Using a double-doping strategy to prepare a bilayer device architecture for high-efficiency red PhOLEDs

Article abstract of DOI:10.1039/c0jm02992k

A simple, bilayered, red phosphorescent organic light-emitting device featuring a doubly-doped emitting layer comprising of the novel hole-transporting host DTAF, the electron-transporting host 27SFBI, and the emitter Os(bpftz)₂(PPhMe₂)₂ covering the interfacial region provides an unusually high current of ca. 1560 mA cm ^-2 at 8.5 V, a maximum brightness of 32700 cd m^-2, external quantum efficiencies as high as 12.3% (10.9% at 1000 cd m ^-2), and a power efficiency of 13.5 lm W^-1. This concise device architecture is very cost-effective and competitive for practical applications.

Full text of DOI:10.1039/c0jm02992k

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/materials | Journal of Materials Chemistry
Using a double-doping strategy to prepare a bilayer device architecture
for high-efficiency red PhOLEDs
a
b
Ming-Tzu Kao, Wen-Yi Hung, Zhen-Han Tsai,b Hong-Wei You,b Hsiao-Fan Chen,a Yun Chic
*
a
*
and Ken-Tsung Wong
Received 8th September 2010, Accepted 18th October 2010
DOI: 10.1039/c0jm02992k
A simple, bilayered, red phosphorescent organic light-emitting device featuring a doubly-doped
emitting layer comprising of the novel hole-transporting host DTAF, the electron-transporting host
27SFBI, and the emitter Os(bpftz)2(PPhMe2)2 covering the interfacial region provides an unusually
high current of ca. 1560 mA cmꢀ2 at 8.5 V, a maximum brightness of 32 700 cd mꢀ2, external quantum
efficiencies as high as 12.3% (10.9% at 1000 cd mꢀ2), and a power efficiency of 13.5 lm Wꢀ1. This concise
device architecture is very cost-effective and competitive for practical applications.
low process yields, adding considerably to the overall production
cost.
To overcome this problem, herein we demonstrate an efficient
Introduction
Phosphorescent organic light-emitting devices (PhOLEDs)
employing transition metal-centered phosphors dispersed in an
appropriate host matrix are attracting much attention because
both singlet and triplet excitons can be harvested to achieve
internal quantum efficiencies of up to 100%.1 Indeed, employing
this strategy may pave the way toward the practical commer-
cialization of OLED displays and lighting. In addition to the use
of efficient phosphors, the nature of the host material also plays
a crucial role affecting the device performance. In most cases, the
host material exhibits either hole-transporting (HT) or electron-
transporting (ET) characteristics. When using a predominantly
HT- or ET-type host, the excitons formed in the device tend to
accumulate close to the interface between the emitting layer
(EML) and the HT or ET layer (HTL or ETL), respectively.
Thus, the inevitable exciton quenching through triplet–triplet
annihilation is usually detrimental to the device efficiency. In
2002, Zhou et al. were the first to develop device structures
possessing double-emission layers (D-EMLs), prepared by
doping both the HT layer and the ET host with a phosphorescent
emitter, to increase the size of the exciton formation zone.2 The
D-EML concept enlarges the exciton recombination zone over
the active layers, leading to significantly reduced levels of triplet–
triplet annihilation relative to those of conventional single
emission layer (S-EML) OLEDs.2,3 To date, most reported
D-EML OLEDs have featured multilayer device structures—
a stepped progression of the highest occupied molecular orbital
(HOMO) and lowest unoccupied molecular orbital (LUMO)
energy levels of the HT materials, host materials, and ET
materials—to facilitate hole and electron injection into the
emissive layer. Such complicated structures require time-
consuming and tedious fabrication processes and usually feature
D-EML red PhOLED configured using a simple bilayer device
structure that does not feature additional HT or ET layers. This
simple device comprises two host materials: a HT-type host
having a high HOMO energy level for hole injection and a high
LUMO energy level for electron blocking, and an ET-type host
having a low LUMO energy level for electron injection and a low
HOMO energy level for hole blocking. We doped the interfacial
zone of these two host materials with a red-emitting Os-based
phosphor to create a wide exciton formation area, resulting in an
efficient device exhibiting a maximum external quantum effi-
ciency of 12.3% (13.5 cd Aꢀ1), a power efficiency of 16.5 lm Wꢀ1
,
a saturated red emission according to the CIE diagram, and
a low driving voltage of 4.5 V to reach a luminance of 1000 cd
mꢀ2. Thus, double-doping is a practical means of fabricating
highly efficient PhOLEDs having a simple bilayer device struc-
ture, potentially reducing the cost of fabrication relative to those
of other contemporary display manufacturing processes.
