High-Energy-Level Blue Phosphor for Solution-Processed White Organic Light-Emitting Diodes with Efficiency Comparable to Fluorescent Tubes.

A high-energy-level blue phosphor FIr-p-OC8 has been developed for solution-processed white organic light-emitting diodes (WOLEDs) with comparable fluorescent tube efficiency. Benefiting from the electron-donating nature of the introduced alkoxy, FIr-p-OC8 shows not only efficient blue light but also elevated highest occupied molecular orbital/lowest unoccupied molecular orbital levels to well match the dendritic host H2. Consequently, the hole scattering between FIr-p-OC8 and H2 can be prevented to favor the direct exciton formation on the blue phosphor, leading to reduced driving voltage and thus improved power efficiency. By exploiting this approach, a maximum power efficiency of 68.5 lm W-1 is achieved for FIr-p-OC8-based white devices, slightly declining to 47.0 lm W-1 at a practical luminance of 1,000 cd m-2. This efficiency can be further raised to 96.3 lm W-1 @ 1,000 cd m-2 when a half-sphere is applied to increase light out-coupling. We believe that our results can compete with commercial fluorescent tubes, representing an important progress in solution-processed WOLEDs.

Introduction

White organic light-emitting diodes (WOn class="Gene">LEDs) have attracted intense academic an class="Disease">nd industrial attention owing to their potential to replace fluorescent tubes for applications in energy-saving lightings (Kido et al., 1995, Sun et al., 2006, Reineke et al., 2009, Park et al., 2008, Su et al., 2008, Wang et al., 2009, Kamtekar et al., 2010, Chang et al., 2013, Ou et al., 2014, Ding et al., 2015, Zhou et al., 2015, Liu et al., 2016, Wu et al., 2008, Wu et al., 2009, Wu et al., 2016, Huang et al., 2009, Zou et al., 2011, Fu et al., 2012, Zhang et al., 2012, Zhang et al., 2013, Aizawa et al., 2014, Chiba et al., 2015). Nowadays most of the attention is focused on vacuum-deposited WOLEDs (v-WOLEDs) (Ou et al., 2014, Ding et al., 2015, Zhou et al., 2015, Liu et al., 2016, Wu et al., 2016). Since the first report in 1995 (Kido et al., 1995), the power efficiency (PE) of v-WOLEDs has been significantly improved and has surpassed those of the commercial fluorescent tubes (60–100 lm W−1, the threshold for practical applications) (Reineke et al., 2009, Ou et al., 2014, Ding et al., 2015). Nevertheless, solution-processed WOLEDs (s-WOLEDs) are under less investigation, although they are believed to be superior to v-WOLEDs in view of the advantages including low cost, large area, and flexible device fabrication (Wu et al., 2008, Wu et al., 2009, Huang et al., 2009, Zou et al., 2011, Fu et al., 2012, Zhang et al., 2012, Zhang et al., 2013, Aizawa et al., 2014, Chiba et al., 2015). Their PE remains far away from the fluorescent tube efficiency even if transition-metal-containing phosphors are used to harvest both singlet and triplet excitons to increase the external quantum efficiency (EQE) (Wu et al., 2008, Wu et al., 2009, Zou et al., 2011, Fu et al., 2012, Zhang et al., 2012, Zhang et al., 2013, Aizawa et al., 2014, Chiba et al., 2015). Therefore, the accompanying high driving voltage may account for the poor PE of s-WOLEDs because it is directly proportional to the ratio of EQE to driving voltage (PE ∝ EQE/V).

