Effects of novel transition metal oxide doped bilayer structure on hole injection and transport characteristics for organic light-emitting diodes

Effects of novel transition metal oxide doped bilayer structure on hole injection and transport characteristics for organic light-emitting diodes

Synthetic Metals 243 (2018) 121–126 Contents lists available at ScienceDirect Synthetic Metals journal homepage: www.elsevier.com/locate/synmet Effe...

2MB Sizes 0 Downloads 9 Views

Synthetic Metals 243 (2018) 121–126

Contents lists available at ScienceDirect

Synthetic Metals journal homepage: www.elsevier.com/locate/synmet

Effects of novel transition metal oxide doped bilayer structure on hole injection and transport characteristics for organic light-emitting diodes

T

Chi-Ting Tsaia, Ya-Han Liua, Jian-Fu Tanga, Po-Ching Kaob, Chung-Hao Chianga, ⁎ Sheng-Yuan Chua,c, a

Department of Electrical Engineering, National Cheng Kung University, Tainan 70101, Taiwan Department of Electrophysics, National Chiayi University, Chiayi 60004, Taiwan c Advanced Optoelectronic Technology Center, National Cheng Kung University, Tainan 70101, Taiwan b

A R T I C LE I N FO

A B S T R A C T

Keywords: Organic light-emitting diodes Transition metal oxide Electrical doping Graded doping Hole injection

A contemporary hole injection bilayer structure (HIBL) based on molybdenum trioxide (MoO3)-doped N,N′-Di(1naphthyl)-N,N′-diphenyl-(1,1′-biphenyl)-4,4′-diamine (NPB) has been demonstrated and compared with several efficient transition metal oxide (TMO)-based hole injection layers (HILs). Device performances of OLEDs was significantly improved by the utilization of this HIBL. Results of electroluminescence (EL) spectra, hole-only current density-voltage test and capacitance measurement by impedance spectroscopy (IS) are indicative of enhanced hole injection and transport characteristics. Moreover, ultraviolet photoelectron spectroscopy (UPS) results authenticated a cascading highest occupied molecular orbital (HOMO) energy level contributed to both improved hole injection and transport properties, therefore leads to better carrier balance and device efficiency in OLEDs.

1. Introduction Innovation of organic light-emitting diode (OLED) has attracted researchers to explore the next generation LEDs. Vast applications like solid state lighting and flat panel display are economically viable by the unique properties of OLEDs [1]. To attain high efficiency in light emission, one primary issue is to overcome the energy losses during electron-photon conversion [2]. Two factors that influence carrier injection and transport characteristics to diminish the energy losses and enhance efficiency are injection barrier at electrode/organic interface and charge carrier transport from this interface to active zone [2,3]. Diverse techniques that assist carrier injection from electrodes to organic layers are being pursued by researchers. To be precise, the effective approaches are interlayer sandwiched between electrode and organic layer [4,5], modification of electrode work function [6–8] and addition of p-type/n-type dopant in carrier transporting layer [9,10]. Particularly, transition metal oxides such as MoO3, WO3, V2O5, ReO3 are utilized for hole injection recently due to their high work function [10–13]. These TMOs fulfill their existence as thin hole injection interlayers and also as p-type dopants in hole transport layers (HTLs). These TMO-based thinfilms are practiced effectively to enhance hole injection characteristics in OLEDs and hole-extraction properties in organic photovoltaic devices (OPVs) [14]. On the other hand,



Corresponding author. E-mail address: [email protected] (S.-Y. Chu).

https://doi.org/10.1016/j.synthmet.2018.06.008 Received 9 February 2018; Received in revised form 17 June 2018; Accepted 18 June 2018 Available online 23 June 2018 0379-6779/ © 2018 Elsevier B.V. All rights reserved.

introducing a cascading or graded interface and transport layer are known to improve carrier flow and device performance [15]. Besides, it has also been found that employing a graded doping profile in emissive layer (EML) can substantially extend OLED lifetime by widening the recombination zone [16]. Significant efforts have been made to focus on the graded doping system based on emitting host-guest system [17–24], organic-organic [15,25,26] and metal-polymer materials [27]. Much less attention was paid to other inorganic dopants, especially, TMOs. Moreover, the underlying mechanisms behind its role in augmenting carrier injection and transport are not completely understood. In this work, we report a novel hole injection bilayer (HIBL) structure based on NPB:MoO3 for gradual tuning the energy barrier distribution of material with respect to altered doping concentration. Hole transport in this HIBL is analyzed by the current density-voltage characteristics of hole-only devices. Observed by ultraviolet photoelectron spectroscopy (UPS), this HIBL exhibits an energetically stepwise structure which accomplishes effective hole injection and transport properties those are desired to achieve improved OLED characteristics. Additionally, capacitance-voltage measurement was conducted by impedance spectroscopy (IS) to authenticate the charge injection ability by monitoring the carrier dynamics in the emitting layer.

