Achieving High Efficiency and Pure Blue Color in Hyperfluorescence Organic Light Emitting Diodes using Organo-Boron Based Emitters

In the field of organic light-emitting diodes, organo-boron based thermally activated delayed fluorescence (TADF) emitters have witnessed outstanding achievements. However, it is still challenging to achieve pure blue color (CIE y < 0.20) along with high efficiencies. To overcome these hurdles, the hyperfluorescence (HF) system suggests a key strategy for future display applications. Here, two TADF host materials, p MDBA-DI and m MDBA-DI, and a pure blue multi-resonance type tert-butyl substituted TADF fluorescence emitter t -Bu- ν -DABNA are reported, for efficient HF devices. Among the synthesized TADF sensitized host materials, the m MDBA-DI HF device exhibits a high external quantum efficiency of 39.1% along with narrow emission with full width at half maximum of 19 nm (CIE y = 0.15). The high device efficiency is mainly attributed to the high molecular orientation factor, enhanced photoluminescence quantum yield, and a good TADF characteristic of t -Bu- ν -DABNA with efficient Förster energy transfer. lifetime (LT 50 ) of 440 h at an initial luminance of 1000 cd m − 2 . In addition, m MDBA-DI based HF device exhibited significantly high EQE max of 39.1% (Lambertian assumption) along with narrow emission spectra of t -Bu- ν -DABNA through an efficient FRET process. Further, we have analyzed contributed portions on such high EQE based on theoretical optical simulation and photophysical properties of the emis-sive layer. Such high efficiency was dominantly attributed to the following criteria; high horizontal orientation factor, reduced self-quenching by the low doping concentration, and additional singlet exciton recycling process from DET or charge trapping emission on the t -Bu- ν -DABNA emitter. Thus, we believe that our design strategy and device analysis would propose a significant way to attain highly efficient and stable pure blue


Design Strategy and Theoretical Calculations
By considering the previously reported blue TADF materials, symmetrical oxygen bridged boron DBA acceptor was one of the promising acceptor candidates for deep-blue TADF materials due to its weak acceptor strength, high PLQY, and high RISC rate constant. [3,[10][11][12] DBA-DI emitter exhibits high efficiency and long device lifetime, however, the PL emission color is too much red shifted, [11] which makes insufficient energy transfer to pure-blue HF system. Thus, the acceptor was modified by introducing weak electron donating groups on DBA moiety, anticipating that acceptor strength can be reduced while increasing the optical band gap. The newly designed boron acceptors, pMDBA and mMDBA were connected with DI donor to develop new TADF emitters, namely pMDBA-DI and mMDBA-DI, respectively. In HF system, attaching bulky or inert group on final emitters can reduce the undesired Dexter energy transfer (DET). Usually, incorporation of tert-butyl groups on the terminal sites of the molecule can effectively prohibit the DET and may help to increase the PLQY by reducing the aggregation-caused self-quenching. Thus, we attached tertbutyl groups on ν-DABNA core to increase the intermolecular distance with nearby molecules and enhance the PLQY. Eventually, t-Bu-ν-DABNA was designed as novel blue MR type emitter for high color purity in HF system. In order to confirm the design strategy and investigate the photophysical, electronic properties of designed molecules, the density functional theory (DFT) calculation for the optimization of the ground state, and time-dependent DFT (TD-DFT) calculation for the excited state were performed. The calculations were obtained by utilizing the Lee-Yang-Parr correlation function (B3LYP) with 6-31 (G) basis set using Schrödinger 2020-4 program. [13] As illustrated in Figure S1 (Supporting Information), the HOMO and LUMO are well-separated, in