Boosting EQE to nearly 35% in acridone-based TADF materials via HOMO delocalization and nonradiation suppression

Abstract

As weak donors, carbazole (Cz) derivatives are particularly suitable for blue thermally activated delayed fluorescence (TADF) emitters. However, due to the large singlet–triplet energy gap (ΔE_ST) and excessively weak charge transfer state, these are unfavorable for efficient TADF emitters. Herein, a series of blue and sky-blue TADF materials, tCz-mPAO, BCz-mPAO and BCz-PAO, were developed by incorporating tert-butyl-carbazole (tCz) or 3,9′-bicarbazole (BCz) as a donor and 10-pyridyl-acridone as an acceptor. Focusing on simultaneously ensuring efficient reverse intersystem crossing (RISC) and increasing the photoluminescence quantum yields (PLQYs), two molecular design strategies related to the donor and acceptor units were deliberately employed, respectively. By varying the donor from tCz to BCz, the HOMO distribution was significantly delocalized while maintaining a minimal HOMO–LUMO overlap, resulting in an enhanced k_r and optimal ΔE_ST. Furthermore, strategic elimination of the flexible methyl group on the acceptor effectively suppresses energy dissipation, thereby restricting the non-radiative transition rate constant (k_nr). Combining these strategies, BCz-PAO demonstrated an optimal k_r of 9.3 × 10⁷ s⁻¹ and a high PLQY of 98.2% in doped films. The sky-blue OLED exhibited a remarkable maximum external quantum efficiency of 34.4% without light out-coupling enhancement or a TADF-sensitized structure. These synergistic strategies establish a generalizable design paradigm for high-performance TADF emitters.

---

1. Introduction

In recent years, organic light-emitting diodes (OLEDs) have revolutionized display technology by offering superior resolution, wider viewing angles, enhanced color contrast, and flexible form factors.^1–3 However, challenges remain in the device structure and material properties, particularly for blue OLEDs which lag behind their red and green counterparts in efficiency and lifetimes.⁴ The development of organic emitters started with fluorescent materials,^5–7 which were later replaced by phosphorescent materials capable of utilizing triplet excitons.⁸ While phosphorescent OLEDs can achieve high performance, their metal–ligand bonds are inherently unstable, leading to rapid degradation^3,9 and high production costs due to noble metal requirements.¹⁰ Recent advances in purely organic thermally activated delayed fluorescent (TADF) materials show promise, as their small singlet–triplet energy gap (ΔE_ST) enables efficient up-conversion of nonluminous triplet excitons (75%) to radiative singlet states via reverse intersystem crossing (RISC), potentially exceeding the 25% internal quantum efficiency (IQE) limit of conventional fluorescent materials.¹¹

A fundamental requirement for an efficient RISC process is maintaining a sufficiently small ΔE_ST to ensure that the triplet excitons of TADF molecules can absorb ambient heat and revert to the singlet state via RISC.¹² To achieve small ΔE_ST, conventional twisted intramolecular charge transfer (TICT) type TADF emitters are typically designed with large torsion angles between the donor (D) and the acceptor (A) to separate the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO).^13–15 Following the above design principle, a series of TICT-TADF emitters were developed featuring benzophenones,^16,17 sulfones,¹⁸ triazines,^8,19 and their derivatives as acceptors, paired with triphenylamines,^20,21 carbazoles,^22–24 acridines,^25,26 and their derivatives as donors. Furthermore, by introducing large sterically hindering groups, the torsion angles between these acceptors and donors can be modulated to theoretically minimize the overlap of the frontier molecular orbitals (FMOs) so as to realize small ΔE_ST.²⁷ Moreover, strategic enhancement of donor/acceptor strengths can also lower the ¹CT energy level, which narrows ΔE_ST and consequently accelerates the RISC process.²⁸

To obtain a high EL efficiency, it is equally critical to achieve superior photoluminescence quantum yields (PLQYs), which depend on both the enhanced radiative transition rate constant (k_r) and the suppressed nonradiative transition rate constant (k_nr). Considering from the viewpoint of quantum chemistry, achieving high k_r typically requires a large HOMO–LUMO overlap, which contradicts the realization of small ΔE_ST.⁸ To address this challenge, Prof. Adachi et al. proposed an alternative strategy that involves extending the delocalization of the HOMO and LUMO in a charge-transfer compound with a well-separated HOMO and LUMO, which can induce a large transition dipole moment while maintaining a small ΔE_ST. Based on this strategy, they developed a series of blue TADF emitters by systematically modulating the HOMO delocalization through the introduction of secondary donors and peripheral modifications, demonstrating the simultaneous achievement of both a high k_r and a small ΔE_ST.²⁹ In 2015, Lee et al. reported a series of triazine-based TADF materials with an evenly distributed HOMO brought by increasing donor units and optimizing terminal groups, resulting in high PLQY close to 100% alongside a marginal reduction of ΔE_ST.³⁰ In 2019, Tao et al. incorporated dibenzothiophene and dibenzofuran as secondary electron donors, effectively extending the HOMO distribution and consequently enhancing PLQYs while maintaining low ΔE_ST.³¹ Besides, expansion of HOMO distributions can also be achieved by increasing the conjugation length of donor groups. It was confirmed by Mei et al. in 2022 that in their acridone-based TADF materials, the conjugation length of donor groups could be increased by introducing peripheral groups such as phenyl and tert-butylphenyl at the 3,6-site of Cz donors, resulting in an expansion of HOMO distribution and enhanced oscillator strength (f). The donor engineering strategy applied to 3,6-DDPhCz-AD enabled significantly enhanced TADF performance, with nearly doubled PLQY (42% for 3,6-DCz-AD and 83% for 3,6-DDPhCz-AD) and an increased RISC rate constant (k_RISC) by two orders of magnitude due to a reduced ΔE_ST.³² This result demonstrates that the adopted strategy successfully reconciles low ΔE_ST with high PLQYs, revealing their fundamental compatibility.

