Donor Extended Blue TADF Dendrimer for High‐Performance Solution‐Processed OLEDs Authors: Mahni Fatahi, Yuka Yasuda, Ryo Kondo, Hironori Kaji, Eli Zysman‐Colman Published: 2026 ABSTRACT ABSTRACTAchieving efficient and stable blue solution‐processed OLEDs remains an outstanding challenge in the field. We introduce a rational donor decoration strategy and apply it to the TADF emitter DOBNA‐SpAc (aka. TDBA‐SAF), exemplified in the emitter DOBNA‐SpAc‐DCz. By introducing ter(tert‐butylcarbazole) units at the 2 and 7 positions of the acridine moiety, solubility and hole‐transport properties are improved without compromising the blue emission endemic to DOBNA‐SpAc. This emitter has a high photoluminescence quantum yield, Φ PL, of 93% in 20 wt.% doped films in PPF (2,8‐bis(diphenyl‐phosphoryl)‐dibenzo[b,d]furan), a very small singlet‐triplet energy gap (ΔE ST = 0.01 eV), and thus fast reverse intersystem crossing (k RISC > 1 × 106 s−1), resulting in a short delayed lifetime of 2 µs. Solution‐processed OLEDs with DOBNA‐SpAc‐DCz reached a maximum external quantum efficiency, EQE max, of 29.4 ± 0.1% at CIE coordinates of (0.145, 0.211). By probing different electron transport materials and comparing to devices using DOBNA‐SpAc, we found that the introduction of carbazole substituents promotes improved hole transport and a more spatially distributed recombination zone, while the faster k RISC suppresses triplet‐related annihilation processes. These results demonstrate that the targeted peripheral donor dendron decoration of spiroacridine‐based TADF emitters is an effective strategy to achieve highly efficient emitters suitable for solution‐processed OLED applications. Introduction Solution processing (SP) promises an attractive route toward cost‐efficient large‐area fabrication of organic light‐emitting diodes (OLEDs). High‐throughput techniques such as slot‐die coating and inkjet printing that are compatible with roll‐to‐roll manufacturing have been used to fabricate these devices [1, 2, 3]. Compared to vacuum deposition, SP uses less energy, and there is not as much material waste [2]. Yet, fabricating SP‐OLEDs that show comparable performance to vacuum‐deposited devices remains a challenge. What is required are high‐performance emitters optimized for solution‐processed devices, where solubility and film‐forming properties must be considered along with optoelectronic demands [2, 3]. Organic thermally activated delayed fluorescence (TADF) emitters are metal‐free and can harvest both singlet and triplet excitons, the latter via reverse intersystem crossing (RISC), enabling TADF OLEDs to reach comparably high internal quantum efficiencies as state‐of‐the‐art phosphorescent devices [4]. For high device performance, TADF emitters must have a fast RISC rate constant (k RISC) and a high photoluminescence quantum yield (Φ PL) [5]. Achieving these photophysical requirements while maintaining robust solution processability is non‐trivial, as molecular designs that minimize the singlet‐triplet energy gap (∆E ST) often compromise morphological stability or solubility as well as oscillator strength that adversely affects Φ PL [6, 7]. Bridging this gap between excited‐state engineering and practical processability remains a central challenge for high‐efficiency SP‐OLEDs. Among the different emission colors, optimizing the blue pixel is particularly challenging, particularly in terms of the required color point and the stability of the device [8]. Notably, several blue and deep‐blue TADF vacuum‐deposited OLEDs already demonstrate high efficiencies as a result of high Φ PL, small ∆E ST, and fast k RISC of the emitter. Of relevance to the current study, in 2020 Lim et al. reported TDBA‐SAF [9] (here referred to as DOBNA‐SpAc; Figure 1), which resembles a spiroacridine‐decorated derivative of the UV‐emitting compound 2a [10], first reported by Hatakeyama et al. (later also referred to as DOBNA [11] or BOO [12]). By decorating the DOBNA core with the spiroacridine donor, the emission is red‐shifted to the deep blue, peaking at a λ PL of 455 nm, and the Φ PL of 90% remains high in 20 wt.