Device Engineering and Degradation Mechanism Study of All-

Phosphorescent White Organic Light-Emitting Diodes

By

Lisong Xu

Submitted in Partial Fulfillment of the

Requirements for the Degree

Doctor of Philosophy

Supervised by Professor Ching W. Tang & Professor Lewis J. Rothberg

Materials Science

Arts, Sciences and Engineering

Edmund A. Hajim School of Engineering and Applied Sciences

University of Rochester

Rochester, New York

2017

ii

Biographical Sketch

Lisong Xu was born in Zhejiang, China in 1986. He received a Bachelor of Science

degree in Materials Science and Engineering from Beihang University in 2009. He

continued to pursue his studies at King Abdullah University of Science and Technology,

Saudi Arabia, where he received his Master of Science degree in Materials Science and

Engineering in 2011. In the fall of 2011, he enrolled in the doctoral program in Materials

Science at the University of Rochester, under the joint supervision of Professor Ching W.

Tang and Professor Lewis J. Rothberg. His field of study is physic, materials and devices

related to organic light-emitting diodes.

List of Publications and Papers Submitted for Publication

[1] L. Xu, C.W. Tang, and L.J. Rothberg, “High efficiency phosphorescent white

organic light-emitting diodes with an ultra-thin red and green co-doped layer and

dual blue emitting layers,” Org. Electron. Physics, Mater. Appl. 32, 54 (2016).

[2] J. Li, L. Xu, C.W. Tang, and A.A. Shestopalov, “High-resolution organic light-

emitting diodes patterned via contact printing,” ACS Appl. Mater. Interfaces 8,

16809 (2016).

[3] J. Li, L. Xu, S. Kim, and A.A. Shestopalov, “Urethane–acrylate polymers in high-

resolution contact printing,” J. Mater. Chem. C 4, 4155 (2016).

iii

[4] S.C. Dong, L. Xu and C.W. Tang, “Chemical degradation mechanism of TAPC as

hole transport layer in blue phosphorescent OLED,” Org. Electron. Physics, Mater.

Accepted Nov. 2016.

[5] L. Xu, C.W. Tang and L.J. Rothberg, “Investigation of phosphorescent blue and

white organic light-emitting diodes with high efficiency and long lifetime,” In

preparation.

[6] L. Xu, J.U. Wallace and C.W. Tang, “Fractionation of nearly osomeric di-

substituted anthracene mixtures upon thermal vacuum deposition,” In preparation.

iv

Acknowledgments

First and foremost, I would like to sincerely thank my advisor Professor Ching W.

Tang and co-advisor Lewis J. Rothberg for their continuous guidance and support

throughout the course of my pursuing the doctorate degree. Their rigorous attitude of

research and scholarship taught me all the necessary attributes to achieve academic goals

and made my study very enjoyable, exciting, fruitful and ultimately, a rich experience. I

would also like to thank them for providing me with an amazing research environment.

In addition, I would like to thank Professor Alex Shestopalov of the Department of

Chemical Engineering and Professor Yongli Gao of the Department of Physics and

Astronomy for serving as my thesis committee members and providing prompt and

valuable feedback on my research.

Special thanks go to Mr. Joseph Madathil who taught me the many techniques of

high vacuum systems that were proven to be very useful for my research work. Without

his gracious assistance and guidance, my research would have been more challenging. I

would also like to thank Dr. David S. Weiss and Dr. Ralph H. Young for their valuable

feedback upon my thesis writing. My gratitude also goes to Mr. Mike Culver and Mr. John

Miller for their help on equipment-related matters. I also deeply thank Mr. Larry Kuntz,

Mrs. Sandra Willison, Mrs. Gina Eagan and all faculty and staff members in the

Department of Chemical Engineering and Program of Materials Science for their

administrative support and assistance.

I would like to acknowledge my fellow lab-mates and colleagues: Dr. Minlu Zhang,

Dr. Wei Xia, Dr. Hao Lin, Dr. Hui Wang, Dr. Hsiang Ning (Sunny) Wu, Dr. Felipe Angel,

v

Charles Chan, Dr. Sang-Min Lee, Dr. Jason Wallace, Prashant Kumar Singh, Laura

Ciammaruchi, Guy Mongelli, Dr. Chris Favaro, Dr. Kevin Klubek, Aanand Thiyagarajan,

Michael Beckley, Thao Nguyen, Sihan (Jonas) Xie, Soyoun Kim and other group alumni,

for their collaboration and insightful discussions throughout my research. Special thanks

go to Dr. Shou-Cheng Dong from Hong Kong University of Science and Technology for

providing the chance of collaboration and for his generous advice, suggestions and

guidance.

Finally, I would like to express deep gratitude to my family for their unconditional

love, understanding and encouragement, not only during my pursuit for higher education,

but throughout my entire life.

vi

Abstract

As a possible next-generation solid-state lighting source, white organic light-

emitting diodes (WOLEDs) have the advantages in high power efficiency, large area and

flat panel form factor applications. Phosphorescent emitters and multiple emitting layer

structures are typically used in high efficiency WOLEDs. However due to the complexity

of the device structure comprising a stack of multiple layers of organic thin films, ten or

more organic materials are usually required, and each of the layers in the stack has to be

optimized to produce the desired electrical and optical functions such that collectively a

WOLED of the highest possible efficiency can be achieved. Moreover, device degradation

mechanisms are still unclear for most OLED systems, especially blue phosphorescent

OLEDs. Such challenges require a deep understanding of the device operating principles

and materials/device degradation mechanisms.

This thesis will focus on achieving high-efficiency and color-stable all-

phosphorescent WOLEDs through optimization of the device structures and material

compositions. The operating principles and the degradation mechanisms specific to all-

phosphorescent WOLED will be studied.

First, we investigated a WOLED where a blue emitter was based on a doped mix-

host system with the archetypal bis(4,6-difluorophenyl-pyridinato-N,C2) picolinate

iridium(III), FIrpic, as the blue dopant. In forming the WOLED, the red and green

components were incorporated in a single layer adjacent to the blue layer. The WOLED

efficiency and color were optimized through variations of the mixed-host compositions to

vii

control the electron-hole recombination zone and the dopant concentrations of the green-

red layers to achieve a balanced white emission.

Second, a WOLED structure with two separate blue layers and an ultra-thin red and

green co-doped layer was studied. Through a systematic investigation of the placement of

the co-doped red and green layer between the blue layers and the material compositions of

these layers, we were able to achieve high-efficiency WOLEDs with controllable white

emission characteristics. We showed that we can use the ultra-thin co-doped layer and two

blue emitting layers to manipulate exciton confinement to certain zones and energy transfer

pathways between the various hosts and dopants.

Third, a blue phosphorescent dopant tris[1-(2,6-diisopropylphenyl)-2-phenyl-1H-

imidazole]iridium(III) (Ir(iprpmi)3) with a low ionization potential (HOMO 4.8 eV) and

propensity for hole-trapping was studied in WOLEDs. In a bipolar host, 2,6-bis(3-

(carbazol-9-yl)phenyl)-pyridine (DCzPPy), Ir(iprpmi)3 was found to trap holes at low

concentrations but transport holes at higher concentrations. By adjusting the dopant

concentration and thereby the location of the recombination zone, we were able to

demonstrate blue and white OLEDs with external quantum efficiencies over 20%. The

fabricated WOLEDs shows high color stability over a wide range of luminance. Moreover,

the device lifetime has also been improved with Ir(iprpmi)3 as the emitter compared to

FIrpic.

Last, we analyzed OLED degradation using Laser Desorption Time-Of-Flight Mass

Spectrometry (LDI-TOF-MS) technique. By carefully and systematically comparing the

LDI-TOF patterns of electrically/optically stressed and controlled (unstressed) OLED

viii

devices, we were able to identify some prominent degradation byproducts and trace

possible chemical pathways involving specific host and dopant materials.

ix

Contributors and Funding Sources

This work was supervised by a dissertation committee consisting of Professor

Ching W. Tang (advisor) and Professor Alexander A. Shestopalov (committee member) of

the Department of Chemical Engineering, Professor Lewis J. Rothberg (co-advisor) of the

Department of Chemistry, Professor Yongli Gao (committee member) of the Department

of Physics and Astronomy, and Professor John C. Lambropoulos (committee chair) of the

Department of Mechanical Engineering.

Throughout the entire thesis, the organic boats used were based on an initial design

by previous fellow lab-member Dr. Sang-min Lee, Mr. Joseph Madathil and Professor

Ching Tang.

For Chapter 4, the data analyses were conducted in part by Professor Ching W.

Tang and Professor Lewis J. Rothberg and were published in 2016, in an article listed in

the Biographical Sketch.

For Chapter 5, the data analyses were conducted in part by Professor Ching W.

Tang and Dr. Shou-Cheng Dong. The results were presented at the 2016 MRS Spring

Meeting & Exhibit in Phoenix, AZ.

For Chapter 6, Dr. Shou-Cheng Dong of HKUST performed TOF/TOF experiment

and DFT calculation of TAPC, which was supported by IAS at HKUST. Part of the the

analyses were conducted in part by Dr. Dong and Professor Tang, and were submitted for

publication in 2016, in an article listed in the Biographical Sketch.

All other work conducted for this dissertation was completed by Lisong Xu

independently.

x

Table of Contents

Biographical Sketch ii

Acknowledgements iv

Abstract vi

Contributors and Funding Source ix

List of Tables xiv

List of Figures xvii

Chapter 1 Background and Introduction 1

1.1. Introduction to White Organic Light Emitting Diodes 1

1.2. Basics of OLEDs 2

1.2.1. Basic Device Physics 2

1.2.2. Fluorescence and Phosphorescence from OLEDs 4

1.2.3. Energy Transfer and Quenching in OLEDs 6

1.3. Performance Characterization of WOLEDs 9

1.4. Status of WOLED Development 12

1.4.1. All-Fluorescent WOLEDs 15

1.4.2. All-Phosphorescent WOLEDs 17

1.4.3. Hybrid WOLEDs 20

1.4.4. TADF WOLEDs 23

1.5. Device Stability and Degradation Mechanism of WOLEDs 25

1.5.1. Instability of Blue Phosphorescent Materials 26

xi

1.5.2. MALDI-TOF-MS 27

1.6. Objectives and Outline of the Thesis 29

References 32

Chapter 2 Experimental Methods and Materials 40

2.1. Vacuum Vapor Deposition Process 40

2.2. Boat Design and Coater Specifications 42

2.3. Device Fabrication Conditions 45

2.4. Device and Material Characterization 48

2.5. Device Lifetime Test 49

2.6. LDI-TOF-MS Analysis 50

2.7. Materials 51

Chapter 3 White Organic Light-Emitting Diodes with FIrpic in a Mixed-Host 56

3.1. Introduction 56

3.2. Results and Discussion 57

3.2.1. Effects of an mCP Buffer Layer 58

3.2.2. Effects of Host Types for FIrpic 60

3.2.3. Effects of Red Dopant Concentration 65

3.2.4. The Role of a Non-Doped Interlayer 68

3.3. Conclusions 70

References 72

Chapter 4 High Efficiency White Organic Light-Emitting Diodes with an Ultra-Thin

Red and Green Co-Doped Layer and Dual Blue Emitting Layers 74

xii

4.1. Introduction 74

4.2. Results and Discussion 75

4.3. Conclusions 87

References 88

Chapter 5 Investigation of Phosphorescent Blue and White Organic Light-Emitting

Diodes with High Efficiency and Improved Lifetime 90

5.1. Introduction 90

5.2. Results and Discussion 91

5.3. Conclusions 102

References 103

Chapter 6 Investigating Chemical Degradation Mechanism of High-Triplet-Energy

Materials in Blue Phosphorescent OLED Using LDI-TOF 106

6.1. Introduction 106

6.2. Results and Discussion 108

6.2.1. Device Performance and Lifetime Evaluation 108

6.2.2. Overall Stability Assessment of the Blue PhOLED 109

6.2.3. Degradation of Blue Dopant 114

6.2.4. Degradation of TAPC 114

6.2.5. Degradation of TCTA, DCzPPy and TmPyPB 121

6.3. Conclusions 127

References 128

Chapter 7 Summary and Future Work 130

xiii

References 137

xiv

List of Tables

Table 2.1: Materials used throughout this thesis. HOMO/LUMO/triplet energies were

taken from literature. 52

Table 3.1: EL performance of WOLEDs with the mCP buffer layer. ITO (110

nm)/TAPC:MoO3 (40%, 10 nm)/HTL (30nm)/TCTA:TPBi:FIrpic(28%:57%:15%,

4nm)/TPBi:Ir(2-phq)2(acac):Ir(ppy)3 (1.5%, 5%, 4 nm)/BPhen (20

nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm). (Measured at a current density of 5

mA/cm2) 59

Table 3.2: EL performance of WOLEDs with a mixed-host for FIrpic. ITO (110

nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (27 nm)/mCP (3 nm)/TCTA:TPBi:FIrpic (x:y,

15%, 4nm)/TPBi:Ir(2-phq)2(acac):Ir(ppy)3 (1.5%, 5%, 4 nm)/BPhen (20

nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm). (Measured at a current density of 5

mA/cm2) 61

Table 3.3: EL performance of WOLEDs with a mixed-host for FIrpic. ITO (110

nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (27 nm)/mCP (3 nm)/TCTA:TmPyPB:FIrpic

(x:y, 15%, 4nm)/TPBi:Ir(2-phq)2(acac):Ir(ppy)3 (1.5%, 5%, 4 nm)/BPhen (20

nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm). (Measured at a current density of 5

mA/cm2) 63

Table 3.4: EL performance of WOLEDs with a mixed-host for FIrpic. ITO (110

nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (30 nm)/mCP (x nm)/TCTA:DCzPPy:FIrpic

xv

(y:z, 15%, 4nm)/DCzPPy:/Ir(2-phq)2(acac):Ir(ppy)3 (1.5%, 5%, 4 nm)/BPhen (20

nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm). (Measured at a current density of 5

mA/cm2) 65

Table 3.5: EL performance of WOLEDs with different red dopant concentrations. ITO (110

nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (30 nm)/mCP (3

nm)/TCTA:DCzPPy:FIrpic(28%:57%:15%, 4nm)/TPBi:Ir(2-phq)2(acac):Ir(ppy)3

(x%, 6%, 6 nm)/BPhen (20 nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm).

(Measured at a current density of 5 mA/cm2) 67

Table 3.6: EL performance of WOLEDs with interlayers. ITO (110 nm)/TAPC:MoO3

(40%, 10 nm)/TAPC (30 nm)/TCTA:FIrpic (85%:15%, 4nm)/Interlayer/Host:/Ir(2-

phq)2(acac):Ir(ppy)3 (1.5%, 5%, 4 nm)/BPhen (20 nm)/BPhen:Cs2CO3 (50%, 10

nm)/Al (100 nm). (Measured at a current density of 5 mA/cm2) 70

Table 4.1: EL Performance of devices with four different ultra-thin layer doping conditions.

ITO (110nm)/HATCN (3 nm)/TAPC(37 nm)/TCTA:FIrpic (15%, 4nm)/TCTA:Ir(2-

phq)2(acac):Ir(ppy)3 (x%:y%, 0.5 nm)/DCzPPy:FIrpic (20%, 3nm)/TmPyPB (10

nm)/TmPyPB:Cs2CO3 (50%, 30 nm)/Al (100 nm). (a: values at 5 mA/cm2; b:

luminance range from 400 to 4000 cd/m2.) 79

Table 4.2: EL Performance of devices with various thicknesses of the ultra-thin red and

green co-doped layer (driven at 5 mA/cm2). ITO (110nm)/HATCN(3 nm)/TAPC (37

nm)/TCTA:FIrpic (15%, 4 nm)/TCTA:Ir(2-phq)2(acac):Ir(ppy)3 (2%:6%, x

xvi

nm)/DCzPPy:FIrpic (20%, 3 nm)/TmPyPB (10 nm)/TmPyPB:Cs2CO3 (50%, 30

nm)/Al (100 nm). 82

Table 4.3: EL Performance of white devices with selectively blue doped emitting layers

(driven at 5 mA/cm2). ITO (110nm)/HATCN (3 nm)/TAPC(37 nm)/TCTA:FIrpic (x%,

4 nm)/TCTA:Ir(2-phq)2(acac):Ir(ppy)3 (2%:6%, 0.5 nm)/DCzPPy:FIrpic (y%, 3

nm)/TmPyPB (10 nm)/TmPyPB:Cs2CO3 (50%, 30 nm)/Al (100 nm). 84

Table 4.4: EL Performance of blue devices with selectively blue doped emitting layers

(driven at 5 mA/cm2). ITO (110nm)/HATCN (3 nm)/TAPC(37 nm)/TCTA:FIrpic (x%,

4 nm)/ /DCzPPy:FIrpic (y%, 3 nm)/TmPyPB (10 nm)/TmPyPB:Cs2CO3 (30 nm)/Al

(100 nm). 86

Table 5.1: EL Performance of WOLEDs with various Ir(iprpmi)3 concentrations.

ITO/HATCN(3 nm)/TAPC(40 nm)/TCTA(4 nm)/TCTA:Ir(2-phq)2(acac):Ir(ppy)3

(2%, 6%, 1 nm)/DCzPPy:Ir(iprpmi)3 (x%, 4 nm)/TmPyPB(10

nm)/TmPyPB:Cs2CO3(50%, 30 nm)/Al (100 nm). (a: values at 5 mA/cm2, b: measured

at current densities from 0.05 mA/cm2 to 20 mA/cm2.) 97

Table 6.1: List of mass peaks and their proposed structures. 112

Table 6.2: List of mass peaks and their proposed structures. 124

xvii

List of Figures

Figure 1.1: (a) Energy level diagram of a single-layer OLED; (b) Modern OLED device

architecture illustration. 3

Figure 1.2: Spin states of electrons. 5

Figure 1.3: Different photon-emitting mechanisms of OLEDs [26]. 6

Figure 1.4: The schematic diagram of Förster resonance energy transfer [26]. 7

Figure 1.5: The schematic diagram of Dexter electron transfer [26]. 8

Figure 1.6: The CIE chromaticity diagram. 11

Figure 1.7: Various device layouts to realize white light emission. (a) vertically stacked

OLEDs, (b) pixelated monochrome OLEDs, (c) single-emitter-based WOLEDs, (d)

blue OLEDs with down-conversion layers, (e) multiple-doped emission layers (EMLs),

and (f) single OLEDs with a sub-layer EML design. (Reprinted with permission from

ref. [7]) 13

Figure 1.8: Characteristics of a WOLED. a) device structure and b) luminance decay curve.

The inset shows the lifetime versus initial luminance relationship. (Reprinted with

permission from ref. [56]) 17

Figure 1.9: Energy level diagram of a WOLED. Solid lines correspond to HOMO and

LUMO energies. The orange color marks intrinsic regions of the emission layer.

(Reprinted with permission from ref. [14]) 19

xviii

Figure 1.10: Device configurations of WOLEDs. The dopants employed are FIrpic for blue

(B), Ir(ppy)2(acac) for green (G), Ir(BT)2(acac) for yellow (Y), and Ir(MDQ)2(acac)

for red (R). (Reprinted with permission from ref. [36]) 20

Figure 1.11: (a) Device architecture and the energy level diagram of the hybrid WOLED.

(b) Lifetime of the device with Bepp2 as the interlayer. (Reprinted with permission

from ref. [68]) 22

Figure 1.12: Energy-level scheme for materials used in the hybrid WOLED, and exciton

energy diagram of the EMLs. R, G, B, and Tm represent Ir(MDQ)2(acac),

Ir(ppy)2(acac), 4P-NPD, and TmPyPB, respectively. (Reprinted with permission from

ref. [69]) 23

Figure 1.13: Materials, energy-level scheme and exciton-energy transfer mechanism of a

hybrid WOLED incorporating a blue TADF material. (Reprinted with permission

from ref. [71]) 25

Figure 1.14: Schematic of a MALDI-TOF-TOF-MS setup [81]. 29

Figure 2.1: Basic design of a vacuum vapor deposition coating system. 41

Figure 2.2: Design, components and boats for organic and inorganic material deposition.

42

Figure 2.3: Boat assembly and sensor configuration. (a) Boats and sensors alignment, (b)

graphical top view of the boats assembly. 44

Figure 2.4: Configuration of ITO pattern on glass substrates. 47

xix

Figure 3.1: (a) Energy level diagram of all materials used in WOLEDs. (b) Device structure

of a typical WOLED. 58

Figure 3.2: (a) EL spectra of Devices B1, B2, B3 and B4. (b) EQE vs. luminance vs. PE of

Devices B1, B2, B3 and B4. (Measured at a current density of 5 mA/cm2) 61

Figure 3.3: (a) EL spectra of Devices C1, C2, C3 and C4. (b) EQE vs. luminance vs. PE of

Devices C1, C2, C3 and C4. (Measured at a current density of 5 mA/cm2) 63

Figure 3.4: (a) EL spectra of Devices D1, D2 and D3. (b) EQE vs. luminance vs. PE of

Devices D1, D2 and C3. (Measured at a current density of 5 mA/cm2) 65

Figure 3.5: (a) EL spectra of Devices E1, E2 and E3. (b) EQE vs. luminance vs. PE of

Devices E1, E2 and E3. (Measured at a current density of 5 mA/cm2) 66

Figure 3.6: (a) EL spectra of Devices F1, F2 and F3. (b) EQE vs. luminance vs. PE of

Devices F1, F2 and F3. (Measured at a current density of 5 mA/cm2) 70

Figure 4.1: Energy level diagram and device architecture of a WOLED with an ultra-thin

red, green co-doped emitting layer (LUMO and HOMO energy levels are labeled

above and below the rectangles, triplet energy levels are indicated in parentheses). 76

Figure 4.2: EQE vs current density of devices with four different ultra-thin layer doping

conditions. (Embedded are the EL spectra of the four devices driven at 5 mA/cm2.)

78

xx

Figure 4.3: Absorption and emission spectra of various materials used in this study

(absorption spectra are normalized at 300 nm and emission spectra are normalized to

their maxima). 81

Figure 4.4: EL Spectra of devices with various thicknesses of the red and green co-doped

layer. 82

Figure 4.5: EQE vs luminance of the devices with selectively blue doped emitting layers.

(Embedded are the EL spectra of the three devices driven at 5 mA/cm2.) 85

Figure 4.6: Transient PL decay of two FIrpic doped films. 85

Figure 4.7: Device lifetime of three blue devices with selectively doped blue emitting

layers. 87

Figure 5.1: Schematic energy level diagram of the materials used in this chapter (LUMO

and HOMO energy levels are labeled above and below rectangles, triplet energy levels

are indicated in parentheses). 92

Figure 5.2: J-V curves of hole-only and electron-only devices with various doping

concentrations of Ir(iprpmi)3. 93

Figure 5.3: Device performance of five blue OLEDs with various Ir(iprpmi)3 dopant

concentrations. (a) Current density vs. voltage, (b) EQE vs. current density, (c) EL

spectra at 5 mA/cm2. 95

xxi

Figure 5.4: Device performance of five WOLEDs with various Ir(iprpmi)3 dopant

concentrations. (a) EQE vs. current density; (b) EL spectra at 5 mA/cm2; (c) color shift

of the device with 9% Ir(iprpmi)3 at current densities from 0.05 to 20 mA/cm2. 99

Figure 5.5: Device lifetime tested at 5 mA/cm2 (WOLEDs EL spectra are in the inset).102

Figure 6.1: Schematic energy diagram for blue PhOLEDs. (The triplet energy is in

parentheses, and HOMO/LUMO energies are below and above the rectangles). 108

Figure 6.2: Efficiencies and lifetime performances of Device A1, A2 and A3. 109

Figure 6.3: Normalized LDI-TOF spectra of Device A3 with and without degradation.110

Figure 6.4: Normalized LDI-TOF spectra of the neat Ir(iprpmi)3 film in the linear mode.

114

Figure 6.5: TOF/TOF spectrum of the TAPC cation. 116

Figure 6.6: LDI-TOF spectra of the neat TAPC film and HATCN/TAPC bilayer. 118

Figure 6.7: Dissociation energy of bonds in the neutral TAPC and TAPC cation. 119

Figure 6.8: Dissociation energy of cracking reactions after cyclohexyl is opened in the

TAPC cation. The dissociation of 1 corresponds to fragments at 570 and 591, and that

of 2 corresponds to the peak at 583. 119

Figure 6.9: Resonant structures (up) of the TAPC cation and ring-opened TAPC cation and

HOMO (down) of TAPC and ring-opened TAPC. 121

Figure 6.10: LDI-TOF-MS spectra of device B1 before and after degradation. 123

xxii

Figure 6.11: LDI-TOF-MS spectra of Device B2 before and after degradation. 125

Figure 6.12: LDI-TOF-MS spectra of Device B3 before and after degradation. 126

1

Chapter 1 Background and Introduction

1.1. Introduction to White Organic Light Emitting Diodes

Ever since their discovery [1], great efforts have been made to develop organic light

emitting diodes (OLEDs) for display applications because OLEDs have superior properties

such as high color contrast, high brightness and power efficiency, mechanical flexibility

and light weight [2–7]. Numerous consumer products, including TVs and mobile phones,

with OLED displays have entered the consumer market. Moreover, as a possible next

generation solid-state lighting source, white OLEDs (WOLEDs) [8] have the advantages

in high-power-efficiency, large-area and flat-panel-form-factor applications [9–13].

