Color stable and highly efficient hybrid white organic light-emitting devices using heavily doped thermally activated delayed fluorescence and ultrathin non-doped phosphorescence layers

Organic Electronics 43 (2017) 112e120

Contents lists available at ScienceDirect

Organic Electronics journal homepage: www.elsevier.com/locate/orgel

Color stable and highly efficient hybrid white organic light-emitting devices using heavily doped thermally activated delayed fluorescence and ultrathin non-doped phosphorescence layers Yige Qi, Zijun Wang, Sihui Hou, Junsheng Yu* State Key Laboratory of Electronic Thin Films and Integrated Devices, School of Optoelectronic Information, University of Electronic Science and Technology of China (UESTC), Chengdu 610054, PR China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 9 December 2016 Received in revised form 5 January 2017 Accepted 8 January 2017 Available online 11 January 2017

Blue/orange complementary fluorescence/phosphorescence hybrid white organic light-emitting devices with excellent color stability and high efficiency have been fabricated, which are based on an easily fabricated multiple emissive layer (EML) configuration with an ultrathin non-doped orange phosphorescence EML selectively inserted between heavily doped blue thermally activated delayed fluorescence (TADF) EMLs. Through systematic investigation and improvement on luminance-dependent color shift and efficiency deterioration, a slight Commission Internationale de 10 Eclairage coordinates shift of (0.008, 0.003) at a practical luminance range from 1000 to 10000 cd/m2, a maximum power efficiency of 45.8 lm/W, a maximum external quantum efficiency (EQE) of 15.7% and an EQE above 12% at 1000 cd/m2 have been achieved. The heavily doped blue TADF emitters which act as the main charge transport channels and recombination sites in the host with high-lying lowest triplet excited state, take advantage of the bipolar transport ability to broaden the major charge recombination region, which alleviates triplet energy loss. The selectively inserted ultrathin non-doped orange EML makes its emission €rster energy transfer, which is effective to keep color stable under different mechanism dominated by Fo drive voltages. © 2017 Elsevier B.V. All rights reserved.

Keywords: Organic light-emitting device Thermally activated delayed fluorescence emitter Phosphorescence emitter Color stability High efficiency

1. Introduction White organic light-emitting devices (WOLEDs) have been paid tremendous research attentions on their applications in full-color flat-panel displays and solid state lighting, due to their advantages such as light weight, high resolution, homogeneous largearea emission and potential application on flexible substrates [1,2]. In the past two decades, WOLEDs have been significantly developed by various approaches, including emitting material system innovation and device configuration optimization, etc [3,4]. Nowadays, three kinds of WOLEDs based on different emitting materials are reported, which are full-fluorescence WOLEDs, fullphosphorescence WOLEDs and fluorescence/phosphorescence hybrid WOLEDs. Since phosphors theoretically enable an internal quantum efficiency (IQE) of 100% by harvesting both singlet and triplet excitons [5,6]. The full-phosphorescence WOLEDs can obtain

* Corresponding author. E-mail address: [email protected] (J. Yu). http://dx.doi.org/10.1016/j.orgel.2017.01.012 1566-1199/© 2017 Elsevier B.V. All rights reserved.

efficiency fourfold higher than that of full-fluorescence ones [7,8]. However, the development of full-phosphorescence WOLEDs is limited by the lack of blue phosphors with both high efficiency and good stability, which results in the poor lifetime of WOLEDs [9,10]. Soon after, hybrid WOLEDs combining blue fluorophores with green and red phosphors have attracted considerable interest, which take advantage of the superior stability of fluorophores and high efficiency of phosphors simultaneously [11e13]. To achieve high-efficiency hybrid WOLEDs, triplet energy loss through the non-radiative triplet states of the blue fluorophores should be prevented. Although the employment of blue fluorophores with high-lying lowest triplet excited state (T1) facilitates the energy transfer from blue fluorophores to phosphors, resulting in an alleviated triplet energy loss [14,15], there are still some triplet excitons inevitably deactivated on blue fluorophores through non-radiative recombination during diffusion process, due to the short radius of Dexter energy transfer [16]. Recently, the blue thermally activated delayed fluorescence (TADF) emitters which were introduced by Adachi et al., are considered as promising alternatives to replace conventional blue fluorophores in hybrid