Suppressing k_nr requires enhanced molecular rigidity to minimize energy dissipation. Practical strategies include intramolecular locking³³ and implementation of intramolecular hydrogen bonding.³⁴ A recent study by our research team confirmed that acridin-9(10H)-one, i.e. acridone (AD), with a large rigid cyclic framework and balanced electron-withdrawing capacity, serves as an excellent acceptor unit for blue emitters. Notably, acridone-based TADF molecules combine synthetic accessibility with inherent structural advantages. The sp²-hybridized carbonyl group enables n–π* electronic transitions while simultaneously restricting molecular conjugation. Furthermore, the unique sp²-hybridization of the 10-site nitrogen atom of the AD ring enforces the coplanarity of the three six-membered rings while maintaining the orthogonal orientation of the 10-site phenyl ring, conferring high molecular rigidity. Therefore, numerous acridone-based deep-blue or pure-blue TADF emitters were developed, exhibiting an exceptional k_r of 10⁷–10⁸ s⁻¹.^15,32 To further increase the molecular rigidity and suppress excited state vibrational relaxation, Mei et al. designed and synthesized a pyridine-substituted acridone by incorporating pyridyl from its 2-position on the AD ring rather than from the 3-position reported in the literature,³⁵ which led to the formation of intramolecular hydrogen bonds between the pyridinic N atom and the 4- or 5-site H atom of the AD ring and thus restricted the k_nr by 1000 times.³⁴ Additionally, a high horizontal dipole ratio is imperative for improving the limited light outcoupling efficiency (η_out)³⁶. Acridone-based TADF molecules adopt a compact rod-like conformation, which promotes favorable alignment of their emission dipole moments.

In this work, three sky-blue TADF emitters, tCz-mPAO, BCz-mPAO, and BCz-PAO, were designed and synthesized using 10-(5-methylpyridin-2-yl)-3-(trifluoromethyl)acridin-9(10H)-one (mPAO) and 10-(pyridin-2-yl)-3-(trifluoromethyl)acridin-9(10H)-one (PAO) as new acceptors paired with the tert-butyl-carbazole (tCz) or 3,9′-bicarbazole (BCz) donor, aiming to explore the effects of two molecular design strategies, i.e. the HOMO distribution expansion and flexible group removal, on the luminescence performance. First, the tCz and BCz donors were selected basically because the small dihedral angle between the five-membered carbazole donor and the AD ring ensures an effective FMO overlap, leading to a large f and high k_r. Meanwhile, through donor engineering, specifically replacing the steric tert-butyl group with a planar carbazole moiety, electron distributions were delocalized, which induced a larger transition dipole moment, thus increasing PLQYs from 77.7% (tCz-mPAO) to 84.4% (BCz-mPAO) while ensuring a high k_RISC by reducing ΔE_ST. Then, the second design strategy was applied to the acceptor by removing the flexible methyl group. Pyridine-substituted acridones were employed as acceptors since the intramolecular hydrogen bonds could enhance the molecular rigidity. In tCz-mPAO and BCz-mPAO, the introduction of a methyl group at the pyridine moiety was designed to attenuate the acceptor's electron-withdrawing strength, thereby enabling emission color tuning. However, this modification simultaneously increased molecular vibrational and rotational freedom, leading to enhanced non-radiative decay pathways. By strategically eliminating the conformationally flexible methyl group from the acceptor, the k_nr value of BCz-PAO was reduced by 10 orders of magnitude to 1.3 × 10⁶ s⁻¹, resulting in a further PLQY increase from 84.4% (BCz-mPAO) to 98.2% (BCz-PAO). Benefitting from both the enhanced k_RISC and PLQY through the HOMO distribution expansion and the methyl removal, the corresponding OLEDs with tCz-mPAO, BCz-mPAO and BCz-PAO as doped emitters exhibited stepwise enhancement in EQE_max values of 25.7%, 27.8%, and 34.4%, respectively. The results demonstrate that these dual strategies establish a universal design paradigm for developing high-efficiency TADF emitters by achieving record-high PLQYs while lowering ΔE_ST.

2. Experimental section

The experimental details are provided in the SI. All of the chemicals for compound syntheses were purchased from commercial sources and used as received unless stated otherwise. The important intermediate 3-fluoro-6-(trifluoromethyl)acridin-9(10H)-one was synthesized according to the procedure reported in the literature.³⁷

2.1 General procedure for the synthesis of 3-F-6-CF₃-AD-mPy and 3-F-6-CF₃-AD-Py

The intermediates 3-fluoro-6-(trifluoromethyl)acridin-9(10H)-one (2 g, 7.11 mmol), 2-bromo-5-methylpyridine (1.49 g, 9.25 mmol) or 2-bromopyridine (1.59 g, 9.25 mmol), potassium carbonate (1.47 g 10.64 mmol), copper(I) iodide (0.135 g, 0.71 mmol), and 2,2,6,6-tetramethyl-3,5-heptanedione (0.249 g, 1.35 mmol) were dissolved in anhydrous DMF in a 100 mL three-necked flask. The mixture was then refluxed at 156 °C under a nitrogen atmosphere for 24 h. After removal of the solvent, the mixture was extracted three times with water and dichloromethane, and the product was finally obtained by chromatographic purification (eluent : PE/EA = 5 : 1).

3-F-6-CF₃-AD-mPy: pale yellow solid (1.70 g), yield: 64.20%. ¹H NMR (500 MHz, CDCl₃) δ 8.73 (s, 1H), 8.65 (d, J = 8.4 Hz, 1H), 8.61–8.52 (m, 1H), 7.95 (d, J = 8.0 Hz, 1H), 7.50 (d, J = 8.4 Hz, 1H), 7.39 (d, J = 7.7 Hz, 1H), 7.03 (t, J = 8.4 Hz, 1H), 6.87 (s, 1H), 6.28 (d, J = 10.9 Hz, 1H), 2.60 (s, 3H).

3-F-6-CF₃-AD-Py: pale yellow solid (1.61 g), yield: 60.69%. ¹H NMR (500 MHz, CDCl₃) δ 9.01–8.91 (m, 1H), 8.66 (d, J = 8.3 Hz, 1H), 8.58 (dd, J = 8.9, 6.5 Hz, 1H), 8.19 (td, J = 7.7, 1.9 Hz, 1H), 7.71 (dd, J = 7.6, 4.9 Hz, 1H), 7.52 (dd, J = 10.3, 8.1 Hz, 2H), 7.10–6.99 (m, 1H), 6.84 (s, 1H), 6.26 (dd, J = 10.8, 2.3 Hz, 1H).