Results and discussion
Design and synthesis
Scheme 1 displays the molecular structures of the host materials,
which both feature a fluorene core. Our design of the HT-type host
9,9-di[N,N-di(p-tolyl)aminophenyl-4-yl] fluorene (DTAF) was
inspired by the structure of 1,1-bis{4-[N,N-di(p-tolyl)-
amino]phenyl}cyclohexane (TAPC),4 a widely used HTL material
for PhOLEDs that exhibits a high triplet energy (ET ¼ 2.87 eV) and
a high hole mobility (mh ¼ ca. 10ꢀ3 cm2 Vꢀ1 sꢀ1). The low glass
transition temperature (Tg ¼ ca. 79 ꢁC) of TAPC, however, limits
its long-term use. To enhance the thermal stability of TAPC, we
modified its structure by replacing the cyclohexane unit with
a fluorene moiety, thereby obtaining DTAF.5 For the ET-type host
material, we adopted phenylbenzimidazole as the electron-
accepting group because we have previously used 2,20-bis(N-phe-
nylbenzimidazole)-9,90-spirobifluorene successfully as an ET
material.6 In this present study, we introduced an additional p-
phenylene ring between the N-phenylbenzimidazole groups and
aDepartment of Chemistry, National Taiwan University, Taipei, 106,
Taiwan. E-mail: kenwong@ntu.edu.tw; Fax: +886 2 33661667; Tel:
+886 2 33661665
bInstitute of Optoelectronic Sciences, National Taiwan Ocean University,
Keelung, 202. E-mail: wenhung@mail.ntou.edu.tw; Fax: +886
24634360; Tel: +886 2 24622192 ext. 6718
2
cDepartment of Chemistry, National Tsing Hua University, Hsinchu, 300,
Taiwan
1846 | J. Mater. Chem., 2011, 21, 1846–1851
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Scheme 1 Molecular structures of the materials used in this study and an energy level diagram of the device.
the central spirobifluorene to extend the p-conjugation length and
adjust the HOMO and LUMO energy levels, thereby obtaining
the new ET-type host 2,20-bis[4-(1-phenylbenzimidazol-2-
yl)phenyl]-9,90-spirobifluorene (22SFBI). For comparison, we
also synthesized the C2-symmetric 2,7-disubstituted counterpart
27SFBI. We used the highly efficient red-emitting complex bis[3-
(trifluoromethyl)-5-(4-tert-butylpyridyl)-1,2,4-triazolate]dimethyl-
phenylphosphine osmium(II) [Os(bpftz)2(PPhMe2)2, OS1]7 as the
red dopant.
respectively. We observed no crystallization for 22SFBI during
the second DSC scan. In contrast, the C2-symmetric 27SFBI
exhibited a crystallization peak of 203 ꢁC and a melting peak of
ꢁ
321 C. We attribute the amorphous behavior of 22SFBI to its
perpendicular configuration, imparted by the orthogonal
arrangement of the substituents around the spirobifluorene core,
We synthesized DTAF in one step from 9-fluorenone in 92%
yield, as we have described previously.5 Scheme 2 displays the
synthetic pathways toward 22SFBI and 27SFBI. We synthesized
the key intermediate for the synthesis of both compounds,
N-phenylbenzimidazole boronic acid (1), according to literature
procedures.8 Accordingly, Suzuki couplings of 1 with 2,7-
dibromo-9,90-spirobifluorene (2) and 2,20-diiodo-9,90-spirobi-
fluorene (3), in the presence of catalytic amounts of Pd(PPh3)4
and P(t-Bu)3, gave good yields of 27SFBI and 22SFBI, respec-
tively.
Molecular properties
The rigid molecular structures of DTAF, 22SFBI, and 27SFBI
provide them with high thermal stabilities,9 as manifested by
their thermal decomposition temperatures (Td, corresponding to
5% weight loss during thermogravimetric analysis) in the range
358–391 ꢁC (Table 1). For DTAF, differential scanning calo-
rimetry (DSC) revealed a glass transition temperature of 124 ꢁC
and a melting peak of 320 ꢁC; 22SFBI and 27SFBI displayed
well-defined glass transition temperatures of 150 and 160 ꢁC,
Scheme 2 The synthesis of 27SFBI and 22SFBI.