Y. Qi et al. / Organic Electronics 43 (2017) 112e120

WOLEDs [16e19], since blue TADF emitters with high T1 and high efficiency have been reported, due to their intrinsic small singlettriplet splits and the potential of realizing unity IQE [20,21]. More important, the triplet energy loss on blue TADF emitters can be significantly decreased, since triplet excitons can convert to singlet ones through reverse inter-system crossing (RISC) process, and the latter can be used for blue fluorescence by radiative decay or €rster energy transfer. transferred to phosphors via long-radius Fo Besides, highly efficient and color stable hybrid WOLEDs are usually based on three typical device configurations, including multiple emissive layers (MEML), single emissive layer with multiple dopants (SEML-MD) and single emissive layer with single dopant (SEML-SD). However, the high-performance WOLEDs were obtained through complex device structures and/or precisely controlled fabrication process, which raise the fabrication cost and decrease reproducibility. For instance, to achieve high color stability, the MEML based WOLEDs need an additional charge or exciton-blocking interlayer [22,23], while the SEML-MD based WOLEDs require precisely controlled doping process for optimum doping concentration ratio between different dopants [19]. Furthermore, to alleviate the deteriorated effect of charge trapping €rster energy on color stability and realize incomplete host-guest Fo transfer, the state-of-the-art blue/yellow or blue/orange complementary WOLEDs based on SEML-SD structure also need precise doping control to maintain an extremely low (0.5 wt%) doping concentration [12,24]. Hence, high-performance WOLEDs with simplified configuration and/or easy fabrication should be further developed. Recently, a blue TADF emitter bis[4-(9,9-dimethyl-9,10dihydroacridine)phenyl]sulfone (DMAC-DPS) is highly attractive due to its high photoluminescence quantum yield (PLQY) of 0.88 in neat film and bipolar transport ability [20,25]. Since methyl groups on the DMAC units inhibit the intermolecular p-p stacking interactions between electron donors, the high PLQY of DMAC-DPS in solid state is attributed to the aggregation-alleviated ability, which enables the highly efficient OLEDs with easily fabricated heavily doped EML. Besides, DMAC-DPS exhibits a rather broad characteristic emission spectrum with a full width at half maximum about 80 nm, which is suitable as blue emission component in blue/orange complementary WOLEDs. Therefore, in this work, we propose blue/orange complementary hybrid WOLEDs with an easily fabricated MEML structure consisting of heavily doped fluorescence EMLs using DMAC-DPS and an ultrathin non-doped phosphorescence EML employing an orange iridium complex bis(4-tert-butyl-2-phenylbenzothiozolato0 N,C2 )iridium(III) (acetylacetonate) [(tbt)2Ir(acac)]. The MEML based WOLEDs without interlayer exhibit excellent color stability and high efficiency with a slight Commission Internationale de 10 Eclairage (CIE) coordinates shift of (0.008, 0.003) from 1000 to 10000 cd/m2 and a maximum external quantum efficiency (EQEmax) of 15.7%, respectively. Meanwhile, the charge transport process and emission mechanism have been systematically investigated by analyzing the electroluminescence (EL) spectra, characteristics and photoluminescence (PL) transient decay curves. 2. Experimental Indium tin oxide (ITO) coated glass substrates with a sheet resistance of 15 U/sq which acted as anode, were pre-cleaned with detergent, acetone, deionized water and ethanol for 15 min at each ultrasonic step. Then oxygen plasma treatment was performed to further clean the ITO surface. Organic functional layers and metallic cathode were thermally evaporated in vacuum under a pressure of 3  104 Pa and 3  103 Pa, respectively. The device configurations in this research are listed in Table 1, and all devices have an active

113

Table 1 Device configurations of blue OLEDs and WOLEDs. Device

MoO3

B1 B2 B3

10

Device

MoO3

W1 W2 W3 W4 Device

TAPC

mCP

mCP: DMAC-DPS

Bphen

40

10

15 (10 wt%) 15 (30 wt%) 15 (50 wt%)

40

TAPC

mCP

mCP: B

Bphen

1 0.3 0.1 0.05

10

40

10

15

MoO3

TAPC

mCP

mCP: B

Ora

mCP: B

Bphen

0.1

5 10 15

40

Ora

DPEPO: B

Bphen

0.1

0 3 6

40

P1 P2 P3

10

40

10

10 5 0

Device

MoO3

TAPC

mCP

DPEPO: B

10

15 12 9

W5 W6 W7

Ora

10

40

40

The above are organic functional layers except ITO anode and 100 nm-thick Mg: Ag cathode, mCP: B ¼ mCP: DMAC-DPS(50 wt%); Ora¼(tbt)2Ir(acac); DPEPO: B ¼ DPEPO: DMAC-DPS(50 wt%). The unit for layer thickness is nm.

area of 0.25 cm2. Therein, molybdenum trioxide (MoO3) was used as hole injection layer, while 4,40 -cyclohexylidenebis[N,N-bis(4methylphenyl)aniline] (TAPC) and 4,7-diphenyl-1,10phenanthroline (Bphen) were employed as hole transport layer and electron transport layer, respectively. Besides, 3,50 -N,N0 -dicarbazole-benzene (mCP) was inserted between TAPC and blue EML to inhibit exciplex formation at the interface of TAPC/DMAC-DPS. The blend films of DMAC-DPS fluorophore doped in mCP or bis[2(diphenylphosphino)phenyl]ether oxide (DPEPO) host matrix acted as blue EMLs. The ultrathin non-doped film of (tbt)2Ir(acac) phosphor was an orange EML. The energy level diagram of WOLEDs along with the chemical structures of the organic materials in EMLs are depicted in Fig. 1. UVevis absorption spectrum was characterized with a SHIMATZU UV-1700 spectrophotometer. PL spectra were recorded with a PerkinElmer LS55 spectrometer. EL spectra and CIE coordinates were measured with an OPT-2000 spectrometer. PL transient decay characteristics were recorded with a HORIBA Scientific Single Photon Counting Controller FluoroHub-B, in which the samples were exited at 370 nm using a NanoLED-370 excitation light source, and the emitted photons were detected by a TBX detector connected to a TBX-PS power supply. Current densityvoltage-luminance (J-V-L) characteristics were tested with a Keithley 4200 source and a luminance meter. All the measurements were performed in air at room temperature without encapsulation except that the PL transient decay characteristics were recorded under nitrogen atmosphere. 3. Results and discussion According to the UVevis absorption spectrum of the (tbt)2Ir(acac) neat film in Fig. 1(b), the absorption peaks from 300 nm to 350 nm, are mostly assigned to ligand-centered (LC) p-p* transitions, while the absorption peaks located at 450 nm and 490 nm are originated from singlet and triplet metal to ligand charge transfer (1MLCT and 3MLCT) transitions, respectively. The PL spectrum of the DMAC-DPS neat film which is excited at 370 nm using an excitation light source, shows a peak at 472 nm originated from singlet intramolecular charge transfer transition. There is a large spectral overlap between the PL spectrum of DMAC-DPS and the 1MLCT €rster absorption band of (tbt)2Ir(acac). It indicates that efficient Fo

