Regulation of aggregation-enhanced thermally activated delayed fluorescence in butterfly-shaped donor–acceptor conjugates

Abstract

We studied how ether-linked benzophenone and dibenzofuran/dibenzothiophene-functionalized benzophenone as auxiliary groups in carbazole–phthalonitrile conjugates influence aggregation-enhanced thermally activated delayed fluorescence (AE-TADF). CDBFPN and CDBTPN exhibited AE-TADF due to their unique butterfly-shaped geometries. In contrast, CBPN faces aggregation-caused quenching, underscoring the critical role of these functionalized groups in boosting the blue TADF properties.

---

Organic luminescent materials that utilize both singlet and triplet excitons hold great potential for organic light-emitting diodes^1,2 and photosensitizers.³ Yet, weak spin–orbit coupling, luminescence quenching in the aggregated state, and nonradiative decay hinder efficient triplet harnessing under ambient conditions.⁴ Thermally activated delayed fluorescence (TADF) offers a solution by enabling triplet excitons to transform into radiative singlet states through reverse intersystem crossing (RISC).^1,5–7 This process relies on a small singlet–triplet energy gap (ΔE_ST) and orbital transformation via vibronic coupling. However, TADF emitters face challenges such as low RISC rates and aggregation-caused quenching (ACQ) at high doping concentrations, leading to inevitable triplet quenching due to triplet–triplet, and triplet–polaron annihilation.⁸ In addition, developing blue TADF emitters is particularly difficult due to the band-gap principle.^9,10

On the other hand, aggregation-induced emission (AIE) offers a promising solution to counteract ACQ.^9,11 AIE emitters typically feature flexible, twisted molecular structures that minimize intermolecular π⋯π interactions in the aggregated state, reducing emission quenching and exciton annihilation. In recent years, several TADF emitters with AIE properties have been developed, making the creation of aggregation-induced TADF enhancement emitters a desirable goal.^12–14 In the past decade, a wide variety of carbazole-based TADF emitters^6,15 were reported, but they face the ACQ effect, which limits their applications. In this work, we studied carbazole–phthalonitrile-based D–A–D conjugates, namely CBPN, CDBFPN, and CDBTPN, appended with benzophenone (BP) and dibenzofuran/dibenzothiophene-functionalized benzophenone (DBF-BP/DBT-BP) groups through ether linkages, resulting in twisted geometries resembling an open butterfly (Fig. 1). Through careful molecular design, we demonstrate that all these conjugates show efficient sky-blue TADF by fine-tuning the orbital nature for vibronic coupling and low energy gaps among S1, T1 and T2. Interestingly, photophysical analysis of CDBFPN and CDBTPN revealed an increase in the lifetimes of the prompt and delayed components with enhanced quantum yields (∼20-fold) due to aggregation-enhanced TADF (AE-TADF) caused by the introduction of DBF/DFT units resulting in unique packing, while CBPN suffers from the ACQ effect. Single-crystal X-ray diffraction (SCXRD) revealed that CDBTPN exhibited negligible π⋯π interactions, unlike CBPN, where such interactions likely cause ACQ. This design strategy may provide valuable insights for the development of next-generation AE-TADF materials.

Fig. 1 (a) Molecular design and (b) structures.

The final compounds were synthesized via two-step aromatic nucleophilic substitution reactions. All the compounds were characterized by NMR spectroscopy, high-resolution mass spectrometry, SCXRD, and HPLC analysis of the vacuum sublimated samples (ESI†).

We grew crystals of CBPN and CDBTPN using a binary solvent mixture of dichloromethane and hexane at room-temperature (RT), though CDBFPN failed to crystallize. SCXRD analysis revealed CBPN crystallizes in the centrosymmetric C2/c, while CDBTPN in the noncentrosymmetric Cc space group (Table S1, ESI†). Both adopt open, butterfly-like shapes with nearly orthogonal carbazole units, creating spatial separation of the HOMO and LUMO due to their twisted geometries (Fig. 2(a) and (b)). CBPN shows intermolecular π⋯π stacking between its carbazole fragments (3.82 Å), while CDBTPN lacks such interaction (Fig. S13, S14 and Table S2, ESI†). Furthermore, Hirshfeld surface analysis¹⁶ revealed key intermolecular interactions in the crystal structures, with C–H⋯π interactions significantly stabilizing both molecules. CBPN showed stronger π⋯π stacking (C⋯C: 10.5%) than CDBTPN (3.4%), while CDBTPN had higher N⋯H (12.1% vs. 7.0%) and unique S⋯H (5.1%) interactions absent in CBPN (Fig. 2(c)–(f) and Fig. S15, Table S3, ESI†).