Synthetic Metals 243 (2018) 121–126

C.-T. Tsai et al.

2. Experimental details

molecular orbital (HOMO) level of NPB (5.5–5.7 eV) [28,29] and work function of MoO3 (5.6–5.7 eV) [11,30], charge transfer (CT) complexes can be easily procured by co-deposition of NPB and MoO3. As shown in Supplementary Fig. S1, an additional absorption peak can be evidently observed in the UV–vis absorption spectra at 495 nm [13], authenticating the presence of CT complexes of MoO3+:NPB− [31]. While the doping concentration of MoO3 increased, intensity of this absorption peak was also increased, which is attributed to more CT complexes produced, suggesting that extra free holes were generated in NPB:MoO3. As seen in the inset of Supplementary Fig. S1, intensity of this additional peak promptly increased as the dopant concentration increased, then seemed to gradually saturate by further rising MoO3 concentration over 33%. In accordance with the increase in free holes, the Fermi level shift of NPB toward the HOMO level was projected [32], hence lowered the energy difference between Fermi level and HOMO energy state [33]. Consequently, by gradually employing different MoO3 doping concentrations from high to low from the anode to HTL, a descending arrangement of HOMO energy level should be expected. To confirm our concept, appropriate concentration selection of each sublayer in HIBL is crucial for the best hole injection. We therefore fabricated a batch of OLEDs with a wide variety of concentration combinations. The thickness of HIBL was nominated as 6 nm, which is at a typical thickness range of injecting layers. Supplementary Fig. S2 presents the J–V characteristics of OLEDs with various HIBL concentration combinations. Device with combination 35%/20% demonstrated the optimum J–V properties compared to other devices under identical processing conditions, indicating superior hole injecting capability. Based on these results, we selected 35%/20% as the concentration combination of our HIBL. Fig. 1 (a) and (b) shows the J–V-L characteristics of the fluorescent green OLEDs with the structure of ITO/HIL/NPB(40 nm)/Alq3(40 nm)/ LiF(1 nm)/Al(140 nm). For comparison, devices with typical TMOsbased HILs consist of pure MoO3, uniformly doped NPB:MoO3 (25%) and also device without HIL (normal device) were fabricated. According to literature, optimum thickness of pure MoO3 HIL is usually around 1∼2 nm, therefore, 1-nm is chosen as the thickness of pure MoO3 HIL. Diverse HIL schematic diagrams of devices A–D are shown in Fig. 2. It can be clearly seen that both light emission and charge injection properties in device D showed inferior behavior. This is attributed to a rather high hole injection barrier between ITO anode and hole transport layer NPB. In contrast, with MoO3-based HILs, devices A–C demonstrated enhanced charge injection characteristics and also much higher luminance at a given current density. Particularly, device with the stepwise structure (device A) showed extraordinary hole injection characteristics and light emission compared to other devices. Regarding J–V curves, driving voltage at current density of 100 mA/cm2 for device A was 5.88 V, while for device B was 6.20 V, for device C was 6.42 V and for device D was 8.89 V. In case of luminance, the maximum luminance achieved for device A was 31,150 cd/m2 (at 8.5 V), which was superior to device B 22,010 cd/m2 (at 10 V), device C 20,311 cd/m2 (at 10 V) and device D 11,930 cd/m2 (at 10.5 V). Moreover, luminance at 300 mA/cm2 for devices A–D were 10943, 9757, 9401 and 9266, respectively. These results denote that device with our stepwise HIBL should exhibit improved luminance efficiency. Fig. 1(c) presents the current efficiency and power efficiency as a function of current-density (η -J) for devices A–D. It is evident that device D showed a lower efficiency (2.59 cd/A at 100 mA/cm2) due to higher hole injection barrier at anode/organic interface. On the other hand, devices A–C showed 3.37, 3.10 and 3.11 at 100 mA/cm2, respectively. The power efficiency also showed similar tendency with efficiencies 2.02, 1.57, 1.45 and 0.98 lm/W for devices A–D respectively at 100 mA/cm2. The enhanced current efficiency is due to more holes injected from anode as hole injection is significantly improved by stepwise HIBL. However, the maximum current efficiency increased was very little, i.e., from 3.13 (Device D at 380 mA/cm2) to 3.65 (Device A at 315 mA/cm2) cd/A, which we inferred it as a result influenced by unbalanced charge effects