J. Mater. Chem., 2011, 21, 1846–1851 | 1847
This journal is ª The Royal Society of Chemistry 2011

View Article Online
Table 1 The physical properties of DTAF, 22SFBI, and 27SFBI
Compound Tga/Tcb/Tm /ꢁC Td/ꢁC (lmax) Abs/nm Sol./Film (lmax) PL/nm Sol./Film HOMO/eVd LUMO/eVd DEgd/eV ET/eV me/cm2 Vꢀ1 sꢀ1
c
DTAF
22SFBI
27SFBI
124/n.d./360
150/n.d./273
160/203/321
374
358
391
308/311
331/337
350/355
383/383
380, 400/394, 413
401, 421/417, 439
–5.22
–5.77
–5.67
–1.75
–2.61
–2.67
3.47
3.16
3.00
2.87
2.51
2.36
(h) 2.5 ꢂ 10ꢀ3
(e) 2.8 ꢂ 10ꢀ6
(e) 3.0 ꢂ 10ꢀ5
a
Glass transition temperature. b Crystallization temperature. c Melting temperature. d HOMO determined using photoelectron yield spectroscopy (AC-
e
2). LUMO ¼ HOMO + Eg, where Eg was calculated from the absorption onset of the solid film. The value of m was determined from the TOF
measurement at E ¼ 5.7 ꢂ 105 V cmꢀ1, (h) for hole, (e) for electron.
disrupting intermolecular interactions and suppressing its
tendency to crystallize. Thus, upon thermal evaporation, we
would expect these host materials to form homogeneous and
stable amorphous films—an essential property for OLED device
applications.
that of 27SFBI, presumably because of the perpendicular
arrangement of the two active sites for electron migration in
22SFBI. The orthogonal configuration leads to greater spatial
hindrance between neighboring 22SFBI molecules and, subse-
quently, increases the carrier hopping distance.11
The ability of the host materials to conduct electrons or holes
is important for maintaining charge balance within the emissive
layer. To evaluate the carrier mobilities of DTAF, 22SFBI, and
27SFBI, we performed time-of-flight (TOF) measurements.10
Fig. 1(a)–(c) displays typical room temperature TOF transients
for these materials under an applied electric field; we observe
a non-dispersive hole transient photocurrent for DTAF and
dispersive electron transient photocurrents for both 22SFBI and
27SFBI. We determined the transit times (tT) from the cusps of
the double-logarithmic plots of the transient photocurrents
[insets to Fig. 1(a)–(c)]; we then calculated the mobility using the
formula, d2/VtT, where d is the sample thickness and V is the
applied voltage. The hole mobility of DTAF (up to 2 ꢂ 10ꢀ3 cm2
Vꢀ1 sꢀ1) is similar to that of TAPC (mh ¼ ca. 10ꢀ3 cm2 Vꢀ1 sꢀ1),
suggesting that the structural modification from TAPC to DTAF
had a limited effect on the hole transport behavior. The observed
electron mobility (me ¼ ca. 3 ꢂ 10ꢀ5 cm2 Vꢀ1 sꢀ1) of 27SFBI was
greater than that of 22SFBI (me ¼ ca. 2.8 ꢂ 10ꢀ6 cm2 Vꢀ1 sꢀ1).
Under the same electric field (E ¼ 5.7 ꢂ 105 V cmꢀ1), the
observed electron mobility of 22SFBI was significantly less than
We measured the UV-vis absorption and photoluminescence
(PL) spectra of DTAF, 22SFBI, and 27SFBI in solution (CH2Cl2)
and in the form of solid films on quartz slides (Fig. 2, Table 1).
DTAF exhibits similar absorption (ca. 310 nm) and emission (ca.
383 nm) maxima in solution and in its solid state film; in the
latter, an additional shoulder emission (>420 nm), which is also
observed for TAPC,12 is due to aggregation of ditolylamino
subunits. The structure of 27SFBI differs from that of 22SFBI by
the presence of two diphenylbenzimidazole groups attached to
the same biphenyl branch of the spirobifluorene core. The
extended p- conjugation length of 27SFBI results in slightly red-
shifted (by ca. 20 nm) absorption, PL, and phosphorescence
spectra relative to those of its 2,20-disubstituted counterpart
22SFBI. The triplet energies (ET) of DTAF, 22SFBI, and 27SFBI
were 2.87, 2.51 and 2.36 eV, respectively, estimated from the
peaks of the shortest wavelength of the phosphorescence spectra
recorded in EtOH at 77 K. The HOMO energy levels of DTAF,
Fig. 1 Representative TOF transients: (a) DTAF (d ¼ 1.6 mm) at E ¼
1.9 ꢂ 105 V cmꢀ1; (b) 22SFBI (d ¼ 1.9 mm) at E ¼ 7.5 ꢂ 105 V cmꢀ1
;
Fig. 2 Room-temperature absorption and emission (PL) spectra of
(a) DTAF, (b) 22SFBI, and (c) 27SFBI in CH2Cl2 solutions (Solu.) and as
neat films (Film), and corresponding phosphorescence (Phos.) spectra
recorded from their EtOH solutions at 77 K.
