Improved efficiency in blue phosphorescent organic light-emitting devices using host materials of lower triplet energy than the phosphorescent blue emitter

Article abstract of DOI:10.1002/adfm.201100586

Data from a series of phosphorescent blue organic light-emitting devices with emissive layers consisting of either 4,4′-bis(N-carbazolyl)-2, 2′-biphenyl (CBP):6% bis[(4,6-difluorophenyl)pyridinato-N,C ²](picolinato)iridium(III) (FIrpic) or bis(

Full text of DOI:10.1002/adfm.201100586

www.afm-journal.de
www.MaterialsViews.com
Improved Efficiency in Blue Phosphorescent Organic Light-
Emitting Devices Using Host Materials of Lower Triplet
Energy than the Phosphorescent Blue Emitter
James S. Swensen, Evgueni Polikarpov, Amber Von Ruden, Liang Wang,
Linda S. Sapochak, and Asanga B. Padmaperuma*
In addition to solid-state lighting, OLEDs
have shown important applications in dis-
plays. Recently, key industrial players have
demonstrated the developments of phos-
phorescent OLEDs for large-area flexible
displays manufactured by cost-effective
approaches such as ink-jet printing. Effi-
ciency and stability are two central fac-
tors fundamental to commercialization
of OLEDs for both display and lighting
purposes. Devices that utilize organome-
tallic phosphors can reach 100% internal
quantum efficiency (IQE) by harvesting
both singlet and triplet excitons[1] and,
thus, have the potential to significantly
exceed the efficiency of current lighting
technology. To date, efficient and stable red
and green phosphorescent OLEDs have
already been demonstrated.[2–6] Recent
results have been reported showing blue
phosphorescent OLEDs reaching the
high efficiency levels of their green and
red counterparts.[7–12] In the continued
development of OLEDs for displays and
lighting, new host materials that provide
long lifetime in conjunction with high
efficiency for blue phosphorescent OLEDs
Data from a series of phosphorescent blue organic light-emitting devices
with emissive layers consisting of either 4,4′-bis(N-carbazolyl)-2,2′-biphenyl
(CBP):6% bis[(4,6-difluorophenyl)pyridinato-N,C2](picolinato)iridium(III)
(FIrpic) or bis(9-carbazolyl)benzene (mCP):6% FIrpic show that the triplet
energy of the hole and electron transport layers can have a larger influence on
the external quantum efficiency of an operating device than the triplet energy
of the host material. A maximum external quantum efficiency of 14% was
obtained for CBP:6% FIrpic devices which is nearly double all other published
CBP:6% FIrpic results. A new host material, 4-(diphenylphosphoryl)-N,N-
di-p -tolylaniline (DHM-A2), which has a triplet energy lower than that of
FIrpic is also reported. Devices fabricated using DHM-A2 show improved
performance (lower drive voltage and higher external quantum efficiency)
over devices using 4-(diphenylphosphoryl)-N,N-diphenylaniline (HM-A1), a
high performance ambipolar DHM-A2 analogue with a triplet energy greater
than FIrpic. Nearly 18% external quantum efficiency was obtained for the
DHM-A2:5% FIrpic devices. The results suggest modified design rules for the
development of high performance host materials: more focus can be placed
on molecular structures that provide good charge transport (ambipolarity for
charge balance) and good molecular stability (for long lifetimes) rather than
first focusing on the triplet energy of the host material.
1. Introduction
are a key need that remains unfulfilled.
The emissive layer (EML) of phosphorescent OLEDs is a
host–guest system because of strong emitter self-quenching by
the phosphorescent dopant in a neat film.[13] One of the path-
ways for quenching of triplet emission occurs when other mol-
ecules within close proximity to the phosphorescent emitter
(such as the host material, the hole transport material, and/or
the electron transport material) possess triplet energy (ET) levels
that are lower than that of the emitter. When such a material is
present, non-radiative energy transfer of triplet excitons from
the emitter onto the lower-lying triplet energy level(s) of these
nearby molecules occurs.[14–17] Below, we will demonstrate that
the triplet energies of the transport layer materials in imme-
diate contact with the emissive layer play a more important role
in maintaining high efficiency in phosphorescent OLEDs than
the triplet energy of the host material.
Phosphorescent organic light-emitting devices (OLEDs) are a
promising solid-state lighting technology due to their potential
for low manufacturing costs and reduced power consumption.
Dr. J. S. Swensen, Dr. E. Polikarpov, A. Von Ruden, Dr. L. Wang,
Dr. A. B. Padmaperuma
Energy and Environment Directorate
Pacific Northwest National Laboratory
902 Battelle Blvd, P.O. Box 999
MSIN K3–59, Richland, WA 99354, USA
E-mail: asanga.padmaperuma@pnl.gov
Dr. L. S. Sapochak
Solid State and Materials Chemistry Program
Division of Materials Research Mathematical
and Physical Sciences Directorate
One of the initial reports demonstrating a blue phosphores-
cent OLED used 4,4′-bis(N-carbazolyl)-2,2′-biphenyl (CBP) as the
host and bis[(4,6-difluorophenyl)pyridinato-N,C2](picolinato)-
National Science Foundation Arlington, VA 22230
DOI: 10.1002/adfm.201100586
©
3250 wileyonlinelibrary.com
2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Funct. Mater. 2011, 21, 3250–3258

www.afm-journal.de
www.MaterialsViews.com
iridium(III) (FIrpic) as the phosphorescent emitter.[14] It was
shown that CBP had a lower lying triplet energy level than
FIrpic, resulting in triplet exciton transfer from FIrpic to
CBP. These devices had a maximum external quantum effi-
ciency (EQE) of ∼6%. To improve the emission efficiency,
high triplet energy hosts were developed and compared to
CBP:FIrpic based OLEDs: bis(9-carbazolyl)benzene (mCP),[15]
4.4′-bis(9-carbazolyl)-2,2′-dimethyl-biphenyl (CDBP),[18,19] and
p-bis(triphenylsilyly)benzene (UGH2).[20,21] OLEDs with the
new, higher triplet energy hosts had EQEs close to 12%, dem-
onstrating that EQE indeed increased when a higher triplet
energy host was used.