Today, a WOLED has reportedly achieved a power efficiency of 120 lm/W, which is higher

than that of a typical fluorescent tube [14].

However, WOLEDs are still facing challenges such as high material and fabrication

costs; complex processing procedures, which typically involve high-vacuum deposition

methods; and lack of accurate theories of exciton formation, diffusion and energy transfer

in WOLEDs [15–17]. Moreover, material-degradation and device-operating-stability

issues further hinder the mass manufacturing of WOLEDs and limit their potential to

compete with LCDs for display applications and LEDs [5, 6, 18–20] for lighting

applications. Therefore, intensive studies are ongoing to achieve high-efficiency, color-

stable and long-lifetime WOLEDs for applications in lighting and displays.

2

1.2. Basics of OLEDs

1.2.1. Basic Device Physics

In an OLED, light is generated by the recombination of injected electrons and holes

in an active organic layer. A very basic device structure needs only one organic layer.

Transparent indium tin oxide (ITO) is typically used as an anode, and a low-work-function

metal, such as Ca and Al, is used as a cathode. Between the sandwich-like arrangement of

electrodes is the active organic layer, which can be either small molecules or polymers.

When a voltage is applied to the device, charge carriers are injected into the active organic

layer from the electrodes, namely holes from the anode and electrons from the cathode

[21].

As shown in the energy level diagram in Figure 1.1(a), for electron injection from

the metal cathode to the organic layer, the electron from the metal needs to overcome the

energy barrier between the metal’s Fermi level and the lowest unoccupied molecular orbital

(LUMO) level of the organic material. Similarly, for hole injection, the hole from ITO

anode needs to overcome the barrier between the ITO’s Fermi level and the highest

occupied molecular orbital (HOMO) level of the organic material. After injection, the

electrons and holes are transported across the organic layer by hopping through the LUMO

and HOMO levels, respectively, of the organic molecules. The recombination of these

injected holes and electrons in the organic layer results in either heat dissipation (non-

radiative recombination) or light emission (radiative recombination or

electroluminescence) characteristic of the organic material. The recombination region

depends on the magnitude of the injection barriers and the relative mobilities of holes and

3

electrons while the color of the emitted light is governed by the HOMO-LUMO energy

difference of the organic material.

Figure 1.1: (a) Energy level diagram of a single-layer OLED; (b) Modern OLED device architecture illustration.

In almost all OLED devices, a commonly adopted structure is a stack of multi-

functional organic layers (Figure 1.1(b)). Each of the layers performs a specific charge-

transport or light-emission function with the aim of achieving the highest

electroluminescent efficiency and the lowest possible voltage and power. In general, an

OLED device is fabricated by vapor deposition with which the entire stack of layers is

sequentially deposited layer by layer on a substrate. For a typical bottom-emitting OLED

(Figure 1.1(b)), the anode is indium tin oxide (ITO) pre-deposited on a glass substrate. The

organic layer stack comprises in sequence 1) a hole-injecting layer (HIL), 2) a hole-

transport layer (HTL), 3) an emitting layer (EML), 4) an electron-transport layer (ETL), 5)

4

an electron-injecting layer (EIL) and a cathode. The cathode is typically a metal with a low

work function, such as Al, Ca or Mg. The HIL and EIL, which are typically a mixture of

strong electron donors and acceptors, act as “Ohmic” buffer layers to facilitate hole and

electron injection from anode and cathode, respectively. The HTL and ETL, which are

typically weak electron donors and acceptors, respectively, serve as media for transporting

holes and electrons to the EML. Holes and electrons recombine at the light-emitting layer

to form excitons, which can decay radiatively or non-radiatively. To enhance radiative

recombination, the emitting layer is typically a dopant-host matrix in which the dopant,

present in various concentrations, is a highly fluorescent or phosphorescent organic

compound, and the host is an organic compound or mixture capable of transporting both

holes and electrons [22].

1.2.2. Fluorescence and Phosphorescence from OLEDs

Excitons are formed by the recombination of injected electron-hole pairs in an

OLED device. There are two spin states in an exciton: total spin S = 0 (singlet: anti-parallel

spin vectors with magnetic quantum number = 0) and total spin = 1 (triplet: parallel spin

vectors with a magnetic quantum number Î[-1, 0, 1]). When an electron in the ground state

is excited, it can follow two different paths: one leading to the singlet state and another to

the triplet state. In the first path, all of the energy is used for exciting the electron, whereas

in the second path, part of the energy is used to unpair the spin. So the triplet state is at a

lower energy level. Generally, excitons in an OLED are created in a ratio of 3:1, i.e., 75%

triplets and 25% singlets, due to spin statistics.

5

Figure 1.2: Spin states of electrons.

Fluorescence-based OLEDs utilize only singlet excitons for light emission because

the transition from the lowest singlet excited state to the ground state (also a singlet) is spin

allowed. The transition from the triplet excited state to the ground state is forbidden by

symmetry; therefore, all the triplet excitons are wasted. Consequently, the internal quantum

efficiency of fluorescent OLEDs is limited to 25% [23].

Phosphorescence-based OLEDs [24] utilize both singlet and triplet excitons for

light emission. In addition to the spin-allowed singlet-transition fluorescence, the

transition from the triplet excited state to the ground singlet state is allowed through spin-

orbital coupling. This is made possible by the use of heavy-metal complexes as dopants

(such as Ir(ppy)3). Therefore, the theoretical internal quantum efficiency is 100% [25].

6

Figure 1.3: Different photon-emitting mechanisms of OLEDs [26].

A Jablonski energy diagram is shown in Figure 1.3 to explain the light-emitting

mechanism in terms of molecular energy levels. The diagram illustrates various light-

emission and exciton energy-loss pathways. Other than fluorescent and phosphorescent

emission, a third light-emitting mechanism (delayed fluorescence) is also shown. Delayed

florescence occurs when triplet excitons are converted to singlets through reverse inter-

system crossing, a process which is highly dependent on the energy gap between the lowest

singlet excited state and the triplet state.

1.2.3. Energy Transfer and Quenching in OLEDs

Excitons, either singlet or triplet, formed by electron-hole recombination in an

OLED device can suffer non-radiative decay, which results in a loss in electroluminescence

efficiency. A typical loss mechanism is quenching by which the energy of the exciton is

transferred to a quencher in the vicinity of the exciton prior to its emission as fluorescence

7

or phosphorescence. For singlet excitons, the quenching process is long range and well

known as Förster resonance energy transfer (FRET), whereas for triplet excitons, the

energy transfer is short range and known as Dexter transfer [27].

Förster resonance energy transfer refers to the phenomenon that an excited donor

transfers energy (not an electron) to an acceptor through a non-radiative process (Figure

1.4). To allow energy transfer, the absorption spectrum of the acceptor must overlap the

fluorescence spectrum of the donor. Moreover, FRET relies on the distance-dependent

transfer of energy from a donor molecule to an acceptor molecule through dipole-dipole

interaction between donor and acceptor. When the conditions are ideal for FRET to occur,

no photons will be emitted, but rather the energy is transferred from the donor-excited

energy level to the acceptor molecule, thus resulting in a decrease in the density of excited

state donors and an increase in the density of excited state acceptors. A complete FRET

energy transfer would result in fluorescence from the acceptor and complete quenching of

the donor fluorescence.

Figure 1.4: The schematic diagram of Förster resonance energy transfer.

8

Dexter electron transfer is another mechanism though which an excited donor and

an acceptor exchange electrons to accomplish the non-radiative process. It is a process

whereby two molecules bilaterally exchange their electrons. The reaction rate constant of

Dexter electron transfer exponentially decays as the distance between these two parties

increases. The exchange mechanism typically occurs within 1 nm, much shorter than the

dipole-dipole interaction in FRET. Furthermore, the exchanged electron should occupy the

orbital of the other party, which means that the exchange energy transfer needs the overlap

of the donor and acceptor wave functions. By Dexter electron transfer mechanism, triplet-

triplet annihilation can occur when two triplets (D* and A*) react to produce two singlet

states, as indicated in Equation 1.1.

!∗ + $∗®! + $ (1.1)

Figure 1.5: The schematic diagram of Dexter electron transfer.

In WOLEDs, such energy transfer from host to guest molecules and between two

different guest molecules can be utilized to shape the white emission spectrum and improve

the electroluminescence efficiency. As in most WOLEDs, white emission is realized with

dopants capable of emitting different colors (e.g., red, green and blue). These color dopants

9

can be all incorporated in a single emitting layer that comprises one or more host materials.

Alternatively, these color dopants can be individually incorporated in separate emitting

layers in the OLED stack. Most device architectures reported in the literature were

designed to provide a mechanism to control the multiple pathways for energy and charge

transfer from the host to the dopants or from one dopant to another, all within an individual

emitting layer or in adjacent emitting layers. This is usually done by tuning the thicknesses

and dopant concentrations in the individual layers, and also by employing multiple doped

emitting layers with undoped interlayers sandwiched in between. To understand these

transfer processes, it is often necessary to model the exciton density and diffusion length

in OLED systems. Based on Fick’s second law for particle diffusion, the following

equation (Equation 1.2) is often used to relate the exciton density at a specific location in

the emitting layer to the exciton (or excited state) lifetime, where Lx is the diffusion length,

n0 is the exciton density at the interface where the electron-hole recombination occurs, D

is the diffusion constant and & is the excited-state lifetime [28].

' ( = '*+,-

./012ℎ4- = !&(1.2)

1.3. Performance Characterization of WOLEDs

The efficiency of an OLED is characterized by its external quantum efficiency

(hext), current efficiency (hL) and luminous efficiency (hp). The external quantum

efficiency (also known as EQE) is defined by the ratio of the number of photons emitted

by an OLED into the space outside of the OLED to the number of electrons injected. For a

typical planar OLED structure, the EQE is based on the total number of photons emitted

10

through the transmissive electrode into the viewing direction. The intrinsic quantum

efficiency (hint) is the ratio of the total number of photons generated inside the structure to

the number of electrons injected. It is the product of charge balance ratio (ge-h ≤ 1), the

fraction of emissive exciton states (hs-p) and the radiative decay efficiency (Φi). The

external efficiency is decreased by a light outcoupling factor (hout), which is the fraction of

photons that can escape the device and is limited by wave guiding in the device’s layers

and the substrate. The relation of these parameters is shown in Equation (1.3).

:;-< = :=><×:@A< = B;,C×:D,E×Φ=×:@A< (1.3)

The power efficiency (PE) is defined as the overall light output per consumed

electric power and is generally considered as the most important figure of merit for

WOLED performance. In addition, two extra parameters are needed to characterize the

color quality of WOLEDs: Commission Internationale de L’Eclairage (CIE) coordinates

and color rendering index (CRI). The CIE 1931 chromaticity diagram is shown in Figure

1.6. Along the curved boundary are monochromatic colors with wavelengths indicated in

nm. By mixing any two monochromatic colors in different proportions, any color with CIE

coordinates located between the two points can be generated. Point (0.33, 0.33) is

considered to be the “colorless” white light; however, a somewhat broad region around this

point can also be considered to be white. The black line in the diagram is defined as the

Planckian locus, which indicates the CIE coordinates that can be considered to be

variations of white colors. Each point has a correlated color temperature (CCT). The higher

11

the CCT is, the bluer the white color appears to human eyes. For most lighting applications,

the CCTs typically range from 2,700 to 5,000 K. Low-CCT or warm white light is suited

for residential lighting, whereas high-CCT or cold white light is more generally used in the

workplace.

Figure 1.6: The CIE chromaticity diagram.

Color rendering index is a parameter that is commonly used to characterize the

quality of lighting or how well the lighting matches a black-body radiation of a specific

color temperature. Color rendering index values range from 0 to 100: a CRI value above

80 is considered to be somewhat adequate for general lighting, and a CRI value greater

than 90 is considered to be excellent. WOLEDs tend to have high CRIs due to their

12

generally broad emission spectra that can be tailored with multiple emitters to more closely

resemble the black-body spectra, particularly those at low color temperatures.

OLED device lifetime t1/2 is typically defined as the time for the luminance output

from the device to drop to half of its original level. Equation 1.4 was first used by Van

Slyke [29] to provide a relationship between operating lifetime and output luminance. The

parameter L0 is the initial luminance of an OLED, t1/2 is the half-life time (time to ½ L0 at

constant current), which is inversely proportional to L0, and C is a constant.

4@×2G H = I(1.4)

This equation provides a strictly inverse relationship between the initial luminance

and half-life, which may only be valid over a narrow luminance range. More often, the

lifetime of an OLED is much shorter when it is operated at high current densities (i.e. high

luminance levels). For a more accurate lifetime projection, a modified equation (Equation

1.5) has been adopted, where n is the acceleration coefficient to account for a steeper rate

of luminance loss at higher luminance values. [30–32].

4@>×2G H = I (1.5)

1.4. Status of WOLED Development

For WOLEDs, the device structure tends to be much more complex due to the fact

that a single organic-molecule-based emitter generally cannot provide a sufficiently broad

spectral range to produce a white-color emission, which requires red, green and blue (RGB)

color components. Because of their nearly 100% IQEs, phosphorescent emitters are

expected to be used in high-efficiency WOLEDs. Currently, commercial WOLEDs use

13

phosphorescent emitters for green and red colors, and fluorescent emitters for blue owing

to the lack of stable phosphorescent blue emitters. To achieve high performance WOLEDs,

various device structures (Figure 1.7) have been adopted, including 1) a single EML with

multiple dopants [33–35], 2) multiple EMLs to improve color renditions [14, 15, 36–38],

3) hybrid (incorporating both fluorescent and phosphorescent emitters) WOLEDs [39–42]

for better device lifetime, and 4) tandem devices to increase lifetime and luminance output

[43, 44]. Some of these device designs can lead to a very complex light generation process

that involves charge- and energy-transfer processes among various molecular species in

their ground or excited states within an individual layer and/or between separate layers.

Figure 1.7: Various device layouts to realize white light emission. (a) vertically stacked OLEDs, (b) pixelated monochrome OLEDs, (c) single-emitter-based WOLEDs, (d) blue

OLEDs with down-conversion layers, (e) multiple-doped emission layers (EMLs), and (f) single OLEDs with a sub-layer EML design. (Reprinted with permission from ref. [7])

14

To improve the performance of WOLEDs, two [4, 45, 46] or three [43, 47, 48]

phosphorescent emitters are necessary. These phosphorescent emitters can be incorporated

into a single layer or distributed in multiple layers. The latter has the advantage of a wider

scope for optimizing the color quality and efficiency of WOLEDs, although the fabrication

process may be more complicated.

To date, most research interests are focused on WOLEDs with a multiple-emitting-

layer (multi-EML) structure. For multi-EML WOLEDs, one of the biggest challenges is to

manage the distribution of excitons among the two or more emitters to realize white

emission with a desired spectrum [49]. There are several approaches: 1) insert an interlayer

to block the undesirable energy transfers between adjacent emitting layers, 2) tune the

dopant concentrations and the individual layer thicknesses to either facilitate or reduce the

energy transfers, and 3) control the recombination location with a host material or a

combination of host materials with specific transport characteristics. Another issue with

multi-EML WOLEDs is color shift due to a shift in recombination zone with a varying

drive voltage. Hence, a careful design of the device layer architecture is needed to produce

a high-performance multi-EML WOLED with a good color quality and stability.

There are mainly four types of WOLED: all-fluorescent WOLED, all-

phosphorescent WOLED, hybrid fluorescent-phosphorescent WOLED, and thermally

activated delayed fluorescence (TADF) WOLED. These types are briefly described below.

15

1.4.1. All-Fluorescent WOLEDs

All fluorescent WOLEDs can be categorized into single-EML WOLEDs and multi-

EML WOLEDs. In a single-EML WOLED, there are two ways to achieve white-color

emission: 1) The EML is composed of a red/yellow fluorescent guest doped into a blue

fluorescent host. The concentration of the guest is typically below 1%. Due to the low

concentration of the guest, exciton energy transfer from the blue host molecules to the

red/yellow guest molecules is incomplete, which results in partial blue emission from the

host and red/green emission from the guest, thus leading to the realization of white

emission [50–52]. 2) A non-emitting material is used as the host with appropriate color

dopants, including a blue dopant. White emission can be obtained by adjusting the

concentrations of color dopants [53].

Single-EML WOLEDs are generally less efficient than multi-EML WOLEDs.

Chuen and Tao [50] reported a single-EML WOLED using 4-{4-[N-(1-naphthyl)-N-

phenylaminophenyl]}-1,7-diphenyl-3,5-dimethyl-1,7-dihydro-dipyrazolo [3,4-b;4’3-e]

pyridine (PAP-NPA) as a blue host and rubrene as a yellow/red dopant. The rubrene

concentration is only 0.5%. The detailed device structure is as follows: ITO/4,4'-bis[N-

(1naphthyl)-N-phenyl- amino]-biphenyl (NPB) (40 nm)/PAP-NPA:rubrene (20

nm)/TPBi(40 nm)/Mg:Ag. The WOLED has a maximum luminance efficacy of 2.92 lm/W

at 6.5 V and a current efficiency of 6.11 cd/A at 7.0 V with a CIE of (0.33, 0.33). Kim et

al. [53] used 9,10-Di(naphth-2-yl)anthracene (ADN) as a host, 4,4'-Bis(9-ethyl-3-

carbazovinylene)-1,1’-biphenyl (BCzVBi), 2,3,6,7-Tetrahydro-1,1,7,7,-tetramethyl-1H

,5H ,11H -10-(2-benzothiazolyl)quinolizino[9,9a ,1gh]coumarin (C545T), and 4-

16

(Dicyanomethylene)-2-tert-butyl-6-(1,1,7,7-tetramethyljulolidin-4-yl-vinyl)-4H-pyran

(DCJTB) for blue, green, and red emission, respectively, in a single-EML WOLED. The

device structure is as follows: ITO/NPB (70 nm)/ADN:7% BCzVBi:0.05% C545T:0.1%

DCJTB (30 nm)/Bphen (30 nm)/Liq (2 nm)/Al (1,200 Å). Optimizing energy transfer

between the guest emitters resulted in a maximum current efficiency of 9.08 cd/A and a

CRI of 82.

For multi-EML WOLEDs, the device architectures are more complex, and

performance optimization involves many factors, including material composition for each

EML layer, its placement with respect to other EML layers, and layer thicknesses. Ho et

al. reported a highly efficient all-fluorescent WOLED in 2007 [54]. They used a dual EML

comprised of 1) 1,4-di-[4-(N,N-diphenyl)amino]styryl-benzene (DSA-ph) as a blue-

emitting host and rubrene as a yellow-emitting guest and 2) a pristine DSA-ph layer

without any dopant. White emission with CIE of (0.32, 0.43) was obtained. Furthermore,

by adopting a p-i-n device architecture (ITO/p-HTL/1,1-bis((di-4-

tolylamino)phenyl)cyclohexane (TAPC) (50 nm)/MADN:0.2% Rubrene:3% DSA-ph:5%

NPB (10 nm)/MADN:3% DSA-Ph:5% NPB (5 nm)/BPhen (10 nm)/n-ETL/LiF (1 nm)/Al

(150 nm)) with TAPC as the hole-transport layer, the device operating voltage was greatly

reduced and a power efficiency of 9.3 lm/W at 1000 cd/m2 and 3.4 V was achieved.

Long WOLED lifetime has been demonstrated in all-fluorescent WOLEDs. Duan

et al. [55] reported a half- lifetime of 150,000 h at an initial brightness of 1,000 cd/m2.

Figure 9 shows the detailed device structure and lifetime performance. The unique feature

of this WOLED is that two blue EMLs are positioned adjacent to each other. Both EMLs

17

use ENPN (6,6�-(1,2-ethenediyl) bis(N- 2-naphthalenyl-N-phenyl-2-naphthalenamine) as

a blue emitter. The blue EML adjacent to the ETL (Alq3) utilizes neat α, β-ADN as the

host. The other blue EML contains a mixture of α, β-ADN and NPB as a host, where NPB

is intended to broaden the recombination region. The third EML layer comprises a yellow

emitter DDAF (3,11-Diphenylamino-7,14-diphenylacenaphtho[1,2-k] fluoranthene) in a

mixed α, β-ADN and NPB host. This WOLED with three EMLs reportedly has a lifetime

that is almost 40 times longer than that of a conventional WOLED.

Figure 1.8: Characteristics of a WOLED. a) device structure and b) luminance decay curve. The inset shows the lifetime versus initial luminance relationship. (Reprinted with

permission from ref. [56])

1.4.2. All-Phosphorescent WOLEDs

In 2009, Wang et al. [56] presented a high-efficiency WOLED that incorporated

two phosphorescent dyes in a single-EML WOLED with a device structure as follows:

ITO/NPB (40 nm)/4,4',4''-Tris(N-carbazolyl)triphenylamine (TCTA) (5 nm)/mCP:6.5%

(a) (b)

18

FIrpic:0.75% (fbi)2Ir(acac) (20 nm)/TAZ (40 nm)/LiF/Al. This study shows that the blue

emission originates from energy transfer (mCP to FIrpic), whereas the orange emission is

a result of direct exciton formation on (fbi)2Ir(acac) due to its low HUMO level that traps

holes. Such a WOLED yields a peak power efficiency of 42.5 lm/W and EQE of 19.3%.

Unipolar host materials such as the hole-transporting material mCP and electron-

transporting material 9,9’-spiro-bisilaanthracene (UGH4) have been used as host materials

of the emitter layers in WOLEDs. Because of their unipolar nature, the recombination

region is confined at the interface adjacent to the HTL or EML, which can lead to more

severe TTA and triplet-polaron quenching [57]. To overcome such a shortcoming, bipolar

host materials or mixed-host systems are used to widen the recombination region [58–61];

2,6-bis(3-(9H-Carbazol-9-yl)phenyl)pyridine (DCzPPy) is a typical bipolar host. The hole

and electron mobilities of DCzPPy are both around 10-5 cm2/V•s. Liu et al. [64] developed

a high-efficiency WOLED with a configuration of ITO/MeO-TPD:F4-TCNQ (100 nm,

4%)/TAPC (20 nm)/DCzPPy:FIrpic:(MPPZ)2Ir(acac) (8 nm, 25%:1%)/TmPyPB (45

nm)/LiF (1 nm)/Al (200 nm) where DCzPPy is the bipolar host. Such a WOLED shows a

power efficiency of 37.1 lm/W at 100 cd/m2 and 31.3 lm/W at 1,000 cd/m2.

Reineke et al. [14] reported a much improved WOLED based on a multi-EML

structure presented in Figure 1.9. Two blue EMLs (FIrpic in TCTA and FIrpic in TPBi, 2

nm each) are sandwiched between a red EML (Ir(MDQ)2(acac)-doped TCTA) and green

EML (Ir(ppy)3-doped TPBi). TCTA and TPBi are hole-transport and electron-transport

materials, respectively. In this WOLED, the recombination zone is confined at the

interface between two blue EMLs. By sandwiching the two blue EMLs with a red EML

19

and a green EML, more excitons can be harvested. Together with an outcoupling structure,

a power efficiency of 90 lm/W at 1,000 cd/m2 was obtained for the WOLED.

Figure 1.9: Energy level diagram of a WOLED. Solid lines correspond to HOMO and LUMO energies. The orange color marks intrinsic regions of the emission layer.

(Reprinted with permission from ref. [14])

Chang et al. [36] fabricated a WOLED with 4,4′-Bis(N-carbazolyl)-1,1′-biphenyl

(CBP) as a common host material for four individual EMLs (red, orange, green and blue).

The device structure is illustrated in Figure 1.10. The exciton recombination region is

located at the CBP/TPBi interface. By utilizing triplet energy conversion, FIrpic excitons

can efficiently transfer energy to green, orange and red dopants sequentially. Such Förster-

type energy transfer was found to have an efficiency of 90%, thus resulting in a WOLED

with an EQE of 20.4% at 5,000 cd/m2.