114

Y. Qi et al. / Organic Electronics 43 (2017) 112e120

Fig. 1. (a) Chemical structures of materials in EMLs. (b) UVevis absorption and PL spectra of neat films of emitters. Schematic energy level diagram of the WOLEDs based on (c) mCP and (d) DPEPO host.

energy transfer from DMAC-DPS to (tbt)2Ir(acac) can occur. To pave a way for high-efficiency WOLEDs, the performance of blue OLEDs is firstly improved by doping DMAC-DPS in mCP and optimizing the doping concentration. The J-V-L characteristics of devices B1-B3 with increased doping concentrations from 10 to 50 wt% are plotted in Fig. S1(a). The current density increases along with the doping concentration at a constant voltage. It is well known that, the EL mechanisms in the OLEDs based on doped EML follow two primary rules, i.e. host-guest energy transfer and direct exciton formation on the dopant. In a host-guest energy transfer system, charge carriers mainly transport and recombine on host molecules, so the J-V curves do not exhibit obvious sensitivity with the variation of doping concentration [26]. For our blue OLEDs, the dependence of J-V curves on doping concentrations indicates that direct exciton formation on the dopant is dominant [27]. It can be inferred from the comparison of electron injection barriers between the interface of mCP/Bphen (0.5 eV) and DMAC-DPS/Bphen (0 eV) shown in Fig. 1(c). Since there exists a large electron injection barrier from Bphen to mCP, heavily doped DMAC-DPS molecules can act as additional injection and transport channels, thus facilitating the energetically favorable electron injection into and electron hopping on dopants, which can be further revealed by the J-V curves of the electron-only devices based on mCP or mCP: DMAC-DPS shown in Fig. S1(b). When DMAC-DPS is doped in mCP, the current density is drastically enhanced at a constant voltage. On the contrary, it can be seen in Fig. S1(b) that, the almost overlapped J-V curves of the hole-only devices using mCP and mCP: DMAC-DPS indicate that holes mainly transport on mCP in blue EML. The L-V characteristics also exhibit doping concentration dependent characteristics similar to J-V curves with a decreased turn-on voltage

(Von) from 4.7 to 3.7 V. The highest maximum luminance and lowest turn-on voltage of device B3 are attributed to sufficient emitting centers and facilitative electron injection. According to the power efficiency-current density-luminance efficiency (PE-J-LE) characteristics shown in Fig. S1(c), the blue OLEDs exhibit dramatically enhanced efficiencies with the increase of doping concentration. The highest PEmax of 17.2 lm/W, EQEmax of 10.8% and EQE of 8.7% at 1000 cd/m2 are obtained with a high doping concentration of 50 wt% (Table S1), which are attributed to the aggregation-alleviated capability of DMAC-DPS [25]. Therefore, as shown in Fig. S1(d), the emission maximum of heavily doped device B3 has a red shift of only 8 nm from 479 to 487 nm compared with that of device B1, which verifies the aggregation-alleviated ability of DMAC-DPS. The blue device with an optimized doping concentration of 50 wt% provides a prerequisite for high-performance blue/orange complementary WOLEDs. To simplify the fabrication process, a (tbt)2Ir(acac)-based ultrathin non-doped orange phosphorescence EML is deposited on the DMAC-DPS-based blue EML, and the thickness of the orange EML has been optimized. The J-V-L curves of devices W1-W4 with the orange EML thickness decreasing from 1 to 0.05 nm are plotted in Fig. S2(a). As the orange EML becomes thinner, the increase of current density at a constant voltage is attributed to the decreasing amount of (tbt)2Ir(acac) molecules, which can act as hole trapped centers due to a much shallower highest occupied molecular orbital (HOMO) level compared with that of mCP. The L-V characteristics exhibit a variation tendency similar to J-V curves with a decreased Von from 3.8 to 3.2 V. The reduction of Von also reveals the role of (tbt)2Ir(acac) molecules as hole trapped centers. Meanwhile, the maximum luminance

