UNIVERSITY OF CALIFORNIA

SANTA CRUZ

PHOSPHORESCENT ORGANIC LIGHT EMITTING DEVICES

A thesis submitted in partial satisfaction of the
requirements for the degree of

MASTER OF SCIENCE

in

PHYSICS

by

Glenn P. Gray

December 2010

The Thesis of Glenn P. Gray

is approved:

Professor Sue A. Carter, Chair

Professor Joshua Deutsch

Professor David Belanger

Tyrus Miller
Vice Provost and Dean of Graduate Studies
Copyright
c by

Glenn P. Gray

2010
Table of Contents

List of Figures iv

Abstract vi

Acknowledgments vii

1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 From OLEDs to PHOLEDs . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Electroluminescence in Matter . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Basic OLED Operation and Structure . . . . . . . . . . . . . . . . . . 8
1.4.1 Double and Multilayered OLEDs . . . . . . . . . . . . . . . . . 10
1.4.2 Single Layer PHOLEDs . . . . . . . . . . . . . . . . . . . . . . 12
1.5 OLED Efficiencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.5.1 Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

2 Experiment 22

3 Results 24

4 Conclusion and Next Steps 29

A Preparing the Mix 30

B Preparing the Substrate 32

C Device Fabrication Procedure 34

D Device Characterization Procedure 36

Bibliography 40

iii
List of Figures

1.1 Energy level diagram showing the two types of luminescence; fluores-
cence and phosphorescence. . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2 Jablonski diagram illustrating the various absorption and emission
processes in molecules. Internal conversion (IC) corresponds to non-
radiative transitions between states of the same multiplicity, while
intersystem crossing (ISC) corresponds to non-radiative transitions
between states of differing multiplicity. Figure from reference[1] p.224. 9
1.3 Simplified energy level diagram for single layer OLED showing charge-
injection EL process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.4 (a) Double layer OLED with the electron transport layer (ETL) as the
emissive layer (EML). (b) Double layer device with the hole transport
layer (HTL) as the EML. (c) Multilayer device where the EML is a
separate layer, not a part of the ETL or HTL. . . . . . . . . . . . . . . 11
1.5 Simplified energy level diagram for single layer PHOLED showing
charge-injection EL process. . . . . . . . . . . . . . . . . . . . . . . . . 12
1.6 Illustration of current J and recombination current Jr . . . . . . . . . . 17
1.7 Radiometric quantities and their photometric analogs. . . . . . . . . . 18
1.8 A graph of the photometric response of the eye superimposed with the
visible spectrum. The peak is centered over green, meaning our eyes
detect greens better than reds or blues, and we don’t see anything
outside the visible spectrum. . . . . . . . . . . . . . . . . . . . . . . . 19
1.9 Photometric responsivity of EG&G UV-100BQ photodiode. . . . . . . 20

3.1 Comparison of photocurrent versus voltage for device made with

PEDOT:PSS and LiF interfacial layers, and device made with PE-
DOT:PSS but no LiF interfacial layers. . . . . . . . . . . . . . . . . . 25
3.2 Current versus voltage and photocurrent versus voltage data for one
device tested three times on three different days. . . . . . . . . . . . . 26
3.3 Current versus voltage and photocurrent versus voltage data for best
device. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

iv
3.4 Spectra from device made with PEDOT:PSS and LiF interfacial layers,
and just PEDOT:PSS layer. . . . . . . . . . . . . . . . . . . . . . . . . 27
3.5 AFM image of morphology for device made which used 1,2
dichloroethane as the solvent. . . . . . . . . . . . . . . . . . . . . . . . 28

D.1 Experimental setup for taking spectra. . . . . . . . . . . . . . . . . . . 37

D.2 Experimental setup for taking current versus voltage and photocurrent
versus voltage curves data. . . . . . . . . . . . . . . . . . . . . . . . . . 37

v
 Abstract

Phosphorescent Organic Light Emitting Devices

by

Glenn P. Gray

Organic Light Emitting Devices (OLEDs), have emerged as promising alternatives

to solid state lighting applications and flat panel displays. In an effort to increase

the internal conversion efficiency of OLEDs, phosphors are added to the active layer

and Phosphorescent Organic Light Emitting Devices (PHOLEDs) are made. Here we

investigate the fabrication and characterization of PHOLEDs doped with the iridium

complex Ir(ppy)3 , a low molecular weight green emitter. First we will motivate the

research and explain the physics of device operation. Then we will discuss device

characterization and fabrication, and finally present the results achieved thus far.

Furthermore we will discuss the addition of the ionic salt Bu4 NBF4 and the use of

thermo-electric annealing (TE annealing) to aid in charge injection, eliminating the

need for additional charge injection aiding layers thus creating a single layered device.
 Acknowledgments

First I would like to thank Sue for allowing me to do this research. I would also like

to thank specific members of the lab who helped me in one way or another; Yvonne

Rodriguez, Jeremy Olson, Chris France, Ben Balaban, and Guanmei Zhang. Last

but not least I’d like to thank my family for their love and support.

vii
Chapter 1

Introduction

Here we present work done on fabricating and characterizing organic

light emitting devices (OLEDs), specifically OLEDs doped with phosphors called

PHOLEDs. First we will motivate the research by highlighting some of the novel

applications of PHOLEDs. Next we will set the stage for the research conducted

here by putting it into the context of work done so far. Then the physics of electro-

luminescence is explained, followed by a review of basic OLED/PHOLED operation

and device architectures. The introduction is concluded with an explanation of effi-

ciencies and measurements used to characterize OLEDs.

The experiment section outlines the procedure for making PHOLEDs1 . The

results section covers the progress made so far, and the conclusion section summarizes

the findings and suggests next steps for continuing the research done here.
1
See Appendix for detailed step-by-step fabrication and characterization procedures.

1
1.1 Motivation

The next generation of flat panel display (FPD) and general solid state

lighting applications has arrived with the OLED. In order for OLEDs to be compet-

itive in these markets they must exhibit exceptional brightness, strong efficiencies,

low operating voltages, and long lifetimes[2]. OLEDs are thin film devices made ei-

ther from thermally evaporating small organic molecules (called SMOLEDs), or are

polymer based (called PLEDs) and solution processable made using wet fabrication

techniques such as spin casting, doctor blading, and ink-jet printing[3].