which HOMO was mainly distributed over the DI donor, and LUMO was located on the MDBA acceptor, leading to good charge transfer characteristics and small ΔE ST . The calculated HOMO and LUMO energy values were 5.50 and 2.69 eV for pMDBA-DI, 5.49 and 2.65 eV for mMDBA-DI, respectively. As expected, the HOMO energy level was similar due to the same donor moiety (DI). In contrast, the LUMO energy level of pMDBA-DI and mMDBA-DI became shallower compared to 2.76 eV of DBA-DI due to methyl substitution on the boron acceptors. The optical band gap was calculated as 2.81 and 2.84 eV for pMDBA-DI and mMDBA-DI, respectively. Thus, we can expect that both pMDBA-DI and mMDBA-DI will show blue-shifted emission compared to DBA-DI. The calculated singlet and triplet energies are 3.03/2.94 eV for pMDBA-DI, and 2.95/3.09 eV for mMDBA-DI, respectively. The calculated ΔE ST values are 0.09 and 0.14 eV for pMDBA-DI and mMDBA-DI, respectively. These small ΔE ST energy values indicate that pMDBA-DI and mMDBA-DI could reveal good TADF performances. Additionally, the theoretical calculations were performed for t-Bu-ν-DABNA and ν-DABNA and compared each. t-Bu-ν-DABNA exhibits atomically separated frontier molecular orbital (FMO) distributions similar to ν-DABNA, where HOMO distributes mainly on nitrogen atoms, and LUMO locates primarily on boron atoms as shown in Figure S2 (Supporting Information). These alternative distributions clearly indicate that it is an effective MR molecular structure. The calculated energy levels of t-Bu-ν-DABNA and ν-DABNA were almost similar. Thus, we can expect that t-Bu-ν-DABNA also will show similar properties.

Material Synthesis
The synthetic routes of pMDBA-DI, mMDBA-DI, and MR type t-Bu-ν-DABNA is shown in Scheme 1. The new acceptors were synthesized using a two-step process, including cyclization in the presence of n-butyllithium (n-BuLi) and boron tribromide (BBr 3 ). The final TADF compounds were synthesized using palladium-catalyzed Buchwald-Hartwig amination reaction between diindolocarbazole donor (DI) and boron acceptors. And t-Bu-ν-DABNA was synthesized using commercially available compounds in a two-step process, including palladiumcatalyzed C-N coupling and one-pot double borylation. The detailed synthetic procedures are given in the supplementary information. The structures of all the synthesized materials were characterized by 1 H and 13 C NMR and high-resolution mass spectrometry (HRMS) (Figures S18-S33, Supporting Information).

Electrochemical Properties
Further, to understand the synthesized materials' electrochemical properties, cyclic voltammetry (CV) measurement was performed. The HOMO energy level was obtained from the oxidation potential. The LUMO energy level was determined by subtracting the energy level of HOMO and the optical bandgap. The calculated HOMO values from the CV were 5.50 eV for pMDBA-DI and 5.48 eV for mMDBA-DI. The optical band gap (E g ) calculated from the absorption onset was 2.89 and 2.91 eV for pMDBA-DI and mMDBA-DI, respectively. The calculated LUMO values were 2.61 and 2.57 eV for pMDBA-DI and mMDBA-DI, respectively, exhibiting shallower LUMO energy levels than 2.71 eV of DBA-DI. [11] Notably, mMDBA-DI showed a 0.04 eV shallower LUMO energy level than pMDBA-DI expected in the DFT calculation. Since HOMO energy levels were similar among the three materials, optical band gaps were increased compared to DBA-DI. All electrochemical measurement data of TADF materials was well-corresponded with the DFT calculation. In the case of t-Bu-ν-DABNA, the calculated HOMO and LUMO energy levels were 5.54 and 2.94 eV, and the optical band gap was 2.60 eV. According to our electrochemical measurement of ν-DABNA, the bandgap energy level was the same but, the HOMO and LUMO energy level became deeper by 0.03 eV. The corresponding graphs of all cyclic voltammetry (CV) for all materials are depicted in Figure S3 (Supporting Information).