In this work, we show that high EQEs can be achieved
using CBP as the host material for the blue phosphorescent
dopant FIrpic. Many reports attribute the poor performance of
CBP:FIrpic OLEDs solely to triplet exciton transfer from FIrpic
to CBP.[7,14–19] Considering an accepted value for the light out-
coupling efficiency of 22% and the 78% PLQY reported by
Kawamura et al.,[16] a maximum EQE of ∼16% is estimated for
CBP:6% FIrpic OLEDs. The published EQE (∼6%) results for
CBP:6% FIrpic devices[14,15,18] do not come close to reaching
this calculated EQE maximum of 16%. It would seem, then,
that transfer of triplet excitons from FIrpic to CBP is not
the only cause for the low EQEs of CBP:6%:FIrpic OLEDs
reported in the literature. In this work, through the utilization
of transport layers that confine both the charge and the triplet
exciton within the emissive layer, we demonstrate CBP:6%
FIrpic based OLEDs with EQEs of nearly 14%. Additionally
we report a new host material 4-(diphenylphosphoryl)-N,N-
di-p-tolylaniline (DHM-A2) with a triplet energy lower than
FIrpic that performs better than either its high triplet energy
analog 4-(diphenylphosphoryl)-N,N-diphenylaniline (HM-A1)
or CBP.
As explained by Adachi et al.,[2] the external quantum effi-
ciency is proportional to the photoluminescent quantum yield
(PLQY) of the emitter. Kawamura et al.[16] showed a maximum
PLQY of 78% for a thin film consisting of CBP:6% FIrpic and
99% for a thin film consisting of mCP:6% FIrpic, further rein-
forcing the accepted view that the host must have a higher triplet
energy than the phosphorescent emitter. This PLQY data shows
that a host material with a higher triplet energy than FIrpic will
have a higher PLQY due to the reduction of quenching path-
ways available to the excited state of FIrpic and that higher
PLQY may correspond to higher EQE in a blue OLED.
Significant research into the development of new wide
bandgap host materials with high triplet energies followed.
As a result, new classes of high triplet energy hosts such as
phosphine oxides developed in our labs,[22–26] 2,2’-bis(p-carbazol-
9-yl-phenyl)-1,1′-biphenyl (4CzPBP),[7–9] 3,5-bis(9-carbozolyl)tet-
raphenylsilane (SimCP),[27,28] and 1,2-trans-di-9-carbazolylcyclob-
utane (DCz)[29] were developed for use in phosphorescent blue
OLEDs. The near exclusive focus on the triplet energy of the
host material came at the expense of other important material
parameters that play a crucial role in obtaining either high EQE
or other desirable device characteristics such as low drive voltage
and long lifetime. For example, charge injection into these wide
bandgap, high triplet energy hosts can be difficult, resulting in
higher drive voltages and lower power efficiencies.[20,30] Addi-
tionally, charge transport in some wide bandgap, high triplet
energy hosts is poor or unipolar, leading to lower EQEs, higher
drive voltages, and likely shorter lifetimes.[11,20,21,27,31] Often the
triplet energy of these wide bandgap host materials is greater
than or equal to the triplet energy of the electron transport layer
(ETL) or hole transport layer (HTL). In this case, triplet excitons
formed on the host material are not confined within the EML
and may diffuse from the EML host into the ETL and/or HTL
and be lost to non-radiative decay pathways.
The maximum EQE estimations mentioned below are based
on a simple light extraction efficiency model with no outcoup-
ling enhancement based on the relationship ηext = ηint * ηout. ηext
is the external quantum efficiency, ηint is the percentage of gen-
erated excitons which decay radiatively in the emissive layer
and can be estimated by PLQY (a maximum of 25% for fluo-
rescent emitters and a maximum of 100% for phosphorescent
emitters), and ηout is the light outcoupling efficiency. For this
OLED architecture, the ηout can be estimated by 1/2n2 where n
is the refractive index of the glass substrate (∼1.5 to 1.6). Given
a maximum ηint of 100% for phosphorescent emitters and ηout
of 20% to 22%, the maximum EQE for this OLED architecture
is ∼20% to 22%.[2,9,32,33]
2. Results and Discussion
2.1. CBP:FIrpic and mCP:FIrpic based OLEDs
We used CBP and mCP as the host materials because of their
similarity in molecular structure yet different triplet energies,
making them suitable for a comparative study when FIrpic is
used as the blue phosphorescent dopant.[15] The ETLs used in
this work were 2,8-bis(dihenylphoshporyl)dibenzo[b,d]furan
(PO14),[26] bis(dihenylphoshporyl)biphenyl (PO1),[34] and
2,9-dimethyl-4,7-diphenyl-1,10-phenanthroline (BCP). All three
ETLs have hole blocking properties for CBP and mCP but have
different exciton blocking properties spanning a triplet energy
range higher than, equal to, and lower than that of FIrpic. The
HTLs used in this work that were in contact with the emis-
sive layer, 4,4′,4′′-tris(carbazol-9-yl)triphenylamine (TCTA)
and 1,1′-bis[(di-4-tolylamino)phenyl]cyclohexane (TAPC), have
adequate electron blocking and exciton blocking properties
for both CBP:FIrpic and mCP:FIrpic systems. In this way, a
systematic study can be performed to understand the role of
exciton confinement in an OLED with hosts having different
triplet energies. The energy of the highest occupied molecular
orbital (EHOMO), energy of the lowest unoccupied molecular
orbital (ELUMO), and the energy of the triplet excited state for
materials used in the study are given in Table 1. See Experi-
mental Section and Supporting information for further details.