A single emitting layer (EML) device structure (Wu et al., 2008, Wu et al., 2009, Huang et al., 2009, Zou et al., 2011, Fu et al., 2012, Zhang et al., 2012, Zhang et al., 2013, Aizawa et al., 2014), which consists of a host don class="Chemical">ped with blue an class="Disease">nd long-wavelength phosphor, is usually adopted in s-WOLEDs since the layer beneath could be dissolved by the solution deposition of a subsequent layer. Considering the negligible contribution from the long-wavelength phosphor with an extremely low doping concentration in EML, the host and blue phosphor are two critical factors that are related to the high driving voltage originating from the limited charge injection and transport. First, a suitable host with a triplet energy higher than that of blue phosphor is desirable to avoid triplet exciton back transfer from the dopant to the host (Ding et al., 2009, Lee and Lee, 2013, Park et al., 2013, Lee et al., 2015, Seino et al., 2014). So an increased driving voltage should be provided to overcome large energy barriers before carriers (holes and/or electrons) are injected into such wide-band-gap host. Second, either the highest occupied molecular orbital (HOMO) level or the lowest unoccupied molecular orbital (LUMO) level of the blue phosphor does not match with those of the host used. The resultant charge trapping or scattering may bring about the enhancement of driving voltage by decreasing the charge transport of the EML (Tsung and So, 2008, Li et al., 2011, Li et al., 2012a, Li et al., 2012b, Li et al., 2014, Schmechel and von Seggern, 2004, Tong et al., 2007, von Malm et al., 2001).

Aiming at low-voltage-driving an class="Disease">nd thus power-effin class="Chemical">cient s-WOLEDs, the first challenge is to develop solution-processible host materials that exhibit appropriate energy levels to ensure efficient charge injection without sacrificing high triplet energy. Recently, a dendritic oligocarbazole molecule H2 has been demonstrated as a good host for the blue phosphor iridium(III)[bis(4,6-difluorophenyl)-pyridinato-N,C2]-picolinate (FIrpic) (Ding et al., 2009). When compared with the archetypical high-triplet-energy polymeric host poly(N-vinylcarbazole) (PVK) with a deep HOMO level of −5.9 eV, H2 possesses a triplet energy as high as 2.9 eV and an elevated HOMO level up to −5.28 eV simultaneously. As a result of the reduced hole injection barrier from PVK to H2, the PE of H2-based blue device is improved from 8.3 to 15.4 lm W−1. Further co-doping a small amount of yellow phosphor iridium(III)[5-trifluoromethyl-2-(9,9-diethylfluoren-2-yl)pyridine] (Ir(Flpy-CF3)3) into the EML leads to a promising white light (Zhang et al., 2012), showing a PE of 23.3 lm W−1 at a practical luminance of 1,000 cd m−2. In this case, although the host is well modified to facilitate hole injection, hole transport could be obviously influenced by the associated hole scattering because the HOMO level of FIrpic (−5.64 eV) is much lower than that of H2. To solve this bottleneck, the second challenge is to design novel blue phosphors with high energy levels so as to match with those of the dendritic host. However, this important issue has rarely been addressed to date.

Here we present the realization of power-effin class="Chemical">cient s-WOn class="Gene">LEDs with PE approaching 100 lm W−1 at a luminance of 1,000 cd m−2 by using a newly developed high-energy-level blue phosphor FIr-p-OC8 (Figure 1). Unlike the commonly used FIrpic (−5.64/-2.62 eV), the para substitution of alkoxy group endows FIr-p-OC8 with elevated HOMO/LUMO levels of −5.28/-2.24 eV while maintaining efficient blue emission. When combined with the dendritic host H2 (Figure 3D), the hole scattering between FIr-p-OC8 and H2 can be effectively prevented because of their matched HOMO levels, thus allowing excitons to be formed directly on FIr-p-OC8. With this approach, low driving voltage and high PE have been successfully achieved for the solution-processed white light-emitting device based on FIr-p-OC8, revealing a state-of-art PE of 68.5 lm W−1 at the maximum and 47.0 lm W−1 at 1,000 cd m−2. Also, when a half-sphere is applied to increase the light out-coupling, these efficiencies can be further raised to 130.1 and 96.3 lm W−1, respectively. To our knowledge, such performance is the highest ever reported so far for s-WOLEDs, which could even compete with those of commercial fluorescent tubes.

Figure 1

Design of High-Energy-Level Blue Phosphors

(A) Molecular structures of blue phosphors (FIrpic, FIr-m-OC8 and FIr-p-OC8).

(B) UV-Vis absorption spectra in dichloromethane (DCM) together with photoluminescent (PL) spectra in toluene (The black circles and arrows indicate the corrsponding vertical axis for each data curves).

(C) PL decay curves in toluene.

(D) Cyclic voltammograms in solution using 0.1M n-Bu4NClO4 as supporting electrolyte at a scan rate of 100 mV s−1.

(E) HOMO/LUMO level alignment of the blue phosphors.

Also see Scheme S1 and Figure S1.