Photophysical Properties
In order to investigate the photophysical properties of synthesized materials, ultraviolet-visible (UV-vis) absorption and photoluminescence (PL) spectra were recorded in toluene solution. Room-temperature photoluminescence (RTPL) spectra were measured at 300 K, and low-temperature photoluminescence (LTPL) spectra were measured at 77 K after 30 ms of a delay from the excitation. In UV-vis absorption, π-π* transitions have occurred from diindolocarbazole (DI) donor moiety around 320 nm in pMDBA-DI and mMDBA-DI. In addition, intramolecular charge transfer (ICT) induced n-π* absorption was appeared at 384 nm for pMDBA-DI and 371 nm for mMDBA-DI, respectively. The PL emission spectra exhibited at 460 and 451 nm for pMDBA-DI and mMDBA-DI, respectively. As expected, both new TADF materials obtained hypsochromicshifted emission compared to 467 nm of DBA-DI. [11] Since the electron-rich electrophilic moiety of methyl can attack electron deficient central boron, the electron-withdrawing ability of boron became weaker, and bandgap can be effectively increased. Significantly, the mMDBA-DI emitter showed a 16 nm blue-shifted emission from DBA-DI. It is attributed that the methyl on the para position of boron can more effectively influence the electron-deficient state of boron, withdrawing ability became much weaker, and bandgap increment is larger. Both materials revealed a positive solvatochromism effect from non-polar n-hexane to polar methylene chloride solvent, and the emission peaks were shifted by 111 nm (pMDBA-DI) and 107 nm (mMDBA-DI), respectively as shown in Figure S4 (Supporting Information). These solvatochromism results indicate that both materials possess strong ICT of donor-acceptor interaction and the weaker acceptor of mMDBA-DI exhibited less solvatochromism than pMDBA-DI. The S 1 and T 1 energy values were determined by calculating the onset point of the RTPL and LTPL spectra, respectively. The S 1 and T 1 values were 2.99/2.92 eV for pMDBA-DI and 3.04/2.92 eV for mMDBA-DI, respectively. The calculated ΔE ST values were 0.07 and 0.12 eV for pMDBA-DI and mMDBA-DI. The ΔE ST values were further analyzed by the Arrhenius plot using the activation energy of RISC and temperature as shown in Figure S5 (Supporting Information). The ΔE ST values evaluated from the slope of temperature dependence of k RISC plot (Arrhenius plot) were found to be 75 meV for pMDBA-DI and 127 meV for mMDBA-DI, respectively. These ΔE ST values are in good agreement with the measured ΔE ST values from RTPL and LTPL. Further, photophysical properties of t-Bu-ν-DABNA were measured. The maximum absorption peak was at 457 nm, and the PL emission peak was noticed at 467 nm with a narrow FWHM of 14 nm in the toluene solution. t-Bu-ν-DABNA exhibited a small solvatochromism effect, shifting only 14 nm from hexane to methylene chloride. These narrow FWHM and small Stoke's shift are attributed to the alternative FMO distribution on t-Buν-DABNA core due to non-bonding characters, which minimizes the vibronic coupling between the S 0 and S 1 states and the vibrational relaxation at the S 1 state. [7] The S 1 and T 1 energy values were calculated to be 2.78 and 2.74 eV, respectively, and the ΔE ST value is 0.04 eV. The corresponding spectra are shown in Figure 1, and the details of the photophysical and electrochemical data of all materials are summarized in Table 1.
In order to confirm the TADF behavior of pMDBA-DI and mMDBA-DI, transient-PL (TRPL) measurements were performed using two different solvents, toluene, and methylene chloride, shown in  This phenomenon is due to the stabilization of the S 1 state in the polar medium. In order to verify the TADF performance in the film state, PLQY and TRPL measurements were recorded in 30% doped pMDBA-DI and mMDBA-DI with dibenzo [b,d] furan-2,8-diylbis(diphenylphosphine oxide) (DBFPO) host.