Figure 1 shows the molecular structure for the materials used
in this paper.
Plots of current density vs. voltage (J–V) and EQE vs. cur-
rent density (EQE-J) are shown in Figure 2 for OLEDs with
the structure of glass/ITO/100 Å CuPc/300 Å α-NPD/50 Å
TCTA/300 Å EML/400 Å ETL/10 Å LiF/1000 Å Al where EML =
CBP:6% FIrpic or mCP:6% FIrpic and ETL = BCP, PO1 or
PO14. The HTL, EML, and ETL thicknesses are identical to
those reported by Holmes et al.[15] The only exceptions in our
©
wileyonlinelibrary.com
Adv. Funct. Mater. 2011, 21, 3250–3258
2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3251

www.afm-journal.de
www.MaterialsViews.com
T a b le 1 . The ELUMO, EHOMO, and triplet energy for HTL, EML, and ETL
materials used to fabricate the OLEDs reported on in this paper. ELUMO
and EHOMO are estimated from cyclic voltammetry and optical data. Tri-
plet energies were obtained from low temperature emission spectra.
material N,N′-Bis(naphthalen-1-yl)-N,N′-bis(phenyl)-benzidine
(α-NPD, ET = 2.3 eV) and the EML to form a triplet exciton
barrier between the EML and α-NPD. An additional differ-
ence between this work to the work of Holmes et al. was not
using the 4-biphenyloxolato aluminum(III)bis(2-methyl-8-
quinolinato)4-phenylphenolate (BAlq) as the ETL because it has
poor hole blocking. Since all of the HTLs and ETLs used in this
work provide charge blocking properties, we are able to ana-
lyze the results solely based on differences in triplet energies.
By comparing the triplet energy of FIrpic to the triplet ener-
gies of the all the materials used in the devices in Figure 2 (see
Table 1), triplet excitons on FIrpic are only allowed to transfer to
CBP (host), BCP (ETL), and/or possibly to PO1 (ETL). All of the
rest of the materials have higher triplet energies than FIrpic.
In Figure 2a–c, an OLED with a CBP or a mCP host and
the same ETL have nearly identical J–V results indicating
that CBP and mCP have similar charge transport proper-
ties. Hence, the differences in the EQE shown in Figure 2d
are attributed to the triplet energy of the host and/or the ETL
rather than the differences in charge transport properties of
Material
PO14
PO1
EHOMO [eV]
−6.5
ELUMO [eV]
−2.5
Triplet Energy [eV]
3.1
2.7
2.6
2.8
2.9
2.6
2.9
2.5
2.7
−7.2
−2.9
BCP
−6.4
−3.0
TCTA
−5.2
−1.7
TAPC
−5.1
−1.7
CBP
−6.2
−2.7
mCP
−5.8
−2.5
Ir(ppy)2acac
FIrpic
−5.1
−2.9
−5.8
−3.2
device structure is the insertion of a thin layer of a higher triplet
energy material, TCTA, between the lower triplet energy HTL
O
O
P
P
N
N
N
N
O
BCP
PO14
TAPC
N
N
N
N
O
Ir
F
O
N
N
CBP
F
2
FIrpic
mCP
O
P
P O
O
Ir
PO1
O
N
N
N
2
Ir(ppy)2acac
NPD
P
N
O
HM-A1
N
N
N
N
N
N
Cu
N
N
N
N
P O
N
N
N
DHM-A2
CuPc
TCTA
Figure 1. Molecular structure for the materials used to fabricate OLEDs.
©
3252 wileyonlinelibrary.com
2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Funct. Mater. 2011, 21, 3250–3258

www.afm-journal.de
www.MaterialsViews.com
(a)
102
101
100
10-1
10-2
10-3
10-4
10-5
10-6
(c)
102
101
100
10-1
10-2
10-3
10-4
10-5
10-6
10-7
mCP - BCP
CBP - BCP
mCP - PO14
CBP - PO14
10-7
0
2
4
6
8
10
0
2
4
6
8
10
(b)
(d)
Volts /V
Volts /V
102
101
14
12
10
8
100
10-1
10-2
10-3
10-4
10-5
10-6
6
4
mCP - PO1
CBP - PO1
2
10-7
0
0
10-2
10-1
100
101
102
2
4
6
8
10
Current Density /mA cm-2
Volts /V
Figure 2. a–c) Current density vs. voltage and d) EQE (%) vs. current density for OLEDs fabricated with the following device architecture: Glass/
ITO/100 Å CuPc/300 Å α-NPD/50 Å TCTA/300 Å EML/400 Å ETL/10 Å LiF/1000 Å Al [EML = mCP:6% FIrpic (solid) or CBP:6% FIrpic (hollow); ETL =
BCP (square), PO1 (circle), or PO14 (triangle)].
2.2. CBP:FIrpic and CBP:Ir(ppy)2acac based OLEDs
the different hosts. Interestingly, EQE as a function of current
density is more strongly influenced by the ETL rather than by
the host material. As seen in Figure 2d and shown quantita-
tively in Table 2, the EQE is reduced by 73% for a CBP:FIrpic
based OLED at 1 mA cm−2 when going from the highest tri-
plet energy ETL PO14 to the lowest triplet energy ETL BCP.
This large drop in EQE is significantly greater than the differ-
ence between mCP:FIrpic or CBP:FIrpic based devices when
the same ETL is used. These same trends hold at all current
densities.
