RGB phosphorescent organic light-emitting diodes by using host materials with heterocyclic cores: Effect of nitrogen atom orientations

Article abstract of DOI:10.1021/cm102975d

A series of host materials 1 - 7 containing various heterocyclic cores, like pyridine, pyrimidine, and pyrazine, were developed for RGB phosphorescent organic light-emitting diodes (OLEDs). Their energy levels can be tuned by the change of heterocyclic cores and their nitrogen atom orientations, and decrease of singlet - triplet exchange energy (ΔE_ST) was achieved with introducing one or two nitrogen atoms into the central arylene; this is also consistent with density functional theory calculations. Their carrier mobilities can also be tuned by the choice of heterocyclic cores, giving improved bipolarity compared with that without any heterocyclic cores. Due to the high triplet energy level of the developed host materials, well confinement of triplet excitons of blue emitter iridium(III) bis(4,6-(difluorophenyl) pyridinato-N,C^2′) picolinate (FIrpic) was achieved except for 7 due to its low E_T. In contrast, triplet energy can be well confined on green emitter fac-tris(2-phenylpyridine) iridium (Ir(PPy)₃) and red emitter tris(1-phenylisoquinolinolato-C²,N)iridium(III) (Ir(piq)₃) for all the hosts, giving comparable lifetime (τ), photoluminescent quantum efficiency (η_PL), and radiative and nonradiative rate constants (k_r and k_nr). Highly efficient blue and green phosphorescent OLEDs were achieved for 2, exhibiting one of the highest ever efficiencies to date, especially at much brighter luminance for lighting applications. In comparison, the highest efficiencies hitherto were achieved for the red phosphorescent OLED based on 6, which can be attributed to its lower-lying LUMO level and the smallest ΔE_ST, giving improved electron injection and carrier balance. Different from the blue and green phosphorescent OLEDs based on FIrpic and Ir(PPy)₃, the host materials with lower-lying LUMO levels seem to be better hosts for a red emitter Ir(piq)₃, achieving improved efficiency and reduced efficiency roll-off at high current density.

Full text of DOI:10.1021/cm102975d

274 Chem. Mater. 2011, 23, 274–284
DOI:10.1021/cm102975d
RGB Phosphorescent Organic Light-Emitting Diodes by Using Host
Materials with Heterocyclic Cores: Effect of Nitrogen Atom Orientations
Shi-Jian Su,*,†,‡ Chao Cai,‡ and Junji Kido*,‡
†Institute of Polymer Optoelectronic Materials and Devices, Key Laboratory of Specially Functional
Materials of the Ministry of Education, South China University of Technology, Guangzhou 510640,
China, and ‡Department of Organic Device Engineering, Graduate School of Science and Engineering,
Yamagata University, 4-3-16 Jonan, Yonezawa, Yamagata 992-8510, Japan
Received October 15, 2010. Revised Manuscript Received December 5, 2010
A series of host materials 1-7 containing various heterocyclic cores, like pyridine, pyrimidine, and
pyrazine, were developed for RGB phosphorescent organic light-emitting diodes (OLEDs). Their
energy levels can be tuned by the change of heterocyclic cores and their nitrogen atom orientations,
and decrease of singlet-triplet exchange energy (ΔEST) was achieved with introducing one or two
nitrogen atoms into the central arylene; this is also consistent with density functional theory
calculations. Their carrier mobilities can also be tuned by the choice of heterocyclic cores, giving
improved bipolarity compared with that without any heterocyclic cores. Due to the high triplet
energy level of the developed host materials, well confinement of triplet excitons of blue emitter
0
iridium(III) bis(4,6-(difluorophenyl)pyridinato-N,C2 ) picolinate (FIrpic) was achieved except for 7
due to its low ET. In contrast, triplet energy can be well confined on green emitter fac-tris-
(2-phenylpyridine) iridium (Ir(PPy)3) and red emitter tris(1-phenylisoquinolinolato-C2,N)iridium-
(III) (Ir(piq)3) for all the hosts, giving comparable lifetime (τ), photoluminescent quantum efficiency
(ηPL), and radiative and nonradiative rate constants (kr and knr). Highly efficient blue and green
phosphorescent OLEDs were achieved for 2, exhibiting one of the highest ever efficiencies to date,
especially at much brighter luminance for lighting applications. In comparison, the highest
efficiencies hitherto were achieved for the red phosphorescent OLED based on 6, which can be
attributed to its lower-lying LUMO level and the smallest ΔEST, giving improved electron injection
and carrier balance. Different from the blue and green phosphorescent OLEDs based on FIrpic and
Ir(PPy)3, the host materials with lower-lying LUMO levels seem to be better hosts for a red emitter
Ir(piq)3, achieving improved efficiency and reduced efficiency roll-off at high current density.
Introduction
emitters to prevent reverse energy transfer from the guest
back to the host and to effectively confine triplet excitons
on the guest molecules.5,6 To meet this requirement,
π-conjugation of the host materials should be limited to
achieve a triplet energy level higher than that of the triplet
emitter. As an example, triphenyl silyl-containing wide-
band-gap insulating host materials have been developed
for the blue phosphorescent emitter.7 However, increased
operating voltage was obtained since the charge hopping
in the emitting layer (EML) occurred between the adja-
cent dopant molecules. As thus, another requirement for
a host material is its carrier transport property, and a lot
of host materials were developed with building blocks of
carbazole, a well-known electron donor.8-12 However, this
may induce an unbalanced carrier injection and transport
into the EML and then limited device performance.
Since the development of transition metal complexes,
like iridium and platinum, by Forrest and Thompson’s
groups, nearly 100% internal quantum efficiency of elec-
troluminescence has been realized by harvesting both
electrogenerated singlet and triplet excitons for emission
from organic light-emitting diodes (OLEDs) which emit
phosphorescence from their triplet states.1-4 To achieve a
highly efficient phosphorescent OLED, triplet emitters
are normally doped into a host material to reduce con-
centration quenching. Development of effective host
materials is thus of importance for efficient phosphores-
cent OLEDs. Generally, triplet excited states of the host
materials should be higher than those of the triplet
*To whom correspondence should be addressed. E-mail: mssjsu@
scut.edu.cn; kid@yz.yamagata-u.ac.jp.
(1) Baldo, M. A.; O’Brien, D. F.; You, Y.; Shoustikov, A.; Sibley, S.;
Thompson, M. E.; Forrest, S. R. Nature (London) 1998, 395, 151.
(2) Baldo, M. A.; Lamansky, S.; Burrows, P. E.; Thompson, M. E.;
Forrest, S. R. Appl. Phys. Lett. 1999, 75, 4.
(3) Adachi, C.; Baldo, M. A.; Thompson, M. E.; Forrest, S. R. J. Appl.
Phys. 2001, 90, 5048.
(5) Holmes, R. J.; Forrest, S. R.; Tung, Y.-T.; Kwong, R. C.; Brown,
J. J.; Garon, S.; Thompson, M. E. Appl. Phys. Lett. 2003, 82, 2422.
(6) Tokito, S.; Iijima, T.; Suzuki, Y.; Kita, H.; Tsuzuki, T.; Sato, F.
Appl. Phys. Lett. 2003, 83, 569.
(7) Ren, X.; Li, J.; Holmes, R. J.; Djurovich, P. I.; Forrest, S. R.;
Thompson, M. E. Chem. Mater. 2004, 16, 4743.