The PLQY values in DBFPO film were 97.8% and 97.3% for pMDBA-DI and mMDBA-DI emitters, respectively. In addition, the delayed lifetime in DBFPO film still exhibited the same tendency with solution results, which were 1.80 and 2.90 µs for pMDBA-DI and mMDBA-DI, respectively. Furthermore, the rate constants of excited-state processes were calculated to analyze the TADF characteristics of both materials quantitatively by using the reported method. [14] Calculated rate constants are listed in Table S1 (Supporting Information). Among them, the pMDBA-DI exhibited high k RISC of 1.14 × 10 6 s −1 and almost 2 times higher than 6.11 × 10 5 s −1 of mMDBA-DI, which is derived of small ΔE ST values and short delayed lifetime of pMDBA-DI. Although their k RISC values were lower than DBA-DI, [11] these values are still high enough to have high TADF performance. Moreover, TRPL and PLQY of t-Bu-ν-DABNA were investigated and measured prompt exciton lifetime (τ p ) and delayed exciton lifetime (τ d ) in toluene were 5.79 ns and 1.55 µs, respectively, while delayed exciton lifetime (τ d ) in methylene chloride became longer as 1.88 µs. In film state, 5% doping of t-Bu-ν-DABNA in DBFPO host exhibited enhanced PLQY value of 91.9% compared to ν-DABNA, and such enhancement is due to the incorporation of tert-butyl group. In TRPL measurement, t-Bu-ν-DABNA showed short delayed exciton lifetime of 2.93 µs, and k RISC was 2.54 × 10 5 s −1 , which is lower than DBA-DI type TADF materials. Thus, TADF performance is expected to be relatively poor. On the other hand, extremely narrow FWHM and high PLQY of t-Bu-ν-DABNA characteristics would be good as a final emitter in the HF system. All measured film TRPL decay lifetime and calculated rate constants are described in Figure S6 and summarized Table S1 (Supporting Information).

Thermal Properties
The thermal properties of synthesized materials were investigated by using thermo-gravimetric analysis (TGA) and differential scanning calorimetry (DSC). The decomposition temperatures (T d ) of the materials were measured at 5% weight reduction, and the observed values were 447 °C (pMDBA-DI), 470 °C (mMDBA-DI), and 520 °C (t-Bu-ν-DABNA), respectively. The glass transition temperature (T g ) was obtained by using DSC measurement. The T g values are 226 °C (pMDBA-DI), 232°C (mMDBA-DI), and 273 °C (t-Buν-DABNA), respectively. These results suggest that all these materials revealed high thermal stabilities. All corresponding TGA and DSC figures are shown in Figures S7 and S8 (Supporting Information).

Energy Transfer Study
Efficient FRET process is a key parameter to activate an effective HF system. Generally, Förster radius distance (R F ) can be interpreted as the capability of FRET, and the R F can be expressed in Equation (1). [15] 9000 ln 10 128 where R F is critical intermolecular distance between donor and acceptor where energy transfer is occurred 50%. Φ D stands for the PLQY of donor without acceptor, and κ represents the dipole orientation and assumed to 0.845 2/3 , while n is refractive index (1.8) . F D (λ) is the normalized emission spectra of the donor to an area of 1 and ε A (λ) is the molar absorption coefficient of the acceptor. It is obvious that R F mainly depends on the Φ D and spectral overlap between the absorption of acceptor and PL emission of donor ( ( ) ( ) Since both materials exhibited similar PLQY, the extent of spectral overlap mainly determines the R F . Thus, the extent of spectral overlap between pMDBA-DI and mMDBA-DI with t-Bu-ν-DABNA in the DBFPO film state were measured. As a result, absorption spectrum of t-Bu-ν-DABNA was largely overlapped with PL emission spectra of both pMDBA-DI and mMDBA-DI compared to DBA-DI as illustrated in Figure S9   UV-vis absorption wavelengths measured in 10 −5 m of toluene solution; b) PL maximum emission measured in toluene at 300 K; c) Calculated by the onset of RTPL spectrum in toluene; d) Calculated by the onset of low temperature PL (LTPL) spectrum in toluene at 77 K with 30 ms of delaying; e) ΔE ST = S 1 − T 1 ; f) Prompt decay lifetime in toluene; g) Delayed decay lifetime in toluene; h) Delayed decay lifetime in methylene chloride; i) HOMO level energy measured by CV method; j) LUMO = HOMO-optical band