Based on triplet energy, mCP:FIrpic devices are expected
to have higher EQE than CBP:FIrpic devices because of the
higher triplet energy of mCP.[14–19] A more recent publica-
tion continues to attribute the low EQE reported in the pub-
lished literature on CBP:FIrpic based OLEDs to the lower tri-
plet energy of the host.[7] Contrary to published results, our
data show that high EQEs can be obtained for OLEDs with
CBP:FIrpic based EMLs when coupled with proper HTL and
ETL materials. Our results demonstrate that the triplet energy
of the ETL has a larger effect on the magnitude of the EQE
than the triplet energy of the host material. In fact, when the
highest triplet energy ETL PO14 is used, both the mCP and
CBP based devices have a maximum EQE of ∼13.5%.
In order to verify that the EQE dependence on the ETL seen
in Figure 2 is in fact due to the triplet energy of the ETLs and
not charge transport behavior, devices were fabricated to com-
pare the high triplet energy blue phosphorescent dopant FIrpic
to the lower triplet energy green phosphorescent dopant bis(2-
phenylpyridine)(acetylacetonato)iridium(III)
(Ir(ppy)2acac)
(Figure 3). CBP was used as the host material for both dopants
in the EML. The device structure was glass/ITO/450 Å
TAPC/300 Å CBP:6% FIrpic or CBP:6% Ir(ppy)2 acac/400 Å
ETL/10 Å LiF/1000 Å Al where ETL is BCP, PO1, or PO14,
keeping the device architecture (i.e., HTL, EML, and ETL thick-
nesses) the same as the devices reported on in Figure 2.
All of the ETLs used in the devices in Figure 3 provide effec-
tive exciton blocking for Ir(ppy)2acac, but, as discussed earlier
and shown in Table 1, the ETLs vary in their exciton blocking
ability with regards to FIrpic. For FIrpic based devices, the data
in Figure 3c shows the same EQE dependence as a function of
ETL as was shown in Figure 2d, namely that EQE is very sensitive
to the triplet energy of the ETL. The CBP:6% Ir(ppy)2acac data
in Figure 3c, however, show that EQE is much less dependent
on the triplet energy of the ETL at all current densities. When
©
wileyonlinelibrary.com
Adv. Funct. Mater. 2011, 21, 3250–3258
2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3253

www.afm-journal.de
www.MaterialsViews.com
T a b le 2 . Extracted EQE (%) from the OLED data shown in Figure 1 at 1 and 100 mA cm−2 . The change in EQE (%) for the different ETL and host
materials is shown at these current densities.
External Quantum Efficiency (%)
100 mA cm−2
% Change (mCP to CBP)
1 mA cm−2
1 mA cm−2
100 mA cm−2
ETL
mCP
13.5
9.1
CBP
12.5
7.6
mCP
7.2
CBP
4.6
PO14
−7%
−76%
−79%
−36%
−37%
−35%
PO1
3.6
2.5
BCP
4.2
3.4
3.1
2.0
% Change (PO14 to PO1)
% Change (PO14 to BCP)
−32%
−69%
−39%
−73%
−50%
−56%
−45%
−55%
PO14 and PO1 are used as the ETL in Ir(ppy)2acac devices, the
EQE-J results are essentially identical, which can be explained
by the fact that their triplet energies are both much greater
than that of Ir(ppy)2acac and thus no triplet excitons escape the
EML to the ETL. We expect the CBP:Ir(ppy)2acac OLEDs with a
BCP ETL to have similar EQE-J results as well because the tri-
plet energy of BCP is also higher than Ir(ppy)2acac. As seen in
Figure 3c, however, this was not the case.
A comparison of the J–V results in Figure 3b to the corre-
sponding EQE-J results in Figure 3c for the CBP:Ir(ppy)2acac
OLEDs with different ETLs may help to explain the lower EQE
in the OLED with a BCP ETL. At low current densities, a higher
voltage is required for BCP devices to achieve the same amount
of current density as the PO1 and PO14 devices. The device
stack for all the blue or green OLEDs is identical except for the
material used as the ETL. The high drive voltage is therefore
(a)
102
101
100
10-1
10-2
10-3
10-4
10-5
10-6
10-7
(c)
16
14
12
10
8
6
4
2
0
10-2
10-1
100
101
102
0
2
4
6
8
10
Current Density /mA cm-2
(b)
Volts /V
102
101
FIrpic - BCP2
100
10-1
10-2
10-3
10-4
10-5
10-6
10-7
FIrpic - PO12
10
FIrpic - PO142
Ir(ppy)2acac - BCP
Ir(ppy)2acac - PO1
Ir(ppy)2acac - PO14
0
0
2
4
6
8
10
10-2
10-1
100
101
102
Volts /V
Figure 3. a,b) Current density vs. voltage and c) EQE (%) vs. current density for OLEDs fabricated with the following device architecture: Glass/
ITO/450 Å TAPC/300 Å EML/400 Å ETL/10 Å LiF/1000 Å Al (EML = CBP:6% FIrpic (solid) or CBP:6% Ir(ppy)2acac (hollow); ETL = BCP (square), PO1
(circle), and PO14 (triangle)).
©
3254 wileyonlinelibrary.com
2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Funct. Mater. 2011, 21, 3250–3258

www.afm-journal.de
www.MaterialsViews.com
due to a barrier for electron injection from BCP into the host
and/or a lower BCP electron mobility in comparison to PO1
and PO14. As a result, a lower electron concentration exists in
the emissive layer at a given current density when BCP is the
ETL, leading to poor charge balance and lower EQE. Surpris-
ingly the EQE of BCP devices increases slightly with current
density up to ∼5 mA cm−2 . This rise in EQE could be attrib-
uted to the improvement of electron injection from BCP into
the host upon increasing electric field. At high current densities
(>5 mA cm−2 ), the J–V of the PO1 devices slowly drops to meet
the level of BCP devices. Correspondingly, the EQE of the PO1
devices rolls off to the level of the BCP devices. Geibink et al.[35]
showed that peak EQE and EQE roll-off are related to charge
balance in the current densities range at which these devices
are operating. The BCP devices start out with poorer charge
balance and therefore have a lower overall EQE at low current
densities. As the current density in the PO1 devices drops from
the level of the PO14 devices to the level of the BCP devices, the
EQE does the same. The charge balance in the PO1 devices is
likely changing from better (PO14) to worse (BCP) causing the
steeper roll-off in EQE.