20

Figure 1.10: Device configurations of WOLEDs. The dopants employed are FIrpic for blue (B), Ir(ppy)2(acac) for green (G), Ir(BT)2(acac) for yellow (Y), and Ir(MDQ)2(acac)

for red (R). (Reprinted with permission from ref. [36])

1.4.3. Hybrid WOLEDs

Although a WOLED with phosphorescent blue emitters typically yields high device

efficiency, the device lifetime is relatively short. One main reason for the short lifetime is

the fast degradation of the blue phosphorescent materials. Moreover, the high triplet energy

level of such blue emitters requires host and transport materials with a large band gap and

a high triplet energy level. Hence, the device operating voltage would increase, which may

lead to more severe electrochemical and thermal degradation of the materials.

To overcome such shortcomings, hybrid WOLEDs with fluorescent blue and

phosphorescent red/green emitters have been widely studied. Using device-structure

engineering and materials selection, it has been proven that triplet excitons of fluorescent

emitters can transfer their energy to triplet states of red and green phosphorescent materials

[62–65]. Such triplet harvesting mechanisms can enhance device efficiency and prolong

device lifetime [66, 67].

21

The biggest challenge to realizing high-efficiency hybrid WOLEDs is separating

the two exciton-harvesting pathways (blue singlet and red/green triplet). This requires the

triplet energy level of the blue fluorescent material to be higher than the triplet energy level

of the red and green phosphorescent materials. If this condition is not met, a specifically

designed device structure (such as one in which a spacer is inserted to block Förster energy

transfer) is needed to alleviate triplet quenching by the blue fluorescent materials.

Liu et al. [68] reported a hybrid WOLED structure with an extremely long lifetime.

The detailed device structure is as follows: ITO/MeO-TPD:F4-TCNQ (100 nm, 4%)/NPB

(20 nm)/MADN:DSA-ph(20 nm, 7%)/Interlayer (3 nm)/Bebq2 :Ir(MDQ)2(acac) (9 nm,

5%)/Bebq2 (25 nm)/LiF (1 nm)/Al (200 nm). The blue fluorescent emitter DSA-ph was

doped into MADN, and the red phosphorescent emitter Ir(MDQ)2(acac) was doped into a

Bebq2. An n-type interlayer was inserted between the blue and red layers. By varying the

types of the interlayer material, it was found that when Bepp2 was used as the interlayer,

the hybrid WOLED had a half-lifetime of 30,000 h (Figure 1.11). However, the device

power efficiency is reported to be only 16.0 lm/W at 100 cd/m2.

22

Figure 1.11: (a) Device architecture and the energy level diagram of the hybrid WOLED. (b) Lifetime of the device with Bepp2 as the interlayer. (Reprinted with permission from

ref. [68])

To further improve device efficiency, high triplet fluorescent materials have been

studied to avoid triplet quenching and the use of an interlayer. N,N�-di-1-naphthalenyl-

N,N-diphenyl-[1,1�:4�,1�:4�,1�-quaterphenyl]-4,4�-diamine (4P-NPD) was first

used by Schwartz et al. [39] to fabricate a hybrid WOLED. 4P-NPD has a triplet energy of

2.3 eV (higher than red/yellow phosphorescent emitters) and a photoluminescence

quantum yield of 92%. Blue triplet harvesting can be realized with proper device

engineering. Sun et al. [69] reported a high-performance hybrid WOLED structure without

(a)

(b)

23

an interlayer between the fluorescent and phosphorescent EMLs. The detailed EML

structure is as follows: TCTA:4% Ir(MDQ)2(acac) (3.5 nm)/TCTA:8%Ir(ppy)2(acac) (5

nm)/TCTA:TmPyPB:4P-NPD(73%:25%:2%, 7 nm). Figure 1.12 illustrates the complete

device stack and device performance characteristics. 4P-NPD was doped into a mixed-host

of TCTA and TmPyPB. The mixed-host broadens the fluorescent blue emission region,

while the blue triplet energy can be transferred to the red and green emitters. The low

concentration of 4P-NPD also minimizes formation of non-luminescent triplet excited

states of 4P-NPD. As a result, such a WOLED has an EQE of 17.0% and a power efficiency

of 34.3 lm/W at 1,000 cd/m2.

Figure 1.12: Energy-level scheme for materials used in the hybrid WOLED, and exciton energy diagram of the EMLs. R, G, B, and Tm represent Ir(MDQ)2(acac), Ir(ppy)2(acac),

4P-NPD, and TmPyPB, respectively. (Reprinted with permission from ref. [69])

1.4.4. TADF WOLEDs

After a breakthrough research reported by Adachi et al. [69] in 2012, thermally

assisted delayed fluorescent (TADF) materials have been actively studied. Such materials

24

have a small singlet and triplet energy split (< 0.1 eV). Therefore, triplets can be thermally

activated and form a reverse intersystem crossing that gives rise to delayed fluorescence.

TADF materials provide a possible alternative method of fabricating high-efficiency and

long-lifetime WOLEDs. A great amount of TADF research [70–75] has been conducted

recently.

Zhang et al. [71] reported a high-efficiency and color-stable hybrid WOLED with

a blue TADF material 4,5-bis(carbazol-9-yl)-1,2-dicyanobenzene (2CzPN) and an orange

phosphorescent material (acetylacetonato)bis[2- (thieno[3,2-c]pyridin-4-

yl)phenyl]iridium(III) (PO-01). The device architecture is demonstrated in Figure 1.12.

2CzPN has a triplet energy level of 2.5 eV, which is higher than that of PO-01 (2.2 eV).

When the exciton recombination zone is designed to be located at the interface of the blue

and orange EMLs, blue triplets can efficiently transfer energy to the orange triplet states,

thus yielding a maximum EQE of 22.5% and a power efficiency of 47.6 lm/W.

25

Figure 1.13: Materials, energy-level scheme and exciton-energy transfer mechanism of a hybrid WOLED incorporating a blue TADF material. (Reprinted with permission from

ref. [71])

1.5. Device Stability and Degradation Mechanism of WOLEDs

Device degradation refers to drive-voltage increase and luminance reduction over

a device’s operation time. In general, there are two pathways of device degradation over

time: 1) extrinsic causes, which include material impurities, poor device encapsulation, etc.

and 2) intrinsic causes, which include device architecture and physical and chemical

degradation of the device. Device architecture usually determines the charge balance of

OLED devices. High concentrations of charges in a thin interface can typically cause

instabilities. On the other hand, material properties (such as glass transition temperature,

energy band gap, bond energy, etc.) can also greatly affect device stability. Compared to

monochrome OLEDs, WOLEDs suffer from two more device stability issues: 1) color shift

26

caused by different lifetimes of emissive materials and 2) possible recombination region

shift and exciton distribution change.

1.5.1. Instability of Blue Phosphorescent Materials

According to Universal Display Corporation (UDC, a leading OLED research

company), red, green and blue phosphorescent OLEDs have half-lifetimes of 900,000,

400,000 and 20,000 h, respectively. Phosphorescent WOLED panels have a half-lifetime

of 30,000 h [76]. The lifetimes of blue OLEDs are one order-of-magnitude shorter

compared to that of their red/green counterparts. Phosphorescent WOLEDs’ lifetime is thus

limited by blue phosphorescent materials. For example, the most commonly studied and

commercially available FIrpic blue phosphorescent material is very unstable. It has been

reported that FIrpic-based OLEDs have lifetimes ranging from minutes to approximately

100 hours [77–81]. In addition, FIrpic is unstable for hole transport [82].

To improve device lifetimes of high-efficiency phosphorescent WOLEDs, new

classes of blue phosphorescent materials with stable chemical and electrochemical

properties are in great need of development. Based on the structure–property relationship

of materials, a new series of Ir complexes with phenyl-imidazole ligands have recently

been studied. With such materials, device lifetimes reaching 10,000 h have been reported

[83–88]. For instance, in a recent report, a blue dopant tris[1-(2,6-diisopropylphenyl)-2-

phenyl-1H-imidazole]iridium(III) (Ir(iprpmi)3) has been studied and found to have a

significantly longer device lifetime compared to FIrpic [79].

27

1.5.2. MALDI-TOF-MS

It has been a difficult task to pinpoint the primary degradation pathways in OLED

systems. Nevertheless, it is now widely accepted that material degradation caused by

chemical reactions during device operation is one of the main reasons for device

degradation. There have been a few techniques reported for chemical analysis of OLEDs

to provide an insight into certain material degradation pathways. Such methods include

optical techniques such as infrared and Raman spectroscopy, surface analysis techniques

such as atomic force microscopy (AFM), depth-profiling techniques such as X-ray

photoelectron spectroscopy (XPS) and secondary ion mass spectrometry (SIMS) and

chemical analysis tools such as high-performance liquid chromatography (HPLC) coupled

with mass spectrometry (MS).

One of the most powerful and successful tools for chemical analysis of OLEDs is

matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-

TOF-MS), which is a technique used to analyze polymers and biomolecules in general [89,

90]. This technique can detect chemical compounds through their mass-to-charge ratio

even at very low concentrations. It has been demonstrated that MALDI-TOF-MS can be

utilized to analyze solid-state thin-film organic semiconductors, with the ability to study

positively or negatively charged ions of the materials and their fragments at high

resolutions.

Figure 1.14 illustrates a schematic of a MALDI apparatus. Samples are typically

dispersed over a large excess of matrix material and applied onto a metal plate. Short-

pulsed laser light (nitrogen laser light, wavelength: 337 nm) then irradiates the sample and

28

triggers desorption and ionization of the target and matrix materials. The ions formed are

then accelerated by a high voltage and enter into a flight tube where they are separated

according to their masses. The time-of-flight detector records the time-of-flight of the ions,

which are converted to mass-to-charge ratio after external or internal calibration. From the

mass-to-charge ratio, the molecular structures of the parent molecule and its fragments can

be deduced. Two types of a TOF detector are used in general: linear and reflectron. In the

linear mode, ions travel directly towards the linear detector; in the reflectron mode, ions

travel through an ion mirror (which is a series of evenly spaced electrodes onto which a

single, linear, electric field is applied) and reach the reflectron detector. A reflectron

corrects for the energy dispersion of ions leaving the source (ions of the same m/z ratio

with different starting kinetic energies), because ions with more kinetic energy penetrate

the reflectron more deeply and spend more time in it, thus compensating for the spread in

kinetic energy. This gives a substantial increase in the resolution of the TOF analyzer.

When analyzing OLED devices, matrix materials are not necessary due to the

abundance of host materials. By using LDI-TOF-MS technique, it has been reported that

in the FIrpic molecular structure, the ancillary picolinate ligand and fluorine-substituted

phenyl-pyridyl ligands are susceptible to photo-induced dissociation [81, 91].

29

Figure 1.14: Schematic of a MALDI-TOF-TOF-MS setup [81].

1.6. Objectives and Outline of the Thesis

Due to the complexity of multi-EML WOLEDs, 10 or more organic materials

(transport, host and guest materials) are typically required to form the layers, and each of

these layers has to be optimized to produce a desired function such that they can

collectively achieve a WOLED with the highest possible quantum efficiency and power

efficiency. Therefore, a deep understanding of device operating principles and mechanisms

is required. Moreover, device degradation mechanisms are still unclear for most OLED

systems. Chemical degradation can be found in almost every organic material used in an

OLED, including charge-carrier-transporting materials, emitters, host materials, etc.

Therefore, it is crucial to understand certain degradation pathways of modern OLED

devices (especially high-performance blue phosphorescent OLEDs) in order to develop

more stable materials and device architectures.

In Chapter 2, the experimental methods for the fabrication and characterization of

various WOLED devices are described.

30

In Chapter 3, a series of multi-EML WOLEDs based on FIrpic in a mixed-host are

investigated. A mixed-host system can help broaden an exciton recombination region and

improve device performance. By varying host types and device structures, device operating

principles in all-phosphorescent multi-EML WOLEDs will be discussed.

Chapter 4 is based on a published paper in Organic Electronics, in which a multi-

EML WOLED comprising two separate blue layers and an ultra-thin red and green co-

doped layer sandwiched in between was studied. Through a systematic investigation of

exciton confinement and various pathways for energy transfer among the hosts and

dopants, it was found that both the ultra-thin co-doped layer and two blue EMLs play a

vital role in achieving high device efficiency and controllable white emission.

In Chapter 5, charge carrier properties of a blue phosphorescent dopant Ir(iprpmi)3

is first studied. Ir(iprpmi)3 is found to be trapping holes when doped at a low concentration

and transporting holes when doped at a high concentration in the bipolar host material

DCzPPy. By varying the blue dopant concentration and controlling the recombination

region, blue and white OLEDs with EQEs over 20% have been achieved. The WOLED

exhibits high color stability over a wide range of luminance. Moreover, device lifetime has

also been improved compared to the common blue dopant FIrpic.

In Chapter 6, we investigate device degradation mechanisms of Ir(iprpmi)3-based

blue OLEDs using the LDI-TOF-MS technique. Materials with high triplet energy (> 2.7

eV) (TAPC, TCTA, TmPyPB and DCzPPy) were selected as host or transport materials.

By carefully and systematically comparing the LDI-TOF patterns of electrically/optically

31

stressed and controlled (unstressed) OLED devices, possible degradation pathways of each

material are proposed and discussed.

In Chapter 7, an overall conclusion of this study will be provided. Finally, future

work for further improving WOLED performance and probing device degradation

mechanisms is proposed.

32

References

[1] C.W. Tang and S.A. Vanslyke, Appl. Phys. Lett. 51, 913 (1987).

[2] G. Zhou, W.Y. Wong, and S. Suo, J. Photochem. Photobiol. C Photochem. Rev. 11,

133 (2010).

[3] M.C. Gather, A. Köhnen, and K. Meerholz, Adv. Mater. 23, 233 (2011).

[4] S.J. Su, E. Gonmori, H. Sasabe, and J. Kido, Adv. Mater. 20, 4189 (2008).

[5] T. Nakayama, K. Hiyama, K. Furukawa, and H. Ohtani, J. Soc. Inf. Disp. 16, 231

(2008).

[6] M. Mizukami, N. Hirohata, T. Iseki, K. Ohtawara, T. Tada, S. Yagyu, T. Abe, T.

Suzuki, Y. Fujisaki, Y. Inoue, S. Tokito, and T. Kurita, IEEE Electron Device Lett.

27, 249 (2006).

[7] S. Reineke, M. Thomschke, B. Lüssem, and K. Leo, Rev. Mod. Phys. 85, 1245

(2013).

[8] J. Kido, M. Kimura, and K. Nagai, Science 267, 1332 (1995).

[9] G. Cheng, Y. Zhang, Y. Zhao, Y. Lin, C. Ruan, S. Liu, T. Fei, Y. Ma, and Y. Cheng,

Appl. Phys. Lett. 89, 043504 (2006).

[10] Y.J. Tung, T. Ngo, M. Hack, J. Brown, N. Koide, Y. Nagara, Y. Kato, and H. Ito,

SID Symp. Dig. Tech. Pap. 35, 48 (2004).

[11] Y.S. Wu, S.W. Hwang, H.H. Chen, M.T. Lee, W.J. Shen, and C.H. Chen, Thin

Solid Films 488, 265 (2005).

[12] Y.C. Tsai and J.H. Jou, Appl. Phys. Lett. 89, 243521 (2006).

33

[13] F. So, J. Kido, and P. Burrows, MRS Bull. 33, 663 (2008).

[14] S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Lüssem, and K. Leo,

Nature 459, 234 (2009).

[15] H. Sasabe, J.I. Takamatsu, T. Motoyama, S. Watanabe, G. Wagenblast, N. Langer,

O. Molt, E. Fuchs, C. Lennartz, and J. Kido, Adv. Mater. 22, 5003 (2010).

[16] J.W. Park, D.C. Shin, and S.H. Park, Semicond. Sci. Technol. 26, 34002 (2011).

[17] A.R. Duggal, J.J. Shiang, D.F. Foust, L.G. Turner, W.F. Nealon, J.C. Bortscheller,

and P. Red, SID Symp. Dig. Tech. Pap. 35, 28 (2005).

[18] B.W. D’Andrade and S.R. Forrest, Adv. Mater. 16, 1585 (2004).

[19] I. Rabelo De Moraes, S. Scholz, B. Lüssem, and K. Leo, Org. Electron. 13, 1900

(2012).

[20] V. Sivasubramaniam, F. Brodkorb, S. Hanning, H.P. Loebl, V. van Elsbergen, H.

Boerner, U. Scherf, and M. Kreyenschmidt, J. Fluor. Chem. 130, 640 (2009).

[21] W. Brütting, S. Berleb, and A.G. Mückl, Org. Electron. 2, 1 (2001).

[22] C.W. Tang, S.A. Vanslyke, and C.H. Chen, J. Appl. Phys. 65, 3610 (1989).

[23] H. Sasabe and J. Kido, European J. Org. Chem. 7653 (2013).

[24] M.A. Baldo1, D.F. O'Brien1, Y. You, A. Shoustikov, S. Sibley, M. E. Thompson

and S. R. Forrest, Nature 395, 151 (1998).

[25] L. Xiao, Z. Chen, B. Qu, J. Luo, S. Kong, Q. Gong, and J. Kido, Adv. Mater. 23,

926 (2011).

[26] http://micro.magnet.fsu.edu/primer/java/jablonski/lightandcolor/index.html.

[27] J.R. Lakowicz, Principles of Fluorescence Spectroscopy (2006).

34

[28] J. Wünsche, S. Reineke, B. Lüssem, and K. Leo, Phys. Rev. B - Condens. Matter

Mater. Phys. 81, (2010).

[29] S.A. Van Slyke, C.H. Chen, and C.W. Tang, Appl. Phys. Lett. 69, 2160 (1996).

[30] C. F́ry, B. Racine, D. Vaufrey, H. Doyeux, and S. Ciǹ, Appl. Phys. Lett. 87, 213502

(2005).

[31] T.Y. Chu, J.F. Chen, S.Y. Chen, and C.H. Chen, Appl. Phys. Lett. 89, 113502

(2006).

[32] J.N. Yu, M.Y. Zhang, C. Li, Y.Z. Shang, Y.F. Lü, B. Wei, and W. Huang, Chinese

Phys. B 21, 83303 (2012).

[33] S. Chen, Q. Wu, M. Kong, X. Zhao, Z. Yu, P. Jia, and W. Huang, J. Mater. Chem.

C 1, 3508 (2013).

[34] B.W. D’Andrade, R.J. Holmes, and S.R. Forrest, Adv. Mater. 16, 624 (2004).

[35] H. Wu, J. Zou, F. Liu, L. Wang, A. Mikhailovsky, G.C. Bazan, W. Yang, and Y.

Cao, Adv. Mater. 20, 696 (2008).

[36] Y.L. Chang, Y. Song, Z. Wang, M.G. Helander, J. Qiu, L. Chai, Z. Liu, G.D.

Scholes, and Z. Lu, Adv. Funct. Mater. 23, 705 (2013).

[37] S. Chen, G. Tan, W.Y. Wong, and H.S. Kwok, Adv. Funct. Mater. 21, 3785 (2011).

[38] Y.H. Son, Y.J. Kim, M.J. Park, and J.H. Kwon, Dig. Tech. Pap. - SID Int. Symp.

44, 1438 (2013).

[39] G. Schwartz, M. Pfeiffer, S. Reineke, K. Walzer, and K. Leo, Adv. Mater. 19, 3672

(2007).

35

[40] Y.R. Sun, N.C. Giebink, H. Kanno, B.W. Ma, M.E. Thompson, and S.R. Forrest,

Nature 440, 908 (2006).

[41] G. Schwartz, S. Reineke, T.C. Rosenow, K. Walzer, and K. Leo, Adv. Funct. Mater.

19, 1319 (2009).

[42] X. Yang, S. Zheng, R. Bottger, H.S. Chae, T. Tanaka, S. Li, A. Mochizuki, and G.E.

Jabbour, J. Phys. Chem. C 115, 14347 (2011).

[43] H. Kanno, R.J. Holmes, Y. Sun, S. Kena-Cohen, and S.R. Forrest, Adv. Mater. 18,

339 (2006).

[44] X. Qi, M. Slootsky, and S. Forrest, Appl. Phys. Lett. 93, 193306 (2008).

[45] X.M. Yu, H.S. Kwok, W.Y. Wong, and G.J. Zhou, Chem. Mater. 18, 5097 (2006).

[46] Q. Wang, J. Ding, D. Ma, Y. Cheng, L. Wang, X. Jing, and F. Wang, Adv. Funct.

Mater. 19, 84 (2009).

[47] Y.S. Park, J.W. Kang, D.M. Kang, J.W. Park, Y.H. Kim, S.K. Kwon, and J.J. Kim,

Adv. Mater. 20, 1957 (2008).

[48] Q. Wang, J. Ding, D. Ma, Y. Cheng, L. Wang, and F. Wang, Adv. Mater. 21, 2397

(2009).

[49] Q. Wang and D. Ma, Chem. Soc. Rev. 39, 2387 (2010).

[50] C.H. Chuen and Y.T. Tao, Appl. Phys. Lett. 81, 4499 (2002).

[51] C. H. Chuen, Y. T. Tao, F. I. Wu, and C. F. Shu. Appl. Phys. Lett. 85, 4609 (2004).

[52] G. Li and J. Shinar, Appl. Phys. Lett. 83, 5359 (2003).

[53] N.H. Kim, Y.H. Kim, J.A. Yoon, S.Y. Lee, D.H. Ryu, R. Wood, C.B. Moon, and

W.Y. Kim, J. Lumin. 143, 723 (2013).

36

[54] M.-H. Ho, S.-F. Hsu, J.-W. Ma, S.-W. Hwang, P.-C. Yeh, and C.H. Chen, Appl.

Phys. Lett. 91, 113518 (2007).

[55] L. Duan, D. Zhang, K. Wu, X. Huang, L. Wang, and Y. Qiu, Adv. Funct. Mater.

21, 3540 (2011).

[56] Q. Wang, J. Ding, Z. Zhang, D. Ma, Y. Cheng, L. Wang, and F. Wang, J. Appl.

Phys. 105, 076101 (2009).

[57] X.H. Zhao, Z.S. Zhang, Y. Qian, M.D. Yi, L.H. Xie, C.P. Hu, G.H. Xie, H. Xu,

C.M. Han, Y. Zhao, and W. Huang, J. Mater. Chem. C 1, 3482 (2013).

[58] S. Gong, Y. Chen, J. Luo, C. Yang, C. Zhong, J. Qin, and D. Ma, Adv. Funct. Mater.

21, 1168 (2011).

[59] H. Huang, X. Yang, Y. Wang, B. Pan, L. Wang, J. Chen, D. Ma, and C. Yang, Org.

Electron. Physics, Mater. Appl. 14, 2573 (2013).

[60] B. Pan, B. Wang, Y. Wang, P. Xu, L. Wang, J. Chen, and D. Ma, J. Mater. Chem.

C 2, 2466 (2014).

[61] B. Liu, L. Wang, M. Xu, H. Tao, X. Xia, J. Zou, Y. Su, D. Gao, L. Lan, and J. Peng,

J. Mater. Chem. C 5870 (2014).

[62] G. Schwartz, S. Reineke, K. Walzer, and K. Leo, Appl. Phys. Lett. 92, 53311 (2008).

[63] G. Schwartz, K. Walzer, M. Pfeiffer, and K. Leo, Proc. SPIE 6192, 61920Q (2006).

[64] J. Ye, C.J. Zheng, X.M. Ou, X.H. Zhang, M.K. Fung, and C.S. Lee, Adv. Mater.

24, 3410 (2012).

37

[65] M.E. Kondakova, J.C. Deaton, T.D. Pawlik, D.J. Giesen, D.Y. Kondakov, R.H.

Young, T.L. Royster, D.L. Comfort, and J.D. Shore, J. Appl. Phys. 107, 014515

(2010).

[66] T. Ishisone, S. Seo, Y. Nonaka, T. Kawata, and N. Ohsawa, Dig. Tech. Pap. - SID

Int. Symp. 45, 762 (2014).

[67] J. Liu, Y. Liu, N. Liu, Y. Han, X. Zhang, H. Huang, Y. Lifshitz, B. Liu, L. Wang,

M. Xu, H. Tao, J. Zou, D. Gao, L. Lan, H. Ning, J. Peng, and Y. Cao, Sci. Rep. 4,

7198 (2014).

[68] N. Sun, Q. Wang, Y. Zhao, Y. Chen, D. Yang, F. Zhao, J. Chen, and D. Ma, Adv.

Mater. 26, 1617 (2014).

[69] H. Uoyama, K. Goushi, K. Shizu, H. Nomura, and C. Adachi, Nature 492, 234

(2012).

[70] B.S. Kim and J.Y. Lee, Org. Electron. Physics, Mater. Appl. 21, 100 (2015).