All materials used in this work were purchased from commercial vendors (purities > 99.9%) and were used as received. Patterned indium tin oxide (ITO)-coated glass substrates (Global Tech International) with sheet resistance of approximately 11 Ω/square were used. Diverse device structures for various purposes were described in the text. ITO was used as anode, molybdenum trioxide (MoO3) was used as anode buffer layer or HIL dopant, N,N′-Di(1-naphthyl)-N,N′-diphenyl-(1,1′biphenyl)-4,4′-diamine (NPB) was used as hole transport layer and the host material in HIL, tris-(8-hydroxyquinoline)aluminum (Alq3) was used as light-emitting layer and electron transport layer, and lithium fluoride (LiF)/Al was used as cathode. In phosphorescent OLEDs, 4,4′cyclohexylidenebis[N,N-bis(4-methylphenyl)benzenamine] (TAPC) was used as hole transport layer, bis[3,5-di(9H-carbazol-9-yl)phenyl]diphenylsilane (SimCP2) and bis[2-(4,6-difluorophenyl)pyridinato-C2,N] (picolinato)iridium(III) (FIrpic) were used as host and guest materials of phosphorescent blue emitting layer, and 2,2′,2″-(1,3,5-Benzinetriyl)tris(1-phenyl-1-H-benzimidazole) (TPBi) was used as electron transport layer. In advance of organic film deposition, the substrates were immersed in an ultrasonic bath of detergent solution, isopropanol, ethanol and de-ionized (DI) water for 15 min each sequentially, then followed by a rinse of DI water. Eventually, the substrates should undergo UVozone (Jelight UVO-42) treatment for 20 min before use. All organic layers, inorganic oxides, metal electrodes and doped layers were deposited at room temperature by thermal evaporation under high vacuum of 5 × 10−7 Torr. Electrical doping of MoO3 and NPB was implemented by thermal co-evaporation from two individual source boats. The active area of devices was 2.5 × 2.5 mm2 defined by the overlap of ITO and Al cathode. Deposition rates, thicknesses and doping concentrations were monitored and controlled via separated quartz crystal oscillators (INFICON, SQM-160) and further calibrated by surface profiling (KLA Tencor, Alpha-Step IQ). The deposition rate of organic materials was 1.0 Å /s, while deposition rates of LiF and Al were 0.1 Å /s and 3–5 Å /s, respectively. For NPB:MoO3 layers, deposition rate of dopant MoO3 was controlled as 0.1 Å /s and the deposition rate of host material was adjusted according to different volume ratio. Devices with different configurations for the same measurement were fabricated in the same batch by switching shadow masks. Current density-voltageluminescence (J–V-L) characteristics were observed using a programcontrolled system incorporated by a source meter (Keithley 2400) and a photometer (Photo Research PR-655) in the ambient atmosphere without protective encapsulation. The UV–vis absorption spectrum was procured by a spectrophotometer (HITACHI U-3310) with scan rate of 600 nm/s. UPS analysis was measured using an electron spectroscopy analysis system (ULVAC-PHI 5000 Versaprobe II) equipped with ultra violet sources providing He-I photons (21.22 eV). The samples with deposition in chronological order for UPS measurement were fabricated in the same batch. To avoid possible ambient oxygen-mediated oxidation and/or ambient-induced chemical changes, samples for UPS were stored in a vacuum sealed box after organic layer deposition, then transferred into the UPS chamber. In order to observe the high binding energy cutoff, a -5 V bias was applied during UPS measurements. Since the depth resolution of UPS is at the range of 0.5∼5 nm, the last deposited layer prepared for the UPS in each sample was controlled to be 5 nm to avoid the interruption of background signals from underlying layers. The impedance spectroscopy analysis was investigated by Agilent E4990 A precision impedance analyzer at frequency range of 20 Hz to 1.2 MHz with AC oscillation level of 100 mV. And the applied DC bias was varied between -2.5 to +5 V. To ensure reliability and reproducibility of our data, all the experiments were repeated several times. 3. Results and discussion Owing to closely related energy values of highest occupied 122

Synthetic Metals 243 (2018) 121–126

C.-T. Tsai et al.