We reason that in the case of a host:guest system such as
CBP:FIrpic, where the host has a slightly lower triplet energy
than the guest, energy transfer does occur from the FIrpic to
the lower lying triplet levels of CBP. However, as Adachi et al.[14]
showed, endothermic transfer of the triplet exciton from CBP
back to FIrpic takes place at operating temperatures. Since there
is a pathway for the triplet exciton to get back to the FIrpic and
decay radiatively, significant quenching does not occur if the
CBP:FIrpic EML is surrounded by high triplet energy transport
layers as shown in our data here and the work by Su et al.[8]
We further reason that when a transport layer has lower triplet
energy than the phosphorescent emitter, significant quenching
does occur. The exciton diffusion length is estimated to range
from 280 to 350 nm in CBP:FIrpic films.[36] This diffusion
length is larger than the EML thickness for nearly all OLEDs.
Hence, an exciton generated in any portion of the EML will
likely come in contact with the HTL/EML and/or the EML/
ETL interface(s). If the triplet energy of the HTL and/or ETL is
lower than the triplet energy of either the host or the phospho-
rescent emitter in the EML, then it is energetically favorable for
the triplet exciton to fall from the EML to the lower lying triplet
energy level(s) of the transport layer(s). When this occurs, the
exciton no longer resides in the EML where the phosphorescent
dopant is present. The triplet exciton has a higher probability
to diffuse away from the EML driven by the gradient of exciton
concentration and to be lost to non-radiative pathways than to
be endothermically excited back into the EML where photon
emission can occur. Thus, the triplet energy of the transport
layers is more important than the triplet energy of the host.
Additionally, Lee et al.[30] use CBP as the HTL interfacing
with an UGH:FIrpic EML. In spite of the sufficiently high
triplet energy of the host in the EML, the EQE is reduced by
∼85% (from 7% to below 1% EQE) in comparison to devices
employing an HTL with triplet energy higher than FIrpic.
These results suggest that when CBP is used as an HTL
rather than as a doped host, there is no endothermic interac-
tion between FIrpic and CBP as is the case when CBP is doped
with FIrpic. Instead, when the triplet excitons are transferred
from FIrpic to the lower lying triplet energy level of the CBP
HTL, they are lost to non-radiative decay pathways. We show
that when CBP:FIrpic EMLs are used in conjunction with high
triplet energy HTLs and ETLs, such drastic quenching is not
observed in spite of the lower triplet energy of the CBP in the
EML. In our system, the exciton is confined within the EML
due to effective exciton blocking properties of the HTL and ETL
allowing it to be endothermically excited from the CBP back to
FIrpic where it can still radiatively decay to the ground state.
2.3. HM-A1:FIrpic and DHM-A2:FIrpic Based OLEDs
We thereby suggest that for developing next generation host
materials, it is crucial to consider multiple properties of the
host materials such as charge transport and stability, rather
than focusing exclusively on the triplet energy. To further prove
this new perspective, we evaluated a new host material, DHM-
A2, with a triplet energy lower than that of the emitter FIrpic.
DHM-A2 was synthesized as described in the experimental sec-
tion. Low temperature delayed luminescence spectra were col-
lected at 77 K to estimate the triplet energy levels of HM-A1 (ET =
2.8 eV), FIrpic (ET = 2.7 eV), and DHM-A2 (ET = 2.6 eV) as
described by Turro[37] (see Figure 4). The EHOMO and the ELUMO
of DHM-A2 were estimated to be –5.37 and –2.46 eV respec-
tively from solution electrochecmial studies (see Supporting
Information).
Figure 5 shows experimental results comparing OLEDs
fabricated with the new host DHM-A2 to OLEDs with its
analogue HM-A1. HM-A1 is an ambipolar host with a triplet
energy (see Figure 4) greater than that of FIrpic and has been
reported on by us previously.[25] The device architecture was
glass/ITO/350 Å PEDOT:PSS/350 Å TAPC/150 Å Host:5%
FIrpic/500 Å PO15/10 Å LiF/1000 Å Al where Host is HM-A1 or
DHM-A2. At 1 mA cm−2 , the drive voltage for DHM-A2 OLEDs
is 0.4 V lower than HM-A1 devices. In fact, the drive voltage is
lower at all current densities for the DHM-A2 OLEDs, which
suggests that DHM-A2 has better charge transport properties
than HM-A1 (Figure 5a). Significantly, in Figure 5b the OLEDs
1.0
HM-A1
FIrpic
DHM-A2
0.8
0.6
0.4
0.2
0.0
3.6
3.2
2.8
2.4
2.0
Emission Energy /eV
Figure 4. Normalized delayed luminescence spectra for HM-A1 (solid),
FIrpic (dashed), and DHM-A2 (dotted) measured in dichloromethane at
77 K.
©
wileyonlinelibrary.com
Adv. Funct. Mater. 2011, 21, 3250–3258
2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3255

www.afm-journal.de
www.MaterialsViews.com
that of the HM-A1 based OLEDs as well. Figure 5c shows the
electroluminescence (EL) spectra for the DHM-A2 and HM-A1
OLEDs. Both devices show nearly identical EL with the emis-
sion coming solely from FIrpic. Even though DHM-A2 has a
lower triplet energy than FIrpic, no emission from DHM-A2 is
observed in the EL spectrum. Thus, by not excluding DHM-A2
as a host just because its triplet energy is lower than FIrpic’s,
we obtained better device performance (lower drive votage
and higher brightness) than using the high triplet energy host
HM-A1.
