernards, A.A. Gorodetsky, S.T. Parker, M. P. Cox, R. Rohl, G.G. Malliaras, S. Flores-Torres, H.D. Abruña, J. Appl. Phys. 99, 074502 (2006) 51. D. Tordera, S. Meier, M. Lenes, R.D. Costa, E. Ortí, W. Sarfert, H.J. Bolink, Adv. Funct. Mater. 24, 897 (2012) 52. Y. Shen, D.D. Kuddes, C.A. Naquin, T. Hesterberg, C. Kusmierz, B.J. Holliday, J.D. Slinker, Appl. Phys. Lett. 102, 203305 (2013) 53. K.J. Suhr, L.D. Bastatas, Y. Shen, L.A. Mitchell, B.J. Holliday, J.D. Slinker, A.C.S. Appl, Mater. Interfaces 8, 8888 (2016) 54. L.D. Bastatas, K. Lin, M.D. Moore, K.J. Suhr, M.H. Bowler, Y. Shen, B.J. Holliday, J.D. Slinker, Langmuir 32, 9468 (2016) 55. S.V. Reenen, R.A. Janssen, M. Kemerink, Adv. Funct. Mater. 22, 4547 (2012) 56. D.R. Blasini, J. Rivnay, D.-M. Smilgies, J.D. Slinker, S. Flores-Torres, H.D. Abruña, G.G. Malliaras, J. Mater. Chem. 17, 1458 (2007) 57. K.J. Suhr, L.D. Bastatas, Y. Shen, L.A. Mitchell, G.A. Frazier, D.W. Taylor, J.D. Slinker, B. J. Holliday, Dalton Trans. 45, 17807 (2016) 58. S.V. Reenen, P. Matyba, A. Dzwilewski, R.A.J. Janssen, L. Edman, M. Kemerink, Adv. Funct. Mater. 21, 1795 (2011) 59. S. Niu, Z. Cao, T. Yan, J. Phys. Chem. B 114, 877 (2010) 60. E. Bandiello, M. Sessolo, H.J. Bolink, J. Mater, Chem. C 4, 10781 (2016)

Chapter 5

Improving Charge Carrier Balance by Incorporating Additives in the Active Layer Hai-Ching Su

Abstract Light-emitting electrochemical cells (LECs) have recently drawn much research interest because of their advantages related to single-layered device architecture, low-bias operation, and employing inert electrodes. Such features enable LECs to be inexpensive and efficient organic light-emitting devices. As mentioned in Chap. 1, the lack of charge injection and transporting layers hinders to achieve balanced carrier mobilities in single-layered LECs. In this chapter, techniques reported in literatures to achieve charge carrier balance in LECs by incorporating additives in the active layer are carefully reviewed. Furthermore, a new optical technique to probe charge carrier balance in LECs is introduced to confirm the effect attributed to the use of additives. This chapter summarizes that improving charge carrier balance of LECs would be essential to realize high device efficiency in spite of the existence of electrochemically doped layers. Keywords Charge carrier balance transport

5.1 5.1.1




Carrier trap




Carrier injection




Carrier

Introduction Characteristics of Light-Emitting Electrochemical Cells (LECs)

Organic light-emitting diodes (OLEDs) have attracted intense scientific interest recently owing to their useful applications for displays and solid-state lighting purposes. Nevertheless, OLEDs suffered from some drawbacks, such as time-consuming evaporation processes for multilayered device structures and employing reactive metals as low-work-function electrodes. To achieve simple H.-C. Su (&) Institute of Lighting and Energy Photonics, National Chiao Tung University, Tainan 71150, Taiwan e-mail: [email protected] © Springer International Publishing AG 2017 R.D. Costa (ed.), Light-Emitting Electrochemical Cells, DOI 10.1007/978-3-319-58613-7_5

121

122

H.-C. Su

fabrication processes and to employ air-stable cathode materials, light-emitting electrochemical cells (LECs) were firstly demonstrated by Pei et al. in 1995 [1]. As compared to OLEDs, LECs show some promising properties, including single-layered device structure, low-bias voltage, high power efficiency, and compatibility with air-stable cathodes. In general, LECs consist of only a single organic layer, which is compatible with solution fabrication methods, e.g., spin coating and inkjet printing. Ionic species in LECs can induce electrochemical doping under a bias, leading to low operating voltage even when a high carrier injection barrier is present. Balanced carrier injection under a low bias consequently results in high power efficiency, which is critical for lighting applications. Furthermore, since charge injection is insensitive to the work function of the electrodes, LECs can use inert electrodes, e.g., gold and silver, and thus can avoid complicated packaging processes. The active layer of an LEC has mobile ions, which can move to electrodes under a bias. Active materials of LECs are classified into four categories, namely fluorescent conjugated light-emitting polymers (CP), phosphorescent ionic transition metal complexes (iTMC), small molecules, and quantum dots—see Sect. III for more details. For non-ionic emitters, the active layers contain ionic salts to offer mobile ions. Ion-conducting polymers, e.g., poly(ethylene oxide) (PEO), are often added into the active layer to reduce phase separation between nonpolar polymers and polar salts. For ionic emitters, ion-conducting material is generally not required, however, ionic liquids, e.g., 1-butyl-3-methylimidazolium hexafluorophosphate ([BMIM][PF6]), are typically used as additives. In addition, electroluminescence (EL) efficiencies of iTMC-based LECs are commonly higher because of the phosphorescent nature of iTMCs. As explained in Sect. 1.2.3 in Chap. 1, two models were adopted to show the working principle of LECs. The first one is electrodynamical (ED) model and the second one is electrochemical doping (ECD) model. Mobile ions added in the active layer of LECs move to electrodes under a bias and thus reduce carrier injection barrier in both models. However, the reasons for reduction in carrier injection barrier are different in these two models. In the ED model (left part of Fig. 5.1), under a bias, ions approach electrodes and electric double layers (EDLs) are formed, leading to a significant drop of the electric potential at the electrode interface. Injected carriers move and recombine to generate light in the field-free layer between the EDLs. In the ECD model (right part of Fig. 5.1), oxidized and reduced species form near anode and cathode, respectively. Ohmic contacts at the electrode interfaces are established to enhance carrier injection. Doped layers penetrated into the intrinsic region at the center of active layer with time and a p-i-n structure is built up finally. Electric potential decreases significantly in the intrinsic region and EL takes place because of carrier recombination. Both models were evidenced by experiment results and simulation data. To clarify the working principle of LECs, planar PLECs were employed since it is not easy to measure the potential profile in the thin emissive layer (90%) dropped at the emissive layer/cathode interface and only a small electric field was present across the bulk of the emissive layer [11]. Similar phenomena were found in planar iTMC-based LECs [12]. These data did not follow the ECD model well but agreed with the ED model [13, 14]. In 2010, van Reenen et al. showed that using two different contacts, i.e., non-injection limited and injection limited contacts, in planar LECs lead to different potential profiles. In non-injection limited ohmic contacts, the devices agreed well with the ECD model, forming a dynamic p-n junction in the bulk of the active layer. On the other hand, in the injection-limited regime, the device followed the ED model [15]. Current-luminance-voltage (L-I-V) measurements were performed to reveal the working mechanism for sandwich LECs based on iTMCs [16]. At a bias voltage above the bandgap of the iTMCs, ion separation led to formation of the p- and n-type doped layers and the central region was still neutral, which corresponded to the ECD model. Nevertheless, at a bias below the bandgap of the iTMCs, current transients exhibited the property following the ED model [16]. These results indicated that the operation model of LECs depends on the injection regimes. If ohmic contacts are formed, the LEC follows the ECD model while ED model fits well with experimental results when no ohmic contact is formed. In case

124

H.-C. Su

only one ohmic contact is formed, only one type of doping occurs in the active layer and recombination zone is close to one of the electrodes.

5.1.2

Charge Carrier Balance in LECs

Charge carrier balance is easily optimized in multilayered OLEDs using proper carrier injection and/or transporting layers to adjust the imbalanced charge carrier mobilities of the active layer. Although balanced carrier injection in LECs is relatively easy to be achieved, hole and electron mobility discrepancy of the emissive material would lead to an off-centered recombination zone. Therefore, exciton quenching near the doped layers reduces device efficiency [16]. In general, LECs are less sensitive to electrode work function since ohmic contacts for carrier injection can be induced by the doped layers. Nevertheless, some literatures revealed that adjusting carrier injection efficiency shows pronounced effect on carrier balance. These results revealed that carrier balance of LECs would be altered by modifying carrier injection ratios. For instance, a self-assembled monolayer (SAM) accompanied with n-type doping near the cathode/polymer interface was shown to offer a doubled increase in device efficiency of CP-based LECs [17]. The n-type doped SAM layer near the indium tin oxide (ITO) layer enhanced electron injection, resulting in improved device efficiency. These data showed that ohmic contact may not always true in LECs. In 2004, Gorodetsky et al. revealed how the electrode metals affect the device characteristics of LECs based on [Ru(bpy)3][PF6]2 (where bpy is 2,2′-bipyridyl) [18]. Measured EL characteristics were insensitive to the electrode metals employed under a forward bias (ITO wired as anode). Nevertheless, different anode metals injected holes at different efficiencies under a reverse bias. The thermal deposition damage at the organic/metal interface resulted in the difference in device performance between forward and reverse biases. These results further reveal that adjusting carrier injection efficiency can modify carrier balance of LECs. Device performance of LECs can be improved by systematically adjusting the charge carrier injection efficiency. Liao et al. demonstrated optimizing device efficiencies of LECs by employing proper carrier injection layers in 2012 [19]. In a standard LEC device architecture [ITO/poly(3,4-ethylenedioxythiophene):poly (styrene sulfonate) (PEDOT:PSS) (40 nm)/emissive layer (200 nm)/Ag (100 nm)], the peak external quantum efficiency (EQE) measured at 3.5 V was 8.52%. Adding a thin hole injection layer (HIL), the hole injection efficiency enhanced and the current flow was increased consequently. Nevertheless, the EQE decreased to 6.76%. It showed that improved hole injection lead to excess holes in the active layer and, in turn, the carrier balance was deteriorated. If an electron injection layer (EIL) was inserted in the standard LEC device, both the device current and device efficiency increased. It showed that holes are present as excess carriers in the standard device and, therefore, increased electrons resulted in better carrier balance. These results indicated that proper adjusting carrier injection efficiency would be

5 Improving Charge Carrier Balance by Incorporating Additives …

125

still essential for LECs to improve device efficiency in spite of electrochemically doped layers. However, employing carrier injection layers complicates fabrication processes and adjusting charge carrier balance in the emissive layer would be a feasible approach.

5.1.3

Organization of this Chapter

In the following sections, works focused on improving charge carrier balance by incorporating additives in the active layer are briefly reviewed. Furthermore, a new optical technique to probe charge carrier balance in LECs is first introduced. It provides direct evidence of shifted recombination zone affected by altered charge carrier balance. Then, the techniques reported to optimize charge carrier balance, i.e., to reach a centered recombination zone, by incorporating additives in the active layer are reviewed. Finally, conclusion and outlook are discussed.

5.2

Optical Technique to Probe Charge Carrier Balance in LECs

To study the effect of incorporating additives on the charge carrier balance of LECs, the technology to probe the recombination zone position in the active layer is an important work. Nevertheless, it is not easy to directly probe the recombination zone in sandwiched LECs because of their thin active layer (generally 1 lm. An example of such devices is presented in Fig. 6.19.

158

G. Hernandez-Sosa et al.

Fig. 6.17 Optoelectronic characterization of a bilayer LEC with a filled lattice pattern and a dry droplet diameter and pitch of 50 and 90 lm, respectively. The device was driven with a constant current of 1 mA, and the active area was 113 mm2. The inset shows the homogenous emission from such a bilayer LEC with a droplet diameter and pitch of 60 and 120 lm, respectively. Reprinted from Ref. [42] Copyright 2014 Wiley

Fig. 6.18 Photograph of the slot-die roll coater during the deposition of the active layer (left). Close-up photograph of the slot-die head during coating of an active layer stripe (right). Adapted by Ref. [17] Copyright 2012 Macmillan Publishers Ltd

The ink formulation utilized by Edman and coworkers was based on a blend composed of SY and a PSE based on PEO:KCF3SO3 dissolved in cyclohexane, at a ratio of 1:1.35:0.25. A maximum efficacy of 0.6 cd/A at a corresponding brightness of 50 cd/m2 was reached for the champion device. This relatively low performance, compared to state of the art, was caused by the high relative content of PSE in the active layer, which was shown to be detrimental for device operation [44]. However, such a high PSE ratio was necessary to ensure optimum coating of layers.

6 Morphology Engineering and Industrial Relevant Device …

159

Fig. 6.19 Photograph of a slot-die coated LEC, illustrating the bidirectional light emission and the device conformability (left). Light emission from a semitransparent slot-die coated LEC following >6 months storage in a glove box (right). The devices were driven at V = 7 V. Adapted by Ref. [17] Copyright 2012 Macmillan Publishers Ltd

Nevertheless, further optimization of the process and materials is expected to lead to the state-of-the art efficiencies attained in laboratory experiments, but with this industrially relevant technique.