Fig. 2 Single crystal X-ray structure of (a) CBPN and (b) CDBTPN; surfaces (mapped over d_norm) of (c) CBPN and (d) CDBTPN; and decomposed fingerprint plots showing the proportion of intermolecular C⋯C interactions in (e) CBPN and (f) CDBTPN.

We analyzed the redox properties of CBPN, CDBFPN, and CDBTPN using cyclic voltammetry in dry DMF, referencing the ferrocene/ferrocenium (Fc/Fc⁺) couple.¹⁷ Oxidation peaks appeared at +0.69 V, +0.73 V and +0.65 V, respectively, corresponding to HOMO levels of −5.49 eV, −5.53 eV, and −5.45 eV. The LUMO levels and band gaps were −3.10 eV (CBPN), −3.12 eV (CDBFPN), and −3.09 eV (CDBTPN) leading to nearly similar band gaps (∼2.4 eV), indicating that all compounds possess comparable electronic structures (Fig. S16 and Table S4, ESI†). Furthermore, TGA and DSC confirmed high thermal stability, with 5% weight loss occurring at 445–460 °C and glass transition temperatures between 300–316 °C, highlighting their durability and potential for optoelectronic applications (Fig. S17 and S18, ESI†).

We used quantum chemistry calculations (DFT/TD-DFT, M06-2X/6-31G(d,p))¹⁸ to investigate the molecule's electronic properties. Ground-state geometry optimization revealed HOMO localization on the carbazole (Cz) units and LUMO localization on the phthalonitrile (PN) segments. The BP, DBF-BP, and DBT-BP parts showed no contribution due to Cz's stronger donor strength. The minimal HOMO–LUMO overlap led to low oscillator strength and a small ΔE_S1–T1 (Fig. S19, ESI†). The calculated vertical excitation energies aligned closely with the experimental absorption spectra, validating the reliability of the computational model (Fig. S1, ESI†). Hole–electron analysis¹⁹ confirmed the CT character of S0 → S1 transition, driven by the H → L transition. S0 → T1 and S0 → T2 transitions revealed localized triplet state-dominated PN and keto groups, respectively (Fig. 3, Fig. S21, S22 and Table S5, ESI†). Small gaps (ΔE_S1–T1: 0.10–0.11 eV; ΔE_T1–T2: 0.2–0.21 eV) suggest efficient RISC through orbital transformation following the El-Sayed rule and vibronic coupling (Fig. S22, ESI†).

Fig. 3 Hole–electron analysis (NTOs: |h⁺〉, blue; 〈e⁻|, green) for transition orbitals corresponding to the S1, T1 and T2 excited states of CBPN at isosurface values of ±0.002 a.u. [M06-2X/6-31g(d,p)].

The ultraviolet-visible absorption spectra of CBPN, CDBFPN, and CDBTPN reveal strong bands at 300–350 nm, showing high molar absorptivity from π → π* transitions. Weaker CT bands appear at 390–450 nm, caused by donor–acceptor coupling.^7,20 A small shoulder in CDBFPN and CDBTPN at 310–340 nm can be attributed to the DBF-BP and DBT-BP segments, which are absent in CBPN (Fig. 4(a)). All compounds display negative solvatochromism in the lower-energy absorption bands, indicating ground state CT, and positive solvatochromism in the emission spectra, confirming CT in the excited state, which aligns with quantum chemistry calculations (Fig. 4(b) and Fig. S23, S24, ESI†). Interestingly, all three conjugates exhibit similar emission behavior in toluene (∼480 nm), indicating that CT emission predominantly originates from the Cz-PN part of the molecules (Fig. S25, ESI†). Deoxygenated conditions significantly boosted the PLQYs: 7.0% to 36.2% (CBPN), 6.1% to 41.5% (CDBFPN), and 5.2% to 45.0% (CDBTPN). Lifetime measurements revealed prompt and delayed components under deoxygenated conditions, which decreased with oxygen exposure, indicating triplet-state involvement (Fig. 4(c), Fig. S26 and Table S6, ESI†). Temperature-dependent fluorescence lifetime analysis under ambient aerated conditions revealed the endothermic nature of RISC (Fig. 4(d), Fig. S27 and Table S7, ESI†). Low ΔE_ST values at 77 K (CBPN: 0.17 eV, CDBFPN: 0.18 eV, CDBTPN: 0.11 eV) further validate these findings (Fig. S28, ESI†).