Figure 3

Analysis of the Hole Scattering-Induced Influence

(A) Working mechanism of FIrpic-based blue device.

(B) Current density-voltage characteristics of FIrpic-based hole-only devices with different doping concentration of 0, 1, 5, and 15 wt.%.

(C) Transient EL decay curves for FIrpic-based blue device with a doping concentration of 15 wt. %.

(D) Working mechanism of FIr-p-OC8-based blue device.

(E) Current density-voltage characteristics of FIr-p-OC8-based hole-only devices with different doping concentration of 0, 1, 5, and 15 wt.%.

(F) Transient EL decay curves for FIr-p-OC8-based blue device with a doping concentration of 15 wt.%.

Also see Figure S4 and Table S1.

(A) Molecular structures of blue phosphors (n class="Chemical">FIrpic, FIr-m-n class="Chemical">OC8 and FIr-p-OC8).

(B) UV-Vis absorption sn class="Chemical">pectra inpan> n class="Chemical">dichloromethane (DCM) together with photoluminescent (PL) spectra in toluene (The black circles and arrows indicate the corrsponding vertical axis for each data curves).

(C) n class="Chemical">PL decay curves in n class="Chemical">toluene.

(D) Cyclic voltammograms in solution using 0.1M n class="Chemical">n-Bu4NClO4 as supportinpan>g electrolyte at a scan rate of 100 mV s−1.

Also see Scheme S1 an class="Disease">nd Figure S1.

Results and Discussion

Molecular Design

To tune the energy level of blue phosphors, the alkoxy group with strong electron-donating capability is introduced into the meta an class="Disease">nd para positions to the n class="Chemical">N atom in the 2-(2′,4′-difluorophenyl)pyridine ligand, and the corresponding homoleptic Ir complexes FIr-m-OC8 and FIr-p-OC8 are synthesized, respectively (Figure 1A). During the procedure (Scheme S1), a convenient post-modification is adopted and begins with the preparation of an emissive Ir core functionalized with three reactive hydroxyl groups (FIr-m-OH and FIr-p-OH), which subsequently reacts with octyl bromide to give the final Ir complexes. Their detailed synthesis and characterization are provided in the Supplemental Information.

Figures 1B an class="Disease">nd 1C plot the steady and transient photoluminescence spectra of the complexes in toluene solution, respectively. With respect to FIrpic (λem = 472 nm), as one can see, FIr-m-OC8 displays a red-shifted emission at 481 nm, whereas a hypsochromic shift (λem = 462 nm) is observed for FIr-p-OC8 (Table 1). In addition, the exited state lifetime in toluene solution of FIr-p-OC8 (0.85 μs) is found to be shorter than those of FIrpic (1.02 μs) and FIr-m-OC8 (1.96 μs), in good agreement with the higher photoluminescence quantum yield (PLQY) in toluene solution of FIr-p-OC8 (0.78) than those of FIrpic (0.60) and FIr-m-OC8 (0.23). Most importantly, as estimated from their cyclic voltammetry curves (Figure 1D), the HOMO/LUMO levels are gradually increased from −5.64/-2.62 eV of FIrpic to −5.44/-2.37 eV of FIr-m-OC8 and −5.28/-2.24 eV of FIr-p-OC8 (Figure 1E). The upshift of HOMO/LUMO is in good agreement with the theoretical calculations (Figure S1), indicating that the position and electron-donating effect of the alkoxy substituent can be employed to finely tailor the photophysical and electrochemical properties of blue phosphors.

Table 1

Summary of Photophysical, Electrochemical, and Thermal Properties for the Dendritic Host and Blue Phosphors

Material	λabs(logɛ)a (nm)	λemb (nm)	PLQYc	τ(μs)d	Eoxe (V)	Erede (V)	HOMOf (eV)	LUMO (eV)	T1h (eV)	Tm/Td (°C)
H2	242 (5.5), 268 (5.1), 288 (5.1), 298 (5.2), 350 (4.5)	391, 408	–	–	0.48	–	−5.28	−1.91g	2.86	−/−
FIrpic	256 (4.7), 381 (3.8), 412 (3.5), 427 (3.3), 456 (2.8)	472, 499	0.60	1.02	0.84	−2.18	−5.64	−2.62f	2.63	ND/330
FIr-m-OC8	244 (4.7), 263 (4.6), 323 (4.2), 371 (3.8), 379 (3.7)	481, 497	0.23	1.96	0.64	−2.43	−5.44	−2.37f	2.58	155/337
FIr-p-OC8	265 (4.8), 347 (4.1), 381 (3.9), 417 (3.4), 448 (2.9)	462, 485	0.78	0.85	0.48	−2.56	−5.28	−2.24f	2.69	222/341

Also see Figures S2–S4, and Table S1.