6.3.5

Spray Coating

Spray coating is a versatile technique that is suitable for a large variety of materials and solvents. It is capable of accurate thickness control, [46] with the possibility of 2-D structuring by means of a shadow mask. It has been proven effective in application fields such as paintings and protective coatings and has also been used for the fabrication of organic electronic devices [47, 48]. In general, the desired morphology of the deposited film can easily be tuned by varying the process and material parameters, such as solid concentration, substrate temperature, carrier-gas flow rate, or nozzle to substrate distance [46]. Sandström et al. used spray coating to fabricate LECs with an emitting area of *200 cm2 using the “spray-sintering” method. [45] This method was developed to produce an active layer with micro-particle morphology, rendering the device tolerant to defects and dust particles. Specifically, the sprayed droplets should start evaporating before reaching the substrate in order to prevent the formation of a wet film once the droplets land on the substrate. The droplets should then “sinter” together during the rest of the drying process forming a particle network capable to encapsulate dust particles or seal pinholes. An example of such an active layer is presented in Fig. 6.20. The image shows circular-shaped features with diameters of  20 lm. The authors also demonstrated the fabrication of multilayer structures and multicolor emission. The versatility of the spray coating technique allowed them to fabricate devices onto arbitrary shapes such as a fork which would not be possible to do with any other sheet or roll based process (see Fig. 6.21). In terms of

160

G. Hernandez-Sosa et al.

Fig. 6.20 Optical microscopy images of the first monolayer deposited during spray-sintering, with clearly distinguishable dry particle features. Reprinted from Ref. [45] Copyright 2014 Wiley

Fig. 6.21 Ambient spray-sintering of entire LEC structures, and the inclusion of light-emission function onto flat and complex-shaped metallic surfaces. Top Schematic depicting the sequential spray-sintering of a bottom electrode, an active layer, and a top electrode. Bottom, left Light emission from an all-spray-sintered LEC driven at V = 5 V. Bottom, middle Light emission from an all-spray-sintered LEC, using an Al plate coated with a layer of PEDOT:PSS as the combined substrate and anode. Bottom, right A light-emitting fork, as realized by spray-sintering a stainless-steel fork with an active layer and a top cathode. Reprinted from Ref. [45] Copyright 2014 Wiley

6 Morphology Engineering and Industrial Relevant Device …

161

performance the spray sintered LEC showed an identical performance to the spin-cast reference reaching an efficacy of 5.8 cd/A at 225 cd/m2 and 6.0 cd/A at 233 cd/m2, respectively.

6.4

Conclusion

In this chapter, we have reviewed several approaches to control morphology of PSE based LECs, as well as examples of LECs processing by industrially scalable printing and coating techniques. In the first part of the chapter, it was shown that the PSE molecular structure is not only relevant for its function in the device (i.e., ion solvating material), but also for its critical role in the morphology and rheological properties of the active layer blend. A change to its molecular weight or monomer ratio was shown to affect phase separation and consequently device performance. Additionally, these properties were also shown to affect the printability of the layer. In the second section, several examples of the fabrication of LECs by industrial relevant techniques were presented. The results of gravure and inkjet printing as well as on slot-die and spray coating demonstrated that the properties of the PSE was decisive for an optimum ink formulation that enables device fabrication. The highlighted results demonstrate that the steps taken to adjust a functional ink formulation will inevitably result in a compromise between performance and printability. In the future, however, the resulting drawbacks could be minimized if the specific characteristics of the printing process are taken into account during the early stages of the material development. The device performances achieved in the presented examples serve as a strong motivation to continue research in the field and predicts a bright future for LEC-based technological applications produced by industrial means. Acknowledgements This work was partially supported by the Federal Ministry for Education and Research grant numbers 03X5526 and 13N11903. The authors are grateful to M. Hamburger, S. Stolz, R. Eckstein, U. Lemmer for fruitful discussions.

References 1. C. Tang, S. VanSlyke, Appl. Phys. Lett. 51, 913 (1987) 2. G. Chansin, K. Ghaffarzadeh, H. Zervos, OLED Display Forecasts 2016–2026: The Rise of Plastic and Flexible Displays (2016) 3. B. Geffroy, P. Le Roy, C. Prat, Pol. Int. 55, 572 (2006) 4. S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Lüssem, K. Leo, Nature 459, 234 (2009) 5. R. Das, P. Harrop, Printed, Organic & Flexible Electronics Forecasts, Players & Opportunities 2016–2026 (2016) 6. P. Matyba, H. Yamaguchi, G. Eda, M. Chhowalla, L. Edman, N.D. Robinson, ACS Nano 4, 637 (2010)

162

G. Hernandez-Sosa et al.

7. L. Edman, Electrochim. Acta 50, 3878 (2005) 8. P. Matyba, K. Maturova, M. Kemerink, N.D. Robinson, L. Edman, Nature Mater. 8, 672 (2009) 9. S.B. Meier, D. Tordera, A. Pertegás, C. Roldán-Carmona, E. Ortí, H.J. Bolink, Mater. Today 17, 217 (2014) 10. A. Sandström, L. Edman, Energy Tech. 3, 329 (2015) 11. S.B. Meier, D. Hartmann, D. Tordera, H.J. Bolink, A. Winnacker, W. Sarfert, Phys. Chem. Chem. Phys. 14, 10886 (2012) 12. R.D. Costa, E. Ortí, H.J. Bolink, Pure Appl. Chem. 83, 2115 (2011) 13. Q. Pei, G. Yu, C. Zhang, Y. Yang, A.J. Heeger, Science 269, 1086 (1995) 14. S.B. Meier, S. van Reenen, B. Lefevre, D. Hartmann, H.J. Bolink, A. Winnacker, W. Sarfert, M. Kemerink, Adv. Funct. Mater. 23, 3531 (2013) 15. F. Huang, H. Wu, Y. Cao, Chem. Soc. Rev. 39, 2500 (2010) 16. S. Stolz, M. Petzoldt, N. Kotadiya, T. Rödlmeier, R. Eckstein, J. Freudenberg, U.H. Bunz, U. Lemmer, E. Mankel, M. Hamburger, et al. J. Mater. Chem. C 4, 11150 (2016) 17. A. Sandström, H.F. Dam, F.C. Krebs, L. Edman, Nature Commun. 3, 1002 (2012) 18. H. Kipphan, Handbook of Print Media: Technologies and Production Methods (Springer Science & Business Media, 2001) 19. S. Tang, L. Edman, J. Phys. Chem. Lett. 1, 2727 (2010) 20. D. Tordera, S. Meier, M. Lenes, R.D. Costa, E. Ortí, W. Sarfert, H.J. Bolink, Adv. Mater. 24, 897 (2012) 21. J. Mindemark, L. Edman, J. Mater. Chem. C 4, 420 (2016) 22. J. Zimmermann, N. Jürgensen, A.J. Morfa, B. Wang, S. Tekoglu, G. Hernandez-Sosa, ACS Sust. Chem. Eng. (2016) 23. G. Hernandez-Sosa, R. Eckstein, S. Tekoglu, T. Becker, F. Mathies, U. Lemmer, N. Mechau, Org. Elec. 14, 2223 (2013) 24. P. Matyba, M.R. Andersson, L. Edman, Org. Elec. 9, 699 (2008) 25. L. Sardone, C.C. Williams, H.L. Anderson, G. Marletta, F. Cacialli, P. Samori, Adv. Funct. Mater. 17, 927 (2007) 26. F.P. Wenzl, P. Pachler, C. Suess, A. Haase, E.J. List, P. Poelt, D. Somitsch, P. Knoll, U. Scherf, G. Leising, Adv. Funct. Mater. 14, 441 (2004) 27. J. Chappell, D.G. Lidzey, P.C. Jukes, A.M. Higgins, R.L. Thompson, S.O’Connor, I. Grizzi, R. Fletcher, J. O’Brien, M. Geoghegan, et al. Nat. Mater. 2, 616 (2003) 28. E. Moons, J. Phys, Condensed Matter 14, 12235 (2002) 29. R.D. Costa, A. Pertegás, E. Ortí, H.J. Bolink, Chem. Mater. 22, 1288 (2010) 30. Y. Shao, G.C. Bazan, A.J. Heeger, Adv. Mater. 19, 365 (2007) 31. J.M. Anderson, M.S. Shive, Adv. Drug Deliv. Rev. 64, 72 (2012) 32. B. Dhandayuthapani, Y. Yoshida, T. Maekawa, D.S. Kumar, Int. J. Polit. Sci. 2011, (2011) 33. H.K. Makadia, S.J. Siegel, Polymers 3, 1377 (2011) 34. M. Irimia-Vladu, E.D. Glowacki, G. Voss, S. Bauer, N.S. Sariciftci, Mater. Today 15, 340 (2012) 35. D.A. Norris, N. Puri, M.E. Labib, P.J. Sinko, J. Control. Rel. 59, 173 (1999) 36. G. Hernandez-Sosa, S. Tekoglu, S. Stolz, R. Eckstein, C. Teusch, J. Trapp, U. Lemmer, M. Hamburger, N. Mechau, Adv. Mater. 26, 3235 (2014) 37. R.R. Søndergaard, M. Hösel, F.C. Krebs, J. Polym. Sci. B: Polym. Phys. 51, 16 (2013) 38. G. Hernandez-Sosa, N. Bornemann, I. Ringle, M. Agari, E. Dörsam, N. Mechau, U. Lemmer, Adv. Funct. Mater. 23, 3164 (2013) 39. S. Tekoglu, G. Hernandez-Sosa, E. Kluge, U. Lemmer, N. Mechau, Org. Elect. 14, 3493 (2013) 40. L. Wu, Z. Dong, F. Li, H. Zhou, Y. Song, Adv. Opt. Mater. (2016) 41. G. Mauthner, K. Landfester, A. Köck, H. Brückl, M. Kast, C. Stepper, E.J. List, Org. Elect. 9, 164 (2008) 42. E.M. Lindh, A. Sandström, L. Edman, Small 10, 4148 (2014)

6 Morphology Engineering and Industrial Relevant Device …

163

43. F.C. Krebs, N. Espinosa, M. Hösel, R.R. Søndergaard, M. Jørgensen, Adv. Mater. 26, 29 (2014) 44. J. Fang, P. Matyba, L. Edman, Adv. Funct. Mater. 19, 2671 (2009) 45. A. Sandström, A. Asadpoordarvish, J. Enevold, L. Edman, Adv. Mater. (2014) 46. A. Abdellah, B. Fabel, P. Lugli, G. Scarpa, Org. Elect. 11, 1031 (2010) 47. A. Falco, A. Zaidi, P. Lugli, A. Abdellah, Org. Elect. 23, 186 (2015) 48. A. Morfa, T. Rödlmeier, N. Jürgensen, S. Stolz, G. Hernandez-Sosa, Cellulose 23, 3809 (2016)

Part III

Traditional and New Electroluminescent Materials

Chapter 7

Development of Cyclometallated Iridium(III) Complexes for Light-Emitting Electrochemical Cells Catherine E. Housecroft and Edwin C. Constable

Abstract This chapter gives an overview of the development of cyclometallated iridium(III) complexes for application in light-emitting electrochemical cells (LECs) and highlights the ligand-design strategies employed to enhance device stability, operating efficiency and (critically for LECs in which the ion mobilities are typically low) turn-on times. Typical iridium-containing ionic transition metal complexes (Ir-iTMCs) belong to the family of [Ir(C^N)2(N^N)]+ complexes in which H(C^N) is a cyclometallating ligand and N^N is a diimine or related chelating ligand. The partitioning of Ir/C^N versus N^N character in the HOMO and LUMO of a [Ir(C^N)2(N^N)]+ complex, respectively, lends itself to a ligand functionalization-driven method of varying the band-gap allowing emission-colour tuning. An important development in addressing device stability has been the design of ligands that can protect the iridium(III) centre through intra-cation p-stacking interactions, and progress in this area is discussed in detail. The need for deep-blue emitters has been addressed by introducing fluoro-substituents into the cyclometallating domain; however, this can lead to lower stability of the LECs and alternative means of shifting the emission into the blue are briefly discussed (see Chap. 8 for an in-depth discussion). Finally, we discuss how a move away from singly charged, cationic Ir-iTMCs can be used to shorten the turn-on times of LECs.