Fig. 4 (a) Absorption in toluene, and (b) solvatochromic emission of CBPN. (c) PL decays of CBPN in toluene at different conditions at RT and (d) temperature-dependent decays of CBPN in toluene (λ_ex = 370 nm; 10 μM).

The aggregation effect in CBPN, CDBFPN, and CDBTPN was studied by increasing the water fraction (f_w) in tetrahydrofuran (THF) solutions. CBPN showed a decrease in emission intensity as f_w increased, indicating ACQ. In contrast, both CDBFPN and CDBTPN exhibited weak emission in THF but displayed a blue-shifted emission with significantly higher intensity and quantum yields when f_w exceeded 50%. At 95% f_w, the PLQY was 48.2% for CDBFPN and 54.0% for CDBTPN, confirming strong AIE (Fig. 5 and Fig. S29, S30, ESI†). As the solvent polarity increases, both luminogens show a bathochromic shift due to the stabilization of the CT excited states. However, at f_w = 60–95%, aggregation restricted molecular motion, inducing a hypsochromic shift. Fluorescence lifetime measurements at f_w = 95% revealed increased delayed components and lifetimes as compared to THF (CDBFPN: A_DF, 10.28 to 29.99%, τ_DF: 290 to 560 ns; CDBTPN: A_DF, 12.77 to 28.45%; τ_DF: 360 to 650 ns), supporting AE-TADF (Fig. S31 and Table S8, ESI†).¹⁴ This effect is attributed to the aggregation of CDBFPN and CDBTPN in water, which restricts intramolecular motion and reduces nonradiative decay. The DBF-BP and DBT-BP moieties in the molecular backbone of CDBFPN and CDBFPN facilitated AE-TADF, unlike CBPN, which lacks these additional heteroatomic groups.¹³

Fig. 5 (a) PL spectra of CDBTPN in a THF/water mixture. (b) Relative PL intensities and PLQYs of CDBTPN at different f_w; inset: photographs (f_w = 0% → 95%) captured under UV light.

Given the exceptional photophysical properties observed in solution, we systematically studied the solid-state PL properties of these molecules in both neat and doped films, using 1,3-bis(N-carbazolyl)benzene (mCP) as a host material at varying doping concentrations (1, 10, 20, and 30 wt%). Films were prepared by spin-coating toluene solutions of the compounds and mCP onto quartz substrates. We analyzed and compared the emission behavior of the compounds in the films. In the 1 wt% doped film, CBPN emitted at ∼475 nm, while CDBFPN and CDBTPN emitted at ∼488 nm (Fig. 6(a)). A significant bathochromic shift of 25–30 nm was observed in CBPN with increasing doping concentration, whereas CDBFPN and CDBTPN showed smaller shifts (12–15 nm) (Fig. S32, ESI†). This suggests that CBPN's emission is more sensitive to host concentration, highlighting differences in their aggregation behaviors. As the doping concentration increased, CBPN's emission became more red-shifted, indicating a concentration-dependent effect. In contrast, CDBFPN and CDBTPN displayed more stable emission behavior with minor shifts, suggesting reduced π⋯π interactions. The neat films showed high PLQY values of 67% and 70% for CDBFPN and CDBTPN, respectively, demonstrating resistance to ACQ. CBPN had a lower PLQY of 33%, attributed to strong intermolecular π⋯π interactions between neighbouring Cz units (Fig. 6(b)). These findings indicate that CDBFPN and CDBTPN exhibit robust PL performance, while CBPN is more susceptible to ACQ due to its molecular structure. Furthermore, fluorescence lifetime analysis of both the neat and 1 wt% doped film revealed a bi-exponential decay pattern, characteristic of both τ_PF and τ_DF (Fig. 6(c), (d) and Fig. S33, S34, ESI†). In the 1 wt% doped film, the decay times were as follows: CBPN (τ_PF = 15.6 ns, τ_DF = 360 ns), CDBFPN (τ_PF = 18.4 ns, τ_DF = 430 ns), and CDBTPN (τ_PF = 16.3 ns, τ_DF = 580 ns). Neat films exhibited significantly longer decay times (CBPN: τ_PF = 21.2 ns, τ_DF = 810 ns; CDBFPN: τ_PF = 34.9 ns, τ_DF = 1030 ns; CDBTPN: τ_PF = 31.4 ns, τ_DF = 970 ns). These extended decay times suggest stabilization of the excited states through intermolecular interactions, supporting the role of aggregation in enhancing TADF.¹⁴ Temperature-dependent fluorescence decay analysis revealed a marked reduction in τ_DF with decreased contribution as the temperature decreased from 300 K to 77 K, confirming the involvement of TADF in all conjugates (Tables S9 and S10, ESI†). Additionally, we calculated k_RISC for both neat and doped films. The k_RISC values (1 wt% doped film) were 1.85 × 10⁷ s⁻¹ (CBPN),⁷ 1.70 × 10⁷ s⁻¹ (CDBFPN), and 1.75 × 10⁷ s⁻¹ (CDBTPN). In the neat films, these values were slightly lower: 1.48 × 10⁷ s⁻¹, 1.03 × 10⁷ s⁻¹, and 1.27 × 10⁷ s⁻¹, respectively. Higher k_RISC values in the doped film suggest faster RISC. Interestingly, the non-radiative decay rates were lower in the CDBFPN and CDBTPN neat films compared to CBPN, highlighting the more efficient PL behavior (Table S11, ESI†).