% doped films in DPEPO (bis[2‐(diphenylphosphino)phenyl]ether oxide). A short delayed lifetime, τ d, of 1.34 µs means that k RISC is fast at 2.10 × 106 s−1. Vacuum‐deposited OLED reached a maximal external quantum efficiency (EQE max) of 28.2% at Commission Internationale de l’Éclairage (CIE) coordinates of (0.142, 0.090). Building on this promising work, Lee et al. designed a spiroacridine donor that contains a second spiro carbon‐linked fluorene group in an attempt to increase the horizontal alignment of the emitter and so enhance the light outcoupling in the device. Similar to DOBNA‐SpAc, tBuOBOtSAc [13] (Figure 1) emits in the deep blue at λ PL of 450 nm with a higher Φ PL of 97% in 10 wt.% doped films in DPEPO. However, this structural modification coincides with a longer τ d, to 3.81 µs, which contributes to the slower k RISC of 0.48 × 106 s−1 compared to DOBNA‐SpAc. Vacuum‐deposited OLEDs with tBuOBOtSAc showed a comparable EQE max of 28.2% [CIE coordinates of (0.149, 0.061)] to those with TDBA‐SAF; however, they suffered from much stronger efficiency roll‐off. Chemical structures and key photophysical properties of DOBNA‐spiroacridine‐based TADF emitters and structure and key photophysical properties of DOBNA‐SpAc (aka TDBA‐SAF) and DOBNA‐SpAc‐DCz. Employing a different decoration strategy of the spiroacridine donor of DOBNA‐SpAc, Pei et al. introduced three derivatives, MH, PH, and PP (Figure 1) [14]. These three compounds contain methyl and/or phenyl groups at the 2, 3’, 6’, and 7 positions of the spiroacridine donor. The authors reported an apparent negative ∆E ST value for the 10 wt.% doped film MH in PPF, likely due to the measurements being conducted at 77 K, which is not low enough to completely suppress the delayed fluorescence in a compound that is expected to have a vanishingly small ΔE ST. The authors, however, found that when both the acridine and the fluorene moieties are decorated with phenyl groups as in PP, the ∆E ST is much larger at 200 meV. This trend coincides with that for τ d, where MH has the shortest delayed lifetime, followed by MP, whereas PP does not show any delayed emission in 10 wt.% doped films in PPF (2,8‐bis(diphenyl‐phosphoryl)‐dibenzo[b,d]furan), ostensibly due to non‐radiative decay outcompeting RISC. This study reveals that substitution at the fluorene moiety of the spiroacridine donor has an undesired impact on the photophysical properties of the emitter. The findings of these latter two studies reveal that only substitutions at the 2 and 7 positions of the spiroacridine of DOBNA‐SpAc produce a desirable change in the photophysical properties, such as an increase in Φ PL and accelerated k RISC. On the contrary, substitution on the fluorene moiety leads to an increase in τ d and subsequently a slower k RISC, rendering this substitution pattern unpromising in the light of device properties. Inspired by these prior reports, we designed the blue‐emitting solution‐processable TADF emitter DOBNA‐SpAc‐ DCz (Figure 1) in which two ter(tert‐butylcarbazoles) decorate the spiroacridine donor at the 2 and 7 positions of the acridine. The motivation for the inclusion of these donor dendron units was to simultaneously decrease the ∆E ST, improve the solubility and thus film‐forming properties, and enhance the hole transport as a counterbalance to the use of n‐type phosphine oxide‐based hosts. DOBNA‐SpAc‐DCz emits at 465 and 474 nm in toluene and 20 wt.