Keywords Light-emitting electrochemical cell Iridium Ligand functionalization Ligand modification Device stability Device turn-on time




7.1







Introduction

Light-emitting electrochemical cells (LECs) are light-emitting devices in which the active material is a charged species. In the prototype devices, luminescent polymers containing ionic salts were used [1]. The first LEC incorporating an ionic transition C.E. Housecroft (&)
E.C. Constable Department of Chemistry, University of Basel, Spitalstrasse 51, 4056 Basel, Switzerland e-mail: [email protected] © Springer International Publishing AG 2017 R.D. Costa (ed.), Light-Emitting Electrochemical Cells, DOI 10.1007/978-3-319-58613-7_7

167

168

C.E. Housecroft and E.C. Constable

Fig. 7.1 The structure [17] and a schematic representation of the archetype complex [Ir(ppy)2(bpy)]+. The trans-arrangement of the N-donors of the two [ppy]– ligands is typical of [Ir(C^N)2(N^N)]+ complexes

Fig. 7.2 A working LEC and a schematic representation of a double-layer LEC. (ITO = indium tin oxide; PEDOT:PSS = poly(3,4-ethylenedioxythiophene):polystyrenesulfonate; IL = ionic liquid). Metals such as Ag or Au may replace Al as the cathode. (Photograph: Dr. Collin D. Morris, University of Basel.)

metal complex (iTMC) as the active compound was reported in 1996 [2] and contained a [Ru(bpy)3]2+-based emitter (bpy = 2,2′-bipyridine). Two problems limit the development of Ru-iTMCs: (i) the difficulty of colour tuning the emission away from the characteristic orange-red of [Ru(bpy)3]2+ and (ii) the low stability of the active materials under LEC device conditions. By replacing ruthenium(II) complexes with those of iridium (from second to third row d-block), not only is the stability enhanced but it also becomes easier to tune the emission energy by systematic functionalization of the ligands. The archetype iridium(III) tris-chelate that replaces [Ru(bpy)3]2+ is [Ir(ppy)2(bpy)]+ (Fig. 7.1) where Hppy = 2phenylpyridine and [ppy]– is a cyclometallating C,N-donor ligand. Figure 7.2 gives a schematic representation of a double-layer LEC device which includes a hole injection layer (poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) =

7 Development of Cyclometallated Iridium(III) Complexes …

169

PEDOT:PSS). In the simplest device architecture, the hole injection layer is absent. The active material is an iTMC and in this chapter, we focus on Ir-iTMCs; [3–14] other chapters review other active materials including polymers, small molecules, copper(I) complexes, quantum dots and nanoparticles (see Chaps. 9, 10, 11, 12 and 13). The electroluminophore is usually blended with an ionic liquid (IL), such as 1-butyl-3-methylimidazolium hexafluoridophosphate ([BMIM][PF6]); use of IL improves the ionic conductivity of the active layer [15, 16]. Some of the challenges that face the LEC research community are: (i) the shortening of the turn-on time of the LEC device (defined either as the time for the device to attain maximum luminance or the time to reach a defined luminance), (ii) increasing the device lifetime (the lifetime, t1/2, is defined as the time for the luminance (or brightness) to decay from its maximum value, Bmax, to half this value, Bmax/2 – see Chap. 1 for more details), and (iii) achieving the holy-grail of white light emitters. If single component white light emitters cannot be effectively developed, then there is a need for deep-blue emitters to complement the widely available palette of orange emitters in order to achieve a two-component white. In this chapter, we pay particular attention to the structural features of [Ir(C^N)2(N^N)]+ complexes (C^N denotes a C,N-donor ligand typically, but not always, of the anionic cyclometallated type typified by [ppy]–, and N^N is a neutral bidentate N,N-donor, typified by 2,2′-bipyridine) that influence device stability, turn-on time, and t1/2. We also demonstrate how the choice of ligand affects the emission wavelength (colour tuning). A significant property of [Ir(C^N)2(N^N)]+ complexes is that, in the ground state, the highest-occupied molecular orbital (HOMO) is localized on the {Ir(C^N)2} domain, whilst the lowest-unoccupied molecular orbital (LUMO) is localized on the N^N ligand (Fig. 7.3). The discrete orbital composition of the [Ir(C^N)2(N^N)]+ HOMO–LUMO manifold readily lends itself to colour tuning of the emission wavelength by designing C^N ligands in order to modify the HOMO energy and by functionalizing N^N ligands to

Fig. 7.3 Representations of the HOMO (left) and LUMO (right) of the archetype Ir-iTMC [Ir (ppy)2(bpy)]+ cation which are localized, respectively, on the {Ir(ppy)2} and N^N domains (calculated at DFT level (B3LYP, 6-31G*) using Spartan 16, Wavefunction Inc.)

170

C.E. Housecroft and E.C. Constable

Table 7.1 Variation in IrIII/IrIV oxidation potential (Eox) with N^N ligand functionalization. Values are referenced with respect to the Fc/Fc+ couple Complexa

Eox (V)

Solvent

Reference

[Ir(ppy)2(bpy)][PF6] +0.84 DMF [24] +0.89 DMF [24] [Ir(ppy)2(6-Mebpy)][PF6] +0.93 DMF [24] [Ir(ppy)2(6,6′-Me2bpy)][PF6] +0.95 DMF [24] [Ir(ppy)2(4,4′-Ph2bpy)][PF6] +0.84 MeCN [25] [Ir(ppy)2(5,5′-Ph2bpy)][PF6] +0.91b MeCN [25] [Ir(ppy)2(4,4′-tBu2bpy)][PF6] +0.72 MeCN [26] [Ir(ppy)2(4,4′-(Me2N)2bpy)][PF6] a Ligand structures are given in Sect. 7.3. bThe literature value was measured with respect to SCE, then adjusted to Fc/Fc + (0.40 V vs. SCE)

modify the LUMO energy level. We shall see this strategy at work throughout this chapter. Since the HOMO is located on the {Ir(C^N)2} unit, the IrIII/IrIV oxidation potential (Eox) gives a measure of the HOMO level, and the examples in Table 7.1 illustrate tuning of the HOMO energy by functionalization of the bpy ligand in [Ir(ppy)2(N^N)][PF6] complexes. In addition, Table 1 in reference [10] gives a useful summary of photophysical and electrochemical properties of a range of Ir-iTMCs. High-level theoretical studies of iTMCs in both the ground and excited states are pivotal to an understanding of experimental results and are invaluable in designing new emissive materials [10, 18–23].

7.2 7.2.1

Synthetic Approaches to [Ir(C^N)2(N^N)]+ Complexes Use of [Ir2(C^N)4(l-Cl)2] Dimers

The most common synthetic route to [Ir(C^N)2(N^N)]+ cations is by formation of the chlorido-bridged [Ir2(C^N)4(l-Cl)2] dimer (Fig. 7.4, left) followed by treatment with the N^N ancillary ligand. The dimer is typically prepared by reaction of IrCl3
xH2O with the cyclometallating ligand, H(C^N), under reflux conditions (Scheme 7.1) following the strategy of Watts and coworkers [27]. Although this methodology is widely applied, it may be inefficient when the cyclometallating ligand contains reactive substituents, for example a sulfone [28]. An alternative route is the reaction of [Ir2(cod)2(l-Cl)2] with H(C^N) (cod = 1,5-cyclooctadiene). This route has, for example, been used to prepare mixtures of heteroleptic dimers [Ir2(ppy)4–n(dfppy)n(l-Cl)2] (Hdfppy = 2-(2,4-difluorophenyl)pyridine) [29]. Salts of [Ir(C^N)2(N^N)]+ are conveniently prepared by treatment of [Ir2(C^N)4(l-Cl)2] with the appropriate N^N ligand in methanol in a microwave reactor at *120 °C, or by heating the reagents in a solvent such as methanol under standard reflux conditions. In most cases, the cyclometallated complexes are isolated as hexafluoridophosphate salts by addition of [NH4][PF6] to the reaction mixture

7 Development of Cyclometallated Iridium(III) Complexes …

171

Fig. 7.4 Left Crystallographically determined structure of [Ir2(ppy)4(l-Cl)2] (CSD refcode RINXES [30]) as an example of an [Ir2(C^N)4(l-Cl)2] dimer. Right Crystallographically determined structure of [Ir(ptrz)2(NCMe)2]+ (CSD refcode QOMGAC [31]), the tetrazole-based cyclometallated complex cation shown in Scheme 7.4

Scheme 7.1 Preparation of [Ir2(ppy)4(l-Cl)2] as an example of the most common route used for the formation of the critical [Ir2(C^N)4(l-Cl)2] precursors

(Scheme 7.2). In a few studies, the effect of the counter-ion on device performance has been examined [32], with precipitation of the appropriate salt by the addition of ammonium or sodium salts of [BF4]–, [BPh4] – or [BIm4] – (Scheme 7.2). Although these routes are used for the vast majority of syntheses of [Ir(C^N)2(N^N)]+ luminophores, traces of residual chloride counter-ion can have a significant and detrimental effect on LEC device performance. This has been demonstrated with the prototypical complex [Ir(ppy)2(bpy)][PF6] where tight ion-pairing between aromatic protons and chloride ion (established both in solution and the solid state, Fig. 7.5) led to chloride ion being carried through the synthesis from [Ir2(ppy)4(l-Cl)2] to [Ir2(ppy)4(bpy)][PF6/Cl] [33]. In the 1H NMR spectrum of the product, characteristic shifting of the bpy H3 proton (which interacts with the Cl– ion, Fig. 7.5), can be used to quantify even small traces of chloride ion.

172

C.E. Housecroft and E.C. Constable

Scheme 7.2 Conversion of [Ir2(ppy)4(l-Cl)2] to [Ir2(ppy)4(bpy)][X] as a typical example for the preparation of salts of [Ir(C^N)2(N^N)]+ complexes

Fig. 7.5 Ion-pairing between [Ir(ppy)2(bpy)]+ and Cl– in the solid-state structure of 2{[Ir (ppy)2(bpy)][Cl]}.2CH2Cl2. [H3O].[Cl] (CSD refcode RISMOX [33])

Scheme 7.3 Solvento-intermediate approach to chloride-free salts of [Ir(C^N)2(N^N)]+, illustrated by the preparation of [Ir(ppy)2(bpy)][PF6]

7.2.2

Solvento Complexes

The use of solvento-iridium(III) precursors is a strategy developed to overcome the unfavourable effects of chloride-impurities described above. Watts and coworkers [34] first introduced this approach by treating [Ir2(ppy)4(l-Cl)2] dimers with soluble silver salts which removed chloride ion by precipitation of AgCl. The method (Scheme 7.3) has subsequently been used to isolate luminophores in LECs which exhibit exceptional stabilities. [35] A solvento approach in which the intermediate involves coordinated MeCN has been used for the preparation of tetrazole-based

7 Development of Cyclometallated Iridium(III) Complexes …

173

Scheme 7.4 Solvento-intermediate approach to tetrazole-based cyclometallated complexes

cyclometallated complexes (Scheme 7.4). In this case, attempts to isolate the corresponding chlorido dimer by the direct reaction between IrCl3
xH2O with the N-substituted 5-phenyltetrazole failed. The solvento-complex shown in Scheme 7.4 was structurally characterized (Fig. 7.4, right) [31].

7.3

Development of Ligand Types in [Ir(C^N)2(N^N)]+ Emitters

The ligands referred to in this section are defined in Schemes 7.5, 7.6, 7.8 and 7.9.

7.3.1

Archetype [Ir(ppy)2(bpy)]+ and [Ir(ppy)2(phen)]+ Complexes

Detailed investigations of the photophysical behaviour of the archetype complexes [Ir(ppy)2(bpy)]+ and [Ir(ppy)2(phen)]+ and their performances in LECs, combined with a theoretical understanding of the ground and excited state orbital compositions and energies have played an important role in defining targets for further ligand development [17]. The photoluminescence (PL) emission spectra of [Ir(ppy)2(bpy)]+ and [Ir(ppy)2(phen)]+ in solution (kmax em = 585 and 530 nm, respectively) and in PMMA-based thin films (PMMA = poly(methyl methacrylate) show unstructured broad bands, typical for metal-to-ligand charge transfer (MLCT) electronic transitions. As has now been observed in many studies, the photoluminescence quantum yields (/) increase on going from fluid solution to thin film, the latter being relevant for the LEC device, although the observation of a photoluminescence is not a guarantee that the same compound in a LEC device will exhibit more efficient electroluminescence. The / increases from 14 to 66% for [Ir (ppy)2(bpy)]+, and from 17 to 71% for [Ir(ppy)2(phen)]+ on moving from fluid solution to PMMA film. Electroluminescence (EL) spectra exhibit maxima at 590 and 578 nm for [Ir(ppy)2(bpy)]+ and [Ir(ppy)2(phen)]+, respectively. Working with structurally simple luminophores, this benchmarking study showed that a LEC

174

C.E. Housecroft and E.C. Constable

Scheme 7.5 Structures and abbreviations of N^N ligands in Sect. 7.3

7 Development of Cyclometallated Iridium(III) Complexes …

175

Scheme 7.6 Structures and abbreviations of cyclometallating ligands in Sect. 7.3

using [Ir(ppy)2(bpy)][PF6] exhibited high luminance (Bmax = 219 cd/m2) and high stability (t1/2 = 668 h), whilst that with [Ir(ppy)2(phen)][PF6] performed less well (Bmax = 40 cd/m2, t1/2 = 567 h) [17]. Both LECs show long turn-on times (ton) indicating poor ion mobility, but these can be shortened by the use of ionic liquid. With ratios of Ir-iTMC:IL of 1:0, 4:1 and 1:1, ton decreases from 70.2 to 7.2 to 0.7 h for [Ir(ppy)2(bpy)][PF6] and from 160 to 6.4 to 0.09 h for [Ir(ppy)2(phen)][PF6]. Both LECs show enhanced Bmax with the use of IL, but the t1/2 are also severely reduced. Using [Ir(ppy)2(bpy)][PF6], enhanced luminances, reduced values of ton and long t1/2 can also be achieved by the addition of [X][PF6] (X+ = NH4+, K+ or Li+) [36]. Addition of Li[PF6] similarly boosts performance of LECs containing [Ir(2,4-Ph2py)2(bpy)][PF6] [37].