Measured in DCM at 298 K with a concentration of 10−5 M.

Measured in toluene at 298 K with a concentration of 10−5 M.

Measured in degassed toluene at 298 K excited at 350 nm with a concentration of 10−5 M using a calibrated integrating sphere.

Measured in degassed toluene at 298 K, and the lifetimes were obtained by a monoexponential fit of PL decay curves.

Eox is the first oxidation onset measured in DCM, and Ered is the first reduction onset measured in DMF.

The HOMO and LUMO energy levels were calculated according to the equations: HOMO = -e(Eox+ 4.8 V) and LUMO = -e(Ered+ 4.8 V).

The LUMO levels are calculated according to the equation: LUMO = HOMO + Eg, where Eg is the optical bandgap estimated from the absorption onset.

The triplet energy levels were calculated according to the phosphorescence peak wavelength positioned on the shortest wavelength side measured in toluene solution with a concentration of 10−5 M.

Summary of Photophysical, Electrochemical, an class="Disease">nd Thermal Pron class="Chemical">perties for the Dendritic Host and Blue Phosphors

Also see Figures S2–S4, an class="Disease">nd Table S1.

Measured in n class="Disease">DCM at 298 K with a concentration of 10−5 M.

Measured in n class="Chemical">toluene at 298 K with a concentration of 10−5 M.

Measured in degassed n class="Chemical">toluene at 298 K excited at 350 nm with a conpan>centrationpan> of 10−5 M usinpan>g a calibrated inpan>tegratinpan>g sphere.

Measured in degassed n class="Chemical">toluene at 298 K, an class="Disease">nd the lifetimes were obtained by a monoexponential fit of PL decay curves.

Eox is the first oxidation onset measured in n class="Disease">DCM, an class="Disease">nd Ered is the first reduction onset measured in DMF.

The HOMO an class="Disease">nd LUMO energy levels were calculated accordinpan>g to the equationpan>s: HOMO = -e(Eox+ 4.8 V) an class="Disease">nd LUMO = -e(Ered+ 4.8 V).

The LUMO levels are calculated according to the equation: LUMO = HOMO + Eg, where Eg is the optical ban class="Disease">ndgap estimated from the absorptionpan> onpan>set.

The trin class="Chemical">plet energy levels were calculated accordinpan>g to the phosphorescence peak wavelength positioned on the shortest wavelength side measured in toluene solution with a concentration of 10−5 M.

Superiority of High-Energy-Level Blue Phosphor

In view of the obtained high-lying HOMO/LUMO levels, blue-shifted emission, high Φn class="Chemical">PL, an class="Disease">nd shorter lifetime, FIr-p-OC8 is anticipated to be a better blue phosphor than FIrpic. To demonstrate this point, two types of monochromic devices are prepared at the same time for comparison: ITO(indium tin oxide)/PEDOT:PSS(poly(ethylenedioxythiophene):poly(styrenesulfonate))/H2:FIrpic/TPCz/TmPyPb/LiF/Al (denoted as device A) and ITO/PEDOT:PSS/H2:FIr-p-OC8/TPCz/TmPyPb/LiF/Al (denoted as device B) (Figure S5). When the doping concentration grows up from 1 wt.% to 15 wt.%, the turn-on voltage for device A is monotonically increased from 3.9 to 4.5 V along with the reduced current density at the same driving voltage (Figure S6 and Table S2). On the contrary, the turn-on voltage for device B is in the range of 3.1–3.3 V, and the current density-voltage (J-V) characteristics remain nearly unchanged, irrespective of the doping concentration. Accordingly, device B exhibits reduced driving voltage and improved PE than device A. As depicted in Figure 2, both the J-V and luminance-voltage (L-V) curves at a 15 wt.% content shift by about 1.5 V toward a negative voltage on going from FIrpic to FIr-p-OC8. Under a luminance of 100 and 1,000 cd m−2, the driving voltages are significantly decreased from 5.6 to 6.3 V to 4.0 and 4.5 V. Meanwhile, the PEs are raised from 17.8 and 16.0 lm W−1 to 26.2 and 20.3 lm W−1.