(4) Adachi, C.; Kwong, R. C.; Djurovich, P.; Adamovich, V.; Baldo,
M. A.; Thompson, M. E.; Forrest, S. R. Appl. Phys. Lett. 2001, 79,
2082.
(8) Shih, P.-I.; Chiang, C.-L.; Dixit, A. K.; Chen, C.-K.; Yuan, M.-C.;
Lee, R.-Y.; Chen, C.-T.; Diau, E. W.-G.; Shu, C.-F. Org. Lett.
2006, 8, 2799.
r
pubs.acs.org/cm
Published on Web 12/29/2010
2010 American Chemical Society

Article
Chem. Mater., Vol. 23, No. 2, 2011 275
To solve this problem, building blocks of the electron
acceptor, including benzimidazole,13-16 phosphine oxide,17,18
triazine,19-21 and oxadiazole,22 were introduced to develop
host materials with improved electron injection and trans-
port. Most recently, some host materials with building blocks
of both electron donor and acceptor, called bipolar host
materials, were developed.23-28 However, it is anticipated
that the introduction of both the electron donor and acceptor
to the host material may lead to an intramolecular charge
transfer, resulting in reduction of the energy band gap of the
molecule. Nevertheless, few works report on the effect of the
structure of the electron donor or electron acceptor on
properties of host materials and thus device performance,
which is of importance to realize highly efficient phosphor-
escent OLEDs for illumination applications utilizing blue,
green, and red triplet emitters.
In this article, we report on a series of host materials
1-7 with similar π conjugation containing the electron
donor of carbazole. These host materials are molecularly
different from each other in the number of nitrogen atoms
and their orientations in the central arylenes. Their carrier
mobilities and energy levels, including the highest occu-
pied molecular orbital (HOMO), the lowest unoccupied
molecular orbital (LUMO), and singlet and triplet en-
ergies, can be tuned by the number of nitrogen atoms and
their orientations, giving improved bipolarity compared
with that without any heterocyclic cores. Blue, green, and
red phosphorescent OLEDs were fabricated with well-
known triplet emitters of iridium(III) bis(4,6-(difluoro-
0
phenyl)pyridinato-N,C2 ) picolinate (FIrpic), fac-tris(2-
phenylpyridine) iridium (Ir(PPy)3), and tris(1-phenyliso-
quinolinolato-C2,N)iridium(III) (Ir(piq)3), respectively.
The effects of the central arylenes on their photophysical
properties, electron/hole mobilities, and device perfor-
mances were studied comprehensively.
Experimental Section
1
General. The H and 13C NMR spectra were recorded on a
Varian 500 (500 MHz) spectrometer. Mass spectra were ob-
tained using a JEOL JMS-K9 mass spectrometer. Differential
scanning calorimetry (DSC) was performed using a Perkin-
Elmer Diamond DSC Pyris instrument under nitrogen atmo-
sphere at a heating rate of 10 °C min-1. Thermogravimetric
analysis (TGA) was undertaken using a SEIKO EXSTAR 6000
TG/DTA 6200 unit under nitrogen atmosphere at a heating
rate of 10 °C min-1. UV-vis spectra were measured using a
Shimadzu UV-3150 UV-vis-NIR spectrophotometer. PL
spectra were obtained using a FluoroMax-2 (Jobin-Yvon-Spex)
luminescence spectrometer. Ionization potentials were deter-
mined by atmospheric ultraviolet photoelectron spectroscopy
(Rikken Keiki AC-3). Time-resolved emission spectra of host
materials were obtained at T = 4.2 K under excitation by a
nitrogen laser (λ = 337 nm, 50 Hz, 800 ps pulses) combined with
a streak scope C4334 (Hamamatsu) and a synchronous delay
generator C4792-02 (Hamamatsu). In comparison, transient
photoluminescence (PL) decays and the corresponding simul-
taneous PL spectra of the phosphorescent emitter-doped films
were recorded at room temperature. For calculation of HOMO
and LUMO energy levels, density functional theory (DFT) calcu-
lations were performed for optimized molecular structures and
single-point energies at the B3LYP/6-31G(d) and B3LYP/
6-311þG(d,p) levels, respectively, using a Gaussian suite of pro-
grams (Gaussian 03W). For calculation of ground states (S0) and
triplet excited states (T1), optimized molecular structures and
single-point energies were also calculated at the B3LYP/6-31G(d)
and B3LYP/6-311þG(d,p) levels, respectively.29
(9) Tsai, M.-H.; Hong, Y.-H.; Chang, C.-H.; Su, H.-C.; Wu, C.-C.;
Matoliukstyte, A.; Simokaitiene, J.; Grigalevicius, S.; Grazulevicius,
J. V.; Hsu, C.-P. Adv. Mater. 2007, 19, 826.
(10) Fukagawa, H.; Watanabe, K.; Tsuzuki, T.; Tokito, S. Appl. Phys.
Lett. 2008, 93, 133312.
(11) Tsai, M.-H.; Ke, T.-H.; Lin, H.-W.; Wu, C.-C.; Chiu, S.-F.; Fang,
F.-C.; Liao, Y.-L.; Wong, K.-T.; Chen, Y.-H.; Wu, C.-I. ACS Appl.
Mater. Interfaces 2009, 1, 567.
(12) Tsai, M.-H.; Lin, H.-W.; Su, H.-C.; Ke, T. H.; Wu, C.-C.; Fang,
F.-C.; Liao, Y.-L.; Wong, K.-T.; Wu, C.-I. Adv. Mater. 2006, 18,
1216.
(13) Chen, C.-H.; Huang, W.-S.; Lai, M.-Y.; Tsao, W.-C.; Lin, J. T.;
Wu, Y.-H.; Ke, T.-H.; Chen, L.-Y.; Wu, C.-C. Adv. Funct. Mater.
2009, 19, 2661.
(14) Takizawa, S.-Y.; Montes, V. A.; Anzenbacher, P., Jr. Chem. Mater.
2009, 21, 2452.
(15) Ge, Z.; Hayakawa, T.; Ando, S.; Ueda, M.; Akiike, T.; Miyamoto,
H.; Kajita, T.; Kakimoto, M. Chem. Mater. 2008, 20, 2532.
(16) Ge, Z.; Hayakawa, T.; Ando, S.; Ueda, M.; Akiike, T.; Miyamoto,
H.; Kajita, T.; Kakimoto, M. Adv. Funct. Mater. 2008, 18, 584.
(17) Padmaperuma, A. B.; Sapochak, L. S.; Burrows, P. E. Chem.
Mater. 2006, 18, 2389.
Device Fabrication and Characterization. Phosphorescent
OLEDs were grown on glass substrates precoated with a ∼110-nm-
thick layer of indium-tin oxide (ITO) having a sheet resistance of
15 Ω/0. The substrates were cleaned with ultrapurified water and
organic solvents and then dry-cleaned for 20 min by exposure to an
UV-ozone ambient. To improve hole injection from the anode,
poly(arylene amine ether sulfone)-containing tetraphenylbenzidine
(TPDPES) doped with 10% (by weight) tris(4-bromophenyl)
aminium hexachloroantimonate (TBPAH) was spun onto the
(18) Vecchi, P. A.; Padmaperuma, A. B.; Qiao, H.; Sapochak, L. S.;
Burrows, P. E. Org. Lett. 2006, 8, 4211.