(c) 27SFBI (d ¼ 2.1 mm) at E ¼ 4.8 ꢂ 105 V cmꢀ1. Insets: Double-loga-
rithmic plots of (a)–(c). (d) Carrier mobilities plotted with respect to E1/2
at ambient temperature.
1848 | J. Mater. Chem., 2011, 21, 1846–1851
This journal is ª The Royal Society of Chemistry 2011

View Article Online
22SFBI, and 27SFBI were ꢀ5.22, ꢀ5.77 and ꢀ5.67 eV, respec-
tively, determined from the photoemission spectra recorded in
air (AC-2); we deduced the LUMO energy levels from the
difference between the HOMO energy levels and the optical band
gaps. Scheme 1 depicts the energy levels of DTAF, 22SFBI,
27SFBI, and the red emitter Os(bpftz)2(PPhMe2)2, OS1. The
HT-type host (DTAF) possesses an appropriate HOMO energy
level (–5.22 eV) to provide a low hole injection barrier and a high-
lying LUMO energy level (–1.75 eV) to block electrons from the
ETL. In contrast, the ET-type hosts (27SFBI, 22SFBI) possess
appropriate LUMO energy levels (ꢀ2.67 and ꢀ2.61 eV, respec-
tively) to provide low electron injection barriers and low-lying
HOMO energy levels (–5.67 and ꢀ5.77 eV, respectively) to block
holes from the HTL. In addition, we expected the relatively high
hole and electron mobilities of these materials to lead to effective
hole and electron transport in the corresponding devices.
Accordingly, for bilayer devices featuring DTAF/27SFBI or
DTAF/22SFBI, we would expect the charge carriers to be
significantly accumulated at the interface. With judicious selec-
tion of a doping phosphor with suitable HOMO/LUMO energy
levels to cover the interfacial zone between the HT-host and the
ET-host, the accumulated holes and electrons would be smoothly
injected into the guest emitter from HT-host and ET-host,
respectively. The triplet energies of these materials allow excel-
lent triplet energy confinement within the emissive layers. The
wide emissive zone could reduce the possibility of triplet–triplet
annihilation and, thereby, improve the device efficiency with
reduced performance roll-off.
Fig. 3 (a) The current density–voltage–luminance (I–V–L) characteris-
tics. (b) External quantum (hext) and power efficiencies (hP) as a function
of brightness. (c) The EL spectra of devices.
Device performance
In our design of the bilayer device, we employed DTAF as the
HT-type host together with 27SFBI or 22SFBI as the ET-type
host. Individually doping the HT-type host or ET-type host with
the Os-based red emitter provided single emitting layer (S-EML)
reference devices for comparison with the devices having a D-
EML. To smooth the ITO electrode surface and improve the
quality of hole injection from the anode, we spun poly(3,4-ethy-
lenedioxythiophene)/poly(styrene sulfonic acid) (PEDOT:PSS)
onto the pre-cleaned (by UV-ozone) substrate to form a 30 nm-
thick polymer buffer layer. We fabricated two reference devices
featuring an S-EML [(I) ITO/PEDOT:PSS/DTAF (15 nm)/
DTAF:6 wt% Os1 (30 nm)/27SFBI (55 nm)/LiF/Al; (II) ITO/
PEDOT:PSS/DTAF (15 nm)/27SFBI:6 wt% Os1 (25 nm)/27SFBI
(55 nm)/LiF/Al] and one double-doped device [(III) ITO/
PEDOT:PSS/DTAF (15 nm)/DTAF:6 wt% Os1 (5 nm)/27SFBI:6
wt% Os1 (20 nm)/27SFBI (55 nm)/LiF/Al]. For comparison, we
also employed 22SFBI as an ET-type host to fabricate the
D-EML device IV: ITO/PEDOT:PSS/DTAF (15 nm)/DTAF:6
wt% Os1 (5 nm)/22SFBI:6 wt% Os1 (25 nm)/22SFBI (55 nm)/
LiF/Al]. Fig. 3 presents the current density–luminance–voltage
(I–L–V) curves and key characteristics of these devices; Table 2
summarizes the data.
D-EML device III (L ¼ 32 700 cd mꢀ2) was significantly higher
than those of the S-EML devices I (L ¼ 14 400 cd mꢀ2) and II
(L ¼ 18 400 cd mꢀ2) in the applied voltage range from 8.5 to 9 V.