2.2 General procedure for the synthesis of tCz-mPAO, BCz-mPAO and BCz-PAO

A mixture of tert-butyl-carbazole (413 mg, 1.48 mmol) or 9H-3,9′-bicarbazole (491 mg, 1.48 mmol), intermediate 3-F-6-CF₃-AD-Py (500 mg, 1.34 mmol) or 3-F-6-CF₃-AD-mPy (480 mg, 1.34 mmol) and Cs₂CO₃ (875 mg, 2.69 mmol) was stirred in anhydrous DMF in a 100 mL three-necked flask and then refluxed at 165 °C under a nitrogen atmosphere for 12 h. After removing the solvent by distillation, chromatographic purification (eluent:DCM) provided the crude product. Repeated recrystallization from CHCl₃/CH₃OH yielded the pure products.

tCz-mPAO: pale yellow solid (700 mg), yield: 82.51%. ¹H NMR (500 MHz, CDCl₃) δ 8.75 (dd, J = 18.4, 8.9 Hz, 3H), 8.12 (s, 2H), 7.91 (d, J = 8.5 Hz, 1H), 7.62–7.53 (m, 2H), 7.46 (t, J = 7.9 Hz, 3H), 7.37 (d, J = 8.7 Hz, 2H), 6.96 (s, 1H), 6.87 (s, 1H), 2.52 (s, 3H), 1.48 (s, 18H). MALDI-HRMS (m/z): calcd for C₄₀H₃₆F₃N₃O 631.2805, found 631.2792 [M⁺]. Anal. calcd for C₄₀H₃₆F₃N₃O: C, 76.05; H, 5.74; N, 6.65; found: C, 76.08; H, 5.72; N, 6.64.

BCz-mPAO: pale yellow solid (800 mg), yield: 85.47%. ¹H NMR (500 MHz, DMSO) δ 8.69 (s, 1H), 8.63 (d, J = 8.5 Hz, 1H), 8.59 (d, J = 8.4 Hz, 1H), 8.49 (s, 1H), 8.25 (t, J = 8.6 Hz, 3H), 8.11 (d, J = 8.2 Hz, 1H), 7.91 (d, J = 8.1 Hz, 1H), 7.75 (d, J = 8.7 Hz, 1H), 7.65 (dd, J = 8.5, 4.7 Hz, 2H), 7.57 (d, J = 9.0 Hz, 1H), 7.49 (d, J = 8.4 Hz, 1H), 7.44 (t, J = 7.8 Hz, 1H), 7.37 (t, J = 7.6 Hz, 2H), 7.32–7.23 (m, 5H), 6.94 (s, 1H), 6.87 (s, 1H), 2.39 (s, 3H). MALDI-TOF-MS (m/z): calcd for C₄₄H₂₇F₃N₄O 684.2131, found 684.2148 [M⁺]. Anal. calcd for C₄₄H₂₇F₃N₄O: C, 77.18; H, 3.97; N, 8.18; found: C, 77.20; H, 3.95; N, 8.19.

BCz-PAO: pale yellow solid (750 mg), yield: 81.56%. ¹H NMR (500 MHz, DMSO) δ 8.93 (d, J = 4.9 Hz, 1H), 8.71 (d, J = 8.5 Hz, 2H), 8.67 (s, 1H), 8.54 (d, J = 2.1 Hz, 3H), 8.38–8.28 (m, 5H), 8.27 (s, 3H), 8.06 (d, J = 7.9 Hz, 3H), 7.86 (dd, J = 8.6, 1.9 Hz, 2H), 7.75 (dd, J = 16.6, 8.2 Hz, 5H), 7.63 (dd, J = 8.7, 2.1 Hz, 2H), 7.55–7.40 (m, 7H), 7.37–7.27 (m, 10H), 6.94 (s, 2H), 6.89 (d, J = 2.0 Hz, 2H). MALDI-TOF-MS (m/z): calcd for C₄₃H₂₅F₃N₄O 670.1975, found 670.1972 [M⁺]. Anal. calcd for C₄₃H₂₅F₃N₄O: C, 77.01; H, 3.76; N, 8.50; found: C, 76.98; H, 3.78; N, 8.50.

3. Results and discussion

3.1 Synthesis and electrochemical properties

The chemical structure and synthetic routes of the target molecules are illustrated in Scheme 1. The intermediate 3-F-6-CF₃-AD was synthesized according to the procedure reported in the literature. Through Ullmann coupling with 2-bromo-5-methylpyridine and 2-bromopyridine, respectively, the two key intermediates 3-F-6-CF₃-AD-mPy and 3-F-6-CF₃-AD-Py were obtained at a high yield of over 60%. Finally, nucleophilic substitution reactions with different donors in anhydrous N,N-dimethylformamide (DMF) at 165 °C afforded the final products tCz-mPAO, BCz-mPAO and BCz-PAO. All compounds were initially purified by column chromatography, followed by recrystallization from methanol/dichloromethane mixtures. This purification process was repeated until spectroscopic purity was achieved prior to optical characterization and vacuum-deposited OLED device fabrication. Detailed structural characterization spectra of all intermediates and final products are provided in the SI.

Scheme 1 Chemical structures and synthetic routes of tCz-mPAO, BCz-mPAO and BCz-PAO.