% doped films in PPF, respectively, and has high Φ PL of 79 and 93%, respectively. In both solution and the film state, the τ d are short at around 2 µs, leading to fast k RISC of 2.70 and 1.71 × 106 s−1, respectively. These promising photophysical properties prompted the fabrication of solution‐processed OLEDs. The SP‐OLED emitted at 474 nm, corresponding to CIE coordinates of (0.145, 0.211), and showed an EQE max of 29.4 ± 0.1%. In further device studies, using an altered device stack, a correlation between the increase in k RISC and enhanced hole mobility of DOBNA‐SpAc‐DCz compared to DOBNA‐SpAc and reduced efficiency roll‐off was found. Results and Discussion Theoretical Calculations To understand the interplay between the different fragments within the emitter design, we first performed calculations at a PBE0/6‐31G(d,p) level of theory in the gas phase. The HOMO/LUMO energy levels for DOBNA‐SpAc‐DCz are −5.09/−1.97 eV, and the orbital density plots show a clear separation of the HOMO, being located on the carbazole‐decorated spiroacridan moiety, and the LUMO, being located on the DOBNA unit (Figure 2). The HOMO of DOBNA‐SpAc‐DCz is destabilized compared to that of DOBNA‐SpAc (also known as TDBA‐SAF) at −5.21 eV, which is attributed to the extended π‐conjugated donor system. The LUMO is stabilized compared to that of DOBNA‐SpAc (−1.73 eV). A comparison of the calculated HOMO/LUMO levels to those of the structurally related compound MH (calculated at B3LYP/def2svp level of theory; Figure 1) reveals almost identical values for the LUMO energy (−1.99 eV) [14] and a stabilized HOMO for DOBNA‐SpAc‐DCz (MH HOMO: −4.98 eV) [14]. Calculated HOMO and LUMO energies and corresponding orbital plots (isovalue = 0.02), excited‐state energy levels and corresponding natural transition orbital plots and spin‐orbit coupling constants calculated on a PBE0/6‐31G(d,p) level of theory for (a) DOBNA‐SpAc and (b) DOBNA‐SpAc‐DCz. The excited‐state energies were calculated using TDA‐DFT at the same level of theory as the ground‐state optimized geometry. The energies of the S1 and the T1 states are predicted to be 2.75 and 2.74 eV, resulting in a ∆E ST of 0.01 eV. The natural transition orbital visualizations indicate that the S1 state has long‐range charge transfer (LRCT) character between carbazole‐decorated spiroacridan, acting as the donor, and the DOBNA unit, acting as the acceptor. A very similar picture exists for the T1 state. Compared to DOBNA‐SpAc (S1/T1 = 2.84/2.83 eV), the S1 and T1 states are stabilized; however, a ∆E ST of 0.01 eV remains the same for both emitters. The spin‐orbit coupling (SOC) between the T1 and S1 states is twofold stronger in DOBNA‐SpAc‐DCz (0.0095 cm−1) than in DOBNA‐SpAc (0.0046 cm−1), yet still weak. Notably, the root mean square difference for the geometry between the ground state (S0) and the optimised S1 state in both molecules remains low at 0.32 and 0.49 Å for DOBNA‐SpAc and DOBNA‐SpAc‐DCz, respectively, implying that the PL spectra of both should be equally narrowband (Figure S1). The ground‐state optimized conformation for both compounds reveals an expectedly large dihedral angle between the DOBNA and the spiroacridine unit, minimizing orbital overlap between the donor and acceptor moieties that explains the very small computed ∆E ST (Figure S1). Synthesis The donor and acceptor moieties of DOBNA‐SpAc‐DCz were synthesised separately prior to being coupled