(a)
102
100
10-2
10-4
10-6
DHM-A2
HM-A1
3. Conclusions
10-8
0
In summary, we demonstrate that CBP:FIrpic devices can
achieve 14% EQE when used in conjunction with high tri-
plet energy transport layer materials. High performance blue
phosphorescent OLEDs (3.75 V, 16.5% EQE, and 440 cd/m2 at
1 mA cm−2 ) were demonstrated using the new host material
DHM-A2, which has a triplet energy lower than that of FIrpic.
Our results illustrate that host materials with triplet energies
equal to or slightly lower than the phosphorescent dopant can
be used to achieve high EQEs in operating OLEDs if coupled
with ETLs and HTLs of sufficiently high triplet energy and
appropriate blocking properties. These results emphasize the
fact that the design of host materials does not need to adhere
strictly to the high triplet energy and wide bandgap rule. We pro-
pose that, when designing hosts for OLEDs, more focus should
be placed on molecular structures that reduce drive voltage and
lengthen operating lifetime rather than focus solely on high tri-
plet energy. By allowing the host to have triplet energy equal to
or slightly lower than the phosphorescent emitter, a wider range
of host molecular structures which were previously undesirable
under the much-emphasized high triplet energy paradigm can
be explored.
2
4
6
Volts /V
(b)
18
16
14
12
10
8
50
40
30
20
10
0
6
4
DHM-A2
HM-A1
2
0
10-2
10-1
100
101
102
Current Density /mA cm-2
(c)
1.0
DHM-A2
HM-A1
0.8
0.6
0.4
0.2
0.0
4. Experimental Section
Synthesis of 4-(Diphenylphosphoryl)-N,N-di-p-tolylaniline (DHM-A2):
All chemicals used in synthetic procedures were obtained from Aldrich
Chemical Co. and used as received unless noted otherwise. NMR
spectra were obtained using a Varian Oxford 500 MHz spectrometer
at the following frequencies: 499.8 MHz (1H), 202.3 MHz (31P), and
125.7 MHz (13C). Tetramethylsilane was used as an internal reference
1
for H and 13C spectra, and the 31P signals were externally referenced to
350 400 450 500 550 600 650 700
85% H3PO4.
The synthesis of DHM-A2 was accomplished using a modified
literature procedure and is shown in Scheme 1 (see Supporting
Information).[38] The starting bromophenyl amine 1 was prepared from a
commercially available amine (TCI America) and 1,4-bromoiodobenzene
using Cu coupling conditions via a literature procedure.[39]
To a 3 neck round bottom flask was added 1 (8.0 g, 0.0228 mol,
1.0 eq) and the flask flushed with argon. Anhydrous tetrahydrofuran
(100 mL) was added and the solution was cooled to −78 °C (dry
ice: acetone bath). N-butyllithium (9.6 mL, 0.0239 mol, 2.5 M in
hexanes, 1.05 eq) was added dropwise maintaining the temperature
below −60 °C. The resulting dark solution was stirred cold for 2 h.
Chlorodiphenylphosphine (4.4 mL, 0.0239 mol, 1.05 eq) was added to
give a pale yellow solution. The reaction mixture was allowed to warm
to room temperature and stirred overnight. The reaction mixture was
quenched with methanol and solvents were removed to give a thick
Wavelength /nm
Figure 5. a) Current density vs. voltage and b) EQE (solid) and power
efficiency (hollow) vs. current density for OLEDs fabricated with the
following device architecture: Glass/ITO/350 Å PEDOT:PSS/350 Å
TAPC/150 Å Host:5% FIrpic/500 Å PO15/10 Å LiF/1000 Å Al (Host =
DHM-A2 (triangles) or HM-A1 (squares)). c) Electroluminescent spectra
for OLEDs using the hosts DHM-A2 (dashed) and HM-A1(solid).
with DHM-A2 have higher EQE than the OLEDs with HM-A1
despite the fact that the triplet energy of DHM-A2 is lower
than that of FIrpic. Due to the lower drive voltage and higher
brightness, power efficiency for the DHM-A2 devices exceeds
©
3256 wileyonlinelibrary.com
2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Funct. Mater. 2011, 21, 3250–3258

www.afm-journal.de
www.MaterialsViews.com
orange oil. The crude diphosphine was loaded onto a SiO2 plug using
1:20 ethyl acetate:hexanes, The phosphine was collected and solvents
were removed to give a white foam. The foam was dissolved in 100 mL
of methylene chloride in a flask affixed with a reflux condenser. To the
stirred solution, 10 mL of 50% hydrogen peroxide was added slowly
to prevent excessive foaming and bumping. The resulting mixture was
stirred overnight at room temperature. The organic layer was washed
with water and brine, and the extract was evaporated to dryness to afford
a light yellow foam which was purified by column chromatography (Rf =
0.34, SiO2, hexanes/methanol/ethyl acetate = 6:1:2) to yield 3.93 g, (36%
over last two steps) of DHM-A2. 1H NMR (500 MHz, CDCl3, δ): 2.31 (s,
6H), 6.94 (dd, J = 9.0 Hz, 2.5 Hz, 2H), 7.03 (d, J = 8.5 Hz, 4H), 7.10 (d, J =
8.0 Hz, 4H), 7.36–7.40 (m, 2H), 7.44 (td, J = 8.0 Hz, 2.5 Hz, 4H), 7.51
(t, J = 6.8 Hz, 2H), 7.67–7.71 (m, 4H); 13C NMR (125 MHz, CDCl3, δ):
Acknowledgements
This project was funded by the Solid State Lighting Program of the U.S.