Cyclic voltammetry (CV) was conducted to investigate the oxidation and reduction potentials of these emitters. As shown in Fig. 1, tCz-mPAO, BCz-mPAO and BCz-PAO all exhibited reversible oxidation and reduction processes, demonstrating the excellent electrochemical stability of their acceptors and donors. The HOMO energy levels were determined from the onset potential (E^onset_ox) of the first oxidation wave according to the formula E_HOMO = −(E^onset_ox + 4.4), and the LUMO energy levels of these compounds were determined from the onset potentials of the first reduction wave (E^onset_red) according to the formula E_LUMO = −(E^onset_red + 4.4).^38,39 In this way, the HOMO and LUMO levels of their molecules were calculated and are summarized in Table 1. The reduction process occurred predominantly on the mPAO and PAO acceptor groups, as evidenced by their nearly identical LUMO levels (−3.04 for tCz-mPAO and −3.05 eV for BCz-mPAO and BCz-PAO). Although the structures of the two acceptors are slightly different (methyl substitution on the pyridine ring), the pyridine ring is almost perpendicular to the AD ring, resulting in LUMO distributions hardly extending to the pyridine ring, and therefore, the methyl group on the pyridine did not affect the LUMO distributions, much less than the LUMO levels, which corresponds to the results of theoretical calculations (vide infra). The HOMO levels, determined by the oxidation characteristics of the donor units, revealed that tCz-mPAO exhibited a HOMO level at −5.61 eV, while BCz-mPAO and BCz-PAO shared similar values of −5.67 eV and −5.68 eV. The calculated E_g values indicated that both molecules feature wide energy gaps, suggesting their potential for blue emission.

Fig. 1 Cyclic voltammograms of tCz-mPAO, BCz-mPAO and BCz-PAO in dilute DCM (anodic) and DMF (cathodic) solutions.

Table 1 The experimentally determined physical parameters of tCz-mPAO, BCz-mPAO and BCz-PAO

Emitter	λabs^a [nm]	λem^a [nm]	ES^b [eV]	ET^b [eV]	ΔEST^b [eV]	HOMO/LUMO^c [eV]	Eg^c [eV]
a Absorption and fluorescence peak wavelengths in toluene solutions at room temperature.b Estimated from the short-wavelength onset of the fluorescence/phosphorescence spectra at 77 K in frozen 2-Me-THF solution, ΔEST = ES − ET.c HOMO levels were determined from cyclic voltammetry measurements. LUMO levels were determined from CV measurements or calculated by using the HOMO levels and Eg.
tCz-mPAO	375/392	450	3.10	2.71	0.39	−5.61/−3.04	2.57
BCz-mPAO	374/393	476	3.06	2.73	0.33	−5.67/−3.05	2.62
BCz-PAO	372/393	480	3.07	2.73	0.34	−5.68/−3.05	2.63

---

3.2 Theoretical calculations

To comprehensively investigate HOMO and LUMO distributions and excited state of these emitters, density functional theory (DFT) and time-dependent DFT (TD-DFT) calculations were performed at the B3LYP level and are illustrated in Fig. 2a. As expected, the optimized ground-state geometries revealed the C–N⋯H distances of 2.75–2.77 Å between the pyridyl N-atom and the 4,5-position hydrogen atoms of the AD core (Fig. S1 in the SI), falling within the effective van der Waals range and consistent with the hydrogen-bonding interactions previously demonstrated by Mei et al. (2023) in similar systems.³⁴ The observed orthogonal orientation of the pyridyl ring contributed to further enhancing molecular rigidity while suppressing vibrational relaxation. Furthermore, the decrease in the S₀ → S₁ root-mean-square deviation (RMSD) from 1.9523 Å for BCz-mPAO to 1.6890 Å for BCz-PAO (Fig. S2 in the SI) confirmed that the removal of the flexible methyl group indeed enhances the molecular rigidity, which should thereby be favorable to suppress the non-radiative transitions. More importantly, the unique orthogonal orientation of the 10-position pyridine ring localized the LUMO predominantly on the acridone fragment of the mPAO/PAO acceptors, with negligible distribution on the pyridine moiety, and thus the terminal methyl group on the pyridine ring has a minimal impact on the LUMO distributions or energy levels. It was also observed that, in all molecules, the small dihedral angle between the donor and acceptor units (θ ≈ 50°) confined the LUMOs primarily to the acceptors, while the HOMOs were predominantly localized on the donors albeit with minor extension to the acceptors. This spatial distribution resulted in partial separation of the FMOs while preserving the appropriate overlap, thereby facilitating large transition dipole moments, which was an intrinsic advantage of carbazole derivatives as donors. Furthermore, the quantitative analysis of the HOMO distribution in the ground states and the hole distribution in the S₁ states for these emitters is shown in Fig. S3 in the SI. Compared to tCz-mPAO for which the carbazole ring contributes 85% to the HOMO, the HOMO uniformly spans the whole BCz donor in BCz-mPAO, with 22% on the inner carbazole and as high as 75.9% on the peripheral carbazole ring. Similarly, the hole cloud of the S₁ state for BCz-mPAO is also mainly dispersed on the terminal carbazole ring (66.8%). Apparently, the HOMO delocalization brought by the BCz donors will result in favorable transition dipole moments and finally lead to larger k_r.

Fig. 2 (a) Chemical structures, optimized geometries, FMO distributions and dihedral angles of tCz-mPAO, BCz-mPAO and BCz-PAO. (b) The NTOs of the S₀ → S₁ and S₀ → T₁–T₃ transitions of three molecules.

The FMO distributions of these three emitters were also simulated using different functionals, such as ωB97XD and M062X, and the results calculated are compared with B3LYP in Fig. S4. And the absorption and fluorescence wavelengths and the excited state energies were also calculated using different functionals and compared with the experimentally obtained data in Table S3. Apparently, all three functionals consistently reproduced the same HOMO delocalization trend (Fig. S4), showing progressive enhancement from tCz-mPAO to BCz-mPAO and BCz-PAO. As shown in Table S3, although all functionals showed deviations from experimental values, B3LYP demonstrated the closest agreement with experimental data for almost all of the key parameters.