One of the biggest advantages of OLED technology is the ability to deposit

the materials onto virtually any substrate such as glass, plastic, or thin metal foils,

resulting in devices that can be made paper thin, flexible, and transparent[4][5][6][7].

Imagine a TV in your living room that has the thickness of a wall poster and appears

to be part of the wall until it is turned on, or a PDA that can be rolled up and stored

in a small space such as a pen. Also, since OLED technology is thin and lightweight

it is ideal for applications in portable displays such as cell phones, which require low

power.

Furthermore, the color of OLEDs can be tuned by introducing differ-

ent organic dopants, and white OLEDs have already been made whose high effi-

ciency and low power consumption make them ideal for general solid state lighting

applications[8]. OLEDs are also scalable, meaning they can be made into large area

devices such as large area billboards or large area white lights, helping to reduce en-

2
ergy consumption, which is a major issue in the United States where 20% of electrical

energy consumption is used for lighting[9].

1.2 From OLEDs to PHOLEDs

Obtaining high efficiencies has been the major driving force in OLED re-

search and development. OLEDs are made from one or multiple organic layers sand-

wiched between a transparent anode and a metallic cathode, and over the last twenty

years new materials, multilayered devices, and dopants have been introduced to in-

crease efficiency. Second to improving efficiency in OLEDs has been understanding

degradation mechanisms and increasing the lifetime of devices, which is required for

competitiveness in the FPD and lighting industries and which will greatly contribute

to understanding device physics. The trends in the literature show reports of high

efficiency devices, followed by methods of increasing device lifetime. Here we review

some of the work done on OLEDs which have influenced current device structures

and our current understanding.

The first OLED was reported in 1965 by Helfrich and Schneider[10] and

was made from single crystal anthracene, however they were not very efficient and

required high operating voltages due to the thickness of the crystal.

The first SMOLED, reported by Tang and VanSlyke in 1987, was a dou-

ble layer device made from layers of aromatic diamine and 8-hydroxyquinoline alu-

minum (Alq3 ), sandwiched between a transparent indium tin oxide (ITO) anode and

3
a magnesium-silver alloy cathode. These devices exhibited a ‘high’ external quan-

tum efficiency, defined as the fraction of photons emitted in the forward viewing

direction to injected electrons, of 1%[11]. The first PLED made from the conju-

gated polymer poly(para-phenylene vinylene) (PPV) was reported by Friend et al.

in 1990[12]. These two seminal reports marked the beginning of the OLED research

and development boom.

A major breakthrough in increasing efficiency was reached in 1998 by Baldo

et al.[13], and by Hoshino and Suzuki[14], with the introduction of a phosphorescent

dopant and the birth of the phosphorescent OLED (PHOLED). In previous works

OLEDs relied on the use of fluorescent materials to generate light. However the

internal quantum efficiency, defined as the ratio of photons generated within the

material to injected electrons, is limited to 25% based on spin statistics[15], which

ultimately restricts the external quantum efficiency and hence device performance.

In phosphorescent devices the internal quantum efficiency can reach 100%, which

increases device efficiency and hence lowers the current required to achieve a desired

brightness[16]. The utilization of phosphorescent dopants in OLEDs was the logical

next step in increasing the efficiency of OLEDs.

The first PHOLEDs used a platinum doped porphine complex PtOEP2 ,

and had internal and external quantum efficiencies of 23% and 4% respectively[13].

The second generation PHOLEDs use iridium complexes such as Ir(ppy)3 as the

phosphorescent dopant[17][18][19], which have internal quantum efficiencies of almost

2
2,3,7,8,12,13,17,18-octaethyl-21H, 23H-porphine platinum(II).

4
100%[20][21]. Devices based on Ir(ppy)3 with external quantum efficiencies of greater

than 10%[22] and lifetimes3 of greater than 10, 000h have been achieved for green and

red PHOLEDs with initial brightness of 600cd/m2 and 300cd/m2 respectively[23].

Device lifetimes in PHOLEDs have been shown to be inversely proportional to initial

brightness which is a typical characteristic of fluorescent OLEDs[24].

1.3 Electroluminescence in Matter

The process of radiative relaxation of molecules at low temperatures is

known as luminescence, and is distinct from incandescence which is a high tempera-

ture process4 . Photoluminescence (PL) is the simplest form in which an electron in a

molecule absorbs a photon, is excited to a higher energy and reemits another photon

upon falling back to the ground state. However there are many forms of lumines-

cence, which depend on how the molecule was excited before luminescing, such as

chemiluminescence, bioluminescence, sonoluminescence, and mechanoluminescence.

Here we focus on charge-injection electroluminescence (EL), in which electrons and

holes are injected into the luminescent material where they form singlet and triplet

excitons, which decay to the ground state and emit light. Before explaining the

charge-injection EL process it is useful to review the energy structure of electrons in

molecules.

The allowed energy levels of an electron in a molecule correspond to sta-

3
Device lifetime is defined as the amount of time a device takes to degrade from one half of its
initial brightness.
4
Luminescence is also known as ‘cold emission’ for this reason.

5
tionary states, which are represented by the eigenfunctions and eigenvalues of the

molecule’s Hamiltonian found after solving the Schrödinger equation. One way to

number these states is to consider their spin multiplicities. States with multiplicity

1 are referred to as singlet states Sn and have total spin angular momentum of 0.

States with multiplicity of 3 are called triplet states Tn and have total spin angular

momentum of 1. Using these labels the energy structure of a molecule can be num-

bered, from lowest energy to highest energy (starting from the ground state) they

are S0 , S1 , . . . , Sn , and T1 , T2 , . . . , Tn . Luminescence occurs when an electron in an

excited singlet or triplet state relaxes to the singlet ground state. The two types

of emission are called fluorescence for Sn → S0 transitions and phosphorescence for

Tn → S0 transitions, see Fig. 1.1.

Figure 1.1: Energy level diagram showing the two types of luminescence; fluorescence
and phosphorescence.

It is important to note that normally the ground state is a singlet state,

and that fluorescence transitions from excited singlet states to the ground state are

spin-allowed; the spin is conserved from the initial to the final state. This is not the

case for phosphorescence, where the transitions are spin-forbidden, meaning the spin

of the electron must flip in order to go from a triplet state to a singlet state. The

6
consequence of these spin-allowed and spin-forbidden transitions can be explained

by the lifetime of the excited states. Electrons like to be in their lowest energy state,

so electrons singlet states decay rapidly (0.1 to 1 ns), while electrons in triplet states

can have much longer lifetimes (sometimes greater than 10s) since the transition is

spin-forbidden. This is why phosphorescent materials are used in glow-in-the-dark

paints, once the paint is ‘charged’ (electrons in the material are excited) it can emit

light over a long period of time, even in the dark.