[71] D.D. Zhang, L.A. Duan, Y.L. Li, D.Q. Zhang, and Y. Qiu, J. Mater. Chem. C 2,

8191 (2014).

[72] B. Seong Kim, K. Soo Yook, J. Yeob Lee, B.S. Kim, K.S. Yook, and J.Y. Lee, Sci.

Rep. 4, 6019 (2014).

[73] J. Nishide, H. Nakanotani, Y. Hiraga, and C. Adachi, Appl. Phys. Lett. 104, 233304

(2014).

[74] T. Higuchi, H. Nakanotani, and C. Adachi, Adv. Mater. 27, 2019 (2015).

[75] Y.J. Cho, K.S. Yook, and J.Y. Lee, Sci. Rep. 5, 7859 (2015).

[76] Universal Display Corporation. http://http://www.udcoled.com. 2014.

38

[77] H. Lee, H. Ahn, and C. Lee, J. Inf. Disp. 12, 219 (2011).

[78] W. Sotoyama, T. Satoh, M. Kinoshita, M. Tobise, K. Kawato, T. Ise, H. Takizawa,

and S. Yamashita, SID Symp. Dig. Tech. Pap. 41, 556 (2010).

[79] K.P. Klubek, S.C. Dong, L.S. Liao, C.W. Tang, and L.J. Rothberg, Org. Electron.

Physics, Mater. Appl. 15, 3127 (2014).

[80] K. S. Yook, O. Y. Kim, and J. Y. Lee, Synth. Met. 161, 2677 (2012).

[81] I.R. De Moraes, S. Scholz, B. Lüssem, and K. Leo, Org. Electron. 12, 341 (2011).

[82] K.P. Klubek, C.W. Tang, and L.J. Rothberg, Org. Electron. Physics, Mater. Appl.

15, 1312 (2014).

[83] H. Yamamoto, J. Brooks, M.S. Weaver, J.J. Brown, T. Murakami, and H. Murata,

Appl. Phys. Lett. 99, 033301 (2011).

[84] K. Udagawa, H. Sasabe, C. Cai, and J. Kido, Adv. Mater. 26, 5062 (2014).

[85] N.C. Giebink, B.W. D’Andrade, M.S. Weaver, P.B. MacKenzie, J.J. Brown, M.E.

Thompson, and S.R. Forrest, J. Appl. Phys. 103, 044509 (2008).

[86] B. D’Andrade, P.B. Mackenzie, M.S. Weaver, and J.J. Brown, US Patent 8,557,399

(2013).

[87] C. Lin, A.B. Dyatkin, Z. Elshenawy, and W. Yeager, US Patent 8,580,402 (2013).

[88] C. Lin, P.B. Mackenzie, R.W. Walters, J.Y. Tsai, C.S. Brown, and J. Deng, US

Patent 7,902,374 (2011).

[89] R. Seifert, I. Rabelo De Moraes, S. Scholz, M.C. Gather, B. Lüssem, and K. Leo,

Org. Electron. Physics, Mater. Appl. 14, 115 (2013).

39

[90] I.R. De Moraes, S. Scholz, B. Lüssem, and K. Leo, Appl. Phys. Lett. 99, 53302

(2011).

[91] E. Baranoff, B.F.E. Curchod, J. Frey, R. Scopelliti, F. Kessler, I. Tavernelli, U.

Rothlisberger, M. Grätzel, and M.K. Nazeeruddin, Inorg. Chem. 51, 215 (2012).

40

Chapter 2 Experimental Methods and Materials

2.1. Vacuum Vapor Deposition Process

Vacuum vapor deposition is a method for coating a thin film or multiple layers of

thin films on a substrate in a vacuum chamber. A large range of materials can be processed

with this method, including metals, high-temperature inorganic materials such as

semiconductors, and low-temperature organic compounds such as low-molecular-weight

or small molecules. For deposition of OLED materials, the temperature required is

generally low, typically below 500 oC, and resistive heating with a suitable crucible is

commonly used.

Figure 2.1 shows a typical vacuum vapor deposition system. The chamber pressure

is on the order of 10-6 torr for vapor deposition, which can be readily achieved with a turbo

or cryo pump, backed by a rotary pump. For deposition of organic materials, the crucible

or boat can be made of pyrex glass or quartz, which can be electrically heated with a

tungsten or nichrome wire. For high-temperature materials, including metals, boats made

of thin tungsten, molybdenum or tantalum foils and crucibles made of graphite, alumina,

or boron nitride are commonly used. Common substrates for OLED devices are glass or

plastic plates, and the substrates are usually kept at ambient temperature to avoid growth

of crystalline films.

41

Figure 2.1: Basic design of a vacuum vapor deposition coating system.

The thin film deposition rate is monitored by a quartz crystal microbalance (QCM)

sensor along with the necessary deposition controller. For OLED materials, the rate is

typically on the order of a few Å/s for the deposition of a single component film. For

deposition of multicomponent films, multiple QCM sensors with independent controllers

are required, and the deposition rate for each material component, which can vary from

below a tenth to several tens of Å /s, must be controlled precisely to produce a film of pre-

determined composition. The deposition rate is dependent on the source temperature which

can be manually or automatically controlled with the deposition controller. A shutter in

between the source and the substrate provides a convenient means of controlling precisely

the thickness of the film deposited on the substrate.

42

2.2. Boat Design and Coater Specifications

WOLED is a multi-layer device comprising a stack of thin organic films, some of

which contain two or more material components, such as the host and dopant in the emitter

layer. Co-deposition of two or more organic materials of various concentrations are needed,

where the concentration can be from as low as less than 1% to over 50%. Therefore,

independent rate control for all material components must be as precise as possible. To

facilitate deposition of multicomponent films in a single deposition chamber of confined

space, a customized boat configuration is required.

Figure 2.2: Design, components and boats for organic and inorganic material deposition.

43

The boat design for the deposition of organic materials was previously developed

by S. Lee of our laboratory. The design, as shown in Figure 2.2, has four parts: a glass test

tube, a coil of nichrome wire, a Macor base, and a pair of electrical contact pins. The glass

test tube (10.3 mm diameter) is cut by a Dremel rotary tool to form an open-sided cylinder

of 4.3 cm. A small side hole (~1 mm diameter) is drilled on the test tube body. The

nichrome wire (22 BNC, 0.0253” nominal diameter) is coiled to a diameter about 10 mm

and tightly fit into the glass test tube. Two holes are drilled in the Macor ceramic base in

order to hold two ends of the nichrome wire. The two ends are clamped with contact pins.

An epoxy resin is applied around the Macor base and the contact pins to bind the four parts

together. An aluminum foil is wrapped around the glass test tube to better conserve heat

and improve deposition rate stability. In general, only about 10 W electrical power is

sufficient to evaporate most organic materials.

The temperature required for most inorganic materials used in OLED devices such

as LiF, MoOx and Cs2CO3 can be as high as above 1000 ºC. Instead of pyrex glass tubes,

Macor, a ceramic material with temperature tolerance up to 1000C, was used for

constructing the boat. A 7/8” diameter Macor rod is machined into a hollowed cylinder of

a dimension similar to the glass cylinder. The bottom of the boat is not drilled through (~

5 mm) in order to hold source materials. Two holes are drilled on the bottom of the tube

for insertion of the nichrome wire, and one hole (~1 mm diameter) is drilled on the side of

boat body for rate monitoring. An aluminum foil is also wrapped around the boat body to

better conserve heat.

44

Figure 2.3: Boat assembly and sensor configuration. (a) Boats and sensors alignment, (b) graphical top view of the boats assembly.

A compact and multi-boat assembly is used for the fabrication of WOLEDs. Figure

2.3(a) shows the arrangement of the boats and the sensor positions. Because of the side

hole on the boats, QCM sensors can be placed on the side of the boats rather than on top,

providing a convenient way of monitoring the individual deposition rates of multiple

materials during co-deposition without cross-talks. Four QCM sensors (two facing back

(a)

(b)

45

and two facing front), are fixed in positions for monitoring up to four material depositions

at the same time. All of the boats (up to 16 boats) are mounted onto a movable aluminum

stage, which can slide from side to side. The boats are evenly distributed on the stage. With

this arrangement, multiple co-depositions can be done without the need to break vacuum

or rearrange the boat positions. The QCM sensors are water cooled to minimize possible

errors caused by radiant heating from the boats.

Figure 2.3(b) is a schematic top view of the boat arrangement. As an example, with

the positions of the four sensors as shown, a thin film with a composition of a blue emitter

doped into a mixed-host of TCTA and DCzPPy can be made by co-evaporation of the three

components. Moving the stage laterally to the right, a red and green co-doped TCTA film

can be achieved. Likewise, MoOx can be doped into TAPC and Cs2CO3 can be doped into

BPhen. It can be seen that this translational boat assembly is quite flexible and well-suited

for fabrication of complex OLED devices, including WOLEDs.

A cryo-pump (Cryo-torr 8) is used with the vacuum chamber. It provides a base

pressure of 5 *10-6 torr within 1.5 hours. The QCM sensors and the power supplies for the

boat sources are controlled using a computer with installed deposition software (INFICON

SQS-242).

2.3. Device Fabrication Conditions

All OLED devices were fabricated on patterned indium-tin-oxide (ITO) coated

glass substrates purchased from Tinwell Electronic Technology Company. The substrate

size is 2 inch by 2 inch. The ITO thickness is 110 nm with a sheet resistance of 15 Ω/sq.

46

The ITO pattern is shown in Figure 2.4. There are 12 narrow ITO stripes and 4 wide ITO

stripes on each substrate. Up to 6 OLEDs with different layer configurations can be

fabricated on one substrate (2 identical OLEDs per layer configuration). This is achieved

by two movable metal shutters beneath the substrate, which control the area of the substrate

that is open to film deposition. All vapor depositions were carried out at a base pressure of

10−6 torr (without breaking vacuum). For host and transport materials, the rate was set to

be around 4 Å/s. For organic dopants, the rate was generally below 1 Å/s depending on the

actual doping percentage. For inorganic materials, both MoOx and LiF rates were set to be

0.5 Å/s. Cs2CO3 deposition rate was below 0.5 Å/s depending on doping percentage.

Aluminum deposition was done using an Alumina coated tungsten boat (from R.D.

Mathies) and was manually controlled with a rate of 10~20 Å/s, up to a total thickness of

1000 Å.

Prior to film deposition, the glass substrates were cleaned in deionized water and

organic (acetone and ethanol with volume proportion of 2:1) baths with ultra-sonication

sequentially. The cleaned substrates were then dried with N2, followed by an O2 plasma

treatment before loading into the vacuum chamber.

47

Figure 2.4: Configuration of ITO pattern on glass substrates.

Figure 2.5(a) shows the organic layers and the aluminum cathode deposited onto a

pre-patterned ITO glass substrate. The overlap of the narrow ITO stripe and aluminum

cathode is the active device area (0.2 cm * 0.5 cm, 0.1 cm2). Figure 2.5(b) shows the two

movable metal shutters that can slide left/right to control the open area of the substrate for

deposition. Such a design allows a maximum of six different layer configurations to be

completed on one substrate, thus offering high throughput productivity in device

fabrication as well as reproducibility in device characteristics by minimizing fabrication

process variables such as chamber conditions and substrate differences. As shown in

48

Figure 2.5(c), by varying just one parameter (blue dopant concentration), multiple WOLED

devices (different color emission) can be fabricated on a single substrate.

Figure 2.5: (a) Photo of fabricated OLED on a ITO coated glass substrate. (b) Two pieces of metal shutters that are movable to fine control deposition conditions on one

substrate. (c) Illuminated devices with different colors on one substrates.

2.4. Device and Material Characterization

Current density-voltage (J-V) data of the OLED devices were obtained using a

Keithley sourcemeter (Keithley 2400). A Photoresearch PR650 was employed to measure

the radiometric and photometric characteristics, including electroluminescent spectra, CIE

co-ordinates, power efficacy (lm/W), current efficiency in candela per ampere (cd/A),

external quantum efficiency in photon per electron (EQE) and other parameters.

Acquisition software was provided by Eastman Kodak Company. A typical device

characterization included stepping up the current density incrementally from 0.01 to 20

mA/cm2 and collection of EL data per each current step. The EL output was measured

normal to the substrate plane with the assumption that angular distribution was Lambertian.

Thanks to the precise deposition control that the coating system provides, devices

(b) (a) (c)

49

fabricated with the same device structure showed good performance reproducibility and

only yielded EQE/PE variations of less than 5%.

In Chapter 4, the UV-Vis absorption spectra were obtained using a Perkin Elmer

Lambda 900 spectrophotometer. Photoluminescence (PL) spectra were recorded on a

Hitachi F-4600 fluorescence spectrophotometer. For exciton transient measurements, a

Quanta-ray GCR-150-10 pulsed Nd:YAG laser with THG (third harmonic generation, 355

nm) output was used to excite the films. Transient PL signals were directed through a

monochromator at 450 nm and detected using a photomultiplier and a Tektronix TDS 3052

oscilloscope.

In Chapter 6, Gauss09 was used for density function theory (DFT) calculations at

b3lyp level with 6-31g(d) as basic set. The sum of the electronic and thermal enthalpies

was used to estimate bond dissociation energy. The calculation was done by Shou-Cheng

Dong at HKUST.

2.5. Device Lifetime Test

For device lifetime evaluation, a completed OLED device (with top electrode) was

transferred to a vacuum assembly after fabrication. In this process the device was exposed

to ambient atmosphere for about 30 s. The assembly was kept at a base pressure of 50

mTorr with a mechanical pump. Such a simple encapsulation method is suitable for OLED

with relatively short lifetime. The devices were driven with a constant current density of 5

mA/cm2 at room temperature.

50

2.6. LDI-TOF-MS Analysis

In Chapter 6, laser desorption/ionization time-of-flight mass spectrometry (LDI-

TOF-MS) analysis was carried out in order to get information about the possible

degradation mechanisms in OLEDs. The spectrometer is a Brüker Autoflex III MALDI-

TOF system. Aged and unaged (control) samples were analyzed under conditions that were

kept as constant as possible. Before loading the samples on a LDI sample plate, the

aluminum cathodes were removed by Kapton tape. The N2 laser frequency was set to be

100 Hz. A total of 500 spectra were acquired at each spot position. The detected mass range

was set to be between 30 and 1,500 Dalton. All data were obtained in positive reflector

mode. To reduce material degradation induced by the laser, the laser power was increased

from 30% of the built-in power step by step (usually by an increment of 5%). Below a

certain laser power threshold, ionization of the sample material could not occur and there

was no signal in the MS spectra. Above the threshold, the MS signals increased with laser

power, usually non-linearly. The MS signal intensities (counts of ions), of the prominent

species were typically adjusted to the order of 104 counts. LDI Mass Spectra were post-

calibrated using molecular mass peaks of materials used in devices as internal standards.

The TOF/TOF experiments were performed by Shou-Cheng Dong on a Brüker

UltrafleXtreme mass spectrometer at Hong Kong University of Science and Technology

(HKUST).

51

2.7. Materials

Table 2.1 lists the acronyms, chemical names, molecular structures, device

functions, HOMO and LUMO levels, and the triplet energy levels for all of the materials

used in this thesis. They are grouped according to the function of the materials.

52

Table 2.1: Materials used throughout this thesis. HOMO/LUMO/triplet energies were taken from literature.

Acronym Chemical Name Molecular Structure Function HOMO

(eV)

LUMO

(eV)

E(T1)

(eV)

MoOx

molybdenum(VI) oxide

- HIL - - -

HATCN

1,4,5,8,9,11-

hexaazatriphenylene-

hexanitrile

HIL - - -

TAPC

1,1-bis((di-4-

tolylamino)phenyl)cyclo

hexane

HTL 5.5 2.3 2.9

mCP 1,3-Bis(N-

carbazolyl)benzene

HTL 5.9 2.4 2.9

53

Acronym Chemical Name Molecular Structure Function HOMO

(eV)

LUMO

(eV)

E(T1)

(eV)

TCTA

4,4,4-tris(N-

carbazolyl)triphenylami

ne

HTM 5.7 2.3 2.9

TmPyPB 1,3,5-tri(m-pyrid-3-yl-

phenyl)-benzene

ETM 6.7 2.8 2.8

TPBi

1,3,5-tris(2-N-

phenylbenzimidazolyl)

benzene

ETM 6.2 2.7 2.74

BPhen 4,7-diphenyl-1,10-

phenanthroline ETL 6.4 3.0 2.6

54

Acronym Chemical Name Molecular Structure Function HOMO

(eV)

LUMO

(eV)

E(T1)

(eV)

DCzPPy 2,6-bis(3-(carbazol-9-

yl)phenyl)-pyridine

Bipolar host 6.68 2.78 2.78

Cs2CO3

cesium carbonate

- EIL - - -

FIrpic

bis(4,6-difluorophenyl-

pyridinato-N,C2)

picolinate- iridium(III)

Blue emitter 5.6 2.5 2.6

Ir(iprpmi)3

tris[1-(2,6-

diisopropylphenyl)-2-

phenyl-1H-

imidazole]iridium(III)

Blue emitter 4.8 2.2 2.66

55

Acronym Chemical Name Molecular Structure Function HOMO

(eV)

LUMO

(eV)

E(T1)

(eV)

Ir(ppy)3 fac-tris(2-phenyl-

pyridinato)-iridium(III)

Green

emitter 5.3 2.9 2.4

Ir(phq)2(acac)

bis(2-phenylquinoline)-

(acetylacetonate)-

iridium(III)

Red emitter 5.3 3.1 2.0

56

Chapter 3 White Organic Light-Emitting Diodes with

FIrpic in a Mixed-Host

3.1. Introduction

Traditional OLEDs and WOLEDs are typically composed of multiple organic

layers, including an HTL, an EML and an ETL [1, 2]. However, in such device structures,

excitons are generally confined in a thin interface in which charge carriers recombine.

Accumulated excitons and charge carriers at such a thin interface can lead to exciton

quenching and induce possible photochemical reactions, and thus reduce device

performance and lifetime [3–6].

With the intent of broadening the recombination region and eliminating interfaces

where charge carriers and excitons are densely accumulated, mixed-host systems have been

reported with improved charge balance, device efficiency and lifetime (especially for blue

phosphorescent OLEDs) [7–12]. Moreover, mixed-host blue EMLs also have been

incorporated into WOLEDs to achieve better device efficiency and stability [13–17].

In this chapter, we present a dual-EML WOLED with one mixed-host blue layer

and one red and green co-doped phosphorescent layer to achieve a warm white color with

a high power efficiency and low drive voltage. The effects of host material types, co-host

composition and dopant concentrations on device performance are studied to better

understand working mechanisms of WOLEDs with a mixed-host blue EML.

57

3.2. Results and Discussion

A p-i-n structure is adopted to help reduce the device operating voltage. A typical

structure of the WOLED is ITO (110 nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (30 nm)/mCP

(3 nm)/blue EML (4 nm)/red and green EML (4 nm)/BPhen (20 nm)/BPhen:Cs2CO3 (50%,

10nm)/Al(100nm). TAPC and BPhen can help confine excitons due to their relatively

higher triplet energy levels (TAPC has ET1 = 2.9 eV, which is higher than that of FIrpic

(2.62 eV); BPhen has ET1 = 2.5 eV, which is higher than that of Ir(ppy)3 (2.4 eV) and Ir(2-

phq)2(acac) (2.0 eV)). Figure 3.1(a) shows the energy level diagram of a dual-EML

WOLED device (numbers in parentheses indicate the triplet energy of each corresponding

material). Figure 3.1(b) shows a typical WOLED stack structure. Noticeably, the light blue

phosphorescent emitter (FIrpic) is doped into a mixed-host composed of a hole-

transporting material (TCTA) and an electron-transporting material (TPBi).

58

Figure 3.1: (a) Energy level diagram of all materials used in WOLEDs. (b) Device structure of a typical WOLED.

3.2.1. Effects of an mCP Buffer Layer

It is known that in OLED devices, there is a charge imbalance due to different

charge-carrier mobilities (hole mobility at the level of 10-3 cm2/(V•s), and electron mobility

in the range of 10-6~10-5 cm2/(V•s)). Such imbalanced charge carriers can limit device

efficiency as discussed in Chapter 1; therefore, we introduced a bi-layered hole-

transporting structure (TAPC + mCP) to improve the charge balance. The total thickness

of HTL is 30 nm. Device A1 had a neat TAPC as the HTL, and Device A4 had a neat mCP

as the HTL. A thin mCP layer was inserted between TAPC and EML1 for Device A2 and

A3 (3 nm and 5 nm respectively). The device structure was as follows: ITO (110

nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (30-x nm)/mCP (x nm)/EML1/EML2/BPhen (20

nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm). A 10 nm TAPC was doped with 40%

MoO3 as an HIL, whereas a 10 nm BPhen was doped with Cs2CO3 as an EIL. Fifteen

59

percent FIrpic was doped into a mixed-host of TCTA and TPBi with a weight ratio of 1:2

as EML1, and 1.5% Ir(2-phq)2(acac) and 5% Ir(ppy)3 were doped into a 4 nm TPBi as

EML2. Table 3.1 summarizes the device performance measured at a current density of 5

mA/cm2.

Table 3.1: EL performance of WOLEDs with the mCP buffer layer. ITO (110 nm)/TAPC:MoO3 (40%, 10 nm)/HTL (30nm)/TCTA:TPBi:FIrpic(28%:57%:15%, 4nm)/TPBi:Ir(2-phq)2(acac):Ir(ppy)3 (1.5%, 5%, 4 nm)/BPhen (20 nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm). (Measured at a current density of 5 mA/cm2)

Device TAPC (nm)

mCP (nm)

Voltage (V)

EQE (%)

PE (lm/W)

Luminance (cd/m2)

1931 CIE x

1931 CIE y

A1 30 0 3.73 10.6 20.3 1209 0.417 0.436 A2 27 3 3.81 13.9 26.0 1907 0.346 0.473 A3 25 5 4.27 15.4 27.8 1933 0.372 0.456 A4 0 30 9.56 12.7 9.5 1452 0.359 0.429

With the introduction of mCP, due to its lower mobility (10-4 cm2/(V•s)) [18]

compared to TAPC (10-2 cm2/(V•s)) [19] and the 0.4 eV lower HOMO level, device drive

voltage increased from 3.73 V for Device A1 to 9.56 for Device A4. Device EQEs also

increased from 10.6% for Device A1 to 15.4% for Device A3, thus achieving an overall

improvement of power efficiency from 20.3 lm/W to 27.8 lm/W. For Device A4, due to

the much higher drive voltage, PE dropped to only 9.5 lm/W. From CIE values, the four

devices exhibited different white color (a blue shift from A1 to A3). Such a color shift

indicated that the recombination region had shifted towards EML1 when mCP was inserted

between TAPC and the blue EML1. Thus, it can be concluded that the mCP layer works

as a buffer layer for holes transporting. With a proper thickness (such as 3 nm), better

60

charge-carrier balance throughout the devices and, hence, higher device efficiencies can be

achieved.

3.2.2. Effects of Host Types for FIrpic

To better study the charge-carrier recombination region, different compositions of

HTM and ETM were used as a mixed-host for FIrpic. FIrpic concentration was fixed at

15%. TCTA and TPBi were chosen as a mixed-host for FIrpic, with the following device

structure: ITO (110 nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (27 nm)/mCP (3 nm)/

TCTA:TPBi:FIrpic (x:y, 15%, 4nm)/TPBi:Ir(2-phq)2(acac):Ir(ppy)3 (1.5%:5%, 4

nm)/BPhen (20 nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm). From Device B1 to Device

B4, the TCTA:TPBi ratio in EML1 was 0:1, 1:5, 1:2 and 1:1, respectively. Detailed device

performance at 5 mA/cm2 is summarized in Table 3.2. In the absence of TCTA (Device

B1), the recombination region is confined to the HTL/EML1 interface. Excitons are

severely quenched, leading to a much lower EQE of 9.9%. In the spectrum (Figure 3.2),

the blue peak dominates, with a low red emission being observed, which mainly comes

from exciton diffusion and energy transfer from FIrpic to Ir(2-phq)2(acac). As the TCTA

concentration in the TCTA:TPBi mixture increases, the device drive voltage decreases;

EQE increases for Devices B2, B3 and B4. More green and red emissions are also observed

in the white spectra. This can be attributed to the shallow HOMO level (5.7 eV) and high

mobility of TCTA that enable the transfer of holes to the EML1/EML2 interface, where

holes can be easily captured by Ir(2-phq)2(acac) and Ir(ppy)3 and form excitons due to their

low energy HOMO levels.

