investigation of blue phosphorescent organic light-emitting diode instability
TRANSCRIPT
Investigation of Blue Phosphorescent Organic Light-Emitting
Diode Instability
By
Kevin P. Klubek
Submitted in Partial Fulfillment of the
Requirements for the Degree
Doctor of Philosophy
Supervised by
Professor Ching W. Tang
Department of Chemical Engineering
Arts, Sciences and Engineering
Edmund A. Hajim School of Engineering and Applied Sciences
University of Rochester
Rochester, New York
2014
ii
Dedicated To
My Wonderful Wife Kimberly
And
Amazing Children Nathaniel, Zachary, and Julia
iii
BIOGRAPHICAL SKETCH
Kevin Paul Klubek was born in 1972 in Lackawanna, New York. In 1996 he
received a Bachelor of Science degree in Medicinal Chemistry from the University of
Buffalo, after which he accepted a position as synthetic chemist at the Kodak Research
Laboratories in Rochester, New York to work on the synthesis of organic materials for
organic light emitting diode applications. From 2007, he began pursuing his doctorate
part-time in Chemical Engineering at the University of Rochester under the supervision
of Professor Ching W. Tang. In 2010 he left Kodak to purse his doctorate full-time. In the
same year, he was awarded a NSF Graduate Research Fellowship. In 2012 he received
his Master of Science Degree. In 2013, he received a NSF East Asia and Pacific Summer
Institute award and spent eight weeks conducting OLED research with Professor Liang-
Sheng Liao at Suzhou University, China. His field of study was in organic materials and
device physics related to Organic Light-Emitting Diodes.
PUBLICATIONS AND PAPERS SUBMITTED FOR
PUBLICATION
1. H. Wang, K. P. Klubek, C.W. Tang, Appl. Phys. Lett., 93 (2008) 093306.
2. K.P. Klubek, C.W. Tang, L.J. Rothberg, Org. Electron., 15 (2014) 1312.
3. K.P. Klubek, S.C. Dong, L.S. Liao, C.W. Tang, L.J. Rothberg, Org. Electron.,
submitted June 2014.
iv
ACKNOWLEDGEMENTS
I would like to first thank my thesis advisor and mentor Professor Ching W. Tang
for providing me with his guidance, support and remarkable insights in conducting
research related to OLED materials, device fabrication, and device physics. I have
learned so much in all the time we have worked together. It was truly an honor and a
privilege to have had the opportunity to work with Professor Tang.
I would like to thank Professor Shaw H. Chen for his support and encouragement
in pursuing my PhD studies at the Department of Chemical Engineering and for the
research collaboration we have undertaken in the DOE project. I also thank Professor
Lewis J. Rothberg, Professor Alexander A. Shestopalov, and Professor Todd D. Krauss
for serving on my thesis committee. I would also like to express my gratitude to Professor
Rothberg for his perspicacity and support throughout our collaboration.
I would like to thank Mr. Joseph Madathil for teaching and helping me with
everything related to vacuum systems/equipment and electronics. It was a pleasure
working with Joe. I could not have imagined building the KOMET coater without all of
his help, guidance and support. The KOMET most definitely did not come from Mars. I
also would like to thank Mike Culver, who was extremely helpful in sharing his technical
and equipment building experience with me. He was always willing to help in any way,
especially with providing all types of lab equipment and supplies. I also like to thank Dr.
Ralph H. Young for our many stimulating conversations. He was extremely patient and
helpful in guiding me in the field of electrostatics. I would like to express my sincere
v
gratitude to Professor Liang-Sheng (Larry) Liao and Dr. Shou-Cheng Dong of Suzhou
University for hosting my visit in Suzhou, and for the continuing collaboration. I would
also like to thank Dr. David Weiss for providing valuable advice related to experiments
and for providing help with proofreading this thesis and in presentations.
My gratitude and appreciation also goes to Master Machinists Mr. John Miller,
Mr. John Gresty, Mr. Richard Fellows and Mr. Jim Despard for all their assistance and
training that they provided to me in the machine shop.
To all the Professors in the Chemical Engineering Department and to Mrs. Sandra
Willison, Mrs. Gina Eagan, Mr. Larry Kuntz, Mrs. Tiffany Landers, Ms. Jennifer Condit,
and Victoria Heberling, I would like to say “thank you” for your support and assistance
with administrative issues.
I also want to acknowledge my fellow lab-mates for their collaboration and
camaraderie: Minlu Zhang, Mohan Ahluwalia, Wei Xia, Hao Lin, Hui Wang, Hsiang Ning
(Sunny) Wu, Felipe Angel, Yung-Hsin (Thomas) Lee, Mathew Smith, Charles Chan, Sang-
Min Lee, Lichang Zeng, Jason Wallace, Prashant Kumar Singh, Laura Ciammaruchi,
Jonathan Welt, Guy Mongelli, Chris Favaro, Erik Glowacki, Lisong Xu, Aanand
Thiyagarajan, Michael Beckley, Thao Nguyen, Meng-huan (Kinneas) Ho, William Finnie,
Sihan (Jonas) Xie, and all of our group alumni.
To Eastman Kodak Company I am grateful for the financial support and “time-off” in
my pursuit of graduate studies.
In addition to all the academic fun, I truly enjoyed all our Group events, going to
the Cantonese House, The Distillery, Sheridans, and cookout Tuesdays. Tuesday will
vi
forever be known as burger day! Mike Culver’s “burger time, burger time” will ring like
a bell. Mike, thank you for the awesome cookies. I’ll never be able to eat another burger
or hotdog without getting a cookie afterwards. And there was tennis! It was great fun
playing tennis with “Bazooka” Joe Madathil, Felipe Angel, Minlu Zhang, Lisong Xu,
Prashant Kumar Singh, and Aanand Thiyagarajan. And to Laura Ciammaruchi, I would
like to thank you for introducing me to squash and for all the fun we had. I’m still not
sure if we’ve really played the “Queen” match yet. Oopha…
To my Mom and Dad, Marian and Gerald, Mother-in-law Diana, and all my
siblings and their families, I thank you all for your support and encouragement.
This thesis would not have been possible without the unconditional love, support
and encouragement provided by my wife Kimberly and children Nathaniel, Zachary, and
Julia. I am extremely grateful to have such an amazing and wonderful family. I love you
all so much.
vii
ABSTRACT
Phosphorescent emitters have played a critical role in the advance of organic
light-emitting diode (OLED) technology. Among the phosphorescent emitters, the
cyclometalated complexes with iridium (III) as the central atom are the most well
developed and successfully commercialized. They are also among the most efficient.
With chelating ligands specifically designed for color tuning, RGB (red, green and blue)
emitters with nearly 100% (internal) quantum efficiency, defined as photons generated
per injected electron, have been demonstrated. The operating lifetime of the
phosphorescent OLED devices, however, remains an issue, particularly for the blue
devices.
This thesis focuses on two phosphorescent blue dopants based on iridium; bis(4,6-
difluorophenyl-pyridinato-N,C2) picolinate iridium (III) (FIrpic), and tris[1-(2,6-
diisopropylphenyl)-2-phenyl-1H-imidazole] iridium (III) (Ir(iprpmi)3). FIrpic is perhaps
the most well-known blue phosphorescent dopant that has demonstrated high quantum
efficiency, but it has been found to produce short operational lifetimes. Ir(iprpmi)3 was
chosen for this study because there were claims in patent literature that it provides
excellent stability in OLED devices.
Using FIrpic as the blue dopant in the emitter layer, we have investigated the
dependence of OLED performance, including device lifetime, on the compositions of the
emitter layer (host-dopant) and the adjacent electron/hole transport layers. Regardless of
the choice of the host materials in the emitter layer, or the materials in the transport
viii
layers, it is shown that the stability of OLED devices is poor (<12 hours) for devices
using FIrpic as the blue phosphorescent dopant. Recombination that occurs on the host or
transport materials is also found to be detrimental to device lifetime.
By tracking voltage and efficiency during operation for model devices with well
known tris(8-hydroxyquinolinato) aluminum (Alq) as the emitter, it is shown that in
addition to its instability in charge recombination processes, FIrpic is also unstable with
respect to hole-transport processes. This indicates that the FIrpic radical cation itself is
unstable. It is also found that the host 3,3'-bis(N-carbazolyl)biphenyl (mCBP) is unstable
to electron-hole recombination processes.
The photophysical properties of Ir(iprpmi)3 were studied. This material has a low
ionization potential of 4.8 eV, which is indicative of a material possessing strong electron
donor characteristics. Electron donating materials are typically associated with hole
injecting or hole transporting materials. By adjusting the concentration of Ir(iprpmi)3
within the host mCBP, we show that Ir(iprpmi)3 is capable of trapping (at low
concentration) and transporting (at high concentration) holes. In this way, the location of
the recombination zone (RZ) is modulated. At low concentrations the RZ is confined near
the hole-transport layer whereas at high concentration the RZ is shifted towards the
electron-transport layer. The external quantum efficiency, EQE, defined as photons
exiting the device per injected electron, can reach >20% as long as the RZ is adjacent a
charge transport material that has a higher triplet energy than Ir(iprpmi)3. Due to
molecular structural differences, device lifetime using Ir(iprpmi)3 as the emitting dopant
is significantly improved compared to FIrpic. However, the lifetime is also highly
ix
dependent on the choice of material for the host and charge transport layers.
x
CONTRIBUTORS AND FUNDING SOURCES
This work was supervised by a dissertation committee consisting of Professors Ching
W. Tang (advisor) and Alexander A. Shestopalov of the Department of Chemical
Engineering, Professors Lewis J. Rothberg and Todd D. Krauss (chair) of the Department of
Chemistry.
For Chapter 2: The organic boat design was based on an initial design by fellow lab-
member Sang-min Lee. Mr. Joseph Madathil and Professor Tang both contributed to the
design and building of the vacuum coater for fabricating OLED devices.
For Chapter 3: My fellow lab-mate Lisong Xu fabricated the undoped TCTA device.
The data analysis was conducted in part by Professor Ching W. 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 2014, in an article listed in the
Biographical Sketch.
For Chapter 5: Dr. Shou-Cheng Dong of Suzhou University in China helped to
acquire photophysical, electrochemical, and theoretical data related to the iridium
cyclometalated dopant Ir(iprpmi)3. The analyses were conducted in part by Dr. Dong,
Professor Tang, Professor Rothberg, and Professor Liang-Sheng Liao of Suzhou University in
China and were submitted for publication in 2014, in an article listed in the Biographical
Sketch.
All other work conducted for this dissertation was completed by Kevin P. Klubek
independently.
This thesis was funded by the National Science Foundation (NSF) under Grant
xi
No. 1316969, the NSF Graduate Research Fellowship Program under Grant No. 0935947
and the Department of Energy (DOE) under Award No. DE-EE0003296. Any opinions,
findings, and conclusions or recommendations expressed in this material are those of the
author and do not necessarily reflect the views of the NSF or DOE.
xii
TABLE OF CONTENTS
Biographical Sketch iii
Acknowledgements iv
Abstract vii
Contributors and Funding Source x
Table of Contents xii
List of Figures xvii
List of Tables xxvi
CHAPTER 1 BACKGROUND AND INTRODUCTION 1
1.1 Organic Electroluminescence 1
1.2 Organic Light-Emitting Diode Operating Principle 1
1.3 Exciton Formation 4
1.4 Device Efficiency 5
1.5 Iridium Cyclometalated Complexes 6
1.6 Device Lifetime 8
1.7 Degradation 10
1.8 Objectives and Outline of Thesis 12
xiii
References 15
CHAPTER 2 DEVICE FABRICATION EXPERIMENTAL DETAILS 21
2.1 Thermal Evaporation Coater 21
2.2 The KOMET Coater 23
2.3 Organic Boats 29
2.4 Device Fabrication 31
2.5 OLED Performance Measurements 34
2.6 Device Lifetime 34
2.7 Materials 37
References 42
CHAPTER 3 FIRPIC AS THE LIGHT-EMITTING DOPANT FOR
PHOSPHORESCENT ORGANIC LIGHT-EMITTING DIODES 45
3.1 Introduction 45
3.2 Experimental 47
3.3 Mixed Host 47
3.3.1 TCTA:UGH3 (1:1) Mixed Host with Varied FIrpic Concentration 49
xiv
3.3.2 Mixed TCTA:UGH3 (x:y) Hosts with Fixed Flrpic Concentration 52
3.3.3 Effect of Electron TransportMaterials 54
3.3.4 Device Lifetime 57
3.4 Single Host 58
3.4.1 TCTA Single Host and FIrpic Concentration Series 59
3.4.2 TCTA Single Host and UGH3 as Electron Transport Material 62
3.4.3 TCTA Host and BAlq as Electron Transport Material 67
3.4.4 mCBP Host and FIrpic Concentration Series 69
3.4.5 mCBP Host and BAlq as Electron Transport Material 73
3.4.6 mCBP Host and NPB as Hole Transport Layer 76
3.5 Conclusions 78
References 81
CHAPTER 4 CHARGE TRANSPORT THROUGH FIRPIC 86
4.1 Introduction 86
4.2 Experimental 87
4.3 Results and Discussion 88
xv
4.3.1 Undoped Hole Transport Layers 89
4.3.2 FIrpic Doped into mCBP Adjacent to the Anode 91
4.3.3 FIrpic Doped mCBP Sandwiched Between Undoped Hole Transport
Materials 94
4.3.4 FIrpic Doped into NPB 96
4.4 Conclusions 97
References 98
CHAPTER 5 INVESTIGATING BLUE PHOSPHORESCENT IRIDIUM DOPANT
WITH PHENYL-IMIDAZOLE LIGANDS 100
5.1 Introduction 100
5.2 Materials and Methods 101
5.3 Results and Discussion 103
5.3.1 Photophysical and Electrochemical Measurements 103
5.3.2 Low EQE OLEDs 105
5.3.3 High EQE OLEDs 107
5.3.4 Probing the Recombination Zone 111
xvi
5.3.5 High EQE, Low Voltage OLEDs 114
5.3.6 Device Lifetime 116
5.4 Conclusions 119
References 120
CHAPTER 6 SUMMARY AND FUTURE RESEARCH 124
6.1 Summary 124
6.2 Future Research 127
References 132
xvii
LIST OF FIGURES
Figure 1.1 OLED configuration and molecular structures used by
Tang and VanSlyke. 2
Figure 1.2 State-of-the-art OLED device structure. 3
Figure 1.3 Blue [Ir(F2ppy)3], green [Ir(ppy)3] and red [Ir(piq)3] Ir-
based phosphorescent dopants and maximum emission
peaks (max). 7
Figure 1.4 Commission Internationale De L'Eclairage (CIE) 1931
color coordinates. 9
Figure 2.1 Basic design of a thermal evaporation coater. 22
Figure 2.2 KOMET coater. 24
Figure 2.3 Inside the KOMET coater chamber. (a) Top of the
chamber, (b) bottom of the chamber, (c) area for depositing
organic materials, and (d) area for depositing
metals/inorganics. 25
Figure 2.4 KOMET coater top plate. 26
Figure 2.5 Features below the KOMET bottom plate. 27
xviii
Figure 2.6 Control cabinet 28
Figure 2.7 Picture of components used to build organic boat. 30
Figure 2.8 Components for assembling organic boat. 30
Figure 2.9 ITO pattern on substrate. 32
Figure 2.10 Pictures of OLED active area and encapsulation technique
(a) ITO pattern, organic deposited through circular mask,
and metal deposited through L-shaped mask, (b) An OLED
device with 4 lit pixels, and (c) Vacuum assembly for
encapsulating devices. 33
Figure 2.11 Output from John Burtis software program (a) current
density (mA/cm2), (b) drive voltage (V), (c) luminance
(cd/m2), (d) CIEx coordinate, (e) CIEy coordinate, and (f)
EQE or p/e (photon/electron). 35
Figure 3.1 Molecular structures for materials discussed in Chapter 3. 48
Figure 3.2 FIrpic concentration series for TCTA:UGH3 mixed host
OLED devices. 49
Figure 3.3 OLED performance data at 5 mA/cm2 for varying FIrpic
concentration in mixed host devices; Anode/35 nm
TAPC/40 nm TCTA+UGH3 (1:1)+x% FIrpic/30 nm
xix
TmPyPB/20 nm BPhen/Cathode, x=9–26%; (a) EL spectra
at 5 mA/cm2, (b) EQE vs. current density, (c) current
density vs. voltage. 51
Figure 3.4 Varying the mixed host doping ratio using FIrpic as
emitting dopant. 52
Figure 3.5 OLED performance data at 5 mA/cm2 for varying mixed
host composition while keeping FIrpic at 14%; Anode/35
nm TAPC/40 nm TCTA+UGH3 (x:y)+14% FIrpic/30 nm
TmPyPB/20 nm BPhen/1 nm Cathode, x+y=86%; (a) EL
spectra at 5 mA/cm2, (b) EQE vs. current density, (c)
current density vs. voltage. 53
Figure 3.6 Removing TmPyPB and using BAlq or BPhen next to the
LEL for TCTA:UGH3 mixed host devices. 54
Figure 3.7 OLED performance data at 5 mA/cm2 for mixed host
devices with alternate ETL’s; Anode/35 nm TAPC/40 nm
TCTA+UGH3 (1:1)+12% FIrpic/ETL/Cathode. ETL is 50
nm BPhen or 10 nm BAlq/50 nm BPhen; (a) EL spectra at
5 mA/cm2, (b) EQE vs. current density, (c) current density
vs. voltage. 55
xx
Figure 3.8 HOMO, LUMO, Triplet Energy levels for materials used to
fabricate OLED devices in Chapter 2. Triplet Energies are
in parenthesis. Shaded boxes indicate typical electron
transport materials. All values in eV. 57
Figure 3.9 Device lifetime taken at 5 mA/cm2 for mixed-host devices;
Anode/35 nm TAPC/40 nm TCTA+UGH3 (1:1)+12%
FIrpic/ETL/Cathode. ETL is 50 nm BPhen, 10 nm BAlq/50
nm BPhen, or 30 nm TmPyPB/20 nm BPhen. 58
Figure 3.10 FIrpic concentration series for TCTA single host OLED
devices. 59
Figure 3.11 EL spectrum at 20 mA/cm2 showing exciplex of
TCTA:TmPyPB for undoped device. Anode/30 nm
TAPC/30 nm TCTA/30nm TmPyPB/Cathode. 61
Figure 3.12 Removing TmPyPB and using UGH3 or BPhen next to the
LEL for single host devices with FIrpic concentration at
1% or 12%. 62
Figure 3.13 EL spectra at 5 mA/cm2 for devices using UGH3 as the
ETL and 1% FIrpic in the LEL. Anode/35 nm TAPC/40 nm
TCTA+12% FIrpic/y nm UGH3/50 nm BPhen/Cathode,
y=0–20 nm. 66
xxi
Figure 3.14 BAlq or BPhen next to the LEL for TCTA single host
devices. 67
Figure 3.15 FIrpic concentration series for mCBP single host devices. 69
Figure 3.16 OLED performance for devices with mCBP as host and 2-
18% FIrpic, Anode/35 nm TAPC/mCBP+x% FIrpic/30 nm
TmPyPB/20 nm BPhen/Cathode. 71
Figure 3.17 EL spectrum at 20 mA/cm2 showing exciplex of
mCBP:TmPyPB. Anode/75 nm mCBP/75 nm
TmPyPB/Cathode. 72
Figure 3.18 Removing TmPyPB and using BAlq or BPhen next to the
LEL for mCBP single host devices with the FIrpic
concentration at 6% or 12%. 73
Figure 3.19 Using BAlq as ETL and NPB as HTL at FIrpic
concentrations of 6% and 12%. 76
Figure 4.1 Molecular structures and HOMO/LUMO energy levels of
materials used in devices. Units in electronvolts (eV). 87
Figure 4.2 EL spectra at 20 mA/cm2 for the following device
architecture: ITO/1.0 nm MoOx/75 nm HTL/75 nm Alq/1.0
nm LiF/100 nm Al, where the HTL’s are: Ref [75 nm
NPB], A1 [75 nm mCBP], B4 [37.5 nm mCBP+12%
xxii
FIrpic|37.5 nm NPB], C1 [37.5 nm mCBP|27.5 nm
mCBP+12% FIrpic|10 nm NPB], D1 [37.5 nm mCBP|37.5
nm mCBP+12% FIrpic], E1 [75 nm NPB+6% FIrpic]. 89
Figure 4.3 OLED performance for devices with an undoped HTL
adjacent to MoOx. ITO/MoOx/HTL/Alq/LiF/Al. Ref[NPB],
A1[mCBP], A2[mCBP/NPB]. (a) Initial current density-
voltage-luminance (J-V-L) characteristics, (b) Device
lifetime at 20 mA/cm2, (c) Operational voltage rise at 20
mA/cm2. 90
Figure 4.4 OLED performance for devices with FIrpic doped mCBP
adjacent to MoOx. ITO/MoOx/mCBP+x%
FIrpic/Alq/LiF/Al. B1[1%], B2[3%], B3[6%], B4[12%]. (a)
Initial current density-voltage-luminance (J-V-L)
characteristics, (b) Device lifetime at 20 mA/cm2, (c)
Operational voltage rise at 20 mA/cm2. 92
Figure 4.5 OLED performance for devices with FIrpic doped mCBP
spaced away from MoOx and for a device with FIrpic
doped in NPB. ITO/MoOx/mCBP/HTL/NPB/Alq/LiF/Al:
C1[mCBP+12%], ITO/MoOx/HTL/Alq/LiF/Al]: D1[mCBP/mCBP
+12% FIrpic] and E1[NPB+6% FIrpic]. (a) Initial current density-
voltage-luminance (J-V-L) characteristics, (b) Device
xxiii
lifetime at 20 mA/cm2, (c) Operational voltage rise at 20
mA/cm2. 95
Figure 5.1 Molecular structures and energy level diagram of the
materials (values in eV). 102
Figure 5.2 UV-Vis absorption and PL spectra of Ir(iprpmi)3 in
CH2Cl2. Inset: PL spectrum of Ir(iprpmi)3 in 2-MeTHF at
77 K. 103
Figure 5.3 Simulated frontier molecular orbitals of Ir(iprpmi)3. (a)
HOMO electron density, (b) ball and stick model, (c)
LUMO electron density. 104
Figure 5.4 OLED performance data for 6% and 12% concentrations of
Ir(iprpmi)3 with NPB as HTL and BAlq as ETL. Anode/40
nm TAPC/30 nm mCBP+ 6% or 12% Ir(iprpmi)3/40 nm
BAlq/Cathode. (a) EL spectra at 5 mA/cm2, (b) current
density vs. voltage, (c) EQE vs. current density. 106
Figure 5.5 OLED performance data for Ir(iprpmi)3 concentration
series with TAPC HTL and BAlq ETL. Anode/40 nm
TAPC/30 nm mCBP+ x% Ir(iprpmi)3/40 nm
BAlq/Cathode. x=0%–24%. (a) EL spectra at 5 mA/cm2,
(b) current density vs. voltage, (c) EQE vs. current density. 108
xxiv
Figure 5.6 OLED performance data when probing the recombination
zone with 6% Ir(iprpmi)3. Anode/40 nm TAPC/x nm
mCBP+6% Ir(iprpmi)3/y nm mCBP/40 nm BAlq/Cathode.
[x=25 nm, y=5 nm], [x=20 nm, y=10 nm],[x=10 nm, y=20
nm], and [x=5 nm, y=25 nm]: (a) current density vs.
voltage, (b) EQE vs. current density. 112
Figure 5.7 OLED performance data when probing the recombination
zone with 25% Ir(iprpmi)3. Anode/40 nm TAPC/x nm
mCBP+25% Ir(iprpmi)3/y nm mCBP/40 nm BAlq/Cathode.
[x=25 nm, y=5 nm],[x=20 nm, y=10 nm], [x=10 nm, y=20
nm], and [x=5 nm, y=25 nm]: (a) current density vs.
voltage [ inset: voltage vs mCBP layer thickness of devices
under current densities of 1, 5, and 10 mA/cm2] (b) EQE
vs. current density. 113
Figure 5.8 OLED performance data for Ir(iprpmi)3 concentration
series with TmPyPB ETL. Anode/40 nm TAPC/30 nm
mCBP + x% Ir(iprpmi)3/40 nm TmPyPB/Cathode. x=3%–
24% (a) current density vs. voltage, (b) EQE vs. current
density. 115
Figure 5.9 Lifetime testing at 5 mA/cm2 for device structure,
Anode/40 nm HTL/30 nm mCBP+x% Ir(iprpmi)3/40 nm
xxv
ETL/Cathode, [NPB-6% dopant-BAlq], [NPB-12% dopant-
BAlq], [TAPC-7% dopant-BAlq], [TAPC-13% dopant-
BAlq], [TAPC-6% dopant-TmPyPB]. 116
Figure 6.1 Different types of blue phosphorescent dopants. 129
xxvi
LIST OF TABLES
Table 2.1 Organic materials discussed throughout this thesis.
HOMO/LUMO/triplet energies were taken from literature.
HOMO energies are based on ionization potential, LUMO
energies based on difference between optical bandgap and
HOMO energy, and triplet energy based on
phosphorescence spectra taken at 77K. 38
Table 3.1 OLED performance data at 5 mA/cm2 for varying FIrpic
concentration in mixed host devices; Anode/35 nm
TAPC/40 nm TCTA+UGH3 (1:1)+x% FIrpic/30 nm
TmPyPB/20 nm BPhen/Cathode, x=9–26%. 51
Table 3.2 Device data taken at 5 mA/cm2 for single host TCTA and
FIrpic concentrations 3-6%. 60
Table 3.3 OLED performance data at 5 mA/cm2 for devices using
UGH3 as the ETL and 12% FIrpic in the LEL; Anode/35
nm TAPC/40 nm TCTA+12% FIrpic/x nm TmPyPB/y nm
UGH3/z nm BPhen/Cathode. 63
Table 3.4 OLED performance data at 5 mA/cm2 for devices using
UGH3 as the ETL and 1% FIrpic in the LEL. Anode/35 nm
xxvii
TAPC/40 nm TCTA+12% FIrpic/x nm TmPyPB/y nm
UGH3/z nm BPhen/Cathode. 65
Table 3.5 OLED performance data for devices using TCTA as single
host and BAlq as the ETL. Anode/35 nm TAPC/40 nm
TCTA+12% FIrpic/x nm BAlq/50 nm BPhen/Cathode. 67
Table 3.6 OLED performance at 5 mA/cm2 for devices using mCBP
as host and varying the FIrpic concentration from 2–18%. 70
Table 3.7 OLED performance data at 5 mA/cm2 for devices using
mCBP as the single host and BAlq as the ETL. Anode/35
nm TAPC/40 nm mCBP+12% FIrpic/x nm BAlq/50 nm
BPhen/Cathode. 74
Table 3.8 OLED performance data at 5 mA/cm2 for devices using
mCBP as the single host and BAlq as the ETL. Anode/35
nm TAPC/40 nm mCBP+6% FIrpic/x nm BAlq/50 nm
BPhen/Cathode. 75
Table 3.9 OLED performance data at 5 mA/cm2 for devices using
NPB as HTL, BAlq as ETL, and x% FIrpic. Anode/35 nm
NPB/40 nm mCBP+x% FIrpic/y nm BAlq/50 nm
BPhen/Cathode. 77
xxviii
Table 4.1 OLED device structure and initial EQE and voltage at 20
mA/cm2. ITO/1 nm MoOx/HTL/75 nm Alq/1 nm LiF/100
nm Al. 88
1
CHAPTER 1
BACKGROUND AND INTRODUCTION
1.1. ORGANIC ELECTROLUMINESCENCE
Light-emission resulting from electrical stimulation of an organic molecule is
known as organic electroluminescence (EL). In 1953, Bernanose et al. [1] adsorbed
acridine dyes on a sheet of cellophane and observed organic EL by applying a strong
alternating electric field that required a threshold voltage of 400-800 V. In 1963 Pope et
al. [2] observed organic EL from single crystal anthracene at ~400 volts. Additional
organic EL research focused on single crystal anthracene where high voltage continued to
be required [3-9]. Vincett et al. [10] was able to reduce the voltage to below 30 V by
exploring anthracene thin films formed by vacuum deposition. While reduced, the voltage
was still much too high for practical applications. In 1987, Tang and VanSlyke [11]
published their seminal work on organic light-emitting diodes (OLED), using thermal
evaporation to deposit an organic bi-layer sandwiched between two electrodes. This
structure achieved 1% external quantum efficiency (EQE) and decreased the voltage to
below 10 V. In 1990, Burrough et al. [12] introduced the first OLED using polymers that
were deposited by spin-coating instead of vapor deposition.