(19) Chen, H.-F.; Yang, S.-J.; Tsai, Z.-H.; Hung, W.-Y.; Wang, T.-C.;
Wong, K.-T. J. Mater. Chem. 2009, 19, 8112.
(20) Inomata, H.; Goushi, K.; Masuko, T.; Konno, T.; Imai, T.; Sasabe,
H.; Brown, J. J.; Adachi, C. Chem. Mater. 2004, 16, 1285.
(21) Rothmann, M. M.; Haneder, S.; Da Como, E.; Lennartz, C.;
Schildknecht, C.; Strohriegl, P. Chem. Mater. 2010, 22, 2403.
(22) Leung, M.-K.; Yang, C.-C.; Lee, J.-H.; Tsai, H.-H.; Lin, C.-F.;
Huang, C.-Y.; Su, Y. O.; Chiu, C.-F. Org. Lett. 2007, 9, 235.
(23) Su, S.-J.; Sasabe, H.; Takeda, T.; Kido, J. Chem. Mater. 2008, 20,
1691.
(24) Tao, Y.; Wang, Q.; Yang, C.; Wang, Q.; Zhang, Z.; Zou, T.; Qin, J.;
Ma, D. Angew. Chem., Int. Ed. 2008, 47, 8104.
(25) Hsu, F.-M.; Chien, C.-H.; Shih, P.-I.; Shu, C.-F. Chem. Mater.
2009, 21, 1017.
(26) Jeon, S. O.; Yook, K. S.; Joo, C. W.; Lee, J. Y. Adv. Funct. Mater.
2009, 19, 3644.
(27) Jeon, S. O.; Yook, K. S.; Joo, C. W.; Lee, J. Y. Adv. Mater. 2010, 22,
1872.
(29) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.;
Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi,
J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.;
Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.;
Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.;
Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian,
H. P.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts,
R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli,
C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.;
Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.;
Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck,
A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.;
Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox,
D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.;
Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong,
M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, revision D.01;
Gaussian, Inc.: Wallingford CT, 2004.
(28) Chou, H.-H.; Cheng, C.-H. Adv. Mater. 2010, 22, 2468.

276 Chem. Mater., Vol. 23, No. 2, 2011
Scheme 1. Synthetic Routes of the Host Materials 1-7a
Su et al.
a i: Cu, K2CO3, DMF. ii: Bis(pinacolato)diboron, Pd2(dba)3, KOAc, PCy3, dioxane. iii: Pd(PPh3)4, 2 M K2CO3, toluene/ethanol, or PdCl2(PPh3)2, 2
M K2CO3, 1,4-dioxane, or Pd2(dba)3, PCy3, K3PO4, 1,4-dioxane/water.
precleaned substrate from its dichloroethane solution to form a
20-nm-thick polymer buffer layer.30 For the red phosphorescent
OLEDs, a 35-nm-thick 1,1-bis(4-(N,N-di(p-tolyl)-amino)phenyl)-
cyclohexane (TAPC) was deposited onto the buffer layer as a hole-
transport layer (HTL). Then, 4% (by weight) Ir(piq)3 was code-
posited with the host materials to form a 10-nm-thick EML.
Finally, a 65-nm-thick electron-transport layer (ETL) of 2,4,6-
tris(30-(pyridin-2-yl)biphenyl-3-yl)-1,3,5-triazine (TPyBPZ)31 was
deposited to block holes and to confine excitons in the emissive
zone. For the green phosphorescent OLEDs, TAPC (30 nm),
Ir(PPy)3 (7 wt %):1-7 (10 nm), and 1,3,5-tri(p-pyrid-3-yl-phe-
nyl)benzene (TpPyPB)32 (50 nm) were successively deposited as
the HTL, EML, and ETL, respectively. 2,20-Bis(m-di-p-tolyl-
aminophenyl)-1,10-biphenyl (3DTAPBP) (30 nm), Firpic (11
wt %):1-7 (10 nm), and 3,5,30,50-tetra(m-pyrid-3-yl)phenyl-
(1,10)-biphenyl (TmPyPBP)33 (40 nm) were successively deposited
as the HTL, EML, and ETL, respectively, for the blue phosphor-
escent OLEDs. Cathodes consisting of a 0.5-nm-thick layer of LiF
followed by a 100-nm-thick layer of Al were patterned using a
shadow mask with an array of 2 ꢀ 2 mm openings. Electrolumi-
nescent (EL) spectra were taken by an optical multichannel
analyzer, Hamamatsu PMA 11. Current density and luminance
versus driving voltage characteristics were measured by Keithley
source-measure unit 2400 and Konica Minolta chroma meter CS-
200, respectively. External quantum efficiencies (ηext) were calcu-
lated from the luminance, current density, and EL spectra,
assuming a Lambertian distribution.
Results and Discussion
Synthesis. The host materials were synthesized by Suzuki-
Miyaura cross-coupling reaction between 9-(3-(4,4,5,5-tetra-
methyl-1,3,2-dioxaborolan-2-yl)phenyl)-carbazole (9)
and aryl halide in the presence of palladium catalyst. 9
was synthesized by cross-coupling between bis(pinacolato)-
diboron and 9-(3-bromophenyl)-carbazole (8), which was
synthesized by Ullmann reaction between 1-bromo-3-iodo-
benzene and carbazole. Among these host materials, 1, 2, and
4 were synthesized with aryl bromide using tetrakis(tri-
phenylphosphine)palladium(0) (Pd(PPh3)4) as the catalyst
in the presence of 2 M K2CO3. In comparison, 5, 6, and 7
were synthesized with aryl chloride using bis(triphenyl-
phosphine)palladium(II) chloride (PdCl2(PPh3)2) as the cat-
alyst. Taking into account the low reactivity of pyridine
chloride, 3 was synthesized with 2,4-dichloropyridine using
tris(dibenzylideneacetone) dipalladium(0) (Pd2(dba)3) as the
catalyst in the presence of tricyclohexylphosphine (PCy3) and
K3PO4 (Scheme 1). All the host materials were obtained as
white powders in good yields and were purified by repeated
temperature gradient vacuum sublimation before character-
ization and device fabrication.
Physical and Optical Properties. UV-vis absorption
and PL spectra of vacuum evaporated films (60 nm) for
host materials 1-7 on quartz substrates measured at
room temperature are shown in Figure 1. Their peak
maxima are summarized in the Supporting Information,
Table S1. Absorption peaks at 286, 297, 331, and 344 nm
can be attributed to π-π* transitions of the carbazole
chromophore. An absorption peak at around 243 nm and
a shoulder at 265 nm can be attributed to π-π* and n-π*
transitions of central arylenes, respectively. Absorbance
at wavelengths longer than 350 nm increases with intro-
ducing one or two nitrogen atoms into the central ary-
lenes. This can be attributed to strong electron affinity of
heterocyclic arylenes and intramolecular charge transfer
formed between the central heterocyclic arylenes and the
For carrier mobility measurement, films of 1-7 (∼10 μm)
were prepared on ITO-coated substrates. A semitransparent Al
layer was patterned using a shadow mask with an array of 2 ꢀ 2
mm openings. Hole/electron mobilities were measured by using
a conventional photoinduced time-of-flight (TOF) technique. A
nitrogen laser was used as the excitation source (λ=337 nm) and
was incident on the sample through the ITO or semitransparent
Al electrode.