together. The t DOBNA‐Br unit was obtained via a two‐step procedure following a previously reported protocol from the literature [9]. The donor moiety was synthesized in two steps from 10H‐Spiro[acridine‐9,9'‐fluorene], a dibromination at the 2‐ and 7‐positions of the acridine using N‐bromosuccinimide proceeded readily in 63% yield, followed by an Ullmann coupling with two equivalents of ter‐tert‐butylcarbazole to form the extended donor moiety SpAc‐DCz in 69% yield. DOBNA‐SpAc‐DCz was obtained in 23% yield following a Buchwald‐Hartwig type reaction between the donor and acceptor fragments (Scheme 1). The structure and purity were confirmed by 1H NMR and 13C NMR spectroscopy, high‐resolution mass spectrometry, melting point determination, and elemental analysis (Figures S14–S24). The high thermal stability was confirmed by thermal gravimetric analysis, which showed a 5% mass loss at a temperature T d, 5% of 520°C (Figure S8), higher than that for DOBNA‐SpAc (T d, 5%: 400°C) [9], while the melting point for DOBNA‐SpAc‐DCz exceeded 400°C. Synthetic route for DOBNA‐SpAc‐DCz. Photophysical Properties The absorption and steady‐state photoluminescence (SS PL) spectra in dilute toluene solution of DOBNA‐SpAc‐DCz and DOBNA‐SpAc are shown in Figure 3a. The dominant low‐energy feature in the absorption spectra is a charge‐transfer band at 380 nm, with a molar extinction coefficient, ε, of 25 × 103 m −1 cm−1 for the former, which effectively has about the same intensity as that in DOBNA‐SpAc (Table 1). The emission maximum of DOBNA‐SpAc‐DCz, λ PL, of 465 nm (FWHM of 56 nm) is slightly red‐shifted to that of DOBNA‐SpAc (λ PL of 450 nm), in line with the computed trend. PL solvatochromism studies reveal that the PL red‐shifts and broadens, reflective of an emissive excited state of long‐range charge transfer (LRCT) character (Figure S4 and Table S2). The Φ PL for DOBNA‐SpAc‐DCz is 79% in degassed toluene, which decreases significantly to 18% under aerated conditions, strongly indicating an involvement of triplet states in the emission mechanism. Time‐resolved PL measurements reveal a rather long prompt lifetime, τ p, of 14.6 ns and a prompt photoluminescence quantum yield, Φ prompt, of 12%, and a rather short delayed lifetime, τ d, of 2.47 µs and a delayed photoluminescence quantum yield, Φ del, of 67% (Figure S5). Excitonic rate constants were extracted from these measurements and are compiled in Table 1. The intersystem and reverse intersystem rate constants, k ISC and k RISC, respectively, are fast at 5.59 × 107 and 2.70 × 106 s−1. (a) Absorption (dashed) and normalized SS PL (solid) spectra of dilute toluene solutions of DOBNA‐SpAc‐DCz (blue; λ exc = 350 nm) and DOBNA‐SpAc (red; λ exc = 380 nm) (all 10−5 m); (b) Normalized SS PL spectra of 20 wt.% doped films of DOBNA‐SpAc‐DCz (blue) and DOBNA‐SpAc (red) in 2,8‐bis(diphenyl‐phosphoryl)‐dibenzo[b,d]furan (PPF); (λ exc = 310 nm) (c) Time‐resolved PL (TRPL) decays of 20 wt.% doped films of DOBNA‐SpAc‐DCz (blue) and DOBNA‐SpAc (red) in PPF; Inset: TRPL decays showing the prompt PL; (d) Temperature‐dependent TRPL of a 20 wt.% doped film of DOBNA‐SpAc‐DCz in PPF. TABLE 1Photophysical properties of dilute solutions of DOBNA‐SpAc‐DCz and 20 wt.% doped films of DOBNA‐SpAc‐DCz and DOBNA‐SpAc in PPF.CompoundMedium λ Abs c /nm λ PL d (FWHM) e /nm E S1 f ,g/eV E T1 f ,g/eV ∆E ST h /eV Φ PL i /% τ p j /ns τ d j /µs k ISC k /× 107s−1 k RISC k /× 106s−1 k s_r k /× 106s−1 k s_nr k /× 106s−1 DOBNA‐SpAc‐DCz Sol. a 380 (25)465(56)2.932.920.0179/1814.62.475.592.708.122.16Film b —474(73)2.902.90<0.0193/6631.42.042.381.716.220.47 DOBNA‐SpAc Sol. a 382 (21)450(50)3.022.930.09———————Film b —467(60)2.912.91<0.0193/7727.93.072.371.0210.70.80 a In toluene solutions (10−5 m). b Measured in spin‐coated thin films consisting of 20 wt.