1.2. ORGANIC LIGHT-EMITTING DIODE OPERATING PRINCIPLE
The OLED device structure and materials utilized by Tang and VanSlyke are
shown in Figure 1.1. The distinguishing feature for this device was utilization of a bi-
2
layer structure to split the charge transport functions between two materials. The bilayer
structure consisted of the di-amine 1,1-bis(di-4-tolylaminophenyl)cyclohexane (TAPC)
and the aluminum complex tris(8-hydroxyquinolinato) aluminium (Alq) sandwiched
between an indium-tin-oxide (ITO) anode and a magnesium:silver cathode. Upon
applying an electric field, electrons are injected from the cathode into the lowest
unoccupied molecular orbital (LUMO) of the electron transport layer (ETL) Alq and
holes are injected from ITO into the highest occupied molecular orbital (HOMO) of the
hole transport layer (HTL) TAPC. The charge carriers recombine at the TAPC/Alq
interface to produce an exciton (bound electron-hole pair) that decays either radiatively to
produce light or non-radiatively to produce heat. In this device, the exciton resides on
Alq, meaning that Alq functions as both the ETL and as the light emitting layer (LEL).
Figure 1.1: OLED configuration and molecular structures used by Tang and VanSlyke.
Since Tang and VanSlyke published their groundbreaking work, there have been intensive
efforts to improve both materials and device structures that have resulted in OLEDs
Glass Substrate
ITO
TAPC
Alq
+
- Mg:Ag
Alq
TAPC
3
achieving commercial success for display and lighting applications [13]. In order to
achieve such improvements, it was necessary to modify the initial device architecture,
adding layers that have very specific functions. Figure 1.2 shows a state-of-the art OLED
that is made up of several layers. In this structure, the cathode is typically reflective so
light emission is observed through the glass substrate.
Figure 1.2: State-of-the-art OLED device structure.
High work-function transparent conducting oxides (TCO) such as indium-tin-
oxide (ITO) are typically used as the anode [14-20]. Low work-function metals such as
Mg, Ca or Ba or bi-layers such as LiF/Al or Li/Al are typically used for the cathode [21-
27]. In order to reduce resistance, various injection and transport layers are used to move
charge carriers from the electrodes to the emitting layer. Hole injection layers (HIL) and
electron injection layers (EIL) often utilize p-type (HIL) or n-type (EIL) doping, [28-31]
respectively, in an effort to better align the electrode work functions with those of the
organic material HOMO/LUMO levels. After injection into the HIL and EIL, charge
Glass Substrate
Anode
Hole Injection Layer
Hole Transport Layer
Emitting Layer
+
-
Electron Transport Layer
Electron Injection Layer
Cathode
4
carriers are then injected into and through the HTL and ETL, after which they recombine
within the LEL. The LEL is typically made up of one or two host materials that are
designed to accept holes and/or electrons from the adjacent transport layers, with the goal
being to ensure that excitons form within the LEL.
1.3. EXCITON FORMATION
If the LEL contains only a single material, then the exciton will form on that
material. Tang et al. [32] introduced the concept of doping the LEL with a fluorescent
material to increase efficiency and to tune the color. Fluorescent dopants are typically
uniformly mixed within the LEL at a concentration of 0.5-2%. Higher concentrations
results in aggregation and self-quenching, resulting in a decrease in efficiency. The
dopant is of lower energy than the host so that excitons initially formed in the host
molecules upon excitation will result in the formation of excitons in the dopant molecules
through an energy transfer process such as Fӧrster energy transfer [33, 34]. Fӧrster
energy transfer is a long-range process that involves a dipole-dipole interaction between a
donor (host) emission transition dipole and an acceptor (dopant) absorption dipole. As a
result, it is important for the donor emission spectrum to overlap with the acceptor
absorption spectrum and also for the donor-acceptor pair to be within 10 nm of each
other. Excitons may also form through a short range (10-20 Å) electron transfer process
known as Dexter energy transfer [33, 34]. This process requires the host and dopant
molecules to have overlapping wave-functions. Excitons may also form directly on the
dopant if electron-hole recombinations occur at the dopant preceded by trapping of either
electron or hole at the dopant molecule.
5
1.4. DEVICE EFFICIENCY
Fluorescent dopants were used in the early stages of OLED development;
however these types of dopants have limitations with respect to EQE. The EQE of an
OLED device is determined by the ratio of the number of photons exiting the device per
the number of electrons injected [35, 36] into the device and is described by equation
(1.1):
(1.1)
Holes and electrons are injected into OLED devices in equal quantities. With the use of
multi-layer OLEDs that can confine charge carriers within the organic stack, it is
generally assumed that every electron injected recombines with its counterpart hole,
resulting in a charge balance factor ( rec) of 100%. The photoluminescence efficiency
(ɳPL) is a property of the emitting material and describes the efficiency of fluorescence
for that material. For the purposes of determining theoretical maximum EQE, this
quantity is generally assumed to be 100%. For OLEDs fabricated on glass substrates and
using reflective cathodes, a large amount of the light is lost within the device and cannot
exit through the glass. This is due predominantly to loss through internal reflections and
wave-guiding, resulting in the optical out-coupling factor (ɳout) to be only ~20% [35].
There is much research devoted to finding ways to improve out-coupling efficiency,
however these techniques will not be discussed here. Spin statistics predicts that
electrically generated excitons produce singlet to triplet excitons in a ratio of 1:3 [37].
Therefore, fluorescent materials (emission from singlet state) have a spin statistic factor
6
(ɳs) of 25%, assuming that only the singlet excitons can decay radiatively. The triplet
excitons are lost through radiationless decay. Taking into consideration an outcoupling
efficiency of 20%, OLEDs based on fluorescent materials are typically assumed to have
an upper limit of 5% EQE. Baldo et al. [38] first utilized a phosphorescent platinum
complex, 2,3,7,8,12,13,17,18-octaethyl-21H,23H-porphine platinum(II) (PtOEP), as the
emitting dopant in an OLED. Porphin complexes without a heavy metal as the central ion
are known to have long-lived triplet states [39] and low phosphorescence efficiencies.
Metalloporphine complexes with a heavy metal ion such as platinum exhibit spin-orbit
coupling effect, which effectively reduces the phosphorescence lifetime and increases the
efficiency of intersystem crossing from the singlet excited state to the triplet state,
resulting in efficient phosphorescent emission. In a highly efficient OLED device, 100%
of the electrically generated excitons are emissive through both singlet and triplet decays,
leading to an EQE upper limit of 20%. Kondakov et al. [40] recently found that triplet
annihilation processes in fluorescent OLEDs produces radiative singlet excited states,
raising the EQE upper limit to 12.5%. Even with this improvement, phosphorescent
materials continue to have higher limits for efficiency and for this reason, there is
continued intensive research aimed at developing phosphorescent materials.
1.5. IRIDIUM CYCLOMETALATED COMPLEXES
While various transition metals such as platinum [38, 41, 42], iridium [43-46],
ruthenium [47-49] and osmium [50-52] have been utilized to form phosphorescent OLED
dopants, iridium cyclometalated dopants have become the most utilized due to their short
phosphorescent lifetimes (microseconds) and high photoluminescence efficiencies.
7
Figure 1.3 shows examples of three different Ir dopants. The emission color can be tuned
based on the choice of ligands. Fac-tris(2-phenylpyridine)iridium(III)Ir(ppy)3 [Ir(ppy)3] is
green emitting with an emission maximum (max) of 515 nm. This material has 2-phenyl-
pyridine as the coordinating ligand. The ligand possesses a donor-acceptor pair, with the
phenyl substituent being the donor and the pyridine substituent being the acceptor. A blue
shift can be produced in 3 different ways, 1) adding electron-donating groups to the
pyridine ring, 2) adding electron withdrawing groups to the phenyl ring, or 3) utilize a
Figure 1.3: Blue [Ir(F2ppy)3], green [Ir(ppy)3] and red [Ir(piq)3] Ir-based
phosphorescent dopants and maximum emission peaks (max).
heteroleptic complex that has a strong electron withdrawing ancillary ligand. As shown in
Figure 1.3, fac-tris-4,6-difluorophenylpyridine Ir(III) [Ir(F2ppy)3] has two electron
withdrawing fluorine substituent’s added to the phenyl ring of Ir(ppy)3, resulting in blue
shift of 45 nm from Ir(ppy)3. A red shift can be affected by extending the conjugation of
the ligand’s chromophore. This is shown using fac-tris(1-phenylisoquinolinato) iridium
(III) [Ir(piq)3] which undergoes a 110 nm red-shift from Ir(ppy)3 by changing the pyridine
to isoquinoline. There are numerous types and combinations of ligands that can be
Ir(piq)3
max = 625 nm
Ir(F2ppy)3
max = 470 nm
Ir(ppy)3
max = 515 nm
8
coordinated to Ir to form highly efficient phosphorescent dopants and this continues to be
an area of intense research [53-58].
Similar to fluorescent dopants, phosphorescent dopants require a host material to
function in an OLED device. In addition to being used for charge transport and exciton
formation, the host materials must have a triplet energy (ET) that is higher than that of the
phosphorescent dopant. If the ET of the host is lower than that of the dopant, then the host
triplets (assuming non-phosphorescent) will act as quenchers for the phosphorescent
dopant triplets, resulting in an overall loss in electroluminescence efficiency.
1.6. DEVICE LIFETIME
There are typically two ways to test device lifetime: 1) Choose a starting
luminance (brightness) and then provide the current density needed to achieve that
brightness. Luminance is the luminous intensity of a surface in a given direction per unit
of projected area, measured in units of candela per square meter, (cd/m2). Current density
is the current per unit area, such as milliamps per square centimeter (mA/cm2). A
commonly used lifetime metric, t50, is the time it takes for the initial luminance to drop by
50% while keeping the current density constant. 2) Choose a specific current density
which will give an initial luminance (or EQE). While keeping the current density fixed,
the luminance will decrease with time and the time taken for the initial luminance to
decrease by 50% is the lifetime (t50).
Universal Display Corporation (UDC) is one of the leading phosphorescent
materials development companies in the world. Without disclosing specific materials or
device architectures, UDC has disclosed t50’s based on a starting luminance of 1,000
9
cd/m2 for red, green and blue phosphorescent OLED’s (PHOLED) of 900,000 hours,
400,000 hours and 20,000 hours, respectively [59]. The blue PHOLED has an order of
Figure 1.4: Commission Internationale De L'Eclairage (CIE) 1931 color coordinates.
magnitude lower t50 which, contrary to red and green PHOLEDs, prevents the blue
PHOLED from being utilized in commercial applications.
Ideally, the red, green and blue (RGB) emitters should be able to produce
saturated red, green and blue colors with a sufficiently wide color gamut for display
applications. For lighting applications, these emitters should be able to produce white
light of various color temperatures. The Commission Internationale De L'Eclairage (CIE)
1931 color space [60] is shown in Figure 1.4 and provides a measure of color saturation
using an x,y coordinate system. The most saturated colors are those at the edge of the CIE
diagram. For the lifetimes mentioned above, the CIEx,y coordinates are 0.64, 0.36 (red),
10
0.31, 0.63 (green) and 0.18, 0.42 (blue). While the red and green coordinates of the
phosphorescent materials are fine for commercial applications, the blue coordinates are
far from where they need to be, with both x and y coordinates needing to be 0.15 or
lower. Considering lifetime and color, the blue PHOLED performance lags far behind
that for red and green.
1.7. DEGRADATION
OLED lifetime is typically affected by two types of degradation, extrinsic and
intrinsic. Extrinsic degradation occurs without the device operating, such as the
electrodes and/or organic materials reacting with water and oxygen present in ambient,
resulting in formation of dark spots and/or luminance quenching species. Extrinsic
degradation has largely been eliminated by using well established encapsulation
techniques. Intrinsic degradation is continuous and results from device operation,
occurring across the device area and inversely dependent on the device current density or
luminance.
Intrinsic degradation is typically attributed to chemical or electrochemical
degradation of specific organic materials in the OLED stack, with the degradation
mechanisms being linked to OLED operational functions such as charge transport and
exciton formation. The identification of the degraded materials has been very difficult
due to the complexity of the OLED layer structure and various interfaces, and the
uncertainty of the material components that are primarily responsible for the device
degradation. As OLED degrades, it is certain that quenchers are formed that can affect
directly the emission efficiency. These quenchers can also affect the transport of charges
11
by acting as traps for either electrons or holes or both, and thus can cause a shift of the
recombination zone as well as the balance of electron and hole densities, which indirectly
can result in a loss of emission efficiency. In most cases, the degraded OLED device will
also require a higher drive voltage to maintain a constant current density due to the
creation of charge traps.
It has been reported that hole and electron transport through Alq produces
unstable Alq radical cations [61] and anions [62], respectively, that form quenching
species in subsequent chemical reactions. It has been reported that electron transport
through 4,4’-bis(3-methylcarbazol-9-yl)-2,2’-biphenyl (BMB) results in luminance loss
and increasing drive voltage [63]. Kondakov et al. [64-67] have shown that charge
recombination at the HTL/LEL interface, where the HTL contains either an aryl-amine or
carbazole moiety, results in the formation of the HTL excited state and subsequent
homolytic C–N bond dissociation. Scholz et al. [68] have shown that the phosphorescent
host 4,4’,4’’-tris(carbazol-9-yl) triphenylamine (TCTA) undergoes dissociation in a
similar manner. Charge recombination at a the LEL/ETL interface with 4,7-diphenyl-
1,10-phenanthroline (BPhen) as the ETL has been shown to result in the formation of
BPhen dimers [68, 69], a process which introduces gap states that affect electron injection
into the LEL. Exciton-polaron interactions can lead to the formation of defects that act as
luminescent quenchers or non-radiative recombination centers [63, 70-74]. These
interactions can also cause molecular aggregation and luminance quenching [75].
Degradation products for iridium cyclometalated complexes have also been
identified. The green emitting dopant Ir(ppy)3 undergoes reversible dissociation of one of
12
the phenyl-pyridine ligands [76], during which time it forms a radical cation that reacts
with the electron-transport material. Dopants such as bis(2-(4,6-difluorophenyl)pyridyl-
N,C2’)iridium(III) picolinate (FIrpic) and bis(2-methyldibenzo-[f,h]quinoxaline)
(acetylacetonate) (Ir(MDQ)2(acac)) also undergo dissociation of their ancillary oxygen
containing ligands, picolinate [77] and acetylacetonate [78, 79], respectively. It has also
been shown that one of the fluorine substituents on the phenyl ring within FIrpic also
dissociates from the molecule [80].
1.8. OBJECTIVES AND OUTLINE OF THESIS
In our power consuming digital world, the need to develop less power intensive
electronic devices is widely recognized. OLED devices are currently incorporated into
commercial products for lighting and displays. As these products compete with
alternative existing technologies such as compact fluorescent bulbs (lighting) and liquid
crystal displays (LCD), there is great motivation to continue on a path of research that
makes these devices more efficient than their counterparts.
Phosphorescent materials are used to make the most efficient OLED devices. Red
and green phosphorescent materials are currently utilized in commercial products. Highly
efficient blue phosphorescent OLEDs have been demonstrated, however short device
lifetime prohibits incorporation of these materials into commercial applications.
Therefore, it is imperative that the degradation mechanisms and processes are understood
so that new blue phosphorescent materials and device structures can be developed to
improve device lifetime to a point where OLED device products are comprised of R,G,B
phosphorescent emitters. The objective of this thesis is to investigate blue phosphorescent
13
OLED instability from both a material and device perspective.
In Chapter 3, a series of OLED devices are fabricated with various host and
transport materials. In all devices, the well-known blue phosphorescent dopant FIrpic is
used as the light-emitter. Depending on the device structure, device lifetimes ranging
from minutes to 12 hours have been achieved. It is found that the formation of the FIrpic
excited state is extremely detrimental to device lifetime. It is also shown that
recombination and emission from the host or electron transport materials is damaging to
device lifetime. Fabrication of stable devices with lifetimes of thousands of hours is not
possible when FIrpic is used as the light emitting material.
In Chapter 4, we investigate how hole transport through FIrpic affects device
performance of model devices that have Alq as the light emitting material. FIrpic is
doped into hole transport materials at various concentrations. Hole transport through
FIrpic forms the FIrpic radical cation, an unstable species that degrades to form hole
trapping products. This results in large increases in the drive voltage. As the voltage rises,
the EQE decreases due to reduced electron-hole recombination occurring within the Alq
emitter. Additionally, the blue phosphorescent host mCBP is found to be stable to hole
transport. However this material contributes to device degradation when participating in
electron-hole recombination.
In Chapter 5, we utilize the blue phosphorescent dopant Ir(iprpmi)3. This dopant
has a very different molecular structure compared to FIrpic. It lacks fluorine substituents
and the picolinate ligand that are present in FIrpic. The photophysical properties of
Ir(iprpmi)3 are reported. A series of devices are fabricated where a range of Ir(iprpmi)3
14
concentrations are studied while using various hole and electron transport materials. It is
found that EQE’s over 20% can be achieved using Ir(iprpmi)3 as the light-emitting
material. A high EQE is also demonstrated for devices using the low triplet energy BAlq
as the ETL. Device lifetime is significantly improved compared to devices using FIrpic as
the light-emitting material.
15
REFERENCES
[1] A. Bernanose, M. Comte, P. Vouaux, J. Chim. Phys. Phys.-Chim. Biol., 50 (1953)
64.
[2] M. Pope, H.P. Kallmann, P. Magnante, J. Chem. Phys., 38 (1963) 2042.
[3] J. Dresner, RCA Rev., 30 (1969) 322.
[4] W. Helfrich, W. Schneider, Phys. Rev. Lett., 14 (1965) 229.
[5] M. Schadt, D.F. Williams, Phys. Status Solidi, 39 (1970) 223.
[6] D.F. Williams, M. Schadt, J. Chem. Phys., 53 (1970) 3480.
[7] D.F. Williams, M. Schadt, Proc. IEEE, 58 (1970) 476.
[8] D.F. Williams, M. Schadt, Proc. Int. Conf. Photocond., 3rd, (1971) 303.
[9] D.F. Williams, M. Schadt, U.S. Patent No. 3,624,632 (1971).
[10] P.S. Vincett, W.A. Barlow, R.A. Hann, G.G. Roberts, Thin Solid Films, 94 (1982)
171.
[11] C.W. Tang, S.A. VanSlyke, Appl. Phys. Lett., 51 (1987) 913.
[12] J.H. Burroughes, D.D.C. Bradley, A.R. Brown, R.N. Marks, K. Mackay, R.H.
Friend, P.L. Burns, A.B. Holmes, Nature (London), 347 (1990) 539.
[13] Mertens, R., www.oled-info.com/devices, 2014
[14] J.-H. Bae, J.-M. Moon, S.W. Jeong, J.-J. Kim, J.-W. Kang, D.-G. Kim, J.-K. Kim,
J.-W. Park, H.-K. Kim, J. Electrochem. Soc., 155 (2007) J1.
[15] J. Cui, A. Wang, N.L. Edleman, J. Ni, P. Lee, N.R. Armstrong, T.J. Marks, Adv.
Mater. (Weinheim, Ger.), 13 (2001) 1476.
[16] H. Kim, C.M. Gilmore, J.S. Horwitz, A. Pique, H. Murata, G.P. Kushto, R. Schlaf,
16
Z.H. Kafafi, D.B. Chrisey, Appl. Phys. Lett., 76 (2000) 259.
[17] H. Kim, J.S. Horwitz, G.P. Kushto, S.B. Qadri, Z.H. Kafafi, D.B. Chrisey, Appl.
Phys. Lett., 78 (2001) 1050.
[18] T.W. Kim, D.C. Choo, Y.S. No, W.K. Choi, E.H. Choi, Appl. Surf. Sci., 253
(2006) 1917.
[19] T.J. Marks, J.G.C. Veinot, J. Cui, H. Yan, A. Wang, N.L. Edleman, J. Ni, Q.
Huang, P. Lee, N.R. Armstrong, Synth. Met., 127 (2002) 29.
[20] K.L. Purvis, G. Lu, J. Schwartz, S.L. Bernasek, J. Am. Chem. Soc., 122 (2000)
1808.
[21] H. Becker, A. Lux, A.B. Holmes, R.H. Friend, Synth. Met., 85 (1997) 1289.
[22] P. Broems, J. Birgersson, N. Johansson, M. Loegdlund, W.R. Salaneck, Synth.
Met., 74 (1995) 179.
[23] B. Chen, J. Hou, S. Xue, S. Liu, Semicond. Photonics Technol., 2 (1996) 293.
[24] P. Dannetun, M. Fahlman, C. Fauquet, K. Kaerijama, Y. Sonoda, R. Lazzaroni,
J.L. Bredas, W.R. Salaneck, Synth. Met., 67 (1994) 133.
[25] L. Duan, K. Xie, Y. Qiu, J. Soc. Inf. Disp., 19 (2011) 453.
[26] M. Stossel, J. Staudigel, F. Steuber, J. Blassing, J. Simmerer, A. Winnacker, H.
Neuner, D. Metzdorf, H.H. Johannes, W. Kowalsky, Synth. Met., 111-112 (2000)
19.
[27] M. Stossel, J. Staudigel, F. Steuber, J. Simmerer, A. Winnacker, Appl. Phys. A, 68
(1999) 387.
[28] R. Meerheim, B. Luessem, K. Leo, Proc. IEEE, 97 (2009) 1606.
17
[29] B. Luessem, M. Riede, K. Leo, Phys. Status Solidi A, 210 (2013) 9.
[30] M. Pfeiffer, K. Leo, X. Zhou, J.S. Huang, M. Hofmann, A. Werner, J. Blochwitz-
Nimoth, Org. Electron., 4 (2003) 89.
[31] K. Walzer, B. Maennig, M. Pfeiffer, K. Leo, Chem. Rev. (Washington, DC, U. S.),
107 (2007) 1233.
[32] C.W. Tang, S.A. VanSlyke, C.H. Chen, J. Appl. Phys., 65 (1989) 3610.
[33] A.A. Shoustikov, Y. You, M.E. Thompson, IEEE J. Sel. Top. Quantum Electron., 4
(1998) 3.
[34] V. Bulovic, M.A. Baldo, S.R. Forrest, Springer Ser. Mater. Sci., 41 (2001) 391.
[35] C. Adachi, M.A. Baldo, M.E. Thompson, S.R. Forrest, J. Appl. Phys., 90 (2001)
5048.
[36] H. Yersin, A.F. Rausch, R. Czerwieniec, T. Hofbeck, T. Fischer, Coord. Chem.
Rev., 255 (2011) 2622.
[37] S.P. McGlynn, T. Azumi, M. Kinoshita, Molecular Spectroscopy of the Triplet
State (Prentice-Hall International Series in Chemistry), 1969.
[38] M.A. Baldo, D.F. O'Brien, Y. You, A. Shoustikov, S. Sibley, M.E. Thompson, S.R.
Forrest, Nature (London), 395 (1998) 151.
[39] M.S. Weaver, Opt. Sci. Eng., 111 (2007) 527.
[40] D.Y. Kondakov, T.D. Pawlik, T.K. Hatwar, J.P. Spindler, J. Appl. Phys., 106
(2009) 124510.
[41] J. Kalinowski, V. Fattori, M. Cocchi, J.A.G. Williams, Coord. Chem. Rev., 255
(2011) 2401.
18
[42] H.-F. Xiang, S.-W. Lai, P.T. Lai, C.-M. Che, Highly Effic. OLEDs Phosphoresc.
Mater., (2008) 259.
[43] C. Adachi, M.A. Baldo, S.R. Forrest, S. Lamansky, M.E. Thompson, R.C.
Kwong, Appl. Phys. Lett., 78 (2001) 1622.
[44] M.A. Baldo, S. Lamansky, P.E. Burrows, M.E. Thompson, S.R. Forrest, Appl.
Phys. Lett., 75 (1999) 4.
[45] S. Lamansky, P.I. Djurovich, D. Murphy, F. Abdel-Razzaq, H.-E. Lee, C. Adachi,
P.E. Burrows, S.R. Forrest, M.E. Thompson, J. Am. Chem. Soc., 123 (2001) 4304.
[46] T. Tsutsui, M.-J. Yang, M. Yahiro, K. Nakamura, T. Watanabe, T. Tsuji, Y. Fukuda,
T. Wakimoto, S. Miyaguchi, Jpn. J. Appl. Phys., 38 (1999) L1502.
[47] W.K. Chan, P.K. Ng, X. Gong, S. Hou, Appl. Phys. Lett., 75 (1999) 3920.
[48] F.G. Gao, A.J. Bard, J. Am. Chem. Soc., 122 (2000) 7426.
[49] J. Yang, K.C. Gordon, Synth. Met., 152 (2005) 213.
[50] B. Carlson, X. Jiang, A.K.Y. Jen, L.R. Dalton, Proc. SPIE-Int. Soc. Opt. Eng.,
4809 (2002) 69.
[51] J. Lu, Y. Tao, Y. Chi, Y. Tung, Synth. Met., 155 (2005) 56.
[52] Y.-L. Tung, P.-C. Wu, C.-S. Liu, Y. Chi, J.-K. Yu, Y.-H. Hu, P.-T. Chou, S.-M.
Peng, G.-H. Lee, Y. Tao, A.J. Carty, C.-F. Shu, F.-I. Wu, Organometallics, 23
(2004) 3745.
[53] Y. Chi, P.-T. Chou, Chem. Soc. Rev., 39 (2010) 638.
[54] P.-T. Chou, Y. Chi, Chem. - Eur. J., 13 (2007) 380.
[55] B. Minaev, G. Baryshnikov, H. Agren, Phys. Chem. Chem. Phys., 16 (2014) 1719.
19
[56] P.I. Djurovich, M.E. Thompson, Highly Effic. OLEDs Phosphoresc. Mater.,
(2008) 131.
[57] L.F. Gildea, J.A.G. Williams, Woodhead Publ. Ser. Electron. Opt. Mater., 36
(2013) 77.
[58] Y. Zhao, D. Ma, Trends Inorg. Chem., 13 (2012) 121.