Fig. 2. The schematic diagrams of various HIL structures in OLEDs.

same HILs referring respectively to devices A to D but in a different device structure of ITO/HIL/TAPC (30 nm)/SimCP2:FIrpic (25 nm, 10 wt%)/TPBi (50 nm)/LiF (1 nm)/Al (110 nm). As shown, the phosphorescent and devices with identical HIL structure demonstrated almost the same tendency as their fluorescent counterparts. Again, the stepwise phosphorescent device (device P-A) exhibited exceptional performances on J–V-L characteristics and efficiencies. This indicates that the stepwise HIBL can be used for a wide range of devices with different device structure. Since changing carrier injecting layer in OLED might come along with location shift of the recombination zone, it is especially important to inspect where the light emission is coming from for each device. Fig. 3(a) shows the normalized electroluminescence (EL) spectra of devices A–D operated at 1000 cd/m2. Device D displayed a slightly redshifted peak located at around 540 nm with a broadened full width at half maximum (FWHM) expanded toward long wavelength regime. The wider red-shifted EL spectrum of device D is expected to be achieved because of the to the lack of additional light absorption from NPB:MoO3 or pure MoO3 at 400∼600 nm and > 400 nm [36], respectively. Besides, device A–C showed almost identical spectrum with an emission peak at 530 nm, indicating negligible shift of the emission zones in these devices. This may be due to a wide recombination zone throughout whole Alq3 layer. Especially for devices who have decent hole injection, the emitting zone is expected to locate in Alq3 layer rather than at NPB/Alq3 interface. Unfortunately, with the EL data shown above, we cannot determine if the emission zones had shifted in these devices. For further clarification, normalized EL spectra of devices P-A to P-D were analyzed in Fig. 3(b). From these EL spectra, two obvious emission peaks located at 475 nm and 502 nm from phosphorescent dopant FIrpic were observed. Noteworthy, a remarkable EL intensity decrease in the 502 nm peak whereas a comparative increase in the 475 nm peak were observed from device P-D to device P-A, respectively. As a result, the EL spectra showed a progressively blueshifted tendency throughout device P-D to device P-A. Owing to wideangle optical interference effect by reflective metal cathode, such an EL spectrum shift could be caused by a recombination zone shift [37]. It has also been reported that by moving the location of emission zone away from the anode, one can continuously tune the EL spectrum of FIrpic-based phosphorescent OLEDs from greenish blue to blue [38]. Thus, as shown in Fig. 3(b), the blue-shift of EL peaks from device P-D to P-A suggested a hole injection/transport enhancement which moves

Fig. 1. J–V, (b) L–V and (c) η -J characteristics for ITO/HIL/ NPB(40 nm)/ Alq3(40 nm)/LiF(1 nm)/Al(140 nm) fluorescent green OLEDs with various HILs.

due to excessive hole injection [34,35]. Furthermore, an OLED with a reversed-stepwise HILB structure (that is, slightly-doped sublayer was adjacent to anode, while highly-doped sublayer was adjacent to HTL) was also tested (Supplementary Fig. S3). But device resistance was significantly increased, indicating that hole injection in this device was considerably crippled by the change of energetic arrangement within the HIBL. The prospective mechanisms and effects behind these HIBLs will be further discussed in the following sections. To further acknowledge the enhancements of stepwise HIBL, we also fabricated a batch of phosphorescent OLEDs. Supplementary Fig. S4 shows the J–V-L and η -J characteristics of devices P-A to P-D. Devices P-A to P-D have 123

Synthetic Metals 243 (2018) 121–126

C.-T. Tsai et al.