Y. Qi et al. / Organic Electronics 43 (2017) 112e120

decreases along with the orange EML thickness due to the decreasing amount of emitting centers. The PE-J-LE characteristics and EL spectra of devices W1-W4 are shown in Fig. S2(b) and Fig. 2(aec), respectively. It is noteworthy that, although the orange EML thickness of device W2 is much thinner than that of device W1, the efficiencies of device W2 are even higher than those of device W1. Since ultrathin EML film of below 0.5 nm can generate an island-like morphology, due to the thinner thickness of film than the diameter of organic molecule [28,29], which allows sufficient space for exciton formation and may minimize triplet-triplet annihilation (TTA) [11]. However, devices W1 and W2 have achieved much higher efficiencies than those of devices W3 and W4, but the blue emission component proportion is negligible compared with that of orange emission (characterized by an emission peak at 560 nm and a shoulder at 598 nm) in EL spectra. The reasons that devices W1 and W2 cannot exhibit white emission, are illuminated as follows: since the hole mobility of TAPC is much higher than the electron mobility of Bphen [30,31], and mCP exhibits strong hole transport ability [32], the major charge recombination region should be narrow and near the mCP/Bphen interface, despite that the heavily doped DMACDPS promotes the electron transport in EML. As an ultrathin nondoped EML is inserted at this interface, the (tbt)2Ir(acac) molecules as hole trapped centers and with exciton energy lower than

115

that of DMAC-DPS become preferred charge recombination sites [33]. Moreover, the (tbt)2Ir(acac) layer with a thickness of 1 or 0.3 nm in device W1 or W2 provides sufficient sites for charge recombination, so that the increasing amount of injected charges along with the increase of drive current cannot exceed the saturation level. That is why the blue emission component ratio is always negligible with the increase of luminance in EL spectra. Fortunately, devices W3 and W4 show warm white EL spectra, and have achieved PEmax above 46 lm/W, EQEmax above 17% and EQE nearly 15% at 1000 cd/m2 (Table S1). However, the large saturationinduced blue shifts in the EL spectra of devices W3 and W4 at a practical luminance range from 1000 to 10000 cd/m2 have been observed, which are corresponding to apparent CIE coordinates variations of (0.025, 0.021) and (0.035, 0.037), respectively. As the charge recombination sites provided by 0.1 or 0.05 nm-thick (tbt)2Ir(acac) layer become saturated with an increase in the amount of injected charges, the amount of excitons formed on and released from DMAC-DPS molecules increases, leading to remarkably enhanced blue emission composition proportion. According to the EL characteristics and spectra of devices W1-W4, the ultrathin non-doped orange EML with a thickness of 0.1 nm is adopted. However, the hole trapping by (tbt)2Ir(acac) molecules should be alleviated to improve EL spectra stability, which can be realized by moving the ultrathin (tbt)2Ir(acac) layer out of the main charge

Fig. 2. EL spectra of OLEDs versus luminance.

116

Y. Qi et al. / Organic Electronics 43 (2017) 112e120

recombination region. To ascertain the location and width of the major charge recombination region and improve the color stability of our WOLEDs, the P-series devices P1-P3 using a 0.1 nm-thick nondoped (tbt)2Ir(acac) layer as a sensitive probe inserted gradually away from the mCP/Bphen interface have been designed. The EL characteristics and spectra of devices P1-P3 are described in Figs. S3(aec) and Fig. 2(def), respectively. Since the blue emission component percentage dramatically enhances in the EL spectra of devices P1 compared with that of device W3, the orange emission composition is not much stronger than that of blue emission any more. This phenomenon suggests that the ultrathin (tbt)2Ir(acac) layer in device P1 is moved out of the major charge recombination region, therefore, the recombination region is mainly located within a 5 nm-width zone from the mCP/Bphen interface. Consequently, DMAC-DPS molecules become the major charge recombination sites, and a majority of excitons on (tbt)2Ir(acac) molecules come from energy transfer from DMAC-DPS rather than direct exciton formation, which can be reflected by the slightly enhanced blue emission composition ratio with the increase of luminance. Since the energy transfer is nearly independent to the amount of injected charges [33,34], the selective insertion of the ultrathin orange EML out of the main charge recombination zone can substantially improve color stability under different driving conditions. As summarized in Table S1, device P1 shows a small CIE coordinates shift of (0.019, 0.004) at a practical luminance range from 1000 to 10000 cd/m2. The slight EL spectra variation is also intuitively shown in CIE1931 chromaticity coordinates diagram (Fig. S3(d)). When the ultrathin (tbt)2Ir(acac) layer is inserted away from the mCP/Bphen interface with a distance of 15 nm, as the excitons on (tbt)2Ir(acac) molecules are totally generated via energy transfer, the EL spectra of device P3 exhibit extremely high color stability with a slight CIE coordinates shift of (0.001, 0.011) from 1000 to 10000 cd/m2. As shown in Fig. 2(def), except for the slightly enhanced blue emission composition ratio, the emission intensity of shoulder peak of (tbt)2Ir(acac) at about 600 nm for devices P1-P3 mildly increases along with the luminance. The increasing amount of excitons formed on and released from DMAC-DPS and (tbt)2Ir(acac) molecules leads to the increase of luminance. Meanwhile, the increase in the amount of triplet excitons results in the increased triplet exciton density and enhanced interaction in (tbt)2Ir(acac) molecules to bring increased emission intensity of shoulder peak. To further evaluate the color stability of our MEML based WOLEDs, a numerical parameter, color variation index (Icv) is adopted according to the previous work of our group [35], which can be expressed as follows:

Z Icv ¼

540

400

Z 760 jN1 ðlÞ  N0 ðlÞjdl þ jN1 ðlÞ  N0 ðlÞjdl 541 Z 760 Z 760 N1 ðlÞdl þ N0 ðlÞdl 400

spectra and a decreased EQE at a constant luminance in EQE-L curves, which may be attributed to aggravated triplet exciton quenching on DMAC-DPS molecules, such as TTA or triplet-polaron annihilation (TPA). The EL mechanism of DMAC-DPS and (tbt)2Ir(acac) in device P1 can be elucidated in Fig. 3. From the schematic diagram, we can see that, both direct exciton formation on (tbt)2Ir(acac) and short-radius Dexter energy transfer from DMACDPS to (tbt)2Ir(acac) are limited, due to the ultrathin (tbt)2Ir(acac) layer out of the major charge recombination region. Singlet and triplet excitons mainly form on DMAC-DPS. Then, triplet excitons convert to singlet ones via RISC, and some of singlet excitons undergo radiative decay to yield prompt and delayed fluorescence. Meanwhile, the other singlet excitons are transferred to (tbt)2Ir(€rster energy transfer. The rate constant of acac) via long-radius Fo €rster energy transfer is around 1010 s1, which is nearly two orFo ders of magnitude faster than those of the inter-system crossing (ISC) and RISC processes of reported TADF molecules (<108 s1) €rster energy transfer will convert the electrically [36,37]. The fast Fo generated singlet excitons on the lowest singlet excited state of DMAC-DPS to the singlet excitons on the 1MLCT state of (tbt)2Ir(acac). Therefore, the energy transfer process from DMAC-DPS to (tbt)2Ir(acac) provides another radiative pathway, which suppresses ISC process to decrease triplet exciton density in the multiple ISC-RISC circulations on DMAC-DPS, so that the concentration quenching of triplet excitons is alleviated [38]. €rster energy transfer process To confirm whether an efficient Fo from DMAC-DPS to (tbt)2Ir(acac) exists, the PL transient decay characteristics of the DMAC-DPS neat and DMAC-DPS: (tbt)2Ir(acac) blend films have been studied under nitrogen atmosphere. DMAC-DPS is excited at a wavelength of 370 nm. DMAC-DPS and (tbt)2Ir(acac) are observed at 472 and 560 nm, respectively. As shown in Fig. 4(a), estimated by decay simulation, the lifetime of the delayed component of DMAC-DPS is markedly decreased with increasing the doping concentration of (tbt)2Ir(acac). Since the Dexter energy transfer from DMAC-DPS to (tbt)2Ir(acac) is negligible due to a low doping concentration, and the triplet excitons on DMAC-DPS are converted from photo-excited singlet ones via ISC process. The shortened lifetime of delayed component is attributed €rster energy transfer from DMAC-DPS to (tbt)2Ir(acac), to efficient Fo

(1)

400

The wavelength at 540 nm is the separating point for blue and orange light. N0(l) and N1(l) represent the photons number distribution calculated from the EL spectra at luminance of 1000 and 10000 cd/m2, respectively. Hence, Eq. (1) is based on the variation of blue and orange photons number from N0(l) to N1(l) and the sum of photons number. According to Eq. (1), Icv values are calculated and shown in inset of Fig. S3(d). As a result, devices W3, P1 and P3 exhibit gradually decreased Icv values of 4.7%, 4.2% and 2.0%, respectively, which are accordant with the EL spectra variation and CIE coordinates shift. However, the P-series devices with the (tbt)2Ir(acac) layer inserted gradually away from the major charge recombination region show a decreased blue emission component percentage in EL

Fig. 3. Schematic EL mechanism diagram of DMAC-DPS and (tbt)2Ir(acac) in device P1.

Y. Qi et al. / Organic Electronics 43 (2017) 112e120

117

Fig. 4. Experimental and simulated PL transient decay characteristics of the DMAC-DPS neat film and DMAC-DPS: (tbt)2Ir(acac) blend films with doping concentrations of 0.5 wt% and 1 wt% observed at (a) 472 nm and (b) 560 nm under nitrogen atmosphere. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

which suppresses the ISC process of DMAC-DPS and decreases triplet exciton population in the multiple ISC-RISC circulations [37,39]. Besides, Fig. 4(b) presents the PL transient decay characteristics of (tbt)2Ir(acac) in the DMAC-DPS: (tbt)2Ir(acac) blend films. The variation tendency of the delayed component of (tbt)2Ir(acac) matches well with that of DMAC-DPS with the increase of doping concentration, which also suggests a fraction of (tbt)2Ir(acac) phosphorescence originated from triplet excitons on €rster energy transfer. DMAC-DPS via up conversion and Fo In device P3, the distance between the (tbt)2Ir(acac) layer and the main charge recombination region should exceed the radius of €rster energy transfer, so the excitons on (tbt)2Ir(acac) come from Fo triplet exciton diffusion followed by Dexter energy transfer. In general, during the diffusion process, parts of triplet excitons are inevitably deactivated through other processes, such as TTA or TPA [16]. According to the above analyses, the emission processes of WOLEDs with the ultrathin (tbt)2Ir(acac) layer gradually away from the mCP/Bphen interface can be described in Fig. 5. When the (tbt)2Ir(acac) layer is located at the interface of mCP/Bphen in

device W3, the excitons on (tbt)2Ir(acac) molecules mainly come from direct exciton formation through hole trapping. It can be seen in Table S1 and Fig. 2(b) that, device W3 achieves high efficiency, but exhibits obvious saturation-induced blue shift with the increase of luminance. When the (tbt)2Ir(acac) layer is located far away from the interface of mCP/Bphen in device P3, the excitons on (tbt)2Ir(acac) totally come from Dexter energy transfer after triplet exciton diffusion. As shown in Table S1 and Fig. 2(f), device P3 realizes excellent color stability, but leads to low efficiency. Accordingly, the insertion of (tbt)2Ir(acac) layer adjacent to the main charge recombination region in device P1 which makes the exci€rster energy transfer, is tons on (tbt)2Ir(acac) mainly come from Fo optimal to the trade-off between color stability and efficiency. The reasons that color stability and efficiency are difficult to be improved simultaneously in WOLEDs with a narrow main charge recombination region, can be illuminated as follows: since DMACDPS molecules are the main sites of exciton formation and radiative decay in P-series devices with high color stability, the emission efficiency of blue EMLs mainly contributes to the EQE of WOLEDs,