Fig. 6 (a) PL spectra of all the blend films (1 wt%). (b) PLQYs of blend samples with varying concentrations and neat films. Temperature-dependent PL decay curve of CDBTPN: (c) 1 wt% and (d) neat film (λ_ex = 370 nm).

In summary, we synthesized three carbazole–phthalonitrile conjugates with benzophenone and its derivatives as pendent arms through ether linkages. All compounds exhibited efficient blue TADF due to small energy gaps (S1–T1, T1–T2) that favor orbital transitions. Dibenzofuran and dibenzothiophene functionalized CDBFPN and CDBTPN demonstrated aggregation-enhanced TADF, with high quantum yields. This enhancement results from reduced non-radiative decay pathways, attributed to diminished π⋯π interactions in the butterfly-like twisted geometries. Further modification of these systems and their device performance are underway.

D. R. is grateful to the Science & Engineering Research Board (SERB) (CRG/2022/000128), Department of Science and Technology (DST), India, the Anusandhan National Research Foundation (ANRF) (ANRF/IRG/2024/000038/CS) and Shiv Nadar Institution of Eminence (SNIoE). M. U. thanks SNIoE for the fellowship. The authors acknowledge the use of the high-performance computing cluster ‘MAGUS’ at SNIoE for providing computational resources.

Data availability

The data supporting this article have been included as part of the ESI† and can be obtained from https://doi.org/10.1039/D5CC00015G.

Conflicts of interest

There are no conflicts to declare.