To further study the transition characteristics, the triplet state spin density distributions (TSDDs, T₁) are presented in Fig. S5 (SI) and natural transition orbitals (NTOs) of three emitters are depicted in Fig. 2b. The TSDDs show that all molecules exhibit predominant electron density localization on the mPAO and PAO acceptors with small contribution from the carbazole donor, which indicates that the T₁ states primarily originate from the local excited (LE) state of the acridone fragment with a small charge transfer component. The NTO analysis of the S₁ state reveals that while the particles of all three compounds are concentrated on the nearly identical acceptors, their hole distributions exhibit significant differences. In tCz-mPAO, the hole orbital is primarily localized on the tCz donor, with tiny extension to the AD ring of the acceptor, overlapping with the particle distribution. This spatial separation results in S₁ mainly exhibiting the charge-transfer (CT) character. In contrast, BCz-mPAO and BCz-PAO with large steric donors display delocalized hole distributions that expand further into the peripheral carbazole unit rather than the AD ring. This enhanced hole delocalization leads to more pronounced CT characteristics of S₁ states, increasing the n → π* transition ratios of the S₁ states, while stabilizing the S₁ energy levels (from 3.08 eV in tCz-mPAO to 2.91/2.89 eV in BCz-mPAO/BCz-PAO), resulting in distinct emission red-shifts. The S₀ → T₁ transitions of three compounds show that the particle distributions are primarily localized on the acceptor units, whereas the holes are distributed across the entire molecular frameworks, indicating the hybrid character of the local-excited triplet state (³LE) and the charge-transfer triplet state (³CT). As governed by the El-Sayed rule,⁴⁰ the CT nature of the S₁ state combined with the partial LE feature of the T₁ state creates optimal conditions for significantly high SOC matrix elements (Table S4), thereby accelerating the RISC rate. It is noteworthy that, while the particle and hole distributions of T₂ states in all molecules resemble the corresponding T₁ states, their small energy gaps between T₂ and T₁ enable T₂ to function as excellent transition states for exciton up-conversion. This configuration facilitates an efficient RISC channel via a T₁ → T₂ → S₁ pathway.

3.3 Photophysical properties

UV-Vis absorption and PL emission spectra were measured to investigate the photophysical properties of these target molecules. The normalized spectra are depicted in Fig. 3, with the corresponding data tabulated in Table 1. All molecules displayed similar spectral profiles and exhibited two prominent absorption bands in the range of 290–320 nm. These bands can be attributed to the π–π* electronic transitions of the carbazole-based donor units, as well as the backbone absorptions of both the donor and acceptor moieties. Furthermore, the absorption bands observed in the range of 350–420 nm for all three molecules should originate from both local n–π* and π–π* transitions of the mPAO/PAO acceptors and intramolecular charge-transfer (ICT) transitions between the tCz/BCz donors and acceptor moieties.

Fig. 3 UV-Vis absorption spectra in dilute toluene solutions and solvatochromic effect on the PL spectra of (a) tCz-mPAO, (b) BCz-mPAO and (c) BCz-PAO, and the low-temperature PL and phosphorescence spectra of (d) tCz-mPAO, (e) BCz-mPAO and (f) BCz-PAO in frozen 2-Me-THF glass at 77 K.

The bathochromic shifts for all three compounds were observed with increasing solvent polarity, further demonstrating the ICT character of these molecules. In dilute toluene solution, the PL spectra showed broad and structureless emission bands, characteristic of predominant CT state emission. All three molecules exhibited blue emission, and BCz-mPAO and BCz-PAO showed slightly red-shifted fluorescence compared to tCz-mPAO, which can be attributed to the more pronounced CT character of the S₁ state in the BCz-based donor systems due to HOMO expansion, as supported by theoretical calculations.

The low-temperature (LT) fluorescence and phosphorescence spectra of the two emitters in 2-methyltetrahydrofuran (2-MeTHF) solution were measured as shown in Fig. 3. Apparently, in the fluorescence measuring mode, the obtained LT-PL spectra all contain a structureless fluorescence band in the 400–450 nm range and a phosphorescence band with well-resolved vibronic features in the 450–550 nm range, the latter of which happens to be the same as the phosphorescence spectra that were measured in the phosphorescence mode. These spectral features confirmed the CT character of the S₁ state and the LE character of the T₁ state, as predicted by the NTO results (Fig. 2b). According to the NTO calculations, the phosphorescence of T₁ mainly arises from the ³LE state of the acridone fragment, which explains the minimal difference observed in the phosphorescence spectra of these molecules. For consistency, the S₁ and T₁ state energies were determined from the short-wavelength onsets of the fluorescence and phosphorescence spectra to be 3.10/2.71 eV, 3.06/2.73 eV and 3.07/2.73 eV for tCz-mPAO, BCz-mPAO and BCz-PAO, with corresponding ΔE_ST values of 0.39 eV, 0.33 eV and 0.34 eV, respectively. The marginally smaller ΔE_ST values of BCz-PAO and BCz-mPAO presumably should stem from the strengthened electron-donating ability of the BCz donors, which downshifted the ¹CT levels. All corresponding data are summarized in Table 1.

In order to further investigate the delayed fluorescence behavior and the luminescence mechanism of these emitters, time-resolved PL (TRPL) experiments were carried out in doped films with a 2,8-bis(diphenylphosphoryl)dibenzo[b,d]furan (PPF) host at concentrations of 20 wt% for tCz-mPAO, 20 wt% for BCz-mPAO and 30 wt% for BCz-mPAO. At room temperature, the transient PL spectra of all doped films displayed a double-exponential decay, consisting of both prompt and delayed components, as shown in Fig. 4. Through a time-correlated single photon counting (TCSPC) technique, the lifetimes of the prompt fluorescence (PF, τ_PF) of tCz-mPAO, BCz-mPAO and BCz-PAO were measured to be 7.1 ns, 8.3 ns and 8.4 ns, respectively. Using the multichannel scanning (MCS) method, the lifetimes of the delayed components (τ_DF) of their doped films were measured, which decreased from 22.4 μs via 11.4 μs to 10.6 μs. The remarkable shortening of τ_DF may be attributed to the reduced ΔE_ST, which, in principle, facilitates a more efficient and faster RISC process. The delayed components of these emitters exhibited spectral profiles identical to the corresponding PF, with a gradual reduction in spectral intensity as the delay time increases (Fig. S6), suggesting the delayed fluorescence (DF) in these components. In addition, in the temperature-dependent transient PL measurements, regular increases in PL intensity with increasing temperature, especially in the short time range (e.g. the first 500 μs for tCz-mPAO and the first 300 μs for BCz-mPAO and BCz-PAO), confirmed the TADF mechanism of the DF for all three emitters.

