Saturated carbazole-embedded BN-aromatic systems as narrowband sky-blue emitters

Abstract

Due to their rigid structure and short-range charge transfer emissive excited states, multiresonant thermally activated delayed fluorescence (MR-TADF) emitters provide exceptional color purity due to their narrowband emission. Many examples are based on nitrogen/boron-doped polycyclic aromatic hydrocarbons. The emission color is in part modulated by the strength of the donor heterocycle. In an effort to shift the emission to the blue, we explore in this study the partial saturation of the carbazole donor. We designed four emitters-tButHCzB, DtButHCzB, SpAc-tButHCzB, and SpAc-DtButHCzB featuring tetrahydrocarbazole donor moieties, which disrupt the planarity and reduce the conjugation length. Theoretical calculations predict moderate singlet–triplet excited-state energy gaps (ΔE_ST = 250–270 meV) and low-lying T₂ and T₃ states, suggesting that these compounds should exhibit TADF. In toluene, the four compounds emit in the sky-blue region (λ_PL = 476–480 nm), with narrow full-width at half-maximum (FWHM) values (18–21 nm) and high photoluminescence quantum yields (Φ_PL) of 86–91%. In 1 wt% doped films in mCP, however, their Φ_PL drops to between 40–45% and their emission slightly red-shifts to λ_PL of 483–491 nm. Temperature-dependent time-resolved PL measurements confirm that these compounds exhibit TADF. This study offers valuable insights into a region of chemical space scantly explored in MR-TADF emitter design.

---

Introduction

Thermally activated delayed fluorescence (TADF) emitters have emerged recently as an alternative and promising class of materials to phosphorescence complexes for use in organic light-emitting diodes (OLEDs).¹ This is largely due to their comparable exciton utilization efficiency in electroluminescent devices to phosphorescent materials. A subclass of TADF emitters, so-called multiresonant TADF (MR-TADF) compounds, provide desirable narrowband emission required for high-definition displays, due to their rigid structure typically based on a p- and n-doped polycyclic aromatic hydrocarbon skeleton. The regiochemistry of the p- and n-dopants produces an excited state of short-range charge-transfer (SRCT) character that results in a moderately small singlet–triplet excited-state energy gap, ΔE_ST, on the order of 200 meV.^1–4 In particular, to meet the blue BT.2020 industry standard, narrowband emission is important both to reduce the amount of UV-light that is being emitted and to decrease the excitation energy that results in greater photochemical stability of the emitter material.^5,6

OLEDs are typically fabricated by vacuum deposition. An alternative fabrication methodology is to use solution-processing technologies, which offer the promise of more cost-effective and less material-wasteful manufacture of the device.^7–10 This requires the use of a solution-processable emitter, which necessitates that this material be suitably soluble in the organic solvents used to process it into a film and that the resulting film be homogeneous.^1,11 A primary strategy to achieve greater solubilities in these rigid, planar molecules, is to disrupt their planarity through the introduction of bulky substituents and/or to increase the conformational flexibility by saturating the rings within the core structure, the latter strategy of which has not been extensively explored. We previously documented that the incorporation of mesityl groups in Mes₃-DiKTa¹² increases the solubility of this compound in chlorobenzene (Mes₃-DiKTa: 20 mg mL⁻¹; DiKTa: 12.5 mg mL⁻¹) and mitigates aggregation-caused quenching (ACQ) in spin-coated films and in the vacuum-deposited device. This substitution also causes a red-shift in the emission from λ_PL of 453 nm for the parent compound DiKTa (aka QAO¹³) to 468 nm for Mes₃-DiKTa in dilute toluene solution. This is due to the inductively electron-withdrawing nature of the mesityl substituents. Wu et al. likewise demonstrated how decorating the DABNA-1¹⁴ core with methyl, mesityl, and adamantyl substituents in A-BN, DA-BN, and A-DBN¹⁵ reduces ACQ, and maintains the blue emission in A-BN with a λ_PL of 462 nm. Yet additional substitutions to prevent quenching even further lead to a red-shifted emission from λ_PL of 462 nm for DABNA-1 to 483 nm for DA-BN and 482 nm for A-DBN in dilute dichloromethane solution.

