1. Materials and Device Fabrication
PEDOT: PSS of CH8000 (Heraeus) used as hole injection layer (HIL) was spin-coated
on a precleaned patterned ITO and then annealed at 160 oC in 15 min. Small molecular cross-linkable hole transport material (x-HTL) [4-(9-phenyl-9H-carbazol-4-yl)
-phenyl]-bis- (4'-vinyl-biphenyl-4-yl)-amine (PbV, from Lumtec), homopolymer 4-Butyl-N,N-diphenylaniline
(PolyTPD, from Sigma) and Poly(9,9-dioctylfluorene-alt-N-(4-sec-butylphenyl)-diphenylamine
(TFB, from Lumtec) were used for HTL. PolyTPD or TFB (0.4 wt% in toluene) was mixed
with PbV (0.4 wt% in toluene) at a concentration of 20 wt% and spin-coated on pre-annealed
PEDOT:PSS and then annealed at 230 oC for 1 h. Solution of host 7,7- dimethyl-5-phenyl-2-(9-phenyl-9H-carbazol-3-yl)-5,7-dihydroindeno
[2,1-b] carbazole (PCIC) [8] or 2,2’,2’’-(1,3,5-Benzinetriyl)-tris(1-phenyl-1-H-benzimidazole) (TPBi) was mixed
tris [2-(p-tolyl)pyridine] iridium (III) (green dopant, Ir(mppy)3) at the weight raito of 9:1 to use as EML. The sovlent for host and dopant is chlorobenzene.
The EML solution was then spin-coated on the annealed HTL to form a smooth film, followed
by being annealed at 105 oC for 10 min. Consequently, the remaining layers were deposited by vacuum evaporation.
An exciton-blocking layer (TPBi) was used to confine exciton inside EML. A 2-[4-(9,10-dinaphthalen-2-yl-anthracen-2-yl)-phenyl]-1-phenyl-1H-benzoimidazole
(DNAPPBi) [5] mixed with Lithium quinolate (Liq, from Sigma) at an equal concentration to serve
as electron transport layer (ETL), and a thin Liq layer of 1 nm was used as electron
injection layer (EIL). Finally, a thick Al (Sigma) was evaporated to be served as
a cathode.
3. Result and Discussion.
Park et al. have exclusively investigated the charge transport characteristics of
soluble p-type host PCIC, which showed higher transport ability for hole compared
to electron owing to the carbazole electron-donating unit in its molecular structure
[5,8]. The thermal evaporated TPBi is widely known as an efficient ETL [9-12] and also as an n-type host material [13-15]. Here, we incorporated both PCIC and TPBi as soluble hosts into solution-processed
OELDs and investigated the dependence of device performance on these hosts. Devices
having the following structures were then fabricated (Fig. 1(a) and (b)):
Fig. 1. Device structures with different host materials: (a) PCIC as host; (b) TPBi as host; (c) Current density and luminance plotted versus voltage; (d) Current efficiency (CE) and power efficiency plotted versus luminance; (e) External quantum efficiency (EQE) plotted versus luminance; (f) Electroluminescent (EL) spectra at 1000 a luminance of cd/m2for Device 1 and Device 2.
Device 1: ITO/ CH8000 (400Å)/ PbV:PolyTPD (20%, 170 Å)/ PCIC:Ir(mppy)3 (10%,300 Å)/ TPBi (50 Å)/ DNAPPBi:Liq (1:1, 450 Å)/ Liq (10 Å)/ Al (100 Å).
Device 2: ITO/ PEDOT:PSS (400 Å)/ PbV: PolyTPD (20%, 170 Å)/ TPBi:Ir(mppy)3 (10%, 300 Å)/ TPBi (50 Å) / DNAPPBi:Liq (1:1, 450 Å)/ Liq (10 Å)/ Al (1000 Å).
A thin TPBi with moderately high triplet energy (T1) of 2.74 eV was used to confine exciton inside EML [16]. Device characteristics are illustrated in Fig. 1 and also in Table 1. The voltage for turning on device (Von) and operational voltage (Vop) were 2.7 V and 4.6 V in Device 1, 3.3 V and 6.4 V in Device 2, respectively. Device
1 also exhibits a better performance with higher maximum current efficiency (CEmax) of 29.5 cd/A and maximum external quantum efficiency (EQEmax) of 8.5 % compared to Device 2 (20.7 cd/A and 6.3 %). The higher Von could be attributed primarily to the high injection barrier for hole caused by the
deep HOMO of 6.2 eV in TPBi [16,17]. This inadequate hole injection and weak hole transport ability induced the recombination
of exciton toward the mixing region at the HTL/EML interface, which would impair device
performance (Fig. 1(c)-(f)). In particular, the spectra of both devices overlapped at low voltage (3.5 V)
implying that the RZ of both devices were located close to each other, probably around
the center of EML. As the voltage further increased, the EL spectra of Device 2 broadened
to the longer wavelength region, while Device 1's EL spectra remained stable (Fig. 2). In the abovementioned voltage range, CIE color coordinates change from (0.3038,
0.6211) to (0.2992, 0.6211) for Device 1 and significantly change from (0.3032, 0.62)
to (0.3101, 0.6167) for Device 2. This is well known to be due to the shift of RZ
toward the HTL side [17]. However, how far the RZ was shifted still has not been extensively studied. Here,
${\Delta}$d which is the distance between the shifted RZ at high voltage and that
at 4 V, will be quantitatively investigated. By applying the planar microcavity effect,
the EL spectra at different voltages were able to be reproduced using the below equation
[13,17]:
where I(${\Delta}$d, ${\lambda}$), expressed as a function of wavelength (${\lambda}$)
and shifted distance ${\Delta}$d, is the EL spectrum intensity at higher voltages,
and Io is the spectrum intensity at low voltage (4 V). Rcathode and n are the reflectance of Al cathode which is assumed to be 0.95, and refractive
index of organic material (typically, n=1.75), respectively.