(30) Sato, Y.; Ogata, T.; Kido, J. Proc. SPIE 2000, 4105, 134.
(31) Su, S.-J.; Sasabe, H.; Pu, Y.-J.; Nakayama, K.-I.; Kido, J. Adv.
Mater. 2010, 22, 3311.
(32) Su, S.-J.; Chiba, T.; Takeda, T.; Kido, J. Adv. Mater. 2008, 20, 2125.
(33) Su, S.-J.; Tanaka, D.; Li, Y.-J.; Sasabe, H.; Takeda, T.; Kido, J.
Org. Lett. 2008, 10, 941.

Article
Chem. Mater., Vol. 23, No. 2, 2011 277
HOMO energy levels of host materials 1-7 determined
by atmospheric photoelectron spectroscopy are between
6.08 and 6.23 eV, and LUMO energy levels estimated
from the corresponding HOMO and Eg are between 2.60
and 3.09 eV. As expected, lower-lying HOMO and
LUMO energy levels are achieved by introducing hetero-
cyclic cores with stronger electron affinity instead of
benzene, and the more nitrogen atoms in the central
heterocyclic core, the lower lying the HOMO and LUMO
levels achieved. In addition, the orientations of nitrogen
atoms in the heterocyclic cores also result in different
energy levels, and one or more nitrogen atoms at the
outside of the molecule seem to induce lower-lying energy
levels. Besides host materials, a similar phenomenon has
also been found for the pyridine-containing ETMs.31,34
To obtain triplet energies (ET) of host materials 1-7,
extremely long PL decay components of their films lasting
for ∼10 ms were recorded at T = 4.2 K. Figure 2 shows
two time-resolved emission spectra of each film. One is
the emission spectrum obtained in the whole range (0 <
t < 10 ms) (black lines), and the other is the emission spec-
trum of the delayed component (2 ms < t < 10 ms) (gray
lines), also called the phosphorescence spectrum of the
corresponding host material. 1-6 exhibit the highest-
energy phosphorescence peaks (S0
v=0
r T1v=0) at 456,
458, 458, 458, 461, and 470 nm, corresponding to ET of
2.72, 2.71, 2.71, 2.71, 2.69, and 2.64 eV, respectively.
Phosphorescence of 7 is too weak to estimate its peak or
onset. Although phosphorescence of 2-4 with a pyridine
core is slightly weaker than that of 1 with a benzene core,
v=0
v=0
they show quite similar S0
r T1
and ET. With
introducing one more nitrogen atom in the central hetero-
cyclic core, a bathochromic shift of phosphorescence was
clearly seen, resulting in lower ET for 5 and 6. Although
phosphorescence of 7 is too weak to estimate its peak or
onset, it is anticipated that its ET should also be lower
than that of 1-4. This will be proven by the following
DFT calculations. Nevertheless, compared with a general
host material of N,N0-dicarbazolyl-4,40-biphenyl (CBP)
(ET=2.56 eV), higher ET can be achieved when the central
arylenes between the outer carbazoles are combined with
each other at their meta-positions (refer to combination
sites).
Figure 1. Room-temperature UV/vis absorption (top) and photolumi-
nescence (bottom) spectra for vacuum-deposited films of host materials
1-7 on quartz substrates.
outer carbazoles. From their absorption edges, energy
band gaps (Eg) of 1-7 can be assumed to be 3.48, 3.40,
3.38, 3.43, 3.17, 3.13, and 3.14 eV, respectively. As anti-
cipated, narrower Eg values are obtained by introducing
heterocyclic arylenes as the cores, and the more nitrogen
atoms in the heterocyclic core, the narrower the Eg values
obtained. For example, Eg of 2 is 3.40 eV, which is 0.23
and 0.26 eV wider than that of 5 and 7, respectively. A
similar phenomenon is also found for the other host
materials.
PL spectrum of 1 peaks at 352, 378, and 400 nm. In
contrast, PL spectra of 2-7 are structureless and peak at
387, 401, 381, 426, 439, and 421 nm, respectively. A batho-
chromic shift of the PL spectrum is found when introducing
one or two nitrogen atoms in the central arylene core, and
the more nitrogen atoms in the heterocyclic core, the longer
the bathochromic shifts btained. Besides the number of
nitrogen atoms, their PL spectra are also affected by
nitrogen atom orientations in the heterocyclic cores. The
host materials that have some nitrogen atoms at the ortho-
position (refer to combination sites) show a longer batho-
chromic shift, and it can be attributed to elongated n-π
conjugation induced by nonpaired electrons of those nitro-
gen atoms.
Compared with the phosphorescence spectra, the
stronger emission spectra at shorter wavelengths are
due to fluorescence of the host materials, which are almost
consistent with those recorded at room temperature. From
v=0
the highest energy fluorescence peaks (S0
r S1v=0),
singlet energies (ES) of host materials 1-7 can be estimated
as3.25, 3.10, 3.08, 3.14, 2.89, 2.82, and 2.94 eV, respectively.
In addition to the requirement of a high ET, a good host
material is required to have favorable exchange energy as
small as possible to allow for both charge injection into
the host and efficient triplet emission from a dispersed
triplet emitter. Here, the energy difference (ΔEST) between
the singlet and triplet emission transitions is taken as an
(34) Su, S.-J.; Takahashi, Y.; Chiba, T.; Takeda, T.; Kido, J. Adv. Funct.
Mater. 2009, 19, 1260.

278 Chem. Mater., Vol. 23, No. 2, 2011
Su et al.
Figure 2. Time-resolved emission spectra of vacuum-deposited films of 1-7 excited by a nitrogen laser (λ = 337 nm, 50 Hz, 800 ps) at T = 4.2 K. Black
lines: emission spectra in the whole range (0 < t < 10 ms). Gray lines: emission spectra of the delayed component (phosphorescence) (2 < t < 10 ms).
Table 1. Singlet Energies (ES, Measured as S0v=0 r S1v=0 at 4.2 K),
Triplet Energies (ET, Measured as S0v=0 r T1v=0 at 4.2 K), and
Singlet-Triplet Energy Difference (ΔEST) for 1-7
estimate for the exchange energy.35 With introducing a
pyridine core instead of a benzene core, a decrease of
ΔEST by 0.10-0.16 eV (compare 1 and 2-4) was achieved
(Table 1). In addition, with introducing one more nitrogen
atom into the central heterocyclic core, further decreases in
ΔEST by 0.20 and 0.18 eV were achieved for 5 and 6,
respectively, indicating the host materials with heterocyclic
cores may be attractive candidates to reduce the driving
voltage of OLEDs.
materials
ES, eV
ET, eV
ΔEST, eV
1
2
3
4
5
6
7
3.25
3.10
3.08
3.14
2.89
2.82
2.94
2.72
2.71
2.71
2.71
2.69
2.64
n.a.
0.53
0.39
0.37
0.43
0.20
0.18
n.a.