61

Figure 3.2: (a) EL spectra of Devices B1, B2, B3 and B4. (b) EQE vs. luminance vs. PE of Devices B1, B2, B3 and B4. (Measured at a current density of 5 mA/cm2)

A well-balanced mixed-host composition in EML1 can evidently broaden the

exciton generation region and alleviate quenching due to charge accumulation in thin

interfaces. With TCTA as a component of EML1, holes are more readily transported to the

EML1/EML2 interface to directly form green and red excitons. Hence, excitons are more

efficiently used for light output, leading to higher EQEs of WOLEDs.

Table 3.2: EL performance of WOLEDs with a mixed-host for FIrpic. ITO (110 nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (27 nm)/mCP (3 nm)/TCTA:TPBi:FIrpic (x:y, 15%, 4nm)/TPBi:Ir(2-phq)2(acac):Ir(ppy)3 (1.5%, 5%, 4 nm)/BPhen (20 nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm). (Measured at a current density of 5 mA/cm2)

Device TCTA: TPBi

Voltage (V)

EQE (%)

PE (lm/W)

Luminance (cd/m2)

1931 CIE x

1931 CIE y

B1 0:1 4.16 9.89 17.1 1138 0.353 0.439 B2 1:5 4.14 11.8 21.5 1355 0.384 0.431 B3 1:2 3.94 13.9 26.1 1633 0.375 0.438 B4 1:1 3.85 15.2 29.1 1784 0.431 0.446

62

We then employed TmPyPB as another ETM to replace TPBi and studied charge

transport of the mixed-host layer. The detailed device structures are as follows: ITO

(110nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (27 nm)/mCP (3 nm)/TCTA:TmPyPB: FIrpic

(x:y, 15%, 4nm)/TPBi:Ir(2-phq)2(acac):Ir(ppy)3(1.5%:5%, 4 nm)/BPhen (20

nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm). From Device C1 to Device C4, the

TCTA:TmPyPB ratio in EML1 ranges was 0:1, 1:5, 1:2 to 1:1, respectively. The device

performance at a current density of 5 mA/cm2 is summarized in Table 3.3. EQEs were

greatly improved with the increase in TCTA composition, with more red and green

emissions being observed in EL spectra (Figure 3.3). This finding is in agreement with the

previous study in which TPBi was used as the ETM.

Noticeably, when only TPBi was used as the host for FIrpic (Device B1), the EQE

was 9.9%. With TmPyPB as the only host material for FIrpic (Device C1), the EQE

dropped to 5.9%. No TCTA was present in both devices; therefore, the highly concentrated

FIrpic (15%) transported holes to some extent due to its HOMO level being the same as

that of mCP (5.9 eV). Electrons had only one path which was to be transported by TPBi or

TmPyPB in each device configuration. Thus, the HTL/EML1 interface was the

recombination region. The formed TmPyPB anions (TmPyPB-) or TPBi anions (TPBi-)

have a probability of quenching FIrpic, thus leading to a decreased EQE. Comparing the

performances of Device B1 and C1, TmPyPB anions appear to be to a more efficient

exciton quencher than TPBi anions.

63

Figure 3.3: (a) EL spectra of Devices C1, C2, C3 and C4. (b) EQE vs. luminance vs. PE of Devices C1, C2, C3 and C4. (Measured at a current density of 5 mA/cm2)

Table 3.3: EL performance of WOLEDs with a mixed-host for FIrpic. ITO (110 nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (27 nm)/mCP (3 nm)/TCTA:TmPyPB:FIrpic (x:y, 15%, 4nm)/TPBi:Ir(2-phq)2(acac):Ir(ppy)3 (1.5%, 5%, 4 nm)/BPhen (20 nm)/BPhen:Cs2CO3 (50%,10 nm)/Al (100 nm). (Measured at a current density of 5 mA/cm2)

Device TCTA:

TmPyPB Voltage

(V) EQE (%)

PE (lm/W)

Luminance (cd/m2)

1931 CIE x

1931 CIE y

C1 0:1 4.18 5.9 9.2 610 0.306 0.383 C2 1:5 4.10 8.8 13.9 908 0.357 0.386 C3 1:2 3.74 12.8 23.4 1414 0.406 0.422 C4 1:1 3.63 14.9 28.6 1654 0.440 0.433

The benefit of mixing HTM and ETM as a host for FIrpic to alleviate exciton

quenching can be realized by adopting a single bipolar host, such as DCzPPy. We

fabricated WOLEDs D1, D2 and D3 with DCzPPy as the universal host for all three

primary dopants. The device structure is as follows: ITO (110 nm)/ TAPC:MoO3 (40%, 10

nm)/TAPC (30 nm)/mCP (x nm)/TCTA:DCzPPy:FIrpic (y:z, 15%, 4 nm)/DCzPPy:Ir(2-

phq)2(acac):Ir(ppy)3 (1.5%, 5%, 4 nm)/BPhen (20 nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al

(100 nm). For Devices D1 and D2, a thin mCP layer was inserted between TAPC and

64

EML1 to balance charge carriers. For Devices D2 and D3, only DCzPPy was used as the

host material for FIrpic, whereas for device D1, mixed TCTA:DCzPPy host (1:2) was used

to control the EL spectra while reducing the drive voltage. Table 3.4 summarizes device

performance at a current density of 5 mA/cm2. For devices with the mCP buffer layer, EQE

was about 20% higher than that of those without mCP (see Devices D2 and D3). Although

the device drive voltage went up by 0.15 V at 5 mA/cm2, the overall PE improved from

21.9 lm/W (Device D3) to 25.3 lm/W (Device D2). With the introduction of TCTA in

Device D1, more holes pass through EML1 and get trapped by red and green dopants.

Therefore, red and green emission peaks appear more prominent in the EL spectra. Due to

the reduced drive voltage, the PE reaches 29.4 lm/W at 5 mA/cm2 with a warm white color.

Compared with the mixed-host of TCTA and TPBi/TmPyPB, the mixed-host of TCTA and

bipolar material DCzPPy balanced devices’ charge carriers more effectively, and a higher

EQE was achieved (~20% EQE). However, there was almost a 1 V drive voltage increase

at 5 mA/cm2 for Device D1, which inhibited further improvement of the PE. The increased

voltage can be contributed to two factors: 1) DCzPPy has a higher LUMO level (2.56 eV)

compared to that of TmPyPB, and a deeper HOMO level (6.05 eV) compared to that of

TCTA. Therefore, electrons and holes both experience higher energy barriers. 2) Charge

carrier mobilities of DCzPPy (hole and electron mobilities in the order of 10-5 cm2/(V•s)

[19]) are lower than those of typical unipolar hosts (such as TCTA and TmPyPB). Figure

3.4(b) illustrates the luminance-EQE-PE curves of Devices D1, D2 and D3. The

TCTA:DCzPPy mixed-host in Device D1 reduced efficiency roll-off at higher current

65

density, mainly due to an improved charge-carrier balance and a broadened recombination

region.

Figure 3.4: (a) EL spectra of Devices D1, D2 and D3. (b) EQE vs. luminance vs. PE of Devices D1, D2 and C3. (Measured at a current density of 5 mA/cm2)

Table 3.4: EL performance of WOLEDs with a mixed-host for FIrpic. ITO (110 nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (30 nm)/mCP (x nm)/TCTA:DCzPPy:FIrpic (y:z, 15%, 4nm)/DCzPPy:/Ir(2-phq)2(acac):Ir(ppy)3 (1.5%, 5%, 4 nm)/BPhen (20 nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm). (Measured at a current density of 5 mA/cm2)

Device mCP (nm)

TCTA: DCzPPy

Voltage (V)

EQE (%)

PE (lm/W)

Luminance (cd/m2)

1931 CIE x

1931 CIE y

D1 3 1:2 4.76 18.5 29.4 2224 0.427 0.461 D2 3 0:1 5.20 18.1 25.3 2098 0.346 0.433 D3 0 0:1 5.05 15.1 21.9 1760 0.350 0.436

3.2.3. Effects of Red Dopant Concentration

Layer EML2 seems to be critical not only for the overall EL spectra, but also device

EQEs. The concentrations of red and green dopants determine the efficiency of energy

transfer from FIrpic to Ir(ppy)3 and Ir(2-phq)2(acac), and also from Ir(ppy)3 to Ir(2-

66

phq)2(acac). Moreover, the low-lying HOMO levels of these two dopants can efficiently

trap holes at increased dopant concentrations. We fixed the green dopant’s concentration

at 6% and varied the red dopant’s concentration for Devices E1, E2 and E3 (1%, 1.5% and

2%, respectively) to study the energy transfer among the three emitters. The thickness of

EML2 (6 nm) was made slightly larger than that of EML1to achieve better control of the

low dopant concentration. The detailed device structures are as follows: ITO (110 nm)/

TAPC:MoO3 (40%, 10 nm)/TAPC (30 nm)/mCP (3 nm)/ TCTA:DCzPPy:FIrpic

(28%:57%:15%, 4 nm)/TPBi:Ir(2-phq)2(acac):Ir(ppy)3 (x%, 6%, 6 nm)/BPhen (20

nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm).

Figure 3.5: (a) EL spectra of Devices E1, E2 and E3. (b) EQE vs. luminance vs. PE of Devices E1, E2 and E3. (Measured at a current density of 5 mA/cm2)

As demonstrated in Figure 3.5(a), the devices’ EL spectra differed significantly

even with a small concentration variation of the red dopant. Device E3 had the smallest

amount of the red dopant; hence, the green emission dominates the EL spectrum. With the

67

slightly higher red dopant concentration of 1.5% for E2, the red emission surpassed the

green emission, indicating a more efficient energy transfer from the green dopant to the red

dopant. The blue emission at 474 nm remains almost unchanged. With the red dopant’s

concentration further increased to 2% for E1, the blue emission is suppressed, and the EL

spectrum is predominantly red. Moreover, a low green emission could be observed. This

finding indicates that with a higher red dopant concentration, energy transfer from both

FIrpic and Ir(ppy)3 to Ir(2-phq)2(acac) becomes more efficient.

Table 3.5 summarizes device performance at a current density of 5 mA/cm2. With

increased red dopant concentration, device drive voltage slightly increases from 4.46 V to

4.53 V. Device EQEs decrease from 14.4% to 12.2%. The combined effect of these two

factors is a significant drop in power efficiency from 26.1 lm/W (Device E3) to 16.2 lm/W

(Device E1). The change in drive voltage and EQE indicates that direct trapping of charge

that occurs at red dopant sites becomes stronger at higher red dopant concentrations.

Table 3.5: EL performance of WOLEDs with different red dopant concentrations. ITO (110 nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (30 nm)/mCP (3 nm)/TCTA:DCzPPy:FIrpic(28%:57%:15%, 4nm)/TPBi:Ir(2-phq)2(acac):Ir(ppy)3 (x%, 6%, 6 nm)/BPhen (20 nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm). (Measured at a current density of 5 mA/cm2)

Device Red:Green

Voltage (V)

EQE (%)

PE (lm/W)

Luminance (cd/m2)

1931 CIE x

1931 CIE y

E1 2%:6% 4.53 12.2 16.2 1167 0.504 0.394 E2 1.5%:6% 4.52 14.7 23.7 1705 0.415 0.442 E3 1%:6% 4.46 14.4 26.1 1852 0.372 0.478

68

The competition between energy transfer from FIrpic excitons to green/red dopants

and the internal energy transfer between green and red dopants largely depends on dopant

concentration in EML2. An optimal dopant concentration would not only improve the

white color quality but also increase a device’s efficiency.

3.2.4. The Role of a Non-Doped Interlayer

As discussed in Chapter 1, the concept of inserting interlayers between EMLs to

control triplet energy transfer has been utilized in multi-EML WOLEDs. Typically, the

introduction of an extra layer would cause an increase in device drive voltage and a

decrease in EQE. The extent of voltage increase and EQE drop depends on the thickness

and the charge transport property of the interlayer.

To study the effects of interlayers on device performance, we fabricated three

WOLEDs (F1, F2 and F3) with an interlayer inserted between EML1 and EML2. To

minimize a possible drive voltage increase, the interlayer material is also used as the host

material for the red and green dopants. Two materials were investigated as the interlayer,

namely TPBi (for Device F1) and DCzPPy (for Device F2). To further study the charge-

carrier transport property of DCzPPy, Device F3 was fabricated in which DCzPPy doped

with 15% FIrpic was the interlayer. For EML1, only TCTA was used as the host to control

charge in carrier recombination region. The thicknesses of both EML1 and EML2 were

fixed at 4 nm. The interlayer thickness was fixed at 2 nm. The device structures were as

follows: ITO (110 nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (30 nm)/ TCTA: FIrpic

69

(85%:15%, 4 nm)/Interlayer/Host:Ir(2-phq)2(acac):Ir(ppy)3 (1.5%, 5%, 4 nm)/BPhen (20

nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm).

Table 3.6 summarizes device performance at 5 mA/cm2. Devices F2 and F3 have

approximately 1 V higher drive voltage than Device F1 in which TPBi is used as the host

and interlayer instead of DCzPPy. Device EQEs of F2 and F3 are almost 30% higher than

that of Device F1. In Figure 3.6(a), EL spectra of the three devices indicate that the

intensities of red and green emissions are almost identical, while the blue emission is

weaker for F2 and F3, indicating that fewer excitons were generated in EML1. DCzPPy

and TPBi have triplet energies of 2.70 eV and 2.74 eV, respectively, which are both higher

than that of FIrpic (2.62 eV); hence, the exciton quenching effect should not be severe. The

change in exciton difference in EML1 for F1 and F2 could be attributed to the bipolar

transporting property of DCzPPy, in which holes can travel through the interlayer and get

trapped by red and green dopants, thus leading to more efficient use of excitons by direct

recombination. However, for Device F1, electron transporting material TPBi was used as

the interlayer, which led to carrier recombination being confined to the EML1/interlayer

interface. Therefore, more FIrpic excitons could be generated. Compared to F2, the

introduction of FIrpic to the interlayer for Device F3 does not affect the overall device

performance (EQE, EL spectra and PE remain almost unchanged). Such findings indicate

that the exciton recombination region is located at the EML1/interlayer interface rather at

the interlayer/EML2 interface for both device F2 and F3. Figure 3.6(b) illustrates an EQE

roll-off at high luminance. The EQE of Device F1 exhibits stronger roll-off. A possible

70

reason is that triplet-polaron quenching is more severe at the EML1/TPBi interface

compared to the EML1/DCzPPy interface.

Figure 3.6: (a) EL spectra of Devices F1, F2 and F3. (b) EQE vs. luminance vs. PE of Devices F1, F2 and F3. (Measured at a current density of 5 mA/cm2)

Table 3.6: EL performance of WOLEDs with interlayers. ITO (110 nm)/TAPC:MoO3 (40%, 10 nm)/TAPC (30 nm)/TCTA:FIrpic (85%:15%, 4nm)/Interlayer/Host:/Ir(2-phq)2(acac):Ir(ppy)3 (1.5%, 5%, 4 nm)/BPhen (20 nm)/BPhen:Cs2CO3 (50%, 10 nm)/Al (100 nm). (Measured at a current density of 5 mA/cm2)

Device Interlayer 2 nm

Voltage (V)

EQE (%)

PE (lm/W)

Luminance (cd/m2)

1931 CIE x

1931 CIE y

F1 TPBi 3.80 11.9 22.9 1389 0.387 0.442 F2 DCzPPy 4.73 15.4 25.4 1907 0.393 0.469 F3 FIrpic:DCzPPy 4.67 15.4 25.7 1907 0.390 0.478

3.3. Conclusions

A systematic and layer-by-layer study was conducted to optimize the performance

of a dual-EML WOLED structure. The hole-buffer layer mCP was found to balance charge

carriers and improve device efficiency. By varying the mixed-host compositions of the blue

layer and dopant concentrations of the green-red layer, the electron-hole recombination

71

zone could be controlled, and a balanced white emission was achieved. The optimized

WOLED structure exhibits a PE of 33 lm/W and an EQE of 18% at 1,000 cd/m2.

72

References

[1] B. Geffroy, P. le Roy, and C. Prat, Polym. Int. 55, 572 (2006).

[2] Y.-S. Tyan, J. Photonics Energy 1, 11009 (2011).

[3] D.Y. Kondakov, J.R. Sandifer, C.W. Tang, and R.H. Young, J. Appl. Phys. 93,

1108 (2003).

[4] T.D. Pawlik, M.E. Kondakova, D.J. Giesen, J.C. Deaton, and D.Y. Kondakov, J.

Soc. Inf. Disp. 17, 279 (2009).

[5] S. Reineke, K. Walzer, and K. Leo, Phys. Rev. B - Condens. Matter Mater. Phys.

75, (2007).

[6] S.J. Su, E. Gonmori, H. Sasabe, and J. Kido, Adv. Mater. 20, 4189 (2008).

[7] M.E. Kondakova, J.C. Deaton, D.Y. Kondakov, T.D. Pawlik, R.H. Young, C.T.

Brown, and D.J. Giesen, 2007 SID Int. Symp. 38, 837 (2007).

[8] S.W. Liu, C.A. Huang, J.H. Lee, K.H. Yang, C.C. Chen, and Y. Chang, Thin Solid

Films 453, 312 (2004).

[9] Z.L. Jiang, W. Tian, Z.Q. Kou, S. Cheng, and Y.H. Li, Opt. Commun. 372, 49

(2016).

[10] C.H. Hsiao, Y.H. Chen, T.C. Lin, C.C. Hsiao, and J.H. Lee, Appl. Phys. Lett. 89,

163511 (2006).

[11] S.H. Kim, J. Jang, K.S. Yook, and J.Y. Lee, Appl. Phys. Lett. 92, 023513 (2008).

[12] J.H. Lee, C.I. Wu, S.W. Liu, C.A. Huang, and Y. Chang, Appl. Phys. Lett. 86,

103506 (2005).

[13] J. Lee, J.I. Lee, and H.Y. Chu, ETRI J. 31, 642 (2009).

73

[14] T.C. Liao, H.T. Chou, F.S. Juang, Y.S. Tsai, L.A. Hong, and Y.Y. Ho, Curr. Appl.

Phys. 13, 152 (2013).

[15] Y.C. Tsai and J.H. Jou, Appl. Phys. Lett. 89, 133509 (2006).

[16] C.W. Seo and J.Y. Lee, Thin Solid Films 520, 5075 (2012).

[17] L.Y. Guo, X.L. Zhang, M.J. Zhuo, C. Liu, W.Y. Chen, B.X. Mi, J. Song, Y.H. Li,

and Z.Q. Gao, Org. Electron. Physics, Mater. Appl. 15, 2964 (2014).

[18] J. Wang, J. Liu, S. Huang, X. Wu, X. Shi, C. Chen, Z. Ye, J. Lu, Y. Su, G. He, and

Y. Zheng, Org. Electron. Physics, Mater. Appl. 14, 2854 (2013).

[19] Y.S. Tsai, L.A. Hong, F.S. Juang, and C.Y. Chen, J. Lumin. 153, 312 (2014).

74

Chapter 4 High Efficiency White Organic Light-

Emitting Diodes with an Ultra-Thin Red and Green Co-

Doped Layer and Dual Blue Emitting Layers

4.1. Introduction

WOLEDs are currently being utilized for both display and lighting applications.

Ever since their first demonstration, the research focus has been on improving the WOLED

efficiency, brightness, and lifetime. To produce high efficiency WOLEDs, phosphorescent

emitters are indispensable, as they provide a pathway of achieving emission with a nearly

100% internal quantum efficiency. Significant enhancement in efficiency has also been

realized in various device layer architectures, including a single-layer emitter with multiple

color dopants [1, 2], a multiple-layer emitter consisting of two or more adjoining EMLs [3-

7], and hybrid WOLEDs [8-10]. To obtain multi-fold improvements in both lifetime and

brightness, tandem structures are often implemented in WOLEDs at the expense of layer

complexity [11-12].

To date, most research interest in WOLEDs is focused on multi-EML structures

because they provide better control of the recombination and emission processes, enabling

a higher efficiency. From the perspective of device fabrication, it is much easier to adopt

an emitter structure in a WOLED consisting of two broadband EMLs producing

complementary blue-green and orange-red color layers. In contrast, WOLEDs with three

primary colors tend to produce white color with a better color rendering index [13-15].

75

Introduction of an extra layer to accommodate three emitters, however, makes it

challenging to manage interlayer charge-transport and energy-transfer between the various

hosts and dopants. Those processes not only control the emission efficiency and the color

balance, but also affect color-stability at various drive voltages [7, 16-21].

In this chapter, we describe a WOLED with a triple-layer emitter structure

consisting of an ultra-thin co-doped red and green layer sandwiched in between two blue

EMLs. By tailoring the doping concentration and layer thicknesses, we can control the

exciton energy transfer amongst the hosts and dopants. Our device structure produces

WOLEDs with an extremely high EQE (over 20%) and a power efficiency of 40 lm/W at

1000 cd/m2 and 3.7 V. At the same time, the color variation is minimal over a wide range

of emission intensities.

4.2. Results and Discussion

The WOLED structure for this study is as follows: ITO (110nm)/HATCN (3

nm)/TAPC (37 nm)/TCTA:FIrpic (4 nm)/red-green co-doped layer (0.5 nm)/ DCzPPy (4

nm)/TmPyPB:FIrpic (10 nm)/TmPyPB:Cs2CO3 (50%, 30 nm)/Al (100 nm). The thickness

of each individual layer was optimized to achieve the highest external quantum yield

possible without necessarily increasing the drive voltage.

To reduce the operating voltage, HATCN was deposited on top of pre-cleaned ITO

substrates as the hole injection layer. TAPC was chosen as the hole transporting material.

The electron transporting material was doped with cesium carbonate (Cs2CO3) to help

increase electron injection efficiency from the cathode. Hole-transporting material TCTA

76

and bipolar-transporting materials DCzPPy were chosen as the two host materials for the

blue emitter FIrpic. In between the two blue EMLs, a red emitter Ir(2-phq)2(acac) and a

green emitter Ir(ppy)3 were doped into an ultra-thin TCTA layer. The deposition rate of

each organic layer was monitored by quartz crystal sensors via a side aperture on the boats.

Selected because of their relatively high triplet energy levels, TAPC (2.9 eV) and TmPyPB

(2.78 eV) serve to confine the triplet excitons generated in FIrpic (2.62 eV), Ir(ppy)3 (2.4

eV) and Ir(2-phq)2(acac) (2.0 eV) within the EMLs. These triplet energy levels are

indicated (in parentheses) in the energy level diagram as shown in Figure 4.1(a), along with

the LUMO and HOMO energy levels (labeled above and below the rectangles) for the

sequence of layers from TAPC to TmPyPB. For clarity, the corresponding WOLED

configuration including the layer thicknesses and dopant concentrations is shown in Figure

4.1(b).

Figure 4.1: Energy level diagram and device architecture of a WOLED with an ultra-thin red, green co-doped emitting layer (LUMO and HOMO energy levels are labeled above

and below the rectangles, triplet energy levels are indicated in parentheses).

77

The injected holes enter the EMLs first through the blue (TCTA:FIrpic) layer and

then the red-green (TCTA:Ir(2-phq)2(acac):Ir(ppy)3) layer. Since both of these layers use

TCTA, a hole-transporting material, as the host, the majority of holes are expected to

traverse these two layers. The injected electrons enter the EMLs through the blue

(DCzPPy:FIrpic) layer, where DCzPPy, a bipolar-transporting material, is the host. As

shown in Figure1(a), the energy offsets between TCTA and DCzPPy are substantial (0.35

eV for HOMO and 0.16 eV for LUMO), providing a suitable interface to localize electron-

hole recombination. Due to this specific arrangement for the EMLs, the long-lived triplet

excitons formed as a result of these recombination events can effectively diffuse in the

TCTA and DCzPPy hosts and are subsequently redistributed between the blue, green and

red dopants commensurate with their concentrations in these hosts and their distance from

the TCTA/DCzPPy interface.

We fabricated four devices, B1/B2, B1/R&G/B2, B1/R/B2, and B1/G/B2, having

different composite EMLs. B1/B2 is a blue device with two different blue EMLs as the

composite emitter comprised of 15% FIrpic doped TCTA (B1) and 20% FIrpic doped

DCzPPy (B2). B1/R/B2, B1/G/B2 and B1/R&G/B2, are devices with three EMLs as the

composite emitter where an ultra-thin red, green or red and green (co-doped) EML is

inserted between the blue EMLs B1 and B2, respectively. The thickness of this interlayer is

only 0.5 nm and the dopant concentrations were adjusted to produce a balanced white

emission with high efficiency.