[59] Universal Display Corporation, www.udcoled.com, 2014
[60] T. Smith, J. Guild, Trans. Optical Soc. (London), 33 (1932) 73.
[61] H. Aziz, Z.D. Popovic, N.-X. Hu, A.-M. Hor, G. Xu, Science (Washington, D. C.),
283 (1999) 1900.
[62] Y. Luo, H. Aziz, G. Xu, Z.D. Popovic, Chem. Mater., 19 (2007) 2079.
[63] N.C. Giebink, B.W. D'Andrade, M.S. Weaver, J.J. Brown, S.R. Forrest, J. Appl.
Phys., 105 (2009) 124514.
[64] D.Y. Kondakov, J. Appl. Phys., 104 (2008) 084520.
[65] D.Y. Kondakov, Proc. SPIE, 7051 (2008) 70510I.
[66] D.Y. Kondakov, T.D. Pawlik, W.F. Nichols, W.C. Lenhart, J. Soc. Inf. Disp., 16
(2008) 37.
[67] D.Y. Kondakov, organic electronics book, (2010) 211.
[68] S. Scholz, K. Walzer, K. Leo, Adv. Funct. Mater., 18 (2008) 2541.
[69] S. Scholz, C. Corten, K. Walzer, D. Kuckling, K. Leo, Org. Electron., 8 (2007)
709.
[70] S. Reineke, K. Walzer, K. Leo, Physical Review B, 75 (2007).
[71] N.C. Giebink, B.W. D'Andrade, M.S. Weaver, P.B. Mackenzie, J.J. Brown, M.E.
20
Thompson, S.R. Forrest, J. Appl. Phys., 103 (2008) 044509.
[72] H. Zamani Siboni, H. Aziz, Appl. Phys. Lett., 101 (2012) 063502.
[73] Q. Wang, H. Aziz, ACS Appl. Mater. Interfaces, 5 (2013) 8733.
[74] Q. Wang, I.W.H. Oswald, M.R. Perez, H. Jia, B.E. Gnade, M.A. Omary, Adv.
Funct. Mater., 23 (2013) 5420.
[75] Q. Wang, B. Sun, H. Aziz, Adv. Funct. Mater., 24 (2014) 2975.
[76] I. Rabelo de Moraes, S. Scholz, B. Luessem, K. Leo, Appl. Phys. Lett., 99 (2011)
053302.
[77] I. Rabelo de Moraes, S. Scholz, B. Luessem, K. Leo, Org. Electron., 12 (2011)
341.
[78] R. Meerheim, S. Scholz, S. Olthof, G. Schwartz, S. Reineke, K. Walzer, K. Leo, J.
Appl. Phys., 104 (2008) 014510.
[79] S. Scholz, R. Meerheim, B. Luessem, K. Leo, Appl. Phys. Lett., 94 (2009)
043314.
[80] V. Sivasubramaniam, F. Brodkorb, S. Hanning, H.P. Loebl, V. van Elsbergen, H.
Boerner, U. Scherf, M. Kreyenschmidt, J. Fluorine Chem., 130 (2009) 640.
21
CHAPTER 2
DEVICE FABRICATION EXPERIMENTAL DETAILS
2.1 THERMAL EVAPORATION COATER
All OLED devices described in this thesis were fabricated using materials
(organic, inorganic, and metal) that were deposited (coated) as thin films using the well
established thermal evaporation technique [1-5]. Figure 2.1 shows the basic design of a
vacuum coater for the deposition of thin films by thermal evaporation. The required
vacuum chamber pressure is typically 10-6
torr or lower.
A thermal evaporation source, often known as a boat, is used to vaporize solid
materials for thin film deposition. A typical boat consists of a resistive heating element,
which is typically made of refractive metals such as tungsten, tantalum or molybdenum in
the form of a sheet or coil. Often the metal sheet is shaped in the form of a boat to hold
the material (known as the source material) to be evaporated. A quartz or ceramic
crucible is also often used to hold the source material with the metal sheet or coil acting
the heating element. Low-voltage electrical power is generally the heating source. Upon
heating in the vacuum chamber to a sufficiently high temperature, the source solid
material will vaporize and transform into isolated molecules, which are ejected from the
boat into the vacuum space of the deposition chamber. Condensation of these molecules
on a substrate placed in a line-of-sight position from the boat source will form film. For
22
the fabrication of OLED devices, the substrate is usually kept at or near room
temperature to promote the formation of uniform and generally amorphous thin films.
The deposition of thin films by thermal evaporation is usually controlled by a
commercially available thin-film deposition controller (DC), which is capable of
measuring the rate of thin film deposition (in Angstrom per sec) using a quartz crystal
microbalance (QCM). The QCM is placed in a line-of-sight position with respect to the
boat source. A feedback loop to the boat power supply is used to control the temperature
of the source material in the boat and thus its rate of evaporation. The DC can monitor
both the film thickness accumulated on the substrate and control the rate of accumulation
by adjusting the electric power supplied to the boat. Independently, the thickness of the
film deposited on the substrate can be measured using a profilometer or ellipsometer.
Figure 2.1: Basic design of a thermal evaporation coater.
Vacuum Pump
Deposition
Controller
Power Supply
Chamber
Substrate
Shutter
Quartz crystal microbalance
Resistively
heated boat
23
A shutter is used to prevent any material from coating the substrate until a well-
controlled deposition rate is established. This enables accurate control of the thickness of
the thin film deposited on the substrate.
2.2 THE KOMET COATER
A custom-designed vacuum coater was designed and built by K. Klubek for the
fabrication of all OLED devices described in this thesis. This coater is known as the
KOMET coater, an acronym based on the names of those who contributed to this system
being built: Klubek (Kevin), Madathil (Joseph), and Tang (Ching). The last two letters of
the acronym are derived from our field of study, Organic Electronics. This coater was
designed to include several important features: 1) compact thermal source design, 2) low
power consumption boats for depositing organic materials, and 3) co-evaporation of
multiple materials.
Figure 2.2 shows a picture of the KOMET coater. This table-top system has a
14”x14” square base and a top plate that also measures 14”x14”. The chamber, placed
between these two plates, consists of a Pyrex collar (12.5” outside diameter) that sits on
top of a metal collar (13-7/8” outside diameter). The bottom plate is attached to 4 posts
(8-7/8” long) that sit on the bench-top. The posts raise the chamber and allow easy access
to various feedthroughs in the bottom plate. Between the top and bottom plates is the
working space of the coater. Mounted on these plates are all the key feedthroughs for
various electrical and mechanical components (electrical power, QCM signals,
turbopump, shutter controls, etc).
24
Figure 2.2: KOMET coater.
The inside of the KOMET coater chamber is shown in Figure 2.3. Figure 2.3(a)
shows the top of the chamber which includes the substrate and mask platters, and the
shutter plate. The substrate platter can accommodate 6 square substrates (1.5”x1.5”). The
mask platter holds two different masks that are used for fabricating OLED devices, one
for depositing organic/inorganic materials and another for depositing metals. The shutter
plate is placed below the mask and substrate platters.
Figure 2.3(b) shows the bottom of the coater which is partitioned into two
separate sides, one for depositing organic materials and another for depositing
metal/inorganic materials. Partitioning the two sides is to prevent cross-contamination
from occurring. Figure 2.3(c) shows the side with the boat assembly for organic material
deposition. The key feature of this coater is the boat assembly (4” diameter) consisting of
Top plate
Base plate
Chamber
12.5”
6-1/16”
14”
14”
8-7/8”
25
Figure 2.3: Inside the KOMET coater chamber. (a) Top of the chamber, (b) bottom of the chamber,
(c) area for depositing organic materials, and (d) area for depositing metals/inorganics.
12 custom-built, low-temperature, filament evaporation boats (organic boats) for
thermally depositing organic materials. The organic boats are based on a glass tube
design originally developed by S. Lee [6] and will be discussed in the following section,
section 2.3. The electrical sockets for the boats are mounted to a rotatable platter. There
are 4 water-cooled QCM’s (MCVAC Manufacturing Company part #’s 700-001, 100-011,
900-004, 900-002, 900-001) fixed in positions with respect to the rotatable platter. This
arrangement allows the deposition of any one material or the co-deposition of two to four
materials, all out of the 12 boat sources for any given thin-film deposition. The QCM’s
12-16-11-DOE site visit
Substrate platterMask platter
Shutter plate
4 Crystal sensors
Organic boats
Common electrode
Individual electrodes
Glass divider
(a) (b)
(c) (d)
Organic deposition
Metal/Inorganicdeposition
Organic source assembly
Metal boat
26
are water cooled to minimize effects of radiant heat from the boats. The rotatable platter
design allows the selection of materials held in designated boats for co-evaporation.
Figure 2.3(d) shows the metal/inorganic side of the coater. The boats on this side
are made from tungsten (R.D. Mathis Company part # ME4-.005W) for the deposition of
aluminum (Al), and from tantalum (R.D. Mathis company part # SB3-AO-TA) lithium
fluoride (LiF) and molybdenum oxide (MoOx) deposition. The current required to
evaporate these materials are: 120-150 amps for Al (at a rate of 10 Å/s), 90-100 amps for
LiF and 55-65 amps for MoOx (at rates of 0.5 Å/s for each). There is a water cooled
QCM positioned above these boats.
Figure 2.4: KOMET coater top plate.
Figure 2.4 shows the top plate. This plate has several feedthroughs, including the
Top plate
Substrate port
Metal shutter
Organic shutter Substrate/
maskrotor
27
substrate/mask rotor which is used to rotate the substrates into position for depositing
either organic or inorganic/metal materials. This plate also has the control for the organic
and metal shutters. Substrates are loaded into the substrate platter through the substrate
port.
Figure 2.5 shows a turbopump (Pfeiffer TPU 170) connected to the bottom plate.
This pump allows the KOMET coater to achieve a base pressure of 2 × 10-7
Torr.
Figure 2.5: Features below the KOMET bottom plate.
Also attached to the bottom plate is the rotor for the organic material evaporation source
assembly and the apparatus for indexing the organic boats. The indexed position relates
the boat positions to the QCM’s and determines which boats can be used to evaporate
materials.
Figure 2.6 shows the control cabinet for the KOMET coater. This cabinet contains
Organic boat rotor
Apparatus for indexing organic boats
Turbopump
Bottom plate
28
Figure 2.6: Control cabinet.
a) the turbopump controller (Pfeiffer TCP 310), b) two deposition controllers, one for
organics (Kurt J. Lesker-FTC-2800) and one for metals/inorganics (Sigma Instruments
SQC-122c), c) organic switchboard connecting the power supplies to 4 out of the 12
organic boats, d) metal/inorganic switchboard for connecting a power supply to one of
the 4 available boats, and e) the power supplies. The deposition controller for the
organics is capable of monitoring the deposition of 1-4 materials at a time. There are 4
power supplies (Sorensen DCS8-125E) used for depositing the organic materials and
there is a single power supply (Sorensen DCS12-250) used for depositing inorganic/metal
materials. These particular power supplies were used out of convenience (they were
donated by Eastman Kodak Company) and not out of necessity as the power
Power supplies
Metal/inorganic
switchboard
Organic switchboard
Deposition controllers
Turbo-pump
controller
29
requirements for the boats are all far lower than what these power supplies can provide.
2.3 ORGANIC BOATS
The components and design of the organic boat are shown in Figure 2.7. The steps
for assembling the boat are shown in Figure 2.8. A disposable glass test tube (VWR
catalog # 47729-570) is cut at both ends using a diamond wheel (SE DW13,
B000P49NCC) to form a cylinder and a side hole is made using a diamond drill bit
(McMaster-Carr part # 4376A12, 0.039” diameter). A nickel:chromium (nichrome) (22
BNC, 0.0253” nominal diameter) coil is used as the resistive heating element. This coil
was made using a 0.325” hollowed-out aluminum spindle attached to a lathe. There are
16 coils, with a space made between coils 6 and 7 (counting from top) to prevent the coil
from blocking the side hole. A ceramic rod made from Macor is machined to make the
base of the boat. The nichrome coil is inserted into the base and the glass cylinder then
surrounds the coil. Contact pins (Summit part# 030 1952 000) are attached to the
nichrome wire with a press. Aluminum foil is wrapped around the outside of the glass
tube, leaving an opening for the side hole. This foil is important because it helps to keep
heat within the tube. Without the foil, material condenses between the coils, resulting in
inconsistent film thicknesses. A high temperature epoxy (Cotronics Corporation, Resbond
907) is used to bind the contact pins to the macor base. This epoxy is also used to bind
the glass cylinder to the Macor base. A heat gun is used during application of the epoxy to
assist with drying. The entire assembly is then heated at 300º C in an oven (in air) to fully
cure the epoxy. There is a ~0.3” gap between the macor base and the bottom of the
nichrome coil. This space holds 100-300 mg of organic material.
30
Figure 2.7: Picture of components used to build organic boat.
Figure 2.8: Components for assembling organic boat.
Test tube
Cylinder
Nichrome coil
Macor
base
Electrical
contact
pins
Assembled boats
Epoxy
Al
shield
0.157”
0.118”
0.407”0.512”
Macor base
Nichrome coilGlass cylinder with
0.039” side hole
Electrical
contact pin
0.41”
1.71”
Space for
material
0.47”
Al shield
0.3”
31
As shown in Figure 2.3(c), the boats with the NiCr wires are fitted into the
ceramic electrical sockets with the side hole directly facing the QCM. As material is
heated within the boat, the vapor exits from the top opening and their relative deposition
rates are monitored by the top and side QCMs. The linear correlation between these rates
is used to establish the side QCM as the rate controller. The calibration of the side QCM
is done by actually measuring the thickness of the deposited film on the substrate position
using a KLA Tencor D100 profilometer. A tooling factor is established for the side QCM
so that its thickness reading can be equated to the measured thickness.
The organic boats require less than 10 W to deposit commonly used OLED
materials. The voltage and current requirements are ~2–4 V and ~1–2 A, respectively.
2.4 DEVICE FABRICATION
All devices were prepared on glass substrates (Tinwell Electronic Technology
Company) pre-patterned with indium-tin-oxide (ITO) having a thickness of 110 nm and a
sheet resistance of about 15 ohm/square. Figure 2.9 shows the ITO patterns. There are 4
test patterns on each substrate, all with an identical structure with etched ITO as
electrodes. The substrates were batch-wise cleaned by, sequentially, soaking and
mechanically scrubbing in detergent solution, washing in deionized water, acetone and
isopropanol using an ultrasonic bath, drying with nitrogen and then further treating with
O2 plasma. All films were prepared by thermal deposition (<5x10-6
Torr) without
breaking vacuum until the OLED device was completed with the top electrode. The
deposition rate for host and transport layers was 4 Å/s. Dopant rates were a percentage of
the host rate and therefore were varied based on the choice of concentration.
32
Figure 2.9: ITO pattern on substrate.
Molybdenum trioxide (MoOx) of 1.2 nm was used as a hole-injection layer (HIL)
[7-11]. It was deposited on top of the ITO using an “organic” mask with a circular
opening that permits deposition over the 4 test patterns. The term “Anode” within this
thesis refers to the sequential thin film sequence consisting of the following bilayer
structure: Glass substrate/110 nm ITO/1.2 nm MoOx. The term “Cathode” within this
thesis refers to the sequential thin film sequence consisting of the following bilayer
structure: 1.0 nm LiF/100 nm Al [12-14]. A typical representation for devices that will be
discussed within this thesis is the following: Anode/organic layers/Cathode, where the
organic layers consist of a sequence of hole-transport, light-emitting, and electron-
33
transport layers.
Figure 2.10(a) shows how the device active area is defined. On top of the ITO
pattern, through the organic mask (0.65” diameter) is deposited in order: MoOx, the
organic layers and then LiF. For the cathode, aluminum is then deposited through an L-
shape “metal” mask. The area where the ITO pattern and Al metal overlap (with organic
material in between) is the active area, which is 0.1 cm2 in our design. Figure 2.10(b)
shows a completed device with all 4 pixels being lit. After fabrication, the OLED devices
incurred a brief exposure to the ambient atmosphere before transferring to a vacuum
assembly. The vacuum assembly shown in Figure 2.10(b) and 2.10(c) consists of a valve
attached to a flange containing an o-ring. The device is placed on the o-ring, a mechanical
pump is attached to the end of the valve, and then the assembly was evacuated to below
50 mTorr using a mechanical pump. This valve assembly provides a a simple
encapsulated environment for the OLED device and is suitable for short-term lifetime
tests.
Figure 2.10: Pictures of OLED active area and encapsulation technique (a) ITO pattern, organic
deposited through circular mask, and metal deposited through L-shaped mask, (b) An OLED device
with 4 lit pixels, and (c) Vacuum assembly for encapsulating devices.
(a) (b) (c)
ValveConnection to vacuum pump
SubstrateO-ringFlange
Flange with O-ring0.65” diameter
34
2.5 OLED PERFORMANCE MEASUREMENTS
The luminance-current-voltage (LIV) characteristics were measured using a
Keithley 2400 source meter and a Photoresearch PR650 SpectraScan Colorimeter. The
source meter and PR650 were computer controlled using a software program developed
by John Burtis from Eastman Kodak Company. To obtain an LIV trace, the source meter
supplies a constant current to the OLED device under test and the corresponding voltage
from the source meter and the luminance value from the PR650 were logged. These
measurements were repeated for a range of currents from 0.1 to 100 mA/cm2. The output
data from the software program is shown in Figure 2.11. They include both radiometric
and photometric data of the electroluminescence as a function of current density and
voltage. The spectral data, as well as the luminous efficiency data in terms of cd/A
(candela per ampere) were also plotted.
The devices were viewed perpendicular to the device plane, and the angular
distribution of the EL was assumed to be Lambertian.
2.6 DEVICE LIFETIME
Device lifetimes for phosphorescent OLEDs in this thesis were measured at a
current density of 5 mA/cm2. EL data cited in the literature are often reported at 5
mA/cm2. This makes it simple to compare data collected within this thesis to those
reported in the literature. Furthermore, this is a practical current density that typically
provides a brightness (500-2000 cd/m2) that is comparable to what would be required for
commercial device applications. For a typical OLED device lifetime measurement, the
initial brightness (cd/m2) was first measured using PR650 at 5 mA/cm
2. The device was
35
Figure 2.11: Output from John Burtis software program (a) current density (mA/cm2), (b) drive voltage (V), (c) luminance (cd/m2), (d) CIEx
coordinate, (e) CIEy coordinate, and (f) EQE or p/e (photon/electron).
(a) (b) (c) (d) (e) (f)
36
then moved to a lifetime test station where it was driven at 5 mA/cm2 and the brightness
was continuously monitored by a silicon photodiode (Hamamatsu S1787-08) placed in
front of the OLED device. The lifetime of the OLED is the time it takes for the brightness
to decrease by 50% (t50). The luminance output was again recorded by the PR650 at the
end of the lifetime.
37
2.7 MATERIALS
The acronym, chemical name, molecular structure, and relevant energy levels for
all organic materials described in this thesis are shown in Table 2.1. The following
materials were purchased from Nichem Fine Technology Company and used as received:
FIrpic (Ni-D306), TCTA (Ni-HB03), and TmPyPB (Ni-E206). The following materials
were purchased from Luminescence Technology Corporation and used as received:
Ir(iprpmi)3 (LT-N643), and UGH3 (LT-N449). The following materials were obtained
from Eastman Kodak Company and used as received: Alq, BAlq, BPhen, NPB, TAPC.
mCBP (ALD-F007) was purchased from Jilin OLED Material Technology Company and
used as received. Aluminum (stock# 43357), LiF (stock# 14056) and molybdenum(VI)
oxide (stock# 11837) were all purchased from Alfa-Aesar and used as is.
Molybdenum(VI) oxide is referred to as MoOx throughout this thesis because the oxygen
oxidation state is variable [15] depending on the deposition conditions.
With the exception of Ir(iprpmi)3, the values of the highest occupied molecular
orbital (HOMO), the lowest unoccupied molecular orbital (LUMO), and the triplet energy
are from the literature. Unless noted otherwise, HOMO energies were determined using
ultraviolet photoelectron spectroscopy and LUMO energies were based on the difference
of the HOMO level and the optical band gap as determined by the onset of absorption.
Triplet energies are estimated experimentally from phosphorescence spectra at 77 K.
38
Table 2.1: Organic materials discussed throughout this thesis. HOMO/LUMO/triplet energies were taken from literature. HOMO energies are based
on ionization potential, LUMO energies based on difference between optical bandgap and HOMO energy, and triplet energy based on
phosphorescence spectra taken at 77K.
aData determined using solution based cyclic voltammetry-HOMO energy from oxidation potential and LUMO energy from reduction potential [16]. bData based on work in this thesis, discussed further in Chapter 5.
Acronym Name Structure HOMO
(eV)
LUMO
(eV)
Triplet
(ev)
Alq tris(8-hydroxyquinolinato) aluminum
5.7[17]a 2.5[17]
a 2.2[17]
BAlq bis-(2-methyl-8-quinolinolate)-4-
(phenylphenolate)aluminum (III)
5.9[18] 2.9[18] 2.2[18]
BMB 4,4'-bis(3-methylcarbazol-9-yl)-2,2'-
biphenyl
BPhen 4,7-diphenyl-1,10-phenanthroline
6.4[19] 3.0[19] 2.6[20]
39
Acronym Name Structure HOMO
(eV)
LUMO
(eV)
Triplet
(ev)
FIrpic bis(4,6-difluorophenyl-pyridinato-
N,C2) picolinate iridium (III)
5.6[21]a
5.9[22]
2.5[21]a
3.0[22] 2.6[22]
Ir(F2ppy)3 fac-tris-4,6-difluorophenylpyridine
Ir(III)
5.7[21] 3.0[21] 2.71[17]
Ir(iprpmi)3
fac-tris[(2,6-diisopropylphenyl)-2-
phenyl-1Himidazol[e]iridium(III)
4.8b 2.2
b 2.66
b
Ir(MDQ)2(acac) bis(2-methyldibenzo-[f,h]quinoxaline)
(acetylacetonate)
5.4[23] 2.8[23] 2.0[24]
Ir(piq)3 fac-tris(1-phenylisoquinolinato)
iridium(III)
5.1[25] 3.1[25] 2.0[26]
40
Acronym Name Structure HOMO
(eV)
LUMO
(eV)
Triplet
(ev)
Ir(ppy)3 fac-tris(2-phenylpyridine)iridium(III)
5.6[27] 3.0[27] 2.49[17]
Ir(tdmp)3
fac-tris[3-(2,6-dimethylphenyl)-7-
methylimidazo[1,2-f] phenanthridine]
iridium (III)
mCBP 3,3'-bis(N-carbazolyl)biphenyl
5.6[28]a
6.0[29]
2.1[28]a
2.4[29] 2.8[28]
NPB 4,4'-bis[N-(1naphthyl)-N-phenyl-
amino]-biphenyl
5.3[17]a
5.4[18]
2.1[17]a
2.3[18] 2.3[18]
PtOEP 2,3,7,8,12,13,17,18-octaethyl-
21H,23H-porphine platinum(II)
41
Acronym Name Structure HOMO
(eV)
LUMO
(eV)
Triplet
(ev)
TAPC 1,1-bis(di-4-tolylaminophenyl)
cyclohexane
5.5[30] 2.0[30] 2.9[30]
TCTA 4,4',4''-Tris(N-
carbazolyl)triphenylamine
5.7[18] 2.3[18] 2.9[18]
TmPyPB 1,3,5-Tris(3-pyridyl-3-phenyl)benzene
6.7[22] 2.8[22] 2.8[22]
UGH3 m-bis(triphenylsilyl)benzene
7.2[31] 2.8[31] 3.5[31]
42
REFERENCES
[1] H.L. Caswell, Semicond. Prod. Solid State Technol., 6 (1963) 23.
[2] M.T. Thomas, Methods Exp. Phys., 14 (1979) 521.
[3] D.R. Biswas, J. Mater. Sci., 21 (1986) 2217.
[4] S. Bosch, Trends Vac. Sci. Technol., 2 (1997) 131.
[5] S.M. Rossnagel, J. Vac. Sci. Technol., A, 21 (2003) S74.
[6] S.M. Lee, PhD Thesis, Department of Chemistry, University of Rochester, 2012
[7] T. Matsushima, Y. Kinoshita, H. Murata, Appl. Phys. Lett., 91 (2007) 253504.
[8] H. You, Y. Dai, Z. Zhang, D. Ma, J. Appl. Phys., 101 (2007) 026105.
[9] T. Matsushima, G.-H. Jin, H. Murata, J. Appl. Phys., 104 (2008) 054501.
[10] G. Xie, Y. Meng, F. Wu, C. Tao, D. Zhang, M. Liu, Q. Xue, W. Chen, Y. Zhao,
Appl. Phys. Lett., 92 (2008) 093305/1.
[11] K.S. Yook, J.Y. Lee, Synth. Met., 159 (2009) 69.
[12] H. Heil, J. Steiger, S. Karg, M. Gastel, H. Ortner, H. von Seggern, M. Stossel, J.
Appl. Phys., 89 (2001) 420.
[13] G. Parthasarathy, C. Shen, A. Kahn, S.R. Forrest, J. Appl. Phys., 89 (2001) 4986.
[14] D.Y. Kondakov, J. Appl. Phys., 99 (2006) 024901.
[15] Irfan, H. Ding, Y. Gao, C. Small, D.Y. Kim, J. Subbiah, F. So, Appl. Phys. Lett.,
96 (2010) 243307.
[16] B.W. D'Andrade, S. Datta, S.R. Forrest, P. Djurovich, E. Polikarpov, M.E.
Thompson, Org. Electron., 6 (2005) 11.
[17] M.E. Kondakova, T.D. Pawlik, R.H. Young, D.J. Giesen, D.Y. Kondakov, C.T.
43
Brown, J.C. Deaton, J.R. Lenhard, K.P. Klubek, J. Appl. Phys., 104 (2008)
094501.
[18] D.P.-K. Tsang, M.-Y. Chan, A.Y.-Y. Tam, V.W.-W. Yam, Org. Electron., 12 (2011)
1114.
[19] Y. Sun, S.R. Forrest, Appl. Phys. Lett., 91 (2008) 263503.
[20] J. Zhao, J. Yu, X. Hu, M. Hou, Y. Jiang, Thin Solid Films, 520 (2012) 4003.
[21] J.J. Brooks, R.C. Kwong, Y.-J. Tung, M.S. Weaver, B.W. D'Andrade, V.
Adamovich, M.E. Thompson, S.R. Forrest, J.J. Brown, Proc. SPIE, 5519 (2004)
35.
[22] J. Lee, J.-I. Lee, J.Y. Lee, H.Y. Chu, Appl. Phys. Lett., 94 (2009) 193305.
[23] R. Meerheim, S. Scholz, S. Olthof, G. Schwartz, S. Reineke, K. Walzer, K. Leo, J.
Appl. Phys., 104 (2008) 014510.
[24] G. Schwartz, S. Reineke, T.C. Rosenow, K. Walzer, K. Leo, Adv. Funct. Mater.,
19 (2009) 1319.
[25] T. Zhang, Y. Liang, J. Cheng, J. Li, J. Mater. Chem. C, 1 (2013) 757.
[26] C. Cai, S.-J. Su, T. Chiba, H. Sasabe, Y.-J. Pu, K. Nakayama, J. Kido, Org.
Electron., 12 (2011) 843.