Fig. 5. Schematic diagram of emission processes in devices W3, P1 and P3.

118

Y. Qi et al. / Organic Electronics 43 (2017) 112e120

which can be reflected by the similar EQE-L characteristics of devices B3 and P1. Meanwhile, since the T1 of DMAC-DPS (2.91 eV) is higher than that of Bphen (2.5 eV), the narrow major charge recombination region near the mCP/Bphen interface could not only bring exciton aggregation-induced quenching, but also lead to severe triplet exciton quenching by Bphen, which decrease the emission efficiency of blue EMLs. Besides, the T1 of DMAC-DPS (2.91 eV) is also higher than that of mCP (2.9 eV), which could cause back energy transfer. To simultaneously achieve good color stability and high efficiency, DPEPO with extremely high T1 level of 3.3 eV has been adopted as host to alleviate the triplet energy loss in blue EMLs [25]. The dominant electron transport capability of DPEPO could promote the balance of charge transport, which may realize a broadened major charge recombination region and a shift of this region towards anode compared with the mCP-based devices, so that the exciton aggregation is alleviated, and the amount of triplet excitons formed near and quenched by Bphen decreases. Meanwhile, the back energy transfer from DMAC-DPS to DPEPO is also inhibited. The EL parameters of blue device B4 are summarized in Table 2. The Von dramatically decreases to 3.0 V. Since large hole and electron injection barriers exist at the interfaces of mCP/DPEPO and Bphen/DPEPO, respectively, due to the deep HOMO and shallow lowest unoccupied molecular orbital (LUMO) levels of DPEPO. The heavily doped DMAC-DPS molecules provide additional injection and transport channels for both holes and electrons without injection barrier, and facilitate injection into and hopping on dopant molecules. Fig. 6(a) exhibits the J-V characteristics of hole-only and electron-only devices based on DPEPO neat and DPEPO: DMAC-DPS blend films, which are composed of ITO/TAPC (40 nm)/mCP (10 nm)/DPEPO (15 nm) or DPEPO: DMAC-DPS (50 wt%, 15 nm)/ TAPC (40 nm)/Ag and Mg: Ag/Bphen (40 nm)/DPEPO (15 nm) or DPEPO: DMAC-DPS (50 wt%, 15 nm)/Bphen (40 nm)/Mg: Ag, respectively. As DMAC-DPS molecules are doped in DPEPO host, the increase of current density at a constant voltage in the J-V curves of both hole-only and electron-only devices reveals the direct charge injection into DMAC-DPS molecules. Therefore, a wide major charge recombination region and a shift of this region towards anode are actually achieved by the heavily doped DMAC-DPS with bipolar transport ability, which promotes the balance of charge transport. The wide recombination region not only alleviates the exciton concentration quenching on DMAC-DPS, but also decreases the amount of triplet excitons quenched by Bphen. As a result, the EQEmax dramatically increases to 15.1%. In view of the dramatic efficiency enhancement of the DPEPO based blue device, WOLEDs W5-W7 with an ultrathin non-doped orange EML inserted gradually away from the DPEPO/Bphen interface have been fabricated. Fig. 6(bed) and Fig. 2(gei) exhibit the EL characteristics and spectra of the DPEPO based WOLEDs, respectively. The blue emission component percentage is higher than that of orange emission in the EL spectra of devices W5-W7, which can be attributed to the broadened major charge

recombination region, leading to the amount of excitons formed on and released from DMAC-DPS molecules more than that on (tbt)2Ir(acac) molecules. It is noteworthy that, the EQEmax and EQE at 1000 cd/m2 of all devices are above 15.5% and 12%, respectively. Accordingly, the efficient blue EMLs can contribute to improving the EQE of WOLEDs. Moreover, the higher efficiencies of devices W5-W7 than that of device B4 also indicate that the ultrathin or€rster energy ange EML alleviates energy loss on DMAC-DPS via Fo transfer. According to the EL spectra, device W5 with the ultrathin orange EML at the DPEPO/Bphen interface shows an apparent red shift with the increase of luminance. The orange EML is speculated to be out of the major charge recombination region due to a shift of this region towards anode. When device is driven by high current, the obvious red shift indicates that the amount of excitons directly formed on (tbt)2Ir(acac) increases, which may be attributed to the balanced charge transport and broadened major charge recombination region. As a result, device W5 exhibits a poor color stability with the highest CIE coordinates variation of (0.058, 0.054) and Icv value of 12.6%. It is noteworthy that, as the ultrathin orange EML is inserted away from the DPEPO/Bphen interface with a distance of 6 nm, device W7 shows a slight blue shift in EL spectra with the increase of luminance. When the ultrathin orange EML is inserted in the main charge recombination region, although (tbt)2Ir(acac) molecules can act as preferred charge recombination sites by hole trapping, the amount of holes trapped by (tbt)2Ir(acac) is negligible, which is attributed to the broadened main charge recombination region, and can be reflected by the almost overlapped J-V curves of devices B4 and W7. Consequently, excitons on (tbt)2Ir(acac) mole€rster energy transfer, so that the color cules mainly come from Fo stability is substantially improved. As a result, device W7 exhibits excellent color stability with slightly varied EL spectra, a drastically decreased CIE coordinates shift of (0.008, 0.003) and a small Icv value of 2.6% from 1000 to 10000 cd/m2.