Besides optical properties, their glass transition tem-
peratures (Tg) are also influenced by the number of
nitrogen atoms in the heterocyclic cores. Slightly im-
proved Tg values are achieved for the host materials with
one or two nitrogen atoms in the central arylene core, and
it can be attributed to increased polarity and thus stronger
molecular interaction. In addition, orientations of nitro-
gen atoms in the heterocyclic cores also affect their Tg. As
an example, Tg values of 3 and 4 are slightly higher than
that of 2 since their nitrogen atoms are located at the
€
(35) Brunner, K.; van Dijken, A.; Borner, H.; Bastiaansen, J. J. A. M.;
(36) Meier, C.; Ziener, U.; Landfester, K.; Weihrich, P. J. Phys. Chem.
B 2005, 109, 21015.
(37) Ziener, U. J. Phys. Chem. B 2008, 112, 14698.
Kiggen, N. M. M.; Langeveld, B. M. W. J. Am. Chem. Soc. 2004,
126, 6035.

Article
Chem. Mater., Vol. 23, No. 2, 2011 279
outside of the molecule to give stronger molecular inter-
action than 2, such as an intermolecular CH-N hydro-
gen-bonding interaction.36,37 All the host materials show
Tg higher than 100 °C, proving high morphologic stability
of the amorphous phase in a deposited film, which is
a prerequisite for their applications in OLEDs.
DFT Calculations. DFT calculations of 1-7 were per-
formed using the Gaussian suite of programs (Gaussian
03W). To obtain their HOMO and LUMO energy levels,
molecular structure optimization and single-point energy
calculations were performed at the B3LYP/6-31G(d) and
B3LYP/6-311þG(d,p) levels, respectively. As shown in
Figure 3, HOMOs of all the host materials are mainly
located at the outer carbazole. In contrast, their LUMOs
are mainly located at the central diphenyl arylene skele-
ton between the outer carbazoles. The calculated HOMO
and LUMO levels are demonstrated in the Supporting
Information, Figure S1, in comparison with the experi-
mental ones. It can be clearly seen that lower-lying
LUMO energy levels are obtained by introducing hetero-
cyclic cores instead of benzene, and the more nitrogen
atoms in the heterocyclic core, the lower lying the LUMO
levels obtained. Moreover, different energy levels are
obtained by the change of nitrogen atom orientations in
the heterocyclic cores. Although the experimental energy
levels are somewhat different from the calculated ones,
their trends are almost consistent. Different from the
calculated LUMO levels, the calculated HOMO levels
are slightly influenced by the number of nitrogen atoms
and their orientations, and this is also consistent with the
experimental results.
Triplet energies of 1-7 are obtained by the difference of
their ground state (S0) and triplet excited state (T1)
energies. Their optimized molecular structures and the
corresponding single-point energies were also calculated
at the B3LYP/6-31G(d) and B3LYP/6-311þG(d,p) le-
vels, respectively. Although we failed in estimating ET of 7
from its phosphorescence spectrum, the calculated one
for 7 is 2.76 eV (Table 2), which is the lowest one among
these host materials. It further proves that more nitrogen
atoms in the central heterocyclic arylene induce lower
triplet energies. Moreover, the trend of the calculated
triplet energy is consistent with that of the experimental
one. The energy difference (ΔE) between Eg and (T1 - S0)
is taken as an estimate for the energy difference between
the singlet and triplet transitions. Similar to the experi-
mental ΔEST, ΔE of 1 with a benzene core is the largest
one among these hosts, and smaller ΔE is achieved with
introducing one or two nitrogen atoms in the central
arylene core. In addition, ΔE of 6 is the smallest one
among these hosts, which is also consistent with the
experimental result.
Figure 3. Calculated spatial distributions (DFT, B3LYP/6-311Gþdp,
Gaussian 03W) of HOMOs and LUMOs of 1-7.
ability, well-known red, green, and blue triplet emitters of
Ir(piq)3, Ir(PPy)3, and FIrpic were doped into 1-7 for
transient PL decay measurements. Their decay curves and
the corresponding simultaneous PL spectra were shown
in Supporting Information, Figures S2-S4, respectively.
It can be seen that FIrpic doped into 1-4 and 6 clearly
exhibits monoexponential decay curves with relatively
long lifetimes between 1.51 and 1.73 μs (Table 3). In
addition, they exhibit high PL quantum efficiency (ηPL
)
over 70%, which can be attributed to good confinement
of triplet energy on the FIrpic molecules and efficient
energy transfer from the host to the guest. Although
transient PL decay of FIrpic doped into 5 is not mono-
exponential, its second exponential decay component is
much smaller than its first one (lower than 1%), giving
Triplet Exciton Confinement. As a host material for a
triplet emitter, its triplet excited state should be higher
than that of the triplet emitter to prevent reverse energy
transfer from the guest to the host. Since decreased ET is
obtained with introducing one or two nitrogen atoms in
the central arylene, it seems in contradiction with this
requirement. To verify their triplet exciton confinement
η
PL of 57% with a first exponential component lifetime of
1.29 μs. Among these host materials, FIrpic doped into 7

280 Chem. Mater., Vol. 23, No. 2, 2011
Table 2. Density Functional Theory (DFT) Calculations of 1-7a
Su et al.
materials
HOMO, eV
LUMO, eV
Eg, eV
T1 - S0, eV
ΔE,b eV
dipole moment, D
1
2
3
4
5
6
7
5.72
5.69
5.73
5.77
5.69
5.73
5.78
1.60
1.89
2.04
1.78
2.30
2.42
2.28
4.12
3.80
3.69
3.99
3.39
3.31
3.50
2.95
2.84
2.86
2.93
2.82
2.78
2.76
1.17
0.96
0.83
1.06
0.57
0.53
0.74
1.460
3.008
1.481
2.455
1.622
2.150
1.588
a Molecular structure optimization and single-point energy calculations were performed at the B3LYP/6-31G(d) and B3LYP/6-311þG(d,p) levels
(DFT, Gaussian 03W), respectively. b The energy difference (ΔE) between Eg and (T1 - S0) taken as an estimate for the energy difference between the
singlet and triplet transitions.
Table 3. Phosphorescence Lifetime (τ), Photoluminescence Quantum
Efficiency (ηPL), and Radiative and Nonradiative Rate Constants (kr and
Ir(PPy)3:(1-6) exhibit high ηPL over 80%. Even for 7 with
knr) of FIrpic, Ir(PPy)3, and Ir(piq)3 Doped into 1-7
by the different hosts compared with those for FIrpic.
the lowest ET, ηPL of the doped film is as high as 78%,
guests
FIrpic
hosts
τ, μs
η
PL, %
kr, ꢀ 105 s-1
knr, ꢀ 105 s-1
slightly lower than the others. It indicates good triplet
energy confinement on the Ir(PPy)3 molecules and effi-
cient energy transfer from the host to the guest, and ET
of 7 should be higher than that of Ir(PPy)3 (ET = 2.41 eV,
corresponding to the highest-energy peak).
For Ir(piq)3 with the lowest triplet energy among these
phosphorescent emitters, all the doped films exhibit clear
monoexponential decay curves and relatively long life-
times of about 1.2 μs. Their kr values are between 4.46 ꢀ
105 and 4.61 ꢀ 105 s-1, and knr values are between 3.59 ꢀ
105 and 3.95 ꢀ 105 s-1. Their ηPL values are around 55%
and are slightly affected by the number of nitrogen atoms
and their orientations in the heterocyclic cores. The
transient PL observation indicates that triplet energy
transfer from Ir(piq)3 to the hosts is completely sup-
pressed, and the energy is well confined on the Ir(piq)3
molecules due to their higher ET than the guest.