gap; k) Calculated by the edge of the absorption spectrum; l) Measured in DBFPO film. pMDBA-DI and mMDBA-DI were 30% doped and t-Bu-ν-DABNA was 5% doped. film were investigated. Additional 1% of t-Bu-ν-DABNA doping, the prompt and delayed decay lifetime was shortened, reducing the delayed portion as illustrated in Figure S10 (Supporting Information). The prompt decay time of pMDBA-DI with 1% of t-Bu-ν-DABNA was decreased from 45.6 to 15.8 ns, and mMDBA-DI was reduced from 47.0 to 10.1 ns. This means that the faster FRET process from TAH to t-Bu-ν-DABNA accelerated the overall prompt energy transfer process, indicating an efficient FRET process. The exact rate constants of FRET in the HF system were calculated by using Equation (2). [15,16] FRET PF r,S I SC where k r,S and k ISC are singlet radiative rate constant, and intersystem crossing rate constant, respectively. The calculated k FRET of 30% mMDBA-DI and pMDBA-DI with 1% of t-Bu-ν-DABNA was 7.77 × 10 7 s −1 and 4.14 × 10 7 s −1 , respectively. These k FRET rates are much faster than calculated , r s k k ISC values as provided in Table S1 (Supporting Information), which is an experimental evidence that FRET is enhanced in this HF system Unlike FRET, the DET from TAH to t-Bu-ν-DABNA in HF system should be minimized. Since the k RISC of t-Bu-ν-DABNA is insufficient and lower than TAH materials, triplet exciton loss would be increased when the DET occurs a lot. The rate constant of DET (k DET ) can be calculated by using Equation (3). [ where k nr,T represents the triplet non-radiative rate constant. The calculated k DET of the same HF system was 2.23 × 10 5 s −1 and 2.01 × 10 5 s −1 , respectively, which were limited to 10 5 s −1 .
Due to the 1% of small doping concentration and large inert bulky phenyl group in terminal site of t-Bu-ν-DABNA, k DET can be limited. However, the k DET of pMDBA-DI was higher, which may be attributed of slower k FRET and smaller spectral overlap. The measured TRPL decay lifetime, constant numbers and calculated FRET and DET rate constants are listed in Table S2 (Supporting Information).

Device Performances
To understand the TADF characteristics of synthesized materials, TADF-OLEDs with synthesized materials were fabricated using previously reported device configuration; [11] ITO (50 nm)/ HATCN (7 nm (TAPC)  and 1,3,5-tris(1-phenyl-1H-benzo [d]imidazole-2-yl) benzene (TPBi) served as hole-and electron-transporting layers, respectively. 3,5-Di(9H-carbazol-9-yl)-N,N-diphenylaniline (DCDPA) and DBFPO were utilized as exciton blocking layers on the each sides of the EML owing to their higher tri-plet energy levels. The DBFPO also served as a host material. The maximum electroluminescence (EL) peak at 1000 cd m −2 of pMDBA-DI was 483 nm, whereas mMDBA-DI exhibited 474 nm, and mMDBA-DI showed 9 nm blue-shifted emission than pMDBA-DI. Measured CIE coordinates were (0.15, 0.31) and (0.14, 0.23), respectively. TADF devices of pMDBA-DI and mMDBA-DI exhibited high EQE max of 33.1% and 32.8%, and these EQEs of both devices were maintained at 32.3% and 28.4% (Lambertian assumption) at 1000 cd m −2 , respectively. Interestingly, mMDBA-DI device exhibited the higher EQE value in the very initial luminance region, which may be attributed to the better charge balance. However, as the charge balance of pMDBA-DI device also gradually matched, pMDBA-DI exhibited the EQE max value near 1000 cd m −2 . Moreover, mMDBA-DI device showed severe efficiency roll-off at higher luminance region compared with those of pMDBA-DI, which can be attributed to larger ΔE ST and two times lower k RISC of mMDBA-DI. Long living triplet excitons from lower k RISC can remain in the EML longer, becoming hot excitons, occurring triplet-triplet annihilation (TTA) or triplet-polaron annihilation (TPA), which are the major reasons of material degradation. In addition, as the amount of exciton generation is getting larger in higher luminance region, their activities can be accelerated. Both TADF materials exhibited surprisingly high EQE values, which cannot be fully explained by only high PLQY and good TADF behavior. Since mMDBA-DI exhibited more blue-shifted emission, we measured the molecular orientation factor (θ) of mMDBA-DI with the previously reported method. [18] As