4. Conclusion In summary, we have reported the hybrid WOLEDs with high color stability and efficiency based on an easily fabricated MEML structure, which is composed of a selective insertion of an ultrathin non-doped orange EML between heavily doped blue EMLs. According to the systematic analyses on the EL spectra and characteristics of several series of hybrid WOLEDs, the luminancedependent color shift and efficiency degradation are investigated and improved. As a result, a slight CIE coordinates variation of (0.008, 0.003) at a practical luminance range from 1000 to 10000 cd/m2, a PEmax of 45.8 lm/W, an EQEmax of 15.7% and an EQE above 12% at 1000 cd/m2 have been obtained. The heavily doped blue TADF emitters which act as the main charge transport channels and recombination sites in the host with high-lying lowest triplet excited state, make full use of the bipolar transport capability to broaden the major charge recombination region, which

Table 2 EL characteristics of blue OLEDs and WOLEDs based on DPEPO host. Device

Vona (V)

Lmax (cd/m2)

PEmax (lm/W)

PEb (lm/W)

LEmax (cd/A)

LEb (cd/A)

EQEmax (%)

EQEb (%)

CIEb (X, Y)

DCIEc (X, Y)

B4 W5 W6 W7

3.0 3.0 3.0 3.1

8500 11800 10050 10150

50.0 46.9 52.0 45.8

8.0 13.4 13.4 12.5

46.9 44.1 48.8 45.2

19.9 27.3 27.6 28.2

15.1 17.1 15.9 15.7

10.9 13.7 12.6 12.1

(0.161,0.252) (0.231,0.311) (0.267,0.353) (0.262,0.355)

e (0.058,0.054) (0.028,0.019) (0.008,0.003)

Abbreviations: Von: turn-on voltage. Lmax, PEmax, LEmax and EQEmax: maximum luminance and efficiencies. a Recorded at 1 cd/m2. b Recorded at 1000 cd/m2. c Recorded from 1000 to 10000 cd/m2.

Y. Qi et al. / Organic Electronics 43 (2017) 112e120

119

Fig. 6. (a) J-V characteristics of the single-carrier devices based on the DPEPO neat film and DPEPO: DMAC-DPS blend film with doping concentration of 50 wt%. (b) J-V-L (c) PE-J-LE and (d) EQE-L characteristics of the blue OLED and WOLEDs based on DPEPO host. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

alleviates triplet exciton quenching. The selectively inserted ultrathin non-doped orange EML makes its emission mechanism €rster energy transfer, which is effective to maintain dominated by Fo EL spectra stability with the increase of luminance. This work introduces an easily fabricated MEML structure and provides a guidance to achieve MEML based hybrid WOLEDs with excellent color stability and high efficiency. Acknowledgements The authors gratefully acknowledge the financial support from National Natural Science Foundation of China (NSFC) (Grant No. 61675041); the Foundation for Innovation Research Groups of the NSFC (Grant No. 61421002); Science and Technology Department of Sichuan Province, China (Grant No. 2016HH0027). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.orgel.2017.01.012. References [1] S.J. Su, E. Gonmori, H. Sasbe, J. Kido, Adv. Mater 20 (2008) 4189. [2] S. Reineke, F. Lindner, G. Schwartz, N. Seidler, K. Walzer, B. Lüssem, K. Leo, Nature 459 (2009) 234. [3] G.M. Farinola, R. Ragni, Chem. Soc. Rev. 40 (2011) 3467. [4] K.T. Kamtekar, A.P. Monkman, M.R. Bryce, Adv. Mater 22 (2010) 572. [5] M.A. Baldo, D.F. O'Brien, Y. You, A. Shoustikov, S. Sibley, M.E. Thompson, S.R. Forrest, Nature 395 (1998) 151. [6] H. Sasabe, J. Takamatsu, T. Motoyama, S. Watanabe, G. Wagenblast, N. Langer, O. Molt, E. Fuchs, C. Lennartz, J. Kido, Adv. Mater 22 (2010) 5003. [7] Q. Wang, J.Q. Ding, D.G. Ma, Y.X. Cheng, L.X. Wang, X.B. Jing, F.S. Wang, Adv. Funct. Mater 19 (2009) 84.