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1.51
1.73
1.53
1.59
1.29
1.53
0.67
1.12
1.24
1.13
1.24
1.07
1.23
1.14
1.21
1.19
1.21
1.23
1.19
1.23
1.20
83
84
71
87
57
73
18
91
85
88
90
83
87
78
55
53
56
55
53
56
54
5.48
4.84
4.65
5.48
4.41
4.78
2.67
8.15
6.86
7.82
7.28
7.73
7.10
6.82
4.53
4.46
4.61
4.49
4.46
4.57
4.50
1.13
0.93
1.90
0.82
3.32
1.77
12.2
0.80
1.22
1.06
0.81
1.58
1.06
1.93
3.71
3.95
3.62
3.67
3.95
3.59
3.83
Ir(PPy)3
Ir(piq)3
Carrier Mobilities. Since the charge carriers are recom-
bined in the EML in an OLED, balanced carrier injection
and transport into the EML is a prerequisite for improved
device performance. As thus, different from the hole-
transport and electron-transport materials, balanced hole
and electron mobility is ideal for a host material. Hole/
electron mobilities of 1-7 were measured by using a
conventional photoinduced TOF technique using a nitro-
gen laser as the excitation source (λ = 337 nm). Figure 4
shows their hole (μh) and electron (μe) mobilities plotted
as a function of the square root of electric field (E). μh of 1
lies in the range from 9.6 ꢀ 10-5 to 1.8 ꢀ 10-4 cm2 V-1 s-1
exhibits a complex decay curve with a short first expo-
nential component lifetime of 0.67 μs. From its simulta-
neous PL spectrum, a high-energy emission peak was
found at 425 nm, which can be attributed to delayed
fluorescence of 7, indicating inefficient energy transfer
from 7 to FIrpic. Since the emissions from the host 7 and
the guest FIrpic are covered with each other, the decay
curve can be hardly devided according to the host or the
guest, but the intensity from FIrpic is much stronger than
that from the host. The first exponential component
lifetime of 0.67 μs can be considered to be that of FIrpic.
Besides the short lifetime, ηPL of FIrpic:7 is only 18%, the
lowest one among all the doped films. Moreover, its
radiative rate constant (kr) is only 2.67 ꢀ 105 s-1, which
is also the smallest one among these doped films and even
at an electric field between 3.0 ꢀ 105 and 7.2 ꢀ 105 V cm-1
,
which is 1 order of magnitude higher than its μe, ranging
from 1.1 ꢀ 10-5 to 5.2 ꢀ 10-5 cm2 V-1 s-1 at an electric
field between 3.6 ꢀ 105 and 8.1 ꢀ 105 V cm-1. In contrast,
μe values of 2-7 are somewhat higher or equal to their μh
values except 6, and the improved μe can be attributed to
the heterocyclic cores with stronger electron affinity
instead of benzene. Generally, μe of organic materials is
several orders of magnitude lower than μh,38 and as a host
for the EML, increased μe may facilitate electron injection
and transport into the EML to give a balanced carrier and
thus higher recombination efficiency. μe of 6 is slightly
much smaller than its nonradiative rate constant (knr
=
12.2 ꢀ 105 s-1). All these results indicate ET of 7 must be
the lowest one among these host materials and should be
equal to or lower than that of FIrpic (ET = 2.61 eV,
corresponding the highest-energy peak), further support-
ing the former experimental and calculation results.
In comparison, Ir(PPy)3 doped into 1-7 clearly exhi-
bits monoexponential decay curves with lifetimes be-
tween 1.07 and 1.24 μs. Their kr values are between
6.82 ꢀ 105 and 8.15 ꢀ 105 s-1, and knr values are between
0.80 ꢀ 105 and 1.93 ꢀ 105 s-1. Both are slightly influenced
(38) Antoniadis, H.; Abkowitz, M. A.; Hsieh, B. R. Appl. Phys. Lett.
1994, 65, 2030.

Article
Chem. Mater., Vol. 23, No. 2, 2011 281
Figure 4. Hole (μh) and electron (μe) mobilities of 1-7 films plotted as a
function of the square root of electric field (E).
lower than its μh, which can be attributed to its stronger
molecular polarity since its two nitrogen atoms in
the heterocyclic cores are located at the outside of the
molecule.
μh of 1 is higher than that of 2-4. In contrast, its μe is
lower or equal to that of 2-4. It can be seen that intro-
ducing the pyridine ring instead of benzene as the core
induces lower μh but slightly affects μe. In addition, there
is almost no effect on μe from the nitrogen atom orienta-
tion in the pyridine cores. A similar phenomenon has also
been found for the pyridine-containing ETMs that their
μe values are nearly independent of the nitrogen atom
orientations.34 Taking into account their energy levels
and carrier mobility, an improved bipolarity has been
achieved for the host materials containing heterocyclic
cores compared with 1.
Phosphorescent OLED Characterization. Blue, green,
and red phosphorescent OLEDs were fabricated by dop-
ing FIrpic, Ir(PPy)3, and Ir(piq)3 into the developed host
materials as the EMLs, and Figures 5-7 show their
efficiency-luminance characteristics, respectively. Their
performance data are summarized in Table 4. For the
blue phosphorescent OLEDs, the highest efficiency was
achieved for the host 2, giving ηext of 24.3% and power
efficiency (PE) of 46.1 lm W-1 at 100 cd m-2 and rolling
to 22.6% and 34.5 lm W-1 at 1000 cd m-2, respectively,
which is one of the highest performances for the FIrpic-
based blue phosphorescent OLEDs.32,34,39 In compari-
son, lower efficiencies of 21.3% and 40.2 lm W-1 were
obtained at 100 cd m-2 for the host 1. Although ηPL
values of FIrpic:1 and FIrpic:2 are quite similar, the
improved performance for the device based on 2 can be
attributed to improved bipolarity of 2 achieving balance
of electron and hole fluxes in the EML. For the devices
based on 3 and 4 with the same pyridine core but different
Figure 5. (a) External quantum efficiency (ηext) and (b) power efficiency
(PE) vs luminance (L) characteristics for ITO/TPDPES:TBPAH (20 nm)/
3DTAPBP (30 nm)/1-7:11 wt % FIrpic (10 nm)/TmPyPBP (40 nm)/LiF
(0.5 nm)/Al (100 nm).
nitrogen atom orientations, the efficiencies at 100 cd m-2
decreased to 14.9%, 28.6 lm W-1, and 19.1%, 34.6
lm W-1, respectively. Although ηPL of FIrpic:4 is slightly
higher than that of FIrpic:2, the decreased efficiency of
the device based on 4 compared with the device based on 2
can be attributed to the lower-lying LUMO level of 4
giving electron-rich EML. Further reduced efficiency was
obtained for the device based on 3 due to lower ηPL of
FIrpic:3. ηext values of the devices based on 5-7 de-
creased further to 16.9%, 9.2%, and 8.4%, respectively.
As anticipated, efficiency of the device based on the host 7
is the lowest one among these host materials due to its
lowest ET and thus the lowest ηPL. Different from its PL
spectrum, no emission from the host 7 was detected in the
EL spectrum (Supporting Information, Figure S5), in-
dicating some of the excitons are generated by a direct
carrier trap by the FIrpic molecules. Since FIrpic is more
like an electron transporter, a carrier trap by the FIrpic
molecules may further induce unbalanced carriers in the
EML containing host materials with low-lying LUMO
levels.