current density at a given voltage was higher than that of device without any HIL by more than two orders of magnitude. According to Fig. 4, J–V characteristics of devices with HILs can be categorized into three regimes of (I) an ohmic conduction ( J ∝ V ) at low-voltage region below 0.7 V, (II) a trap-limited conduction ( J ∝ V n, n > 2 ) at applied voltage of 1.8 to 3.4 V, and (III) a space-charge limited conduction ( J ∝ V n, n ∼ 2 ) at higher voltages over 3.4 V. In regime (I), all devices showed ohmic conduction behavior. Further increasing the bias leaded to a considerable surge of current density after threshold voltage. From Fig. 4, the threshold voltage for stepwise device was 1.3 V, which is lower than that of uniformly doped device (1.4 V), pure MoO3 device (1.6 V) and normal device (2 V). This suggests that the stepwise HIBL exhibits improved hole injection capability. Besides, in the terminating of part trap-limited region, slope of the J–V plot for stepwise device ( J ∝ V 7.9 ) was much steeper than other devices with HILs ( J ∝ V 4.3 for uniformly doped NPB:MoO3 and J ∝ V 4.3 for pure MoO3, respectively), suggesting enhanced hole transport via stepwise HIBL. It is notable that in both MoO3-doped NPB cases, anomalous J–V characteristics were observed at low-voltage region ranging from 0.5 to 3 V, where a negative differential resistance (NDR), was observed. NDR has reported to be related to guest hopping sites (GHSs) and phonon scattering effects in certain doped organic systems [39,40]. In addition, it has been revealed that the dopant MoOx can be served as hole hopping sites in organic doped system [41]. This strongly implies that the NDR effect observed here was credited to doping of MoO3. However, it was not until the bias V > 3 V that the current increased dramatically in these NPB:MoO3 cases, indicating that the doping induced GHSs as a significant source of OLED performance enhancement is unlikely. To clarify the origin of hole injection and transport improvements, energy levels and interface effects of the stepwise HIBL were investigated by UPS measurements. Fig. 5(a) shows the UPS spectra of the stepwise structure sequentially deposited in the configuration of NPB:MoO3 (35%)/NPB:MoO3 (20%)/NPB on the ITO substrates. The spectrum of ITO/NPB is also shown. Fig. 5(b) and (c) depicts the energy level diagrams of ITO/NPB interface and the stepwise structure. The values of work function, interfacial vacuum level (VL) shift and HOMO level for the deposited layers were obtained from the measured UPS spectra presented in Fig. 5(a). By knowing the band gap of NPB (3.1 eV) [42], the lowest unoccupied molecular orbital (LUMO) values were computed. The VL shift contributed from band-bending on the ITO/ NPB:MoO3 (35%) interface was determined by ITO/NPB and ITO/ NPB:MoO3 (35%) spectra. As shown, hole injection barrier was decreased from 1.022 eV to 0.4467 eV on the ITO/organic interface attributed to the high doping rate of MoO3 (35%). Further deposition of 20% NPB:MoO3 led to HOMO edge 0.6712 eV below the Fermi level. As a result, the so-called stepwise bilayer structure smoothly distributed the injection barrier between ITO and HTL, thus achieved excellent hole injection properties. These results specify that the utilization of stepwise HIBL provides not only superior hole injection on the anode/organic interface but also remarkable carrier transport characteristics. Since it has been declared previously that the significant current (injected carriers) increase under high bias (V > 3 V ) is owing to improved charge transport through HIBL, we conclude this stepwise HOMO arrangement is responsible for the fulfillment of enhanced OLED performance. Furthermore, this also explains the substandard OLED performance of device with reversed-stepwise HILB (Supplementary Fig. S3). Opposite to device A, holes injected in reversed-stepwise device should undergo a higher injection barrier at the ITO/organic interface (∼0.67 eV), and after injected into the second sublayer, the holes have to overcome a large barrier [∼0.5 eV (0.9417 eV-0.4467 eV)]. This may lead to carrier accumulation on the interfaces due to high barrier, which results in poor performance of OLEDs. Finally, capacitance-voltage (C–V) measurement based on impedance spectroscopy was also conducted to observe the dynamics of the mobile charge in devices. In this type of measurements, each organic layer in device can be represented by a parallel RC unit because of

Fig. 3. Normalized electroluminescence (EL) spectra of (a) devices A to D and (b) devices P-A to P-D at 1000 cd/m2.

the recombination zone toward the cathode side. To elucidate the effects of the stepwise HIBL on hole injection, hole only devices in a configuration of ITO/HIL/NPB (30 nm)/MoO3 (1 nm)/ Al (140 nm) were prepared and tested. A variety of hole-injection layers corresponding to devices A to D were inserted into the hole-only devices. Fig. 4 displays the double logarithmic J–V characteristics of the hole-only devices. As shown, significant improvement of J–V characteristics was observed on device with stepwise HIBL compared to other devices in the full range of applied voltage, which contributes to outstanding OLED performance enhancements. With stepwise HIBL, the

Fig. 4. Double-logarithmic J–V characteristics of hole-only devices with different HILs. 124

Synthetic Metals 243 (2018) 121–126

C.-T. Tsai et al.