Figure 4.2 shows the plot of external quantum efficiency (EQE) versus current

density for the four devices. Table 1 summarizes the performance data at 5 mA/cm2. It can

78

be seen that all four devices exhibit high EQE ranging from 17.5% for the blue device

B1/B2 to 20.3% for predominately green device B1/G/B2. The drive voltages for these

devices are also very similar, approximately 3.8 ± 0.1 V (at 5 mA/cm2). The power

efficiency varies substantially due to a large variation in emission colors from these

devices, ranging from 31 lm/W for B1/B2 to 52 lm/W for B1/G/B2. The B1/R&G/B2 device

provides a warm white emission with color co-ordinates of (0.458, 0.448) that shift only

slightly over a luminance range of 400-4000 cd/m2. In contrast, device B1/R/B2 shows a

cool white emission with color ordinates of (0.382, 0.400) that vary marginally over the

same luminance range.

Figure 4.2: EQE vs current density of devices with four different ultra-thin layer doping conditions. (Embedded are the EL spectra of the four devices driven at 5 mA/cm2.)

79

The inset in Figure 4.2 shows the spectral response of these four devices driven at

a current density of 5 mA/cm2. The blue B1/B2 device exhibits only FIrpic emission with a

peak at 474 nm. Device B1/R/B2 with a red-doped interlayer shows a cold white color due

to the lack of green emission whereas the B1/G/B2 device with a green-doped interlayer

exhibits mostly green emission with a peak at 510 nm. Both devices retain blue FIrpic

emission due to incomplete energy transfer. It is worth noting that the FIrpic emission is

suppressed in the B1/G/B2 device compared to the B1/R/B1 device. This feature simply

indicates that energy transfer from blue FIrpic is more efficient to Ir(ppy)3 at a higher

concentration compared to Ir(2-phq)2(acac) at a much lower concentration. However, with

a red and green co-doped layer, the red emission from the B1/R&G/B2 device is enhanced

as a result of triplet energy transfer from the green to red dopants. This transfer is in

addition to the direct channel from FIrpic to the red dopant.

Table 4.1: EL Performance of devices with four different ultra-thin layer doping conditions. ITO (110nm)/HATCN (3 nm)/TAPC(37 nm)/TCTA:FIrpic (15%, 4nm)/TCTA:Ir(2-phq)2(acac):Ir(ppy)3 (x%:y%, 0.5 nm)/DCzPPy:FIrpic (20%, 3nm)/TmPyPB (10 nm)/TmPyPB:Cs2CO3 (50%, 30 nm)/Al (100 nm). (a: values at 5 mA/cm2; b: luminance range from 400 to 4000 cd/m2.)

Device Voltage (V)a

EQE (%)a

PE (lm/W)a

Luminance (cd/m2)a

1931 CIE xa

1931 CIE ya

1931 CIE Dxb

1931 CIE Dyb

B1/R&G/B2 3.91 18.5 34.80 2170 0.458 0.448 ±0.018 ±0.012 B1/R/B2 3.88 19.2 33.73 2082 0.382 0.400 ±0.028 ±0.003 B1/G/B2 3.84 20.3 51.71 3158 0.250 0.563 ±0.007 ±0.016 B1/B2 3.77 17.5 31.67 1899 0.155 0.373 ±0.001 ±0.002

To further understand the underlying sequential exciton energy transfer mechanism

from host to dopant and dopant to dopant, we measured the photoluminescence and

80

absorption of the hosts and dopants. Figure 4.3 shows the PL and absorption spectra of the

host and dopant materials. Both TCTA and DCzPPy have a PL peak centered around 400

nm, which overlaps with the absorption of FIrpic and Ir(ppy)3, indicating that energy

transfer from these two hosts to the blue and green dopants should be efficient. In contrast,

the energy transfer to the red dopant is inefficient as the absorption of Ir(2-phq)2(acac),

which centers at 440 nm and 520 nm, has little overlap with the host PL. The PL of FIrpic

and Ir(ppy)3 peaks at 470 nm and 520 nm, respectively and overlaps well with the

absorption of Ir(2-phq)2(acac), therefore the energy transfer from the blue and green

dopants to the red dopant can be efficient. Moreover, the overlap between FIrpic emission

and Ir(2-phq)2(acac) absorption is larger than that between FIrpic emission and Ir(ppy)3

absorption. Hence, the weaker blue emission in the B1/G/B2 device compared to B1/R/B2

device can mostly be attributed to the higher Ir(ppy)3 doping concentration (6%) compared

to the rather low Ir(2-phq)2(acac) doping concentration (2%).

81

Figure 4.3: Absorption and emission spectra of various materials used in this study (absorption spectra are normalized at 300 nm and emission spectra are normalized to

their maxima).

To investigate the effects of the red and green co-doped layer on the white emission

spectrum, we fabricated three devices where the thickness of the co-doped layer is 0.5, 0.75

and 1 nm respectively. As shown in Figure 4.4, it was found that EQE is practically

identical for all three devices (see Table 4.2 for detailed EL performance). Nonetheless,

increasing the thickness of the co-doped layer causes the red emission to increase relative

to the blue emission. This indicates increased energy transfer from FIrpic and TCTA

excitons to red and green dopants. Since the blue emission from FIrpic exciton formed at

the TCTA/DCzPPy interface must be balanced with the red and green emission from the

co-doped EML, the tri-layer design presented here where we can control both composition

82

and thicknesses is an extremely flexible architecture to engineer white emission with

specific color temperatures.

Table 4.2: EL Performance of devices with various thicknesses of the ultra-thin red and green co-doped layer (driven at 5 mA/cm2). ITO (110nm)/HATCN(3 nm)/TAPC (37 nm)/TCTA:FIrpic (15%, 4 nm)/TCTA:Ir(2-phq)2(acac):Ir(ppy)3 (2%:6%, x nm)/DCzPPy:FIrpic (20%, 3 nm)/TmPyPB (10 nm)/TmPyPB:Cs2CO3 (50%, 30 nm)/Al (100 nm).

Co-doped layer thickness (nm)

Voltage (V)

EQE (%)

PE (lm/W)

Luminance (cd/m2)

1931 CIE x

1931 CIE y

0.5 3.86 17.2 33.7 2074 0.414 0.454 0.75 3.88 17.4 33.9 2091 0.468 0.458 1.0 3.86 18.5 35.4 2170 0.484 0.452

Figure 4.4: EL Spectra of devices with various thicknesses of the red and green co-doped layer.

83

To understand the function of the dual blue EMLs in the exciton formation and

energy transfer processes, we examined three white devices with the following emitter

structures:

W1: TCTA:FIrpic / TCTA:Ir(2-phq)2(acac):Ir(ppy)3 / DCzPPy,

W2: TCTA:FIrpic / TCTA:Ir(2-phq)2(acac):Ir(ppy)3 / DCzPPy:FIrpic and

W3: TCTA / TCTA:Ir(2-phq)2(acac):Ir(ppy)3 / DCzPPy:FIrpic.

For Device W1, only the TCTA layer was doped with FIrpic whereas the DCzPPy

layer was undoped; for Device W2, both the TCTA and DCzPPy layers were doped; and

for Device W3, only the DCzPPy layer was doped. The thickness of the doped or undoped

layer was 4 nm and the FIrpic concentration in TCTA and DCzPPy was 15% and 20%,

respectively. The detailed EL performance of the devices is summarized in Table 4.3. As

shown in Figure 4.5, the EQEs are almost identical regardless of the variation of the emitter

structures. However, it can be seen that the spectral responses (inset of Figure 4.5) are quite

different, especially in the blue region. Relative to the red and green emissions, Device W1

shows the strongest blue emission whereas it is the weakest in Device W3. Emissions from

the green and red dopants can come from three different pathways: 1) direct energy transfer

from host DCzPPy or TCTA to the dopants, 2) direct exciton formation at the dopants due

to hole trapping, 3) indirect energy transfer from host DCzPPy or TCTA to the dopants via

FIrpic. Pathways 1) and 2) should contribute to the red and green emission irrespective of

the emitter structures, while pathway 3) would lead to different blue emission intensity if

FIrpic excitons were to have a different lifetime in the TCTA and DCzPPy hosts. The

84

lower blue emission intensity observed in the FIrpic doped DCzPPy (Device W3) suggests

that FIrpic excitons are longer lived in DCzPPy than in TCTA.

The photoluminescence lifetime of FIrpic in TCTA and DCzPPy hosts was

measured on films (50 nm) of compositions identical to those of FIrpic doped layers used

in the devices. FIrpic in DCzPPy was found to have a lifetime of 0.94 µs compared to 0.61

µs for FIrpic in TCTA (see Figure 4.6). These lifetime results are in agreement with the

device data that the longer-lived FIrpic excitons in DCzPPy more likely undergo energy

transfer to the adjacent red and green dopants than FIrpic in TCTA.

Table 4.3: EL Performance of white devices with selectively blue doped emitting layers (driven at 5 mA/cm2). ITO (110nm)/HATCN (3 nm)/TAPC(37 nm)/TCTA:FIrpic (x%, 4 nm)/TCTA:Ir(2-phq)2(acac):Ir(ppy)3 (2%:6%, 0.5 nm)/DCzPPy:FIrpic (y%, 3 nm)/TmPyPB (10 nm)/TmPyPB:Cs2CO3 (50%, 30 nm)/Al (100 nm).

Device Voltage (V)

EQE (%)

PE (lm/W)

Luminance (cd/m2)

1931 CIE x

1931 CIE y

W1 4.04 18.6 32.1 2135 0.419 0.428 W2 3.91 18.5 34.8 2170 0.459 0.453 W3 4.05 19.2 35.2 2265 0.491 0.453

85

Figure 4.5: EQE vs luminance of the devices with selectively blue doped emitting layers. (Embedded are the EL spectra of the three devices driven at 5 mA/cm2.)

Figure 4.6: Transient PL decay of two FIrpic doped films.

86

Furthermore, we fabricated pure blue devices (without the green and red co-doped

layer) for device lifetime studies. It is known that FIrpic molecules are susceptible to

excited state dissociation [22-24] releasing the ancillary picolinate ligand. We therefore

expect a device with FIrpic doped DCzPPy as the blue emitting layer to be less stable than

one with FIrpic doped TCTA due to the longer excited state lifetime in the former. This

conjecture is consistent with the relative device lifetimes observed in three blue devices of

the following emitter structures:

B1: TCTA:FIrpic (15%) / DCzPPy,

B2: TCTA:FIrpic (15%) / DCzPPy:FIrpic (20%), and

B3: TCTA / DCzPPy:FIrpic (20%).

All three devices show blue emissions from FIrpic with relatively similar EQE of

16.8%, 15.8%, and 17.7%, respectively. However, when tested at 5 mA/cm2 with an initial

luminance of about 1800 cd/m2, Device B3 with FIrpic doped DCzPPy exhibited a

considerably lower half-life of only about 8 minutes compared to about 28 minutes for

Devices B1 and B2 where FIrpic was doped in TCTA. (Table 4.4 includes device EL

performance and Figure 4.7 shows the lifetime test data.)

Table 4.4: EL Performance of blue devices with selectively blue doped emitting layers (driven at 5 mA/cm2). ITO (110nm)/HATCN (3 nm)/TAPC(37 nm)/TCTA:FIrpic (x%, 4 nm)/ /DCzPPy:FIrpic (y%, 3 nm)/TmPyPB (10 nm)/TmPyPB:Cs2CO3 (30 nm)/Al (100 nm).

Device Voltage (V)

EQE (%)

PE (lm/W)

Luminance (cd/m2)

1931 CIE x

1931 CIE y

B1 4.60 17.7 26.7 1959 0.161 0.384 B2 4.62 16.8 25.6 1882 0.166 0.389 B3 4.95 15.8 21.8 1717 0.157 0.375

87

Figure 4.7: Device lifetime of three blue devices with selectively doped blue emitting layers.

4.3. Conclusions

We have successfully demonstrated high-efficiency WOLEDs with an emitter

structure consisting of an ultra-thin red and green co-doped layer sandwiched in between

two blue layers. Using this flexible architecture, we were able to adjust the compositions

and thicknesses of the individual layers to realize WOLEDs with EQE of nearly 20% and

luminance efficacy of over 40 lm/W (at 1000 cd/m2) and minimal color shift over a large

range of intensities (400-4000 cd/m2). We also found that the device degradation is related

to the lifetime of FIrpic excited states, which is dependent on the host materials.

88

References

[1] J. Kido, M. Kimura, and K. Nagai, Science 267, 1332 (1995).

[2] B.W. D’Andrade, R.J. Holmes, and S.R. Forrest, Adv. Mater. 16, 624 (2004).

[3] S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Lüssem, and K. Leo,

Nature 459, 234 (2009).

[4] Y.L. Chang, Y. Song, Z. Wang, M.G. Helander, J. Qiu, L. Chai, Z. Liu, G.D.

Scholes, and Z. Lu, Adv. Funct. Mater. 23, 705 (2013).

[5] H. Sasabe, J.I. Takamatsu, T. Motoyama, S. Watanabe, G. Wagenblast, N. Langer,

O. Molt, E. Fuchs, C. Lennartz, and J. Kido, Adv. Mater. 22, 5003 (2010).

[6] S. Chen, G. Tan, W.Y. Wong, and H.S. Kwok, Adv. Funct. Mater. 21, 3785 (2011).

[7] Y.H. Son, Y.J. Kim, M.J. Park, and J.H. Kwon, Dig. Tech. Pap. - SID Int. Symp.

44, 1438 (2013).

[8] G. Schwartz, M. Pfeiffer, S. Reineke, K. Walzer, and K. Leo, Adv. Mater. 19, 3672

(2007).

[9] Y.R. Sun, N.C. Giebink, H. Kanno, B.W. Ma, M.E. Thompson, and S.R. Forrest,

Nature 440, 908 (2006).

[10] G. Schwartz, S. Reineke, T.C. Rosenow, K. Walzer, and K. Leo, Adv. Funct. Mater.

19, 1319 (2009).

[11] H. Kanno, R.J. Holmes, Y. Sun, S. Kena-Cohen, and S.R. Forrest, Adv. Mater. 18,

339 (2006).

[12] X. Qi, M. Slootsky, and S. Forrest, Appl. Phys. Lett. 93, 193306 (2008).

[13] S.J. Su, E. Gonmori, H. Sasabe, and J. Kido, Adv. Mater. 20, 4189 (2008).

89

[14] Y.S. Seo and D.G. Moon, J. Soc. Inf. Disp. 18, 1010 (2010).

[15] R. Wang, D. Liu, H. Ren, T. Zhang, H. Yin, G. Liu, and J. Li, Adv. Mater. 23, 2823

(2011).

[16] J. Lee, J.-W. Lee, N.S. Cho, J. Hwang, C.W. Joo, W.J. Sung, H.Y. Chu, and J.-I.

Lee, Curr. Appl. Phys. 14, S84 (2014).

[17] Y. Zhao, L. Zhu, J. Chen, and D. Ma, Org. Electron. 13, 1340 (2012).

[18] Y. Sun and S.R. Forrest, Appl. Phys. Lett. 91, 263503 (2007).

[19] P. Tyagi, R. Srivastava, A. Kumar, S. Tuli, and M.N. Kamalasanan, J. Lumin. 136,

249 (2013).

[20] S. Chen, Q. Wu, M. Kong, X. Zhao, Z. Yu, P. Jia, and W. Huang, J. Mater. Chem.

C 1, 3508 (2013).

[21] Q. Wang and D. Ma, Chem. Soc. Rev. 39, 2387 (2010).

[22] V. Sivasubramaniam, F. Brodkorb, S. Hanning, H.P. Loebl, V. van Elsbergen, H.

Boerner, U. Scherf, and M. Kreyenschmidt, J. Fluor. Chem. 130, 640 (2009).

[23] I.R. De Moraes, S. Scholz, B. Lüssem, and K. Leo, Org. Electron. 12, 341 (2011).

[24] E. Baranoff, B.F.E. Curchod, J. Frey, R. Scopelliti, F. Kessler, I. Tavernelli, U.

Rothlisberger, M. Grätzel, and M.K. Nazeeruddin, Inorg. Chem. 51, 215 (2012).

90

Chapter 5 Investigation of Phosphorescent Blue and

White Organic Light-Emitting Diodes with High

Efficiency and Improved Lifetime

5.1. Introduction

After decades of development, OLEDs have successfully been used for display

applications in mobile phones and TVs. However, as a potential solid-state lighting source

to replace traditional light sources such as LEDs, incandescent light bulbs and fluorescent

tubes [1–5], WOLEDs are still in the development stage. To achieve high-efficiency

WOLEDs, phosphorescent emitters based on heavy metals such as iridium(III) and

platinum are inevitable, thanks to their nearly 100% IQEs [6–8]. Although a number of red

and green phosphorescent materials with a high efficiency and long lifetime (> 100,000 h)

have been developed [9, 10], blue phosphorescent materials are still suffering from much

shorter device lifetimes [12, 13]. Such device instability can be attributed to various forms

of material degradation caused by chemical reactions and bond cleavage in the excited state

of the molecules (including wide band gap dopant, host and transport materials) [14–17].

Therefore, the lifetime of a high-performance WOLED is greatly restricted by the choice

of blue phosphorescent materials.

Typical iridium-based blue phosphorescent dopants, such as FIrpic and FIr6, have

been intensely studied and devices reported with EQEs above 20% [18–21] and lifetimes

ranging from minutes [22, 23] to hours [24, 25] (depending on device structures and stress

91

conditions). It has been shown that the fluorine ligand in these two materials (FIrpic and

FIr6), the ancillary picolinate ligand in FIrpic [14, 26] and the pyrazolyl-borates ligand in

FIr6 [14, 27] are susceptible to dissociation. Moreover, FIrpic has been reported to be

unstable with respect to hole transport [17].

More stable blue dopants such as Ir(iprpmi)3 with an imidazole-phenol ligand, were

first reported by Lin et al. [28]. OLED lifetimes of up to 1,000 h have been achieved [28–

30]. A recent report from our laboratory showed that the efficiency and lifetime of

Ir(iprpmi)3-based OLEDs were highly dependent on the choice of HTM and ETM [31]. In

this chapter, we describe in more detail how the properties of the blue phosphorescent

dopant (Ir(iprpmi)3) affect the performance of the blue and white OLEDs, including

improvements in lifetime over the FIrpic-based devices.

5.2. Results and Discussion

As illustrated in Figure 5.1, HATCN was used as an HIL, and TAPC as an HTL.

The EML was either Ir(iprpmi)3 or FIrpic doped into a bipolar host DCzPPy. Adjacent to

the EML was an undoped TmPyPB layer for electron transport. Cesium carbonate

(Cs2CO3) doped TmPyPB (50%) was the EIL, and aluminum was the cathode. For

WOLEDs, TCTA was the host for the red dopant (Ir(2-phq)2(acac)) and green dopant

(Ir(ppy)3). The HOMO and LUMO levels (labeled above and below the rectangles) and the

triplet energy levels (in parentheses) for the materials used are also indicated in Figure 5.1.

92

Figure 5.1: Schematic energy level diagram of the materials used in this chapter (LUMO and HOMO energy levels are labeled above and below rectangles, triplet energy levels

are indicated in parentheses).

For the study of the charge-carrier transport properties of Ir(iprpmi)3, hole-only (A)

and electron-only (B) devices with the following layer structures were fabricated: (A)

ITO/HATCN(3 nm)/TAPC(40 nm)/DCzPPy:Ir(iprpmi)3 (x%, 30 nm)/HATCN(3 nm)/Al,

and (B) ITO/TmPyPB(10 nm)/DCzPPy:Ir(iprpmi)3(x%, 30 nm)/TmPyPB(40 nm)/LiF(1

nm)/Al. For both devices, a bipolar host DCzPPy was used and the concentration of the

dopant Ir(iprpmi)3 was varied from 0% to 20% (0%, 3%, 6%, 9%, 15% and 20% namely).

Figure 5.2 summarizes the current density-drive voltage (J-V) characteristics. For

the hole-only device without Ir(iprpmi)3, the drive voltage was the lowest. With 3%

Ir(iprpmi)3, the drive voltage was substantially increased to 11 V at 5 mA/cm2. With further

increase in Ir(iprpmi)3 concentration, the drive voltage began to decrease, but only to a

level above that of the device without Ir(iprpmi)3. This J-V behavior shows that Ir(iprpmi)3

93

is an effective hole trap when doped at a low concentration in DCzPPy, which has a high

HOMO level at 6.05 eV. At higher concentrations of Ir(iprpmi)3, the hole transport took

place in Ir(iprpmi)3 in conjunction with the transport in the host, although with a lower

mobility. In contrast, for the electron-only devices (as illustrated in Figure 5.2(b)), the drive

voltage increased proportionally with increasing Ir(iprpmi)3 concentration over the entire

range of 0% to 20%. This indicates that Ir(iprpmi)3 does not support electron transport in

DCzPPy owing to the relatively higher LUMO level of Ir(iprpmi)3 (2.2 eV) compared to

that of DCzPPy (2.56 eV).

Figure 5.2: J-V curves of hole-only and electron-only devices with various doping concentrations of Ir(iprpmi)3.

A set of five blue OLEDs were fabricated with the following device structures:

ITO/HATCN(3 nm)/TAPC(40 nm)/TCTA(4 nm)/DCzPPy:Ir(iprpmi)3 (x%, 10

nm)/TmPyPB(10 nm)/TmPyPB:Cs2CO3(50%, 30 nm)/Al(100 nm), where the dopant

Ir(iprpmi)3 concentration in host DCzPPy was varied from 3% to 20%. A 4 nm TCTA

(a)

(b)

94

layer was the buffer between TAPC and the EML to provide balanced recombination.

Figure 5.3(a) shows the J-V characteristics of the five blue OLEDs. As expected, the drive

voltage decreased proportionally with an increasing Ir(iprpmi)3 concentration. At 5

mA/cm2, the drive voltage for the device with 20% Ir(iprpmi)3 was approximately 0.6 V

lower than that of the device with 3% Ir(iprpmi)3. This modest decrease in voltage was in

part due to a thinner EML (10 nm). With such a thin EML, the device’s EQE is expected

to be more sensitive to the dopant concentration. As demonstrated in Figure 5.3(b), at

5mA/cm2, devices with 15% and 20% Ir(iprpmi)3 exhibited lower EQEs compared to the

moderately doped devices (6% and 9% Ir(iprpmi)3). This can be attributed to increased

self-quenching at high dopant concentrations. The emission spectra (Figure 5.2(c)) indicate

a slight red shift with increasing Ir(iprpmi)3 concentration, which can be an indication of a

shift of the recombination region towards the EML/ETL interface.

95

Figure 5.3: Device performance of five blue OLEDs with various Ir(iprpmi)3 dopant concentrations. (a) Current density vs. voltage, (b) EQE vs. current density, (c) EL

spectra at 5 mA/cm2.

As illustrated in Figure 5.3(b), the EQE of the devices with various Ir(iprpmi)3

concentrations exhibits very different behaviors at low current densities. With low

concentrations (3%), the EQE is only about 12% at the current density of 0.01 mA/cm2. At

high Ir(iprpmi)3 dopant concentrations (15% and 20%), the EQE is even lower (at about

5%). However, with medium Ir(iprpmi)3 concentrations (6% and 9%), the EQE is the

highest at 20%. Such phenomena can be explained as follows. At a low Ir(iprpmi)3

(a) (b)

(c)

96

concentration, holes are predominantly trapped at the HTL/EML interface, and electron

dominates as the transport in the EML. Therefore, the recombination zone is confined to

the HTL/EML interface, which leads to possible polaron-triplet quenching and triplet-

triplet annihilations and, therefore, a low EQE. As the current density increases, holes

transport starts in the EML, thus resulting in the broadening of the recombination region

and, consequently, a gradual increase in EQE. For the highly doped devices (15% and 20%

Ir(iprpmi)3), holes are readily transported across the EML via the Ir(iprpmi)3 dopant in the

DCzPPy host, whereas electron injection from TmPyPB to DCzPPy is impeded by a 0.2

eV energy barrier. As the current density increases at a higher bias voltage, electrons are

more easily injected and transported through the EML, thus causing the recombination

region to shift towards the HTL/EML interface and an increased EQE. With moderate

Ir(iprpmi)3 concentrations (6% and 9%), the recombination zone is more extended inside

the EML as the holes do not get trapped at the HTL/EML interface because of assisted hole

transport via Ir(iprpmi)3), and they provide a more balanced recombination with the

injected electrons. EQEs above 20% are therefore achieved as a consequence.

Based on the structures of the preceding blue devices, a set of five WOLEDs were

fabricated with an addition of a red and green co-doped thin layer. The detailed layer

structure is as follows: ITO/HATCN(3 nm)/TAPC(40 nm)/TCTA(4 nm)/TCTA:Ir(2-

phq)2(acac):Ir(ppy)3 (2%, 6%, 1 nm)/DCzPPy:Ir(iprpmi)3 (x%, 4 nm)/TmPyPB(10

nm)/TmPyPB:Cs2CO3(50%, 30 nm)/Al(100 nm), where the Ir(iprpmi)3 concentration is

varied from 3% to 20% in the blue EML. The thickness (1nm) and doping concentration

97

for the red (2%) and green (6%) co-doped layer were fixed at optimized values based on

our previous study [32].