A different approach was taken by Du et al. for TImBN¹⁶ (Fig. 1) in which they truncated the carbazole units of CzBN¹⁷ and incorporated instead 2,3-dimethyl-1H-indole moieties. By truncating the aromatic system and effectively decreasing the conjugation length, this compound shows a slightly blue-shifted emission peaking at λ_PL of 475 nm compared to DtBuCzB¹⁸ (λ_PL = 481 nm) in dilute toluene solution. However, TImBN possesses a significantly larger ΔE_ST of 390 meV compared to DtBuCzB (ΔE_ST = 130 meV),¹⁸ which makes the thermally promoted exciton upconversion kinetics much less favourable. The authors also demonstrated that a blue-shift in the emission to λ_PL of 470 nm in dilute toluene could be induced by attaching a 3,6-di-tert-butylcarbazole donor moiety para to the boron atom on the central ring. However, the T₁ level remains unaffected by this substitution, leading to an even larger ΔE_ST of 420 meV, which means that this compound is unlikely to emit by TADF, despite the authors’ claims to the contrary.

Fig. 1 Chemical structures of relevant emitters in the literature^16,18–20 and the four novel emitters DtButHCzB, tButHCzB, SpAc-DtButHCzB and SpAc-tButHCzB introduced in this study.

Here, we investigate an approach to introduce saturation into one of the rings of the carbazole moieties of the parent DtBuCzB emitter in order to blue-shift the emission and increase the solubility to make these derivatives compatible for solution-processed OLEDs (SP-OLEDs). We designed and synthesized four emitters, tButHCzB, DtButHCzB, SpAc-tButHCzB, and SpAc-DtButHCzB, where we replaced the 3,6-di-tert-butyl carbazole with either 6-(tert-butyl)-2,3,4,9-tetrahydrocarbazole (tButHCz) or 3,6-di-tert-butyl-2,3,4,9-tetrahydrocarbazole (DtButHCz) donor moieties (Fig. 1). By saturating one of the phenyl rings of the carbazole, this moiety becomes less planar, and the size of the conjugated system decreases, hence promoting an expected and desired blue-shift of the emission. Two derivatives of these two compounds, SpAc-tButHCzB and SpAc-DtButHCzB, also contain a spiro[acridine-9,9′-fluorene] donor moiety positioned para to the boron atom on the central benzene ring to promote a further blue-shift of the emission. Theoretical calculations at the RI-wPBEPP866/cc-pVDZ level of theory predict moderate ΔE_ST of 250–270 meV and close-lying T₂ and T₃ states that suggest that these four compounds should be TADF. In dilute toluene solution, the four emitters show sky-blue photoluminescence (PL) at λ_PL of 476–480 nm and very narrowband emission with full-width half maxima (FWHM) of 18–21 nm. All four emitters have high photoluminescence quantum yields, Φ_PL, of 86–91%, though no delayed emission was detected in solution. The photophysics of the spin-coated 1 wt% doped films in 1,3-bis(N-carbazolyl)benzene (mCP) revealed an undesired red-shifting and broadening of the emission band (λ_PL of 483–491 nm and FWHM of 26–45 nm), ascribed to the formation of π-stacked aggregates, even at this low doping concentration, that are also present in the single crystal X-ray structures. This is accompanied by a noted significant decrease in the Φ_PL to 40–45%, thus decreasing their value as emitters for OLEDs. This study reveals for the first time, structure–property relationships in MR-TADF emitters containing partially saturated N-heterocyclic donor moieties.