Meanwhile, D-EML device III exhibited a significantly higher
current density (I ¼ 1560 mA cmꢀ2) than those of the S-EML
devices I (I ¼ 1000 mA cmꢀ2) and II (I ¼ 930 mA cmꢀ2). For the
S-EML device I, based on the observed carrier-transport
behavior and the HOMO/LUMO energy levels of DTAF and
27SFBI, it is likely that the electrons that accumulated at the
DTAF–27SFBI interface were injected directly into Os1, which
was present only in the DTAF layer, followed by recombination
with holes within the DTAF layer. A similar mechanism might
have operated in the S-EML device II, in which direct hole
injection from DTAF into Os1 was possible because the differ-
ence in HOMO energy levels between DTAF and Os1 is much
smaller than the difference between DTAF and 27SFBI. There-
fore, the S-EML devices I and II displayed similar properties. In
contrast, we would expect the carrier recombination zone of the
D-EML device III to cover the interfacial area between the two
active layers, mainly due to the mismatched HOMO/LUMO
energy levels. The carriers that reached the interface would have
been preferably trapped by the dopant, suppressing the loss of
luminance efficiency during prolonged device operation. For
these reasons, it would be reasonable to expect a higher lumi-
nance and current density for the D-EML device III, relative to
those of the S-EML devices I and II. As a result, Fig. 3(b) reveals
that the two S-EML devices possessed lower efficiencies relative
The I–L–V characteristics of devices I–IV reveal that holes and
electrons were both readily injected from the anode and cathode,
respectively, to the organic layers; as a result, the operating
voltages were low for both the S-EML and D-EML devices. All
of the L–V curves of these devices feature a steep increase after
the turn-on voltage reached 2 V. The maximum luminance of the
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 1846–1851 | 1849

View Article Online
Table 2 The electroluminescence data of the devices prepared at various host material
Imax
/
max.
hext [%]
max.
hl/cd Aꢀ1
max. hp
[lm Wꢀ1
L ¼ 1000
CIE coordinates
[x, y]
Host
Vona/V
Lmax/cd mꢀ2
mA cmꢀ2
]
nit [V; %]
I
II
III
IV
DTAF
27SFBI
DTAF/27SFBI
DTAF/22SFBI
2
2
2
2
14 400 (9 V)
1000
930
1560
1760
7.3
9.9
12.3
12.4
8.8
10.5
13.5
12.9
9.5
10.6
16.5
11.6
5.7; 5.5
5.2; 7.7
4.5; 10.9
5.9; 8.2
0.65, 0.34
0.66, 0.34
0.66, 0.34
0.66, 0.34
18 400 (8.5 V)
32 700 (8.5 V)
25 700 (10.5 V)
a
Defined as the voltage at which the EL is rapidly enhanced.
to that of the D-EML device III. The S-EML device I exhibited
maximum values of hext and hP of 7.3% (8.8 cd Aꢀ1) and 9.5 lm
Wꢀ1, respectively, with the S-EML device II providing corre-
sponding values of 9.9% (10.5 cd Aꢀ1) and 10.6 lm Wꢀ1, respec-
tively; both these sets of values are far from those provided by the
double-doped to cover the HT–ET interfacial zone, rendering
direct charge injection and trapping on the triplet emitter
possible. This design prevents the accumulation of charge
carriers at the interface, with the wide emission zone reducing the
degree of triplet–triplet annihilation, leading to improved device
efficiency. Optimally, we obtained a red device (CIE coordinates:
x ¼ 0.66, y ¼ 0.34) exhibiting a high external quantum efficiency
of 12.3% (10.9% at 1000 cd mꢀ2) and a power efficiency of 13.5 lm
Wꢀ1. Our simple bilayer device design provides a practical means
of using the double-doping strategy to fabricate highly efficient
PhOLEDs possessing high device efficiency and operating at
a low operation voltage; therefore, it is potentially cost-
competitive with other contemporary display manufacturing
processes.
D-EML device III [12.3% (13.5 cd Aꢀ1) and 16.5 lm Wꢀ1
,
respectively]. The lower current efficiencies of the S-EML devices
are an indication of unbalanced charge recombination, due to the
unipolar characteristics of the active layers used. The improved
efficiency in the D-EML device can be rationalized by consid-
ering that the wide triplet-exciton generation region across the
interfacial area significantly reduced the number of possible
quenching pathways and avoided charge carrier loss. As
a consequence, the D-EML OLED exhibited higher efficiency
and stability relative to those of the conventional S-EML
OLEDs.