Fig. 4 Transient PL decay curves of (a) tCz-mPAO, (b) BCz-mPAO and (c) BCz-PAO doped in PPF films at 298 K, and their temperature-dependent transient PL decay curves (d)–(f) in doped films.

The PLQYs (Φ_PL) of these emitters were measured for doped films and are summarized in Table 2. Remarkably, all these emitters exhibit high PLQYs in the solid films. Based on these results, the radiative transition rate constant k_rs, the non-radiative transition rate constant k_nr, the intersystem crossing rate constant k_ISC, and the RISC rate constants k_RISCs for these emitters were calculated using the methodology provided in the SI and summarized in Table 2. As expected, compared with tCz-mPAO (7.4 × 10⁷ s⁻¹), the k_r values of BCz-mPAO and BCz-PAO increased to 8.6 × 10⁷ s⁻¹ and 9.3 × 10⁷ s⁻¹, respectively, which is mainly owing to the extended HOMO distribution on the BCz moiety. Besides, due to the removal of flexible groups, the k_nr of BCz-PAO was significantly reduced to 0.13 × 10⁷ s⁻¹, showing at least a 10 time decrease relative to BCz-mPAO and leading to a further increase of PLQY up to 98.2% for BCz-PAO. Furthermore, both BCz derivatives exhibited increased k_RISC values relative to tCz-mPAO (Table 2), indicating the enhanced TADF characteristics. This obviously benefits from the reduced ΔE_ST and the favorable SOC effects.

Table 2 Excited-state dynamic parameters of TADF emitters in doped PPF films at concentrations of 20 wt% for tCz-mPAO, 20 wt% for BCz-mPAO and 30 wt% for BCz-mPAO

Emitter	τPF/τDF [ns μs^−1]	ΦPL [%]	ΦPF/ΦDF [%]	kr [10^7 s^−1]	kISC [10^7 s^−1]	kRISC [10^4 s^−1]	knr [10^7 s^−1]
tCz-mPAO	7.1/22.4	77.7	52.4/25.3	7.4	4.6	6.6	2.12
BCz-mPAO	8.3/11.4	84.4	72.6/11.8	8.6	1.9	10.4	1.60
BCz-PAO	8.4/10.6	98.2	77.9/20.7	9.3	2.5	11.9	0.13

---

3.4 Electroluminescence

To evaluate the electroluminescence (EL) performance of three TADF emitters, multi-layer OLEDs were fabricated with the structure of ITO/HAT-CN (5 nm)/TAPC (20 nm)/TCTA (5 nm)/mCP (5 nm)/EML (20 nm)/PPF (5 nm)/TmPyPb (40 nm)/LiF (1 nm)/Al (200 nm) (devices B1, B2 and B3 for tCz-mPAO, BCz-mPAO and BCz-PAO, respectively), as shown in Fig. 5. EML represents the emitting layer with these emitters doped in the PPF host at a concentration of 20 wt% for BCz-mPAO and 30 wt% for BCz-PAO. Herein, HAT-CN (1,4,5,8,9,11-hexaazatriphenylenehex-acarbonitrile) acted as a hole injection layer. TAPC (1,1-bis[(di-4-tolylamino)phenyl]cyclohexane) and TmPyPB served as the hole-transporting (HT) and electron-transporting (ET) layers, respectively. With high triplet energies, TCTA (4,4,4-tris(N-carbazolyl)triphenylamine) was used as an exciton-blocking layer (EBL) to confine the triplet excitons within the EML. The electron-transporting PPF was chosen as the host in EML because of its high triplet state energy level and suitable HOMO/LUMO energy levels. The selected doping concentration of 20% and the device configuration for B1 and B2 were obtained after a systematic optimization exploration of fabrication conditions using tCz-mPAO as an example, as shown in Fig. S7. Different from other two analogues, the BCz-PAO device achieved the optimized performance with higher efficiencies at a relatively higher doping concentration of 30 wt% (device B3) than at 20 wt% (device B4, Fig. S8). The energy level diagram and chemical structures of these functional layers are illustrated in Fig. 5(a) and (b). And the EL spectra, current density–voltage–luminance (J–V–L) characteristics and efficiency curves are shown in Fig. 5(c) and (d), with pertinent EL data summarized in Table 3.

Fig. 5 (a) Device configuration and energy level diagrams, (b) chemical structures of materials used in devices, (c) J–V–B characteristics, and (d) EQE curves and EL spectra (inset) of the blue TADF-OLEDs in devices B1–B3.

Table 3 EL data summary of blue OLEDs in devices B1–B3

Device	Emitter	Von^a [V]	Lmax^b [cd m^−2]	ηc^c [cd A^−1]	ηp^c [lm W^−1]	ηext^c [%]	λEL^d [nm]	CIE^e [x, y]
a Von, turn-on voltage at 1 cd m^−2.b Lmax, maximum luminance.c ηc/ηp/ηext, the maximum current efficiency/power efficiency/external quantum efficiency, and the values after the slash are the efficiencies at 100 cd m^−2.d λEL, EL peak wavelength at 7 V.e CIE (x, y), Commission Internationale de I’Eclairage coordinates at 7 V.
B1	tCz-mPAO	3.1	3580	35.5/11.2	25.7/11.6	25.7/11.7	460	(0.15, 0.17)
B2	BCz-mPAO	3.0	7124	53.6/25.5	56.0/20.7	27.8/13.0	470	(0.19, 0.29)
B3	BCz-PAO	2.8	14 448	73.4/36.7	82.3/33.2	34.4/17.2	478	(0.20, 0.32)