7.3.2

Functionalizing N^N Ligands in [Ir(ppy)2(bpy)]+ and [Ir(ppy)2(phen)]+ with Bulky Substituents

In this section, we look at the effects of introducing substituents into the archetype N^N ligands, with a focus on sterically demanding substituents introduced in the 3,3′-, 4,4′- or 5,5′-positions of bpy, or 4,7-positions of phen. Such functionalization typically leads to stable and efficient LEC devices. The PL and EL spectra of [Ir(ppy)2(4,7-Ph2phen)][PF6] show similar maxima at *600 nm. A LEC with [Ir(ppy)2(4,7-Ph2phen)][PF6] achieves a t1/2 = 65 h, which at the time (2006) was the longest lived solid-state LEC device using an Ir-iTMC [38]. The decay in luminance is accompanied by a decay in current density. It was concluded that the long t1/2 of the LEC containing [Ir(ppy)2(4,7-Ph2phen)]+ is largely due to the hydrophobicity and steric hindrance of the peripheral phenyl rings in 4,7-Ph2phen which suppress reactions with water and its degradation products. A systematic investigation of the photophysical properties of [Ir(ppy)2(bpy)]+,

176

C.E. Housecroft and E.C. Constable

Table 7.2 Effects of substituents and use of ionic liquid on LEC device t1/2; LEC configuration = ITO/PEDOT:PSS/Ir-iTMC:IL/Al and IL = [BMIM][PF6]. Each complex was used as a [PF6]– salt. Data from Ref. [24] Complex

Ir-iTMC:IL

t1/2 (h)

[Ir(ppy)2(bpy)]+ [Ir(ppy)2(bpy)]+ [Ir(ppy)2(6-Mebpy)]+ [Ir(ppy)2(6-Mebpy)]+ [Ir(ppy)2(6,6′-Me2bpy)]+ [Ir(ppy)2(6,6′-Me2bpy)]+ [Ir(ppy)2(6,6′-Me2-4,4′-Ph2bpy)]+ [Ir(ppy)2(6,6′-Me2-4,4′-Ph2bpy)]+

4 1 4 1 4 1 4 1

70 7.8 269 28 223 41 356 95

: : : : : : : :

1 1 1 1 1 1 1 1

[Ir(ppy)2(6-Mebpy)]+, [Ir(ppy)2(6,6′-Me2bpy)]+ and [Ir(ppy)2(6,6′-Me2-4,4′Ph2bpy)]+ and of their performances in LECs has been complemented by a theoretical investigation [24]. Values of t1/2 are significantly enhanced with increased substitution and the PL of the Ir-iTMC in a film of the same composition as that used in the LEC approximately doubles (34–74%) on going from [Ir(ppy)2(bpy)]+ to [Ir(ppy)2(6,6′-Me2-4,4′-Ph2bpy)]+. The increased substitution also has a beneficial effect on the LEC stability, which can be further improved by altering the ratio of Ir-iTMC:IL (Table 7.2). Similarly, in a series of [Ir(ppy)2(2,9-X2phen)]+ luminophores in which X = Me, nBu or Ph, increasing the size of the substituents leads to improved luminance [39]. A 5,5′-substitution pattern in bpy is also advantageous. Single-layer LECs of configuration ITO/Ir-iTMC/Au with an active component of [Ir(ppy)2(5,5′-Ph2bpy)][PF6] exhibit high stability with values of t1/2 * 107 h [25]. This surpasses the t1/2 of a similarly configured LEC containing [Ir(ppy)2(4,4′-tBu2bpy)][PF6], but the time to reach the maximum emission is far longer for the LEC with [Ir(ppy)2(5,5′-Ph2bpy)][PF6] (140 h) than [Ir(ppy)2(4,4′-tBu2bpy)][PF6] (2 h). Replacing the tBu substituents by strongly electron-donating N,N-dimethylamino groups in [Ir(ppy)2(4,4′-(Me2N)2bpy)][PF6] leads to a blue-green emitter, the shift in PL emission wavelength arising from t destabilization of the LUMO; kmax em = 581 nm in [Ir(ppy)2(4,4′- Bu2bpy)][PF6] and 520 nm with 491(sh) nm in [Ir(ppy)2(4,4′-(Me2N)2bpy)][PF6] [26]. These single-layer LECs emit light with an onset voltage as low as 2.5 V. The incorporation of 9,9-dihexylfluorenyl groups into the periphery of the N^N ligands 3-fluorphen and 3,8-(fluor)2phen (Scheme 7.5) and analogous N^N ligands in which the n-hexyl chains are terminated by carbazole units, gave a series of ligands designed to extend the p-conjugation in the Ir-iTMC. Increased degree of substitution of phen leads to increased excited state lifetimes (s), and the bias required to turn on and drive LECs of configuration ITO/PEDOT:PSS/Ir-iTMC/Al increased with the size and number of side-chains. The LEC device stability were in the range of t1/2 = 6–20 h under a 6 V bias [40, 41]. The N^N ligand diazaspiro (Scheme 7.5 and Fig. 7.6) has been designed to incorporate both electron- and

7 Development of Cyclometallated Iridium(III) Complexes …

177

Fig. 7.6 The solid-state structure of [Ir (ppy)2(diazaspiro)]+ (CSD refcode EJIZUT [42]) which incorporates both electronand hole-transporting domains in the N^N ligand

hole-transporting domains. Disappointingly, the / of [Ir(ppy)2(diazaspiro)][PF6] in the solid-state is 3000 h, compared to *30 h for a LEC containing [Ir(ppy)2(bpy)]+. Replacing Phbpy by Phphen (Scheme 7.10) has a similar stabilizing influence [69]. Interestingly, however, incorporating a second p-stacking interaction by going from [Ir(ppy)2(Phbpy)]+ to [Ir(ppy)2(6,6′-Ph2bpy)]+ (Fig. 7.8) did not lead to a further enhanced LEC performance, although the LEC with [Ir(ppy)2(6,6′-Ph2bpy)]+ remains long-lived (t1/2 * 1300 h) [70]. The strategy of using 6-phenyl-substituted bpy ligands in [Ir(C^N)2(N^N)]+ luminophores has lead to some exceptional results, in particular when this is combined with the use of bulky, peripheral substituents which inhibit close cation…cation approach in the active layer, thus reducing exciton quenching. This is nicely demonstrated by a comparison of the performances of LECs which use [Ir(ppy)2(Phbpy)]+, [Ir(ppy)2(MeO-Phbpy)]+, [Ir(ppy)2(decylO-Phbpy)]+ or [Ir(ppy)2(denO-Phbpy)]+ (see Fig. 7.8 and Scheme 7.10) combined with ionic liquid (Ir-iTMC:[BMIM][PF6] = 4 : 1) in the active layer. At a driving voltage of 3 V, the values of Bmax increase with steric demands of the peripheral group: [Ir(ppy)2(decylO-Phbpy)]+ (284 cd/m2) > [Ir(ppy)2(MeO-Phbpy)]+ (183 cd/m2) > [Ir(ppy)2(Phbpy)]+ (109 cd/m2). Device stability are in the range of t1/2 = 660– 1290 h, the longest being observed with [Ir(ppy)2(Phbpy)]+. However, even after being held under a bias of 3 or 4 V for 48 h, the LEC containing [Ir(ppy)2(denO-Phbpy)]+ failed to turn on, indicating that there is a limit to the size of the substituents that can be introduced; denO-Phbpy contains a highly sterically demanding dendritic domain (Scheme 7.10) [71]. However, one must be cautious about applying a simple sterics versus performance correlation. Development of ligands for Ir-iTMCs goes hand-in-hand with changes and improvements in the

186

C.E. Housecroft and E.C. Constable

ways in which LECs are driven, and the application of a pulsed current driving method [72, 73] has had an impact on turn-on times and also stabilizes the doped regions in the device resulting in longer t1/2. Using such a pulsed driving mode, a LEC containing [Ir(ppy)2(Ph2N-Phbpy)]+ (Scheme 7.10) with the sterically demanding 4-bis(3,5,-(C6H4-4-NPh2)2phenyl group reached a luminance of 75 cd/m2 in 0.4 h and a final Bmax of 83 cd/m2, and had a t1/2 of 250 h [74]. It has also been concluded that the presence of bromo-substituents in the Ir-iTMC is detrimental to LEC performance. On going from [Ir(ppy)2(Ph2N-Phbpy)]+ to [Ir(ppy)2(Br2-Phbpy)]+ (Scheme 7.10), Bmax increases from 83 to 101 cd/m2, but t1/2 decreases from 250 to 2.9 h [74]. Exceptionally stable LECs have also been achieved by replacing the 6-phenyl substituent in the bpy ligand by a 2-naphthyl unit (Scheme 7.10). Again, a pulsed driving mode was applied, and a LEC in configuration ITO/PEDOT:PSS/[Ir(ppy)2(Naphbpy)][PF6]:[BMIM][PF6]/Al (Ir-iTMC:[BMIM][PF6] = 4:1) reached Bmax > 300 cd/m2 with no decay in luminance over a 350 h period. Significantly, this bright and stable LEC was achieved even though the solution / was only 7.7%, a fact that was explained in terms of the presence of a low-lying triplet state associated with the 2-naphthyl unit which is close to the MLCT/LLCT lowest-energy emitting triplet [75]. Appending an aryl group adjacent to the metal-binding site is not restricted to the N^N ligand, and aryl-substituted C^N ligands also lead to intra-cation p-stacking interactions, although with variable consequences. Li et al. have reported that LECs in configuration ITO/PEDOT:PSS/Ir-iTMC: [BMIM][PF6]/Al (Ir-iTMC:[BMIM] [PF6] = 4:1) driven under a bias of 3 V and with [Ir(2,6-Ph2py)2(phen)]+ or [Ir(2,4,6-Ph3py)2(phen)]+ in the active layer do not turn on even after 24 h; increasing the bias to 8 V did not lead to light emission. The double p-stack in the [Ir(2,4,6-Ph3py)2(phen)]+ cation is shown on the left in Fig. 7.9 and leads to the phen ligand being sandwiched between two phenyl rings; in turn, this causes the iridium (III) coordination environment to be noticeably distorted. DFT/TD-DFT calculations reveal that in both [Ir(2,6-Ph2py)2(phen)]+ and [Ir(2,4,6-Ph3py)2(phen)]+, the p-stacking contacts do not prevent the iridium(III) coordination sphere from opening up in the 3MC state [76]. This is in contrast to the beneficial effects seen when the aryl substituent is attached to the 6-position of bpy and the p-stacking interaction is between the pendant aryl ring and a cyclometallated ring (see Figs. 7.7 and 7.8). With this in mind, the ligand 2-(3,5-diphenyl)phenylpyridine (H(Ph2ppy), Scheme 7.10) was designed to give two separate p-stacking interactions in the complex [Ir(Ph2ppy)2(bpy)]+; the two cyclometallating ligands embrace one another as shown in the centre of Fig. 7.9. A LEC with a blend of [Ir(Ph2ppy)2(bpy)][PF6] and [BMIM][PF6] (4 : 1) in the active layer and operated under a pulsed current driving mode, reached a maximum luminance of 425 cd/m2 in 1.21 h and the device was very stable (t1/2 = 360 h). Paradoxically, however, a much brighter (Bmax = 1024 cd/m2) and more stable (t1/2 = 2800 h) LEC was fabricated using [Ir(Phppy)2(bpy)][PF6]. The H(Phppy) ligand lacks the pendant phenyl ring of H(Ph2ppy) and as a result, there are no intra-cation p-stacking contacts in [Ir(Phppy)2(bpy)]+ (Fig. 7.9, right). Significantly, a combination of a pendant 6-phenyl ring in the bpy domain with either H(Phppy) or H(Ph2ppy) as the

7 Development of Cyclometallated Iridium(III) Complexes …

187

Fig. 7.9 Left the structure of the [Ir(2,4,6-Ph3py)2(phen)]+ cation from the [PF6]– salt (CSD refcode YODFII) [76]. Middle the structure of [Ir(Ph2ppy)2(bpy)]+ cation from the [PF6]– salt (CSD refcode BOYPAI) with the two C^N ligands shown in red and pale blue respectively [35]. Right the structure of the [Ir(Phppy)2(bpy)]+ cation from the [PF6]– salt (CSD refcode BOYPEM) [35]. In each figure, the N^N ligand (phen or bpy) is on the right-hand side, and p-stacking interactions are shown in space-filling representations

Table 7.3 Performances of LECs (ITO/PEDOT:PSS/Ir-iTMC: [BMIM][PF6] (4:1)/Al) containing multiple phenyl substituents on the C^N and N^N domains in [Ir(C^N)2(N^N)]+; Phbpy, H(Phppy) and H(Ph2ppy) are defined in Fig. 7.8 and Scheme 7.10. Each complex was a [PF6]– salt, and devices were driven by applying a pulsed current density of 300 A/m2. Data from Ref. [35] ton (h)a

Luminophore

Bmax (cd/m2)