Strictly speaking transitions from triplet states to singlet states are spin-

forbidden and therefore have zero intensity, nonetheless phosphorescent transitions

occur. The reason for this has to do with spin-orbit coupling. When the spin-orbit

coupling term HSO is added to the molecule’s Hamiltonian and treated as a per-

turbation, the new stationary states of the new hamiltonian are mixtures of singlet

and triplet states from the old Hamiltonian, resulting in what are called “impure

singlet” and “impure triplet” states. The amount of mixing depends on the strength

of the spin-orbit coupling term; more mixing corresponds to strong spin-orbit cou-

pling. When the spin-orbit coupling term in the Hamiltonian is strong, transitions

between these triplets and singlets (phosphorescence) is possible. The strength of

the spin-orbit coupling also determines the lifetimes of the excited states, making it

possible for some materials with strong spin-orbit coupling to have phosphorescence

transitions with the same lifetimes as fluorescence.

The phenomena of fluorescence and phosphorescence is most readily illus-

trated using a Jablonski diagram (Fig. 1.2), which shows the different processes that

7
can lead to luminescence in molecules. The diagram is more detailed than Fig. 1.1

because it includes vibrational energy levels, which are located above each of the sin-

glet and triplet levels. In the diagram absorption (A) and emission (F - fluorescence,

P - phosphorescence) are shown as straight lines, and the wavy lines correspond to

non-radiative processes (IC - internal conversion, VR - vibrational relaxation, ISC -

intersystem crossing).

As Fig. 1.2 shows, phosphorescence occurs after an electron in an excited

singlet state undergoes ISC to an excited triplet state (or one of the higher vibrational

levels of that state) after which it phosphoresces, falling down to the ground state,

or undergoes VR and does not emit a photon. The rate of ISC is determined by how

much mixing of singlet and triplet states occur due to spin-orbit coupling, and must

be fast for efficient phosphorescent materials. Lifetimes of triplet states and triplet

energy transfer have been studied in Alq3 based devices[25].

1.4 Basic OLED Operation and Structure

OLEDs are made up of one or multiple thin film organic semiconducting

layers sandwiched between a metallic cathode and transparent anode.

The simple structure shown in Fig. 1.3 illustrates the charge-injection EL

process similar to almost all OLEDs. There are five processes involved in charge-

injection EL;(i) charge injection, (ii) charge migration, (iii) exciton formation, (iv)

light emission, and (v) light extraction. Electrons and holes are injected from the

8
Figure 1.2: Jablonski diagram illustrating the various absorption and emission pro-
cesses in molecules. Internal conversion (IC) corresponds to non-radiative transitions
between states of the same multiplicity, while intersystem crossing (ISC) corresponds
to non-radiative transitions between states of differing multiplicity. Figure from
reference[1] p.224.

9
Figure 1.3: Simplified energy level diagram for single layer OLED showing charge-
injection EL process.

cathode and anode, respectively, and travel through the material in opposite di-

rections. Charge transport through the organic material can be characterized by

hopping between adjacent molecules, and holes and electrons are assumed to recom-

bine and emit light on a specific molecule [26][27][28]. If the electron and hole meet

on a single molecule, singlet and triplet excitons are formed. Because the excitons

have relatively high binding energies of ∼ 1eV they can travel throughout the ma-

terial without dissociating. Not all excitons radiatively recombine, for example in

fluorescent OLEDs only 25% of excitons contribute to light generation.

1.4.1 Double and Multilayered OLEDs

As previously mentioned, OLEDs are made up of one or more thin semicon-

ducting organic layers sandwiched between a transparent anode (ITO) and a metal

cathode. Double layer devices are made up of an electron transport layer (ETL)

10
and a hole transport layer (HTL), which are chosen for their good electron and hole

transport and charge injection from the cathode and anode respectively. The ETL or

the HTL can be doped so that it becomes the emissive layer (EML). So, in a double

layer device, electrons and holes are injected into the ETL and HTL then meet at the

EML to recombine and form excitons, followed by luminescence. Figure 1.4 shows

device architectures for double and multilayered devices.

Figure 1.4: (a) Double layer OLED with the electron transport layer (ETL) as the
emissive layer (EML). (b) Double layer device with the hole transport layer (HTL)
as the EML. (c) Multilayer device where the EML is a separate layer, not a part of
the ETL or HTL.

Multilayer devices incorporate more layers to control and improve device

efficiency[29]. For example, electron and/or hole blocking layers are introduced to

aid in increasing the recombination current, and hence the charge balance factor, by

confining excitons to the EML[30]. And devices made with double emission layers

11
have been shown to be highly efficient[31]. Furthermore, interfacial layers between

the cathode and ETL and/or anode and HTL are commonly introduced to reduce

the charge injection barrier, aiding in charge injection.

1.4.2 Single Layer PHOLEDs

Single layer PHOLEDs are made from a blend of a electron transport mate-

rial (ETM), a hole transport material (HTM), and a phosphorescent dye dispersed in

a appropriate solvent(s). The ETM and HTM are doped polymers with good charge

transport characteristics and make up the polymer matrix in which the phospho-

rescent dye is dispersed. The phosphorescent dye chosen has fast ISC rates (strong

spin-orbit coupling) and has a low molecular weight to aid in the dispersion of the

dye into the polymer matrix. Figure 1.5 shows a diagram of the energy structure of

a single layer PHOLED.

Figure 1.5: Simplified energy level diagram for single layer PHOLED showing charge-
injection EL process.

12
 One of the biggest factors that determines the performance of a single layer

PHOLED is the morphology; i.e. the phase separation of the different ingredients

of the blend. A uniform morphology, which can be characterized by a smooth AFM

image of the surface, yields the best devices with the most uniform light emission

and highest efficiency. As such, the choice of the solvent(s) is very important in

controlling the morphology of the single layer PHOLED film.

When choosing a solvent two things must be kept in mind. First and

foremost the solvent must properly dissolve all of the materials. This is difficult to

achieve with a one part solvent, so using a binary (two-part) solvent is recommended.