[27] J.-H. Jou, W.-B. Wang, M.-F. Hsu, J.-J. Shyue, C.-H. Chiu, I.M. Lai, S.-Z. Chen,
P.-H. Wu, C.-C. Chen, C.-P. Liu, S.-M. Shen, ACS Nano, 4 (2010) 4054.
[28] S. Gong, X. He, Y. Chen, Z. Jiang, C. Zhong, D. Ma, J. Qin, C. Yang, J. Mater.
Chem., 22 (2012) 2894.
[29] H. Nakanotani, K. Masui, J. Nishide, T. Shibata, C. Adachi, Sci Rep, 3 (2013)
44
2127.
[30] J. Lee, J.-I. Lee, J.Y. Lee, H.Y. Chu, Appl. Phys. Lett., 95 (2009) 253304.
[31] X. Ren, J. Li, R.J. Holmes, P.I. Djurovich, S.R. Forrest, M.E. Thompson, Chem.
Mater., 16 (2004) 4743.
45
CHAPTER 3
FIRPIC AS THE LIGHT-EMITTING DOPANT FOR
PHOSPHORESCENT ORGANIC LIGHT-EMITTING
DIODES
3.1 INTRODUCTION
Organic light emitting diodes (OLED) have recently achieved commercial success
as a display and lighting technology [1]. The impact of using phosphorescent materials
has been discussed in Chapter 1. While recently introduced commercial products employ
phosphorescent red and green emitting materials, the poor performance of blue
phosphorescent emitters, notably the lifetime, has prevented adaptation into commercial
products. Universal Display Corporation (UDC), a leading PHOLED materials
development company, has reported light blue PHOLEDs that have 1931 Commission
Internationale de l’Eclairage (CIE) color coordinates of 0.18, 0.42, luminous efficiency of
50 cd/A and operating lifetime (t50; time for initial luminance to drop by 50%) of 20,000
hours, all at an initial luminance of 1,000 cd/m2 [2]. This lifetime is far inferior to those
of red and green devices reported by UDC, which have lifetimes of 900,000 and 400,000
hours, respectively. The specific materials and device architectures used by UDC have
not been disclosed.
Giebink et al. [3] achieved ~9% EQE and t50 of 700 hours at a current density of
46
~7 mA/cm2
using the following device architecture, Indium tin oxide (ITO)/10 nm
undisclosed material/30 nm NPB/30 nm BMB + 9% Ir(tdmp)3/5 nm BMB/40 nm
Alq/LiF/Al. In this architecture, ITO is the anode, the undisclosed material is the hole
injecting layer (HIL), NPB is the hole transport layer (HTL), BMB is the light-emitting
layer (LEL) host with Ir(tdmp)3 as the dopant, undoped BMB is the electron transport
layer (ETL), Alq is used as both the electron injection layer (EIL) and ETL, and LiF/Al is
the cathode. The focus of this research revolved around exciton-polaron annihilation
reactions and the effect on the device voltage and lifetime. Yamamoto et al. [4] reported
similar device architecture and performance as Giebink et al., however they focused on
how lifetime is improved by fabricating devices under ultra high vacuum (6.5 × 10-7
torr).
In a U.S. patent, D’Andrade et al. [5] disclosed a similar device structure using the same
materials as Giebink et al. and achieved an EQE of ~7% and a t50 of 10,000 hours at ~4
mA/cm2. While device performance was briefly discussed in each of these references,
none related device lifetime and/or EQE to specific materials or the device’s layer
structure.
FIrpic is a phosphorescent blue emitter that has been widely used in blue
PHOLED devices. Recently, the stability of FIrpic has been examined using laser-
desorption/ionization time-of-flight mass spectrometry (LDI-TOF-MS) [6]. The results
indicated that FIrpic undergoes chemical dissociation, including loss of the picolinate
ligand and loss of CO2. Fragments of FIrpic were shown to react with other materials,
which could contribute to device instability. In another study with LDI-MS, the F atoms
on FIrpic were found to dissociate during device operation [7]. Depending on device
47
architecture and the choice of host and transport materials, lifetimes ranging from
minutes to 110 hours have been reported with FIrpic as the phosphorescent blue emitter
[6, 8-10]. While there have been few reports on PHOLED device lifetime, there have
been numerous reports on FIrpic achieving an external quantum efficiency (EQE) greater
than 20% using a variety of host and transport materials [11-16]. Relatively few
publications report device lifetime for blue PHOLEDs, especially for devices optimized
specifically for achieving the highest EQE.
In this chapter, device performance is evaluated, including device lifetime, with
FIrpic as the phosphorescent dopant. The goal is to correlate the device performance with
variations in device architecture and material composition. Three different hosts are
compared, TCTA:UGH3 mixed host, TCTA single host, and mCBP single host. For each
of these hosts, a FIrpic concentration series is examined. The affect of varying the ETL
material is also explored. Two different HTL materials for devices with mCBP as the host
are compared.
3.2 EXPERIMENTAL
Device fabrication was discussed in Chapter 2. Molecular structures for all
materials are listed in Figure 3.1.
3.3 MIXED HOST
Mixed host technology has been used to improve charge balance, device
efficiency and lifetime for both fluorescent [17-20] and phosphorescent [21-27] OLEDs.
Instead of using a single host – a host comprising a pure material, a mixed host typically
48
comprises two material components, usually mixed uniformly in a solid film, one for
Figure 3.1: Molecular structures of materials used in Chapter 3.
transporting electrons and the other for holes. In this way, the electron and hole transport
functions are supported separately in the two host components. Essentially, the mixed
host spatially extends the recombination region by allowing holes to inject into the LEL
through the hole-transport host while electrons are injected into the LEL through the
electron-transport host.
Mixed-host LEL’s have been shown to improve the device lifetime by extending
the recombination zone within the LEL to reduce both charge build-up and recombination
HTL Materials
Host Materials Blue Dopant
ETL Materials
NPB TAPC
TCTA UGH3 mCBP FIrpic
BPhen TmPyPB UGH3 BAlq
49
cycles per molecule at the LEL/transport layer interface. Recently, an EQE of 21.6% was
achieved using a mixed-host of TCTA and UGH3 with FIrpic as the blue dopant [14].
However, similar to other publications, this article only highlighted the extraordinarily
high EQE achievable in a blue PHOLED without any reference to the device lifetime.
Previously, Lee et al. [27] of our laboratory studied the same mixed-host using a graded
structure, however the lifetime was not studied. In this section, device performance is
evaluated using TCTA:UGH3 as a mixed host. TCTA has a strong electron donating
character with the triphenylamine core substituted with carbazole substituents. This
material is the hole transporting component of the mixed host. UGH3 has
tetraphenylsilane as the core moiety. It has been described as the electron transporting
component for this mixed host system. Both materials have triplet energies higher than
that of FIrpic.
3.3.1 TCTA:UGH3 (1:1) mixed host with varied FIrpic concentration
Figure 3.2: FIrpic concentration series for TCTA:UGH3 mixed host OLED devices.
A series of PHOLED devices was fabricated with a 1:1 ratio of TCTA:UGH3 as
35 nm HTL
TAPC
30 nm ETL
TmPyPB
20 nm EIL
BPhen
40 nm Mixed host
TCTA:UGH3 (1:1)
Emitting dopant
FIrpic
Concentration
9-26%
Anode Cathode
50
the mixed host where the concentration of FIrpic was varied from 9%–26%. The device
layer structure is as follows: Anode/35 nm TAPC/40 nm TCTA+UGH3 (1:1)+x%
FIrpic/30 nm TmPyPB/20 nm BPhen/Cathode (Figure 3.2). TAPC and TmPyPB have
triplet energies of 2.90 eV and 2.78 eV, respectively, and both are higher than that of
FIrpic (2.62 eV) to prevent exciton quenching. BPhen is known to be a good EIL. The
emission spectra, EQE, and current-voltage characteristics are shown in Figure 3.3 and
the device data at 5 mA/cm2 are shown in Table 3.1.
The electroluminescence (EL) spectra for devices having the lowest (9%) and
highest (26%) FIrpic concentrations as shown in Figure 3.3(a) indicate that the EL
emission originates from FIrpic. Essentially the EL spectra are independent of FIrpic
concentration. The slightly broader emission from the 26% FIrpic device can be
attributed to the solid-state solvation effect [28, 29]. Figure 3.3(b) shows that for all the
devices relatively high EQE’s of ~15-20% are obtained in the low current density range
of 0.1-5 mA/cm2, with the EQE then decreasing steadily with increasing current density.
At 20 mA/cm2, an EQE above 15% is maintained for all the devices except the device
with 26% FIrpic. Even this device, which presumably has a lower EQE due to
concentration quenching [30, 31] or increased triplet-triplet annihilation [32], attains a
high EQE in the range of 13-15% throughout the entire current range. The J-V
characteristics shown in Figure 3.3(c) indicate negligible differences for FIrpic
concentrations between 9%–17%. The shift to slightly higher voltage for the device with
26% FIrpic is likely due FIrpic interfering with the hole transport through the TCTA host.
51
Table 3.1: OLED performance data at 5 mA/cm2 for varying FIrpic concentration in mixed host
devices; Anode/35 nm TAPC/40 nm TCTA+UGH3 (1:1)+x% FIrpic/30 nm TmPyPB/20 nm
BPhen/Cathode, x=9–26%.
FIrpic (%)
V cd/m2 EQE
(%)
9 6.4 1824 17.2
12 6.4 1987 18.3
17 6.5 1879 17.0
26 6.8 1598 13.8
Figure 3.3: OLED performance data at 5 mA/cm2 for varying FIrpic concentration in mixed host
devices; Anode/35 nm TAPC/40 nm TCTA+UGH3 (1:1)+x% FIrpic/30 nm TmPyPB/20 nm
BPhen/Cathode, x=9–26%; (a) EL spectra at 5 mA/cm2, (b) EQE vs. current density, (c) current
density vs. voltage.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
400 500 600 700
Rela
tiv
e I
nte
nsi
ty (a
.u.)
Wavelength (nm)
9%
26%
FIrpic
(a)
0
5
10
15
20
0.1 1 10 100
EQ
E (
%)
Current Density (mA/cm2)
9%
12%
17%
26%
FIrpic
(b)
0.1
1
10
100
4 6 8 10
Cu
rren
t D
en
sity
(m
A/c
m2)
Voltage (V)
9%
12%
17%
26%
FIrpic
(c)
52
3.3.2 Mixed TCTA:UGH3 (x:y) hosts with fixed Flrpic concentration
Figure 3.4: Varying the mixed host doping ratio using FIrpic as emitting dopant.
The previous devices had a fixed host ratio while the FIrpic concentration was
varied. In the following set of devices, the host ratio was varied while fixing the FIrpic
concentration at 14%. The device layer structure is as follows: Anode/35 nm TAPC/40
nm TCTA+UGH3 (x:y)+14% FIrpic/30 nm TmPyPB/20 nm BPhen/Cathode (Figure 3.4).
The sum of x and y is equal to 86%. The TCTA and UGH3 concentrations were varied in
the range of 16%–70%. Figure 3.5(a) shows that the device EL spectra are invariant with
host compositions. Similar to what was observed by Lee et al., the EQE vs J curves
shown in Figure 3.5(b) indicate only minor differences with host composition. From 0.1–
5 mA/cm2, the EQE is 15–18% before decreasing to ~14–15% at 20 mA/cm
2. The J-V
data shown in Figure 3.5(c) show that the voltage is generally lower with increasing
TCTA concentration. The decrease in drive voltage is more than 1 V (at 5 mA/cm2) going
from low to high TCTA concentration. Devices using a mixed host of TCTA:UGH3 are
dominated by hole transport through the TCTA component with the UGH3 behaving
essentially as an inert component responsible for lowering the hole mobility in TCTA.
35 nm HTL
TAPC
30 nm ETL
TmPyPB
20 nm EIL
BPhen
40 nm Mixed host
TCTA:UGH3 (x:y)
Emitting dopant
14% FIrpic
Host Ratio (x:y)
70:16 – 16:70
Anode Cathode
53
Figure 3.5: OLED performance data at 5 mA/cm2 for varying mixed host composition while
keeping FIrpic at 14%; Anode/35 nm TAPC/40 nm TCTA+UGH3 (x:y)+14% FIrpic/30 nm
TmPyPB/20 nm BPhen/1 nm Cathode, x+y=86%; (a) EL spectra at 5 mA/cm2, (b) EQE vs. current
density, (c) current density vs. voltage.
The role of UGH3 will be discussed later within this chapter. As the TCTA concentration
becomes too low, i.e. 31% and 16%, hole transport is effectively impeded and the drive
voltage increases.
0.1
1
10
100
4 6 8 10
Cu
rren
t D
en
sity
(m
A/c
m2)
Voltage (V)
70%
55%
31%
16%
TCTA (x)
0
5
10
15
20
0.1 1 10 100
EQ
E (
%)
Current Density (mA/cm2)
70%
55%
31%
16%
TCTA (x)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
400 500 600 700R
ela
tiv
e I
nte
nsi
ty (a
.u.)
Wavelength (nm)
70%
16%
TCTA (x)
(a)
54
3.3.3 Effect of electron transport materials
Figure 3.6: Removing TmPyPB and using BAlq or BPhen next to the LEL for TCTA:UGH3 mixed
host devices.
It has been reported that the recombination zone for devices utilizing an LEL of
TCTA:UGH3:FIrpic is at or near the LEL/ETL interface [27]. Aza-aromatic compounds
such as TmPyPB that are commonly used as ETL materials have long been suspected of
contributing significantly to OLED device degradation [33]. To assess the degradation
due to the ETL materials, devices with alternate ETL materials were fabricated with the
following layer structure: Anode/35 nm TAPC/40 nm TCTA+UGH3 (1:1)+12%
FIrpic/ETL/Cathode (Figure 3.6). The ETL is either a single layer of 50 nm BPhen or a
bilayer of 10 nm BAlq/50 nm BPhen with BAlq adjacent to the LEL. The EL spectra for
both devices as shown in Figure 3.7(a) indicate that the EL emissions from FIrpic are
slightly modified. The FIrpic EL spectrum typically has a peak at 472 nm with a shoulder
at 496 nm. The EL spectrum for the device with BAlq/BPhen as the ETL has peaks at 472
nm and 496 nm of almost equal intensity. This may be due to 2 reasons; 1) an optical
interference effect [22, 34], and/or 2) emission from BAlq [35]. Light emission from the
LEL is isotropic. For device structures studied in this thesis, light reflects from the Al
35 nm HTL
TAPC
50 nm EIL
BPhen
40 nm Mixed host
TCTA:UGH3 (1:1)
Emitting dopant
12% FIrpic
BAlq
0 nm or 10 nm
ETLAnode Cathode
55
Figure 3.7: OLED performance data at 5 mA/cm2 for mixed host devices with alternate ETL’s;
Anode/35 nm TAPC/40 nm TCTA+UGH3 (1:1)+12% FIrpic/ETL/Cathode. ETL is 50 nm BPhen or
10 nm BAlq/50 nm BPhen; (a) EL spectra at 5 mA/cm2, (b) EQE vs. current density, (c) current
density vs. voltage.
cathode. There is an optimal distance from the LEL to the cathode that prevents optical
interference as light reflects from the Al. As the thickness between the LEL and cathode
changes, the EQE can also change due to this optical interference. Since recombination is
expected to occur nearest the LEL/ETL interface, it is also possible that BAlq can
contribute to the EL spectra, increasing the peak at 496 nm and cause a slight broadening
of the spectra. The EQE-J data for these two devices with modified ETL’s is shown in
Figure 3.7(b). At 0.1 mA/cm2, the EQE’s are 3-6% before rising to a maximum of 8-9%
at 20 mA/cm2. These EQE values are far less that what was observed for previous devices
0.0
0.2
0.4
0.6
0.8
1.0
1.2
400 500 600 700R
ela
tiv
e I
nte
nsi
ty (a
.u.)
Wavelength (nm)
BPhen
BAlq/BPhen
(a)
0.0
5.0
10.0
0.1 1 10 100
EQ
E (
%)
Current Density (mA/cm2)
BPhen
BAlq/BPhen
0.1
1
10
100
2 6 10
Cu
rren
t D
en
sity
(m
A/c
m2)
Voltage (V)
BPhen
BAlq/BPhen
(b) (c)
56
using TmPyPB/BPhen as the ETL. As discussed in Chapter 1, the transport layers should
have higher triplet energy (ET) than the dopant so as to prevent the triplet excitons from
being quenched. Unlike TmPyPB (ET~2.8 eV), neither BPhen (ET~2.6 eV) nor BAlq (ET
~ 2.2 eV) have triplet energies exceeding that of FIrpic (ET~2.6 eV), so these materials
are able to quench triplet excitons, thereby reducing the EQE. These devices are
dominated by hole current at low current densities, therefore the EQE is suppressed due
to recombination occurring predominantly at the LEL/ETL interface. As the current
density increases, the electron injection from the ETL into the LEL is increased, resulting
in a shift of the recombination away from the LEL/ETL interface and consequently
reducing exciton quenching. The device with BPhen as the ETL has a lower EQE
compared to the device with BAlq/BPhen. This is somewhat surprising considering that
BPhen has a higher ET compared to BAlq. One possible explanation is that the host TCTA
forms an exciplex with BPhen, which is a stronger acceptor compared to BAlq, resulting
in the formation of a lower energy complex that can result in additional exciton
quenching. The J-V data shown in Figure 3.7(c) indicates that the devices using BPhen
and BAlq/BPhen as the ETL have lower voltages than those using TmPyPB/BPhen. At 5
mA/cm2, BPhen, BAlq/BPhen, and TmPyPB/BPhen devices have voltages of 5.0 V, 5.8
V, and 6.4 V, respectively. The reason for this trend can be explained by the relative
alignment of the LUMO levels of these materials as shown in Figure 3.8. For all three
devices, electrons are injected into and transported through BPhen. Without the BAlq or
TmPyPB layer, electrons can inject directly from BPhen into the LEL layer and
recombine directly on FIrpic. With BAlq and TmPyPB, there are barriers of 0.1 eV and
57
0.2 eV for electron injection from BPhen into BAlq and TmPyPB, respectively, resulting
in higher drive voltages.
Figure 3.8: HOMO, LUMO, Triplet Energy levels for materials used to fabricate OLED devices in
Chapter 2. Triplet Energies are in parenthesis. Shaded boxes indicate typical electron transport
materials. All values in eV.
3.3.4 Device lifetime
Device lifetime, t50, measured at a fixed current density of 5 mA/cm2, refers to the
time duration for the initial EQE to decrease by 50%. The lifetime data for the mixed host
devices are shown in Figure 3.9. Regardless of the composition of the mixed hosts and
Flrpic concentrations, all devices using TmPyPB/BPhen as the ETL have a lifetime of
less than 20 minutes as represented by a single curve in Figure 3.9. The devices with
BPhen and BAlq/BPhen as the ETL have longer lifetimes of ~4 hours and ~12 hours,
respectively. The very short lifetime for the TmPyPB/BPhen devices is likely a result of
two factors: 1) these devices have high EQE’s, suggesting that the probability of FIrpic
excited states undergoing dissociation is increased; and 2) electron-hole recombination in
TmPyPB could result in bond breaking and the formation of quencher species. The
2.0
5.5
TCTA
(2.9)
TAPC
(2.9) TmPyPB
(2.8)FIrpic
(2.6)
2.3
5.7
3.0
5.9
6.7
2.8
7.2
2.8
UGH3
(3.5)
2.9
5.9
BAlq
(2.2) BPhen
(2.6)
3.0
6.4
HOMO
LUMO2.4
6.0
mCBP
(2.8)
2.3
5.4
NPB
(2.3)
58
device with the BPhen ETL has a slightly longer lifetime. The device with BAlq has the
longest lifetime in agreement with previous reports with BAlq used as the ETL. The
stability of BAlq may be attributed to its lower excited state energy as well as the stability
of its radical cation compared to TmPyPB.
Figure 3.9: Device lifetime taken at 5 mA/cm2 for mixed-host devices; Anode/35 nm TAPC/40 nm
TCTA+UGH3 (1:1)+12% FIrpic/ETL/Cathode. ETL is 50 nm BPhen, 10 nm BAlq/50 nm BPhen,
or 30 nm TmPyPB/20 nm BPhen.
3.4 SINGLE HOST
While a mixed host is known to improve device performance, it is often beneficial
to simplify the device structure for the ease of fabrication and for a better understanding
of the device operation. To this end, there has been significant research effort aimed at
designing and synthesizing bipolar host materials that can perform both hole transport
and electron transport functions [36-41]. The HOMO/LUMO levels can be tuned based
on the functional groups within the molecule, and typically include both hole transport
and electron transport moieties that are electronically isolated from each other using
various spacer groups or substitution patterns. In this way, only a single host material is
0.5
0.6
0.7
0.8
0.9
1
1.1
0 5 10
Rela
tiv
e E
QE
Time (hours)
TmPyPB/BPhen
BAlq/BPhen
BPhen
ETL
59
required within the LEL.
Carbazole-based derivatives have been used extensively as host materials for
phosphorescent device [21, 42-44]. CBP is one such derivative that reportedly has bipolar
characteristics [45, 46]. While this material does not have an electron withdrawing
moiety, the central biphenyl group linking the two carbazole units allows for -orbital
overlap that is conducive for electron transport. CBP has a triplet energy of 2.6 eV; which
is low enough that it can quench FIrpic triplet excitons. We fabricate a series of devices to
compare the performance of two different single host carbazole-based materials, TCTA
and mCBP. Both materials have triplet energies higher than FIrpic. TCTA was used as the
hole transporting component in the mixed-host system in section 3.3. This material lacks
a biphenyl moiety and is not considered a bipolar host material. In contrast, mCBP is
similar to CBP and has been used to transport both holes and electrons [5, 47]. Similar to
section 3.3, the ETL materials are also varied for these two host materials. The HTL is
varied for devices with mCBP as the host.
3.4.1 TCTA single host and FIrpic concentration series
Figure 3.10: FIrpic concentration series for TCTA single host OLED devices.
35 nm HTL
TAPC
30 nm ETL
TmPyPB
20 nm EIL
BPhen
40 nm Single host
TCTA
Emitting dopant
FIrpic
Concentration
3-26%
Anode Cathode
60
In this series TCTA is used as the single host component in a layer structure as
follows: Anode/35 nm TAPC/40 nm TCTA+x% FIrpic/30 nm TmPyPB/20 nm
BPhen/Cathode (Figure 3.10). The range of Flrpic concentration in TCTA is 3–26%.
Device data taken at 5 mA/cm2 is shown in Table 3.2.
Table 3.2: Device data taken at 5 mA/cm2 for single host TCTA and FIrpic concentrations 3-6%.
FIrpic (%)
V cd/m2
EQE (%)
t50
(min.)
3 6.6 1176 10.9 3
6 6.7 1137 10.6 4
11 6.5 1523 13.8 7
26 6.4 1264 10.8 25
Similar to the mixed host devices, devices utilizing TCTA as a single host are
dominated by hole transport, with recombination occurring at the LEL/ETL interface.
The drive voltage is relatively independent of the FIrpic concentration. The EQE ranges
from 11–14%, with the highest efficiency occurring at a concentration of 12% FIrpic.
These EQE values are considerably lower than what was observed for mixed host devices
(17-18%). A possible explanation for the lower EQE is the formation of exciplex between
TCTA and TmPyPB. Exciplexes are known to form between hole transporting materials
and electron transporting materials [22, 48, 49]. With TCTA as the single host,
recombination is expected to occur at TCTA/TmPyBP interface and therefore exciplex
formation at this interface will likely affect the EQE. We examined the possibility of
exciplex formation using the structure: Anode/30 nm TAPC/30 nm TCTA/30 nm
TmPyPB/Cathode, where recombination is expected to occur at the TCTA/TmPyPB
61
interface. The EL spectrum shown in Figure 3.11 shows a broad peak centered at 500 nm
which can be attributed to the TCTA/TmPyPB exciplex emission. The 415nm peak is
from TCTA [48]. The TCTA/TmPyPB exciplex, which has a lower energy than FIrpic, is
expected to quench FIrpic excitons, resulting in a lower EQE. An alternative mechanism
for the lower EQE is quenching due to TCTA cations [50]. A large build-up of TCTA
cations at the TCTA/TmPyPB interface could result in FIrpic exciton quenching due to
polaron-exciton interactions.
Figure 3.11: EL spectrum at 20 mA/cm2 showing exciplex of TCTA:TmPyPB for undoped device.
Anode/30 nm TAPC/30 nm TCTA/30nm TmPyPB/Cathode.
As shown in Table 3.2, the lifetime of the single-host devices is very short,
ranging from 3 to 25 minutes. There is a weak dependence on FIrpic concentration with
lifetime increasing as FIrpic concentration increases. This trend suggests that FIrpic is not
the material limiting the lifetime. As mentioned earlier, it is quite likely that TmPyPB is
the key culprit contributing to the short device lifetime.
0
0.2
0.4
0.6
0.8
1
1.2
400 500 600 700
Rela
tiv
e I
nte
nsi
ty (a
.u.)
Wavelength (nm)
TCTA:TmPyPB
Exciplex emission peak
Anode/TAPC/TCTA/TmPyPB/Cathode
62
3.4.2 TCTA single host and UGH3 as electron transport material
Figure 3.12: Removing TmPyPB and using UGH3 or BPhen next to the LEL for single host
devices with FIrpic concentration at 1% or 12%.
In this series of devices, TmPyPB was replaced by UGH3 as the ETL. UGH3 has
been used in mixed host systems [14, 51] as well as for the ETL [51-53]. There have been
no reports of device lifetime specifically related to this material. UGH3 lacks
heterocyclic moieties that have been linked to instability. The device structure is as
follows: Anode/35 nm TAPC/40 nm TCTA+12% FIrpic/x nm UGH3/50 nm
BPhen/Cathode (Figure 3.12). The UGH3 thickness is varied from 0–30 nm. Device data
at 5 mA/cm2 is shown in Table 3.3.
Without the UGH3 layer, BPhen is in direct contact with the LEL, supporting both
electron injection and transport. The drive voltage is only 4.7 V, apparently due to the
absence of barriers to electron transport from BPhen to the LEL. The EQE is 4.7%, which
is low due to triplet exciton quenching by BPhen. Inserting a 2.5 nm UGH3 layer raised
the voltage by 0.4 V to 5.1 V. The drive voltage rises to 8.8 V with a 20 nm UGH3 layer.
35 nm HTL
TAPC
50 nm EIL
BPhen
UGH3
0–30 nm
ETL40 nm Single host
TCTA
Emitting dopant
FIrpic
Concentration
1% or 12%
Anode Cathode
63
Table 3.3: OLED performance data at 5 mA/cm2 for devices using UGH3 as the ETL and 12%
FIrpic in the LEL; Anode/35 nm TAPC/40 nm TCTA+12% FIrpic/x nm TmPyPB/y nm UGH3/z
nm BPhen/Cathode.
TmPyPB (x nm)
UGH3 (y nm)
BPhen (z nm)
V cd/m2
EQE (%)
t50 (min.)