Notes and references

1. (a) S. Dey, M. Hasan, A. Shukla, N. Acharya, M. Upadhyay, S.-C. Lo, E. B. Namdas and D. Ray, J. Phys. Chem. C, 2022, 126, 5649–5657; (b) P. L. Dos Santos, J. S. Ward, D. G. Congrave, A. S. Batsanov, J. Eng, J. E. Stacey, T. J. Penfold, A. P. Monkman and M. R. Bryce, Adv. Sci., 2018, 5, 1700989.
2. H. Noda, X.-K. Chen, H. Nakanotani, T. Hosokai, M. Miyajima, N. Notsuka, Y. Kashima, J.-L. Brédas and C. Adachi, Nat. Mater., 2019, 18, 1084–1090.
3. M. A. Bryden and E. Zysman-Colman, Chem. Soc. Rev., 2021, 50, 7587–7680.
4. N. J. Turro, V. Ramamurthy and J. C. Scaiano, Principles of molecular photochemistry: an introduction, University Science Books, 2009.
5. (a) N. Acharya, M. Upadhyay, R. Deka and D. Ray, J. Phys. Chem. C, 2024, 128, 8750–8758; (b) R. Deka, S. Dey, M. Upadhyay, S. Chawla and D. Ray, J. Phys. Chem. A, 2024, 128, 581–589.
6. (a) I. Bhattacharjee, N. Acharya, H. Bhatia and D. Ray, J. Phys. Chem. Lett., 2018, 9, 2733–2738; (b) H. Uoyama, K. Goushi, K. Shizu, H. Nomura and C. Adachi, Nature, 2012, 492, 234–238.
7. S. Dey, R. Deka, M. Upadhyay, S. Peethambaran and D. Ray, J. Phys. Chem. Lett., 2024, 15, 3135–3141.
8. M. A. Baldo, C. Adachi and S. R. Forrest, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 62, 10967.
9. (a) H. Bhatia, I. Bhattacharjee and D. Ray, J. Phys. Chem. Lett., 2018, 9, 3808–3813; (b) H. Bhatia and D. Ray, J. Phys. Chem. C, 2019, 123, 22104–22113.
10. E. Tankelevičiūtė, I. D. Samuel and E. Zysman-Colman, J. Phys. Chem. Lett., 2024, 15, 1034–1047.
11. (a) H. Bhatia, S. Dey and D. Ray, ACS Omega, 2021, 6, 3858–3865; (b) Y. Hong, J. W. Lam and B. Z. Tang, Chem. Soc. Rev., 2011, 40, 5361–5388; (c) S. Karmakar, S. Dey, M. Upadhyay and D. Ray, ACS Omega, 2022, 7, 16827–16836; (d) J. Mei, N. L. Leung, R. T. Kwok, J. W. Lam and B. Z. Tang, Chem. Rev., 2015, 115, 11718–11940.
12. (a) H. Dai, Y. Liang, X. Long, T. Tang, H. Xie, Z. Ma, G. Li, Z. Yang, J. Zhao and Z. Chi, Chem. Sci., 2025, 16, 156–162; (b) Y. Matsui, Y. Yokoyama, T. Ogaki, K. Ishiharaguchi, A. Niwa, E. Ohta, M. Saigo, K. Miyata, K. Onda and H. Naito, J. Mater. Chem. C, 2022, 10, 4607–4613; (c) F. Di Maiolo, D. A. P. Huu, D. Giavazzi, A. Landi, O. Racchi and A. Painelli, Chem. Sci., 2024, 15, 5434–5450; (d) Y.-Y. Hsieh, R. S. Sánchez, G. Raffy, J.-J. Shyue, L. Hirsch, A. Del Guerzo, K.-T. Wong and D. M. Bassani, Chem. Commun., 2022, 58, 1163–1166.
13. J. Guo, X. L. Li, H. Nie, W. Luo, S. Gan, S. Hu, R. Hu, A. Qin, Z. Zhao and S. J. Su, Adv. Funct. Mater., 2017, 27, 1606458.
14. W. Zhang, S. Li, Y. Gong, J. Zhang, Y. Zhou, J. Kong, H. Fu and M. Zhou, Angew. Chem., Int. Ed., 2024, 63, e202404978.
15. (a) D. Liu, J. Y. Wei, W. W. Tian, W. Jiang, Y. M. Sun, Z. Zhao and B. Z. Tang, Chem. Sci., 2020, 11, 7194–7203; (b) B. Wex and B. R. Kaafarani, J. Mater. Chem. C, 2017, 5, 8622–8653.
16. P. R. Spackman, M. J. Turner, J. J. McKinnon, S. K. Wolff, D. J. Grimwood, D. Jayatilaka and M. A. Spackman, J. Appl. Crystallogr., 2021, 54, 1006–1011.
17. (a) C. M. Cardona, W. Li, A. E. Kaifer, D. Stockdale and G. C. Bazan, Adv. Mater., 2011, 23, 2367–2371; (b) N. Elgrishi, K. J. Rountree, B. D. McCarthy, E. S. Rountree, T. T. Eisenhart and J. L. Dempsey, J. Chem. Educ., 2018, 95, 197–206.
18. M. J. Frisch, et al., Gaussian 16, Revision C.01. 2016, Gaussian Inc., Wallingford, CT, 2016.
19. T. Lu and F. Chen, J. Comput. Chem., 2012, 33, 580–592.
20. S. Dey, A. K. Pal, M. Upadhyay, A. Datta and D. Ray, J. Phys. Chem. B, 2023, 127, 9833–9840.

---

Footnote

† Electronic supplementary information (ESI) available: Synthesis, characterization, SCXRD, electrochemical, thermal, computational analysis and PL studies provided. CCDC 2413934 and 2413935. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d5cc00015g

---