Effmax (cd/A)

t1/2 (h)b

+

[Ir(Phppy)2(bpy)] 0.14 1024 3.5 2800 1.21 425 1.4 360 [Ir(Ph2ppy)2(bpy)]+ 0.42 676 2.2 1204 [Ir(Phppy)2(Phbpy)]+ 0.05 261 0.7 >2800 [Ir(Ph2ppy)2(Phbpy)]+ a ton = time to reach Bmax. b t1/2 = time for luminance to decay from Bmax to Bmax/2 Note The terms ton, t1/2, Bmax, Effmax refer to turn-on time, lifetime, maximum luminance, maximum efficiency, respectively. The definition of these terms are provided in Chap. 1

cyclometallating ligand leads to highly efficient LECs; the performance parameters of devices with [Ir(Phppy)2(Phbpy)][PF6] or [Ir(Ph2ppy)2(Phbpy)][PF6] (and IL) in the emissive layer are summarized in Table 7.3 along with those for LECs with [Ir(Phppy)2(bpy)][PF6] and [Ir(Ph2ppy)2(bpy)][PF6] [35]. The exceptional stability (Table 7.3) of the device with [Ir(Ph2ppy)2(Phbpy)]+ in the active layer is assumed to be associated with the face-to-face p-stacking interactions shown in Fig. 7.10 which minimize attack at the metal centre by nucleophiles. Taking all the data discussed in this section into account, it appears that the p-contact between the phenyl ring of the Phbpy ligand and one of the cyclometallating rings plays a key role. The design of [Ir(C^N)2(N^N)]+ luminophores involving p-stacking between a pendant phenyl ring of the N^N ligand and one cyclometallating ring is indeed a successful strategy for achieving stable LECs. A comparison of the performances of

188

C.E. Housecroft and E.C. Constable

Fig. 7.10 The three face-to-face p-stacking interactions in [Ir(Ph2ppy)2(Phbpy)]+. There are two pyridyl…pendant phenyl interactions (shown in pale blue and orange) and one face-to-face p-contact between the phenyl ring of the Phbpy ligand and one cyclometallating ring (shown in red). Data from CSD refcode BOYPIQ [35]

LECs containing [Ir(ppy)2(Phbpy)]+ (Fig. 7.8) or [Ir(dmppz)2(Phbpy)]+ (Fig. 7.11, left) in configuration ITO/PEDOT:PSS/Ir-iTMC:[BMIM][PF6]/Al (Ir-iTMC: [BMIM][PF6] = 4:1) and under a 3 V bias confirms that the device stability observed for [Ir(ppy)2(Phbpy)]+ (discussed earlier in this section) is replicated in [Ir(dmppz)2(Phbpy)]+. The t1/2 is further increased (from 1288 to 2000 h on going from [Ir(ppy)2(Phbpy)]+ to [Ir(dmppz)2(Phbpy)]+) by the effects of the methyl substituents in the pyrazole ring which hinder the approach of nucleophiles. [77] A pendant phenyl substituent adjacent to the metal-binding site is also effective when the N^N ligand is N-Phpypz (Fig. 7.11). The [Ir(ppy)2(N-Phpzpy)]+ cation exhibits the expected stacking of the pendant phenyl group and one cyclometallating ring (Fig. 7.11, right). The effect of introducing the phenyl substituent was assessed by comparing LECs containing [Ir(ppy)2(N-Phpzpy)][PF6] and [Ir(ppy)2(pzpy)][PF6] (see Scheme 7.8 for pzpy). In a configuration ITO/PEDOT:PSS/Ir-iTMC: BMIMPF6/Al (Ir-iTMC:[BMIM][PF6] = 2:1) and under a 3 V bias, the LEC with [Ir(ppy)2(pzpy)][PF6] reached a maximum luminance of 1.4 cd/m2 in 3.8 min and had a t1/2 of * 300 min. When the pendant phenyl group is introduced, ton, Bmax and t1/2 of the blue-green LEC dramatically increase (325 min to reach 37 cd/m2, and t1/2 = 950 min) [78]. However, He et al. note that at the time of their achievement (2011), the brightness and t1/2 of this blue-green LEC remain significantly lower than those of the most stable orange-red LECs then known. Whilst fluoro-substituents in the cyclometallating ligands are used to realize blue emitters (see Sect. 7.5), Chen et al. have sought to enhance intra-cation p-stacking by using H(dfppz) which contains an electron-poor difluorophenyl cyclometallating ring (Fig. 7.12). Each of [Ir(dfppz)2(sp)][PF6], [Ir(dfppz)2(Phsp)][PF6] or [Ir(dfppz)2(Ph2sp)][PF6] was blended with ionic liquid ([BMIM][PF6]) in the active layer of a LEC. The t1/2 undergoes a 4-fold increase on going from [Ir(dfppz)2(sp)]+ to [Ir(dfppz)2(Phsp)]+, increasing from 141 min to 585 min, but introducing two phenyl rings (and therefore two p-stacking interactions) does not have a beneficial effect (t1/2 = 102 min) [79]. This contrasts with the fact that LECs with both

7 Development of Cyclometallated Iridium(III) Complexes …

189

Fig. 7.11 Left the structure of the cyclometallating ligand Hdmppz and of the [Ir(dmppz)2(Phbpy)]+ cation (CSD refcode YUWWOD) [77]. Right the N^N ligand N-Phpzpy and the [Ir(ppy)2(N-Phpzpy)]+ cation (CSD refcode UTADAV) [78]

[Ir(ppy)2(Phbpy)]+ (one p-stacking contact) and [Ir(ppy)2(6,6′-Ph2bpy)]+ (two p-stacking interactions) are long-lived, even though there is no advantage to there being an additional p-stacking interaction (see above) [70]. Structurally, 2,2′:6′,2′′-terpyridine (tpy, Scheme 7.10) is closely related to 6-phenyl-2,2′-bipyridine (Phbpy). Although tpy is most often encountered as a bis (chelating) N,N′,N′′-donor ligand [80], it can only act as a bidentate ligand in an [Ir(C^N)2(tpy)]+ complex because of the restriction to 6-coordination at the metal centre. One pyridyl ring in the bidentate tpy in [Ir(ppy)2(tpy)]+ therefore mimics the pendant phenyl ring in [Ir(ppy)2(Phbpy)]+. As expected, complexes of the type [Ir(ppy)2(Xtpy)]+ (X represents a peripheral functional group) show face-to-face p-stacking of the pendant pyridyl ring with one of the cyclometallating rings. Although this appeared promising for the design of luminophores for LECs, and despite the fact that the emissions were in the red-orange region, [Ir(ppy)2(Xtpy)][PF6] complexes showed values of / in solution or thin films of 3-msppy * 5-msppy). Despite the brightest luminance being 940 cd/m2, all the LECs are short-lived (  1.5 h). LECs with [Ir (5-msppy)2(pypz)]+ and [Ir(4-msppy)2(pypz)]+ in the active layer also show low stability with values of t1/2 of between 2.4 and 6.2 min; in this series, maximum luminances are also low (  141 cd/m2). Although the introduction of methylsulfonyl groups does not appear to aid device stability, positional isomerism has been shown to be an effective means of colour tuning the emission [28, 87]. An alternative strategy for fluorine-free blue-green emitters comes from Zysman-Coleman and Bolink [88]. A combination of a cyclometallated pyridyl (rather than phenyl) ring with electron-donating functionalities (in this case methoxy) has been shown to be electronically similar to using a fluoro-substituted cyclometallated [dfppy]– ligand. Turn-on times for devices driven under pulsed current conditions were very fast ( 100 nm) observed for complex 22, consistent with the smaller LUMO destabilization (0.2 eV) caused by pyim than that (> 0.4 eV) caused by pzpy. The emission from complex 35 exhibited dominant 3LLCT (ppy ! pyim) and 3MLCT (Ir ! pyim) character, in contrast to the dominant ppy-centered 3p-p* character for complex 22. Therefore, both the extent of the blue-shifting effect and the nature of the emissive triplet state depend on the relative positions of the electron-donating nitrogen atoms in the five-membered N-heterocycle of the ancillary ligand [33]. With the dfppy C^N ligand, complex 36 showed further blue-shifted emission, with the emission peak at 489 nm in acetonitrile solution [32]. Moreover, the complexes with pyridine-imidazole ancillary ligands showed reversible electrochemical reductions, indicating their good electrochemical stability and electron transport in LECs. The LEC ITO/PEDOT:PSS (40 nm)/36: [BMIM][PF6] (molar ratio: 1:0.35) (100 nm)/Al emitted blue-green light centered at 497 nm, [32] with a maximum brightness of 39 cd/m2 and a peak current efficiency of 8.4 cd/A under 4.0 V. White LEC were assembled based on the blue-green LEC, which showed a peak current efficiency of 11.2 cd/A at 3.5 V. He and co-authors further developed a 2,2’-biimidazole N^N ligand and used it to assemble complex 37 [32]. The biimidazole ligand exhibits a strong

8 Recent Advances on Blue-Emitting Iridium(III) Complexes … + [PF6]-

+ [PF6]-

R

N N

N

N

N

N

N

Ir

R

N

N

N

N

NH

N

R

R

R=H, 35; R=F, 36

37 + [X]-

F

Ir

N R

F

F

R=adamantyl, X=[PF6], 40a; R=adamantyl, X=[BF4], 40b; R=CH2Ph, X=[PF6], 41a; R=CH2Ph, X=[BF4], 41b

N

R1

Ir

F

N N N

N N

F

+ [PF6]-

R2

N N

N

R=H, 38; R=F, 39; + [X]-

F

N F F

N

R

Ir

Ir

R

+ [PF6]-

R

N

N R

213

Ir

R1

N N N

N R

R=Ph, X=[PF6], 42a; R=Ph, X=[BF4], 42b; R=biphenyl, X=[PF6], 43a; R=biphenyl, X=[BF4], 43b

N N

N

4

CN

R2

R1=Me, R2=H, 44; R1=R2=F, 45

+

R2

[PF6]N N

R1

Ir

R1

N

N N

N

3

R2

R1=Me, R 2=H, 46; R1=R2=F, 47

Fig. 8.4 Ionic iridium(III) complexes using ancillary ligands beyond the bpy skeleton

electron-donating property because it involves two nitrogen atoms of which the long pair electrons are conjugated into the ligand plane. Indeed, compared to complex 1, complex 37 showed a significantly destabilized (ca. 0.6 eV) LUMO. In acetonitrile solution, complex 37 emitted blue-green light peaked at 496 nm, with considerable ppy-centered 3p-p* character. Nevertheless, complex 37 showed a low luminescent efficiency, owing to the twisting of the biimidazole plane caused by steric hindrance between the two methyl groups. The LEC ITO/PEDOT:PSS (40 nm)/37: [BMIM][PF6] (molar ratio: 1:0.35) (100 nm)/Al emitted green light peaked at 524 and 497 nm, with a maximum brightness of 68 cd/m2 and a peak current efficiency of 0.4 cd/A at 5.0 V [32]. The low luminescent efficiency, high LUMO level and irreversible electrochemical reduction of complex 37 should be responsible for its low LEC performance. Sunesh and co-authors developed complexes 38 and 39 with a pyridine-imidazole ancillary ligand of 2-(4-ethyl-2-pyridyl)-1H-imidazole [34]. In acetonitrile solution, complexes 38 and 39 emitted blue-green and blue light,

214

L. He

respectively. The emission showed dominant ligand-centered 3p − p* character, in contrast to the dominant 3LLCT/3MLCT character for the emission of complexes 35–36. Therefore, the N-H in 2-(4-ethyl-2-pyridyl)-1H-imidazole appears to have a stronger electron-donating property than N-phenyl in pyim. In neat films, the emission of complexes 38–39 was red-shifted by ca. 40 nm relative to that in solution. The LEC ITO/PEDOT:PSS (40 nm)/38 or 39 (75 nm)/Al emitted green and blue-green light, with the emission maximums at 522 and 500 nm, for complexes 38 and 39, respectively [34]. Driven under sweeping voltages, the blue-green LEC incorporating complex 39 gave a peak current efficiency of 0.88 cd/A. Mydlak and co-authors developed ancillary ligands with a skeleton of 2-(1H1,2,3-triazol-4-yl)pyridine [35]. Similar to the pyridine-pyrazole ligands, these pyridine-triazole ligands had high LUMO levels. Using the pyridine-triazole type N^N ligands and the dfppy C^N ligands, complexes 40–43 were developed, which emitted blue light peaked around 452 and 484 nm in dichloromethane solution [35]. Nevertheless, complexes 40–43 showed irreversible or not fully reversible oxidation and reduction processes. LECs were fabricated with a structure of ITO/PEDOT:PSS (100 nm)/40–43: TBATf (molar ratio: 1:1) (70 nm)/Al, with tetrabutylammonium trifluoromethanesulfonate (TBATf) doped as an ionic liquid. The devices emitted blue light featuring a spectrum with a maximum and a shoulder at around 488 and 460 nm, respectively. The EL spectra exhibited only small red-shifts with respect to the PL spectra of the emissive layers. Under 5.0 V, the LECs showed maximum brightness of 14.5 − 44.9 cd/m2 and half lifetimes of 3.4 − 38 min. Chen and co-authors developed pyridine-triazole ancillary ligands with a skeleton of 2-(4H-1,2,4-triazol-3-yl)pyridine, and used them to construct blue-green and/or blue-emitting complexes 44–47 [36]. In acetonitrile solution, complexes 44 and 46 using the ppy C^N ligands emitted blue-green light peaked at 480 nm, and complexes 45 and 47 using the dfppy C^N ligands emitted blue light peaked at 454 nm. From the solution to the neat films, the emission exhibited only small red-shifts. Reversible oxidation and quasi-reversible reduction were observed for the complexes in solution. The LECs ITO/PEDOT:PSS/44–47: 14 wt% [BMIM] [PF6] (100 nm)/Al gave blue-green light for complexes 44 and 46 and sky blue light for complexes 45 and 47 [36]. Under 4.5 V, LECs with complex 44 showed a maximum brightness of 308 cd/m2 and a peak current efficiency of 6.8 cd/A, and the values for the LEC with complex 45 were 20.2 cd/m2 and 2.2 cd/A. It was observed that LECs with complex 44 showed faster response, larger current density, higher brightness and higher efficiency than LECs with complex 46. Similar results were also found between complexes 45 and 47. The enhanced performances for the LECs incorporating complexes 44 and 45 were tentatively ascribed to the higher ion mobility in the LECs, which could be rendered by the cyanogen groups in the complexes.