― 0 50 4.7 491 4.7 200
― 2.5 50 5.1 1203 11.5 35
― 5.0 50 5.3 1336 13.1 10
― 10.0 50 6.6 281 2.7 11
― 20.0 50 8.8 352 3.0 120
30 ― 20 6.5 1523 13.8 7
― 30 20 10.2 318 3.2 200
In order to understand the voltage dependence, the following two devices were
fabricated: Anode/35 nm TAPC/40 nm TCTA+12% FIrpic/30 nm UGH3 or TmPyPB/20
nm BPhen/Cathode. Table 3.3 shows the device performance for these two devices at 5
mA/cm2. The voltages are 6.5 V and 10.2 V for the TmPyPB and UGH3 devices,
respectively. Despite the total ETL/BPhen thicknesses being the same, the UGH3 device
voltage is almost 4 V higher than the TmPyPB device. There is a barrier of 0.2 eV (Figure
3.8) for electron injection from BPhen into either UGH3 or TmPyPB. Considering that
the electron barrier heights are the same and the UGH3 device has a higher voltage
compared to the TmPyPB device, this would indicate that UGH3 has a much lower
electron mobility compared to TmPyPB.
With 2.5 nm and 5.0 nm layers of UGH3, the EQE’s are 9.2% and 11.7%, more
than double that of the device without UGH3. Further increasing the UGH3 thickness to
10.0–20.0 nm results in a significant decrease in EQE (to 2‒3%). It would seem that a
thin UGH3 layer, i.e. 2.5 nm or 5.0 nm, is effective for reducing triplet exciton quenching
by BPhen while still allowing for electrons to transit the UGH3 layer and to recombine
64
with the holes at the LEL/ETL interface. It is inferred that a minimum thickness of 5 nm
of UGH3 is required to separate the LEL and BPhen layers. In fact, it is possible that as
the UGH3 thickness increases from 10.0 to 20.0 nm, electron injection from the BPhen
layer is so impeded that the holes are drawn into the UGH3 to cause electron-hole
recombination within the UGH3 layer or at the UGH3/BPhen interface.
Lifetimes for all devices with UGH3 as the ETL are improved compared to
devices with TmPyPB as the ETL. For the device using BPhen as the ETL and EIL, the
lifetime is highest, reaching 200 minutes. The lifetime decreases to 35 min. for the device
with 2.5 nm of UGH3 and levels out at 10 minutes for UGH3 thicknesses between 5.0–
10.0 nm. With UGH3 thicknesses of 20 nm and 30 nm, the lifetime increases to 120
minutes and 200 minutes, respectively. While lifetime can be improved by using UGH3
instead of TmPyPB, the EQE is generally lower due to UGH3 being a poor electron
transporting material.
The data shown here suggests that UGH3 does not support electron transport in a
mixed host with TCTA. It is more likely that UGH3 impedes hole transport in the mixed
host, and thus causes a shift of recombination away from the LEL interface. UGH3 could
also reduce the formation of exciplex between TCTA and TmPyPB. For these reasons, the
EQE’s for the mixed TCTA/UGH3 host systems are higher than those of the TCTA single
host systems.
To further evaluate UGH3 as the ETL, the following structure was fabricated:
Anode/35 nm TAPC/40 nm TCTA+1% FIrpic/x nm UGH3/50 nm BPhen/Cathode, where
the UGH3 thickness is varied from 0–20 nm. Device data is shown in Table 3.4. Flrpic
65
concentration was kept at 1%. By using a low FIrpic concentration, recombination
occurring directly on FIrpic is reduced. Device lifetime will become more dependent on
recombination occurring on other materials. If the device lifetime increases at low FIrpic
concentrations, FIrpic is limiting the lifetime. However, if device lifetime decreases at
low FIrpic concentrations, device lifetime is being limited by other materials within the
device.
Table 3.4: OLED performance data at 5 mA/cm2 for devices using UGH3 as the ETL and 1%
FIrpic in the LEL. Anode/35 nm TAPC/40 nm TCTA+12% FIrpic/x nm TmPyPB/y nm UGH3/z nm
BPhen/Cathode.
UGH3
(nm) V cd/m
2
EQE (%)
t50
(min.)
0 4.7 41 0.5 12
2.5 5.2 290 2.9 8
5.0 5.4 419 4.2 4
10.0 6.4 121 1.2 1
20.0 9.0 299 2.6 56
The EL spectra shown in Figure 3.13 indicate that the emission occurs
predominantly from FIrpic for all devices regardless of the UGH3 thicknesses. All the
devices with UGH3 show additional emission occurring around 410 nm that is attributed
to TCTA (Figure 3.13, inset). The device without UGH3 has a broad shoulder from 400-
450 nm that is possibly due to an exciplex formed between TCTA and BPhen. It is not
possible for all excitons to form on FIrpic with the concentration being so low. As a
result, the excitons are formed on the materials at the LEL/ETL interface.
Since TCTA dominates hole transport and recombination occurs at the TCTA/ETL
interface, the voltages for these low FIrpic concentration devices range from 4.7–9.0 V,
66
Figure 3.13: EL spectra at 5 mA/cm2 for devices using UGH3 as the ETL and 1% FIrpic in the
LEL. Anode/35 nm TAPC/40 nm TCTA+12% FIrpic/y nm UGH3/50 nm BPhen/Cathode, y=0–20
nm. Inset: Enhanced region from 400-450 nm showing TCTA emission.
which is comparable to the devices that had 12% FIrpic (4.7–8.8 V). The EQE’s for these
1% devices are significantly reduced compared to the 12% devices, reaching a maximum
of 4.2% for the device with 5 nm of UGH3. As shown by the EL spectra, the lower EQE
is due to a reduction in the concentration of highly efficient FIrpic recombination centers,
resulting in incomplete capture of the triplet excitons by the FIrpic molecules. Emission
is observed from the inefficient transport layers. The device lifetimes for the 1% devices
range from 1–56 minutes, much shorter than what was observed for the 12% devices (7–
200 minutes). With device lifetime decreasing as the FIrpic concentration is reduced,
FIrpic is presumably not the key factor in limiting the device lifetime. This implicates the
host and transport layers as culprits for causing device degradation if recombination
occurs on these materials.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
400 500 600 700
Rel
ati
ve
Inte
nsi
ty (a.u
.)
Wavelength (nm)
0 nm
5 nm
20 nm
UGH3
0.00
0.02
0.04
0.06
0.08
400 425 450
Rela
tiv
e I
nte
nsi
ty
Wavelength (nm)
67
3.4.3 TCTA host and BAlq as electron transport material
Figure 3.14: BAlq or BPhen next to the LEL for TCTA single host devices.
BAlq was used as the ETL for the mixed host devices (section 3.3.3) to improve
device lifetime. Similar improvements were observed in single host devices with the
following structure: Anode/35 nm TAPC/40 nm TCTA+12% FIrpic/x nm BAlq/50 nm
BPhen/Cathode (Figure 3.14). This structure is similar to the devices using UGH3 as the
ETL. The BAlq thickness is varied from 0–20 nm. Device data is shown in Table 3.5.
Electroluminescence originates from FIrpic.
Table 3.5: OLED performance data at 5mA/cm2 for devices using TCTA as single host and BAlq as
the ETL. Anode/35 nm TAPC/40 nm TCTA+12% FIrpic/x nm BAlq/50 nm BPhen/Cathode.
BAlq (nm)
V cd/m2
EQE (%)
t50
(min.)
0 4.6 539 5.0 200
2.5 4.9 705 5.6 450
5.0 5.1 843 7.3 390
10.0 5.6 844 7.2 350
20.0 6.6 821 6.5 390
The voltage increases with increasing BAlq thickness, which is the same behavior
35 nm HTL
TAPC
50 nm EIL
BPhen
40 nm Mixed host
TCTA:UGH3 (1:1)
Emitting dopant
12% FIrpic
BAlq
0 nm or 10 nm
ETLAnode Cathode
68
that was observed for the UGH3 series of devices. This is most likely due to the fact that
with a sufficiently thin layer, both UGH3 and BAlq behave as tunneling layers separating
the BPhen layer from the LEL layer, and as such the drive voltages are similar. For
thicker layers of 10.0 nm and 20.0 nm, the voltages for BAlq devices are 5.6 V and 6.6 V
respectively compared to 6.7 V and 8.8 V for the UGH3 devices. The lower voltages for
BAlq based devices are likely due to higher electron mobility in BAlq compared to
UGH3. With thinner films (2.5 to 5.0 nm), the BAlq devices have lower EQE’s compared
to those with UGH3 because BAlq, with a triplet energy lower than that of FIrpic, is
capable of quenching FIrpic excitons. For BAlq or UGH3 thicknesses of 10.0 nm or 20.0
nm, the BAlq devices have higher EQE’s than the UGH3 devices. The field across BAlq
is lower than that of UGH3, resulting in improved electron transport and
recombination/emission occurring from the LEL/ETL interface. The larger field across
UGH3 results in reduced recombination at the LEL/ETL interface, with the possibility
that some percentage of recombination occurs within the UGH3 layer or possibly at the
UGH3/BPhen interface. While BAlq (ET=2.2 eV) has a lower triplet energy than BPhen
(ET=2.6 eV) and would be expected to at least quench FIrpic excitons to the same degree,
the devices with BAlq have higher EQE’s than BPhen only devices. It is possible that due
to a lower electron injection barrier into the LEL, recombination for the BAlq devices is
moved slightly away from the quenching interface.
The lifetimes for devices with BAlq reach 6–8 hours, which is a significant
improvement over devices with UGH3 (10 minutes to 3 hours). This lifetime
improvement is due to two reasons. 1) For thin layers, i.e. 2.5 nm and 5.0 nm, the BAlq
69
devices have lower EQE’s which is due to fewer FIrpic molecules undergoing radiative
emission. As a result, fewer FIrpic molecules degrade and lifetime increases. 2) For
thicker layers, i.e. 10.0 nm and 20.0 nm, the EQE for the BAlq devices doesn’t shift
much, so lifetime is relatively constant. For the UGH3 devices, the EQE decreases
dramatically due to recombination likely occurring within UGH3. Recombination outside
the LEL is more detrimental to device lifetime than recombination occurring within the
LEL.
3.4.4 mCBP host and FIrpic concentration series
Figure 3.15: FIrpic concentration series for mCBP single host devices.
As mentioned earlier, stable PHOLEDs using mCBP as the host material with
Ir(tdmp)3 as the blue phosphorescent dopant have been demonstrated. At an initial
luminance of 500 cd/m2, device lifetimes up to 10,000 hours have been achieved. The
HOMO and LUMO energy values for mCBP are 6.0 eV and 2.4 eV, respectively, both
deeper than TCTA (Figure 3.8). The triplet energy of 2.8 eV is higher than FIrpic. In the
following structure, TCTA was replaced by mCBP in the LEL: Anode/35 nm
TAPC/mCBP+x% FIrpic/30 nm TmPyPB/20 nm BPhen/Cathode (Figure 3.15), where the
35 nm HTL
TAPC
30 nm ETL
TmPyPB
20 nm EIL
BPhen
40 nm Single host
mCBP
Emitting dopant
FIrpic
Concentration
2-18%
Anode Cathode
70
FIrpic concentration is varied from 2–18%. Device performance is shown in Figure 3.16
and data taken at 5 mA/cm2 is shown in Table 3.6.
Table 3.6: OLED performance at 5 mA/cm2 for devices using mCBP as host and varying the FIrpic
concentration from 2–18%.
FIrpic (%)
V cd/m2 EQE
(%)
t50
(min.)
2 9.2 1661 15.5 -
6 9.0 1762 16.3 4
12 8.7 1619 14.4 6
18 8.4 1231 10.4 9
The EL spectra shown in Figure 3.16(a) indicate that emission occurs from FIrpic
for all devices and is independent of FIrpic concentration. Figure 3.16(b) show the J-V
data. As the concentration of FIrpic increases from 2 to18%, the voltage decreases
monotonically from 9.2 V to 8.4 V at 5 mA/cm2, which is different from the TCTA-based
devices. In TCTA devices, there was essentially no change in voltage as FIrpic
concentration increased. Furthermore, the mCBP-based devices require significantly
higher voltages regardless of the dopant concentration. TCTA-based devices are
dominated by hole transport, with holes building up at the TCTA/ETL interface where
recombination occurs. As a result, there was little change in voltage even if the dopant
concentration was over 20%. The higher voltage for mCBP devices results from two
factors, 1) with mCBP as the host, there is a 0.5 eV hole injection barrier from TAPC into
mCBP, creating a field across the LEL and 2) mCBP lacks the triarylamine moiety that is
present in TCTA, making it likely that mCBP has a lower hole mobility. These two
71
factors lead to fewer holes building up at the LEL/ETL interface, allowing electrons to
more easily inject from TmPyPB directly onto the FIrpic molecules. As the FIrpic
concentration increases, electron injection is also easier and the voltage decreases.
Figure 3.16(c) shows the EQE-J data. At 0.1 mA/cm2 the EQE’s for the 2–12%
FIrpic devices range from 8%–14%. The EQE’s rise to ~14–17% at 2 mA/cm2 before
decreasing again at higher current densities. These EQE’s are generally higher than what
was observed for TCTA single-host devices (~14% maximum EQE) and lower than the
mixed-host devices (~20% maximum EQE). The TCTA single-host devices had triplet
quenching due to exciplex formation between TCTA and TmPyPB whereas the UGH3
Figure 3.16: OLED performance for devices with mCBP as host and 2-18% FIrpic, Anode/35 nm
TAPC/mCBP+x% FIrpic/30 nm TmPyPB/20 nm BPhen/Cathode.
0.1
1
10
100
4 8 12
Cu
rren
t D
en
sity
(m
A/c
m2)
Voltage (V)
2%
6%
12%
18%
0
4
8
12
16
0.1 1 10 100
EQ
E (
%)
Current Density (mA/cm2)
2%
6%
12%
18%
0.0
0.2
0.4
0.6
0.8
1.0
1.2
400 500 600 700
Rela
tiv
e I
nte
nsi
ty (a
.u.)
Wavelength (nm)
2%
6%
12%
18%
(a)
(b) (c)
72
component of the mixed host helps to prevent the TCTA:TmPyPB exciplex from
forming. The following simplified device was fabricated to investigate whether mCBP
also forms an exciplex with TmPyPB: Anode/75 nm mCBP/75 nm TmPyPB/Cathode.
Figure 3.17 shows the EL spectrum for this device taken at 20 mA/cm2. The peak
centered at ~550 nm is not due to either mCBP [54] or TmPyPB [55]. This indicates that
this is due to an exciplex formed between these two materials. The formation of this
exciplex is likely the reason why the EQE’s are lower than what was observed for mixed-
host devices. Since fewer holes arrive at the mCBP/TmPyPB interface compared to the
TCTA/TmPyPB interface, the mCBP:TmPyPB exciplex may not be as strong as the
TCTA:TmPyPB exciplex, resulting in higher EQE’s for the mCBP devices compared to
Figure 3.17: EL spectrum at 20 mA/cm2 showing exciplex of mCBP:TmPyPB. Anode/75 nm
mCBP/75 nm TmPyPB/Cathode.
TCTA single-host devices. This exciplex may also explain the EQE trend observed with
current density. At low current densities, hole transport is expected to dominate over
electron transport, with recombination occurring at the mCBP/TmPyPB interface. As the
current density increases, more electrons arrive at the mCBP/TmPyPB interface which
0.0
0.2
0.4
0.6
0.8
1.0
1.2
400 500 600 700
Rela
tiv
e I
nte
nsi
ty (a
.u.)
Wavelength (nm)
Anode/mCBP/TmPyPB/Cathode
mCBP:TmPyPB
Exciplex emission peak
73
decreases the formation of exciplex and increases the EQE. The EQE for the device with
18% FIrpic stays between 8–10% regardless of the current density. This low EQE is
attributed to either dopant self-quenching or triplet-triplet annihilation.
The device lifetimes using mCBP as host are all very short, less than 10 minutes
at 5 mA/cm2. This is similar to what was observed for other devices utilizing FIrpic as the
emitting dopant with TmPyPB as the electron transport layer.
3.4.5 mCBP host and BAlq as electron transport material
Figure 3.18: Removing TmPyPB and using BAlq or BPhen next to the LEL for mCBP single host
devices with the FIrpic concentration at 6% or 12%.
BAlq was used as the electron transport for both mixed-host and TCTA single-
host devices to improve lifetime. It was also used to evaluate lifetime with mCBP as the
host in the following device structure: Anode/35 nm TAPC/40 nm mCBP+12% FIrpic/x
nm BAlq/50 nm BPhen/Cathode (Figure 3.18). The BAlq ETL is varied from 0‒20 nm.
The FIrpic concentration was fixed at 12%, the same as what was used in the TCTA
35 nm HTL
TAPC
50 nm EIL
BPhen
BAlq
0–20 nm
ETL40 nm Single host
mCBP
Emitting dopant
FIrpic
Concentration
6% or 12%
Anode Cathode
74
device structure with BAlq. Device data at 5 mA/cm2 are shown in Table 3.7.
Electroluminescence originates from FIrpic.
As the BAlq thickness increases from 0–10 nm, the voltage increases from 6.7–
8.5 V. This trend, which was also observed for the devices with TCTA as the host and
BAlq as the ETL, arises due to the generation of an electric field across BAlq based on
the electron injection barrier of 0.2 eV from BPhen into BAlq (Figure 3.8). With BAlq
quenching triplet excitons, the EQE ranges from 4.8–6.7%, which is about half of what
was achieved compared to devices with TmPyPB as the ETL. The lifetime ranges from
75–220 minutes, an improvement due to the lower EQE and the removal of TmPyPB
from the device.
Table 3.7: OLED performance data at 5 mA/cm2 for devices using mCBP as the single host and
BAlq as the ETL. Anode/35 nm TAPC/40 nm mCBP+12% FIrpic/x nm BAlq/50 nm
BPhen/Cathode.
BAlq (nm)
V cd/m2 EQE
(%)
t50
(min.)
0 6.7 736 6.6 210
2.5 7.0 771 6.7 90
5.0 7.1 770 6.6 220
10.0 7.4 762 6.4 75
20.0 8.5 620 4.8 110
While the previous devices that replaced TmPyPB with BAlq had FIrpic doped at
12% to compare with the TCTA devices, doping 6% FIrpic yields the optimum EQE for
mCBP devices. In the following device structure, BAlq was varied from 0–20 nm while
FIrpic was fixed at 6%: Anode/35 nm TAPC/40 nm mCBP+6% FIrpic/x nm BAlq/50 nm
BPhen/Cathode. Device performance is shown in Figure 3.13.
75
As the BAlq thickness increases from 0–10 nm, the voltage increases from 7.8–
8.9 V, a trend that has been observed and described for other devices using BAlq. BAlq
also quenches the EQE, resulting in a range from ~6.6-8%. These values are
approximately half those achieved compared to devices using TmPyPB as the ETL. The
lifetime ranges from 250-600 minutes, which is similar to the TCTA devices that use
BAlq, and an improvement over the mCBP devices that 12% FIrpic and have BAlq as the
ETL. By decreasing the FIrpic concentration, the lifetime improves.
Table 3.8: OLED performance data at 5 mA/cm2 for devices using mCBP as the single host and
BAlq as the ETL. Anode/35 nm TAPC/40 nm mCBP+6% FIrpic/x nm BAlq/50 nm
BPhen/Cathode.
BAlq (nm)
V cd/m2
EQE t50
(min.)
0 7.8 848 8.0 250
2.5 8.0 714 6.6 600
5.0 8.0 830 7.4 425
10.0 8.3 856 7.5 400
20.0 8.9 793 6.6 500
76
3.4.6 mCBP host and NPB as hole transport layer
Figure 3.19: Using BAlq as ETL and NPB as HTL at FIrpic concentrations of 6% and 12%.
In the prior experiments, it was found that modification of the ETL significantly
impacts device performance. In order to understand how the choice of HTL impacts
device performance, TAPC was replaced with NPB in the following device structure:
Anode/35 nm NPB/40 nm mCBP+12% FIrpic/x nm BAlq/50 nm BPhen/Cathode,
whereas the BAlq thickness varies from 5–20 nm. The FIrpic concentration was fixed at
12%. NPB has a lower triplet energy (2.3 eV) compared to TAPC (2.9 eV), however
TAPC has been found to be less stable than NPB as a hole transporting material. Device
data taken at 5 mA/cm2 is shown in Figure 3.16.
Electroluminescence originates from FIrpic. Compared to the devices in Table 3.7
that had TAPC as the HTL, these devices have a voltage that is approximately 1 V higher
while the EQE drops by ~50%. The higher voltage may be due to a higher injection
barrier from NPB into the mCBP host while the lower EQE indicates that NPB quenches
50 nm EIL
BPhen
BAlq
5–20 nm
ETL40 nm Single host
mCBP
Emitting dopant
FIrpic
Concentration
6% or 12%
35 nm HTL
NPB
Anode Cathode
77
the triplet excitons.
The devices with NPB and BAlq have a 4-6x lifetime improvement compared to
those with TAPC and BAlq shown in Table 3.7. This is likely due to a decrease in
radiative recombination (from NPB quenching excitons) and a subsequent decrease in the
resulting excited state reactions. It also may be due to NPB being more stable compared
to TAPC.
In order to understand how NPB affects device performance with a FIrpic
concentration of 6%, devices were fabricated using the following structure: Anode/35 nm
NPB/40 nm mCBP+6% FIrpic/x nm BAlq/50 nm BPhen/Cathode. The BAlq thickness
was either 5 nm or 10 nm. Device performance at 5 mA/cm2 is shown in Table 3.9.
Table 3.9: OLED performance data at 5 mA/cm2 for devices using NPB as HTL, BAlq as ETL, and
x% FIrpic. Anode/35 nm NPB/40 nm mCBP+x% FIrpic/y nm BAlq/50 nm BPhen/Cathode.
FIrpic (%)
BAlq (nm)
V cd/m2 EQE
(%)
t50
(min.)
12 5.0 8.1 341 2.9 450
12 10.0 8.5 338 2.8 500
12 20.0 9.5 286 2.3 550
6 5.0 8.7 677 6.5 380
6 10.0 9.0 755 7.1 380
The 5 nm and 10 nm BAlq devices have voltages of 8.7 V and 9.0 V, respectively.
The EQE’s are ~7% while the lifetimes are around 400 minutes. With a FIrpic
concentration of 12%, replacing TAPC with NPB resulted in a voltage increase of 1.0 V
and an EQE decrease of 50%. By adjusting the FIrpic concentration to 6%, replacing
TAPC with NPB resulted in the voltage increasing by 0.7 V while the EQE decrease was
78
marginal. As mentioned earlier, the voltage increase is attributed to the hole injection
barrier into the LEL being larger for NPB. It is expected that the devices using mCBP as
the HTL are still dominated by hole transport, with recombination occurring primarily at
the LEL/ETL interface. Switching to NPB didn’t significantly affect the EQE for the 6%
FIrpic devices, unlike what was observed for the 12% devices, indicating a shorter
exciton diffusion length for decreased FIrpic concentrations. The reduced FIrpic
concentration results in fewer excitons diffusing to and getting quenched at the NPB/LEL
interface. The lifetime for the 6% FIrpic devices are also very similar, regardless if the
HTL is NPB or TAPC.
3.5 CONCLUSIONS
With the use of high triplet charge transport materials, TAPC as the HTL and
TmPyPB as the ETL, high EQE’s over 15% were achieved with FIrpic as the light-
emitting dopant. While the efficiency is high the device lifetimes are all very short (less
than 20 minutes at 5 mA/cm2) regardless of the host configuration for the emitting layer.
Device lifetime can be improved to 2–11 hours by replacing TmPyPB with
alternate ETL materials such as BAlq, BPhen, and UGH3. However the EQE is reduced
significantly by utilizing these materials. BAlq and BPhen have lower triplet energies
than FIrpic and quench triplet excitons. UGH3 has a high triplet energy of 3.5 eV and
does not quench FIrpic excitons. However, due to a large electron injection barrier into
UGH3 (from BPhen) accompanied by low electron mobility, UGH3 is a poor electron
transport material and prevents efficient recombination from occurring in the LEL.
However, a thin layer of UGH3 between the LEL and BPhen can be used effectively as a
79
buffer to reduce quenching from the lower triplet energy BPhen (2.6 eV).
Use of a mixed host LEL (TCTA:UGH3) in combination with the low triplet
energy BAlq ETL resulted in the best combination for EQE (~9%) and lifetime (~12
hours). This lifetime is far greater compared to the lifetime of a few minutes which is
observed by using TmPyPB as the ETL. With UGH3 being a poor electron transport
material, its role within the mixed host LEL is that of a spacer group, slowing down hole
transport to mitigate recombination occurring at the LEL/ETL interface.
A study of undoped devices indicates that TmPyPB forms charge transfer
complexes (exciplex) with both TCTA and mCBP. This explains the enhanced efficiency
in mixed host devices utilizing UGH3 and TCTA. With UGH3 being used primarily as a
spacer component, the direct contact between TCTA and TmPyPB is decreased, which
limits or prevents exciplex formation.
It was found that devices with 1% FIrpic doped into TCTA have shorter lifetimes
than devices with 12% FIrpic. At 1% doping, there are not enough dopant molecules to
capture all excitons. Instead, excitons form on the host and/or transport materials and
emission occurs from these materials. This indicates that recombination and emission
occurring on the host/transport materials is detrimental to device lifetime.
Using mCBP as a single host does not improve lifetime. The voltage is higher if
mCBP is used as the host compared to TCTA, indicating that the hole injection barrier
into mCBP is higher and/or the hole mobility in mCBP is decreased. The EQE’s for
mCBP devices are higher than TCTA devices, which is explained by 1) a reduction in
charge quenching by accumulated holes at the LEL/ETL interface; and 2) a decrease in
80
the formation of the mCBP:TmPyPB exciplex (compared to TCTA:TmPyPB exciplex)
that can quench triplet excitons.
The effect of replacing the high triplet energy HTL TAPC (2.9 eV) with low
triplet energy NPB (2.3 eV) was dependent on the concentration of FIrpic within the
LEL. For devices with 12% of FIrpic doped into mCBP, replacing TAPC with NPB
results in the EQE decreasing by ~50%. Reducing the FIrpic concentration to 6%, there is
minimal loss in EQE after replacing TAPC with NPB. This difference in the quenching
between the two different FIrpic concentrations is attributed to a reduction of triplet
excitons (which need to diffuse through FIrpic molecules) at the HTL/LEL interface.
With higher concentrations of FIrpic, the triplet exciton formed at the LEL/ETL interface
is more likely to diffuse to the HTL/LEL interface to get quenched. There is also ~1 V
increase in drive voltage after replacing TAPC with NPB, indicating increased hole
accumulation at the NPB/LEL interface that can also adversely affect the emission
efficiency. Device lifetime is approximately the same regardless if NPB or TAPC is used.
It is concluded that the blue phosphorescent dopant FIrpic along with the transport
and host materials all contribute to device instability. FIrpic is extremely unstable as a
light-emitting material, regardless of the host or transport materials. Low triplet energy
ETL’s such as BAlq or BPhen quench triplet excitons, effectively slowing down excited
state reactions involving FIrpic. Recombination occurring on the host or aza-aromatic
electron transport materials is also detrimental to device lifetime, indicating that these
materials are also unstable in the excited state.
81
REFERENCES
[1] Mertens, R., www.oled-info.com/devices, 2014
[2] Universal Display Corporation, www.udcoled.com, 2014
[3] N.C. Giebink, B.W. D'Andrade, M.S. Weaver, P.B. Mackenzie, J.J. Brown, M.E.
Thompson, S.R. Forrest, J. Appl. Phys., 103 (2008) 044509.