Park et al. used a binary solvent of 1,2-dichloroethane and chloroform (3:1 mixing

weight ratio) to achieve bright uniform EL[32][33][34]. Secondly, it is useful from an

fabrication point of view to choose solvents that have a high boiling point. Solvents

with lower boiling points are not ideal for large scale wet deposition techniques such

as ink-jet printing.

To lower the charge injection barrier you can do two things; either use

interfacial layers between the cathode/organic layers and organic/anode layers, or

add a small amount of salt additive and do a thermo-electric (TE) anneal as shown

by Park et al.[32][33][34]. For example PEDOT:PSS5 can be used as the interfacial

layer between the anode (ITO) and the organic layer, and lithium fluoride (LiF) can

be used as the interfacial layer between the anode (Al) and the organic layer[35].

The addition of the ionic salt tetrabutylammonium tetrafluoroborate

5
poly(3,4-ethylenedioxythiophene) :poly(styrene sulfonate)

13
(Bu4 NBF4 ) and using the TE anneal can also lower the charge injection barriers.

The salt is added in small amounts, anywhere from 0.0025 wt. % to 0.01 wt. %.

After the devices are made the device is placed on a hot plate and heated to T ◦ C

and biased with a voltage V. The temperature of the anneal is chosen so that it is

below the glass transition temperature of the materials used. Once the device lights

up the field is terminated and it is taken off the hot plate and cooled. The use of

heat and electric field acts to split the salt into its cation and anion and allow for the

adsorption of them toward the cathode and anode respectively. It is important that

the field points in the right direction or the ions will adsorb in the wrong direction.

1.5 OLED Efficiencies

It is important to standardize and define the different measurements used

to characterize the luminescence of OLEDs[36]. Here we define and explain the

standard measurements quoted in the literature.

First the external quantum efficiency, ηext , is defined as the ratio of

the number of photons emitted by the OLED in the forward viewing direction to

the number of electrons injected. This is different from the internal quantum

efficiency ηint , which is defined as the ratio of the total number of photons generated

within the structure of the OLED to the number of electrons injected. The difference

arises from the fact that not all photons generated within the structure of the OLED

make it out of the device; some photons are re-absorbed by the material and some

14
are emitted in a direction other than the forward viewing direction. We introduce

the light output coupling factor, ηc to relate the external and internal quantum

efficiencies;

ηext = ηc · ηint . (1.1)

Note that in a perfect system the same number of electrons and holes are injected,

and all electrons and holes recombine to create excitons which all radiatively decay

and are coupled out of the device in the forward direction. A more quantitative

description of the internal and external quantum efficiencies can be obtained after

making a few assumptions.

First of all, we assume that no charge is present in the active material

without charge injection, so we only need to focus on charges injected from the elec-

trodes. Second, molecular excited states; singlet and triplet excitons, are generated

by charge recombination. This means that excitons are generated when a hole from

the anode combines with an electron from the cathode to create singlet and triplet

excitons. Emission occurs when the singlet or triplet excitons radiatively decay to

the ground state. Thirdly we assume that luminescence of the active material has a

high fluorescence quantum yield φf , meaning the material is highly fluorescent (as

opposed to phosphorescent), i.e. most singlet excitons radiatively decay6 .

The internal quantum efficiency ηint is defined as

ηint = γβφf , (1.2)

6
The efficiencies of PHOLEDs can be explained in a similar way to fluorescent OLEDs, however
the rate of intersystem crossing must be taken into account as well as the phosphorescence quantum
yield φp .

15
where γ is the charge balance factor, β is the production efficiency of emissive ex-

citations, and φf is the fluorescence quantum yield. The charge balance factor is a

measure of the efficiency of charge recombination; if electrons and holes injected do

not form excitons they merely travel through the material and out.

In terms of the injected current the charge balance factor is defined as

Jr
γ= , (1.3)
J

where Jr is the electron-hole recombination current density, and J is the measured

current density. If we define the amount of current injected from the cathode and the

anode as Je and Jh , and the amount of current that does not contribute to exciton

formation (i.e. the current that makes it through the device) as Je 0 and Jh 0 , we can

write down expressions for J and Jr (see Fig. 1.6);

J = Jh + Je 0 = Je + Jh 0 , (1.4)

Jr = Jh − Jh 0 = Je − Je 0 . (1.5)

If all injected electrons and holes recombine within the emissive layer,

J = Jr then γ = 1, and γ < 1 if Je 0 or Jh 0 is non-zero. In order to increase the

efficiency of charge recombination (increase γ) a balance between injected holes and

electrons must be obtained, and they must be contained within the emissive region

of the device7 .
7
In single layer devices this amounts to adjusting the charge injection barrier at the elec-
trode/organic interface and/or balancing the mobilities of the electrons and holes. In multilayer
devices electron blocking and hole blocking layers can be introduced to increase the efficiency of
charge recombination by aiding in the confinement of injected electrons and holes.

16
 Figure 1.6: Illustration of current J and recombination current Jr .

The efficiency of the production of emissive excitations β is defined as the

branching ratio for singlets and triplets, and is taken to be π based on simple spin

statistics.

An estimate of the light output coupling factor χ can be derived from simple

ray optics,
r
1
χ=1− 1− , (1.6)
n2

where n is the index of refraction of the material. A typical value for n in OLEDs is

1.7, giving the frequently used approximation for χ as 0.2.

Plugging in some numbers gives an estimate of the maximum external quan-

tum efficiency of the OLED. Taking χ = 0.2, β = 0.25, γ = 1, and φf = 1, the max-

imum ηext = 5.0%. However it must be noted that this is not an exact ceiling value

for efficiencies of OLEDs, but is a useful measure to aid in the understanding of

17
OLED efficiency.

1.5.1 Measurements

Next we will define and explain the standard measurements used to char-

acterize OLED performance. The field of optical measurements is broken into two

sub-fields; radiometry and photometry. Radiometry deals with the measurement of

energy per time, and hence has units of watts (W). However radiometric quantities

are ill-suited for describing the performance of light emitting displays since they do

not account for the sensitivity of our personal light detectors; our eyes (see Fig. 1.8).

Photometric quantities have analog radiometric quantities (see Fig. 1.7), except new

units are introduced to account for the sensitivity of our eyes.

Figure 1.7: Radiometric quantities and their photometric analogs.