Fig. 5. (a) UPS spectra of the sequentially deposited stepwise structure. The inserted numbers are the VL shifts (ITO as reference), (b) energy level diagrams of ITO/ NPB and, (c) energy level diagram of ITO/NPB:MoO3(35%)/NPB:MoO3(20%)/NPB. Φh and Φe represent the hole and electron activation energy (injection barriers at ITO/organic interface), Δ represents the VL shift caused by interface dipoles, Eion represents the ionized energy and Eg the band gap of NPB.

materials allow us to distinguish discrete signals contributed from each specific material [45,46]. In this well-established NPB/Alq3 model, the capacitance in low frequency region directly reflects the capacitance of Alq3 layer [47]. Moreover, Alq3 layer also acts as emission layer in our devices, that is, we can perceive the carrier behaviors such as injection, accumulation and recombination by monitoring the capacitance of Alq3 [48]. Fig. 6 shows the capacitance-voltage plots of devices A to D obtained at 20 Hz. When the devices operated under negative bias, it showed a constant capacitance, corresponding to the geometric capacitance of total organic layers in device:

Cgeo =

ε0 εr A d

where A is the active area of devices, d is entire thickness of organic layers, ε0 is the permittivity of free space, and εr is the relative dielectric constant. For example, for device D, with εr ≈ 3, A = 6.25 × 10−6 m2 , and d = 80 nm , we can obtain Cgeo ≈ 2 nF . With the addition of HILs, the capacitance reduced lightly due to higher organic layer thickness, and varied with different doping concentration. When the bias voltage was set to transition voltage Vt , holes started to inject from ITO, which resulted in a raise of capacitance. It can be seen in Fig. 6 that device A exhibits Vt of -1.1 V, while devices B–D show -0.9, -0.9, and -0.1 V,

Fig. 6. Capacitance versus voltage characteristics of devices A to D by impedance spectroscopy.

the inherent conductivity and dielectric properties of organic materials [43,44]. Diverse AC respond characteristics of different organic 125

Synthetic Metals 243 (2018) 121–126

C.-T. Tsai et al.

respectively. This is indicative of exceptional hole injection ability of stepwise HIBL. Further raising the voltage led to the accumulation of charges at NPB/Alq3 interface, hence the capacitance reached to a saturation value gradually. At bias V = Vbi (built-in voltage), Alq3 layer reached its flat-band potential, the injected holes and electrons started to recombine and the device began to emit light [48]. At V > Vbi , the capacitance of emitter Alq3 decreased drastically while we continually increased the bias owing to annihilation of charges by recombination. All devices showed an equivalent built-in voltage near 2.7 V, which is relevant to the inherent built-in potential of NPB/Alq3 heterojunction. At voltage beyond Vbi , the dropping speed of capacitance appeared to be the slowest in device A. We contribute this to excellent hole injection and transport characteristics which persistently injects abundant holes into device. It is worth noting that the capacitance measured at V > Vbi shows exactly identical tendency to the luminance properties shown in Fig. 1(b), which verifies the reliability of our C–V measurement.