Fig. 2. The normalized EL spectra and the zoomed-in spectra in (a)-(b) Device 1; (c)-(d) Device 2; (e) Overlapping EL spectra at low voltage (3.5 V) of both devices; (e) Hole mobility of PbV:polyTPD and PbV:TFB calculated by space-charge limitted current model from hole only device with structure: ITO/PEDOT:PSS (40 nm)/HTL (20 nm)/MoO3(10 nm)/Al (100 nm).
Fig. 3(a)-(c) displays the calculated result of ${\Delta}$d and an illustration of the shift
of RZ at higher voltages. The overlapping spectra of both devices at low voltage (Fig. 2(e)) suggested that the RZ in both devices could be initially situated in close proximity
to one another. Stated differently, they should be dispersed throughout EML's center.
As voltage increased to 6 V and 8 V, the RZ in Device 2 was shifted toward the HTL
side by a distance of 12 nm and 17 nm, respectively. At 8 V, the RZ was most likely
pinned at the mixing region (Fig. 3(c)). For further details, when voltage increases from 4 V to 5 V, 6 V, 7 V, and 8 V,
the RZ gradually was shifted toward the HTL by a distance of 7 nm, 12 nm, 15 nm, and
17 nm, respectively. The significant movement of RZ with applied voltage in Device
2 could be induced by the presence of the interface mixing, the stronger field-dependent
mobility of TPBi [18] compared to PCIC [5], or the carrier trapping on dopant due to deep HOMO of TPBi [19].
Fig. 3. Fitting of normalized EL spectra at (a) 8 V; (b) 6 V by using Eq.(1); (c) The shifted distance (∆d) of RZ at higher voltages in Device 2.
To overcome the issue of RZ shift and spectral instability, the HTL was modified by
replacing PolyTPD with TFB having higher hole transport ability and deeper HOMO. This
could facilitate the flow of hole into EML, and further improve the charge balance
due to enhanced hole mobility of PbV:TFB (~9.7${\times}$10-4 cm2/(Vs)) compared to PbV:PolyTPD (~5${\times}$10-4 cm2/(Vs) at an electric field of 2${\times}$105 V/cm (Fig. 2(f)) [7]. The device structures are shown below (Fig. 3(a) and (b)):
Device 3: ITO/ CH8000 (400 Å)/ PbV:TFB (20%, 170 Å)/ PCIC:Ir(mppy)3 (10%,300 Å)/ TPBi (50 Å)/ DNAPPBi:Liq (1:1, 450 Å)/ Liq (10 Å)/ Al (100 Å).
Device 4: ITO/ PEDOT:PSS (400 Å)/ PbV:TFB (20%, 170 Å)/ TPBi:Ir(mppy) (10%, 300 Å)/
TPBi (50 Å) / DNAPPBi:Liq (1:1, 450 Å)/ Liq (10 Å)/ Al (1000 Å).
Fig. 3(c)-(f) depict the device performances with PbV:TFB as hole transport layer (Table 1). By employing PbV:TFB, the Von /Vop were reduced to 2.4 V/5.3 V and 3.1 V/5.8 V, respectively, for Device 3 and Device
4. The CEmax and EQEmax were measured to be 27 cd/A and 7.5 % for Device 3, 22 cd/A and 6.6 % for Device
4, respectively. The enhanced performance of Device 4 was ascribed to the increase
of hole current flowing into EML being ascribed to the reduced hole injection barrier
and higher mobility of PbV:TFB (Fig. 2(f)), which is beneficial for charge balance. As the applied voltage increases from 3.5
V to 8 V, the CIE color coordinates change from (0.3046, 0.6201) to (0.2998, 0.62)
for Device 3 and from (0.315, 0.6148) to (0.3134, 0.6136) for Device 4 (Table 1). Device 4 shows negligible change in color coordinates implying that shifting RZ
away from interface mixing could result in significant improvement in spectral stability.
Fig. 4. Device structures with different host materials: (a) PCIC as host; (b) TPBi as host; (c) Current density and luminance plotted versus Voltage; (d) Current efficiency and power efficiency plotted versus luminance; (e) external quantum efficiency (EQE) plotted versus luminance; (f) EL spectra of the two devices at 1000 cd/m2(Device 3 and Device 4).
Fig. 5. The EL spectra and the zoom in the region at different applied voltages of (a)-(b) Device 3; (c)-(d) Device 4.
Table 1. Summary of performances for Devices 1-4
|
Device
|
Von /Vop
(V)
|
CE (cd/A)
|
PE (lm/W)
|
EQE (%)
|
CIE (x, y)
|
|
Max
|
1000 (cd/m2)
|
Max
|
1000 (cd/m2)
|
Max
|
1000
(cd/m2)
|
3.5 V
|
8 V
|
|
Device 1
|
2.7/4.6
|
29.5
|
29.1
|
23.5
|
18.4
|
8.5
|
8.5
|
(0.3038, 0.6211)
|
(0.2992, 0.6211)
|
|
Device2
|
3.3/6.4
|
20.7
|
20.1
|
16.7
|
10.4
|
6.3
|
5.7
|
(0.3032, 0.6200)
|
(0.3101, 0.6167)
|
|
Device3
|
2.4/5.3
|
27.0
|
27.0
|
19.3
|
15.5
|
7.5
|
7.4
|
(0.3046, 0.6201)
|
(0.2998, 0.6200)
|
|
Device 4
|
3.1/5.8
|
22
|
21.6
|
17.4
|
11.4
|
6.6
|
6.2
|
(0.315, 0.6148)
|
(0.3134, 0.6136)
|