---

As shown in the EL spectra in Fig. 4(d), devices B1, B2 and B3 exhibited pure-blue to sky-blue EL with emission peaks at 460 nm, 470 nm and 478 nm and the corresponding CIE coordinates of (0.15, 0.17), (0.19, 0.29) and (0.20, 0.32). Notably, all three devices turned on (to deliver a brightness of 1 cd m⁻²) at low voltages (V_on) of 3.1, 3.0 and 2.8 V, respectively. Evidenced by the J–V–B characteristics in Fig. 4(c), tCz-mPAO based device B1 exhibited the lowest current density and luminance in the whole voltage range, consequently resulting in the lowest efficiencies (Fig. 4(c) and Fig. S9), e.g. with a maximum EQE (η_ext) of 25.7%. This slightly inferior device performance should primarily correlate with the relatively lower PLQY and a larger ΔE_ST that restricts k_RISC. As anticipated, a regular increase of EL efficiencies was observed for these three analogues, corresponding to their gradually improved PLQY and k_RISC. In particular, BCz-PAO based device B3 realized the optimal performance with an EQE_max of 34.4%. Evidently, the EQE_max of 34.4% for BCz-PAO is not only the highest efficiency reported for acridone based blue TADF emitters so far,^{7,15,32,34,37,41–44} but also among the highest values reported for sky-blue TICT-TADF emitters and OLEDs that do not adopt the light out-coupling enhancement or sensitizer structures.^8,35,45,46 This remarkable performance definitely benefits from the suitable molecular engineering by HOMO distribution delocalization and the flexible methyl group elimination, which collectively contribute to both high PLQY and enhanced k_RISC.

4. Conclusions

In summary, a series of blue TADF emitters, tCz-mPAO, BCz-mPAO, and BCz-PAO, were designed by integrating tert-butyl-carbazole (tCz) or 3,9′-bicarbazole (BCz) as a donor unit along with strategically modified acridone acceptors, to achieve both high PLQYs and efficient RISC process through two major design strategies, i.e. HOMO delocalization and nonradiation suppression. Based on identical molecular skeletons, through the implementation of more conjugated BCz donors, a large delocalization of molecular orbitals in structures with well-separated HOMO and LUMO was obtained, which can effectively increase k_r while simultaneously reducing ΔE_ST. Therefore, BCz-mPAO achieved a 1.5-fold enhancement in k_RISC and an increase in k_r and PLQY (from 77.7% to 84.4%), ultimately boosting the device efficiency from 25.7% to 27.8%. Furthermore, one flexible methyl group was eliminated from the acridone acceptor to enhance molecular rigidity and suppress non-radiative decay pathways, which resulted in a remarkable reduction of k_nr by at least 10 times and thus a final increase of the PLQY up to 98.4% for BCz-PAO. By means of these stepwise molecular engineering strategies, the BCz-PAO emitter realized a remarkable EQE_max of 34.4% in simple-structure sky-blue TADF-OLEDs, exceeding those of most of the reported acridone based TADF emitters and many other blue TADF materials with similar CIE coordinates under conditions without light out-coupling enhancement or sensitizer structures. For TICT-TADF emitters, enlarging the dihedral angle between donor (D) and acceptor (A) units is typically employed to reduce the HOMO–LUMO overlap, thereby decreasing the ΔE_ST and increasing k_RISC, but sacrificing the k_r or PLQY. Herein, these high-performance TADF materials demonstrated that strategic HOMO delocalization is a dual-functional design approach, concurrently facilitating efficient RISC process and achieving high PLQYs.

Author contributions

Jingjie Yang: synthesis, characterization, data curation, investigation, and writing – original draft. Jing Jin: investigation, fabricating and optimization of the devices. Jiuyan Li: funding acquisition, investigation, supervision, and writing – review and editing. Jiahui Wang: software and formal analysis. Hongyu Chen: investigation and data curation. Meiling Mao: synthesis and characterization. Yongqiang Mei: investigation, supervision, and writing – review and editing. Di Liu: investigation, project administration, resources, and supervision.

Conflicts of interest

There are no conflicts of interest to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d5tc03112e.

Acknowledgements

We thank the National Natural Science Foundation of China (22478063 and 22408035) and the Fundamental Research Funds for the Central Universities (DUT22LAB610) for financial support of this work.