[7] C. Wu, C. Wu, J. Sturm, A. Kahn, Appl. Phys. Lett. 70 (1997) 1348. [8] J.A. Bardecker, H. Ma, T. Kim, F. Huang, M.S. Liu, Y.J. Cheng, G. Ting, A.K.Y. Jen, Adv. Funct. Mater. 18 (2008) 3964. [9] F. Wang, X. Qiao, T. Xiong, D. Ma, Org. Electron. 9 (2008) 985. [10] D.-S. Leem, H.-D. Park, J.-W. Kang, J.-H. Lee, J.W. Kim, J.-J. Kim, Appl. Phys. Lett. 91 (2007) 011113. [11] T. Matsushima, Y. Kinoshita, H. Murata, Appl. Phys. Lett. 91 (2007) 253504. [12] J. Meyer, K. Zilberberg, T. Riedl, A. Kahn, J. Appl. Phys. 110 (2011) 033710. [13] C.-H. Gao, X.-Z. Zhu, L. Zhang, D.-Y. Zhou, Z.-K. Wang, L.-S. Liao, Appl. Phys. Lett. 102 (2013) 61. [14] K. Zilberberg, S. Trost, J. Meyer, A. Kahn, A. Behrendt, D. Lützenkirchen‐Hecht, R. Frahm, T. Riedl, Adv. Funct. Mater. 21 (2011) 4776. [15] H. Riel, S. Barth, T.A. Beierlein, W. Bruetting, S. Karg, P. Mueller, W. Riess, Grading Interfaces: A New Concept to Improve Device Performance in Organic Multilayer Light-Emitting Diodes, International Society for Optics and Photonics, 2001 p. 167. [16] M. Baldo, C. Adachi, S. Forrest, Phys. Rev. B 62 (2000) 10967. [17] J.Y. Lee, Appl. Phys. Lett. 89 (2006) 153503. [18] J. Lee, H.-F. Chen, T. Batagoda, C. Coburn, P.I. Djurovich, M.E. Thompson, S.R. Forrest, Nat. Mater. 15 (2016) 92. [19] U.S. Bhansali, H. Jia, I.W. Oswald, M.A. Omary, B.E. Gnade, Appl. Phys. Lett. 100 (2012) 101. [20] H.-H. Chang, C.-C. Wu, C.-C. Yang, C.-W. Chen, C.-C. Lee, Appl. Phys. Lett. 78 (2001) 574. [21] Y. Zhang, J. Lee, S.R. Forrest, Nat. Commun. 5 (2014) ncomms6008. [22] S.H. Kim, J. Jang, J.Y. Lee, Appl. Phys. Lett. 90 (2007) 203511. [23] S.-H. Leeb, J.K. Kima, C.H. Leec, Role of Carrier Mobility, Exciton Diffusion, and Their Interplay for Charge Balance and Improved Properties of Organic Electrophosphorescent Device, (2006) p. 633315. [24] W. Gao, K. Yang, H. Liu, J. Feng, J. Hou, S. Liu, Graded Doping in Active Layer for Achievement of High Brightness and Efficiency Organic Light-Emitting Devices, (2001) p. 386. [25] D. Ma, J.M. Lupton, R. Beavington, P.L. Burn, I.D. Samuel, Adv. Funct. Mater. 12 (2002) 507. [26] H. Yang, Y. Yoon, T. Kim, K. Kwack, J. Kim, J. Seo, Y. Kim, Solid State Commun. 137 (2006) 87. [27] S. Ciná, D. Vaufrey, C. Fery, B. Racine, H. Doyeux, A. Bettinelli, J.C. Martinez, P‐135: Efficient Electron Injection from PEDOT‐PSS into a Graded‐n‐doped Electron Transporting Layer in an Inverted OLED Structure, Wiley Online Library, 2005 p. 819. [28] J. Hou, J. Wu, Z. Xie, L. Wang, Appl. Phys. Lett. 95 (2009) 203508. [29] J.-H. Lee, D.-S. Leem, H.-J. Kim, J.-J. Kim, Appl. Phys. Lett. 94 (2009) 93. [30] D. Zhang, M. Cai, Y. Zhang, D. Zhang, L. Duan, ACS Appl. Mater. Interfaces 7 (2015) 28693. [31] W.-J. Shin, J.-Y. Lee, J.C. Kim, T.-H. Yoon, T.-S. Kim, O.-K. Song, Org. Electron. 9 (2008) 333. [32] S. Hamwi, J. Meyer, T. Winkler, T. Riedl, W. Kowalsky, Appl. Phys. Lett. 94 (2009) 174. [33] J.-P. Yang, Q.-Y. Bao, Y. Xiao, Y.-H. Deng, Y.-Q. Li, S.-T. Lee, J.-X. Tang, Org. Electron. 13 (2012) 2243. [34] N. Giebink, S. Forrest, Phys. Rev. B 77 (2008) 235215. [35] S. Höfle, A. Schienle, C. Bernhard, M. Bruns, U. Lemmer, A. Colsmann, Adv. Mater. 26 (2014) 5155. [36] J. Yun, J. Yang, Y. Hong, C. Lee, W.J. Song, Y.J. Sung, J. Korean Phys. Soc. 53 (2008). [37] T.-L. Chiu, P.-Y. Lee, J.-H. Lee, C.-H. Hsiao, M.-K. Leung, C.-C. Lee, C.-Y. Chen, C.C. Yang, J. Appl. Phys. 109 (2011) 084520. [38] Z. Wu, L. Wang, G. Lei, Y. Qiu, J. Appl. Phys. 97 (2005) 103105. [39] Y.-K. Fang, Y.-T. Chiang, S.-F. Chen, C.-Y. Lin, S.-C. Hou, C.-S. Hung, T.-Y. Tsai, S.H. Chang, T.-H. Chou, J. Phys. Chem. Solids 69 (2008) 738. [40] M.M. Sarjidan, S. Basri, N. Za’Aba, M. Zaini, W.A. Majid, Bull. Mater. Sci. 38 (2015) 235. [41] X. Qiao, J. Chen, X. Li, D. Ma, J. Appl. Phys. 107 (2010) 104505. [42] P. Jeon, H. Lee, J. Lee, K. Jeong, J. Lee, Y. Yi, Appl. Phys. Lett. 99 (2011) 169. [43] M.-T. Hsieh, C.-C. Chang, J.-F. Chen, C.H. Chen, Appl. Phys. Lett. 89 (2006) 103510. [44] M.-T. Hsieh, M.-H. Ho, K.-H. Lin, J.-F. Chen, T.-M. Chen, C.H. Chen, Appl. Phys. Lett. 95 (2009) 033501. [45] M.-T. Hsieh, M.-H. Ho, K.-H. Lin, J.-F. Chen, T.-M. Chen, C.H. Chen, Appl. Phys. Lett. 96 (2010) 66. [46] J. Zhao, Y. Cai, J.-P. Yang, H.-X. Wei, Y.-H. Deng, Y.-Q. Li, S.-T. Lee, J.-X. Tang, Appl. Phys. Lett. 101 (2012) 193303. [47] S. Nowy, W. Ren, J. Wagner, J.A. Weber, W. Brütting, Impedance Spectroscopy of Organic Hetero-Layer OLEDs as a Probe for Charge Carrier Injection and Device Degradation, International Society for Optics and Photonics, 2009 p. 74150G. [48] S. Nowy, W. Ren, A. Elschner, W. Lövenich, W. Brütting, J. Appl. Phys. 107 (2010) 054501.