8 Recent Advances on Blue-Emitting Iridium(III) Complexes …

8.2.3

215

Using Ancillary Ligands with Strong Ligand Field Strength

For iridium(III) complexes, the ligand field of the ligands affects directly the dd splitting on the iridium ions, thus impacting the d-involved HOMOs, the energy of the 3MC states and the excited-state properties of the complexes. Li and co-authors systematically investigated the influence of the ligand field strength of the ancillary ligands on the energy levels and emission properties of iridium(III) complexes [37]. They found that, compared to the ligands with weak ligand field strength, the ligands with strong ligand field strength split the d orbitals of the iridium ions to a larger extent, stabilizing the HOMO levels and blue-shifting the emission of the complexes. They also concluded that the increase of the emission energy led to a decrease on the non-radiative decay rate, based on the energy gap law, and a large d-d splitting destabilized the 3MC states away from the emitting triplet state, further suppressing the non-radiative deactivation for the emitting triplet state. Therefore, ancillary ligands with strong ligand field strength can blue-shift the emission and meanwhile can enhance the luminescent efficiency of the complex. On the other hand, ancillary ligands with strong ligand field strength reduce the MLCT character in the emitting triplet excited-state, leading to decreased radiative decay rates and thus long-lived triplet states [37]. The use of ancillary ligands with strong ligand field strength accounts for an important avenue to blue-shift the emission of ionic iridium(III) complexes. The ancillary ligands used for this purpose are in general monodentate ligands such as pyrazole, :CN−,:CN-R, phosphine and carbene. For complexes including these ancillary ligands, the HOMOs distribute on the phenyl rings of the ppy-type C^N ligands and the iridium ions, similar to that in complex 1; however, the LUMOs are mainly delocalized over the pyridine moieties of the ppy-type C^N ligands rather than on the ancillary ligands. The migrations of the LUMOs from the ancillary ligands to the main ligands should be ascribed to the high energy levels of the unoccupied molecular orbitals on the ancillary ligands. Therefore, compared to complex 1, these complexes have intrinsically destabilized LUMO levels, widened energy gaps and blue-shifted emission. In coincidence with the LUMO migration, the lowest emitting triplet states of the complexes usually exhibit mixed C^N-centered 3p − p* and 3MLCT (Ir ! C^N) character, in contrast to the 3LLCT (ppy ! bpy)/3MLCT (Ir ! bpy) character for complex 1. Hereafter, all the complexes described in this section are displayed in Figs. 8.5 and 8.6. The cyanide anion (:CN−) is a known monodenate ligand with strong ligand field strength. Nazeeruddin et al. and Li et al. used it for constructing anionic iridium(III) complexes [Ir(ppy)2(CN)2]−A+, with tetrabutylammonium as the counter cation (A+) [37, 38]. Chen and co-authors developed an anionic iridium(III) complex 48, with a sodium ion as the counter cation [39]. Compared to complex 1, complex 48 showed a significantly destabilized (0.9 eV) LUMO level. In acetonitrile solution, it emitted blue-green light peaked at 472 and 502 nm, with a high

216

L. He F

Na+

R

N

N CN

R R

Ir

F

CN

N

R

F

R2

N

R1 R1

N

N

R2

Ir N

R1

N

N

F

N

F

N

F

Ir

+ [PF6]-

+ [PF6]-

F

N N

N

R=CH3, A=[PF6], 50a; R=CH3, A=[BF4],50b; R=C4H9, A=[PF6], 51a; R=C4H9, A=[BF4], 51b

+ [PF6]-

R1

N

N

R=Me, 48; R=H, 49

N N

F

Ir

+ A-

N

N

N N

Ir

F

N

N

F

N

N N

F

55

R1=R2=H, 52; R 1=F, R2=H, 53; R1=F, R2= COOCH3, 54

N

F

56

Fig. 8.5 Ionic iridium(III) complexes using ancillary ligands with strong ligand field strength

+

+ [X]-

F

N

N

[PF6]-

F

F

F

N

n-Bu N n-Bu N

Ir N

N

n-Bu n-Bu

Ir

N

N

N P

X=I, 57; X=[PF6], 58

Ph Ph P

P P=

O P Ph Ph

xantphos

P

Ph Ph P

59;

O P Ph Ph

dpephos

Ph Ph P

60;

61 P Ph Ph

dppe

Fig. 8.6 Ionic iridium(III) complexes using ancillary ligands with strong ligand field strength

luminescent efficiency of 0.70 and a relatively long excited-state lifetime (s) of 4.0 ls. The LEC ITO/PEDOT:PSS/48 (100 nm)/Al was non-emissive under electrical biases, which was attributed to the poor film quality of the emissive layer, caused by the poor solubility of complex 48 in organic solvent. Dumura and co-authors developed an analogous complex 49 and fabricated a LEC ITO/PEDOT: PSS (40 nm)/poly(N-vinylcarbazole) (PVK) (60 nm)/49 (50 nm)/Al (60 nm) [40]. Because of the insertion of the PVK film, the pinholes could be avoided and

8 Recent Advances on Blue-Emitting Iridium(III) Complexes …

217

meanwhile hole injection was facilitated for the LEC. The LEC emitted green-blue light centered at 488 nm, with x/y CIE color coordinates of 0.29/0.45 and a peak current efficiency of 0.059 cd/A under sweeping voltages. Yang and co-authors developed blue-emitting complexes 50–51, using bis-carbene ancillary ligands and the dfppy C^N ligands [41]. In dichloromethane solution, complexes 50–51 emitted blue light peaked at 452 and 482 nm arising from the dfppy ligands, and gave luminescent efficiencies of around 0.3. In neat films, the complexes still emitted blue light, exhibiting only small red-shifts with respect to their PL spectra in solution, but suffered from severe concentrationquenching. The LECs ITO/PEDOT:PSS (100 nm)/50 or 51: TBATf (molar ratio: 1:1) (70 nm)/Al gave blue-green light peaked at 488 nm [41]. The best x/y CIE color coordinates for the EL were 0.20/0.34 for complex 51a. At 5.0 V, the LECs gave peak current efficiencies at 0.37 − 0.85 cd/A and maximum brightness at 13.4 − 25.7 cd/m2. One reason for the relatively low device performances could be the less efficient electron transport as indicated by the irreversible electrochemical reductions of complexes 50–51. By using carbene-pyridine type ancillary ligands, Kessler and co-authors developed blue-green to blue-emitting complexes 52–55 [42]. Due to the very high LUMO levels of the carbene-pyridine ligands, the emission of the complexes originated from the C^N ligands. Upon adding more electron-withdrawing groups to the phenyl rings of the C^N ligands, the emission was more blue-shifted, with the emission peak blue-shifted from 471, to 450, to 447, and to 435 nm for complexes 52, 53, 54, and 55, respectively. In the solution, complex 55 exhibited the highest efficiency of 0.20 and the longest s of 8.5 ls. For complexes 52–55, the 3MC states concerning the rupture of the Ir-N bond between the iridium and the carbenepyridine ligand accounted for an efficient non-radiative deactivation channel for the emitting triplet states. The LEC ITO/PEDOT:PSS (100 nm)/53: [BMIM][PF6] (molar ratio:1:0.33) (150 nm)/Al gave green emission centered at 512 nm [42]. The red-shift between PL and EL was tentatively attributed to either a polarization effect caused by electrical excitation or a change of the emissive excited-state in solid-state. By using a bis-carbene ancillary ligand and 2’,6’-difluoro-2,3’-bipyridine as the C^N ligand, Meier and co-authors developed a deep-blue-emitting complex 56 [43]. In acetonitrile solution, complex 56 emitted blue light, with emission peaks around 440 and 480 nm and a luminescent efficiency of 0.13. In thin films, complex 56 still emitted blue light, with only small red-shifts relative to those recorded in solution, but showed severe concentration-quenching. Notably, the bis-carbene ancillary ligands eliminated the low-lying 3MC states existing in complexes 52-55 with the pyridine-carbene ancillary ligands. The LEC ITO/PEDOT:PSS (100 nm)/ 56:[BMIM][PF6] (molar ratio:1:0.33) (100 nm)/Al emitted greenish light centered at 500 nm [43]. The EL was gradually red-shifted under the continuous electrical excitation, as revealed by the change of x/y CIE color coordinates from 0.27/0.43 to 0.36/0.45 after ten minutes of operation, due to the broadening of the EL spectrum. Under 6.0 V, the LEC reached a maximum brightness of 15 cd/m2 and a peak current efficiency of 1.76 cd/A. Biased at a constant current of 2.5 mA/cm2, the

218

L. He

LEC showed a better performance, with a maximum brightness of 113 cd/m2 and a peak current efficiency of 4.7 cd/A. The LECs showed very limited stability, with half lifetimes of only several minutes. For complexes 50–56 with carbene-type ancillary ligands, the non-carbene C^N ligands dominated the emission, because of their smaller energy gaps compared to the carbene ancillary ligands. Darmawan and co-authors followed a different approach, employing a tridentate bis-carbene ligand: 1,3-phenylene-jC2)bis (1-butylimidazol-2-ylidene), to construct homoleptic cationic bis-pincer Ir(III) complexes 57–58 [44]. The tridentate bis-carbene ligands had very strong ligand field strength and formed a rigid bis-terdentate coordination with the central iridium ion. For such complexes, the emission was mainly dominated by the tridentate bis-carbene ligands. In acetonitrile solution, complexes 57 and 58 gave similar near-UV emission peaked around 384 and 406 nm, with luminescent efficiencies of 0.41, 0.38 and long s of 8.9, 9.4 ls, respectively. In films, both complexes 57 and 58 showed a new low-energy emission at around 500 nm along with the blue emission at around 400 nm. This was tentatively attributed to the formation of new trapping species associated with the counter anions. The LEC ITO/PEDOT:PSS (80 nm)/PMMA: 50 wt% 58 (80 nm)/Al emitted near-UV light peaked at 386 and 406 nm [44]. The EQE of the LEC was below 1%. This LEC represents the bluest LEC incorporating iridium(III) complexes reported so far. With strong r-donating character, phosphine has high ligand field strength [14, 45]. Recently, Martir and co-authors conducted a systematical investigation on the structure-property relationship for cationic iridium(III) complexes with different bidendtate phosphine ancillary ligands [45]. Complexes 59-61 used biphosphine ligands of xantphos, dpephos and dppe as the ancillary ligands and 2(4,6-difluorophenyl)-4-mesitylpyridine as the C^N ligand (Fig. 8.6). In acetonitrile solution, complexes 59, 60, and 61 emitted blue-green light peaked at 489, 484, and 471 nm, respectively. Complex 61 showed the bluest emission, the highest luminescent efficiency (0.52) and the longest s value (13.5 ls), because the five-membered rigid chelation ring in complex 61 rendered stronger Ir-P bonds (thus stronger ligand field strength) than the flexible eight-membered chelation rings formed in complexes 59 and 60. In neat films, the emission exhibited only small red-shifts compared to the PL in solution and moderate concentration-quenching of luminescence, due to the bulky mesityl and biphosphine ligands. The LEC ITO/PEDOT:PSS (80 nm)/59–61: [BMIM][PF6] (molar ratio: 1:0.25) (100 nm)/Al, however, showed very low luminescence level even under a high-pulsed current of 765 A m−2, which was attributed to the irreversibility of the phosphine ligand oxidation and the high driving current density [45]. The host-guest LEC ITO/PEDOT:PSS (80 nm)/host: complex 61 (mass ratio: 0.45:0.45:0.1) (100 nm)/ Al showed sky blue emission peaked at 479 nm, with a maximum brightness around 8 cd/m2 under a pulsed current density of 100 A m−2 [45].