For the green phosphorescent OLEDs, the highest
efficiency was also achieved for the host 2, exhibiting
ηext of 26.9% and PE of 102 lm W-1 at 100 cd m-2
.
Although ηPL of Ir(PPy)3:1 is higher than Ir(PPy)3:2,
(39) Su, S.-J.; Gonmori, E.; Sasabe, H.; Kido, J. Adv. Mater. 2008, 20,
4189.
efficiencies of 22.0% and 83.6 lm W-1 were obtained at

282 Chem. Mater., Vol. 23, No. 2, 2011
Su et al.
100 cd m-2 for the device based on 1, and both are the
lowest one among these hosts. The improved efficiency
for the host materials 2-7 can also be attributed to their
improved bipolarity by introducing heterocyclic cores
instead of benzene. Compared with the device based on
2, slightly lower ηext values of 23.9% and 25.9% were
obtained at 100 cd m-2 for the devices based on 3 and 4
with the same pyridine core but different nitrogen atom
orientations, but PE of the device based on 4 is as high
as that of the 2-based device at brightness lower than 200
cd m-2. Besides the high efficiencies achieved by the host
materials 2-7, their ηext values roll very slightly from a
display-relevant luminance of 100 cd m-2 to an illumina-
tion-relevant luminance of 1000 cd m-2. Even at a much
brighter luminance of 10000 cd m-2, ηext remains 23.4%
(corresponding to 83.9 cd/A) with a PE of 62.3 lm W-1
for the device based on 2, and both are the highest ever
values for the Ir(PPy)3-based green phosphorescent
OLEDs.31-33 Besides 2, ηext values higher than 20% were
also achieved at 10 000 cd m-2 for the devices based on 5
and 7. It is of interest that all three host materials have a
nitrogen atom located between the combination sites of
the central heterocyclic core.
For the red phosphorescent OLEDs, all the devices
show very high ηext as 17%∼19% at low current density
(∼0.01 mA cm-2, luminance of ∼1 cd m-2 is detected).
Note that ηPL values of Ir(piq)3:1-7 are around 55%.
Taking into account that the light-out-coupling efficiency
of an OLED is around 30%,31,32 similar to the blue and
green phosphorescent OLEDs, a well-balanced carrier
was achieved at low current density for all the devices.
With increase of current density, a steep decrease of ηext
was found for the device based on 1. The steep decrease in
ηext can be attributed to reduced carrier balance at high
current density. Note that the concentration of Ir(piq)3 is
as low as 4 wt %. At low current density, carriers may be
injected into the EML through the Ir(piq)3 dopant mole-
cules, but at high current density, more carriers must be
injected into the EML through the host molecules. Rela-
tively high-lying HOMO and LUMO levels and the high-
er hole mobility of 1 facilitate hole injection and transport
into the EML and thus a reduced carrier balance. The
reduced carrier balance induces decreased carrier recom-
bination efficiency and thus deduced ηext. In addition, the
imbalance of the carriers also causes domination of the
hole polaron (hþ), and a large number of holes and
polarons accumulate at the interface of EML and ETL
and/or in the active layer, resulting in triplet exciton-
polaron quenching and thus further efficiency roll-off.40
On this result, ηext rolls off to 8.36% at an illumination-
relevant luminance of 1000 cd m-2. Due to the high
driving voltage at this luminance, PE further rolls off to
3.56 lm W-1. As a control device based on a general
host material CBP without any heterocyclic core, a
similar steep decrease in ηext and PE was also found due
to this reason, and they are 7.44% and 3.86 lm W-1 at
1000 cd m-2, respectively.
Figure 6. (a) External quantum efficiency (ηext) and (b) power efficiency
(PE) vs luminance (L) characteristics for ITO/TPDPES:TBPAH (20 nm)/
TAPC (30 nm)/1-7:8 wt % Ir(PPy)3 (10 nm)/TpPyPhB (50 nm)/LiF
(0.5 nm)/Al (100 nm).
Figure 7. (a) External quantum efficiency (ηext) and (b) power efficiency
(PE) vs luminance (L) characteristics for ITO/TPDPES:TBPAH (20 nm)/
TAPC (35 nm)/1-7 or CBP:4 wt % Ir(piq)3 (10 nm)/TPyBPZ (65 nm)/
LiF (0.5 nm)/Al (100 nm).
(40) Reineke, S.; Walzer, K.; Leo, K. Phys. Rev. B 2007, 75, 125328.

Article
Chem. Mater., Vol. 23, No. 2, 2011 283
Table 4. Performance Data of RGB Phosphorescent OLEDs Using 1-7 as the Host Materials
at 10 cd m-2
PE, lm W-1
at 100 cd m-2
PE, lm W-1
at 1000 cd m-2
PE, lm W -1
guests
FIrpic
hosts
V, V
η
ext, %
V, V
η
ext, %
V, V
ηext, %
1
2
3
4
5
6
7
1
2
3
4
5
6
7
3.13
3.10
3.05
3.17
3.24
3.08
3.17
2.77
2.76
2.69
2.69
2.80
2.58
2.71
2.65
2.70
2.88
2.80
2.76
2.71
2.48
2.66
48.5
50.0
42.9
43.1
42.0
31.8
26.2
91.2
108
100
109
102
106
22.2
22.8
19.3
20.4
19.5
13.5
11.6
22.4
26.5
23.9
26.3
25.2
24.3
24.4
17.2
18.2
15.8
15.7
15.3
17.9
18.4
18.5
3.61
3.59
3.51
3.70
3.78
3.64
3.75
2.96
2.97
2.85
2.82
3.03
2.80
2.91
3.29
3.62
3.57
3.55
3.54
3.30
3.02
3.15
40.2
46.1
28.6
34.6
31.2
18.3
15.9
83.6
102
21.3
24.3
14.9
19.1
16.9
9.25
8.43
22.0
26.9
23.9
25.9
25.5
24.1
25.5
13.6
13.6
12.6
13.2
10.5
17.1
17.6
17.3
4.50
4.46
4.35
4.67
4.88
4.88
5.00
3.30
3.33
3.22
3.10
3.43
3.20
3.28
5.53
6.62
5.21
5.50
5.63
4.92
4.60
4.56
28.0
34.5
18.5
24.1
19.9
11.7
14.6
68.7
88.4
79.1
86.8
80.2
76.2
86.0
3.86
3.56
5.36
5.92
3.90
8.65
9.54
9.00
18.6
22.6
11.9
17.9
14.0
7.93
6.55
20.1
26.2
22.6
24.2
24.1
21.6
24.8
7.44
8.36
9.90
11.6
7.82
15.5
16.1
15.2
Ir(PPy)3
94.5
102
96.0
97.3
99.5
11.9
10.6
9.96
10.5
8.36
14.2
15.9
14.9
102
Ir(piq)3
CBP
1
2
3
4
5
6
7
18.7
19.0
15.5
15.7
15.6
18.2
20.3
18.8
Although efficiency roll-off was also found for the
devices based on 2-4 with a pyridine core, it was sup-
pressed compared with the devices based on CBP and 1.