4. Conclusion In summary, we demonstrated a novel stepwise hole injecting bilayer structure by utilizing MoO3-doped NPB. Different types of MoO3based HIL were used for comparison of hole injection and transport effects. The J–V-L characteristics of OLEDs indicated that the stepwise HIBL is superior by showing improved device performances on both fluorescent and phosphorescent devices. EL spectra revealed a shifted carrier recombination zone which is attributed to enhanced hole injection. J–V investigation of hole-only devices unveiled an increased current density higher than that of the normal device by more than two orders of magnitude, which is contributed by better hole injection and transport via HIBL. The charge dynamics was also monitored by impedance spectroscopy-based capacitance-voltage measurement, and again presented excellent hole injection ability of our HIBL. Furthermore, UPS studies showed energy level diagrams that not only lowered hole-injection barrier at anode/organic interface from 1.022 eV to 0.4467 eV, but also exhibited a cascading energy level which efficiently distributed the energy barrier from ITO to NPB, therefore enhanced both hole injection and transport efficiency and led to better carrier balance and luminance efficiencies in OLEDs. Acknowledgment This work was financially supported by Industrial Technology Research Institute of Taiwan under Contract No. 107-EC-17-A-24-1303. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.synthmet.2018.06. 008. References [1] D.P.-K. Tsang, T. Matsushima, C. Adachi, Sci. Rep. 6 (2016) 22463. [2] S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Lüssem, K. Leo, Nature 459 (2009) 234. [3] P.J. Jesuraj, R. Parameshwari, K. Kanthasamy, J. Koch, H. Pfnür, K. Jeganathan, Appl. Surf. Sci. 397 (2017) 144. [4] Y. Li, D.-Q. Zhang, L. Duan, R. Zhang, L.-D. Wang, Y. Qiu, Appl. Phys. Lett. 90 (2007) 012119. [5] C.E. Small, S.W. Tsang, J. Kido, S.K. So, F. So, Adv. Funct. Mater. 22 (2012) 3261. [6] K. Sugiyama, H. Ishii, Y. Ouchi, K. Seki, J. Appl. Phys. 87 (2000) 295.

126