Experimental
Next, we compared the behavior of the D-EML structures
employing either 27SFBI or 22SFBI as the ET-type host. The I–
V–L characteristics revealed that the D-EML III incorporating
27SFBI as the ET-type host exhibited a higher luminance and
current density at the same applied voltage than those of the D-
EML device IV incorporating 22SFBI as the ET-type host. At
a luminance of 1000 cd mꢀ2, device III provided values of hext and
hP of 10.9% and 8.3 lm Wꢀ1 at 4.5 V, respectively, considerably
better than those of device IV (8.2% and 4.5 lm Wꢀ1 at 5.9 V,
respectively). These values clearly indicate that the electron
mobilities (Table 1) of 27SFBI and 22SFBI have pronounced
effects on the corresponding device performances. We ascribe the
superior performance of the D-EML device III over that of
device IV to the more balanced charge recombination of the
former. The normalized EL spectra of all four of these devices
(I–IV) displayed [Fig. 3(c)] the deep-red emissions of the dopant
Os(bpftz)2(PPhMe2)2, each with CIE coordinates (CIE1931) of
(0.66, 0.34). No other detectable peaks appeared in the EL
spectra upon increasing the applied voltage.
Synthesis
2,7-Bis[4-(1-phenylbenzimidazol-2-yl)phenyl]-9,90-spirobifluorene
(27SFBI). Tri-tert-butylphosphine (6.4 mL) and Na2CO3(aq) (2 M,
20 mL) were added to a solution of 2,7-dibromo-9,90-spirobi-
fluorene (2) (1.0 g, 2.1 mmol), 4-(1-phenylbenzimidazol-2-yl)phe-
nylboronic acid (1) (1.66 g, 5.28 mmol), and Pd(PPh3)4 (100 mg,
0.086 mmol) in toluene (50 mL) under Ar and then the mixture
was heated under reflux for 48 h. After cooling to room temper-
ature, the reaction mixture was partitioned between CH2Cl2 and
brine. The organic layer was dried (MgSO4) and concentrated.
The crude product was washed with hexane to afford pure 27SFBI
1
(1.51 g, 84%) as a white solid. Mp 321 ꢁC (DSC); H NMR
(CDCl3, 400 MHz) d 7.91 (d, J ¼ 8.0 Hz, 4H), 7.86 (d, J ¼ 6.0 Hz,
2H), 7.63 (d, J ¼ 8.0 Hz, 2H), 7.54 (d, J ¼ 8.0 Hz, 4H), 7.48 (d, J ¼
8.0 Hz, 4H), 7.46–7.35 (m, 8H), 7.32 (d, J ¼ 8.0 Hz, 4H), 7.29 (d,
J ¼ 6.0 Hz, 2H), 7.22 (d, J ¼ 8.0 Hz, 2H), 7.12 (t, J ¼ 8.0 Hz, 2H),
6.91 (s, 2H), 6.79 (d, J ¼ 8.0 Hz, 2H);13C NMR(CDCl3, 100 MHz)
d 151.4, 149.4, 147.9, 142.3, 141.3, 141.2, 140.5, 139.4, 136.8, 136.5,
129.5, 129.2, 128.2, 128.1, 127.6, 127.5, 127.0, 126.6, 126.4, 123.8,
123.0, 122.7, 122.0, 120.2, 119.8, 119.4, 110.1, 66.1; HRMS (m/z,
FAB+) Calcd. for C63H40N4 852.3253, found 852.3245.
Conclusions
We have demonstrated a new strategy—using a simple bilayer
device architecture with double-doped emissive layers—for the
fabrication of highly efficient deep-red PhOLEDs. The success of
this approach is based on the use of tailor-made host materials
and careful selection of the red dopant. The different charge
transport properties and the mismatched HOMO/LUMO energy
levels of the HT-type host DTAF and the ET-type hosts 22SFBI
and 27SFBI allow the charge carriers to accumulate at the HT–
ET interfacial region. The red emitter Os(bpftz)2(PPhMe2)2,
which possesses appropriate HOMO/LUMO energy levels, was
2,20-Bis[4-(1-phenylbenzimidazol-2-yl)phenyl]-9,90-spirobifluorene
(22SFBI). Tri-tert-butylphosphine (5.3 mL) and Na2CO3(aq) (2 M,
15 mL) were added to a solution of 2,20-diiodo-9,90-spirobifluorene
(3) (1.00 g, 1.76 mmol), 4-(1-phenylbenzimidazol-2-yl)phenylbor-
onic acid (1) (1.40 g, 4.46 mmol), and Pd(PPh3)4 (100 mg, 0.086
mmol) in toluene (50 mL) under Ar and then the mixture was
heated under reflux for 48 h. After cooling to room temperature,
the reaction mixture was partitioned between CH2Cl2 and brine.