% emitter in PPF host. c Lowest energy absorbance peak, Absorptivity (ε) in parentheses (/103 m −1 cm−1). d Steady‐state emission maximum at 300 K. λ exc = 350 nm (sol, DOBNA‐SpAc‐DCz), λ exc = 380 nm (sol, DOBNA‐SpAc), λ exc = 310 nm (film). e Full width at half maximum of emission peak. f S1 and T1 energies were obtained from the onsets of the respective steady state PL and phosphorescence spectra (delay: 1 ms (film)/10 ms (sol); gate time: 8.5 ms (film)/85 ms (sol)) at 77 K. λ exc = 350 nm (sol, DOBNA‐SpAc‐DCz), λ exc = 365 nm (sol, DOBNA‐SpAc), λ exc = 310 nm (film). g Solution samples for ΔE ST measurements were prepared in 2‐MeTHF (10−6 m). h ΔE ST = E(S1)−E(T1). i Relative Φ PL in solutions was measured by a comparative method using quinine sulfate as a standard (Φr = 54.6% in 1 N H2SO4) [17], while absolute Φ PL of the thin films was measured using an integrating sphere. j Prompt and delayed lifetimes were obtained by TCSPC and MCS, respectively. λ exc = 375 nm. k Intersystem and reverse intersystem crossing rate constants were calculated using the steady‐state approximation method as described in the literature [18].John Wiley & Sons, Ltd. Photophysical properties of dilute solutions of DOBNA‐SpAc‐DCz and 20 wt.% doped films of DOBNA‐SpAc‐DCz and DOBNA‐SpAc in PPF. In toluene solutions (10−5 m). Measured in spin‐coated thin films consisting of 20 wt.% emitter in PPF host. Lowest energy absorbance peak, Absorptivity (ε) in parentheses (/103 m −1 cm−1). Steady‐state emission maximum at 300 K. λ exc = 350 nm (sol, DOBNA‐SpAc‐DCz), λ exc = 380 nm (sol, DOBNA‐SpAc), λ exc = 310 nm (film). Full width at half maximum of emission peak. S1 and T1 energies were obtained from the onsets of the respective steady state PL and phosphorescence spectra (delay: 1 ms (film)/10 ms (sol); gate time: 8.5 ms (film)/85 ms (sol)) at 77 K. λ exc = 350 nm (sol, DOBNA‐SpAc‐DCz), λ exc = 365 nm (sol, DOBNA‐SpAc), λ exc = 310 nm (film). Solution samples for ΔE ST measurements were prepared in 2‐MeTHF (10−6 m). ΔE ST = E(S1)−E(T1). Relative Φ PL in solutions was measured by a comparative method using quinine sulfate as a standard (Φr = 54.6% in 1 N H2SO4) [17], while absolute Φ PL of the thin films was measured using an integrating sphere. Prompt and delayed lifetimes were obtained by TCSPC and MCS, respectively. λ exc = 375 nm. Intersystem and reverse intersystem crossing rate constants were calculated using the steady‐state approximation method as described in the literature [18]. The S1 and T1 energies were determined from the onsets of the SS PL and gated emission in 2‐methyltetrahydrofuran (2‐MeTHF) glass at 77 K (Figure S9) and are 2.93 and 2.92 eV, respectively; the corresponding ∆E ST is 10 meV, matching the calculated value. By contrast, the S1 energy of DOBNA‐SpAc (Figure S9) is higher at 3.02 eV, while the T1 energy is almost unchanged at 2.93 eV, meaning that the ∆E ST for this compound is larger at 90 meV. Based on the promising photophysical properties in solution, we next investigated the photophysical properties of spin‐coated thin films. Here, we discuss the photophysical properties of 20 wt.