Department of Energy (US DOE), within the Building Technologies
Program (BT) (Award No. M6743231, managed by the National Energy
Technology Laboratory (NETL)). A portion of this research was performed
using EMSL, which is a national scientific user facility sponsored by the
Department of Energy’s Office of Biological and Environmental Research
and is located at Pacific Northwest National Laboratory (PNNL). PNNL
is operated by Battelle Memorial Institute for the U.S. DOE (under
Contract DE_AC06–76RLO 1830). This material was based on work
supported by the National Science Foundation, while working at the
foundation. Any opinion, finding, and conclusions or recommendations
expressed in this material are those of the author and do not necessarily
reflect the views of the National Science Foundation.
21.1, 119.1 (JCP = 12.8 Hz), 122.0 (JCP = 111 Hz), 126.1, 128.5 (JCP
12.1 Hz), 130.3, 131.8 (JCP = 2.0 Hz), 132.2 (JCP = 10.8 Hz), 133.3 (JCP
=
=
103.6 Hz), 133.3 (JCP = 11.3 Hz), 134.4, 144.1, 151.6 (JCP = 2.6 Hz); 31P
NMR (200 MHz, CDCl3, δ): 29.8.
Received: March 15, 2011
Published online: July 4, 2011
Electrochemical and UV–Visible Measurements: Electrochemical
measurements were carried out using Princeton Applied Research
potentiostat model 263A with a silver wire as the pseudo-reference
electrode, a Pt wire as the counter electrode, and glassy carbon as the
working electrode. N,N-dimethylformamide was used as a solvent
for reduction potential measurements. Oxidation potentials were
measured in dichloromethane distilled over CaH2. Tetrabutylammonium
hexafluorophosphate was used as a supporting electrolyte. UV-visible
absorption spectra were collected on a Varian Cary 5 UV–vis–NIR
spectrophotometer. LUMO energy levels were estimated from solution
electrochemical reduction potentials. Highest occupied molecular
orbital HOMO levels were estimated from electrochemical oxidation
potentials or the optical band gap where electrochemical measurements
were not feasible. The boundary orbital energy values were derived from
the electrochemical data according to the published methods.[40,41] See
Supporting Information for cyclic voltammograms of HM-A1 and DHM-A2
as well as further discussion on estimating the HOMO and LUMO levels.
OLED Fabrication: Glass/ITO substrates were cleaned by sonication
in a sequential series of solvents, including a dilute Tergitol solution,
de-ionized water, trichloroethane, acetone, and 2-propanol. The substrates
were then dried with flowing nitrogen. As a final cleaning step before
device fabrication, the substrates were treated with UV ozone (UVO-
Cleaner, Jelight Co., Inc.) at 15 mW cm−2 for 15 min. The substrates were
then loaded into a nitrogen glove box (<1.5 ppm O2, –79 °C dew point)
coupled to a multichamber vacuum deposition system. Organic layers
were sequentially deposited onto the glass/ITO substrates by thermal
evaporation from tantalum boats in a high vacuum chamber with a
base pressure below 3 × 10−7 Torr. Cathodes were defined by thermally
depositing a 1 nm thick layer of LiF immediately followed by a 100 nm
thick layer of Al through a shadow mask with 1 mm diameter circular
openings. Quartz crystal oscillators were used to monitor the thicknesses
of the films, which were calibrated ex situ using ellipsometry.
[1] M. Baldo, D. O’brien, Y. You, A. Shoustikov, S. Sibley, M. Thompson,
S. Forrest, Nature 1998, 395, 151–154.
[2] C. Adachi, M. A. Baldo, M. E. Thompson, S. R. Forrest, J. Appl. Phys.
2001, 90, 5048–5051.
[3] R. C. Kwong, M. R. Nugent, L. Michalski, T. Ngo, K. Rajan, Y. J. Tung,
M. S. Weaver, T. X. Zhou, M. Hack, M. E. Thompson, S. R. Forrest,
J. J. Brown, Appl. Phys. Lett. 2002, 81, 162–164.
[4] V. I. Adamovich, M. S. Weaver, R. C. Kwong, J. J. Brown, Curr. Appl.
Phys. 2005, 5, 15–18.
[5] Y.-Y. Lyu, J. Kwak, W. S. Jeon, Y. Byun, H. S. Lee, D. Kim, C. Lee,
K. Char, Adv. Funct. Mater. 2009, 19, 420–427.
[6] C.-H. Wu, P.-I. Shih, C.-F. Shu, Y. Chi, Appl. Phys. Lett. 2008, 92,
233303.
[7] D. Tanaka, Y. Agata, T. Takeda, S. Watanabe, J. Kido, Jpn. J. Appl.
Phys., Part 2 2007, 46, L117-L119.
[8] S. J. Su, T. Chiba, T. Takeda, J. Kido, Adv. Mater. 2008, 20,
2125–2130.
[9] L. X. Xiao, S. J. Su, Y. Agata, H. L. Lan, J. Kido, Adv. Mater. 2009, 21,
1271–1274.
[10] Y. Zheng, S. H. Eom, N. Chopra, J. W. Lee, F. So, J. G. Xue, Appl.
Phys. Lett. 2008, 92, 223301.
[11] N. Chopra, J. Lee, Y. Zheng, S.-H. Eom, J. Xue, F. So, ACS Appl.
Mater. Interfaces 2009, 1, 1169–1172.
[12] H. H. Chou, C. H. Cheng, Adv. Mater. 2010, 22, 2468–2471.
[13] M. A. Baldo, M. E. Thompson, S. R. Forrest, Pure Appl. Chem. 1999,
71, 2095–2106.