The need for photometric quantities arises from the fact that our eyes only

see light from the visible spectrum, while photodetectors can (theoretically) detect

the entire spectrum. For example an OLED that emits in the UV (or anywhere

outside the visible spectrum) will have a measurable radiometric power efficiency,

but zero photometric power efficiency.

These new units in the photometric quantities are the lumen (lm) and the

candela (cd). The lumen is defined as the measure of the power of light perceived by

18
the human eye, the lumen can also be defined in terms of the candela; 1lm = 1cd·sr.

Hence a light source that emits 1 candela uniformly in all directions emits a total of

4π lumens.

Figure 1.8: A graph of the photometric response of the eye superimposed with the
visible spectrum. The peak is centered over green, meaning our eyes detect greens
better than reds or blues, and we don’t see anything outside the visible spectrum.

When testing devices, two testing setups are needed due to the limitations

of the equipment being used. The first measurement setup is used to measure the

spectra of the device, which amounts to biasing the device so that it is turned on

and counting the number of photons emitted at each wavelength. The second setup

involves measuring the current and photocurrent of the device as a function of bias

voltage. The photocurrent is measured by a photodiode, which has its own photopic

responsivity R(λ) that needs to be taken into account.

The external quantum efficiency ηext is defined as

R
q λIdet (λ)dλ
ηext = R (1.7)
hcf IOLED R(λ)dλ

19
where R(λ) is the photodiode responsivity (shown in Fig. 1.9), Idet (λ) is the pho-

tocurrent from the photodetector, f is the fraction of light emitted from the device

to the light coupled into the detector, q is the charge of the electron, h is Plank’s

constant, and c is the speed of light. Experimentally, it is best for the photodetector

to be larger in area and as close as possible to the device in order to maximize f .

Figure 1.9: Photometric responsivity of EG&G UV-100BQ photodiode.

The luminous efficiency ηL is measured in candelas per amp, and is very

similar to ηext except that the luminous efficiency takes the photopic response of the

eye into account.

AL
ηL = (1.8)
IOLED

where L is the luminance of the OLED (measured in [cd/m2 ]), A is the device area,

and IOLED is the current through the OLED.

The luminous power efficiency, or luminosity, ηP measured in lumens per

watt is defined as the ratio of the luminous power emitted in the forward direction

20
LP to the power used by the OLED at a particular voltage,

LP
ηP = (1.9)
IOLED V

Alternatively we can define ηP explicitly in terms of the photopic response of the

eye,
R
φ0 g(λ)Idet (λ)/R(λ)dλ
ηP = (1.10)
f IOLED V

where φ0 = 683lm/W is the peak photopic response at λ = 555nm.

21
Chapter 2

Experiment

Green PHOLEDs were fabricated by spincasting a solution of active ma-

terial onto a glass slide printed with indium tin oxide (ITO), and evaporating an

aluminum top electrode. The active material consisted of a blend of the hole trans-

porting material TPD1 , the electron transporting material Bu-PBD2 , the green emit-

ter Ir(ppy)3 3 , and the hole transporting host material PVK4 in 1,2 dichloroethane.

Specifically the blend consisted of

TPD 0.08 wt. %
Bu-PBD 0.32 wt. %
Ir(ppy)3 0.06 wt. %
PVK 0.34 wt. %
where the weight percentages are with respect to the solvent. The small amount of

Ir(ppy)3 was chosen to prevent concentration quenching, and to aid in solubility as

Ir(ppy)3 has been shown to have poor solubility in commonly used solvents[37] (such
1
N,N’-diphenyl-N,N’-bis(3-methylphenyl)-1,1’biphenyl-4,4’-diamine
2
2-(4-biphenyl)-5-(4-tert-butylphenyl)-1,3,4-oxadiazole
3
tris(2-phenylpyridinato) iridium
4
Poly(9-vinylcarbazole)

22
as toluene, chlorobenzene, tetrahydrofuran, and 1,2-dichloroethane).

Devices incorporated PEDOT:PSS and/or LiF interfacial layers to aid in

charge injection into the active layer. The interfacial layer PEDOT:PSS and active

material were spun in a fumehood, and annealed in a vacuum oven. Typically5 150µL

of filtered PEDOT:PSS was spun at 1500rpm and annealed at 100-110 ◦ C for 20-

30min. The active layer was formed by spincasting 60µL of the blend and vacuum

annealing for 30min at 100-110 ◦ C. The LiF interfacial layer and aluminum top

electrode were evaporated in a vacuum chamber below 2×10−6 T orr at approximately

1Å/sec to achieve thicknesses of 1nm and 50-100nm respectively.

Single layer devices incorporated 0.001 - 0.005wt.% Bu4 NBF4 in the active

solution and were TE annealed. The TE anneal was performed in the glovebox on a

hot plate at 85-95 ◦ C while being sourced by 7V.

Spectra, current versus voltage, and photocurrent versus voltage measure-

ments were taken in the glovebox, see Fig. D.1 and Fig. D.2. Procedures for prepar-

ing the active blend, preparing the substrates, depositing the materials, and testing

devices can be found in the Appendices.

5
The amount and spin speed chosen yielded full coverage of the glass substrate.

23
Chapter 3

Results

The performance of devices made with PEDOT:PSS and LiF interfacial

layers were compared. Two device structures were used to investigate the role of the

interfacial layers:

ITO/PEDOT:PSS/active material without salt/LiF/Al

ITO/PEDOT:PSS/active material without salt/Al

The active layer was a mix of

TPD 0.08 wt. %
Bu-PBD 0.32 wt. %
Ir(ppy)3 0.06 wt. %
PVK 0.34 wt. %
in 1,2 dichloroethane.

As expected, the turn on voltage for the devices made with both PE-

DOT and LiF interfacial layers was lower than devices made with only PEDOT

(see Fig. 3.1), and the spectra for each device was almost identical (see Fig. 3.4).

On another note, during testing of the best device O2 levels in the glovebox

fluctuated between 6ppm and 100ppm, indicating that such fluctuations should not

24
Figure 3.1: Comparison of photocurrent versus voltage for device made with PE-
DOT:PSS and LiF interfacial layers, and device made with PEDOT:PSS but no LiF
interfacial layers.

significantly effect device performance, as was previously thought, see Fig. 3.2.

The JV and PJV data for the best device is shown in Fig. 3.3.

Furthermore, all devices that were made worked in one way or another.