1850 | J. Mater. Chem., 2011, 21, 1846–1851
This journal is ª The Royal Society of Chemistry 2011

View Article Online
The organic layer was dried (MgSO4) and concentrated. The crude
product was washed with hexane to afford pure 22SFBI (1.53 g,
85%) as a white solid. 1H NMR (CDCl3, 400 MHz) d 7.90 (d, J ¼ 8
Hz, 2H), 7.86 (d, J ¼ 8 Hz, 4H), 7.63 (d, J ¼ 8.0 Hz, 2H), 7.52 (d, J
¼ 8.0 Hz, 4H), 7.47 (d, J ¼ 6.0 Hz, 4H), 7.45–7.35 (m, 8H), 7.32 (d,
J ¼ 8.0 Hz, 4H), 7.31 (d, J ¼ 8.0 Hz, 2H), 7.24–7.14 (m, 4H), 7.10
(t, J ¼ 7.2 Hz, 2H), 6.96 (s, 2H), 6.74 (d, J ¼ 8.0 Hz, 2H); 13C NMR
(CDCl3, 100 MHz) d 151.7, 140.1, 148.8, 142.7, 141.4, 141.4 141.0,
139.6, 137.1, 136.8, 131.8, 131.7 129.7, 129.5, 129.0, 128.6, 128.4,
128.2, 128.2, 127.9, 127.7, 127.2, 126.8, 126.6, 126.5, 123.9, 123.2,
122.9, 122.3, 120.3, 120.1, 119.6, 110.3, 66.0; HRMS (m/z, FAB+)
Calcd. for C63H40N4 852.3253, found 852.3253.
cleaned ultrasonically—sequentially with acetone, methanol, and
deionized water—and then it was treated with UV-ozone. A
hole-injection layer of poly(3,4-ethylenedioxythiophene)-poly-
(4-stylenesurfonate) (PEDOT:PSS) was spin-coated onto the
substrates and dried at 130 ꢁC for 30 min to remove residual
water. Organic layers were then vacuum deposited at a deposi-
tion rate of ca. 1–2 A sꢀ1. Subsequently, LiF was deposited at
ꢀ
ꢀ1
0.1 A s and then capped with Al (ca. 5 A sꢀ1) through shadow
masking without breaking the vacuum. The I–V–L characteris-
tics of the devices were measured simultaneously using a Keithley
6430 source meter and a Keithley 6487 picoammeter equipped
with a calibration Si-photodiode in a glovebox system. EL
spectra were measured using a photodiode array (OTO SD1000)
with a spectral range from 200 to 850 nm and a resolution of
2 nm.
ꢀ
ꢀ
Photophysical measurements
Steady state spectroscopic measurements were conducted both in
solution and solid films prepared by vacuum (2 ꢂ 10ꢀ6 torr)
deposition on a quartz plate (1.6 ꢂ 1.0 cm). Absorption spectra
were recorded with a U2800A spectrophotometer (Hitachi).
Fluorescence spectra at 300 K and phosphorescent spectra at
77 K were measured on a Hitachi F-4500 spectrophotometer
upon exciting at the absorption maxima. Quantum efficiency
measurements were recorded with an integration sphere coupled
with a photonic multi-channel analyzer (Hamamatsu C9920),
which gave anthracene a quantum yield of 23%. The experi-
mental values of HOMO levels were determined with a Riken
AC-2 photoemission spectrometer (PES), and those of LUMO
levels were estimated by subtracting the optical energy gap from
the measured HOMO.
Acknowledgements
This study was supported financially by the National Science
Council of Taiwan.
References
1 (a) M. A. Baldo, S. Lamansky, P. E. Burrows, M. E. Thompson and
S. R. Forrest, Appl. Phys. Lett., 1999, 75, 4; (b) M. A. Baldo,
M. E. Thompson and S. R. Forrest, Nature, 2000, 403, 750.
2 X. Zhou, D. S. Qin, M. Peiffer, J. Blochwitz-Nimoth, A. Werner,
J. Drechsel, B. Maennig, K. Leo, M. Bold, P. Erk and
H. Hartmann, Appl. Phys. Lett., 2002, 81, 4070.
3 (a) G. He, M. Pfeiffer, K. Leo, M. Hofmann, J. Birnstock, R. Pudzich
and J. Salbeck, Appl. Phys. Lett., 2004, 85, 3911; (b) S.-J. Su,
E. Gonmori, H. Sasabe and J. Kido, Adv. Mater., 2008, 20, 4189;
(c) J.-H. Lee, C.-L. Huang, C.-H. Hsiao, M.-K. Leung, C.-C. Yang
and C.-C. Chao, Appl. Phys. Lett., 2009, 94, 223301; (d)
H. Fukagawa, K. Watanabe, T. Tsuzuki and S. Tokito, Appl. Phys.
Lett., 2008, 93, 133312; (e) T. H. Zheng and W. C. H. Choy, J.