Figure 2

Performance Comparison between FIrpic- and FIr-p-OC8-Based Blue Devices with the Same Doping Concentration (15 wt. %)

(A) Current density-voltage-luminance characteristics (The black circles and arrows indicate the corrsponding vertical axis for each data curves).

(B) Power efficiency-luminance characteristics.

Also see Figures S5 and S6, and Table S2.

n class="Chemical">Performance Comparisonpan> between n class="Chemical">FIrpic- and FIr-p-OC8-Based Blue Devices with the Same Doping Concentration (15 wt. %)

(A) Current density-voltage-luminance characteristics (The black circles an class="Disease">nd arrows indicate the corrsponding vertical axis for each data curves).

(B) Power effin class="Chemical">ciency-luminpan>ance characteristics.

Also see Figures S5 an class="Disease">nd S6, an class="Disease">nd Table S2.

Given the same configuration, the observed difference of the device n class="Chemical">performance between n class="Chemical">FIrpic and FIr-p-OC8 can be reasonably ascribed to their different energy alignments with regard to H2. In device A, holes are injected and transported in H2 but not in FIrpic since FIrpic has a deeper HOMO level than H2 (Figure 3A). On one hand, FIrpic can act as hole scatters and cause an increase of the total transit path and thus the decrease of the hole mobility. As verified by the hole-only devices (Figure 3B), the hole current at the same driving voltage is greatly lowered with increasing content of FIrpic. On the other hand, electrons can be injected from TPCz/TmPyPb either into FIrpic preferentially or into H2 through electrostatic attraction and then trapped by FIrpic as a result of the lower lying LUMO of FIrpic relative to H2. These captured electrons on FIrpic can neither escape from FIrpic due to the large electron-trapping depth between FIrpic and H2 (0.70 eV) nor form the radiative excitons because holes occur scarcely in FIrpic due to the existing hole scattering. Consequently, electrons would be inevitably stored on FIrpic molecules to some extent, which is further confirmed by time-resolved electroluminescence measurements (Liu et al., 2011, Weichsel et al., 2012). As one can see in Figure 3C, a distinct transient overshoot, corresponding to an electron storage, is distinctly observed for device A after voltage turn-off. When a negative postpulse bias is applied from 0 to −3 and −5 V, the overshoot time is gradually shortened, associated with the enhanced overshoot intensity. In other words, the existence of hole scattering in device A would cause both reduced hole mobility and electron accumulation on dopant, which is responsible for the inferior device performance of FIrpic.

Analysis of the Hole Scattering-In class="Disease">nduced Inpan>fluence

(A) Working mechanism of n class="Chemical">FIrpic-based blue device.

(B) Current density-voltage characteristics of n class="Chemical">FIrpic-based hole-onpan>ly devices with different dopinpan>g conpan>centrationpan> of 0, 1, 5, an class="Disease">nd 15 wt.%.

(C) Transient EL decay curves for n class="Chemical">FIrpic-based blue device with a dopinpan>g conpan>centrationpan> of 15 wt. %.

(D) Working mechanism of n class="Chemical">FIr-p-OC8-based blue device.

(E) Current density-voltage characteristics of n class="Chemical">FIr-p-OC8-based hole-onpan>ly devices with different dopinpan>g conpan>centrationpan> of 0, 1, 5, an class="Disease">nd 15 wt.%.

(F) Transient EL decay curves for n class="Chemical">FIr-p-OC8-based blue device with a dopinpan>g conpan>centrationpan> of 15 wt.%.

Also see Figure S4 an class="Disease">nd Table S1.