[4] H. Yamamoto, J. Brooks, M.S. Weaver, J.J. Brown, T. Murakami, H. Murata,
Appl. Phys. Lett., 99 (2011) 033301.
[5] B. D'Andrade, P.B. MacKenzie, M.S. Weaver, J.J. Brown, U.S. Patent No.
8,557,399 (2013).
[6] I. Rabelo de Moraes, S. Scholz, B. Luessem, K. Leo, Org. Electron., 12 (2011)
341.
[7] V. Sivasubramaniam, F. Brodkorb, S. Hanning, H.P. Loebl, V. van Elsbergen, H.
Boerner, U. Scherf, M. Kreyenschmidt, J. Fluorine Chem., 130 (2009) 640.
[8] W. Sotoyama, T. Satoh, M. Kinoshita, M. Tobise, K. Kawato, T. Ise, H. Takizawa,
S. Yamashita, Dig. Tech. Pap. - Soc. Inf. Disp. Int. Symp., 41 (2010) 556.
[9] H. Lee, H. Ahn, C. Lee, J. Inf. Disp., 12 (2011) 219.
[10] K.S. Yook, O.Y. Kim, J.Y. Lee, Synth. Met., 161 (2011) 2677.
[11] N. Chopra, J. Lee, Y. Zheng, S.-H. Eom, J. Xue, F. So, Appl. Phys. Lett., 93
(2008) 143307.
[12] H. Sasabe, E. Gonmori, T. Chiba, Y.-J. Li, D. Tanaka, S.-J. Su, T. Takeda, Y.-J. Pu,
K.-i. Nakayama, J. Kido, Chem. Mater., 20 (2008) 5951.
[13] S.-J. Su, E. Gonmori, H. Sasabe, J. Kido, Adv. Mater., 20 (2008) 4189.
82
[14] J. Lee, J.-I. Lee, J.Y. Lee, H.Y. Chu, Appl. Phys. Lett., 94 (2009) 193305.
[15] L. Xiao, S.-J. Su, Y. Agata, H. Lan, J. Kido, Adv. Mater., 21 (2009) 1271.
[16] C.W. Lee, J.Y. Lee, Adv. Mater., 25 (2013) 1.
[17] V.-E. Choong, S. Shi, J. Curless, C.-L. Shieh, H.-C. Lee, F. So, J. Shen, J. Yang,
Appl. Phys. Lett., 75 (1999) 172.
[18] A.B. Chwang, R.C. Kwong, J.J. Brown, Appl. Phys. Lett., 80 (2002) 725.
[19] D. Ma, C.S. Lee, S.T. Lee, L.S. Hung, Appl. Phys. Lett., 80 (2002) 3641.
[20] Y. Shao, Y. Yang, Appl. Phys. Lett., 83 (2003) 2453.
[21] S.H. Kim, J. Jang, K.S. Yook, J.Y. Lee, M.-S. Gong, S. Ryu, G.-k. Chang, H.J.
Chang, J. Appl. Phys., 103 (2008) 054502/1.
[22] M.E. Kondakova, T.D. Pawlik, R.H. Young, D.J. Giesen, D.Y. Kondakov, C.T.
Brown, J.C. Deaton, J.R. Lenhard, K.P. Klubek, J. Appl. Phys., 104 (2008)
094501.
[23] J. Lee, J.-I. Lee, J.Y. Lee, H.Y. Chu, Org. Electron., 10 (2009) 1529.
[24] K.H. Kim, J.Y. Lee, T.J. Park, W.S. Jeon, G.P. Kennedy, J.H. Kwon, Synth. Met.,
160 (2010) 631.
[25] D.P.-K. Tsang, M.-Y. Chan, A.Y.-Y. Tam, V.W.-W. Yam, Org. Electron., 12 (2011)
1114.
[26] C.W. Seo, J.Y. Lee, Thin Solid Films, 520 (2012) 7022.
[27] S. Lee, C.W. Tang, L.J. Rothberg, Appl. Phys. Lett., 101 (2012) 043303.
[28] V. Bulovic, A. Shoustikov, M.A. Baldo, E. Bose, V.G. Kozlov, M.E. Thompson,
S.R. Forrest, Chem. Phys. Lett., 287 (1998) 455.
83
[29] V. Bulovic, R. Deshpande, M.E. Thompson, S.R. Forrest, Chem. Phys. Lett., 308
(1999) 317.
[30] Y. Kawamura, J. Brooks, J.J. Brown, H. Sasabe, C. Adachi, Phys. Rev. Lett., 96
(2006) 017404.
[31] G. Ramos-Ortiz, Y. Oki, B. Domercq, B. Kippelen, Phys. Chem. Chem. Phys., 4
(2002) 4109.
[32] M.A. Baldo, M.E. Thompson, S.R. Forrest, Nature (London), 403 (2000) 750.
[33] V.V. Jarikov, K.P. Klubek, L.-S. Liao, C.T. Brown, J. Appl. Phys., 104 (2008)
074914.
[34] C.-J. Lee, D.-Y. Yoon, D.-G. Moon, D.-K. Choi, J.-I. Han, ECS Trans., 16 (2009)
17.
[35] N.H. Kim, Y.-H. Kim, J.-A. Yoon, S.Y. Lee, H.H. Yu, A. Turak, W.Y. Kim, MRS
Online Proc. Libr., 1567 (2013) opl 2013 734.
[36] C.-H. Chen, W.-S. Huang, M.-Y. Lai, W.-C. Tsao, J.T. Lin, Y.-H. Wu, T.-H. Ke,
L.-Y. Chen, C.-C. Wu, Adv. Funct. Mater., 19 (2009) 2661.
[37] S.-y. Takizawa, V.A. Montes, P. Anzenbacher, Jr., Chem. Mater., 21 (2009) 2452.
[38] Y.-M. Chen, W.-Y. Hung, H.-W. You, A. Chaskar, H.-C. Ting, H.-F. Chen, K.-T.
Wong, Y.-H. Liu, J. Mater. Chem., 21 (2011) 14971.
[39] S. Gong, Y. Chen, J. Luo, C. Yang, C. Zhong, J. Qin, D. Ma, Adv. Funct. Mater.,
21 (2011) 1168.
[40] W. Jiang, L. Duan, J. Qiao, G. Dong, L. Wang, Y. Qiu, Org. Lett., 13 (2011) 3146.
[41] C. Fan, F. Zhao, P. Gan, S. Yang, T. Liu, C. Zhong, D. Ma, J. Qin, C. Yang, Chem.
84
- Eur. J., 18 (2012) 5510.
[42] K. Brunner, A. Van Dijken, H. Boerner, J.J.A.M. Bastiaansen, N.M.M. Kiggen,
B.M.W. Langeveld, J. Am. Chem. Soc., 126 (2004) 6035.
[43] P.-I. Shih, C.-L. Chiang, A.K. Dixit, C.-K. Chen, M.-C. Yuan, R.-Y. Lee, C.-T.
Chen, E.W.-G. Diau, C.-F. Shu, Org. Lett., 8 (2006) 2799.
[44] L.-S. Cui, S.-C. Dong, Y. Liu, Q. Li, Z.-Q. Jiang, L.-S. Liao, J. Mater. Chem. C, 1
(2013) 3967.
[45] T. Yamada, F. Suzuki, A. Goto, T. Sato, K. Tanaka, H. Kaji, Org. Electron., 12
(2011) 169.
[46] T. Zhang, Y. Liang, J. Cheng, J. Li, J. Mater. Chem. C, 1 (2013) 757.
[47] D.B. Knowles, C. Lin, P.B. Mackenzie, J.-Y. Tsai, R.W. Walters, S.A. Beers, C.S.
Brown, W.H. Yeager, U.S. Patent No. 7,915,415 (2011).
[48] Y.-S. Park, K.-H. Kim, J.-J. Kim, Appl. Phys. Lett., 102 (2013) 153306.
[49] Y.-S. Park, S. Lee, K.-H. Kim, S.-Y. Kim, J.-H. Lee, J.-J. Kim, Adv. Funct. Mater.,
23 (2013) 4914.
[50] R.H. Young, J.R. Lenhard, D.Y. Kondakov, T.K. Hatwar, SID Int. Symp. Digest
Tech. Papers, 39 (2008) 705.
[51] J. Lee, J.-I. Lee, H.Y. Chu, Synth. Met., 159 (2009) 991.
[52] J. Lee, J.-I. Lee, K.-I. Song, S.J. Lee, H.Y. Chu, Appl. Phys. Lett., 92 (2008).
[53] J. Lee, J.-I. Lee, J.Y. Lee, H.Y. Chu, Synth. Met., 159 (2009) 1956.
[54] P. Schroegel, N. Langer, C. Schildknecht, G. Wagenblast, C. Lennartz, P.
Strohriegl, Org. Electron., 12 (2011) 2047.
85
[55] S.-J. Su, T. Chiba, T. Takeda, J. Kido, Adv. Mater., 20 (2008) 2125.
86
CHAPTER 4
CHARGE TRANSPORT THROUGH FIRPIC
4.1 INTRODUCTION
As discussed in Chapter 1 and Chapter 3, FIrpic is one of the most widely used Ir
cyclometalated dopants used for testing new blue emitting PHOLED device architectures
and materials. High external quantum efficiencies (EQE) exceeding 20% can be achieved
with this material used as an emitter doped into a host matrix. However, regardless of the
choice of transport and host materials, use of this material as an emitter results in OLED
devices that have lifetimes well below 50 hours, which is a lifetime far removed from
what is required for display and solid-state lighting applications.
A recent study [3] showed that FIrpic undergoes irreversible chemical dissociation in
an operating device, producing fragments that can quench excitons and reduce the OLED
efficiency. Several possible pathways for FIrpic degradation have been proposed,
suggesting mainly the involvement of FIrpic excited states formed as a result of electron-
hole recombination in the dopant/host matrix. In this chapter, the effect of charge
transport in a FIrpic doped mCBP matrix on the stability of the OLED device is
investigated. We conclude that hole transport on FIrpic and charge recombination on
mCBP both lead to rapid device degradation and derive insight into materials selection
criteria.
87
4.2 EXPERIMENTAL
Molecular structures along with highest occupied molecular orbital (HOMO) and
lowest unoccupied molecular orbital (LUMO) energy levels for all materials in this
chapter are shown in Figure 4.1. Additional information for these materials are shown in
Table 2.1 in Chapter 2. Device fabrication procedures were discussed in Chapter 2. A 1.0
nm thin film of MoOx was deposited on ITO as a hole-injecting layer (HIL). Neat NPB,
FIrpic doped NPB, neat mCBP and FIrpic doped mCBP were used as the hole transport
layers (HTL). Alq served as the electron transport layer (ETL) as well as the light emitter.
Figure 4.1: Molecular structures and HOMO/LUMO energy levels of materials used in devices.
Units in electronvolts (eV).
2.1 2.1
5.3
NPB
+
FIrpic
5.6
ITO
4.7
LiF/Al
4.1
MoOx
2.5
Alq
5.65.7
2.5
mCBP
+
FIrpic
2.5
AlqNPB
FIrpic
mCBP
88
4.3 RESULTS AND DISCUSSION
The layer structures of a series of OLEDs designed to probe the effect of HTL
formulation on OLED performance are listed in Table 4.1. All devices have in common
an HTL/ETL device structure with Alq as the ETL and emitter. The thicknesses of the
HTL and ETL are 75 nm each.
Table 4.1: OLED device structure and initial EQE and voltage at 20 mA/cm2. ITO/1 nm
MoOx/HTL/75 nm Alq/1 nm LiF/100 nm Al.
Device HTL: Voltage
(V)
EQE
(%)
Ref. 75 nm NPB 8.3 1.2
A1 75 nm mCBP 8.3 1.5
A2 37.5 nm mCBP|37.5 nm NPB 7.5 1.1
B1 37.5 nm mCBP+1% FIrpic|37.5 nm NPB 9.0 1.1
B2 37.5 nm mCBP+3% FIrpic|37.5 nm NPB 9.3 1.0
B3 37.5 nm mCBP+6% FIrpic|37.5 nm NPB 9.8 1.1
B4 37.5 nm mCBP+12% FIrpic|37.5 nm NPB 9.2 1.3
C1 37.5 nm mCBP|27.5 nm mCBP+12% FIrpic|10 nm NPB 8.4 1.3
D1 37.5 nm mCBP|37.5 nm mCBP+12% FIrpic 7.9 1.5
E1 75 nm NPB+6% FIrpic 7.9 1.1
Electroluminescence originates from Alq for all devices (Figure 4.2). Alq has an
EL maximum around 530 nm. Each of the HTL materials used in this study emits at a
shorter wavelength. The EL maximum is 470 nm for NPB [6] and ~468-472 nm for
FIrpic [7, 8]. mCBP has a PL maximum at 340 nm in solution [9]. It is clear from Figure
4.2 that the EL spectra are solely characteristic of Alq without any contribution from the
HTL materials. While it is generally possible for electrons to inject into the HTL from the
emitter layer in OLED devices, the devices in this study have a large energy barrier of 0.4
eV for electron injection from Alq into either mCBP or NPB (Figure 4.1). The electron
concentration in the HTL should be small, if any. If electrons were to penetrate deep into
89
the HTL, then there would be emission from the HTL material. However, as already
noted, there is no such emission. Recombination at the HTL/Alq interface yields emission
only from Alq due to its smaller energy gap.
Figure 4.2: EL spectra at 20 mA/cm2 for the following device architecture: ITO/1.0 nm MoOx/75
nm HTL/75 nm Alq/1.0 nm LiF/100 nm Al, where the HTL’s are: Ref [75 nm NPB], A1 [75 nm
mCBP], B4 [37.5 nm mCBP+12% FIrpic|37.5 nm NPB], C1 [37.5 nm mCBP|27.5 nm mCBP+12%
FIrpic|10 nm NPB], D1 [37.5 nm mCBP|37.5 nm mCBP+12% FIrpic], E1 [75 nm NPB+6%
FIrpic].
4.3.1 Undoped hole transport layers
The HTL of the reference (Ref), A1 and A2 devices are NPB, mCBP, and
mCBP/NPB, respectively. The Ref and A1 have a single-layer HTL sandwiched between
Alq and the ITO/MoOx anode. A2 has a bi-layer HTL with NPB adjacent to Alq and
mCBP in contact with MoOx. As shown in Figure 4.3(a), the J-V-L characteristics are
very similar for all three devices, independent of the HTL formulation. However, as
shown in Figure 4.3(b), the EQE decay (lifetime) is quite different with A1 degrading
much faster than Ref and A2. The t50 (time to 50% decay in EQE) for A1 is only about 90
hours, whereas the t80 (time to 80% decay in EQE) for Ref and A2 is greater than 200
0
0.2
0.4
0.6
0.8
1
1.2
350 450 550 650 750
Rel
ati
ve
Inte
nsi
ty (a.u
.)
Wavelength (nm)
Ref.
A1
B4
C1
D1
E1
90
Figure 4.3: OLED performance for devices with an undoped HTL adjacent to MoOx.
ITO/MoOx/HTL/Alq/LiF/Al. Ref[NPB], A1[mCBP], A2[mCBP/NPB]. (a) Initial current density-
voltage-luminance (J-V-L) characteristics, (b) Device lifetime at 20 mA/cm2, (c) Operational
voltage rise at 20 mA/cm2.
hours. These lifetime results support previous findings by Kondakov et al. [10, 11] that
electron-hole recombination at the HTL/Alq interface produces HTL excited states which
can lead to the dissociation of the HTL molecules. The probability for dissociation is
1.E-02
1.E-01
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
0
20
40
60
80
100
120
2.0 6.0 10.0 14.0 18.0
Lu
min
ance
(cd
/m2 )
Cu
rren
t D
en
sity
(m
A/c
m2 )
Voltage (V)
Ref.
A1
A2
(a)
0.5
0.6
0.7
0.8
0.9
1.0
1.1
0 50 100 150 200 250 300
EQ
E (
a.u
.)
Time (hrs.)
Ref.
A1
A2
(b)
4
6
8
10
12
14
0 50 100 150 200 250 300
Vo
ltag
e (V
)
Time (hrs.)
Ref.
A1
A2
(c)
91
higher if the excited state energy is larger than the bond dissociation energy, which is the
case for carbazole-based HTL’s such as mCBP, where C-N or C-C homolytic bond
cleavage can occur. As shown in Figure 4.3(c), there is essentially no difference between
A1 and A2 in voltage rise, which is in distinct contrast to the vast difference in EQE
decay. This indicates that mCBP and NPB are equally stable with respect to hole
transport, but mCBP is unstable with respect to charge recombination at the mCBP/Alq
interface. Evidently, luminescence quenchers are more likely to form at the mCBP/Alq
interface.
4.3.2 FIrpic doped into mCBP adjacent to the anode
The next set of devices, B1–B4 of Table 1, were designed for the purpose of
evaluating the stability of FIrpic with respect to hole transport only. Similar to A2, the
HTL of B1–B4 is a bi-layer mCBP/NPB stack. Unlike A2, which has an undoped mCBP
layer, the mCBP layer of B1–B4 is doped with FIrpic ranging in concentration from 1 to
12%. Since this doped layer is spaced away from the Alq emitter layer by a layer of NPB,
it presumes to have little effect on charge recombination at the NPB/Alq interface. As
shown in Figure 4.4(a) and Table 1, the FIrpic doped devices generally require a higher
drive voltage (9.0–9.8 V compared to 7.9 V for the undoped device A2). This increase in
drive voltage may indicate: 1) reduced hole injection efficiency from the MoOx in the
presence of FIrpic, and/or 2) increased hole trapping in the Flrpic doped mCPB layer. It
has been reported that hole injection efficiency from ITO/MoOx into the HTL can be
92
Figure 4.4: OLED performance for devices with FIrpic doped mCBP adjacent to MoOx.
ITO/MoOx/mCBP+x% FIrpic/Alq/LiF/Al. B1[1%], B2[3%], B3[6%], B4[12%]. (a) Initial
current density-voltage-luminance (J-V-L) characteristics, (b) Device lifetime at 20 mA/cm2, (c)
Operational voltage rise at 20 mA/cm2.
affected by electron transfer from the HTL to MoOx [12]. mCBP and FIrpic
coincidentally have a similar HOMO of 5.6 eV, and thus FIrpic can accept direct hole
injection from MoOx and consequently can affect the hole injection efficiency as well as
1.E-02
1.E-01
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
0
20
40
60
80
100
120
2.0 6.0 10.0 14.0 18.0
Lu
min
ance
(cd
/m2 )
Cu
rren
t Den
sity
(m
A/c
m2 )
Voltage (V)
Ref.
B1
B2
B3
B4
(a)
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
0 25 50 75 100 125 150
EQ
E (
a.u
.)
Time (hrs.)
Ref.
B1
B2
B3
B4
(b)
5.0
7.0
9.0
11.0
13.0
15.0
17.0
19.0
0 25 50 75 100 125 150
Vo
ltag
e (V
)
Time (hrs.)
Ref.
B1
B2
B3
B4
(c)
93
the hole transport in mCPB.
As shown in Figure 4.4(b) the EQE degradation is only modestly dependent on
the concentration of FIrpic in mCBP. At low FIrpic concentrations, devices B1 (1%), B2
(3%) and B3 (6%) show a monotonic decrease in EQE at a rate similar to the Ref device,
indicating that the FIpic doped mCBP layer, which is remote from the NPB/Alq interface,
has little effect on EQE. With a higher FIrpic concentration, device B4 (12%) exhibits a
slight rise in EQE before undergoing a steeper decay compared to devices B1–B3. In any
case, the degradation in EQE is relatively minor compared to the large rise in drive
voltage. Figure 4.4(c) shows that the drive voltage increases for all devices at a rate that
increases with FIrpic concentration. The voltage increase (after 200 h of operation at 20
mA/cm2) is as high as 88% for B4 with 12% FIrpic compared to 6.7% for the Ref device
without FIrpic, and 12%, 16.7% and 23.1% for B1, B2, and B3, respectively. This
suggests that FIrpic radical cations are unstable in the mCBP matrix and deep hole traps
are formed as a result of current stress. Since FIrpic can effectively compete with mCBP
for hole transport, more hole traps are likely to form in the HTL with a higher
concentration of FIrpic, which in turn will give rise to a larger increase in the drive
voltage in the OLED devices. On the other hand, the change in EQE reflects the
degradation at the NPB/Alq interface where electron-hole recombination is confined and
should be independent of the degradation in the FIrpic doped mCBP layer, which is
remote from the NPB/Alq interface. This is the case for B1–B3 devices over the course of
200 h. In the B4 device, the initial rise in EQE with drive voltage may be attributed to a
reduction of hole accumulation at the NPB/Alq interface which affects favorably the
94
charge balance factor and thus a reduction in quenching by NPB radical cations [13]. The
subsequent steep decrease in EQE coincides with a faster rise in drive voltage, indicating
that hole traps are formed at a faster rate in the FIrpic doped mCBP layer, which in turn
could cause electric field induced degradation or even electron-hole recombination in this
layer with decreased emission efficiency.
4.3.3 FIrpic doped mCBP sandwiched between undoped hole transport materials
In order to further understand the cause of the voltage rise, we fabricated device
C1 where the HTL consists of a tri-layer with a 12% FIrpic doped mCBP layer (27.5nm)
sandwiched between an undoped mCBP layer (37.5 nm) and an NPB layer (10 nm). As
shown in Figure 4.5(a), the J-V-L characteristics of C1 are similar to the Ref. device. The
fact that there is little voltage shift provides further support that the presence of FIrpic
doped mCBP near MoOx in the HTL affects hole injection as observed in the B1–4 series
of devices. Figure 4.5(b) shows that the EQE for C1 increases modestly to ~1.05 within
the first 130 hours then undergoes a steep decrease. This behavior is similar to B4 which
also has a high concentration of FIrpic (12%) in mCBP.
Also shown in Figure 4.5(a) and Figure 4.5(b) are the J-V-L characteristics and
EQE decay, respectively, for device D1 where, unlike C1, the FIrpic doped mCBP layer is
in direct contact with Alq. Whereas the J-V-L characteristics of C1 and D1 are similar,
their EQE decay behaviors are dramatically different. D1 suffers a severe loss of EQE
and reaches t50 at 48 h as compared to a 10% increase in EQE for C1 up to 115 h, after
which EQE starts to decrease. Figure 4.5(c) shows that the voltage rise during operation
95
Figure 4.5: OLED performance for devices with FIrpic doped mCBP spaced away from MoOx and
for a device with FIrpic doped in NPB. ITO/MoOx/mCBP/HTL/NPB/Alq/LiF/Al:
C1[mCBP+12%], ITO/MoOx/HTL/Alq/LiF/Al]: D1[mCBP/mCBP+12% FIrpic] and
E1[NPB+6% FIrpic]. (a) Initial current density-voltage-luminance (J-V-L) characteristics, (b)
Device lifetime at 20 mA/cm2, (c) Operational voltage rise at 20 mA/cm2.
correlates with the EQE decay for both C1 and D1. The voltage for D1 increases by 90%
at the t50 for EQE decay. The voltage change in C1 is fairly modest during the first 115
hours before undergoing a steep increase that correlates with the EQE loss, similar to
1.E-02
1.E-01
1.E+00
1.E+01
1.E+02
1.E+03
1.E+04
0
20
40
60
80
100
120
2.0 6.0 10.0 14.0 18.0
Lu
min
ance
(cd
/m2 )
Cu
rren
t Den
sity
(m
A/c
m2 )
Voltage (V)
Ref.
C1
D1
E1
(a)
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
0 50 100 150 200 250 300
EQ
E (
a.u
.)
Time (hrs.)
Ref.
C1
D1
E1
(b)
5.0
7.0
9.0
11.0
13.0
15.0
17.0
19.0
0 50 100 150 200 250 300
Vo
ltage (
V)
Time (hrs.)
Ref.
C1
D1
E1
(c)
96
what was observed for B4. It is also noted that the initial drive voltages for both C1 and
D1 are lower compared to B1–B4. This is due to the fact the FIpic doped layer in these
two devices is not in contact with MoOx, providing further evidence that direct contact
between MoOx and FIrpic doped mCBP impedes hole injection.
For device D1, because the FIrpic doped mCBP layer is in direct contact with Alq,
it is likely that recombination at this interface involves FIrpic, resulting in the formation
of quencher species originating from FIrpic radical cations. This would lead to a loss in
EQE and a rise in drive voltage that is more severe compared to C1 where FIrpic is
spaced away from the recombination interface by a layer of NPB and thus not involved in
direct recombination. The fact that the t50 for D1 is even less than that of A1 which has a
neat mCBP layer as the HTL also indicates that recombination directly on FIrpic is a
channel for EQE degradation.
4.3.4 FIrpic doped into NPB
To further validate that FIrpic is unstable with respect to hole transport, device E1
was fabricated where the HTL is a layer of NPB doped with 6% Flrpic. Figure 4.5(a)
shows that the J-V-L of E1 is comparable to the Ref device. Comparing the EQE losses in
D1 and E1 in Figure 4.5(b), it can be seen that E1 is far more stable and closely
resembles the Ref device. Similarly the drive voltage rise for E1, as shown in Figure
4.5(c), is relatively modest with very little difference between E1 and the Ref device. As
previously mentioned, FIrpic and mCBP have similar HOMO’s, so they both can
contribute to hole transport in an HTL containing FIrpic doped in mCBP. In contrast, in
an HTL containing Flrpic doped NPB, holes can only transport in NPB because FIrpic,
97
with a much deeper HOMO than NPB, should be incapable of transporting or trapping
holes. In fact, this also explains why the initial drive voltage for E1 is low compared to
B1–B4 devices even though the FIrpic doped NPB layer is in direct contact with MoOx.
FIrpic is precluded to compete with NBP for holes injected from MoOx because of its
deeper HOMO.
4.4 CONCLUSIONS
In summary, we have shown that hole transport and electron-hole recombination
involving FIrpic both result in OLED degradation. Using FIrpic doped mCBP as an HTL
in conjunction with Alq as the emitter/ETL layer, we found that hole transport in FIrpic
primarily results in voltage rise caused by hole trap formation in the HTL. In contrast,
electron-hole recombination in FIrpic results in a large EQE loss as well as significant
voltage rise. We attribute the loss in device performance to the instability of FIrpic radical
cations. One possible consequence of FIrpic oxidation is reportedly the breaking of the Ir-
O bond, followed by loss of CO2 and formation of FIrpic+ [3]. This product would
contribute to the voltage rise we observed in our devices. The detailed chemical pathway
for FIrpic degradation requires further investigation. These results indicate that
FIrpic/mCBP is unsuitable for practical use as an emitter in OLED devices. A possible
solution is to avoid Flrpic radical cation formation by pairing Flrpic with a host with
which emission is through energy transfer from the host to Flrpic rather than through
hole-trapping or transporting in Flrpic.
98
REFERENCES
[1] R.J. Holmes, S.R. Forrest, Y.J. Tung, R.C. Kwong, J.J. Brown, S. Garon, M.E.
Thompson, Appl. Phys. Lett., 82 (2003) 2422.
[2] J.J. Brooks, R.C. Kwong, Y.-J. Tung, M.S. Weaver, B.W. D'Andrade, V.
Adamovich, M.E. Thompson, S.R. Forrest, J.J. Brown, Proc. SPIE, 5519 (2004)
35.
[3] I. Rabelo de Moraes, S. Scholz, B. Luessem, K. Leo, Org. Electron., 12 (2011)
341.