% doped films of DOBNA‐SpAc‐DCz in a PPF matrix since this host‐guest combination has the highest Φ PL (93% under nitrogen / 66% in air) of the tested OLED‐relevant hosts and doping concentrations (Figure S7 and Table S3). These films emit at λ PL of 474 nm (FWHM of 73 nm). Compared to toluene, the PPF host provides a more polar local environment, which stabilizes the emissive excited state more strongly than the ground state. This stabilizes the S1 energy, which produces a red shift of the PL spectrum. The distribution of local microenvironments in the solid host (site‐to‐site variations in polarity and local electric fields) leads to inhomogeneous broadening, resulting in a wider emission band [15, 16]. In the film, the τ p increased by a factor of two to 31.4 ns (Φ prompt = 20%) compared to the solution‐state measurements (τ p = 14.6 ns), while the τ d remained unaffected (2.04 µs, Φ del = 73%). The S1 and T1 energy levels, measured from the onsets of the steady‐state and time‐gated emission spectra at 77 K, are both 2.92 eV, resulting in an ∆E ST that was too small to determine using our measurement setup. The negligible ∆E ST is in good agreement with the solution‐state measurements as well as with the values reported for tBuOBOtSAc (0.08 eV, 10 wt.% in DPEPO) [13] and MP (0.02 eV, 10 wt.% in PPF) [14]. The excitonic rate constants for these films are similar to those for the solution‐state measurements; however, the non‐radiative decay rate constant from the singlet state, k s_nr, is different, where in the film it is smaller by a factor of 4, reflected in the higher Φ PL (93 vs. 79%) compared to that in solution. The k RISC is 1.71 × 106 s−1. Compared to doped films of DOBNA‐SpAc at the same concentration, the λ PL of DOBNA‐SpAc‐DCz is only slightly red‐shifted and broadened (Figure 3 and Table 1); the ΔE ST is the same, while the τ d decreased from 3.07 µs for DOBNA‐SpAc to 2.04 µs for DOBNA‐SpAc‐DCz. Although both emitters have the same Φ PL of 93%, DOBNA‐SpAc shows a larger Φ prompt (30%) and a smaller Φ del (63%) compared to DOBNA‐SpAc‐DCz (20/73%), indicating more efficient triplet upconversion in the latter. The resulting k RISC increases from 1.02 × 106 s−1 for DOBNA‐SpAc to 1.71 × 106 s−1 for DOBNA‐SpAc‐DCz. Temperature‐dependent TRPL measurements of both films reveal a weak temperature dependence of the delayed emission, confirming the TADF character of the emission (Figure 3d; Figure S6). Electrochemistry Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) experiments in dichloromethane (DCM) using [ n Bu4N]PF6 (0.1 m) as the supporting electrolyte and ferrocene/ferrocenium (Fc/Fc+) as an internal standard were used to infer the energies of the frontier molecular orbitals. The CVs of both DOBNA‐SpAc‐DCz and DOBNA‐SpAc show irreversible reduction waves at similar reduction potentials, E red, of −1.84 and −1.86 V vs. a standard calomel electrode (SCE) (Figure S3) [19]. The CVs of both compounds show quasi‐reversible oxidation waves, with the oxidation potential, E ox, of DOBNA‐SpAc‐DCz being cathodically shifted to 0.97 V vs. SCE compared to that of DOBNA‐SpAc (1.08 V vs. SCE) (Figure S3); this trend is in agreement with the DFT study of the HOMO levels of these two compounds. The LUMO energies for DOBNA‐SpAc‐DCz and DOBNA‐SpAc are similar at −2.50 and −2.49 eV, respectively, while the HOMO for DOBNA‐SpAc‐DCz (−5.31 eV) is destabilized compared to that of DOBNA‐SpAc (−5.42 eV). The corresponding HOMO and LUMO energies are in good agreement with the ones reported for DOBNA‐SpAc (aka. TDBA‐SAF) in the literature (−5.5/−2.4 eV) [9]. When compared to two other DOBNA‐containing emitters (TBDA‐PAS [20] and BO‐tCzPhICz [21]; Figure S2), the LUMO energies are similar (−2.44 [20] and −2.14 eV [21], respectively), indicating that the LUMO is located on the DOBNA moiety. Device Properties Encouraged by the promising photophysical properties, we fabricated OLEDs using DOBNA‐SpAc‐DCz as the emitter. All devices were prepared by solution‐processing up to the emitting layer (EML). The detailed