expected, high θ of 0.82 value was obtained as illustrated in Figure S11a (Supporting Information). The additional device performances depending on the TADF concentration is provided in Figure S12 and Table S3 (Supporting Information). Additionally, the TADF performance of t-Bu-ν-DABNA was also investigated. The maximum electroluminescence (EL) peak at 1000 cd m −2 was 474 nm, and FWHM was narrow as 18 nm, following CIE y coordinate was 0.15. The measured EQE max value of t-Bu-ν-DABNA in TADF device was high as 36.3% (Lambertian assumption) even though PLQY and TADF performances fall behind pMDBA-DI or mMDBA-DI. To understand such high efficiency, the orientation factor of t-Buν-DABNA was measured and found to be the high value as 0.92, as shown in Figure S11b (Supporting Information). Such a higher value is interpreted due to the rigid and long planar structure of t-Bu-ν-DABNA. In addition, tert-butyl groups in the planar structure of ν-DABNA result in higher PLQY by preventing aggregation-induced self-quenching. Owing to these two reasons, the high TADF device efficiency of t-Bu-ν-DABNA could be observed. However, despite such high efficiency, severe efficiency roll-off of this t-Bu-ν-DABNA existed because of low k RISC (EQE value dropped to 16.5% at 1000 cd m −2 ). Detailed device performances of ν-DABNA and t-Bu-ν-DABNA are illustrated in Figure S13 and Table S4 (Supporting Information). From the TADF device performances, pMDBA-DI and mMDBA-DI based devices revealed excellent TADF efficiencies, and t-Buν-DABNA manifested a narrow FWHM. From these results, our newly synthesized three materials would be ideal for good blue HF devices as both pMDBA-DI and mMDBA-DI have high k FRET of energy transfer. Thus, HF-OLEDs were fabricated with 30% TAH and 1% of t-Bu-ν-DABNA emitter in DBFPO host. All the other layers were used without any change of TADF device www.afm-journal.de www.advancedsciencenews.com 2110356 (7 of 10) © 2021 Wiley-VCH GmbH configuration. Since the HOMO energy levels of three materials are similar, holes are prone to transfer via TAH materials, while electrons can move into t-Bu-ν-DABNA. However, as shown in Figure 2b, the current density curves of TADF and HF device were similar, indicating of almost no charge trapping emission in HF device. HF devices with pMDBA-DI and mMDBA-DI TADF hosts exhibited 474 nm of maximum EL spectrum with narrow FWHM of 19 nm. Notably, the mMDBA-DI HF device exhibited a clear t-Bu-ν-DABNA EL spectrum with minor emission of mMDBA-DI, as shown in Figure 2d. As expected from the energy transfer study, a large spectral overlapping and fast energy transfer rate of mMDBA-DI is crucial to getting the real final emitter spectrum in HF-OLED. Obtained CIE coordinates of pMDBA-DI based HF device was of (0.13, 0.19), while those of mMDBA-DI was (0.12, 0.15). Different CIE coordinates were derived from the different emission intensities and wavelengths of TAH. Surprisingly, pMDBA-DI and mMDBA-DI based HF devices showed much enhanced EQE max of 37.6% and 40.7% than conventional TADF devices. At the same time, we noticed that pMDBA-DI and mMDBA-DI based HF devices remained high EQE values of 33.9% and 35.8% at 1000 cd m −2 , respectively. The corresponding plot of current density (J), and luminance (L) versus voltage (V), and EQE versus luminance, electroluminescence (EL) spectra of each TADF and HF device is depicted in Figure 2. Since both TAHs and t-Bu-ν-DABNA have high horizontal orientation factor, the possibility of sur-face plasmon coupling with cathode would be reduced, light out coupling efficiency (n out ) can be intensified. [19] Thus, the light emission intensity coming from front is reinforced, and the angular distribution of emission became non-Lambertian shape. Therefore, it results in an error in the calculated EQE values from Lambertian assumption measurement method. Thereby, the angular distributions of the devices were measured as described in Figure S14 (Supporting Information), and the EQE values were recalculated by integrating all angular-dependent emissions. From this calculation, corrected EQE max values for pMDBA-DI and mMDBA-DI HF-OLEDs were 36.2% and 39.1%, respectively. The detailed device performances are described in Table 2.