In contrast, hole scattering is not present anymore in device B because the HOMO level of n class="Chemical">FIr-p-OC8 matches well with n class="Chemical">H2. Holes can hop freely between H2 and FIr-p-OC8, and hole transport is not affected by the doped blue phosphor (Figure 3D). This is consistent with the hole-only devices (Figure 3E), where the hole current is found to be independent of the doping concentration of FIr-p-OC8. Meanwhile, the electrons that are energetically injected from TPCz/TmPyPb to FIr-p-OC8 owing to their close LUMO levels would not accumulate on FIr-p-OC8, but recombine with the holes to generate excitons directly. As shown in Figure 3F, no transient overshoot is detected for device B, implying the vanishment of electron accumulation. Therefore, the unwanted hole scattering can be eliminated in device B to favor the direct exciton generation on the blue phosphors, leading to the reduction of driving voltage and enhancement of PE.

Solution-Processed White Devices

Promoted by the sun class="Chemical">periority of n class="Chemical">FIr-p-OC8 to FIrpic in terms of the lower driving voltage and higher PE, FIr-p-OC8 is then used to fabricate power-efficient s-WOLEDs. Considering the fact that the PE of FIr-p-OC8 does not reach the maximum value in the range of 1–15 wt.% (Table S2), the FIr-p-OC8 content doped into H2 is tuned to optimize the monochromic device performance. It is found that the best device performance is obtained at a much higher doping concentration of 25 wt.%, giving a peak PE of 34.2 lm W−1 (37.8 cd A−1, 19.6%) with Commission Internationale de l’Eclairage (CIE) coordinates of (0.15, 0.26) (Figures 4A and S7, Tables 2 and S3). Even at a luminance of 100 and 1,000 cd m−2, the PE still remains at 33.1 and 26.9 lm W−1, respectively, indicative of the gentle PE roll-off. It could be ascribed to the very low driving voltage of only about 3.6/4.1 V at a luminance of 100/1,000 cd m−2. In addition, compared with a device with 15 wt. % FIr-p-OC8, the driving voltage at 1,000 cd m−2 for a device with 25 wt. % FIr-p-OC8 is decreased by about 0.4 V, and the PE is correspondingly increased by about 33%. As discussed above, the formation of excitons directly on blue phosphor is mainly responsible for the efficient blue light after the elimination of hole scattering. Thus the higher content of the blue dopant into H2 could contribute more to this pathway, and result in PE improvement.

Figure 4

Performance of FIr-p-OC8-Based Blue and White Devices with an EML Composed of H2, 25 wt.%; FIr-p-OC8, x wt.% Ir(Flpy-CF3)3

(A–C) EL spectra at 1,000 cd m−2; current density-voltage-luminance characteristics; and power efficiency as a function of luminance for devices without out-coupling (The black circles and arrows indicate the corrsponding vertical axis for each data curves).

(D–F) EL spectra at 1,000 cd m−2; current density-voltage-luminance characteristics; and power efficiency as a function of luminance for devices with light out-coupling (The black circles and arrows indicate the corrsponding vertical axis for each data curves).

Also see Figures S5 and S7–S15, and Tables S3 and S4.

Table 2

Summary of Device Performance for FIr-p-OC8-Based Blue and White Light-Emitting Devices with an EML Composed of H2: 25 wt.% and FIr-p-OC8: x wt.% Ir(Flpy-CF3)3

Content of Ir(Flpy-CF3)3	Von (V)	Max Performance				Device Performance at 1000/5000 cd m−2
		L (cd m−2)	LE (cd A−1)	PE (lm W−1)	EQE (%)	Vd (V)	LE (cd A−1)	PE (lm W−1)	EQE (%)	CIE (x, y) at 1000 cd m−2
x = 0 (Without out-coupling)	2.9	17,655	37.8	34.2	19.6	4.1/4.7	34.9/27.1	26.9/17.5	18.1/14.1	(0.15, 0.26)
x = 0 (With out-coupling)	2.9	38,580	70.9	63.0	34.2	4.1/4.6	66.9/56.2	50.9/38.6	32.2/27.0	(0.16, 0.28)
x = 0.4 wt.% (Without out-coupling)	2.9	23,392	60.4	59.4	21.8	4.1/4.6	52.7/37.7	40.2/25.8	19.6/14.5	(0.34, 0.39)
x = 0.4 wt.% (With out-coupling)	2.8	44,177	120.3	117.5	42.8	3.9/4.3	110.5/85.0	88.7/61.9	41.0/32.2	(0.36, 0.40)
x = 0.7 wt.% (Without out-coupling)	2.8	23,085	69.1	68.5	23.6	4.1/4.6	59.7/43.1	47.0/29.8	20.6/15.4	(0.43, 0.44)
x = 0.7 wt.% (With out-coupling)	2.8	44,485	131.5	130.1	44.6	3.9/4.3	119.5/95.0	96.3/69.0	40.9/33.0	(0.44, 0.44)