performance of the devices is summarized in Table 2. The device structures for Devices A and B were indium tin oxide (ITO, 50 nm)/poly(styrenesulfonic acid)‐doped poly(3,4‐ethylenedioxythiophene) (PEDOT:PSS (CH8000), 35 nm)/poly(N‐vinylcarbazole) (PVK, 15 nm)/EML (30 nm)/PPF (5 nm)/1,3,5‐Tris(3‐pyridyl‐3‐phenyl)benzene (TmPyPB, 35 nm)/8‐Quinolinolato lithium (Liq, 1 nm)/Al (100 nm) (Figure 4a). PEDOT:PSS CH8000 was employed as the hole‐injection layer due to its high work function, PVK as the hole‐transport and electron‐blocking layer, PPF as the electron‐transport and hole/exciton‐blocking layer, TmPyPB as the electron‐transport layer, and Liq as the electron‐injection layer. The EMLs consisted of PPF doped with 20 wt.% of the emitter; DOBNA‐SpAc‐DCz was used for Device A, while DOBNA‐SpAc was used for Device B as a reference. TABLE 2Device performances of solution‐processed OLEDs using DOBNA‐SpAc‐DCz and DOBNA‐SpAc as emitters.DeviceEmitterElectron‐transporter λ EL a (FWHM) b /nm EQE (Max/@100 cd m−2) c (%) L max d /cd m−2 V on e /VCIE (x, y) f A DOBNA‐SpAc‐DCz TmPyPB474(73)29.4 ± 0.1/5.3 ± 0.5287 ± 366.0(0.145, 0.211) B DOBNA‐SpAc TmPyPB466(68)33.9 ± 0.9/4.9 ± 0.1321 ± 05.6(0.140, 0.136) C DOBNA‐SpAc‐DCz TPBi473(71)13.2 ± 2.8/5.0± 1.0400 ± 186.8(0.147, 0.231) D DOBNA‐SpAc TPBi474(66)28.7 ± 0.9/3.5 ± 0.1229 ± 46.4(0.141, 0.143) a Electroluminescence peak wavelength. b Full width at half maximum. c Average external quantum efficiency of two devices at the maximum value and at 100 cd m − 2. d Average maximum luminance of two devices. e Turn‐on voltage at 1 cd m−2. f CIE coordinates at 1 mA cm−2.John Wiley & Sons, Ltd. Device performances of solution‐processed OLEDs using DOBNA‐SpAc‐DCz and DOBNA‐SpAc as emitters. Electroluminescence peak wavelength. Full width at half maximum. Average external quantum efficiency of two devices at the maximum value and at 100 cd m − 2. Average maximum luminance of two devices. Turn‐on voltage at 1 cd m−2. CIE coordinates at 1 mA cm−2. (a) Device structure and chemical structures of the materials used in solution‐processed OLEDs; (b) Normalized electroluminescence spectra of Device A (DOBNA‐SpAc‐DCz, blue) and Device B (DOBNA‐SpAc, red) employing TmPyPB as the electron‐transport layer; (c) Current density–voltage–luminance characteristics of Devices A and B; (d) External quantum efficiency‐luminance characteristics of Devices A and B; (e) External quantum efficiency‐luminance characteristics of Devices C and D employing TPBi as the electron‐transport layer. Devices A and B exhibited blue emission with peak wavelengths, λ EL, of 474 and 464 nm, respectively, which are in good agreement with the λ PL values of the corresponding doped films (Figures 4b and 3b). The current density‐voltage plots of Devices A and B clearly show four distinct regimes (Figure 4c). Such behavior was not observed in devices employing other combinations of hole‐transport (i.e., OTPB and X‐DCDPA in Device Z; Figure S10) and host materials (i.e., CzSi in Device X or DPEPO in Device Y; Figure S10), indicating that this phenomenon is specific to the interface between PVK and the phosphine‐oxide‐based host material, PPF. Because the LUMO levels of the ETL and the EML are well aligned, electron injection into the EML is efficient. Consequently, at low driving voltages, electrons tend to accumulate near the PVK/PPF interface. By contrast, the large mismatch between the HOMO levels of PVK and PPF is likely to promote localization of hole traps at their interface. As the applied voltage increases, these interfacial hole traps are progressively filled, and once they are filled, holes begin to flow rapidly into