8 Recent Advances on Blue-Emitting Iridium(III) Complexes …

8.2.4

219

Using Cyclometalating Ligands Beyond the ppy Skeleton

The C^N ligands play a critical role on the emission of iridium(III) complexes, because the HOMOs distribute partially on their phenyl rings. The typical C^N ligand used for iridium(III) complexes is ppy. The complexes with the ppy C^N ligands, such as complex 1, have narrow energy gaps. Attaching electron-withdrawing groups, such as fluorine to the phenyl rings of ppy, blue-shifts the emission, but it deteriorates the stability of the complexes in LECs (vide supra). Other C^N ligands beyond the ppy skeleton are thus desired for blue-emitting complexes. Hereafter, all the complexes described in this section are displayed in Figs. 8.7, 8.8 and 8.9. Researchers have replaced the pyridine ring in the ppy with a nitrogen-rich five-membered heterocycle, such as pyrazole, triazole or tetrazole, to develop new ionic iridium(III) complexes. In most cases, such modifications stabilize the HOMO to some extent, thus blue-shifting the emission of the complex [33]. The HOMO stabilization is usually caused by the high ligand field strength or the electron-deficient characteristic of the nitrogen-rich five-membered heterocycle. Tamayo and co-authors developed complexes 62–63 by using the phenyl-pyrazole C^N ligands and the dtb-bpy N^N ligand [46]. With the 1-phenyl-1H-pyrazole (ppz) C^N ligand, complex 62 emitted yellow light centered at 555 nm in acetonitrile solution, blue-shifted by about 30 nm relative to that of complex 2, owing to the HOMO stabilization (ca. 0.1 eV) caused by ppz. With the fluorinated ppz, i.e., 1-(2,4-difluorophenyl)-1H-pyrazole (dfppz), as the C^N ligand, complex 63

+ [PF6]-

R N

Ir

F

N N

N

F N

N

R1

R2 N

F

N

R3

N

R1

F

R1=R2=H, R3=phenyl, 70; R1=tBu, R 2=H, R3=phenyl, 71; R1=tBu, R 2=R3=phenyl, 72; R1=N(CH3)2, R2=R3=H, 73

N

R2

F

R1=R2=H, 67; R1=H, R 2=ph, 68; R1=R2=ph, 69 + [PF6]-

F N N

Ir

F

N N

N N

F

Ir

F

+ [PF6]-

N

F

F

N

R1

Ir

N

R,R=2,2'-biphenyl, 64; R,R=4-CH3OPh, 4-CH3Oph, 65; R,R=C 2H5, C2H5, 66

R=H, 62; R=F, 63

N F

R

N

N

R

R

+ [PF6]-

F

N N

F

Ir

R

N

N N

R

+ [PF6]-

F

N N

N

N

F

74

Fig. 8.7 Ionic iridium(III) complexes using cyclometalating ligands beyond the ppy skeleton

220

L. He

emitted blue-green light peaked at 495 nm in acetonitrile solution, blue-shifted by 17 nm compared to that of complex 6. The LEC ITO/PEDOT:PSS (40 nm)/63 (70 − 90 nm)/Al gave blue-green light centered at 492 nm, with x/y CIE color coordinates of 0.20/0.40 and a peak EQE of 4.6% under 3.0 V [46]. Since reported, complex 63 has been widely employed as a blue-green emitter for LECs. Liao and co-authors doped a hole-trapper of 3,3’-diethyl-2,2’-oxathiacarbocyanine iodide (DOTCI) in the blue-green LECs incorporating complex 63 [47]. The LEC structure was ITO/PEDOT:PSS (30 nm)/63: DOTCI: [BMIM][PF6] [mass ratio: (80-x): x: 20] (200 nm)/Ag. Compared to complex 63, DOTCl had a much higher HOMO level but a similar LUMO level, thus serving as a hole-trapper in the emissive layer. Because complex 63 preferred to transport holes, doping DOTCl impeded the hole mobility and improved the balance between the hole and electron transport, which moved the carrier recombination zone toward the center of the emissive layer and improved the device efficiency. Upon doping 0.01 wt% DOTCl, the maximum brightness and the peak EQE at 3.3 V were increased from 10.9 cd/m2 to 16.6 cd/m2 and from 9.06% to 12.75%, respectively. Using complex 63, the same group also assembled a white LEC of glass/CCL/ITO (120 nm)/ PEDOT:PSS (30 nm)/63: 20 wt% [BMIM][PF6] (350 nm)/Ag (100 nm), by inserting a red color-conversion layer (CCL) between the ITO and the glass substrate, [48] which showed a maximum brightness of 32 cd/m2 and a peak EQE of 12.5% under 3.7 V. As shown below, dfppz has been widely used for constructing blue-emitting iridium(III) complexes. Su and co-authors developed dfppz-based iridium(III) emitters 64-66 with 4,5-diazafluorene (daf) as the N^N ligand [49]. In dichloromethane solution, complexes 64-66 emitted blue-green light centered at 491– 499 nm, with high luminescent efficiencies of 0.46–0.66. In neat films, the luminescent efficiencies maintained at 0.20–0.28. These high luminescent efficiencies should be ascribed to the rigidity of the daf ligands. Complex 66 showed the bluest emission among the three complexes and was chosen to assemble a white LEC, which showed a maximum brightness of 2.5 cd/m2 and a peak EQE of 4.0% under 2.9 V [49]. Chen and co-authors developed another series of dfppz-based blue-green emitters 67–69, with 4,5-diaza-9,9’-spirobifluorene as the N^N ligand, in which one or two pendant phenyl rings were attached ortho- to the Ir-N bonds [50]. In acetonitrile solution, complexes 67–69 emitted blue-green light centered at 500– 505 nm, with luminescent efficiencies of 0.38–0.46. In neat films, the emission was red-shifted to 510–512 nm. The intermolecular p-p stacking interactions were formed between the pendant phenyl rings of the ancillary ligands and the phenyl rings of dfppz, as observed from the crystal structures [5]. The LECs ITO/PEDOT: PSS/67–69: [BMIM][PF6] (mass ratio: 80:20) (400 nm)/Al showed green-blue emission with x/y CIE color coordinates of 0.23/0.47, 0.28/0.50, and 0.28/0.54 for complexes 67, 68 and 69, respectively [50]. The brightness reached the maxima of 25.4, 5.8 and 10.6 cd/m2 with peak EQEs of 0.82, 1.13, and 1.43% and half lifetimes of 141, 585 and 102 min for devices incorporating complexes 67, 68, and 69 at 3.4 V, respectively. The LECs with complexes 68 and 69 showed slower

8 Recent Advances on Blue-Emitting Iridium(III) Complexes …

221

response and lower current densities, presumably due to the suppressed carrier injection/transport in the supermolecular-caged structures of densely extended p-stacking and low ionic mobility. Although containing intramolecular p-p stacking interactions, complexes 67–69 showed limited stability in LECs. Baranoff and co-authors developed dfppz-based blue-green emitters 70–73 using bpy-type ancillary ligands, in which pendant phenyl rings were attached ortho- to the Ir-N bonds (complexes 70–72) or dimethylamino substituents were attached para to the Ir-N bonds (complex 73) [51]. For complexes 70–72, strong intramolecular p − p stacking interactions were formed. In acetonitrile solution, complexes 70–73 emitted blue-green light with the emission maximums at 517, 505, 501, and 493 nm, respectively. The complexes showed very high luminescent efficiencies of 0.6 − 1.0 in diluted films. The LECs ITO/PEDOT:PSS (90 nm)/ 70-73:[BMIM][PF6] (molar ratio = 1:0.25) (90 nm)/Al emitted yellow or orange light peaked at 545–574 nm [51]. Because the PL spectra of the emissive layers were centered around 500 nm, the large red-shifts between the EL and PL should be caused by electrical excitation. The underlying reasons for such large red-shifts were unclear and could be related to the morphological effects and/or light-out-coupling effects. He and co-authors developed a dfppz-based blue-green emitting complex 74 by using a bulky pyridine-imidazole ancillary ligand [52]. In acetonitrile solution, complex 74 emitted blue-green light peaked at 494 and 472 nm. This emission was blue-shifted compared to the emission (497 nm) from complex 36 with the dfppy C^N ligand. In neat films, the PL of complex 74 was only slightly red-shifted and the luminescent efficiency maintained as high as 0.54, due to the significantly suppressed intermolecular interactions in the film caused by the bulky 4-tritylphenyl group attached at the ancillary ligand. The LEC ITO/PEDOT:PSS (40 nm)/74: [BMIM][PF6] (molar ratio: 1:1) (100 nm)/Al featured blue-green light, with x/y CIE color coordinates of 0.22/0.41 [52]. Under 3.2 V, the LEC reached a maximum brightness of 14.5 cd/m2 and a peak EQE of 7.6%. The efficiency was largely enhanced compared to the LEC incorporating complex 36 due to the suppressed luminescence concentration-quenching for complex 74. White LECs assembled with complex 74 showed a peak current efficiency of 11.4 cd/A and a peak power conversion efficiency of 11.2 lm/W [52]. Shan and co-authors developed a dfppz-based blue-green emitting complex 75, using a pyridine-triazole ancillary ligand [53]. Using complex 75, Wu and co-authors assembled a flexible LEC of PET/ITO/75: [BMIM][PF6] (mass ratio: 2:1)/Al, where PET was polyethylene terephthalate as a flexible substrate [54]. The LEC gave blue-green light centered at 503 nm with x/y CIE color coordinates of 0.25/0.48. Under 5.0 V, the blue-green LEC reached a maximum brightness of 36.5 cd/m2 and a peak current efficiency of 10.7 cd/A. The flexible white LEC were fabricated with complex 75, showing a peak current efficiency of 9.8 cd/A at 7.0 V. The efficiencies of both the blue-green and white LECs featured no obvious degradation after bending the devices at 10 mm curvature radius for 200 times.

222

L. He

Other electron-withdrawing groups rather than fluorine, such as sulfonyl groups (vide supra), have been attached to ppz for fluorine-free blue-emitting complexes. Tordera and co-authors developed complexes 76–81 by using methylsulfonyl substituted ppz as the C^N ligands and bpy derivatives as the N^N ligands [18]. In acetonitrile solution, complexes 76–81 emitted blue-green to green light centered between 492–518 nm, with luminescent efficiencies of 0.38–0.80. Complex 79 exhibited the bluest emission centered at 492 nm, due to the attachment of the electron-donating dimethylamino group onto bpy. LECs were fabricated incorporating complexes 76–78 and 80–81 with a configuration of ITO/PEDOT:PSS (90 nm)/complex:[BMIM][PF6] (molar ratio: 1: 0.25) (90 nm)/Al [18]. The LECs featured green light peaked around 500 nm for complexes 77 and 80 and 515 nm for complexes 76, 78 and 81. Upon using a driven mode based on applying pulsed current densities of 25-100 A/m2, LECs based on complexes 76 and 77 gave better performances with peak current efficiencies of around 15 cd/A. Although complexes 76–81 were fluorine-free and contained intramolecular p-p stacking interactions, the LECs showed limited stability. The longest half lifetime of 54 min at the maximum brightness of 334 cd/m2 was achieved by the LEC incorporating complex 76. Similar to the phenyl-pyrazole ligand, the phenyl-triazole ligand contains a nitrogen-rich five-membered heterocycle and has been employed as the C^N ligand for ionic iridium(III) complexes. For example, ligands with scaffolds of 1-decyl-4-phenyl-1H-1,2,3-triazole, [55] 2-phenyl-2H-1,2,3-triazole, [56] or 1-benzyl-4-phenyl-1H-1,2,3-triazole (phtl) [57] have been tested for this purpose. The phtl-type ligands have been shown to blue-shift the emission of the complex by +

F N

[PF6]N N

F

N N

N

+ [PF6]-

Ir

MeO2S

N N

R

N N

MeO2S

Ir

F

N

N N

N

R

F

R=H, 76; R=tBu, 77; R=SCH3, 78; R=N(CH3)2, 79

75

N

+ R1 [PF6]

N

R2

N

MeO2 S

Ir

MeO2S N

N

Ph

R2

N

R1=R2=tBu, 80; R1=H, R 2=SCH3, 81

R2

N

N N N N

R

Ir

Ir

R1

+ [BF4]-

N N N N N

N

R1

N

+ [PF6]-

CH2Ph N N N

N

R3

N CH2Ph

R1=F, R2=H, R3=Ph, 82; R1=F, R2=H, R3=CH2Ph, 83; R1=R2=F, R3=Ph, 84; R1=R2=F, R3=CH2Ph, 85

N N N

R

R=H, 86; R=tBu, 87

Fig. 8.8 Ionic iridium(III) complexes using cyclometalating ligands beyond the ppy skeleton