Their ηext values at 1000 cd m-2 luminance are 9.90%,
11.6%, and 7.82% with PEs of 5.36, 5.92, and 3.90
lm W-1, respectively, which are improved in comparison
to the devices based on CBP and 1. The suppressed
efficiency roll-off can be attributed to lower-lying energy
levels and improved bipolarity of 2-4 compared with 1,
giving improved carrier balance and thus suppressed triplet
exciton-polaron quenching at high current density.
Efficiency roll-off was further suppressed for the de-
vices based on 5-7 with pyrimidine or pyrazine cores.
Their ηext values are 15.5%, 16.1%, and 15.2% at 1000
cd m-2 luminance, respectively, which are significantly
improved in comparison to the devices based on CBP and
1-4. As a factor concerning power consuming especially
for high luminance applications, their PEs are improved
to 8.65, 9.54, and 9.00 lm W-1 at 1000 cd m-2 luminance,
respectively, and are over two times higher than those of
the devices based on 1 and CBP. Moreover, decreased
driving voltages were achieved for the devices based on
5-7 in comparison with those based on CBP and 1-4,
which can be attributed to their decreased ΔEST achieved
with introducing one or two nitrogen atoms into the
central arylene cores. This further proves that more
carriers must be injected into the EML through the hosts
rather than direct carrier trap by the dopant, especially at
were achieved at 10 cd m-2 and slightly roll off to 17.6%
and 15.9 lm W-1 at 100 cd m-2. Both are the highest
efficiencies for the red phosphorescent OLEDs based on
Ir(piq)3 hitherto.41-44 Even at an illumination-relevant
luminance of 1000 cd m-2, the efficiencies maintain at
16.1% and 9.54 lm W-1, respectively, compared with the
highest ever values of 12.2% and 10.6 lm W-1 recorded at
only 100 cd m-2 41
. The suppressed efficiency roll-off can be
attributed to low-lying LUMO levels of 5-7, giving im-
proved electron injection and carrier balance and thus
reduced triplet exciton-polaron quenching at high current
density. Different from the blue phosphorescent OLEDs
based on FIrpic, the host materials with lower-lying
LUMO levels seem to be better hosts for Ir(piq)3 since
FIrpic is a well-known electron-transport dopant and is
more like a carrier trapper.
The emission color is deep red with an EL emission
peak at 621 nm and CIE-1931 color coordinates of (0.673,
0.327), independent of the choice of hosts. This proves the
emission directly from the triplet emitter Ir(piq)3. Similar
phenomena were also found for the blue and green
phosphorescent OLEDs (Supporting Information, Fig-
ure S5). Note that the thickness of the current EMLs is
only 10 nm, much thinner than in most of the phosphor-
escent OLEDs previously reported. Different from most
of the efforts to reduce efficiency roll-off by using multi-
EMLs to widen the carrier recombination zone,45-47 the
current study indicates that improved efficiency and
reduced efficiency roll-off can be achieved for a device
with a single EML and a narrow recombination zone by
high current density. Taking the device based on 6 (ΔEST
=
0.18 eV) as an example, the luminance reaches values of
10 and 100 cd m-2 at 2.48 and 3.02 V, respectively, which
are even equal to or lower than those of the device based on
a p-i-n structure.41,42 ηext and PE of 18.4% and 20.3 lm W-1
(43) Tsuzuki, T.; Tokito, S. Adv. Mater. 2007, 19, 276.
(44) Kanno, H.; Ishikawa, K.; Nishio, Y.; Endo, A.; Adachi, C.;
Shibata, K. Appl. Phys. Lett. 2007, 90, 123509.
(45) He, G.; Pfeiffer, M.; Leo, K.; Hofmann, M.; Birnstock, J.; Pudzich,
R.; Salbeck, J. Appl. Phys. Lett. 2004, 85, 3911.
(46) Watanabe, S.; Ide, N.; Kido, J. Jpn. J. Appl. Phys., Part 1 2007, 46,
1186.
(41) Meerheim, R.; Walzer, K.; He, G.; Pfeiffer, M.; Leo, K. Proc. SPIE
2006, 6192, 61920P.
(42) Meerheim, R.; Walzer, K.; Pfeiffer, M.; Leo, K. Appl. Phys. Lett.
2006, 89, 061111.
(47) Sun, Y.; Forrest, S. R. Appl. Phys. Lett. 2007, 91, 263503.

284 Chem. Mater., Vol. 23, No. 2, 2011
Su et al.
the development of host materials with ideal energy
levels.
attributed to its low-lying LUMO level and the smallest
ΔEST, giving improved electron injection and carrier
balance and thus reduced triplet exciton-polaron
quenching at high current density. Different from the
blue and green phosphorescent OLEDs based on FIrpic
and Ir(PPy)3, the host materials with lower-lying LUMO
levels seem to be better hosts for Ir(piq)3 since more
carriers must be injected through the host especially at
high current density. To realize OLEDs for the next-
generation illumination applications, development of
highly efficient RGB triplet emitter-based phosphores-
cent OLEDs is the only route beyond fluorescent tube and
inorganic LED efficacy. The current study gives an
approach to such a goal by molecular design of host
materials for specific triplet emitters.
Conclusions
We reported on a series of host materials containing
building blocks of carbazole and arylenes, like benzene,
pyridine, pyrimidine, and pyrazine. Their energy levels
including HOMO, LUMO, singlet, and triplet energies
can be tuned by the choice of central heterocyclic arylenes
and also their nitrogen atom orientations, and this is also
supported by the DFT calculations. A decrease of ΔEST
was achieved with introducing one or two nitrogen atoms
into the central arylene. Their carrier mobilities can also
be tuned by different heterocyclic cores and nitrogen
atom orientations, giving improved bipolarity. For the
doped films FIrpic:1-7, triplet energy can be well con-
fined on the guest molecules except for 7 due to its low ET.
In contrast, triplet energy can be well confined on Ir-
(PPy)3 and Ir(piq)3 molecules for all the hosts, giving
comparable τ, ηPL, kr, and knr slightly influenced by the
heterocyclic arylenes. Highly efficient RGB phosphores-
cent OLEDs were fabricated with a single thin EML
containing red, green, and blue phosphors of Ir(piq)3,
Ir(PPy)3, and FIrpic and the developed hosts. One of the
highest ever efficiencies to date was achieved for the blue
and green phosphorescent OLEDs based on 2 with a
pyridine core, especially at a much brighter luminance for
lighting applications. In comparison, the highest efficien-
cies hitherto were achieved for the red phosphorescent
OLEDs based on 6 with a pyrimidine core, which can be
Acknowledgment. This work was partly supported by the
New Energy and Industrial Technology Development Orga-
nization (NEDO) through the “Advanced Organic Device
Project”. S.J.S. greatly appreciates the financial support
from the National Natural Science Foundation of China
(Project No. 51073057) and the Ministry of Science and
Technology (Project No. 2009CB930604).
Supporting Information Available: Experimental procedures
for the synthesis of 1-7. Characterization data of 1-7 including
EIMS and 1H and 13C NMR. Summary of physical properties of
1-7. Calculated and experimental energy levels of 1-7. Tran-
sient PL decays and simultaneous PL spectra of vacuum code-
posited films of FIrpic:1-7, Ir(PPy)3:1-7, and Ir(piq)3:1-7. EL
spectra of RGB phosphorescent OLEDs (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.