Also see Figures S10 and S13, and Table S4.

n class="Chemical">Performance of n class="Chemical">FIr-p-OC8-Based Blue and White Devices with an EML Composed of H2, 25 wt.%; FIr-p-OC8, x wt.% Ir(Flpy-CF3)3

(A–C) EL sn class="Chemical">pectra at 1,000 n class="Chemical">cd m−2; current density-voltage-luminance characteristics; and power efficiency as a function of luminance for devices without out-coupling (The black circles and arrows indicate the corrsponding vertical axis for each data curves).

(D–F) EL sn class="Chemical">pectra at 1,000 n class="Chemical">cd m−2; current density-voltage-luminance characteristics; and power efficiency as a function of luminance for devices with light out-coupling (The black circles and arrows indicate the corrsponding vertical axis for each data curves).

Also see Figures S5 an class="Disease">nd S7–n class="Gene">S15, and Tables S3 and S4.

Summary of Device n class="Chemical">Performance for n class="Chemical">FIr-p-OC8-Based Blue and White Light-Emitting Devices with an EML Composed of H2: 25 wt.% and FIr-p-OC8: x wt.% Ir(Flpy-CF3)3

Also see Figures n class="Gene">S10 an class="Disease">nd S13, and Table S4.

Based on such an optimized composition (n class="Chemical">H2: 25 wt.% n class="Chemical">FIr-p-OC8), subsequently, an additional yellow phosphor Ir(Flpy-CF3)3 is added into the EML for the fabrication of s-WOLEDs (Figure S5). The absolute concentration of Ir(Flpy-CF3)3 is initially adjusted to be 0.4 wt.% so as to balance the blue and yellow components for the realization of a satisfactory white emission with CIE coordinates of (0.34, 0.39) (Figure 4A), close to the standard white point, viz., (0.33, 0.33). At such a low doping concentration, Ir(Flpy-CF3)3 is considered to have no influence on the device working mechanism, which is evidenced by nearly the same J-V characteristics between monochromic and two-color white lightemitting devices (Figure 4B). Similar to the monochromic device, hole scattering between host and dopant is avoided and most of the excitons are produced directly on the blue phosphor in the white device, followed by a subsequent energy transfer to the yellow phosphor to generate the white light. As a consequence, the advantages of the monochromic device, such as the ultralow driving voltage, high PE, and small efficiency roll-off, could be inherited by the white device. For example, the driving voltage is as low as 2.9, 4.1, and 4.6 V at 1, 1,000 and 5,000 cd m−2, respectively. The maximum PE is 59.4 lm W−1, slightly declining to 40.2 and 25.8 lm W−1 at a luminance of 1,000 and 5,000 cd m−2, respectively (Figure 4C). Further increasing the doping concentration of Ir(Flpy-CF3)3 up to 0.7 wt.% results in a warm white light with CIE coordinates of (0.43, 0.44). Correspondingly, the PE is enhanced to 68.5 lm W−1 at the maximum; 47.0 lm W−1 at 1,000 cd m−2; and 29.8 lm W−1 at 5,000 cd m−2. When compared with FIrpic-based white-light-emitting devices showing hole scattering (23.3 lm W−1 at 1,000 cd m−2) (Zhang et al., 2012), the PE at high luminance is increased by about two-fold for FIr-p-OC8 free of hole scattering. The improvement clearly highlights the importance of high-energy-level blue phosphors in maximizing the PE of solution-processed white-emitting devices by the elimination of hole scattering.