8 Recent Advances on Blue-Emitting Iridium(III) Complexes …

223

10–20 nm, as compared to ppy [55, 57]. However, the extent of the blue-shift was smaller than that caused by ppz. With fluorinated phtl as the C^N ligand and derivatives of pyridine-1,2,3-triazole (pytl) as the N^N ligands, FernándezHernández and co-authors developed blue-emitting complexes 82–85 [58]. In acetonitrile solution, complexes 82–85 emitted sky blue or deep-blue light peaked at 487, 485, 461, and 452 nm, respectively. The excited-states featured a dominant 3 LLCT (phtl ! pytl)/3MLCT (Ir ! pytl) character. In neat films, the emission was still in the blue region centered at 451–473 nm. Nevertheless, the blue emission from complexes 82–85 was very weak in both solutions or diluted films. This was tentatively attributed to the luminescence-quenching caused by the non-radiative 3 MC states. The LECs ITO/PEDOT:PSS (90 − 100 nm)/82–85/LiF (1 nm)/Al emitted yellowish green to blue light, which was largely red-shifted compared to the PL in solution or films [58]. This red-shift between the EL and PL was tentatively ascribed to the excimer formation in the film under electrical excitation. The LEC based on complex 83 gave the bluest emission centered at 487 nm with x/y CIE coordinates of 0.26/0.36. Under 9.0 V, the LECs showed maximum brightness at 3 − 25 cd/m2 and peak current efficiencies at 0.3 − 1.4 cd/A. The phenyl-tetrazole ligand contains a heterocycle with four nitrogen atoms and has also been employed as the C^N ligand for ionic iridium(III) complexes. Monti and co-authors developed complexes 86-87 by using the phenyl-tetrazole C^N ligand and the bpy-type N^N ligands [59]. With a high ligand field strength, the phenyl-tetrazole ligands stabilized the HOMOs of the complexes by 0.3 eV, as compared to the ppy ligands, which was larger than that caused by ppz (0.1 eV). In acetonitrile solution, complexes 86 and 87 emitted green light, peaked at 545 and 530 nm, respectively. They were significantly blue-shifted (45-60 nm) compared to that of complex 1. This blue-shifting effect was also larger than that caused by ppz. The LEC ITO/PEDOT:PSS (80 nm)/87: [BMIM][PF6] (90 nm)/Al emitted yellow-green light centered at 552 nm, [59] which was red-shifted compared to the PL in solution, owing to strong intermolecular interactions in the emissive layer. Although no fluorine substitutions were presented in complex 87, the LEC showed a limited stability, largely due to the irreversible oxidation of the complex induced by the tetrazole heterocycle. The phenyl-imidazole ligand contains a nitrogen-rich imidazole ring. He and co-authors examined the use of phenyl-imidazole as the C^N ligand for cationic iridium(III) complexes [60]. Different from the phenyl-pyrazole or phenyl-triazole, the ligand with a scaffold of 2-phenyl-1H-imidazole did not stabilize the HOMOs of the complexes. Instead, the phenyl-imidazole ligands destabilized simultaneously the HOMO and LUMO levels, presumably due to the electron-donating property of the imidazole ring within the complexes. With the 1,2-diphenyl-1H-imidazole (dphim) C^N ligand and the bpy N^N ligand, complex 88 gave orange-red light very similar to that of complex 1. Therefore, the nitrogen-rich five-membered heterocycle does not always blue-shift the emission of the complex and the blue-shifting effect is determined by the relative positions of the nitrogen atoms in the heterocycle [33].

224

L. He Ph

+ [PF6]-

N

R

+ [PF6]-

N

N

Ir N

N

PBu3

N

N

N

Ir

PF 6-

Ir N

N

N

N

N Ph

Bu3P

PF 6-

N

N

R

N

R=Ph, 89; R=n-Bu, 90

88

PF 6- PBu3 93

+

[PF6]-

O N

Ir

N

N N

O N

N N = bpy, 91; N N = pzpy, 92

N

Fig. 8.9 Ionic iridium(III) complexes (except 93) using cyclometalating ligands beyond the ppy skeleton

With the phenyl-imidazole C^N ligand and the pzpy N^N ligand, complexes 89 and 90 were developed, which emitted blue-green light centered around 505 nm in acetonitrile solution, with dominant 3LLCT (C^N ! N^N)/3MLCT (Ir ! N^N)) character. The LECs ITO/PEDOT:PSS (40 nm)/89 or 90: [BMIM][PF6] (molar ratio: 1: 0.5) (100 nm)/Al emitted green-blue light centered around 510 nm, with maximum brightness of 40.2, 17.8 cd/m2 and peak current efficiencies of 6.3, 2.4 cd/A for complexes 89 and 90, respectively, under 4.0 V. Other heterocycles, such as oxadiazole (oxd), have been used to replace the pyridine ring in ppy. Wang and co-authors developed complexes 91–92, by using 2,5-diphenyl-1,3,4-oxadiazole (dphoxd) as the C^N ligand [61]. Compared to the pyridine ring in ppy, the oxd heterocycle was strongly electron-deficient, which stabilized the HOMO levels of the complexes by more than 0.4 eV. In acetonitrile solution, complexes 91 and 92 emitted similar blue-green light peaked around 490 and 520 nm, with the same character of mixed dphoxd-centered 3p − p* and 3 MLCT (Ir ! dphoxd), irrespective of the different ancillary ligands (bpy or pzpy) used for the complexes. The emission was blue-shifted by about 100 nm relative to that of complex 1, owing to the HOMO stabilization induced by oxd. The LECs ITO/PEDOT:PSS (40 nm)/91 or 92: [BMIM][PF6] (molar ratio: 1: 0.8) (100 nm)/ Al showed yellow light centered at 556 nm for complex 91 and green-blue light peaked at 492, 530 nm for complex 92 [61]. The large red-shift between the EL and the solution PL for complex 91 was attributed to the change of the emitting triplet state from dphoxd-centered 3p − p* in solution to the charge-transfer state in the concentrated film, similar to that occurred in complex 5 [13]. The green-blue LEC

8 Recent Advances on Blue-Emitting Iridium(III) Complexes …

225

incorporating complex 92 showed a maximum brightness of 30 cd/m2 and a peak current efficiency of 5.2 cd/A under 3.5 V. Complexes 1–92 all contain two anionic C^N ligands and their ionic nature is rendered by the ancillary ligands. Bolink and co-authors developed an ionic iridium (III) complex 93 by ionizing the neutral complex Ir(ppy)3 through attaching cationic tri-n-butylphosphonium side groups in the periphery of the ppy ligands [62]. In acetonitrile solution, complex 93 emitted blue-green light peaked at 480 nm, which was blue-shifted by 35 nm relative to that of Ir(ppy)3, owing to the HOMO stabilization caused by the electron-withdrawing tri-butyl phosphor group. The LEC ITO/93:20 wt% PMMA (100–200 nm)/Au emitted green-blue light with a maximum at 487 nm right after the turn-on of the device [62]. However, the EL was gradually red-shifted to the yellow region upon the continuous operation of the LEC. This red-shift of EL was tentatively attributed to the chemical degradation of the complex under the electrical excitation.

8.3 8.3.1

Conclusion and Outlook Current Status

So far, a wide variety of blue-emitting iridium(III) complexes have been developed and applied for LECs. The performances of the blue or blue-green LECs with the EL maximums around or below 500 nm have been summarized in Table 8.1. As summarized in Table 8.1, great progress has been made to improve the color purity, efficiency and brightness, turn-on time and stability of blue LECs. Nevertheless, problems still remain and further improvements to these parameters are needed for applications. We provide a brief description as follows. 1. Color purity Tremendous efforts have been devoted to blue-shifting the emission of ionic iridium (III) complexes for LECs. So far, good color purity has been achieved. For example, Darmawan et al. reported the near-UV emitting complexes (57–58) and LECs, which showed EL maximums at 386 and 406 nm [44]. In general, for constructing warm-white LECs, blue-green or sky blue LECs would fulfill the requirements, of which the x/y CIE color coordinates should be around or better than 0.20/0.40 [32, 48, 49, 52, 54]. Many blue or blue-green LECs have already reached this standard (Table 8.1). Nevertheless, blue-emitting complexes do not always lead to blue LECs. Many complexes are good blue or blue-green emitters in solution or in diluted films, but their corresponding LECs feature electroluminescence responses that are red-shifted to the green or yellow region [12, 13, 20, 21, 23, 26, 29, 30, 42, 51, 58, 59, 61]. This red-shift between the PL and EL can be caused by several factors, such as (i) the downward shift of the triplet states as a result of the strong intermolecular

24

nr

nr

5.0 V

Sweeping voltage Sweeping voltage at 0.5 V/s Sweeping voltage at 0.5 V/s Sweeping voltage at 0.5 V/s Pulsed 100 A/m2

ITO/PEDOT:PSS/ 30–33: IL (1: 0.25) (100 nm)/Al ITO/PEDOT:PSS/ 34: IL (1:0.5) (100 nm)/Al

ITO/PEDOT:PSS/ 29/Al

3.0 V

426

6.5 V

300

0.5–1.3

nr

nr

nr

432

5.0 V

ITO/PEDOT:PSS/ 28/Al

nr

nr

3.5 V

37

49–141

nr

nr

94

39

52

200

33

nr

−6.0 V

ITO/4 (70–80 nm)/ Au ITO/20 (100 nm)/ Ag ITO/PEDOT:PSS/ 22 (75 nm)/Al ITO/PEDOT:PSS/ 23 (75 nm)/Al ITO/22: IL (1: 0.35) (120 nm)/Al ITO/PEDOT:PSS/ 24 (75 nm)/Al ITO/PEDOT:PSS/ 25 (75 nm)/Al

Bmax (cd/m2)

ton (min)

Driving mode

LECa

8.7

nr

0.34

0.31

0.38

0.46

4.3

0.65

0.96

0.43

nr

Effmax (cd/A)

3.4

0.2–0.4

nr

nr

nr

nr

1.6

0.28

0.36

0.20

0.16

EQEm (%)

Table 8.1 Summary of the blue or blue-green LECs incorporating iridium(III) complexes

460,486

420b

950

2.4–6.2

nr

nr

nr

nr

486,512

492–496, 530 –547

511

503

463,491

488,516

486,512

486,508

300b

nr

520,491

500

ELpeak (nm)

nr

30

t1/2 (min)

(0.26/0.48)

nr

(0.24/0.53)

(0.21/0.51)

(0.24/0.42)

(0.28/0.50)

(0.29/0.50)

(0.20/0.28)

(0.27/0.50)

nr

(0.20/0.51)

CIE (x/y)

(continued)

[31]

[30]

[29]

[29]

[27]

[27]

[26]

[26]

[26]

[24]

[12]

Ref.

226 L. He

15–45

4.5

21.6– 25.7

24.0–13.4

2.0 –14.3

nr

nr

nr

nr

4.5 –10.6

5.3 –7.8

4.0 V

4.0 V

4.5 V

4.5 V

5.0 V

5.0 V

ITO/PEDOT:PSS/ 40–43: TBATf (1:1) (70 nm)/Al ITO/PEDOT:PSS/ 44: 14 wt% IL (100 nm)/Al ITO/PEDOT:PSS/ 45: 14 wt% IL (100 nm)/Al ITO/PEDOT:PSS/ 46: 14 wt% IL (100 nm)/Al ITO/PEDOT:PSS/ 47: 14 wt% IL (100 nm)/Al ITO/PEDOT:PSS/ 50a or 50b: TBATf (1:1) (70 nm)/Al ITO/PEDOT:PSS/ 51a or 51b: TBATf (1:1) (70 nm)/Al

nr

nr

Sweeping voltage at 0.5 V/s 5.0 V

175

5.3

97

39

222

4.0 V

Bmax (cd/m2)

ITO/PEDOT:PSS/ 36: IL(1:0.35) (100 nm)/Al ITO/PEDOT:PSS/ 39 (75 nm)/Al

ton (min)

Driving mode

LECa

Table 8.1 (continued)

0.37–0.85

0.73–0.50

0.26

1.7

2.6

3.5

nr

0.88

8.4

Effmax (cd/A)

nr

nr

0.27

0.67

2.7

1.4

nr

nr

3.4

EQEm (%)

9.8–16.7

8.0–53.8

nr

nr

nr

nr

3.4 –38

488,456

488

457,486

483,515

458,488

484,517

488,460

500

497

600d

nr

ELpeak (nm)

t1/2 (min)

(0.20–0.22/0.34–0.38)

(0.27/0.43)

nr

nr

(0.21/0.33)

(0.31/0.53)

nr

(0.24/0.44)

(0.25/0.46)

CIE (x/y)

(continued)

[41]

[41]

[36]

[36]

[36]

[36]

[35]

[34]

[32]

Ref.

8 Recent Advances on Blue-Emitting Iridium(III) Complexes … 227

nr

7.8b

1700

0.28

nr

10b

100b

49

62

54–390

29

23

2.5 mA/cm2

Sweeping voltage

Pulsed 100 A/m2

3.0 V

3.3 V

3.3 V

3.4 V

3.2 V

5.0 V

36.5

14.5

5.8–25.4

16.6

10.9

113

15

3.58

6.0 V

Bmax (cd/m2)

ITO/PEDOT:PSS/ 56: IL (1: 0.33) (100 nm)/Al ITO/PEDOT:PSS/ 56: IL (1: 0.33) (100 nm)/Al ITO/PEDOT: PSS/PMMA: 50 wt % 58 (80 nm)/Al ITO/PEDOT: PSS/Host: 10 wt% 61 (100 nm)/Al ITO/PEDOT:PSS/ 63 (70–90 nm)/Al ITO/PEDOT:PSS/ 63: 20 wt% IL (200 nm)/Ag ITO/PEDOT:PSS/ 63: 0.01 wt% DOTCI: 20 wt% IL (200 nm)/Ag ITO/PEDOT:PSS/ 67–69: 20 wt% IL (400 nm)/Al ITO/PEDOT:PSS/ 74: IL (1:1) (100 nm)/Al PET/ITO/75: 33 wt % IL/Al

ton (min)

Driving mode

LECa

Table 8.1 (continued)

10.7

18.3

2.1–4.3

30

20

10.5

0.08b

nr

4.71

1.76

Effmax (cd/A)

nr

7.6

0.8–1.4

12.8

9.1

4.6

nr