[14] C. Adachi, R. C. Kwong, P. Djurovich, V. Adamovich, M. A. Baldo,
M. E. Thompson, S. R. Forrest, Appl. Phys. Lett. 2001, 79,
2082–2084.
[15] R. J. Holmes, S. R. Forrest, Y. J. Tung, R. C. Kwong,
J. J. Brown, S. Garon, M. E. Thompson, Appl. Phys. Lett. 2003, 82,
2422–2424.
[16] Y. Kawamura, K. Goushi, J. Brooks, J. J. Brown, H. Sasabe, C. Adachi,
Appl. Phys. Lett. 2005, 86, 071104–1-071104–3.
[17] T. Tsuboi, H. Murayama, S. J. Yeh, C. T. Chen, Opt. Mater. 2007, 29,
1299–1304.
OLED Characterization: Optical and electrical characteristics of
the devices were determined in air, with electrical contact made via
a tungsten probe tip on the ITO anode and a 0.002 in. diameter gold
wire directly probing the Al cathode. Current-voltage characteristics
were measured with an Agilent Technologies 4155B semiconductor
parameter analyzer. The light output was detected using a 1 cm2 Si
photodetector placed behind the OLED, and the device brightness was
directly measured using a Newport multifunction optical meter. No
corrections were made for light wave guided in the organic thin films
or the substrate. Electroluminescence spectra were recorded with an
Acton Research Corporation (ARC) SpectrumMM charge coupled device
detector on an ARC Spectra 150 dual grating monochromator.
[18] S. Tokito, T. Iijima, Y. Suzuri, H. Kita, T. Tsuzuki, F. Sato, Appl. Phys.
Lett. 2003, 83, 569–571.
[19] I. Tanaka, Y. Tabata, S. Tokito, Chem Phys Lett 2004, 400, 86–89.
[20] R. J. Holmes, B. W. D’Andrade, S. R. Forrest, X. Ren, J. Li,
M. E. Thompson, Appl. Phys. Lett. 2003, 83, 3818–3820.
[21] X. F. Ren, J. Li, R. J. Holmes, P. I. Djurovich, S. R. Forrest,
M. E. Thompson, Chem. Mater. 2004, 16, 4743–4747.
[22] X. Y. Cai, A. B. Padmaperuma, L. S. Sapochak, P. A. Vecchi,
P. E. Burrows, Appl. Phys. Lett. 2008, 92, 083308.
Supporting Information
Supporting Information is available from the Wiley Online Library or
from the author.
©
wileyonlinelibrary.com
Adv. Funct. Mater. 2011, 21, 3250–3258
2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3257

www.afm-journal.de
www.MaterialsViews.com
[23] L. S. Sapochak, A. B. Padmaperuma, X. Y. Cai, J. L. Male,
P. E. Burrows, J. Phys. Chem. C 2008, 112, 7989–7996.
[24] F. M. Hsu, C. H. Chien, P. I. Shih, C. F. Shu, Chem. Mater. 2009, 21,
1017–1022.
[32] N. C. Greenham, R. H. Friend, D. D. C. Bradley, Adv. Mater. 1994, 6,
491–494.
[33] J. S. Kim, P. K. H. Ho, N. C. Greenham, R. H. Friend, J. Appl. Phys.
2000, 88, 1073–1081.
[25] E. Polikarpov, J. S. Swensen, N. Chopra, F. So, A. B. Padmaperuma,
Appl. Phys. Lett. 2009, 94, 223304.
[26] L. S. Sapochak, A. B. Padmaperuma, P. A. Vecchi, X. Cai,
P. E. Burrows, Proc. SPIE 2007, 6655, 665506.
[27] M.-F. Wu, S.-J. Yeh, C.-T. Chen, H. Murayama, T. Tsuboi, W.-S. Li,
I. Chao, S.-W. Liu, J.-K. Wang, Adv. Funct. Mater. 2007, 17,
1887–1895.
[34] L. S. Sapochak, A. B. Padmaperuma, P. A. Vecchi, H. Qiao,
P. E. Burrows, Proc. SPIE 2006, 6333, 63330F.
[35] N. C. Giebink, S. R. Forrest, Phys. Rev. B 2008, 77, 235215.
[36] N. Matsusue, S. Ikame, Y. Suzuki, H. Naito, J. Appl. Phys. 2005, 97,
123512.
[37] N. J. Turro, Modern Molecular Photochemistry University of Science
Books, Sausalito, CA 1991, pp. 146–148.
[28] 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. 2005, 17, 285–289.
[29] D. R. Whang, Y. You, S. H. Kim, W. I. Jeong, Y. S. Park, J. J. Kim,
S. Y. Park, Appl. Phys. Lett. 2007, 91, 233501.
[38] R. A. Baldwin, M. T. Cheng, J. Organic Chem. 1967, 32,
1572–1577.
[39] H. B. Goodbrand, N.-X. Hu, J. Organic Chem. 1998, 64,
670–674.
[30] J. Lee, J. I. Lee, K. I. Song, S. J. Lee, H. Y. Chu, Appl. Phys. Lett. 2008,
92, 203305.
[40] D. Kolosov, V. Adamovich, P. Djurovich, M. E. Thompson, C. Adachi,
J. Am. Chem. Soc. 2002, 124, 9945–9954.
[31] J. Lee, J.-I. Lee, K.-I. Song, S. J. Lee, H. Y. Chu, Appl. Phys. Lett. 2008,
92, 133304.
[41] B. W. D’Andrade, S. Datta, S. R. Forrest, P. I. Djurovich, E. Polikarpov,
M. E. Thompson, Org. Electr. 2005, 6, 11–20.
©
3258 wileyonlinelibrary.com
2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Funct. Mater. 2011, 21, 3250–3258