[4] M.E. Kondakova, T.D. Pawlik, R.H. Young, D.J. Giesen, D.Y. Kondakov, C.T.
Brown, J.C. Deaton, J.R. Lenhard, K.P. Klubek, J. Appl. Phys., 104 (2008)
094501.
[5] S. Gong, X. He, Y. Chen, Z. Jiang, C. Zhong, D. Ma, J. Qin, C. Yang, J. Mater.
Chem., 22 (2012) 2894.
[6] S.C. Tse, K.C. Kwok, S.K. So, Appl. Phys. Lett., 89 (2006) 262102.
[7] C.W. Lee, K.S. Yook, J.Y. Lee, Org. Electron., 14 (2013) 1009.
[8] J. Zhuang, W. Li, W. Su, Y. Liu, Q. Shen, L. Liao, M. Zhou, Org. Electron., 14
(2013) 2596.
[9] P. Schroegel, N. Langer, C. Schildknecht, G. Wagenblast, C. Lennartz, P.
Strohriegl, Org. Electron., 12 (2011) 2047.
[10] D.Y. Kondakov, J. Appl. Phys., 104 (2008) 084520.
[11] D.Y. Kondakov, T.D. Pawlik, W.F. Nichols, W.C. Lenhart, J. Soc. Inf. Disp., 16
(2008) 37.
99
[12] T. Matsushima, Y. Kinoshita, H. Murata, Appl. Phys. Lett., 91 (2007) 253504.
[13] R.H. Young, J.R. Lenhard, D.Y. Kondakov, T.K. Hatwar, SID Int. Symp. Digest
Tech. Papers, 39 (2008) 705.
100
CHAPTER 5
INVESTIGATING BLUE PHOSPHORESCENT IRIDIUM
DOPANT WITH PHENYL-IMIDAZOLE LIGANDS
5.1 INTRODUCTION
As discussed in Chapters 1 and 3, excellent device lifetime (>100,000 h) has been
achieved for red and green PHOLEDs while the lifetime for blue PHOLEDs remains
relatively short (typically less than 10,000 h) [1] and inadequate for most applications,
such as display and lighting. Therefore, considerable research efforts have focused on
developing efficient and stable blue phosphorescent PHOLEDs [2]. However, there are
few reports on the structure-property relationships for phosphorescent molecules as they
pertain to device lifetime. FIrpic is among the most studied phosphorescent molecules.
As discussed in Chapter 3, when FIrpic is used as a dopant in blue PHOLEDs, high
external quantum efficiencies (EQE) can be achieved and there are also reports that
EQE’s greater than 20% have been achieved [3-8]. However, FIrpic doped PHOLEDs are
very unstable, with reported lifetimes in the range of 0.1-110 hours depending on the
device architecture and test conditions [9-12]. Based on lifetime testing at a current
density of 5 mA/cm2 that was outlined in Chapter 3, we have observed FIrpic-based
devices having lifetimes ranging from minutes up to 12 hours depending on the choice of
host and transport materials. The instability of FIrpic doped PHOLEDs has been largely
101
attributed to the intrinsic instability of Flrpic molecules upon excitation. Both the
ancillary picolinate ligand and the fluorine substituted phenyl-pyridyl ligands are
susceptible to photo-induced dissociation [11, 13, 14]. As discussed in Chapter 4, FIrpic
is also unstable to hole transport [15]. To avoid the detrimental structural features present
in FIrpic, an Ir complex with phenyl-imidizole ligands has been examined, one which
reportedly can provide PHOLEDs with lifetimes as long as 10,000 hours [16-22].
In this chapter, a detailed study of Ir(iprpmi)3 is performed. The results presented
here will show that its performance as a blue phosphorescent dopant is not only related to
its intrinsic photophysical properties, but also highly dependent on the compositions of
the host matrix and the adjacent transport layers.
5.2 MATERIAL AND METHODS
Figure 5.1 shows the molecular structures and the energy level diagram for all
materials used in this work. Additional details regarding these materials were discussed in
Chapter 2. The device fabrication procedures were also outlined in Chapter 2.
UV-Vis absorption spectra were obtained using a Perkin Elmer Lambda 750
spectrophotometer. PL and phosphorescent spectra were recorded on a Hitachi F-4600
fluorescence spectrophotometer. The quantum efficiency was obtained by using fac-
Ir(ppy)3 as a standard [23]. The transient spectrum was acquired with an Edinburgh
FLS920 spectrometer using the time-correlated single-photon counting option and data
was analyzed by F900 software (Edinburgh Instruments). The transient spectrum and
quantum efficiency were conducted at room temperature in degassed dichloromethane.
Degassed solutions were prepared by purging with argon for 30 minutes. Cyclic
102
voltammetry (CV) was carried out on a CHI600 voltammetric analyzer at room
temperature with a conventional three-electrode configuration consisting of a platinum
disk working electrode, a platinum wire auxiliary electrode, and an Ag wire pseudo-
reference electrode with ferrocenium-ferrocene (Fc+/Fc) as the internal standard.
Deaerated dichloromethane was used as solvent. UPS analyses were carried out with an
unfiltered HeI (21.2 eV) gas discharge lamp and a hemispherical analyzer. DFT
calculations were performed at the B3LYP level. The 6-31g(d) basis set was employed for
H, C, N atoms and a “double-ζ” quality basis set, LANL2DZ, was employed for the
Ir(III) metal atom. All calculations were carried out using Gaussian 03 [24].
Figure 5.1 Molecular structures and energy level diagram of the materials (values in eV).
Hole Transport Materials Electron Transport Materials
LEL Host LEL Dopant
TmPyPB
6.7
2.8
2.3
5.4
mCBP
+
dopant
NPB
2.4
6.0
ITO/MoOx
LiF/Al2.9
BAlq
5.9
2.2
4.8
TAPC
5.5
2.0
NPB TAPC TmPyPB BAlq
Ir(iprpmi)3mCBP
103
5.3 RESULTS AND DISCUSSION
5.3.1 Photophysical and electrochemical measurements
Figure 5.2 shows the UV-Vis absorption and photoluminescence (PL) spectra of
Ir(iprpmi)3. The absorption bands below 330 nm are due to the spin-allowed ligand-
centered (LC) 1(-
*) transitions of the phenyl-imidizole moiety [18, 25-27]. The band at
359 nm is attributed to a spin-allowed metal-to-ligand charge transfer (MLCT) transition
[27] while bands longer than 359 nm are due to mixtures of spin-forbidden LC and
MLCT transitions. The PL spectrum has maximum peak intensity at 474 nm which is
similar to what has been reported for FIrpic (470-472 nm) [18, 28-30]. There is a
Figure 5.2: UV-Vis absorption and PL spectra of Ir(iprpmi)3 in CH2Cl2. Inset: PL spectrum of
Ir(iprpmi)3 in 2-MeTHF at 77 K.
secondary peak at 504 nm and a shoulder at ~550 nm. The solution quantum efficiency
for Ir(iprpmi)3 in CH2Cl2 is 0.57. The triplet energy (ET) is 2.66 eV as determined by the
0
0.2
0.4
0.6
0.8
1
1.2
0
0.1
0.2
0.3
0.4
0.5
250 350 450 550 650 750 850
Rel
ativ
e P
L I
nte
nsi
ty (
a.u
.)
Ex
tin
ctio
n C
oef
fici
ent (1
05M
-1cm
-1)
Wavelength (nm)
Absorption
Photoluminescence
0
0.2
0.4
0.6
0.8
1
1.2
400 500 600
Rela
tiv
e P
L I
nte
nsi
ty (a
.u.)
Wavelength (nm)
104
peak intensity of the shortest wavelength peak in the PL spectrum at 77 K (inset of Figure
5.2). The transient PL taken in degassed CH2Cl2 indicates a lifetime of 1.19 s
(Appendix, Figure A1). The highest occupied molecular orbital (HOMO) is 4.8 eV based
on ultraviolet photoelectron spectroscopy (UPS) while the lowest unoccupied molecular
orbital (LUMO) is 2.2 eV considering a HOMO-LUMO gap of 2.6 eV as determined by
the 0-0 absorption band (469 nm). Cyclic voltammetry (CV) indicates that Ir(iprpmi)3
undergoes a reversible one-electron oxidation process with an onset of -0.15 V
(Appendix, Figure A2). The HOMO based on CV is 4.7 eV as calculated according to
E E Eonseto wherein EFC is the HOMO level of ferrocene (4.8 eV).
Figure 5.3 depicts the HOMO and LUMO obtained from density functional theory
(DFT) calculations. The HOMO is localized in the Ir d-orbital and the -orbital of the
phenyl moiety on the ligand whereas the LUMO is distributed on the phenyl-imidizole
ligand. These electron densities are similar to what has been reported for other
phenylimidazolinato Ir (III) complexes [27]. The calculated HOMO level is 4.3 eV.
Figure 5.3: Simulated frontier molecular orbitals of Ir(iprpmi)3. (a) HOMO electron density, (b)
ball and stick model, (c) LUMO electron density.
(a) (b) (c)
105
5.3.2 Low EQE OLEDs
Ir(iprpmi)3 as a blue phosphorescent dopant in PHOLEDs was first reported in the
patent literature by Lin et al.[19]. They showed that Ir(iprpmi)3 and similar dopants with
an imidazole as the ligand generally provide a longer lifetime than other blue
phosphorescent dopants, such as lrpic. ollowing Lin’s work, we e amined in more
detail not only the effect of the dopant and host composition of the emitter layer, but also
the composition of the adjacent transport layers on the device performance. Two devices
were fabricated with a layer structure as follows: Anode/40 nm NPB/30 nm mCBP+x%
Ir(iprpmi)3/40 nm BAlq/Cathode. In this structure, which is similar to that used by Lin,
NPB and BAlq are the hole-transport and electron-transport layers, respectively, and
mCBP is the host matrix for the Ir(iprpmi)3 in the emitter layer. The concentration of
Ir(iprpmi)3 in mCBP is 6% and 12%. Figure 5.4 compares the electroluminescence
characteristics of these two devices. The EL spectra as shown in Figure 5.4(a) are
characteristic of Ir(iprpmi)3 emission and relatively independent of concentration. The
current-voltage curves of Figure 5.4(b) indicates a shift to lower drive voltage with
increasing Ir(iprpmi)3 concentration in the emitter layer, by about 1.3 V at 5 mA/cm2. As
shown in Figure 5.4(c), the EQE dependence on drive current density is different for the
two devices. Although both devices reach an EQE maximum of about 14%, they achieve
their maxima at different current densities. At 5 mA/cm2, the EQE is 8.4% for the device
with 6% Ir(iprpmi)3 compared to 13.7% for the device with 12% Ir(iprpmi)3. Similar to
what has been reported by Lin et al, the maximum EQE values (~14%) are substantially
lower than what can be expected from similar Ir-based dopants. The triplet energy levels
106
Figure 5.4: OLED performance data for 6% and 12% concentrations of Ir(iprpmi)3 with NPB as
HTL and BAlq as ETL. Anode/40 nm TAPC/30 nm mCBP+ 6% or 12% Ir(iprpmi)3/40 nm
BAlq/Cathode. (a) EL spectra at 5 mA/cm2, (b) current density vs. voltage, (c) EQE vs. current
density.
of NPB and BAlq are 2.41eV and 2.18 eV, respectively, which are lower than that of
Ir(iprpmi)3 (2.66 eV). Thus both NPB and BAlq can cause quenching of the Ir(iprpmi)3
triplet excitons generated by electron-hole recombination in the emitter layer and the
extent of quenching will depend on the proximity of the recombination zone relative to
the interfaces with NPB and BAlq. The effect of shifting the recombination zone on
device performance is explored further in the following sections.
0.0
0.2
0.4
0.6
0.8
1.0
1.2
400 500 600 700R
ela
tiv
e I
nte
nsi
tyWavelength (nm)
6%
12%
0.1
1
10
100
0 4 8 12 16
Cu
rren
t D
en
sity
(m
A/c
m2)
Voltage (V)
6%
12%
0
4
8
12
16
0.1 1 10 100
EQ
E (
%)
Current Density (mA/cm2)
6%
12%
(a)
(b)(c)
Ir(iprpmi)3
Ir(iprpmi)3
Ir(iprpmi)3
107
5.3.3 High EQE OLEDs
In the layer structure described above, the Ir(iprpmi)3 doped emitter layer was
sandwiched between the hole-transport layer NPB and the electron-transport layer BAlq.
To examine exciton quenching by the hole-transport layer, we fabricated devices similar
to those described above, except that NPB is substituted by TAPC for the hole-transport
layer. TAPC was chosen to eliminate the possibility of exciton quenching since the triplet
energy of TAPC (2.96 eV) is significantly higher than that of Ir(iprpmi)3 (2.66 eV). In
this series, the Ir(iprpmi)3 concentration range was broadened to cover from 0%
(undoped) to 24%.
The EL spectrum for the undoped device is shown in Figure 5.5(a). The spectrum
is invariant with current density and shows a broad peak centered at 484 nm, which can
be attributed to BAlq [31]. A weak shoulder appears around 400 nm, which can be due to
emissions from the host mCBP [32, 33], or the hole-transport layer TAPC [34], or both.
Qualitatively, the EL spectrum indicates that electron-hole recombination occurs
throughout the emitter layer and at both TAPC and BAlq interfaces adjacent to the
emitting layer. The relative intensities of the broad peak and the weak shoulder may
indicate that the location of the recombination zone is at or near the BAlq interface.
However, this is not necessarily the case as shown in experiments involving doped
devices where the recombination zone can be better located. In any case, the EQE of the
undoped device is very low compared with the Ir(iprpmi)3 doped devices.
The EL spectra for Ir(iprpmi)3 doped devices are also shown in Figure 5.5(a).
Regardless of the dopant concentration, the EL spectra are similar and characteristic of
108
Figure 5.5: OLED performance data for Ir(iprpmi)3 concentration series with TAPC HTL and
BAlq ETL. Anode/40 nm TAPC/30 nm mCBP+ x% Ir(iprpmi)3/40 nm BAlq/Cathode. x=0%–24%.
(a) EL spectra at 5 mA/cm2, (b) current density vs. voltage, (c) EQE vs. current density.
Ir(iprpmi)3 emission. The corresponding CIEx,y coordinates are 0.17, 0.37.
The J-V plots in Figure 5.5(b) show that the drive voltage is dependent on the
Ir(iprpmi)3 dopant concentration in the emitting layer. At low concentrations (3% and
7%), the drive voltage (at 5 mA/cm2) is 10.9 V, which is as much as 2 volts higher than
the undoped device (9.0 V) at the same current density. With higher concentrations, the
drive voltage is reduced. For the device with 24% Ir(iprpmi)3, the drive voltage is about 1
volt less than the undoped cell. These J-V characteristics indicate that Ir(iprpmi)3 may act
as a charge trap or as a conductive channel. It will be shown that it is the hole-transport in
host mCBP that is primarily affected by Ir(iprpmi)3. As shown in Figure 5.5(c), the EQE
0
4
8
12
16
20
24
0.1 1 10 100 1000
EQ
E (
%)
Current Density (mA/cm2)
0%
3%
7%
13%
17%
24%
0.0
0.2
0.4
0.6
0.8
1.0
1.2
380 480 580 680
Rela
tiv
e I
nte
nsi
ty (a
.u.)
Wavelength (nm)
0%
3%
7%
13%
17%
24%
0.1
1
10
100
0 5 10 15
Cu
rren
t D
en
sity
(m
A/c
m2)
Voltage (V)
0%
3%
7%
13%
17%
24%
(c)
(a)
Ir(iprpmi)3
(b)
Ir(iprpmi)3 Ir(iprpmi)3
109
is highly sensitive to the Ir(iprpmi)3 concentration in the emitter layer. At low
concentrations (3% and 7%), the EQE is high and relatively constant, ranging from 18 to
22% over a wide range of current densities from 0.1 to 20 mA/cm2 and decreasing only
slightly with increasing current density. At high concentrations (13%, 17%, and 24%),
the EQE is much reduced, but rises rapidly with increasing current density. For the device
with 13% Ir(iprpmi)3, the EQE increases from 8% at 0.1 mA/cm2 to a plateau of nearly
18% beyond 5 mA/cm2. For devices with 17% and 24% of Ir(iprpmi)3, the EQE is further
reduced and the fall-off at low current densities is much worse. The EQE at 0.1 mA/cm2
is only 0.8% and 1.7% for Ir(iprpmi)3 concentrations of 17% and 24%, respectively.
The fact that the devices with a lower Ir(iprpmi)3 dopant concentration (3% and
7%) exhibit a higher EQE and require a higher drive voltage indicates that electron-hole
recombination and subsequent generation and emission of Ir(iprpmi)3 excitons occurs
mainly near the interface with the hole-transport TAPC layer. Because BAlq is capable of
quenching Ir(iprpmi)3 triplet excitons, a lower EQE would have been observed if
electron-hole recombination was to occur at the BAlq interface. Ir(iprpmi)3 can affect
carrier transport in the emitter layer. Due to its HOMO (4.8 eV) and LUMO (2.2 eV)
energies, Ir(iprpmi)3 mainly affects the transport of holes in an mCBP matrix, but not
electrons. At low Ir(iprpmi)3 concentrations, holes can be effectively trapped by
Ir(iprpmi)3, causing recombination to be more localized at the TAPC interface. As a
result, the EQE is enhanced as observed in devices with 3% and 7% of Ir(iprpmi)3, since
TAPC, unlike BAlq, is incapable of quenching Ir(iprpmi)3 triplet excitons. As the
Ir(iprpmi)3 concentration increases, both the drive voltage and EQE are decreased. The
110
voltage is decreased because Ir(iprpmi)3 molecules at a high concentration can form a
continuous phase to accept hole injection from TAPC and to assist the hole transport
through the mCBP host matrix. The decrease in EQE can be attributed to two reasons: 1)
Ir(iprpmi)3 triplet excitons can diffuse more easily through the Ir(iprpmi)3 phase towards
the BAlq interface, where they suffer quenching by BAlq; and 2) increased hole transport
through the emitter layer results in increased recombination and exciton quenching at the
BAlq interface. As will be shown in section 5.3.5, self-quenching and/or triplet-triplet
annihilation play only a minor role in reducing EQE in Ir(iprpmi)3 doped emitters. .
The EQE dependence on current density, Figure 5.5(c), may also be correlated to
the hole trapping and transport characteristics of Ir(iprpmi)3. For devices with a low
concentration of Ir(iprpmi)3 (such as 3% and 7%), the total current through the emitter
layer is presumably dominated by the electron current as holes are trapped by Ir(iprpmi)3
molecules and the recombination zone is effectively localized at TAPC interface. High
EQE’s are maintained in these devices, regardless of the current density, as long as the
electron current dominates. With increasing concentration of Ir(iprpmi)3 in the emitter
layer (7%–24%), the charge transport becomes more bi-polar as more holes are
transported through Ir(iprpmi)3 in addition to electron transport through mCBP. Since
there is no barrier for hole injection from TAPC to Ir(iprpmi)3 in contrast to a large
barrier (0.5 eV) for electron injection from BAlq into mCBP, it is likely that the hole
current is transport limited whereas the electron current is injection limited, leading to a
space distribution of electrons and holes across the emitter layer that is dependent on the
bias voltage. At a low bias voltage, the hole current dominates as the voltage is not high
111
enough to cause electron injection from BAlq to mCBP, causing the recombination zone
to shift towards the BAlq interface, resulting in a decrease in EQE. With increasing bias,
more electron injection takes place, causing the recombination zone to shift away from
the BAlq interface towards the TAPC interface, resulting in increasing EQE. The extent
of this shift in the recombination zone is dependent on the Ir(iprpmi)3 concentration, with
the highest concentration resulting in the smallest shift away from the BAlq interface and
thus resulting in the lowest EQE.
5.3.4 Probing the recombination zone
In section 5.3.3 the dopant concentration for each device was fixed throughout the
entire mCBP layer. To gain a better understanding of where recombination and emission
occurs in the emitter layer, devices were designed and fabricated where the emitter layer
was split into two adjoining layers consisting of a layer of Ir(iprpmi)3 doped mCBP and a
layer of undoped mCBP. The general device structure is as follows: Anode/40 nm
TAPC/x nm mCBP+6% Ir(iprpmi)3/y nm mCBP/40 nm BAlq/Cathode, where x and y
refer to the thicknesses of the adjoined emitting layer. The Ir(iprpmi)3 was kept at 6%,
representing the low dopant concentration and optimal EQE derived from devices
discussed in section 5.3.3. The total thickness (x+y) of the doped (x) and undoped (y)
layers was kept at 30 nm with x and y varied from 5 to 25 nm. Each of these devices
exhibit identical EL spectra (not shown), similar to that of the device with 7% Ir(iprpmi)3
shown in Figure 5.5(a), indicating that emission is from Ir(iprpmi)3 and the
recombination zone is adjacent to the TAPC interface. Figure 5.6(a) and Figure 5.6(b)
show the device J-V and EQE-J data respectively. Essentially there are only minor
112
differences between the devices and their performance is similar to the 7% Ir(iprpmi)3
device in section 5.3.3; there is little difference in drive voltage and the EQE varied from
15 to 20% among the devices with a mild dependence on current density over a current
density range from 0.1 to 20 mA/cm2. This indicates that the recombination zone is
located within the first 5 nm of the TAPC interface. Furthermore, this also provides
evidence that triplet Ir(iprpmi)3 excitons are unable to diffuse to the BAlq interface and
explains why BAlq can be utilized as the ETL without affecting the device efficiency as
long as the Ir(iprpmi)3 dopant concentration in mCBP is relatively low (~6%).
Figure 5.6: OLED performance data when probing the recombination zone with 6% Ir(iprpmi)3.
Anode/40 nm TAPC/x nm mCBP+6% Ir(iprpmi)3/y nm mCBP/40 nm BAlq/Cathode. [x=25 nm,
y=5 nm], [x=20 nm, y=10 nm],[x=10 nm, y=20 nm], and [x=5 nm, y=25 nm]: (a) current density
vs. voltage, (b) EQE vs. current density.
To further illustrate the effect of Ir(iprpmi)3 on the recombination zone shift in the
emitter layer, devices were again fabricated with the emitter layer being split into two
adjoining layers consisting of a layer of Ir(iprpmi)3 doped mCBP and a layer of undoped
mCBP, except that the Ir(iprpmi)3 dopant concentration in the emitter layer was increased
to 25%. The EL spectra (not shown) for these devices are identical and are the same as
0.1
1
10
100
0 5 10 15
Cu
rren
t D
en
sity
(m
A/c
m2)
Voltage (V)
5 nm
10 nm
20 nm
25 nm
0
4
8
12
16
20
24
0.1 1 10 100
EQ
E (
%)
Current Density (mA/cm2)
5 nm
10 nm
20 nm
25 nm(a) (b)
Undoped mCBP
Undoped mCBP
113
what was observed for the 6% doped devices, indicating the emission is from Ir(iprpmi)3.
Also similar to the 6% doped devices, the EQE, Figure 5.7(b), for these 25% doped
devices is uniformly high (15-18%) and relatively independent of the current density.
This is in direct contrast with the devices having an emitter configuration where the
highly doped layer is in direct contact with the BAlq electron-transport layer. As shown
in the device where a 25% Ir(iprpmi) doped layer is in contact with BAlq (section 5.3.3),
the EQE is low (< 7%) and highly dependent on current density. Evidently, it is critical to
Figure 5.7: OLED performance data when probing the recombination zone with 25% Ir(iprpmi)3.
Anode/40 nm TAPC/x nm mCBP+25% Ir(iprpmi)3/y nm mCBP/40 nm BAlq/Cathode. [x=25 nm,
y=5 nm],[x=20 nm, y=10 nm], [x=10 nm, y=20 nm], and [x=5 nm, y=25 nm]: (a) current density
vs. voltage [ inset: voltage vs mCBP layer thickness of devices under current densities of 1, 5, and
10 mA/cm2] (b) EQE vs. current density.
separate the doped mCBP layer from the BAlq interface at high Ir(iprpmi)3
concentrations in order to avoid exciton quenching by the BAlq and achieve high EQE
values. However, the presence of an undoped mCBP layer can significantly increase the
drive voltage depending on the layer thickness. As shown in Figure 5.7(a), the increase in
0.1
1
10
100
1000
0 5 10 15
Cu
rren
t D
en
sity
(m
A/c
m2)
Voltage (V)
5 nm
10 nm
20 nm
25 nm
0
4
8
12
16
20
24
0.1 1 10 100
EQ
E (
%)
Current Density (mA/cm2)
5 nm
10 nm
20 nm
25 nm(a) (b)
Undoped mCBP
Undoped mCBP
6
8
10
12
0 10 20 30 40
Undoped mCBP (nm)
J (mA/cm2)
10
5
1
Voltage
114
voltage can be as large as 2 volts with a 25 nm undoped mCBP layer, which apparently
blocks hole injection from the mCBP doped layer. The inset of Figure 5.7(a) shows a
linear relationship between device voltage and the undoped mCBP layer thickness,
indicating the electric field across undoped mCBP is constant.These 25% doped devices
all have high EQE’s because the undoped mCBP layer prevents recombination from
occurring at the EML/BAlq interface while also preventing triplet excitons from diffusing
to this interface. There are noteworthy trends among the EQE-J curves displayed in
Figure 5.7(b). Depending on current density, the EQE ranges from 13-18% (device with
y=5 nm), 16-20% (devices with y=10 nm and y=20 nm), and 13-17% (device with y=25
nm). Devices with y=10 nm and y=20 nm have the highest EQE’s. Apparently an
undoped mCBP layer of 10-20 nm thick is needed to completely eliminate the influence
of BAlq. The device with only 5 nm of undoped mCBP has a lower efficiency
presumably from BAlq quenching. The device with 25 nm of undoped mCBP is
sufficiently thick to prevent BAlq quenching and yet the EQE is somewhat lower than
devices with 10 nm and 20 nm thick undoped mCBP. A possible explanation is that the
device with 25 nm of undoped mCBP has a thin doped layer which, being only 5 nm
thick, can lead to a higher density of triplet excitons and consequently a higher rate of
triplet-triplet annihilation and/or exciton-polaron interactions.
5.3.5 High EQE, low voltage OLEDs
Of all the devices described so far, those in section 5.3.3 with 3% and 7% of
Ir(iprpmi)3 have the highest EQE’s but also relatively high drive voltages due to
primarily the use of BAlq as the electron-transport layer, which has a low electron
115
mobility of 10-5
cm2/Vs [35]. By replacing BAlq with TmPyPB, a known electron-
transport material of high mobility (10-3
cm2/Vs) [36], we were able to demonstrate both
high EQE and low voltage in devices that had the following structure: Anode/40 nm
TAPC/30 nm mCBP+x% Ir(iprpmi)3/40 nm TmPyPB/Cathode. The Ir(iprpmi)3
concentration ranged from 3%–24%.
Figure 5.8: OLED performance data for Ir(iprpmi)3 concentration series with TmPyPB ETL.
Anode/40 nm TAPC/30 nm mCBP + x% Ir(iprpmi)3/40 nm TmPyPB/Cathode. x=3%–24% (a)
current density vs. voltage, (b) EQE vs. current density.