References

1. H. Uoyama, K. Goushi, K. Shizu, H. Nomura and C. Adachi, Nature, 2012, 492, 234–238.
2. N. Thejo Kalyani and S. J. Dhoble, Renewable Sustainable Energy Rev., 2012, 16, 2696–2723.
3. J. Jiang and J. Y. Lee, Mater. Today, 2023, 68, 204–233.
4. G. Hong, X. Gan, C. Leonhardt, Z. Zhang, J. Seibert, J. M. Busch and S. Bräse, Adv. Mater., 2021, 33, 2005630.
5. Z. Xu, B. Z. Tang, Y. Wang and D. Ma, J. Mater. Chem. C, 2020, 8, 2614–2642.
6. C. Adachi, Jpn. J. Appl. Phys., 2014, 53, 060101.
7. M. Xu, J. Li, D. Liu, Y. Mei, B. Liu, Q. Wu, Y. Li, J. Wang and H. Wan, Dyes Pigm., 2023, 218, 111503.
8. K. Shizu, H. Noda, H. Tanaka, M. Taneda, M. Uejima, T. Sato, K. Tanaka, H. Kaji and C. Adachi, J. Phys. Chem. C, 2015, 119, 26283–26289.
9. S. Schmidbauer, A. Hohenleutner and B. König, Adv. Mater., 2013, 25, 2114–2129.
10. H. Tian, D. Liu, J. Li, M. Ma, Y. Lan, W. Wei, R. Niu and K. Song, New J. Chem., 2021, 45, 11253–11260.
11. M. Ma, J. Li, D. Liu, Y. Mei and R. Dong, ACS Appl. Mater. Interfaces, 2021, 13, 44615–44627.
12. H. Xia, Y. Tang, Y. Zhang, F. Ni, Y. Qiu, C.-W. Huang, C.-C. Wu and C. Yang, J. Mater. Chem. C, 2021, 10, 4614–4619.
13. K. Matsuo and T. Yasuda, Chem. Sci., 2019, 10, 10687–10697.
14. J. Liu, Y. Feng, H.-J. Yin, T. He, K.-Z. Wang, G. Long and C.-J. Yao, J. Phys. Chem. C, 2023, 127, 5567–5575.
15. Y. Mei, D. Liu, J. Li, R. Dong, M. Ma, W. Wei and Y. Lan, Mater. Today Chem., 2021, 22, 100645.
16. P. Rajamalli, D. Rota Martir and E. Zysman-Colman, J. Photonics Energy, 2018, 8, 032106.
17. M. Mahmoudi, J. Keruckas, D. Volyniuk, V. Andrulevičienė, R. Keruckienė, E. Narbutaitis, Y.-C. Chao, M. Rutkis and J. V. Grazulevicius, Dyes Pigm., 2021, 197, 109868.
18. N. Jürgensen, A. Kretzschmar, S. Höfle, J. Freudenberg, U. H. F. Bunz and G. Hernandez-Sosa, Chem. Mater., 2017, 29, 9154–9161.
19. S. O. Jeon, K. H. Lee, J. S. Kim, S.-G. Ihn, Y. S. Chung, J. W. Kim, H. Lee, S. Kim, H. Choi and J. Y. Lee, Nat. Photonics, 2021, 15, 208–215.
20. H. Zhuo, M. Zhang, H. Zhang, H. Lin, G. Yang, S. Tao, C. Zheng and X. Zhang, Chin. Chem. Lett., 2024, 36, 110760.
21. P. Chang, H. Wang, B. Zhao, Y. Man, Q. Qi, C. Han and H. Xu, J. Phys. Chem. C, 2024, 128, 10802–10809.
22. S. Sharma and S. Sengupta, Chem. – Eur. J., 2023, 30, e202303754.
23. S.-J. Woo, Y.-H. Ha, Y.-H. Kim and J.-J. Kim, J. Mater. Chem. C, 2020, 8, 12075–12084.
24. H. Zhang, L. Duan and D. Zhang, Chem. – Asian J., 2024, 20, e202401234.
25. Y.-Z. Li, F.-E. Szu, H.-Y. Szu, C.-C. Wu, Y.-Y. Zhang, Z.-H. Li, J.-H. Lee, T.-L. Chiu and M.-K. Leung, J. Mater. Chem. C, 2025, 13, 14527–14538.
26. Y. Duan, H. Sude, S. Pei, K. Zhang, L. Wen, S. Li and Z. Su, Spectrochim. Acta, Part A, 2025, 333, 125844.
27. T.-L. Wu, S.-H. Lo, Y.-C. Chang, M.-J. Huang and C.-H. Cheng, ACS Appl. Mater. Interfaces, 2019, 11, 10768–10776.
28. J. Li, M. Zhang, T. Li, D. Guo, T. Tian and H. Zhang, J. Mater. Chem. C, 2022, 10, 13124–13136.
29. S. Hirata, Y. Sakai, K. Masui, H. Tanaka, S. Y. Lee, H. Nomura, N. Nakamura, M. Yasumatsu, H. Nakanotani, Q. Zhang, K. Shizu, H. Miyazaki and C. Adachi, Nat. Mater., 2014, 14, 330–336.
30. D. R. Lee, M. Kim, S. K. Jeon, S.-H. Hwang, C. W. Lee and J. Y. Lee, Adv. Mater., 2015, 27, 5861–5867.
31. W.-W. Tao, K. Wang, J.-X. Chen, Y.-Z. Shi, W. Liu, C.-J. Zheng, Y.-Q. Li, J. Yu, X.-M. Ou and X.-H. Zhang, J. Mater. Chem. C, 2019, 7, 4475–4483.
32. Y. Mei, D. Liu, J. Li and J. Wang, J. Mater. Chem. C, 2022, 10, 16524–16535.
33. A. Khan, X. Tang, C. Zhong, Q. Wang, S. Y. Yang, F. C. Kong, S. Yuan, A. S. D. Sandanayaka, C. Adachi, Z. Q. Jiang and L. S. Liao, Adv. Funct. Mater., 2021, 31, 2009488.
34. Y. Mei, Y. Lan, D. Li, J. Wang, L. Xie, X. Peng, J. Li, D. Liu and S.-J. Su, Chem. Eng. J., 2023, 480, 148351.
35. J. Xu, X. Wu, J. Guo, Z. Zhao and B. Z. Tang, J. Mater. Chem. C, 2021, 9, 15505–15510.
36. Y. Chen, D. Zhang, Y. Zhang, X. Zeng, T. Huang, Z. Liu, G. Li and L. Duan, Adv. Mater., 2021, 23, 2103293.
37. Y. Mei, D. Liu, J. Li and J. Wang, Dyes Pigm., 2022, 207, 110701.
38. Q. Wu, J. Li, D. Liu, Y. Mei, B. Liu, J. Wang, M. Xu and Y. Li, Dyes Pigm., 2023, 217, 111421.
39. J. Yang, Z. Chen, M. Fang and Z. Li, Smart Mol., 2024, 2, e20240034.
40. M. K. Etherington, F. Franchello, J. Gibson, T. Northey, J. Santos, J. S. Ward, H. F. Higginbotham, P. Data, A. Kurowska, P. L. Dos Santos, D. R. Graves, A. S. Batsanov, F. B. Dias, M. R. Bryce, T. J. Penfold and A. P. Monkman, Nat. Commun., 2017, 8, 14987.
41. Y. Mei, J. Li, D. Liu, M. Xu, Y. Li and H. Wan, Adv. Opt. Mater., 2024, 13, 2402964.
42. Y. Li, Y. Mei, B. Liu, M. Xu, H. Sun, Q. Wu, H. Wan, D. Liu and J. Li, Dyes Pigm., 2023, 217, 111391.
43. Y. Guo, J. Zhao, L. Chen, H. Zhao, S. Li, Y. Liu, S. Yan and Z. Ren, Small, 2025, 21, 2502892.
44. Y. Mei, D. Liu, J. Li, M. Xu, J. Wang, J. Jin, L. Xie and H. Wan, Smart Mol., 2024, e20240055.
45. H. Min, I. S. Park and T. Yasuda, Adv. Opt. Mater., 2022, 10, 2200290.
46. T. Ohsawa, H. Sasabe, T. Watanabe, K. Nakao, R. Komatsu, Y. Hayashi, Y. Hayasaka and J. Kido, Adv. Opt. Mater., 2019, 7, 1801282.

---