Phys. D: Appl. Phys., 2008, 41, 055103.
4 (a) K. Goushi, R. Kwong, J. J. Brown, H. Sasabe and C. Adachi, J.
Appl. Phys., 2001, 95, 5048; (b) D. Tanaka, H. Sasabe, Y.-J. Li, S.-
J. Su, T. Takeda and J. Kido, Jpn. J. Appl. Phys., 2007, 46, L10; (c)
D. Tanaka, N. Ide and J. Kido, Jpn. J. Appl. Phys., 2007, 46, 1186;
(d) H. Sasabe, T. Chiba, S.-J. Su, Y.-J. Pu, K. Nakayama and
J. Kido, Chem. Comm., 2008, 5821; (e) H. Sasabe, E. Gonmori,
T. Chiba, Y.-J. Li, D. Tanaka, S.-J. Su, T. Takeda, Y.-J. Pu,
K. Nakayama and J. Kido, Chem. Mater., 2008, 20, 5951.
5 W.-Y. Hung, L.-C. Chi, W.-J. Chen, Y.-M. Chen, S.-H. Chou and K.-
T. Wong, J. Mater. Chem., 2010, 45, 10113.
6 Y.-L. Liao, C.-Y. Lin, K.-T. Wong, T.-H. Hou and W.-Y. Hung, Org.
Lett., 2007, 9, 4155.
7 Y. L. Tung, S. W. Lee, Y. Chi, Y. T. Tao, C. H. Chien, Y. M. Cheng,
P. T. Chou, S. M. Peng and C. S. Liu, J. Mater. Chem., 2005, 15, 460.
8 (a) Z. Ge, T. Hayakawa, S. Ando, M. Ueda, T. Akiike, H. Miyamoto,
T. Kajita and M. Kakimoto, Adv. Funct. Mater., 2008, 18, 584; (b)
M. Debeaux, M. W. Thesen, D. Schneidenbach, H. Hopf,
Time-of-flight (TOF) mobility measurements
The samples for the TOF measurements were prepared through
vacuum deposition in the configuration glass/Ag (30 nm)/organic
(2–3 mm)/Al (150 nm); they were then placed inside a cryostat and
maintained under vacuum. The thickness of organic film was
monitored in situ with a quartz crystal sensor and calibrated by
a thin film thickness measurement (K-MAC ST2000). A pulsed
nitrogen laser was used as the excitation light source through the
transparent electrode (ITO) induced photogeneration of a thin
sheet of excess carriers. Under an applied dc bias, the transient
photocurrent was swept across the bulk of the organic film
toward the collection electrode (Al), and then recorded with
a digital storage oscilloscope. Depending on the polarity of the
applied bias, selected carriers (holes or electrons) are swept
across the sample with a transit time of tT. With the applied bias
V and the sample thickness D, the applied electric field E is V/D,
and the carrier mobility is then given by m ¼ D/(tT E) ¼ D2/(VtT),
in which the carrier transit time, tT, can be extracted from the
intersection point of two asymptotes to the plateau and the tail
sections in double-logarithmic plots.
€
S. Janietz, H. Kruger, A. Wedel, W. Kowalsky and H. Johannes,
Adv. Funct. Mater., 2010, 20, 399.
9 T. P. I. Saragi, T. Spehr, A. Siebert, T. Fuhrmann-Lieker and
J. Salbeck, Chem. Rev., 2007, 107, 1011.
10 P. M. Borsenberger and D. S. Weiss, Organic Photoreceptors for
Imaging Systems; Marcel Dekker: New York, 1993.
11 Y.-L. Liao, W.-Y. Hung, T.-H. Hou, C.-Y. Lin and K.-T. Wong,
Chem. Mater., 2007, 19, 6350.
12 (a) W. Stampor, Chem. Phys., 2000, 256, 351; (b) J. Kalinowski,
G. Giro, M. Cocchi, V. Fattori and P. Di Marco, Appl. Phys. Lett.,
2000, 76, 2352.
OLED device fabrications
All chemicals were purified through vacuum sublimation prior to
use. The OLEDs were fabricated through vacuum deposition of
the materials at 10ꢀ6 torr onto ITO-coated glass substrates
having a sheet resistance of 15 U sqrꢀ1. The ITO surface was
This journal is ª The Royal Society of Chemistry 2011
J. Mater. Chem., 2011, 21, 1846–1851 | 1851