Table 5.1 summarizes the performance of the WOLEDs. Among them, the drive

voltage was the highest for 3% Ir(iprpmi)3 at 4.98 V and lowest for 20% Ir(iprpmi)3 at 4.07

V. This voltage trend is in agreement with the blue devices described earlier. The EQEs of

the low (3% Ir(iprpmi)3) and moderately doped (6% and 9%) devices were all above 20%,

indicating a highly effective exciton confinement with minimal exciton quenching. On the

contrary, for higher-doped devices (15% and 20% Ir(iprpmi)3), the EQEs were down to

17.3% and 14.4%, respectively. This can be partially attributed to an increase in self-

quenching at higher dopant concentration, as in the blue devices. However, with the red

and green co-doped layer between TCTA and the blue EML, the triplet energy transfer and

exciton distribution among the three dopants are more complicated and, therefore, can

strongly affect the dependence of the device’s EQEs and colors on drive conditions.

Table 5.1: EL Performance of WOLEDs with various Ir(iprpmi)3 concentrations. ITO/HATCN(3 nm)/TAPC(40 nm)/TCTA(4 nm)/TCTA:Ir(2-phq)2(acac):Ir(ppy)3 (2%, 6%, 1 nm)/DCzPPy:Ir(iprpmi)3 (x%, 4 nm)/TmPyPB(10 nm)/TmPyPB:Cs2CO3(50%, 30 nm)/Al (100 nm). (a: values at 5 mA/cm2, b: measured at current densities from 0.05 mA/cm2 to 20 mA/cm2.)

Ir(iprpmi)3 (%)

Voltage (V)a

EQE (%)a

PE (lm/W)a

Luminance (cd/m2)a

CIE xa

CIE ya

CIE Dxb

CIE Dyb

3 4.98 20.4 29.0 2299 0.494 0.428 ±0.016 ±0.010 6 4.65 21.7 33.6 2481 0.462 0.453 ±0.006 ±0.008 9 4.57 20.7 33.5 2438 0.407 0.434 ±0.001 ±0.005

15 4.39 17.3 30.3 2118 0.342 0.444 ±0.024 ±0.003 20 4.07 14.4 27.9 1805 0.309 0.449 ±0.015 ±0.005

98

Figure 5.4(a) illustrates EQE’s dependence on the current density. WOLEDs with

3%, 6% and 9% Ir(iprpmi)3 all exhibit very high EQE (about 25%) at 0.01 mA/cm2. Such

a phenomena can be explained as follows: Due to hole trapping by Ir(iprpmi)3 at low

concentrations, the recombination region is located near the red and green co-doped

layer/blue layer interface, resulting in efficient triplet energy transfer from Ir(iprpmi)3 to

Ir(ppy)3 and Ir(2-phq)2(acac), and effectively little loss of excitons generated by the

recombination processes. EQEs for these three WOLEDs only slightly roll off to 18% at

high current densities, mainly due to the usual charge quenching. For the highly doped

WOLEDs (15% and 20%), the hole current is expected to dominate at low bias due to

assisted hole transport via Ir(iprpmi)3. Hence, the carrier recombination region in these

WOLEDs is mainly located near the blue layer/ETL interface, which leads to reduced

efficiency of energy transfer from Ir(iprpmi)3 to Ir(ppy)3 and Ir(2-phq)2(acac). Since blue

excitons cannot be efficiently utilized, the EQEs are, therefore, lower at low current

densities. As the bias increases, electrons can be more easily injected and transported in

the blue layer, thus resulting in a broad recombination region and a gradual increase in

EQEs.

99

Figure 5.4: Device performance of five WOLEDs with various Ir(iprpmi)3 dopant concentrations. (a) EQE vs. current density; (b) EL spectra at 5 mA/cm2; (c) color shift of

the device with 9% Ir(iprpmi)3 at current densities from 0.05 to 20 mA/cm2.

The electroluminescence spectra of the WOLEDs are illustrated in Figure 5.4(b).

At low Ir(iprpmi)3 concentrations (3% and 6%), the red emission from Ir(2-phq)2(acac) is

predominant, whereas at higher Ir(iprpmi)3 concentrations, the blue emission gains

intensity and eventually dominates the spectrum. These spectral behaviors are a clear

indication of the shift of the recombination region from the HTL side towards the ETL side

as Ir(iprpmi)3 concentration increases.

(a) (b)

(c)

100

The details of WOLED performance are summarized in Table 5.1. The power

efficacy at 5 mA/cm2 for all five WOLEDs are remarkably high as it peaks at about 33

lm/W for 6% and 9% Ir(iprpmi)3. Table 5.1 also illustrates the CIE shifts (Dx and Dy) at

various current densities (from 0.05 to 20 mA/cm2). Surprisingly, the moderately

Ir(iprpmi)3-doped devices (6% and 9%) exhibit very small color shift (< 0.01, < 0.01) while

the low Ir(iprpmi)3-doped device (3%) and high-doped devices (15% and 20%) indicate a

marginal color shift (±0.02 and ±0.01, respectively). The color shift of WOLEDs originates

from the shift of the recombination region and redistribution of excitons among the red,

green and blue emitters at various bias values. Due to the fact that Ir(iprpmi)3 in DCzPPy

can play the role of hole trapping and transport depending on the concentration, the shift

of the recombination region can be minimized by optimizing the concentration to achieve

a WOLED with a stable white emission at various current densities. Figure 5.4(c) illustrates

the normalized spectra of a WOLED with 9% Ir(iprpmi)3 biased from 0.05 mA/cm2 to 20

mA/cm2. It can be seen that the white emission is extremely stable.

To examine the devices lifetimes, we fabricated both blue and white OLEDs with

FIrpic or Ir(iprpmi)3 as the blue dopant. The blue device layer structure is as follows:

ITO/HATCN(3 nm)/TAPC(40 nm)/TCTA(4 nm)/DCzPPy:Blue Dopant (x%, 30

nm)/TmPyPB(10 nm)/TmPyPB:Cs2CO3(50%, 30 nm)/Al(100 nm). The dopant

concentration was optimized to be 20% for FIrpic and 9% for Ir(iprpmi)3, with DCzPPy

being the host. The EQEs of the FIrpic- and Ir(iprpmi)3-based blue OLEDs were 15.8%

and 18.6% at 5 mA/cm2, respectively. For WOLEDs, in order to achieve a similar white

spectrum (see the inset of Figure 5.5), the FIrpic-based device structure was as follows:

101

ITO/HATCN(3 nm)/TAPC(40 nm)/TCTA:FIrpic(15%, 4 nm)/TCTA:Ir(2-

phq)2(acac):Ir(ppy)3 (2%, 6%, 1 nm)/DCzPPy:FIrpic(20%, 4 nm)/TmPyPB(10

nm)/TmPyPB+Cs2CO3(50%, 30 nm)/Al (100 nm), whereas the Ir(iprpmi)3-based device

structure had an Ir(iprpmi)3 concentration of 9%. The EL spectra of the two WOLEDs at 5

mA/cm2 are illustrated in the inset of Figure 5.5. The EQEs of the Flrpic- and Ir(iprpmi)3-

based WOLEDs were 16.7% and 19.2%, respectively. For the lifetime test, all devices were

driven at a constant current density of 5 mA/cm2. Figure 5.5 illustrates that the FIrpic-based

blue device is the shortest-lived with a half-lifetime of only 10 min, whereas the

Ir(iprpmi)3-based blue device had a much improved half-lifetime of 2.5 h. For the

WOLEDs, the half-lifetime is considerably better compared to the blue devices. The

FIrpic-based WOLED had a half-lifetime of about 5 h, and the Ir(iprpmi)3-based WOLED

had a half-lifetime approximately 20 h. FIrpic is highly unstable due to the loss of fluorine

substituents and the breakdown of the picolinate ligand during device operation [14, 15].

On the contrary, the phenyl-imidazole ligands in Ir(iprpmi)3 are believed to be more

electro-chemically stable, which leads to an order-of-magnitude improvement in device

lifetime. However, due to the instability of the wide band gaps of the hole transport material

(TAPC) [33, 34] and electron transport material (TmPyPB) [31], the WOLED lifetime may

also be limited by transport layers.

102

Figure 5.5: Device lifetime tested at 5 mA/cm2 (WOLEDs EL spectra are in the inset).

5.3. Conclusions

By investigating the charge-carrier transporting properties of Ir(iprpmi)3 in hole-

only and electron-only devices, we have shown that the Ir(iprpmi)3 dopant at a low

concentration traps holes, and it transports holes at a high concentration in the DCzPPy

bipolar host material. Based on this property, we varied the Ir(iprpmi)3 concentration to

control the recombination region and thereby achieved high-efficiency blue and white

OLEDs with EQEs over 20%. The optimized WOLEDs showed power efficiency close to

40 lm/W at 1,000 cd/m2 and exhibited a minimal color shift (±0.001, ±0.005) over a large

current density range of 0.05–20 mA/cm2. Compared to devices with FIrpic as the blue

dopant, Ir(iprpmi)3-based blue and white OLEDs had significantly improved device

lifetimes under the same stress conditions.

103

References

[1] B.W. D’Andrade and S.R. Forrest, Adv. Mater. 16, 1585 (2004).

[2] M.C. Gather, A. Köhnen, and K. Meerholz, Adv. Mater. 23, 233 (2011).

[3] S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Lüssem, and K. Leo,

Nature 459, 234 (2009).

[4] G. Zhou, W.-Y. Wong, and S. Suo, J. Photochem. Photobiol. C Photochem. Rev.

11, 133 (2010).

[5] K.T. Kamtekar, A.P. Monkman, and M.R. Bryce, Adv. Mater. 22, 572 (2010).

[6] Y.E. Al Baldo, M. A.; O’Brien, DF; You, Nature 395, 151 (1998).

[7] C. Adachi, M.A. Baldo, M.E. Thompson, and S.R. Forrest, J. Appl. Phys. 90, 5048

(2001).

[8] E. Turner, N. Bakken, and J. Li, Inorg. Chem. 52, 7344 (2013).

[9] K. Nishimura, Y. Kawamura, T. Kato, M. Numata, and M. Kawamura, SID Symp.

Dig. Tech. Pap. 40, 310 (2009).

[10] T. Tsuji, S. Kawami, S. Miyaguchi, T. Naijo, T. Yuki, S. Matsuo, and H. Miyazaki,

J. Soc. Inf. Disp. 13, 117 (2005).

[11] J. Birnstock, M. Hofmann, S. Murano, M. Vehse, J. Blochwitz-Nimoth, Q. Huang,

G. He, M. Pfeiffer, and K. Leo, Dig. Tech. Pap. - SID Int. Symp. 36, 40 (2005).

[12] M. Segal, C. Mulder, K. Celebi, M. Singh, K. Rivoire, S. Difley, T. Van Voorhis,

and M.A. Baldo, Proc. SPIE 6999, 699912 (2008).

[13] T. Nakayama, K. Hiyama, K. Furukawa, and H. Ohtani, J. Soc. Inf. Disp. 16, 231

(2008).

104

[14] I.R. De Moraes, S. Scholz, B. Lüssem, and K. Leo, Org. Electron. 12, 341 (2011).

[15] V. Sivasubramaniam, F. Brodkorb, S. Hanning, H.P. Loebl, V. van Elsbergen, H.

Boerner, U. Scherf, and M. Kreyenschmidt, J. Fluor. Chem. 130, 640 (2009).

[16] K.S. Yook, O.Y. Kim, and J.Y. Lee, Synth. Met. 161, 2677 (2012).

[17] K.P. Klubek, C.W. Tang, and L.J. Rothberg, Org. Electron. Physics, Mater. Appl.

15, 1312 (2014).

[18] C.W. Lee and J.Y. Lee, Adv. Mater. 25, 5450 (2013).

[19] J. Lee, J.I. Lee, J.Y. Lee, and H.Y. Chu, Appl. Phys. Lett. 94, 193305 (2009).

[20] L. Xiao, S.J. Su, Y. Agata, H. Lan, and J. Kido, Adv. Mater. 21, 1271 (2009).

[21] F.P. Wu, Y.M. Xie, L.S. Cui, X.Y. Liu, Q. Li, Z.Q. Jiang, and L.S. Liao, Synth.

Met. 205, 11 (2015).

[22] J. Zhuang, W. Li, W. Su, Y. Liu, Q. Shen, L. Liao, and M. Zhou, Org. Electron.

Physics, Mater. Appl. 14, 2596 (2013).

[23] S.E. Lee, H.W. Lee, S.J. Lee, J.-R. Koo, D.H. Lee, H.J. Yang, H.J. Kim, S.S. Yoon,

and Y.K. Kim, J. Nanosci. Nanotechnol. 15, 1456 (2015).

[24] M.S. Weaver, Y.-J. Tung, B. D’Andrade, J. Esler, J.J. Brown, C. Lin, P.B.

Mackenzie, R.W. Walters, J.-Y. Tsai, C.S. Brown, S.R. Forrest, and M.E.

Thompson, SID Symp. Dig. Tech. Pap. 37, 127 (2006).

[25] W. Sotoyama, T. Satoh, M. Kinoshita, M. Tobise, K. Kawato, T. Ise, H. Takizawa,

and S. Yamashita, SID Symp. Dig. Tech. Pap. 41, 556 (2010).

[26] V. Sivasubramaniam, F. Brodkorb, S. Hanning, H.P. Loebl, V. Elsbergen, H.

Boerner, U. Scherf, and M. Kreyenschmidt, Cent. Eur. J. Chem. 7, 836 (2009).

105

[27] R. Seifert, I. Rabelo De Moraes, S. Scholz, M.C. Gather, B. Lüssem, and K. Leo,

Org. Electron. Physics, Mater. Appl. 14, 115 (2013).

[28] C. Lin, P.B. MacKenzie, R.W. Walters, J.-Y. Tsai, C.S. Brown, J. Deng, U.S. Patent

No. 7,902,374 (2011).�

[29] N.C. Giebink, B.W. D’Andrade, M.S. Weaver, P.B. MacKenzie, J.J. Brown, M.E.

Thompson, and S.R. Forrest, J. Appl. Phys. 103, (2008).

[30] B. D’Andrade, P.B. MacKenzie, M.S. Weaver, J.J. Brown, U.S. Patent No.

8,557,399 (2013).�

[31] K.P. Klubek, S.C. Dong, L.S. Liao, C.W. Tang, and L.J. Rothberg, Org. Electron.

Physics, Mater. Appl. 15, 3127 (2014).

[32] L. Xu, C.W. Tang, and L.J. Rothberg, Org. Electron. Physics, Mater. Appl. 32, 54

(2016).

[33] D.Y. Kondakov, J. Appl. Phys. 104, 084520 (2008).

[34] S. Schmidbauer, A. Hohenleutner, and B. König, Adv. Mater. 25, 2114 (2013).

106

Chapter 6 Investigating Chemical Degradation

Mechanism of High-Triplet-Energy Materials in Blue

Phosphorescent OLED Using LDI-TOF

6.1. Introduction

Phosphorescent OLED (PhOLED) has already been adopted in commercial OLED

panels for green and red pixels because of its superior device efficiency and lifetime. Blue

PhOLED, however, is yet to deliver sufficient device lifetime to replace conventional

fluorescent OLED that is currently being used in commercial products [1–4].

The longevity of blue PhOLED is highly dependent on the phosphorescent emitters

and host and transport materials used. In previous studies, it has been concluded that FIrpic,

which is an efficient blue phosphorescent dopant that is widely used in PhOLED research,

is chemically unstable during device operation [5]. The device lifetime of FIrpic is typically

within several hours. Disassociation of the picolinate auxiliary ligand has been identified

as the main degradation pathway of FIrpic through the LDI-TOF technique [6, 7]. In our

recent study, we demonstrated that using homoleptic iridium complex Ir(iprpmi)3 as the

blue dopant can significantly improve device lifetime. We also found out that transport

materials with lower triplet energies tend to result in longer lifetimes and lower efficiencies

for devices, while transport materials with higher triplet energies, i.e. TAPC and TmPyPB,

result in short-lived devices, although with higher efficiencies and lower drive voltages [8].

To make blue PhOLED useful in practice, both high efficiency and long lifetime need to

107

be achieved simultaneously. Understanding the degradation mechanisms of a high-triplet-

energy host and transport materials is crucial to realizing an applicable blue PhOLED. LDI-

TOF, first adopted in OLED degradation analysis by Leo K. et al. [9], has proven to be an

effective tool in chemical degradation analysis. Besides FIrpic, degradation patterns of

various OLED materials have been studied in situ using LDI-TOF [10, 11].

In this work, we investigate the degradation pattern of high-efficiency Ir(iprpmi)3-

based blue phosphorescent OLED comprising materials with high triplet energy (> 2.7 eV).

TAPC, TCTA, TmPyPB and DCzPPy were selected as a host or transport material. We

started with evaluating device performance and lifetime by varying host types for

Ir(iprpmi)3. Degradation products of each material (transport, host and emitter) have been

systematically analyzed by LDI-TOF. The chemical compositions of aged devices were

probed in situ using the LDI-TOF technique. The results suggest that chemical degradation

mainly occurs at the HTL/EML or EML/ETL interface, where the exciton density is high.

From fragment structures and theoretically calculated bond dissociation energies, a cation-

induced ring-open mechanism was deduced as an alternative chemical degradation

pathway of TAPC. TCTA as a host degrades by means of C-N dissociation. The TmPyPB

electron transport material mainly undergoes protonation at recombination interfaces.

Bipolar transporting material DCzPPy, when used as a host, exhibits fewer tendencies of

C-N bond cleavage and is relatively stable.

108

6.2. Results and Discussion

6.2.1. Device Performance and Lifetime Evaluation

Figure 6.1: Schematic energy diagram for blue PhOLEDs. (The triplet energy is in parentheses, and HOMO/LUMO energies are below and above the rectangles).

Three blue PhOLEDs with different host materials were fabricated side by side with

a device structure of ITO/HATCN (3 nm)/TAPC (40 nm)/TCTA (4 nm)/Host:Ir(iprpmi)3

(9%, 10 nm)/TmPyPB (10 nm)/TmPyPB:Cs2CO3 (50%, 30 nm)/Al. The three host

materials were TmPyPB, TCTA, DCzPPy. The molecular structure and triplet energy level

of materials used are illustrated in Figure 6.1. HATCN was the HIL, and TAPC was the

HTL. TCTA was used as an exciton blocking layer that isolated TAPC from the EML. The

DCzPPy bipolar host [12] was used to sensitize Ir(iprpmi)3, which is a relatively stable blue

dopant, as we have previously reported [8]. Neat and heavily doped TmPyPB were utilized

as the ETL and EIL, respectively.

The device EQE and lifetime performance are illustrated in Figure 6.2, in which

Device A3 exhibits an EQE of over 20%, while Devices A1 and A2 only have EQEs of

109

approximately 15%. This can be partially attributed to the bipolar carrier transport property

of DCzPPy, which leads to a broadened recombination region throughout the EML. With

Devices A1 and A2, the recombination region is confined to the TCTA/TmPyPB interface,

where self-quenching might happen. Moreover, from the device lifetime test (illustrated in

Figure 6.2(b)), it is evident that the brightness of all three devices drops to under 30% in

less than 20 h operation at 5 mA/cm2. Nevertheless, Device A3, in which DCzPPy is the

host, exhibits a relatively longer lifetime compared to Devices A1 and A2.

Figure 6.2: Efficiencies and lifetime performances of Device A1, A2 and A3.

6.2.2. Overall Stability Assessment of the Blue PhOLED

To investigate the cause of the blue PhOLED’s short device lifetime, an aged

Device A3 was subjected to LDI-TOF analysis along with an unaged device as reference.

The aged sample was driven at a constant current density of 5 mA/cm2 for 24 h in a vacuum

assembly. Since the laser (wavelength: 337 nm) can induce photo-fragmentation of OLED

110

materials, both samples were analyzed in one run at the same laser intensity and mode.

Figure 6.3 illustrates the LDI-TOF spectra of the two samples.

Figure 6.3: Normalized LDI-TOF spectra of Device A3 with and without degradation.

Based on the LDI-TOF and TOF/TOF spectra of the individual materials, most of

the peaks in Figure 6.3 are assigned to proposed molecular structures. Mass peaks at 626,

740 and 1103 correspond to molecular masses of TAPC, TCTA and Ir(iprpmi)3,

respectively. HATCN was not detected because of its high ionization potential. DCzPPy

was absent due to its lower ionization potential (IP) compared to TAPC. TmPyPB was also

absent due to its lack of absorption at the laser wavelength (337 nm). All other peaks in the

111

spectra are fragments and adducts, the proposed structures that are summarized in Table

6.1.

112

Table 6.1: List of mass peaks and their proposed structures.

Mass (m/z) Origin Nature Proposed structure

431 TAPC Fragment

465 TAPC Fragment

499 TCTA Fragment

536 TAPC Fragment

557 TAPC Fragment

570 TAPC Fragment

583 TAPC Fragment

591 TAPC Adduct

113

Mass (m/z) Origin Nature Proposed structure

626 TAPC Molecular mass

717 TAPC Adduct

740 TCTA Molecular mass

799 Ir(iprpmi)3 Fragment

855 Ir(iprpmi)3 Fragment

Product ion formed by dissociation of meta-stable Ir(iprpmi)3

+ in post-source decay (PSD); it possibly has the same structure as 799.

1057 TAPC Adduct

1103 Ir(iprpmi)3 Molecular mass

a. Peaks at 855 have abnormally low resolutions (large FWHM) and is absent in the linear

mode, which indicates that it is a product ion from a precursor dissociation in the PSD.

114

6.2.3. Degradation of Blue Dopant

Ir(iprpmi)3 undergoes simple Ir-C and Ir-N dissociations, which result in

Ir(iprpmi)2+ at 799 m/z. The peak at 855 m/z also corresponds to the Ir species, possibly

still Ir(iprpmi)2+, formed by meta-stable parent ions dissociating in the reflectron. This can

be supported by data from the linear-mode test (which does not involve a reflectron) of a

neat Ir(iprpmi)3 film. Figure 6.4 illustrates the LDI-TOF results of UV aged (254 nm, 24 h)

Ir(iprpmi)3 films tested at reflectron mode and linear mode respectively. As shown in

Figure 6.4, the peak at 855 m/z disappeared in the linear mode.

Figure 6.4: Normalized LDI-TOF spectra of the neat Ir(iprpmi)3 film in the linear mode.

6.2.4. Degradation of TAPC

TAPC is one of the earliest HTMs used in OLEDs. Its high ionization potential and

hole mobility ensure good hole injection and transport. Its wide band gap is also beneficial

to exciton confinement, particularly in blue phosphorescent and TADF OLEDs. To date,

115

TAPC is still widely used in OLED research to produce some of the highest device

efficiencies [13]. However, devices using TAPC had much shorter lifetimes compared to

devices using other HTMs, e.g., NPB. Kondakov et al. conducted an extensive degradation

study on TAPC- and NPB-based OLED devices [14]. They found that TAPC chemically

degraded more than NPB did both at the recombination interface and in the bulk. From

identified byproducts, C-N bond dissociation was reconstructed as the main degradation

pathway with C-C bond cleavage between phenyl group and cyclohexyl ring as a minor

pathway. The high exciton energy of TAPC was attributed as the driven force of these

degradation pathways. According to this mechanism, the initial degradation started at the

interface with homolytic bond dissociation of neutral TAPC and the degradation in the bulk

was suggested to be caused by radical chain reactions as evidenced by the formation of

high molecular weight byproducts, other potential pathways such as rupture of cyclohexyl

ring, or simply a higher degree of interface deterioration.

Because it has a lower IP than TCTA and DCzPPy, TAPC almost completely

overshadowed other peaks in both aged and unaged samples, except Ir(iprpmi)3, which has

the lowest IP in this set of materials. In the aged device, peaks from TCTA and DCzPPy

were revealed, and the intensity of TAPC fragments increased, thus indicating chemical

degradation of TAPC. From fragment structures listed in Table 6.1, two fragmentation

pathways can be reconstructed for TAPC: 1) C-N bond cleavage and 2) cyclohexyl rupture.

The C-N cleavage gives fragments at 431 and 536, and adducts at 717 and 1056.

Cyclohexyl rupture results in fragments at 465, 557, 570 and 583. A combination of both

116

fragmentation pathways yields an adduct at 591. Some of these fragments are protonated

most likely by interactions with hydrogen-abundant fragments from cyclohexyl.