As for the white devices built in a stan class="Disease">ndard substrate-emitting architecture, the light out-coun class="Chemical">pling efficiency is approximately 20%, and the remaining 80% of the light is trapped in organic and substrate modes. Therefore, similar to the literature report (Reineke et al., 2009), a half-sphere is pasted on the bottom of the ITO glass to unlock the trapped light as much as possible. As one can see in Figure 4D, the adoption of such an out-coupling technique has little effect on the electroluminescent (EL) spectra. The CIE coordinates slightly shift from (0.34, 0.39) and (0.43, 0.44) to (0.36, 0.40) and (0.44, 0.44) for the white devices containing 0.4 wt.% and 0.7 wt.% Ir(Flpy-CF3)3, respectively. Moreover, the luminance at the same driving voltage is found to be significantly increased, albeit the independent current density (Figures 4E and S13A), leading to PE improvement (Figures 4F and S13C). With the white device containing 0.7 wt.% Ir(Flpy-CF3)3 as an example, the maximum luminance is from 23,085 to 44,485 cd m−2 accompanied by the enhanced maximum PE from 68.5 to 130.1 lm W−1 (Table 2). Noticeably, the PE at a high luminance of 1,000 and 5,000 cd m−2 could also reach up to 96.3 and 69.0 lm W−1, respectively. These values are the highest ever reported for s-WOLEDs, which are comparable to those of v-WOLEDs and even competent with the commercial fluorescent tubes (Table 3). Interestingly, the related correlated color temperature and CIE coordinates can be well tuned to satisfy different illumination requirements without considerably sacrificing the PE (Figure S15 and Table S4).

Table 3

Performance Comparison of the Phosphorescent WOLEDs

Year	Reference	PE at the Maximum (lm W−1)	PE at 1,000 cd m−2 (lm W−1)	CIE (x, y)
s-WOLEDs
2008	Wu et al., 2008	10.0	–	(0.30, 0.47)
2009	Huang et al., 2009	23.4	–	(0.38, 0.38)
2009	Wu et al., 2009	20.3	16.8	(0.40, 0.45)
2011	Zou et al., 2011	37.4	20.7	(0.31, 0.50)
2012	Zhang et al., 2012	–	23.3	(0.38, 0.43)
2013	Zhang et al., 2013	–	22.6	(0.35, 0.41)
2014	Aizawa et al., 2014	45	18	(0.45, 0.42)
2015	Wang et al., 2015	41.8	22.8	(0.43, 0.43)
2018	This work	59.4 (w/o)117.5 (w/-)	40.288.7	(0.34, 0.39)(0.36, 0.40)
2018	This work	68.5 (w/o)130.1 (w/-)	4796.3	(0.43, 0.44)(0.44, 0.44)
v-WOLEDs
2008	Su et al., 2008	–	44	(0.34, 0.40)
2009	Reineke et al., 2009	-- (w/o)-- (w/-)	33124	(0.45, 0.47)(0.41, 0.49)
2011	Wang et al., 2011	40.7	37.1	(0.42, 0.44)
2013	Chang et al., 2013	–	33.8	(0.44, 0.46)
2014	Zhu et al., 2014	63.2	53.4	(0.35, 0.42)
2014	Ou et al., 2014	68.8 (w/o)132.8 (w/-)	60.0123.4	(0.35, 0.46)(0.35, 0.46)
2015	Ding et al., 2015	64.9	49.6	(0.35, 0.46)
2015	Zhu et al., 2015	46.6	41.3	(0.45, 0.44)
2017	Wu et al., 2017	105.0	59.5	(0.40,0.48)

w/o, without out-coupling; w/-, with out-coupling.

n class="Chemical">Performance Comparisonpan> of the Phosphorescent WOn class="Gene">LEDs

w/o, without out-coun class="Chemical">plinpan>g; w/-, with out-coun class="Chemical">pling.

Conclusion

In summary, we report power-effin class="Chemical">cient s-WOn class="Gene">LEDs with PE approaching 100 lm W−1 at l,000 cd m−2 luminance. This is obtained by newly developing a high-energy-level blue phosphor FIr-p-OC8 in combination with the dendritic host H2, where hole scattering can be eliminated to allow the direct formation of excitons. As a result, the solution-processed white-light-emitting device based on FIr-p-OC8 realizes a record-high PE of 68.5 lm W−1 at the maximum and 47.0 lm W−1 at a practical luminance of 1,000 cd m−2, which can be further improved to 130.1 and 96.3 lm W−1, respectively, under light out-coupling. We believe that our work represents an important progress in s-WOLEDs and demonstrates their bright future in low-cost, large-area, and flexible energy-saving lightings.

Methods

All methods can be foun class="Disease">nd inpan> the accompanyinpan>g Transparent Methods supn class="Chemical">plemental file.