S.-R. Park et al. / Dyes and Pigments 141 (2017) 217e224
223
To understand the reason of the difference in device perfor-
mances, we fabricated hole only devices (HODs) as well as electron
only devices (EODs) as follow (see also Fig. 4(b) and (c)):
HOD A: ITO/PEDOT:PSS (55 nm)/TAPC (20 nm)/mCP (10 nm)/
mCPPy: FIrpic (10%, 25 nm)/MoO
HOD B: ITO/PEDOT:PSS (55 nm)/TAPC (20 nm)/mCP (10 nm)/
CPDPy: FIrpic (10%, 25 nm) MoO (10 nm)/Al (100 nm).
3
(10 nm)/Al (100 nm).
3
HOD C: ITO/PEDOT:PSS (55 nm)/TAPC (20 nm)/mCP (10 nm)/
DCPPy: FIrpic (10%, 25 nm)/MoO3 (10 nm)/Al (100 nm).
EOD A: ITO/TmPyPB (10 nm)/mCPPy: FIrpic (10%, 25 nm)/
TmPyPB (35 nm)/LiF (1 nm)/Al (100 nm).
EOD B: ITO/TmPyPB (10 nm)/CPDPy: FIrpic (10%, 25 nm)/
TmPyPB (35 nm)/LiF (1 nm)/Al (100 nm).
EOD C: ITO/TmPyPB (10 nm)/DCPPy: FIrpic (10%, 25 nm)/
TmPyPB (35 nm)/LiF (1 nm)/Al (100 nm).
3
Here, the purpose of molybdenum oxide (MoO ) and TmPyPB in
HOD and EOD devices is to block the injection of electron and hole
carriers from the Al (~ꢂ4.30 eV) and ITO (~ꢂ2.90 eV) electrode,
respectively. Actually, we used CH8000 (PEDOT:PSS) as a HIL to
reduce the leakage current level mentioned above (see also
Fig. 4(a). Commonly, CH8000 based devices exhibit maximum
current efficiency at low current density, while the efficiency drops
sharply at high current density which might be due to the seriously
increased amount of PSS acting as an insulator [28,29]. As a result,
exciton may be formed at the interface close to the anode side since
the hole injection into the EML is pretty limited even at the high
current density condition. Hence, it is very important to inject the
electron carriers deeply into an anode side to increase the proba-
bility of charge recombination for outstanding device perfor-
mances. To estimate the relative position for charge recombination,
we prepared the half devices with EML doped by 10 wt% of FIrpic.
Particularly, we plotted the relative current density of those devices
by dividing the hole current density of HODs by the electron cur-
rent density of the EODs [e.g., J(hole)/J(electron) where J: current
Fig. 7. Normalized EL spectra of fabricated blue PHOLEDs (at a brightness of 1000 cd/
m ).
2
Device B: ITO/PEDOT:PSS (55 nm)/TAPC (20 nm)/mCP (10 nm)/
CPDPy: FIrpic (10%, 25 nm)/TmPyPB (35 nm)/LiF (1 nm)/Al
(
100 nm).
Device C: ITO/PEDOT:PSS (55 nm)/TAPC (20 nm)/mCP (10 nm)/
DCPPy: FIrpic (10%, 25 nm)/TmPyPB (35 nm)/LiF (1 nm)/Al
100 nm).
(
We used CH8000 (PEDOT:PSS) as a HIL to reduce the leakage
current level of the devices. Meanwhile, we used TAPC and TmPyPB
as a HTL and an ETL to suppress the exciton energy quenching of
EML because they show higher triplet energy levels [T
1
(
(
TAPC) ¼ 2.9 eV; T
¼ 2.7 eV).
Fig. 5(a) shows the current density - voltage - luminance (J e V e
1
(TmPyPB) ¼ 2.8 eV] than that of FIrpic
T
1
2
density, mA/cm ] as shown in Fig. 6(b), where the J - V character-
istics shown in Fig. 6(a) were collected from the HODs and EODs,
respectively. From this plot, we could expect that the DCPPy could
give the most desirable device performance because the relative
charge density value of the device fabricated with DCPPy showed
fastest electron transporting behavior. Very interestingly, the de-
vices fabricated with CPDPy also showed fast electron transporting
behavior so that the full device also show relatively high efficiency
values. In contrast, the devices containing EML with mCPPy
showed poor electron transporting behavior which cannot gives
high performances when it was applied to the full devices.
L) characteristics of the devices fabricated in this study. At a given
constant voltage of 5.0 V, current density values were recorded to
2
0
.39, 0.49, 0.26, and 0.12 mA/cm from Reference, Device A, Device
B and Device C, respectively. The turn-on voltages (Von) of 4.0, 4.0,
.0, and 4.1 V were obtained from Reference, Device A, Device B
and Device C, respectively. Meanwhile, the operation voltages (Vop
4
)
2
to reach 1000 cd/m were 6.2, 6.3, 6.3, and 6.5 V for Reference,
Device A, Device B, and Device C, respectively (see Table 2). At a
given constant luminance of 1000 cd/m , the current and power
2
efficiencies were 33.6 cd/A and 16.1 lm/W for Reference, 26.0 cd/A
and 12.3 lm/W for Device A, 33.9 cd/A and 16.4 lm/W for Device B
and 43.2 cd/A and 19.4 lm/W for Device C, respectively, as shown in
Fig. 5(b) And, those are also summarized in Table 2. Those efficiency
data correspond to 17.0, 13.4, 17.2 and 21.6% of EQEs for Reference
and Device A, B, and C, respectively. Very interestingly, the EQE
values increased the order of Device A < Reference < Device
Fig. 7 shows the normalized EL spectra of Reference and Device
2
A-C at the brightness of 1000 cd/m . All spectra have FIrpic emis-
sion with a peak wavelength (lmax) at 472 nm for Reference and
Device A-C, respectively.
2
4. Conclusions
B < Device C at the given brightness of 1000 cd/m as shown in
Fig. 5(c) Likewise, the maximum current and power efficiencies
were 36.3 cd/A and 27.2 lm/W for Reference, 31.9 cd/A and 22.3 lm/
W for Device A, 37.9 cd/A and 26.5 lm/W for Device B, 43.3 cd/A
and 24.4 lm/W for Device C, respectively, as summarized in Table 2.
In particular, maximum EQE values also increased in the order of
Device A < Reference < Device B < Device C (e.g., Reference:
In this study, we found that the introduction of carbazole moi-
eties and dipyridyl amine moieties is very useful to prepare the host
materials for highly efficient blue PHOLEDs. We could control the
electron transporting properties by changing the numbers of
dipyridyl amine moieties. New host materials showed moderate
current efficiency as well as EQE values when it was applied to the
blue PHOLEDs. Especially, we found that the selection of fast elec-
tron moving EML is profitable to prepare highly efficient blue
PHOLEDs. In this study, DCPPy was selected as the best candidate
for highly efficient blue PHOLEDs which could give high EQE up to
18.3%; Device A: 16.3%; Device B: 18.9%; Device C: 21.6%). As a
result, we found that Device C (DCPPy) showed a better efficiency
behavior than that of the other materials. Meanwhile, the device
performance of Device A (mCPPy) and Device B (CPDPy) is very
lower or similar to that of Reference (mCP).
21.6%.