Figure 6.5: TOF/TOF spectrum of the TAPC cation.

In the structure of Device A3, TAPC is merely used as an HTM isolated from the

recombination interface by TCTA. Excitons formed in the EML can hardly affect the

TAPC layer. For it to degrade chemically, a self-initiated reaction is necessary. During

operation, the TAPC molecule exists in the neutral state and cationic state. Thus, the

stability of the TAPC cation is crucial to device stability. TOF/TOF, which is a tandem

mass technique, provides a useful way to study the intrinsic fragmentation of cations in

vacuum. As illustrated in Figure 6.5, the mass peak at 583, which is the product ion of

117

cyclohexyl rupture process, appears to be the dominating fragment in the TOF/TOF

spectrum of the TAPC cation. It strongly suggests that TAPC mainly follows a cation-

induced ring-rupture chemical degradation pathway, rather than the exciton-provoked C-

N cleavage mechanism proposed by Kondakov [14]. Additionally, the complete

suppression of TAPC precursor peaks indicates a high fragmentation ratio of the TAPC

cation; hence, a higher chance of degradation when TAPC is used as an HTM.

To elucidate the effect of positive charge on the fragmentation of TAPC in solid

films, a comparison LDI-TOF experiment was conducted on a neat TAPC film (20 nm)

and an HATCN (3 nm)/TAPC (20 nm) bilayer structure. HATCN, which is a strong

electron acceptor, can form charge a transfer complex with electron donating TAPC at the

interface, thus producing TAPC cations in situ. When irradiated by a laser with a relatively

lower intensity in LDI-TOF, these two samples showed different fragmentation patterns

(Figure 6.6). The neat TAPC film mainly underwent C-N cleavage fragmentation, whereas

HATCN/TAPC showed more fragments from the cyclohexyl rupture. Further increase of

laser intensity also increased cyclohexyl rupture fragments in the neat TAPC film. But the

fragment intensity is much lower than that in the HATCN/TAPC sample. It can be

concluded that direct laser irradiation on a neutral TAPC causes a C-N bond dissociation,

whereas the irradiation on a TAPC cation induces cyclohexyl ring-open reaction.

118

Figure 6.6: LDI-TOF spectra of the neat TAPC film and HATCN/TAPC bilayer.

To explain the different fragmentation pathways of neutral and cationic TAPC, the

density function theory (DFT) method was utilized to calculate the dissociation energy of

each broken bond in both the neat TAPC and the TAPC cation (Figure 6.7). In the neutral

TAPC, the dissociation energies of both C-N bonds, which are comparable to the value

reported by Kondakov [14], are lower than that of the C-C bond in cyclohexyl ring.

However, in the TAPC cation, C-N bond dissociation energies increase while, the

cyclohexyl ring-opening energy drastically decreases to 1.4 eV. Once the cyclohexyl ring

is open, further cracking reactions ensue due to the high reactivity of the radical cation

(Figure 6.8). The low energy barrier significantly increases the chance of degradation of

the TAPC cation, which is in agreement with TOF/TOF and LDI-TOF results.

119

Figure 6.7: Dissociation energy of bonds in the neutral TAPC and TAPC cation.

Figure 6.8: Dissociation energy of cracking reactions after cyclohexyl is opened in the TAPC cation. The dissociation of 1 corresponds to fragments at 570 and 591, and that of

2 corresponds to the peak at 583.

To understand the origin of the low ring-opening energy in the TAPC cation, the

resonant forms of the TAPC cation and ring-opened TAPC cation, together with calculated

highest occupied molecular orbitals (HOMOs) of TAPC and ring-opened TAPC, are

120

illustrated in Figure 6.9. In the TAPC cation, positive charge is located on each of the two

amine groups, separated by the center cyclohexyl ring. But once the ring is open, the central

carbon acts like a bridge, delocalizing the positive charge on both amine groups. The

extended resonance lowers the energy of the reactive radical cation and, therefore, lowers

the energy barrier between the TAPC cation and ring-opened TAPC cation. In TAPC, the

HOMO is separated by the cyclohexyl ring. While in ring-opened TAPC, the HOMO on

two amines is connected by the central carbon radical, thus reasserting the cation-induced

ring-opening mechanism. However, the driving force of this ring opening reaction in

OLED device is still not clear. It is possible that the TAPC radical cations at excited state,

which may be produced by photo-excitation from ambient light, is involved in the

degradation.

121

Figure 6.9: Resonant structures (up) of the TAPC cation and ring-opened TAPC cation and HOMO (down) of TAPC and ring-opened TAPC.

6.2.5. Degradation of TCTA, DCzPPy and TmPyPB

Fragmentations of TCTA were observed in the degraded sample at peaks 499 and

799, as illustrated in Figure 6.3. However, it is hard to draw any conclusion about the origin

of these fragmentations due to the matrix effect and IP hierarchy of analytes [15]. The peak

intensity may not necessarily reflect the actual abundance of each species. Signals from

TAPC and its fragments can overshadow other peaks, which makes it hard to analyze

possible degradation of other materials (such as TCTA, DCzPPy and TmPyPB) in the

presence of TAPC.

122

To eliminate the swamping effect of TAPC signals and study the chemical reaction

at the interface of the host materials, a series of bilayer structured OLED devices were

fabricated without the blue phosphorescent dopant. Device B1 had the following structure:

ITO/HATCN (3 nm)/TCTA (40 nm)/TmPyPB (40 nm)/LiF (1 nm)/Al. Excitons are formed

at TCTA/TmPyPB interface where TCTA+ and TmPyPB- are accumulated and

recombined. Device B1 was driven at a constant current density of 5 mA/cm2 for 24 h in a

vacuum assembly and underwent LDI-TOF analysis. Figure 6.10 shows the spectra Device

B1 before and after electrical aging. Without TAPC, a mass peak at 740 from the TCTA

parent is the dominant signal in both samples. Mass peak 499, which corresponds to a

fragment of TCTA ([TCTA-PhCz]+), can also be observed. Moreover, one more peak (at

982 m/z) shows up in the spectra of the two samples. This peak is identified as an adduct

of TCTA and a fragment of TCTA [TCTA+PhCz]+. Such an observation reveals an

obvious TCTA degradation pattern, where the C-N bond located at the central amine

moiety breaks down to produce phenol-carbazole (PhCz) and [TCTA-PhCz] radicals. This

homolytic dissociation mechanism is the main degradation pathway of TCTA, which

agrees with the finding in other reported studies [9, 14, 16].

123

Figure 6.10: LDI-TOF-MS spectra of device B1 before and after degradation.

In the aged sample, four more peaks become prominent, namely those at 538, 779,

1073 and 1277. These four peaks are identified as a protonated TmPyPB cation

([TmPyPB+H]+), adduct between TmPyPB and a fragment of TCTA ([TmPyPB+PhCz]+),

protonated TmPyPB dimer ([TmPyPB+TmPyPB]+) and adduct between TCTA and

TmPyPB ([TCTA+TmPyPB]+), respectively. Table 6.2 lists the mass values, fragment

origins and proposed structures of the peaks. The protonation process of TmPyPB is still

unclear. However, it can be concluded that at the TCTA/TmPyPB interface, there exist

various species of degradation products, such as neutral radicals, charged ions and reaction

products. Therefore, non-radiative recombination and exciton quenching would lead to a

drop in device efficiency and a short lifetime. Such phenomena can also explain the shorter

124

device lifetime of Devices A2 and A3 in which the recombination is confined to the

TCTA/TmPyPB interface.

Table 6.2: List of mass peaks and their proposed structures.

Mass (m/z) Origin Nature Proposed structure

539 TmPyPB Adduct

779 TCTA,

TmPyPB Fragment

982 TCTA Adduct

1073 TmPyPB Adduct

1277 TCTA,

TmPyPB Adduct

125

As a comparison, Device B2 (in which TmPyPB is replaced with DCzPPy) was

fabricated with the following structure: ITO/HATCN (3 nm)/TCTA (40 nm)/DCzPPy (40

nm)/LiF (1 nm)/Al. In this structure, hole are transported through TCTA and DCzPPy, and

electrons are transported via DCzPPy. Recombination is extended throughout the DCzPPy

layer. Two samples of Device B2 (before and after electrical aging) were analyzed by LDI-

TOF. The resultant spectra are illustrated in Figure 6.11. Compared to Device B1, Device

B2 exhibited little difference between samples before and after aging. Three major peaks

(499 and 741 belong to TCTA, and 562 belongs to DCzPPy) are observed in the two

samples without noticeable new peaks after degradation. This finding means that the

TCTA/DCzPPy interface is more stable than the TCTA/TmPyPB interface. Relative

intensity of peak 562 (DCzPPy) increases after 24 h of aging, indicating that DCzPPy is

more stable than TCTA in supporting electron and hole transport.

Figure 6.11: LDI-TOF-MS spectra of Device B2 before and after degradation.

126

The LDI-TOF-MS study of Devices B1 and B2 with TmPyPB as a strong electron

acceptor shows that this material is very unstable when in direct contact with a strong

electron donor such as TCTA. Protonated TmPyPB ([TmPyPB+H]+) and reaction products

between TCTA radicals and TmPyPB radicals can be observed after electrical aging. To

solidify this finding, two samples of Device B3 were fabricated with TAPC as the HTL.

The detailed device structure is as follows: ITO/HATCN (3 nm)/TAPC (40 nm)/DCzPPy

(40 nm)/LiF (1 nm)/Al. Peaks 465, 557, 570, 591 and 626 are from TAPC and its

fragments, all of which were observed in both B3 samples. However, after 24 h of aging,

peak 538 (protonated TmPyPB) appears in the spectra, which was also found in aged

Device B1.

Figure 6.12: LDI-TOF-MS spectra of Device B3 before and after degradation.

127

6.3. Conclusions

We have established a methodology to study the degradation of a set of OLED

materials through LDI-TOF-MS, TOF/TOF measurement, and DFT calculations. We

proposed that TAPC as a hole transport material undergoes a cation-induced cyclohexyl

ring-opening reaction as a degradation pathway during device operation. This finding

suggests that TAPC is not a desirable HTM for OLED devices due to the instability of its

radical cation. The Ir(iprpmi)3 blue phosphorescent dopant’s degradation is mainly induced

by metal-organic ligand dissociation. Host material TCTA mainly undergoes C-N bond

breaking at the central amine moiety. TCTA radicals and fragments can react with

TmPyPB radicals when the recombination is located at the TCTA/TmPyPB interface.

TmPyPB undergoes protonation and is likely to also undergo dimerization. DCzPPy is

relatively stable as a transport and host material. These findings are in agreement with the

blue PhOLED device performance and lifetime data.

128

References

[1] J. Zhuang, W. Li, W. Su, Y. Liu, Q. Shen, L. Liao, and M. Zhou, Org. Electron.

Physics, Mater. Appl. 14, 2596 (2013).

[2] H. Lee, H. Ahn, and C. Lee, J. Inf. Disp. 12, 219 (2011).

[3] K.S. Yook, O.Y. Kim, and J.Y. Lee, Synth. Met. 161(23), 2677 (2012).

[4] K.P. Klubek, C.W. Tang, and L.J. Rothberg, Org. Electron. Physics, Mater. Appl.

15, 1312 (2014).

[5] V. Sivasubramaniam, F. Brodkorb, S. Hanning, H.P. Loebl, V. van Elsbergen, H.

Boerner, U. Scherf, and M. Kreyenschmidt, J. Fluor. Chem. 130, 640 (2009).

[6] R. Seifert, I. Rabelo De Moraes, S. Scholz, M.C. Gather, B. Lüssem, and K. Leo,

Org. Electron. Physics, Mater. Appl. 14, 115 (2013).

[7] I.R. De Moraes, S. Scholz, B. Lüssem, and K. Leo, Org. Electron. 12, 341 (2011).

[8] K.P. Klubek, S.C. Dong, L.S. Liao, C.W. Tang, and L.J. Rothberg, Org. Electron.

Physics, Mater. Appl. 15, 3127 (2014).

[9] S. Scholz, K. Walzer, and K. Leo, Adv. Funct. Mater. 18, 2541 (2008).

[10] I. Rabelo De Moraes, S. Scholz, B. Lüssem, and K. Leo, Org. Electron. 13, 1900

(2012).

[11] S. Scholz, R. Meerheim, K. Walzer, and K. Leo, Proc. SPIE 6999, 6999 1B (2008).

[12] S.J. Su, C. Cai, and J. Kido, Chem. Mater. 23, 274 (2011).

[13] L.S. Cui, J.U. Kim, H. Nomura, H. Nakanotani, and C. Adachi, Angew. Chemie -

Int. Ed. 55, 6864 (2016).

[14] D.Y. Kondakov, J. Appl. Phys. 104, 084520 (2008).

129

[15] A. Hoteling, W. Nichols, D. Giesen, J. Lenhard, and R. Knochenmuss, Eur. J. Mass

Spectrom. 12, 345 (2006).

[16] V. Sivasubramaniam, F. Brodkorb, S. Hanning, O. Buttler, H.P. Loebl, V. van

Elsbergen, H. Boerner, U. Scherf, and M. Kreyenschmidt, Solid State Sci. 11, 1933

(2009).

130

Chapter 7 Summary and Future Work

This thesis is mainly focused on device engineering of WOLEDs based on

phosphorescent red, green and blue emitters. High-efficiency WOLEDs were achieved by

optimizing the layer structures and material compositions and gaining a better

understanding of the device operation and degradation mechanisms, including the possible

chemical degradation pathways of various material components. A systematic device

degradation study was conducted utilizing the LDI-TOF technique.

In Chapter 3, a dual-EML WOLED structure was studied. The effects of a host

material type, compositions of the mixed-host system, dopant concentration of the

red/green co-doped layer and non-emissive interlayer were investigated. By optimizing all

of these parameters, WOLEDs with high power efficiency (~33 lm/W at 1,000 cd/m2) and

high EQE (~18%) have been achieved.

In Chapter 4, we successfully fabricated high-efficiency WOLEDs with a reduced

color shift. An emitter structure consisting of an ultra-thin red and green co-doped layer

sandwiched in between two blue layers was developed. This emitter structure provides the

flexibility to fine tune the WOLED performance, including the color shift without

compromising the efficiency. By adjusting the compositions and thicknesses of the

individual layers, an EQE of almost 20% and a luminance of over 40 lm/W (at 1,000 cd/m2)

have been achieved with minimal color shift over a large range of luminance (400–4000

cd/m2). The function of all three EMLs and exciton energy-transfer mechanisms among

hosts and guests were systematically studied. We also found that the device degradation

was related to the lifetime of excited FIrpic states and dependent on the host materials.

131

WOLEDs fabricated in Chapters 3 and 4 were based on the common FIrpic blue

phosphorescent dopant. Although high-efficiency WOLEDs were successfully

demonstrated with FIrpic, they were very short-lived. In Chapter 5, we investigated another

iridium emitter with phenyl-imidazole ligands, namely Ir(iprpmi)3, which is known to be

relatively more stable. We conducted studies of the charge-carrier transport properties of

Ir(iprpmi)3 in hole-only and electron-only devices and found that Ir(iprpmi)3 trapped holes

efficiently at low concentrations but transported holes at high concentrations in the bipolar

host material, DCzPPy. With Ir(iprpmi)3/DCzPPy as a blue emitter, we fabricated high-

efficiency blue and white OLEDs with EQEs over 20%. Moreover, the power efficiency

of WOLEDs was close to 40 lm/W at 1,000 cd/m2. By varying the concentration of the

hole-trapping Ir(iprpmi)3 in the bipolar DCzPPy host, high-efficiency WOLEDs with

minimal color shift (±0.001, ±0.005) over a current density range of 0.05–20 mA/cm2 were

fabricated. Compared to FIrpic-based devices, Ir(iprpmi)3-based WOLEDs exhibited

significantly longer lifetimes under similar test conditions.

Although the WOLEDs discussed in Chapter 5 exhibited much improved lifetimes

compared to FIrpic-based devices, their half-lifetimes were still below 100 h. To

understand the causes of device instability, we investigated in Chapter 6 the possible

chemical degradation pathways using LDI-TOF, TOF/TOF techniques and DFT

calculations. From the LDI-TOF fragmentation patterns, we found that Ir(iprpmi)3 was

more stable than FIrpic, possibly due to the lack of fluorine substituents and the ancillary

picolinate ligand in Ir(iprpmi)3. The main degradation of Ir(iprpmi)3 was induced by metal-

organic ligand dissociation, which was caused by the Ir-C and Ir-N bonds’ cleavage. We

132

found that TAPC as a hole transport material was very unstable during device operation.

In OLED devices, TAPC can undergo a cation-induced cyclohexyl ring-opening reaction

as the main degradation pathway, and the C-N bond dissociation appears to be a minor

degradation pathway. By contrast, TCTA mainly undergoes C-N bond dissociation at the

central amine moiety. Bipolar host material DCzPPy is relatively stable. There were hardly

any noticeable differences in the LDI-TOF fragmentation patterns of the aged and unaged

devices. We also found that electron-transporting material TmPyPB was another main

cause of device degradation. The LDI-TOF-MS analysis of TCTA/TmPyPB indicated that

the TCTA radical and its various fragments can react with TmPyPB radicals at the

TCTA/TmPyPB interface. We also found that TmPyPB can undergo protonation and

dimerization.

Although we have successfully demonstrated high-efficiency and color-stable

WOLEDs, the findings of this study indicated that there is a critical need to further improve

the device lifetime while maintaining high efficiency. Proposed future work is discussed

as follows:

1) From a device engineering perspective, it is desirable to broaden the electron-

hole recombination region. As discussed in Chapter 3, a wide recombination

region can improve the overall device efficiency by alleviating charge-carrier

and exciton accumulation at the emitting layer interfaces. The LDI-TOF

analysis in Chapter 6 indicated that exciton quenchers formed at either the

EML/HTL or EML/ETL interface can contribute to a reduced device efficiency

and lifetime. To reduce quenching at these interfaces, a bipolar material or a

133

mixture of electron donors and acceptors should be used as the host of the EML,

especially for the blue layer. Because the mobility of electron-transport

materials is generally much higher than the mobility of hole-transport materials,

it is not easy to find a mixed-host system without an excessive accumulation of

holes near the ETL/EML interface. A bipolar host material [1–5] with proper

hole-transporting and electron-transporting moieties can perform the multiple

functions of balanced injection, transportation and recombination of charge

carriers. It would be beneficial to investigate charge-carrier transporting

properties and material stabilities of bipolar host materials to find useful host-

material candidates for stable WOLEDs. Other than DCzPPy, other noteworthy

bipolar host materials include carbazole/diphenyl phosphine oxide (mCPPO1)

[6], phenyl-carbazole/pyridine (PCz-BFP) [7] and triphenyl

amine/benzimidazole (p-BISiTPA) [8].

2) The choice of a blue phosphorescent dopant is another key factor to achieving

stable WOLEDs. Unfortunately, there have been few reports of stable blue

phosphorescent dopants. As found in this work, Ir(iprpmi)3 exhibits an

improved device lifetime. Compared to FIrpic, Ir(iprpmi)3 has a more rigid and

bulky molecular structure in the phenyl-imidazole ligands, which may be the

reason for the improved stability observed in the Ir(iprpmi)3-based blue and

white OLED devices (Chapter 5). Based on this finding, we suggest that a

detailed structure–property relationship should be studied with a particular

attention on the effect of molecular rigidity on the blue OLED lifetime. There

134

have been a few reports of high-efficiency blue OLEDs based on Ir phenyl-

imidazole derivatives [9–12]. Therefore, device-lifetime tests and degradation

analysis can be conducted using these materials to determine the relationships

between molecular structures and device performances. Moreover, LDI-TOF

showed that one of the weakest bonds in iridium-based emitters is the Ir-C or

Ir-N bond. Therefore, non-metal-based dopants such as TADF may provide a

new material class for simultaneously achieving high efficiencies and long

lifetimes in WOLEDs. Although there have been few reports on lifetime studies

based on blue TADF materials [13, 14], WOLEDs incorporating TADF blue

emitters are worth investigating.

3) LDI-TOF analyses, including ours, have found that, in addition to degradation

of dopant materials with electrical stress, there are considerable evidence of

degradation in transport and host materials, including reactions between

dopants and transport materials. We have demonstrated that the prototypical

hole-transport material TAPC, commonly used to achieve high efficiency in

blue and white OLEDs, is unstable because of the C-C bond dissociation of its

radical cation. TCTA is relatively stable compared to TAPC, but it is also

susceptible to C-N bond dissociation. Therefore, the search for hole-transport

materials should go beyond the amine classes. LDI-TOF on HATCN-induced

cations can provide useful information to study the stability of cationic species

of hole-transport materials.

135

4) Most high-efficiency OLEDs utilize materials with aromatic pyridyl nitrogen

moieties for the ETL. These materials tend to produce high triplet energy and,

therefore, can help confine the triplet excitons to the EML. However, these

materials also tend to be unstable when used in blue OLED devices. As shown

in Chapter 6, chemical reaction products of TmPyPB were detected in degraded

devices with a TCTA/TmPyPB bilayer structure. While LDI-TOF can be

readily used for analyzing degradation products from electron donors such as

the hole-transport aromatic amines and Ir dopants, it is not particularly useful

for detecting degradation products from electron acceptors such as TmPyPB.

LDI-TOF may be used to probe the interface degradation between the ETL and

HTL, where the degradation products involving complexation between donor

and acceptor may have sufficiently low ionization potentials to be detectable.

5) LDI typically involves complex chemistry processes, including intramolecular

fragmentation, intermolecular energy and electron transfer, and reactions

between active species. The observed difference in the fragmentation patterns

between aged and unaged OLED samples may account for the presence of a

minute amount of degraded materials. Moreover, due to the high energy of the

nitrogen laser (337 nm) used in LDI, molecular fragmentation will inevitably

occur during the LDI process, making it extraordinarily difficult to assess the

degradation that results from device aging alone. Therefore, it is important to

perform in situ analysis in which control and test samples can be carefully

analyzed under identical conditions. Furthermore, it would be helpful to further

136

refine the LDI techniques in which quantitative results can be obtained. Other

techniques such as liquid chromatography mass spectrometry [15–17] should

also be employed to supplement the LDI analysis in detecting photochemical

and degradation products in OLED devices.

137

References

[1] A. Chaskar, H.F. Chen, and K.T. Wong, Adv. Mater. 23, 3876 (2011).

[2] H.H. Chou and C.H. Cheng, Adv. Mater. 22, 2468 (2010).

[3] D. Wagner, S.T. Hoffmann, U. Heinemeyer, I. Münster, A. Köhler, and P.

Strohriegl, Chem. Mater. 25, 3758 (2013).

[4] S.J. Su, H. Sasabe, T. Takeda, and J. Kido, Chem. Mater. 20, 1691 (2008).

[5] F.-M. Hsu, C.-H. Chien, C.-F. Shu, C.-H. Lai, C.-C. Hsieh, K.-W. Wang, and P.-T.

Chou, Adv. Funct. Mater. 19, 2834 (2009).

[6] S.O. Jeon, S.E. Jang, H.S. Son, and J.Y. Lee, Adv. Mater. 23, 1436 (2011).

[7] C.W. Lee and J.Y. Lee, Adv Mater 25, 596 (2013).

[8] S. Gong, Y. Chen, C. Yang, C. Zhong, J. Qin and D. Ma, Adv Mater, 22, 5370

(2010).

[9] K. Udagawa, H. Sasabe, C. Cai, and J. Kido, Adv. Mater. 26, 5062 (2014).

[10] J. Zhuang, W. Li, W. Wu, M. Song, W. Su, M. Zhou, and Z. Cui, New J. Chem.

39, 246 (2015).

[11] T. Karatsu, M. Takahashi, S. Yagai, and A. Kitamura, Inorg. Chem. 52, 12338

(2013).

[12] B. D’Andrade, P. B. MacKenzie, M. S. Weaver, and J. J. Brown, U.S. Patent No.

8,557,399 (2013)

[13] M. Kim, S.K. Jeon, S.H. Hwang, and J.Y. Lee, Adv. Mater. 27, 2515 (2015).

[14] Q. Zhang, B. Li, S. Huang, H. Nomura, H. Tanaka, and C. Adachi, Nat. Photonics

8, 1 (2014).

138

[15] D.Y. Kondakov, J. Appl. Phys. 104, 084520 (2008).

[16] V. Sivasubramaniam, F. Brodkorb, S. Hanning, H.P. Loebl, V. Elsbergen, H.

Boerner, U. Scherf, and M. Kreyenschmidt, Cent. Eur. J. Chem. 7, 836 (2009).

[17] D. Kondakov, Proc. SPIE 7051, 70510I (2008).