The J-V characteristics are shown in Figure 5.8(a) and indicate a clear reduction
in drive voltage with TmPyPB as the ETL. Furthermore, similar to what was observed for
devices in section 5.3.3 with BAlq as the ETL, the drive voltage decreases as the dopant
concentration increases. A reduction of approximately 1 V is realized by increasing the
Ir(iprpmi)3 concentration from 3% to 24%. As shown in Figure 5.8(b), high EQE’s were
obtained for all devices, ranging from 23% at low current densities to 15% at high current
densities. The efficiency remains high regardless of dopant concentration because
TmPyPB has a triplet energy of 2.78 eV and is unable to quench the Ir(iprpmi)3 triplet
0.1
1
10
100
0 2 4 6 8
Cu
rren
t D
en
sity
(m
A/c
m2)
Voltage (V)
3%
6%
13%
18%
24%
0
5
10
15
20
25
0.1 1 10 100
EQ
E (
%)
Current Density (mA/cm2)
3%
6%
13%
18%
24%
Ir(iprpmi)3
(a) (b)
Ir(iprpmi)3
116
excitons. This also indicates that self-quenching and/or triplet-triplet annihilation play
only a minor role in reducing the EQE.
5.3.6 Device lifetime
Lifetime testing was conducted on several representative Ir(iprpmi)3 doped
devices containing various hole and electron transport layers. For two devices, NPB was
used as the hole-transport layer, BAlq was used as the electron-tranport layer, and the
Ir(iprpmi)3 concentration was 6% and 12%. Two additional devices had the same device
structure except that TAPC replaced NPB as the hole-transport layer. For the last device,
TAPC and TmPyPB were the hole-transport layer and electron-transport layer,
respectively, while the Ir(iprpmi)3 concentration was 6%. It can be seen from Figure 5.9
Figure 5.9: Lifetime testing at 5 mA/cm2 for device structure, Anode/40 nm HTL/30 nm
mCBP+x% Ir(iprpmi)3/40 nm ETL/Cathode, [NPB-6% dopant-BAlq], [NPB-12% dopant-BAlq],
[TAPC-7% dopant-BAlq], [TAPC-13% dopant-BAlq], [TAPC-6% dopant-TmPyPB].
0.5
0.6
0.7
0.8
0.9
1
1.1
0 20 40 60 80 100
No
rmali
zed
Lu
min
ance
Time (hours)
NPB, 6% dopant, BAlq
NPB, 12% dopant, BAlq
TAPC, 7% dopant, BAlq
TAPC, 13% dopant, BAlq
0.5
0.6
0.7
0.8
0.9
1
0 1 2
No
rmal
ized
Lu
min
ance
Time (hours)
117
that the half-life (at a constant current of 5 mA/cm2), while quite short for all devices –
less than 100 hours, is rather dependent on the transport layers. Among the devices, the
device with TAPC and TmPyPB as the hole and electron transport layers, respectively, is
the most short-lived with a half-life of less than ½ hour (see Figure 5.9 insert), whereas
devices with NPB and BAlq replacing TAPC and TmPyPb, respectively, are the most
stable with a half-life in the range of 50-80 hours. Devices with TAPC as the hole-
transport layer and BAlq as the electron-transport layer have an intermediate half-life in
the range of 2-10 hours, respectively. The variation in lifetimes with the transport layers
indicates that the instability in these phosphorescent blue devices may not necessarily be
due to the emitter layer. It is known that fluorescent OLED devices are more stable with
NPB instead of TAPC as the hole-transport layer [50, 51]. As an electron-transport
material, BAlq is known to improve device stability [37]. The better stability of A1 and
A2 are in large part due to the use of NPB and BAlq as the respective hole and electron
transport layers, suggesting that recombination in these layers or at their interfaces with
the emitter layer, be it a fluorescent or phosphorescent emitter, are most detrimental to the
operational stability. Comparing the devices with NPB as the HTL, the device with 6%
Ir(iprpmi)3 has a longer half-life (80 vs. 50 hours) but a lower EQE (7% vs. 13%) than
the device with 12% Ir(iprpmi)3. This correlation of higher stability with lower efficiency
also applies to the devices with TAPC as the HTL, implying that the formation of triplet
excitons invariably also leads to device degradation.
The observed lifetimes for the devices with NPB as the HTL are inferior to what
was reported by Lin et al. [19] where lifetimes reached ~400-700 hrs. using Ir(iprpmi)3 as
118
the emitter. These disparities in lifetime can be attributed to different device architectures
and initial EQE’s.
The lifetime data presented here show that Ir(iprpmi)3-based devices with suitable
transport layers have longer lifetimes compared to FIrpic-based devices. In Flrpic-based
devices, the half-life is on the order of an hour or less regardless of the transport layer,
which has been attributed to the electrochemical degradation of Flrpic molecules related
to its fluorine substituents and the picolinate ligand. These structural features are not
present in Ir(iprpmi)3. Since recombination and emission can be engineered to take place
largely at the hole-transport interface using low-concentration phosphorescent dopants as
hole traps, the ETL material has no significant impact on device lifetime. Thus device
stability is more hinged on the selection of hole-transport materials as well as the host
materials for the emitter, assuming that the Ir-based dopant is relatively stable. We have
shown in this study how important a role the HTL material plays for both device EQE
and lifetime. Since amines and carbazoles are known to undergo homolytic C-N bond
cleavage in the excited state [38-40], it is likely that these common classes of hole-
transport materials are prompt to degrade in blue phosphorescent emitters because of
possible recombination in the hole-transport layer or interface resulting in excitons with
energy exceeding the dissociation energy of the C-N bonds. It follows that none of the
HTL materials used in this study are practical for achieving lifetimes that match those of
green and red phosphorescent devices, where the low-energy dopants can suppress
recombination in both hole and electron transport layers. To significantly improve the
stability of blue phosphorescent devices, new hole-transport materials will be needed.
119
5.4 CONCLUSIONS
Ir(iprpmi)3 as a dopant in blue phosphorescent emitters has been investigated with
mCBP as the host and various materials as the hole and electron transport layers. We
have shown that Ir(iprpmi)3 as a dopant in host mCBP is capable of trapping (at low
concentration) and transporting (at high concentration) holes. Utilizing Ir(iprpmi)3 to trap
holes, the recombination zone can be confined at or near the hole-transport layer. As a
result, EQE values obtained are highly dependent on the HTL used. With TAPC as HTL,
which does not quench Ir(iprpmi)3 triplet e citons, EQE’s over 20% have been achieved.
Utilizing Ir(iprpmi)3 for hole transport, the recombination zone can be shifted towards the
ETL. With low-triplet BAlq as the ETL, much reduced EQE’s are obtained that vary not
only with dopant concentration but also electric field, indicating various extent of exciton
quenching by BAlq. Quenching by the ETL can be eliminated and high EQE’s retained
by inserting a thin undoped mCBP layer between the ETL and EML. We show that
device lifetime using Ir(iprpmi)3 as the emitting dopant is significantly improved
compared to when FIrpic is the emitting dopant. Furthermore, device lifetime is highly
dependent on the choice of host and hole/electron transport layers.
120
REFERENCES
[1] M. Segal, C. Mulder, K. Celebi, M. Singh, K. Rivoire, S. Difley, T. Van Voorhis,
M.A. Baldo, Proc. SPIE, 6999 (2008) 699912.
[2] K.S. Yook, J.Y. Lee, Adv. Mater., 24 (2012) 3169.
[3] N. Chopra, J. Lee, Y. Zheng, S.-H. Eom, J. Xue, F. So, Appl. Phys. Lett., 93
(2008) 143307.
[4] H. Sasabe, E. Gonmori, T. Chiba, Y.-J. Li, D. Tanaka, S.-J. Su, T. Takeda, Y.-J. Pu,
K.-i. Nakayama, J. Kido, Chem. Mater., 20 (2008) 5951.
[5] S.-J. Su, E. Gonmori, H. Sasabe, J. Kido, Adv. Mater., 20 (2008) 4189.
[6] J. Lee, J.-I. Lee, J.Y. Lee, H.Y. Chu, Appl. Phys. Lett., 94 (2009) 193305.
[7] L. Xiao, S.-J. Su, Y. Agata, H. Lan, J. Kido, Adv. Mater., 21 (2009) 1271.
[8] C.W. Lee, J.Y. Lee, Adv. Mater., 25 (2013) 1.
[9] W. Sotoyama, T. Satoh, M. Kinoshita, M. Tobise, K. Kawato, T. Ise, H. Takizawa,
S. Yamashita, Dig. Tech. Pap. - Soc. Inf. Disp. Int. Symp., 41 (2010) 556.
[10] H. Lee, H. Ahn, C. Lee, J. Inf. Disp., 12 (2011) 219.
[11] I. Rabelo de Moraes, S. Scholz, B. Luessem, K. Leo, Org. Electron., 12 (2011)
341.
[12] K.S. Yook, O.Y. Kim, J.Y. Lee, Synth. Met., 161 (2011) 2677.
[13] V. Sivasubramaniam, F. Brodkorb, S. Hanning, H.P. Loebl, V. van Elsbergen, H.
Boerner, U. Scherf, M. Kreyenschmidt, J. Fluorine Chem., 130 (2009) 640.
[14] E. Baranoff, B.F.E. Curchod, J. Frey, R. Scopelliti, F. Kessler, I. Tavernelli, U.
Rothlisberger, M. Gratzel, M.K. Nazeeruddin, Inorg. Chem., 51 (2012) 215.
121
[15] K.P. Klubek, C.W. Tang, L.J. Rothberg, Org. Electron., 15 (2014) 1312.
[16] N.C. Giebink, B.W. D'Andrade, M.S. Weaver, P.B. Mackenzie, J.J. Brown, M.E.
Thompson, S.R. Forrest, J. Appl. Phys., 103 (2008) 044509.
[17] H. Yamamoto, J. Brooks, M.S. Weaver, J.J. Brown, T. Murakami, H. Murata,
Appl. Phys. Lett., 99 (2011) 033301.
[18] J. Zhuang, W. Li, W. Su, Y. Liu, Q. Shen, L. Liao, M. Zhou, Org. Electron., 14
(2013) 2596.
[19] C. Lin, P.B. MacKenzie, R.W. Walters, J.-Y. Tsai, C.S. Brown, J. Deng, U.S.
Patent No. 7,902,374 (2011).
[20] D.B. Knowles, C. Lin, P.B. Mackenzie, J.-Y. Tsai, R.W. Walters, S.A. Beers, C.S.
Brown, W.H. Yeager, U.S. Patent No. 7,915,415 (2011).
[21] V. Adamovich, M.S. Weaver, B. D'Andrade, U.S. Patent No. 8,580,402 (2013).
[22] B. D'Andrade, P.B. MacKenzie, M.S. Weaver, J.J. Brown, U.S. Patent No.
8,557,399 (2013).
[23] A.B. Tamayo, B.D. Alleyne, P.I. Djurovich, S. Lamansky, I. Tsyba, N.N. Ho, R.
Bau, M.E. Thompson, J. Am. Chem. Soc., 125 (2003) 7377.
[24] M.J.T. Frisch, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman,
J. R.; Montgomery, Jr., J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J.
M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.;
Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.;
Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai,
H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.;
122
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A.
J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G.
A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A.
D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.;
Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.;
Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez,
C.; and Pople, J. A., Gaussian 03, Revision C.02, Gaussian, Inc., Wallingford CT,
2004.
[25] T. Fei, X. Gu, M. Zhang, C. Wang, M. Hanif, H. Zhang, Y. Ma, Synth. Met., 159
(2009) 113.
[26] S.J. Lee, K.-M. Park, K. Yang, Y. Kang, Inorg. Chem., 48 (2009) 1030.
[27] T. Karatsu, M. Takahashi, S. Yagai, A. Kitamura, Inorg. Chem., 52 (2013) 12338.
[28] J.J. Brooks, R.C. Kwong, Y.-J. Tung, M.S. Weaver, B.W. D'Andrade, V.
Adamovich, M.E. Thompson, S.R. Forrest, J.J. Brown, Proc. SPIE, 5519 (2004)
35.
[29] S.-J. Yeh, M.-F. Wu, C.-T. Chen, Y.-H. Song, Y. Chi, M.-H. Ho, S.-F. Hsu, C.H.
Chen, Adv. Mater., 17 (2005) 285.
[30] F.-C. Fang, Y.-L. Liao, K.-T. Wong, M.-H. Tsai, H.-W. Lin, H.-C. Su, C.-c. Wu,
C.-I. Wu, Proc. SPIE-Int. Soc. Opt. Eng., 6192 (2006) 61921R.
[31] J.-h. Niu, W.-l. Li, H.-z. Wei, M.-t. Li, W.-m. Su, Q. Xin, Z.-q. Zhang, Z.-z. Hu, J.
123
Phys. D Appl. Phys., 38 (2005) 1136.
[32] P. Schroegel, N. Langer, C. Schildknecht, G. Wagenblast, C. Lennartz, P.
Strohriegl, Org. Electron., 12 (2011) 2047.
[33] S. Gong, X. He, Y. Chen, Z. Jiang, C. Zhong, D. Ma, J. Qin, C. Yang, J. Mater.
Chem., 22 (2012) 2894.
[34] K.B. Kahen, J.R. Vargas, D.Y. Kondakov, C.T. Brown, L. Cosimbescu, V. Jarikov,
U.S. Patent No. 6,876,684 (2005).
[35] J.H. Seo, K.H. Lee, B.M. Seo, J.R. Koo, S.J. Moon, J.K. Park, S.S. Yoon, Y.K.
Kim, Org. Electron., 11 (2010) 1605.
[36] S.-J. Su, T. Chiba, T. Takeda, J. Kido, Adv. Mater., 20 (2008) 2125.
[37] T. Watanabe, K. Nakamura, S. Kawami, Y. Fukuda, T. Tsuji, T. Wakimoto, S.
Miyaguchi, Proc. SPIE-Int. Soc. Opt. Eng., 4105 (2001) 175.
[38] D.Y. Kondakov, W.C. Lenhart, W.F. Nichols, J. Appl. Phys., 101 (2007) 024512.
[39] D.Y. Kondakov, J. Appl. Phys., 104 (2008) 084520.
[40] D.Y. Kondakov, Proc. SPIE, 7051 (2008) 70510I.
124
CHAPTER 6
SUMMARY AND FUTURE RESEARCH
6.1 SUMMARY
Even as OLEDs are currently achieving commercial success, significant research
effort is focused on improving device performance metrics such as efficiency, voltage and
color. The use of iridium (Ir) cyclometalated phosphorescent emitters has been shown to
produce the highest efficiency in OLED devices. In fact, red and green Ir-based emitters
have already been adopted into consumer products due to their ability to produce devices
that have a combination of high efficiency and excellent lifetime. While high efficiency
has also been demonstrated for PHOLEDs using blue Ir-based emitters, the device
lifetimes lag far behind those of the red and green. This thesis focused on blue PHOLED
instability mechanisms from a materials and device perspective. In particular, two key
blue emitting iridium compounds were extensively studied, bis(4,6-difluorophenyl-
pyridinato-N,C2) picolinate iridium (FIrpic) and tris[1-(2,6-diisopropylphenyl)-2-phenyl-
1H-imidazole]iridium(III) (Ir(iprpmi)3).
FIrpic is the most commonly employed blue emitting phosphorescent dopant and
is often used to demonstrate high efficiencies for researchers developing new host and
transport materials. However device lifetime is typically not reported. In Chapter 3,
FIrpic was studied as a light-emitting dopant using both single and mixed host materials.
Devices incorporating high triplet energy host and transport materials yielded devices
125
with EQE’s over 15%, however lifetimes for these devices were all very short (less than
20 minutes at 5 mA/cm2). Switching to charge transport materials with triplet energies
lower than FIrpic reduced the EQE by 50% or more, however the lifetime was improved.
A mixed host LEL (TCTA:UGH3) in combination with the low triplet energy BAlq as the
ETL was found to provide the best combination for EQE (~9%) and lifetime (~12 hours).
While this lifetime is significantly improved compared to devices having the highest
EQE’s, it is ultimately concluded that achieving stable blue phosphorescent devices for
commercial applications is not possible when using FIrpic as the light-emitting dopant.
It was also found that the supposedly electron-transporting co-host UGH3 is in
fact a poor electron transporting material. Its role within the LEL is determined to be that
of a spacer group, serving to slow down hole transport and minimizing exciplex
formation between the hole transport host and the adjacent electron transport material. It
was also found that emission occurring from the host or transport materials is detrimental
to device lifetime. The host mCBP was used in device structures that have been reported
in the patent literature to achieve blue PHOLED device lifetime of 10,000 hours.
However, we found that mCBP as a host for FIrpic does not improve device lifetimes
over those of single host TCTA or mixed host TCTA:UGH3 devices.
In Chapter 4, hole transport in FIrpic doped mCBP and the effect on the device
operational voltage and stability were investigated. The host mCBP was found to be
stable in supporting hole transport but unstable with respect to electron-hole
recombination. As a dopant, FIrpic was found to be unstable with respect to both hole
transport and charge recombination processes. The results from Chapter 3 and Chapter 4
126
provide clear evidence that FIrpic is unsuitable to use for generating light and for
conducting hole transport within OLED devices.
In Chapter 5, the blue phosphorescent dopant Ir(iprpmi)3 was investigated.
Ir(iprpmi)3 lacks specific structural features present within FIrpic, namely fluorine
substituents and the ancillary picolinate ligand, that have been attributed to molecular
degradation. The photophysical properties of this material were reported. PHOLED
devices using Ir(iprpmi)3 as the light-emitting dopant and mCBP as the host have been
evaluated. By optimizing the dopant concentration and the materials for the electron and
hole-transport layers, external quantum efficiencies greater than 20% have been achieved.
Device lifetimes are significantly improved compared to FIrpic-based devices, with
lifetimes approaching 100 hours having been achieved. In addition to the differences in
molecular design, these improvements are attributed to the control of the electron-hole
recombination and emission regions within the emitter layer as well as the choice of
material for the transport layers.
As discussed in Chapters 3 and 5, high triplet energy transport materials are
necessary to confine triplet excitons within the light-emitting layer for devices. TmPyPB,
with a high ET of 2.8 eV, was used as ETL. Presumably, triplet excitons formed on FIrpic
(ET = 2.6 eV) or Ir(iprpmi)3 (ET = 2.66 eV) within the LEL near the LEL/ETL interface
will not be quenched by TmPyPB. However, it is possible that non-thermalized FIrpic or
Ir(iprpmi)3 triplet excitons with a triplet energy higher than 2.8 eV could produce
TmPyPB triplet excitons through energy transfer. Based on a Boltzmann distribution,
there is also a finite probability for the formation of TmPyPB excited states at the
127
LEL/ETL interface. Regardless of how it is formed, the TmPyPB excited state has high
enough energy to promote bond breaking, resulting in the formation of quencher species
capable of impacting device lifetime. This concept can be further extended to include the
host and hole transport materials. The creation of high energy excited states in these
materials may also contribute significantly to the poor lifetimes that we observed for the
highest EQE devices.
As discussed in Chapter 2, the KOMET coater is a bench-top thermal evaporation
system that we designed and built for the fabrication of all OLED devices described in
this thesis. The key feature for this system is the compact thermal source design utilizing
multiple, low power consumption tube boats for depositing organic materials. In addition
to low-power consumption, these boats are highly efficient with respect to material
utilization. The KOMET coater produced OLED devices with excellent reproducibility.
With a compact design, the KOMET coater (<$40,000) is more affordable compared to
traditional evaporation coaters (>$100,000).
6.2 FUTURE RESEARCH
Based on the results presented in this thesis, it is clear that FIrpic should be
avoided as the quest to improve blue PHOLED device performance continues.
Researchers should instead opt for alternative phosphorescent blue dopants such as
Ir(iprpmi)3. While we have shown that using this dopant has produced devices having
significantly improved lifetime compared to FIrpic, the lifetimes we have achieved (~100
hours) are still far too low for commercial applications. Using Ir(iprpmi)3 as the emitter in
a different device architecture, Lin et al. [1] showed the device lifetime can approach
128
400-700 hours. Additional studies need to be conducted with Ir(iprpmi)3 to investigate
whether or not this material is intrinsically unstable or if device stability is instead limited
by the choice of charge transport and host materials. Laser desorption ionization mass
spectrometry (LDI-MS) has been utilized to determine that FIrpic degradation results in
the loss of the picolinate ligand [2]. LDI-MS can similarly be used to investigate the
molecular stability of Ir(iprpmi)3 upon device degradation. Wtih Ir-based phosphorescent
devices generally using relatively large concentrations (6-20%) of the dopant, it should
also be easier to look for MS signals related to these types of materials compared to their
fluorescent counterparts that typically use dopant concentrations of 1-3%. An alternative
known technique [3-5] may be to extract the materials after degradation using solvents
and then use liquid chromatography mass spectrometry (LC-MS) in an attempt to identify
the degradation products related to Ir(iprpmi)3 or derivatives of imidazole-based Ir-
cyclometalated complexes. In Chapter 3 we found that FIrpic was unstable to hole
transport. Since Ir(iprpmi)3 has a shallow HOMO and effectively traps or transports holes
depending on concentration, additional studies are needed to determine how stable this
dopant is to charge transport.
While using Ir(iprpmi)3 is an excellent alternative to FIrpic for evaluating blue
PHOLED devices due to its structural differences and commercial availability, alternative
dopants should also be studied (Figure 6.1). As discussed in Chapter 3, imidazole
derivative Ir(tdmp)3, which is not commercially available, has been reportedly used as a
129
Figure 6.1: Different types of blue phosphorescent dopants.
dopant to achieve a lifetime as high as 10,000 hours [6] in a device structure similar to
those reported by Lin et al. Ir(tdmp)3 has a more rigid structure compared to Ir(iprpmi)3
and also has a bulky dimethyl-phenyl substituent attached to the double bond of the
imidazole ring. It is possible that by making the molecule more rigid and bulky, it is less
susceptible to degradation simply because potentially reactive moieties such as the
nitrogen atoms or the double bond are more shielded from neighboring molecules.
Carbene complexes [7-10] are another class of blue emitting material that are showing
promise and can be compared with the imidazole derivatives. In any case, PHOLED
device and analytical (degradation) studies need to be conducted on these different
classes of dopants and their derivatives to compare and contrast how the differences in
molecular structure affects device performance, especially lifetime.
In addition to studying new blue phosphorescent dopants, there is also a need to
study new charge transport and host materials. In this thesis, triarylamine and carbazole
derivatives such as NPB, TAPC, TCTA, and mCBP have been used as HTL and host
materials. Since these types of materials are known to undergo C-N bond cleavage in
Blue Phosphorescent Dopants
FIrpic Ir(iprpmi)3Ir(tdmp)3 Carbene complex
130
OLED devices, alternative materials that are either free of C-N bonds or that are
sterically hindered so that the C-N bond is less prone to cleavage are needed. From a
device engineering perspective, preventing electron-hole recombination from occurring
on the hole transport materials will also limit or prevent these types of materials from
degrading. For example, the HOMO level of Ir(iprpmi)3 is very shallow and therefore this
material is a very strong hole trap. This type of material may be able to prevent
recombination from occurring on any other hole transport materials. However, the
electron transport through the host also needs to be considered. With the LUMO level of
Ir(iprpmi)3 being rather high, electron transport occurs preferentially on the host material.
This requires the host to be stable to electron transport. We used mCBP for transporting
electrons, additional studies need to be conducted to determine if this material is stable
for this role. Electron-only devices can be made and the voltage rise with time will help
to determine the charge transport stability of mCBP. LDI-MS or LC-MS can then be used
in an attempt to identify degradation products. It is also important to use host materials
that are stable in the excited state. In Chapter 4 we found that forming the mCBP excited
state in a model OLED with Alq as the emitter resulted in enhanced device degradation.
For devices using mCBP as the host for Ir(iprpmi)3, further study is required to assess
whether mCBP is entering the excited state and factoring into the short device lifetimes.
Depending on the type of phosphorescent blue dopant and host materials, the
electron transport material can also play a critical role with respect to device
performance. If recombination occurs at the LEL/ETL interface, then high triplet energy
ETL materials are typically used to prevent exciton quenching. In order to have both high
131
triplet energy and high electron mobility, ETL materials typically utilize pyridine-like
nitrogen atoms within the molecular structure. Unfortunately, these types of materials
often contribute to device instability. If recombination can be engineered to occur away
from the ETL, then degradation related to the ETL material may be avoided. However, a
host material that is stable to electron transport is required.
Since iridium cyclometalated complexes have been shown to achieve EQE’s
greater than 20%, there is often a desire to maintain these high efficiencies while also
utilizing device structures designed to enhance the lifetime. As we showed in Chapter 5,
often the correlation between efficiency and lifetime is unfortunately inverse – decreased
lifetimes for higher efficiency devices. In order to improve the device lifetime, one
strategy is to move away from Ir complexes with a high triplet energy that would produce
deeper blue emissions to Ir complexes with a lower triplet energy that would produce
lighter blue emissions. Although the color of the blue emission is somewhat
compromised, the lifetimes of the light blue PHOLEDs are often significantly higher than
the deep blue PHOLEDs. Engineering stable light-blue and high-efficiency PHOLEDs
may be a more expedient path to practical applications. Furthermore, deeper blue
emissions can be obtained with these light blue emitters by using device structures having
internal microcavity or external color filters. By easing the triplet energy requirements
and allowing for lower EQE blue devices, many new phosphorescent and host materials
are open to investigation and the prospects for achieving practical PHOLEDs should be
much higher.
132
REFERENCES
[1] C. Lin, P.B. MacKenzie, R.W. Walters, J.-Y. Tsai, C.S. Brown, J. Deng, U.S.
Patent No. 7,902,374 (2011).
[2] I. Rabelo de Moraes, S. Scholz, B. Luessem, K. Leo, Org. Electron., 12 (2011)
341.
[3] D.Y. Kondakov, J. Appl. Phys., 104 (2008) 084520.
[4] D.Y. Kondakov, Proc. SPIE, 7051 (2008) 70510I.
[5] D.Y. Kondakov, T.D. Pawlik, W.F. Nichols, W.C. Lenhart, J. Soc. Inf. Disp., 16
(2008) 37.
[6] B. D'Andrade, P.B. MacKenzie, M.S. Weaver, J.J. Brown, U.S. Patent No.
8,557,399 (2013).
[7] P. Erk, M. Bold, M. Egen, E. Fuchs, T. Gessner, K. Kahle, C. Lennartz, O. Molt,
S. Nord, H. Reichelt, C. Schildknecht, H.-H. Johannes, W. Kowalsky, Dig. Tech.
Pap. - Soc. Inf. Disp. Int. Symp., 37 (2006) 131.
[8] C.-F. Chang, Y.-M. Cheng, Y. Chi, Y.-C. Chiu, C.-C. Lin, G.-H. Lee, P.-T. Chou,
C.-C. Chen, C.-H. Chang, C.-C. Wu, Angew. Chem., Int. Ed., 47 (2008) 4542.
[9] H. Sasabe, J.-i. Takamatsu, T. Motoyama, S. Watanabe, G. Wagenblast, N. Langer,
O. Molt, E. Fuchs, C. Lennartz, J. Kido, Adv. Mater. (Weinheim, Ger.), 22 (2010)
5003.
[10] C.-H. Hsieh, F.-I. Wu, C.-H. Fan, M.-J. Huang, K.-Y. Lu, P.-Y. Chou, Y.-H.O.
Yang, S.-H. Wu, I.C. Chen, S.-H. Chou, K.-T. Wong, C.-H. Cheng, Chem. - Eur.
J., 17 (2011) 9180.