[8] H.B. Wu, G.J. Zhou, J.H. Zou, C.L. Ho, W.Y. Wong, W. Yang, J.B. Peng, Y. Cao, Adv. Mater 21 (2009) 4181. [9] G. Cheng, Y.F. Zhang, Y. Zhao, Y.Y. Lin, C.Y. Ruan, S.Y. Liu, T. Fei, Y.G. Ma, Y.X. Cheng, Appl. Phys. Lett. 89 (2006) 043504. [10] G. Schwartz, S. Reineke, T.C. Rosenow, K. Walzer, K. Leo, Adv. Funct. Mater 19 (2009) 1319. [11] Y.R. Sun, N.C. Giebink, H. Kanno, B. Ma, M.E. Thompson, S.R. Forrest, Nature 440 (2006) 908. [12] J. Ye, C.J. Zheng, X.M. Ou, X.H. Zhang, M.K. Fung, C.S. Lee, Adv. Mater 24 (2012) 3410. [13] X.H. Ouyang, X.L. Li, L. Ai, D.B. Mi, Z.Y. Ge, S.J. Su, ACS Appl. Mater. Interfaces 7 (2015) 7869. [14] C.J. Zheng, J. Wang, J. Ye, M.F. Lo, X.K. Liu, M.K. Fung, X.H. Zhang, C.S. Lee, Adv. Mater 25 (2013) 2205. [15] J. Ye, Z. Chen, F.F. An, M.L. Sun, H.W. Mo, X.H. Zhang, C.S. Lee, ACS Appl. Mater. Interfaces 6 (2014) 8964. [16] D.D. Zhang, M.H. Cai, Y.G. Zhang, D.Q. Zhang, L. Duan, ACS Appl. Mater. Interfaces 7 (2015) 28693. [17] H. Uoyama, K. Goushi, K. Shizu, H. Nomura, C. Adachi, Nature 492 (2012) 234. [18] X.K. Liu, Z. Chen, J. Qing, W.J. Zhang, B. Wu, H.L. Tam, F.R. Zhu, X.H. Zhang, C.S. Lee, Adv. Mater 27 (2015) 7079. [19] D.D. Zhang, L. Duan, Y.G. Zhang, M.H. Cai, D.Q. Zhang, Y. Qiu, Light Sci. Appl. 4 (2015) e232. [20] Q. Zhang, B. Li, S. Huang, H. Nomura, H. Tanaka, C. Adachi, Nat. Phot. 8 (2014) 326. [21] T.A. Lin, T. Chatterjee, W.L. Tsai, K.T. Wong, C.C. Wu, Adv. Mater 28 (2016) 6976. [22] B.P. Yan, C.C.C. Cheung, S.C.F. Kui, S.J. Xu, C.M. Che, Adv. Mater 19 (2007) 3599. [23] B.Q. Liu, Z.P. Xu, J.H. Zou, H. Tao, L. Wang, J. Ind. Eng. Chem. 27 (2015) 240. [24] Y.C. Li, Z.H. Wang, X.L. Li, D.C. Chen, J.B. Peng, Y. Cao, S.J. Su, Chem. Mater 27 (2015) 1100. [25] Q. Zhang, D. Tsang, H. Kuwabara, Y. Hatae, B. Li, T. Takahashi, S. Lee, T. Yasuda, C. Adachi, Adv. Mater 27 (2015) 2096. [26] M. Uchida, C. Adachi, T. Koyama, Y. Taniguchi, J. Appl. Phys. 86 (1999) 1680. [27] C. Adachi, M.A. Baldo, M.E. Thompson, S.R. Forrest, J. Appl. Phys. 90 (2001) 5048. [28] J.N. Yu, H. Lin, L. Tong, C. Li, H. Zhang, J.H. Zhang, Z.X. Wang, B. Wei, Phys. Status Solidi A 210 (2) (2013) 408. [29] S. Lamansky, P. Djurovich, D. Murphy, F. Abdel-Razzaq, R. Kwong, I. Tsyba, M. Bortz, B. Mui, R. Bau, M.E. Thompson, Inorg. Chem. 40 (2001) 1704. [30] P.M. Borsenberger, L. Pautmeier, R. Richert, H. Bassler, J. Chem. Phys. 94 (1991)

120

Y. Qi et al. / Organic Electronics 43 (2017) 112e120

8276. [31] C.H. Hsiao, S.W. Liu, C.T. Chen, J.H. Lee, Org. Electron 11 (2010) 1500. [32] R.J. Holmes, S.R. Forrest, J.J. Brown, S. Garon, M.E. Thompson, Appl. Phys. Lett. 82 (2003) 2422. [33] C.H. Hsiao, Y.H. Lan, P.Y. Lee, T.L. Chiu, J.H. Lee, Org. Electron 12 (2011) 547. [34] Q. Wang, D.G. Ma, J.Q. Ding, H.P. Jia, B.E. Gnade, J.H. Halbert, Org. Electron 19 (2015) 157.

[35] S.Q. Liu, J. Li, C.L. Du, J.S. Yu, Appl. Phys. Lett. 107 (2015) 041109. [36] H. Wang, L.Q. Meng, X.X. Shen, X.F. Wei, X.L. Zheng, X.P. Lv, Y.P. Yi, Y. Wang, P.F. Wang, Adv. Mater 27 (2015) 4041. [37] T. Furukawa, H. Nakanotani, M. Inoue, C. Adachi, Sci. Rep. 5 (2015) 8429. [38] C. Li, L. Duan, D.D. Zhang, Y. Qiu, ACS Appl. Mater. Interfaces 7 (2015) 15154. [39] D.D. Zhang, L. Duan, C. Li, Y.L. Li, H.Y. Li, D.Q. Zhang, Y. Qiu, Adv. Mater 26 (2014) 5050.