This means that all of the devices emitted light in one of two ways; either the pixel

had a uniform glow or there was a very bright spot on the pixel and the rest of the

pixel was dim. Pixels with bright spots degraded quickly after multiple JV/PJV

testing runs, and the bright spots turned into dark spots after multiple JV/PJV

tests. Finally the turn on voltages for the devices made here were too high, and

currents were too high indicating poor efficiency.

25
Figure 3.2: Current versus voltage and photocurrent versus voltage data for one
device tested three times on three different days.

26
Figure 3.3: Current versus voltage and photocurrent versus voltage data for best
device.

Figure 3.4: Spectra from device made with PEDOT:PSS and LiF interfacial layers,
and just PEDOT:PSS layer.

27
 One explanation for the high power and non-uniform light emission is poor

morphology of the films. As previously stated it is hard to dissolve multiple materials

into one solvent. Hence binary solvents with different mixing weight ratios were used

to improve morphology, however an ideal morphology was not achieved. Fig. 3.5

shows an AFM image of the morphology of the best device, which has aggregates

on the order of a few hundred nanometers. Decreasing the size of these aggregates,

hence improving the morphology, should improve device characteristics.

Figure 3.5: AFM image of morphology for device made which used 1,2 dichloroethane
as the solvent.

28
Chapter 4

Conclusion and Next Steps

In conclusion, PHOLEDs made from the iridium complex Ir(ppy)3 were

fabricated, however they lacked the desired efficiency exhibited by devices made in

other groups. Future research should focus on improving the morphology of the films,

optimizing the dispersion of the materials in the blend by choosing and appropriate

solvent(s). Improving morphology will yield a significant improvement in efficiency

by lowering the currents and voltages required for bright light emission.

Furthermore, once the proper solvent(s) and desired morphology have been

found more work can be done on devices made with the salt additive and the TE

anneal, with the goal of achieving turn on voltages and operating currents comparable

to devices made without salt and with PEDOT:PSS and LiF interfacial layers.

29
Appendix A

Preparing the Mix

Preparing the mix of ingredients for the active layer consists of choosing a

solvent, calculating how much of each material to use from the weight of the solvent,

weighing each component out, putting everything in a vial with a stir-bar, and leaving

the mix on a stir-plate for 24 hours before making devices.

The amount of solvent is chosen so that the material with the smallest

wt. % required can be weighed out on the scale. The scale can measure accurately

down to about 1mg, so the amount of solvent must be chosen accordingly. Here

is an example, the weight percentages are the same as what is in the Single Layer

PHOLED section, the mix used 4ml of 1,2-dichloroethane, which has a density of

1.253 g/ml. So the mix uses 5.012g of solvent and

TPD 4.01mg
Bu-PBD 16.03mg
Ir(ppy)3 3.00mg
PVK 17.04mg
First weigh out the different materials and add them to the vial, then pipet the

30
solvent into the vial, add a stir-bar, and put it on the stir-plate over night1 To add

a small amount of salt, say 0.005 wt. % of Bu4 NBF4 - which translates into 0.25mg

of salt, make a stock solution because the scale is not accurate enough to weigh such

a small amount of material. To make a stock solution, weigh out a small amount of

material - say 1mg - put it in a vial with a stir bar and add 1ml of solvent. Now the

stock solution has a density of 1mg/ml. To add 0.25mg of salt pipet 250µl of the

stock solution into the mix2

1
Leave the mix on the stir-plate long enough so that all the materials can be dissolved, one way
to check is to hold the vial up to the light and see if there are any materials still floating around.
2
The mix can also be made by making stock solutions for each of the different materials, this
method uses less material but requires more bookkeeping.

31
Appendix B

Preparing the Substrate

The first thing that needs to be done is clean the ITO coated glass slides.

Cleaning five slides at once is common, but up to ten can be cleaned at once if they

are doubled up in the staining jar1 . A commonly used slide cleaning procedure is as

follows2 :

1. Using a razor blade, scrape off old material3 in a petri dish filled with Alconox,

and dispose of waste in the hazardous liquids container in the fumehood.

2. Sonicate in Alconox for 60min4

3. Rinse off slides with de-ionized water, DI H2 O, three times over the sink

4. Sonicate in DI H2 O for 30min

1
If you do clean ten at once, make sure the ITO sides are not touching, place the slides in the
staining jar so they are glass-to-glass.
2
Again this cleaning procedure is not hard and fast, for instance the amount of time in each step
is variable. Letting the slides sonicate for a long time is advantageous for two reasons: first I’m
confident they are clean, and second I can be doing other work while the slides are cleaning
3
If the slides are new this step is unnecessary.
4
I usually set the temperature of the sonicator to 40◦ C during the cleaning process.

32
 5. Using a cotton tipped swab, swab both sides of each slide in a petri dish filled

with ethanol

6. Sonicate in ethanol for 30min

7. Dry slides with filtered nitrogen

Now that the slides are clean devices can be made. It is a good idea to turn

on the vacuum oven at least 30min before making devices as it takes that long to get

to the correct temperature.

33
Appendix C

Device Fabrication Procedure

The procedure here will include the deposition of PEDOT:PSS and LiF

layers. Here is the general order of operations, there are four basic steps: (1) deposit

PEDOT:PSS interfacial layer, (2) deposit active layer, (3) evaporate LiF interfacial

layer, and (4) evaporate Al top electrode. Steps (1) and (2) are done at the spinning

station in the fumehood, and steps (3) and (4) are done in the evaporation chamber

in the glovebox.

The device fabrication procedure is as follows:

1. Filter PEDOT:PSS using 45µm syringe tip filter.

2. Spin down 150µl PEDOT:PSS at 1500rpm1

3. Anneal in the vacuum oven at 100◦ C to 110◦ C for 30min.

1
The important thing here is getting a nice uniform film. Using clean slides will help, but also use
more material if you’re not getting good coverage. Also the spin speed can be increased or decreased
to change the thickness of the different layers.

34
 4. Using a cotton tipped swab dipped in ethanol, carefully clean off the outer

edges of the device to expose the ITO pads.

5. Spin down 60µl of active material at 1000rpm2 for 45sec.

6. Anneal in the vacuum oven for 30min at 100◦ C to 110◦ C.

7. Using a cotton tipped swab dipped in ethanol, clean off the outer area of

the device exposing the ITO pads, now devices are ready for top electrode

evaporation.