the EML. As a result, the accumulated electrons and the injected holes undergo rapid recombination. Exceptionally high EQE max values of 29.4 ± 0.1% for Device A and 33.9 ± 0.9% for Device B were obtained (Figure 4d). However, localized charge accumulation and concentrated recombination also induce a pronounced efficiency roll‐off at high luminance and current density in both devices. Since both devices reached nearly the same maximum luminance, the observed efficiency roll‐off was attributed not to emitter‐specific processes but rather to the host material common to both devices. To isolate the impact of emitter modification on device performance, additional Devices C and D were fabricated by replacing TmPyPB with TPBi (2,2′,2''‐(1,3,5‐Benzinetriyl)‐tris(1‐phenyl‐1‐H‐benzimidazole)) as the electron‐transport layer (Figure S12). TPBi has an electron mobility about two orders of magnitude lower than that of TmPyPB [22, 23]. These devices were designed to reduce the electron density at low operating voltages, thereby suppressing electron accumulation at the interface. The maximum luminance of Device C increased to 400 ± 18 cd m−2, and the efficiency roll‐off was suppressed compared to the corresponding Device D (Figure 4e). Moreover, the introduction of additional carbazole units in DOBNA‐SpAc‐DCz enhances the hole‐transport properties, leading to a more spatially distributed recombination zone within the EML (Figure S11). In addition, the fast RISC of DOBNA‐SpAc‐DCz reduces the triplet exciton density, thereby suppressing exciton‐exciton and exciton‐charge annihilation processes that are responsible for efficiency roll‐off [5]. Conclusions In this work, we demonstrated that by introducing donor dendrons onto the spiroacridine moiety of the blue‐emitting donor‐acceptor TADF emitter, DOBNA‐SpAc, an acceleration of k RISC in DOBNA‐SpAc‐DCz was achieved. DOBNA‐SpAc‐DCz possesses a very small ΔE ST of less than 10 meV and short delayed lifetimes of 2 µs, resulting in high k RISC values of > 1 × 106 s−1 in both toluene solution and 20 wt.% doped film in PPF, making it a promising candidate for solution‐processed OLEDs. The solution‐processed OLEDs with DOBNA‐SpAc‐DCz and DOBNA‐SpAc showed high EQE max of 29.4 ± 0.1% and 33.9 ± 0.9% at CIE coordinates of (0.145, 0.211) and (0.140, 0.136), respectively. The pronounced efficiency roll‐off observed in Devices A and B originates predominantly from interfacial charge accumulation at the PVK/PPF interface rather than intrinsic emitter limitations, as evidenced by the modified device architecture employing TPBi. Notably, the additional carbazole units and faster RISC of DOBNA‐SpAc‐DCz promote improved hole transport and reduced triplet exciton accumulation, enabling a more balanced recombination profile. This leads to a reduced efficiency roll‐off, and the device using DOBNA‐SpAc‐DCz exhibits a higher EQE 100 of 5.0 ± 0.1% and a higher maximum luminance of 400 ± 18 cd m−2 compared to DOBNA‐SpAc (EQE 100 = 3.5 ± 0.1%; L max = 229 ± 4 cd m−2). Author Contributions M.F.: Conceptualization, investigation, methodology, validation, formal analysis, visualization, writing – original draft, writing – review & editing. Y.Y.: Investigation, writing – original draft, writing – review & editing. R.K.: Investigation, writing – review & editing. H.K.: Project administration, resources, supervision, writing – review & editing. E.Z‐C.: Validation, formal analysis, project administration, resources, supervision, writing – review & editing. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Supporting File 1: advs76130‐sup‐0001‐SuppMat.pdf. Supporting File 2: advs76130‐sup‐0002‐Data.zip.