Analysis of High Device Efficiency
The pMDBA-DI and mMDBA-DI based HF devices obtained much higher EQE max values than individual TADF devices. Especially, mMDBA-DI achieved 39.1% of maximum EQE value with pure blue emission color of t-Bu-ν-DABNA. To understand the root origins of such efficiency enhancement in HF devices, we analyzed EQE-related factors. Generally, obtained EQE can be expressed by Equation (4). [20]   where γ is the charge balance factor, η S/T is the singlet-triplet factor, q eff is the effective quantum yield, and η out is the outcoupling efficiency of the emitted light. The η out is not only influenced by the device structure, but also by the horizontal orientation factor of the emitter. High out-coupling efficiency can effectively help improving device efficiency compared to isotropic emitters (η out 0.2-0.3). Considering the high orientation value of 0.82 of mMDBA-DI, the η out value was calculated to be 0.350 in this device by the optical simulation (using Setfos 4.3 program). Since the orientation value of t-Bu-ν-DABNA was measured to be 0.92, η out value could be increased, and the calculated possible device efficiency was 37.8% as shown in Figure 3. Thus, 6% of the EQE value was enhanced by the increased orientation factor of t-Bu-ν-DABNA. However, such molecular orientation effect cannot be fully explained by the 39.1% of high efficiency. So, the low doping concentration merit of the HF device was considered. When the device doping percentage decreased from 5% to 1%, PLQY tends to increase. Our measured PLQY was increased from 91.9% to 93.1% due to a reduction of self-quenching. This PLQY increase could enhance our device EQE additionally about 0.5% in our HF system. Although all possible ways of EQE enhancement were considered, our device EQE is still high about 0.8%. Although this value can be considered as within in the margin of error, additional singlet exciton generation on t-Bu-ν-DABNA was taken into consideration since t-Bu-ν-DABNA has TADF characteristics and can harvest excessive singlet excitons via RISC of triplet excitons formed by direct charge trapping or DET from the host. When the TADF material is used as an assistant host, the total internal EL efficiency (φ EL(int) ) can be expressed by Equation (5). [21] 1 EL int 0 1 where η 0 (S 1 ) is the singlet exciton ratio, and η 0 (T 1 ) is the triplet exciton ratio. φ ISC,H is the ISC efficiency of the TAH material, and φ PL,D stands for the PL quantum efficiency of the emitter. φ FRET is the FRET efficiency from TAH to the final fluorescent emitter. Due to the large spectral overlap, high φ FRET was possible, t-Bu-ν-DABNA can have large amount of mMDBA-DI recycled singlet excitons. Simultaneously, t-Bu-ν-DABNA can harvest additional singlet exciton through triplet exciton recycling via RISC process from charge trapping or DET. Considering both energy transfer and triplet exciton recycling by t-Bu-ν-DABNA, internal quantum efficiency (φ EL(int) ) of t-Bu-ν-DABNA HF system can be described by Equation (6).
where η 0 (T 1 ) DET and η 0 (T 1 ) TRAP represent the triplet exciton occurred from DET and charge trapping, respectively. exciton fraction from t-Bu-ν-DABNA via DET or charge trapping. In our system, the remaining 0.8% of EQE enhancement may be originated from these two effects. The trace of HF device efficiency increment is illustrated in Figure 3. On the other hand, pMDBA-DI TADF exhibited the less effective device efficiency improvement in the HF systems. Due to less spectrum overlap, k FRET was relatively slower, the improvement of η out value was limited. Hence, singlet emission from direct TAH was enlarged in total HF EL spectrum.