Now that the top electrode is ready to be evaporated, first transfer the

slides to the glovebox and load the slides into the evaporation mask. Then raise the

bell jar, load the devices, put LiF and Al in their boats, and close the bell jar. Then

evaporate 1nm to 10nm of LiF (at about .5Å/s, the slower the deposition rate the

more uniform the layer will be). Next evaporate 100nm of Al at about 1Å/s. Now

the devices are made and ready to be tested3 .

2
Again, for thicker devices reduce the spin speed. If devices are shorting try a thicker layer.
3
These devices do not like oxygen, but they survived in the glovebox for up to seven days with
the O2 concentration between 5ppm and 45ppm. This means that devices can be made at night
and then tested the next morning, which is advantageous for efficiently using alloted time on lab
equipment during business hours during the week.

35
Appendix D

Device Characterization Procedure

The first thing to do before you test the devices is see if they work. Load

the slide into the device holder and pin out to one pixel, then using the Keithely

2400 source with current and measure the voltage. Remember that these devices

only work in forward bias, so make sure that you pin out to the pixel correctly. Set

the device holder near the window of the glovebox so that you can see it, and ramp

up the source current while watching for the device to light up1 . If the pixel lights

up the device is ready to be tested.

There are two measurement setups; one for taking the spectra of the pixel,

and the other for taking current versus voltage and photocurrent versus voltage

curves. Figures D.1 and D.2 show the setups for the two measurements.

To take spectra of the device the Ocean Optics mini spectrometer and

spectrum.vi program are used. The basic idea consists of sourcing the device with
1
Devices reported here had high turn on voltages and high currents, hence start at a low current
of 1mA, and ramp up in steps of 1mA

36
 Figure D.1: Experimental setup for taking spectra.

Figure D.2: Experimental setup for taking current versus voltage and photocurrent
versus voltage curves data.

37
enough voltage so that it is very bright and the spectrum.vi program gives a plot of

the spectrum.

First turn on the spectrum.vi program and load the device into the device

holder. Next, using a small allen-head screwdriver and the marking tape, align and

fix the fiber holder to the device holder. Once the fiber holder is attached, source

the pixel of the device that will be used for spectrum data and move the hole in

the fiber holder over the desired pixel. Then place the Ocean Optics fiber into the

hole on the fiber holder and take the spectrum. Now, spectrum.vi (the icon on the

desktop says “run spectrum”) is setup to take different types of spectra; it allows you

to choose what combination of fiber size, grating, and integrating sphere. Use the

600µm fiber and the UV-Vis grating, and no sphere. When ready to take spectra,

click start and the spectra will appear on the screen. Good spectra data counts are

in the 3000 range, and the number of counts can be adjusted by using the ’averaging’

and integration time to increase the signal, and also ramp up the source voltage to

make the device brighter. Once the spectra of the pixel is taken the JV and PJV

curves can be taken for the rest of the device, there is no need to take spectra data

for each pixel.

To take JV and PJV curves use the multisweep program (an Igor program)

on the left hand side computer2 . Fix the photodiode to the device tester and pin

out to the pixel to be tested. The pins are connected to the Keithley 2400 source

meter and the photodiode is connected to the Keithley 485 picoammeter. Set the
2
Remember, visually check to see if the pixel lights up before taking JV/PJV data

38
ammeter so that it is on auto-range mode. Make sure the GPIB cables from both

of the Keithley’s are connected to the left hand side computer. Then specify the file

path where data is to be saved, and a name for the data about to be taken. To set the

voltage range create a text file (for example V in.txt), with the step size, starting

voltage, and ending voltage at the top separated by commas, and tell multisweep

where the file is. Then click “acquire photocurrent” to measure photocurrent and

current versus voltage at the same time. Finally click start and multisweep will take

the data.

39
Bibliography

[1] M. Klessinger and J. Michl, Excited States and Photochemistry of Organic

Molecules. New York: VCH, 1995.

[2] S. R. Forrest, “The road to high efficiency organic light emitting devices,” Or-

ganic Electronics, vol. 4, pp. 45–48, 2003.

[3] B. Geffroy, P. le Roy, and C. Prat, “Organic light-emitting diode (oled) technol-

ogy: Materials, devices and display technologies,” Polymer International, pp.

572–582, 2006.

[4] P. E. Burrows, G. Gu, V. Bulović, Z. Shen, S. R. Forrest, and M. E. Thompson,

“Achieving full-color organic light-emitting devices for lightweight, flat-panel

displays,” IEEE Transactions on Electronic Devices, vol. 44, no. 8, pp. 1188–

1202, 1997.

[5] Z. Shen, P. E. Burrows, V. Bulovicć, S. R. Forrest, and M. E. Thompson, “Three-

color, tunable, organic light-emitting devices,” Science, vol. 276, pp. 2009–2011,

1997.

40
 [6] G. Gustafsson, G. M. Treacy, Y. Cao, F. Klavetter, N. Colaneri, and A. J.

Heeger, “The “plastic” led: A flexible light-emitting device using a polyaniline

transparent electrode,” Synthetic Metals, pp. 4122–4127, 1993.

[7] L. Zhou, A. Wanga, S. Wu, J. Sun, S. Park, and T. N. Jackson, “All-organic

active matrix flexible display,” Applied Physics Letters, vol. 88, 2006.

[8] A. R. Duggal, Organic Electroluminescence. CRC Press and The International

Society for Optical Engineering, 2005, ch. Organic Electroluminescent Devices

for Solid State Lighting.

[9] N. C. Inc., “U.s. lighting market characterization volume 1: National lighting

inventory and energy consumption estimate,” U.S. Department of Energy, 2002.

[10] W. Helfrich and W. G. Schneider, “Recombination radiation in anthracene crys-

tals,” Physical Review Letters, vol. 14, pp. 229–231, 1965.

[11] C. W. Tang and S. A. VanSlyke, “Organic electroluminescent diodes,” Applied

Physics Letters, vol. 51, pp. 913–915, 1987.

[12] J. H. Burroughes, D. D. C. Bradley, A. R. Brown, R. N. Marks, K. Mackay,

R. H. Friend, P. L. Burns, and A. B. Holmes, “Light-emitting diodes based on

conjugated polymers,” Nature, vol. 347, pp. 539–541, 1990.

[13] M. A. Baldo, D. F. O’Brien, Y. You, A. Shoustikov, S. Sibley, M. E. Thomp-

son, and S. R. Forrest, “Highly efficient phosphorescent emission from organic

electroluminescent devices,” Nature, vol. 395, pp. 151–154, 1998.

41
[14] S. Hoshino and H. Suzuki, “Electroluminescence from triplet excited states of

benzophenone,” Applied Physics Letters, vol. 69, pp. 224–226, 1996.

[15] M. A. Baldo, D. F. O’Brien, M. E. Thompson, and S. R. Forrest, “Excitonic

singlet-triplet ratio in a semiconducting organic thin film,” Physical Review B,

vol. 60, no. 20, pp. 14 422–14 428, 1999.

[16] M. Baldo, S. Forrest, and M. Thompson, Organic Electroluminescence. CRC

Press and The International Society for Optical Engineering, 2005, ch. Organic

Electrophosphorescence.

[17] C. Lee, K. B. Lee, and J. Kim, “Polymer phosphorescent light-emitting devices

doped with tris(2-phenylpyridine) iridium as a triplet emitter,” Applied Physics

Letters, vol. 77, no. 15, pp. 2280–2282, 2000.

[18] M. A. Baldo, S. Lamansky, P. E. Burrows, M. E. Thompson, and S. R. Forrest,

“Very high-efficiency green organic light-emitting devices based on electrophos-

phorescence,” Applied Physics Letters, vol. 75, no. 1, pp. 4–6, 1999.

[19] S. Tokito, M. Suzuki, F. Sato, M. Kamachi, and K. Shirane, “High-efficiency

phosphorescent polymer light-emitting devices,” Organic Electronics, vol. 4, pp.

105–111, 2003.

[20] C. Adachi, M. A. Baldo, M. E. Thompson, and S. R. Forrest, “Nearly 100%

internal phosphorescence efficiency in an organic light emitting device,” Journal

of Applied Physics, vol. 90, no. 10, pp. 5048–5051, 2001.

42
[21] C. Adachi, M. A. Baldo, S. R. Forrest, and M. E. Thompson, “High-efficiency

organic electrophosphorescent devices with tris(2-phenylpyridine)iridium doped

into electron-transporting materials,” Applied Physics Letters, vol. 77, no. 6, pp.

904–906, 2000.

[22] C. Adachi, M. E. Thompson, and S. R. Forrest, “Architectures for efficient

electrophosphorescent organic light-emitting devices,” IEEE Journal on Selected

Topics in Quantum Electronics, vol. 8, no. 2, pp. 372–376, 2002.

[23] R. C. Kwong, M. S. Weaver, M. M. Lu, Y. Tung, A. B. Chwang, T. X. Zhou,

M. Hack, and J. J. Brown, “Current status of electrophosphorescent device

stability,” Organic Electronics, vol. 4, pp. 155–164, 2003.

[24] R. C. Kwong, M. R. Nugent, L. Michalski, T. Ngo, K. Rajan, Y. Tung, M. S.

Weaver, T. X. Zhou, M. Hack, M. E. Thompson, S. R. Forrest, and J. J. Brown,

“High operational stability of electrophosphorescent devices,” Applied Physics

Letters, vol. 81, no. 1, pp. 162–164, 2002.

[25] M. A. Baldo and S. R. Forrest, “Transient analysis of organic electrophospho-

rescence: I. transient analysis of triplet energy transfer,” Physical Review B,

vol. 62, no. 16, pp. 10 958–10 966, 2000.

[26] T. T. Tsui, Organic Electroluminescence. CRC Press and The International So-

ciety for Optical Engineering, 2005, ch. Electroluminescence in Small Molecules.

[27] J. Kalinowski, Organic Electroluminescence. CRC Press and The International

43
 Society for Optical Engineering, 2005, ch. Emission Mechanisms in Organic

Light-Emitting Diodes.

[28] J. Shinar, Organic Light-Emitting Devices: A Survey. New York: Springer-

Verlag, 2004.

[29] S. A. VanSlyke, C. H. Chen, and C. W. Tang, “Organic electroluminescent

devices with improved stability,” Applied Physics Letters, vol. 69, pp. 2160–

2162, 1996.

[30] M. Ikai, S. Tokito, Y. Sakamoto, T. Suzuki, and Y. Taga, “Highly efficient phos-

phorescence from organic light-emitting devices with an exciton-block layer,”

Applied Physics Letters, vol. 79, no. 2, 2001.

[31] G. He, M. Pfeiffer, K. Leo, M. Hofmann, J. Birnstock, R. Pudzich, and J. Sal-

beck, “High-efficiency and low-voltage p-i-n electrophosphorescent organic light-

emitting diodes with double-emission layers,” Applied Physics Letters, vol. 85,

no. 17, pp. 3911–3913, 2004.

[32] S. W. Kim, J. H. Park, S. S. Oh, D. Y. Kim, E. H. Choi, G. S. Cho, Y. H.

Seo, S. O. Kang, B. Park, Y. Saito, N. Watanabe, H. Takezoe, and J. Watan-

abe, “Highly efficient green phosphorescent single-layered organic light-emitting

devices,” Applied Physics Letters, vol. 89, 2006.

[33] S. S. Oh, J. H. Park, S. W. Kim, and B. Park, “Enhanced light emission

from phosphorescent single-layered organic light-emitting devices doped with

44
 ionic salt by simultaneous thermal and electrical annealing,” Journal of Applied

Physics, vol. 102, 2007.

[34] B. Park, M. Y. Han, and S. S. Oh, “Solution processable ionic p-i-n phospho-

rescent organic light-emitting diodes,” Applied Physics Letters, vol. 93, 2008.

[35] X. H. Yang and D. Neher, “Polymer electrophosphorescence devices with high

power conversion efficiencies,” Applied Physics Letters, vol. 84, no. 14, pp. 2476–

2478, 2004.

[36] S. R. Forrest, D. D. C. Bradley, and M. E. Thompson, “Measuring the efficiency

of organic light-emitting devices,” Advanced Materials, vol. 15, no. 13, pp. 1043–

1048, 2003.

[37] H. Liu, J. He, P. Wang, H. Xie, X. Zhang, C. Lee, B. Sun, and Y. Xia, “High-

efficiency polymer electrophosphorescent diodes based on an ir (iii) complex,”

Applied Physics Letters, vol. 87, 2005.

45