Device Lifetime
In order to investigate the operational lifetime effect in the devices by methyl substitutions in DBA core, additional TADF and HF devices, including our reported TDBA-DI and DBA-DI materials, [10,11] were fabricated. Since TAPC, DBFPO, and DCDPA materials possess weak BDEs, the operational lifetimes of previously fabricated devices were very poor. Therefore, a new device configuration with stable materials was designed, as shown in Figure S15 (Supporting Information). Unlike TDBA acceptor, the bond dissociation energy (BDE) of methyl CC bonding in MDBA acceptor is higher than 3.0 eV in all neutral, anion, cation states ( Figure S16, Supporting Information), and improved device lifetime was expected in both TADF and HF devices. In addition, MDBA-DIs exhibited the hypsochromic shifted emission than DBA-DI, enhanced color purity in HF device was obtained. Since the triplet energy level of 2,6-di(9H-carbazol-9-yl)pyridine (DCzPy) host is similar to TADF emitter, [22] some triplet excitons could be quenched, resulting in decreased device efficiency. Even though some efficiency dropping, fabricated TADF devices with 30% doping of pMDBA-DI and mMDBA-DI exhibited 23.0% and 21.3% of EQE max values, respectively. The TADF device lifetime (LT 50 ) at constant current operation at the 1000 cd m −2 initial luminance was 332 and 35 h, respectively. Due to the large ΔE ST and longer k RISC , the triplet excitons density in mMDBA-DI TADF device can be higher in EML, undesired hot excitons, and destructive TTA, TPA process can easily be generated, revealing the shorter device lifetime. To confirm the similar effect in HF devices, both HF devices with 1% doping of t-Bu-ν-DABNA were fabricated. The pMDBA-DI and mMDBA-DI devices exhibited the increased device efficiency of 26.6% and 28.1% with CIE y of 0.20 and 0.16, respectively. Measured device lifetimes of both HF devices at the same condition were 440 and 133 h, respectively. Although the absolute device lifetime of the mMDBA-DI HF device was shorter than pMDBA-DI, the extent of improvement was more remarkable, which may be due to the help of large spectral overlap and efficient FRET. Importantly, pMDBA-DI based TADF device achieved the 1.8 times improved device lifetime than TDBA-DI based device due to the enhanced BDE of MDBA acceptor. Similarly, the mMDBA-DI HF device also resulted in 1.6 times increased device operational lifetime, as shown in Table S6 (Supporting Information). However, both HF devices' operational lifetime was limited than DBA-DI based HF device (573 h), CIE y of DBA-DI HF device has deteriorated to 0.23. All the device performances and device lifetime curves are illustrated in Figure S17 and Table S6 (Supporting Information). Introducing the methyl groups on oxygen para position to DBA acceptor, the pMDBA-DI HF device could have good color purity of CIE y = 0.20 and an elongated lifetime of 440 h.

Conclusion
In summary, two new highly efficient blue TADF materials and narrow spectral width t-Bu-ν-DABNA emitter were designed and synthesized for efficient and pure blue HF-OLEDs. Synthesized pMDBA-DI and mMDBA-DI by the methyl electrophilic substitution effect exhibited hypsochromic shifted emission, leading to good spectral overlapping with the t-Bu-ν-DABNA emitter. The t-Bu-ν-DABNA demonstrated high PLQY and narrow FWHM as a good candidate emitter for efficient HF-OLEDs. The HF-OLED device with pMDBA-DI resulted in an exemplary device lifetime (LT 50 ) of 440 h at an initial luminance of 1000 cd m −2 . In addition, mMDBA-DI based HF device exhibited significantly high EQE max of 39.1% (Lambertian assumption) along with narrow emission spectra of t-Bu-ν-DABNA through an efficient FRET process. Further, we have analyzed contributed portions on such high EQE based on theoretical optical simulation and photophysical properties of the emissive layer. Such high efficiency was dominantly attributed to the following criteria; high horizontal orientation factor, reduced self-quenching by the low doping concentration, and additional singlet exciton recycling process from DET or charge trapping emission on the t-Bu-ν-DABNA emitter. Thus, we believe that our design strategy and device analysis would propose a significant way to attain highly efficient and stable pure blue HF-OLEDs.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.