Results and discussion

Synthesis and characterisation

The synthesis of tButHCzB and DtButHCzB was carried out in three steps as shown in Scheme 1, and for SpAc-tButHCzB and SpAc-DtButHCzB in four steps. The first step for all four emitters was a Fisher-indole synthesis using (4-(tert-butyl)phenyl)hydrazine hydrochloride and 4-(tert-butyl)cyclohexan-1-one or cyclohexan-1-one to construct the tetrahydrocarbazole derivative 1 or 2 in yields of 74 and 90%, respectively. For tButHCzB and DtButHCzB, an S_NAr reaction between 2-bromo-1,3-difluorobenzene and two equivalents of the respective tetrahydrocarbazole gave intermediates 3 and 4 in 45 and 20% yields, respectively. Spiro-acridine-containing derivatives SpAc-tButHCzB and SpAc-DtButHCzB were obtained by first reacting 1 and 2 with 5-bromo-2-chloro-1,3-difluorobenzene under S_NAr conditions to give intermediates 5 and 6 in 77 and 71% yields, respectively, followed by a Buchwald–Hartwig coupling with 10H-spiro[acridine-9,9′-fluorene] to afford 7 and 8 in 89 and 71% yields, respectively. The final step for all four emitters consisted of a one-pot borylation using tert-butyl lithium, boron tribromide and diisopropylethylamine¹⁴ to afford the target emitters tButHCzB, DtButHCzB, SpAc-tButHCzB and SpAc-DtButHCzB in 13, 22, 29, and 39%, respectively. Their structure and purity were verified using a combination of ¹H NMR and ¹³C NMR spectroscopy, high-resolution mass spectrometry, elemental analysis, single-crystal XRD, and melting point determination. Emitters containing only two tert-butyl groups have significantly lower melting points (tButHCzB and SpAc-tButHCzB showed melting points of 315–318 and 330–334 °C, respectively) than their analogues with four tert-butyl groups (DtButHCzB and SpAc-DtButHCzB showed melting points of 357–361 and >400 °C, respectively). This is not only due to the increased molecular weight but also that the second tert-butyl group locks the conformation of the cyclohexenyl ring. In thermogravimetric analysis a different trend is observed, here tButHCzB has the lowest degradation temperature, T_d at 5% weight loss (354 °C) followed by DtButHCzB (375 °C). Surprisingly tButHCzB shows the highest T_d of 497 °C (Fig. S9).

Scheme 1 Synthesis of tButHCzB, DtButHCzB, SpAc-tButHCzB, and SpAc-tButHCzB.

X-ray crystal structures

Crystals of three emitters, tButHCzB, SpAc-DtButHCzB, and SpAc-tButHCzB, suitable for single crystal diffraction were grown by a slow diffusion of antisolvent (methanol or iso-propanol) into a saturated solution of the emitter in either dichloromethane or toluene (SI for details). Crystals of DtButHCzB were also obtained but were poorly diffracting and gave unsatisfactory data. A solution from this data did show connectivity consistent with the expected structure (Fig. S1). The structures obtained are consistent with the proposed molecular structures (Fig. 2 and Fig. S2–S4), with SpAc-tButHCzB, and SpAc-DtButHCzB crystallising as toluene and hexane solvates, respectively. The phenylborane-bridged tetrahydrocarbazole core in all three structures is generally planar, showing a slight cupping angle between the pyrrole rings (6.33(13)–7.77(6)°). The fused cyclohexene ring of the tetrahydrocarbazole adopts a half-chair conformation in all cases. In the structure where the cyclohexene ring is substituted, SpAc-DtButHCzB, the C3 carbon of the tetrahydrocarbazole moiety is oriented either above or below the plane, with the tert-butyl group in a pseudo-equatorial position. For both SpAc-DtButHCzB and SpAc-tButHCzB, the acridine unit is oriented approximately perpendicularly to the central benzene ring (angle between benzene and acridine central ring 80.10(6) and 83.46(9)°, respectively), with the acridine itself bent (angle between benzene rings of each acridine 14.88(7) and 22.19(10)°, respectively), and the fluorene unit is arranged nearly orthogonal (angle between central rings of acridine and fluorene 87.13(8) and 88.79(10)°, respectively).

Fig. 2 Face-to-face and edge-to-face (green dashed lines) stacking in the crystal structures of (a) SpAc-DtButHCzB, (b) SpAc-tButHCzB, and (c) tButHCzB. Hydrogen atoms, minor parts of disorder, and solvates (except for one toluene in SpAc-tButHCzB) are omitted for clarity.

In the structures of tButHCzB and SpAc-DtButHCzB, the planar fused ring systems stack face-to-face (centroid⋯centroid 3.5304(10) and 3.7983(16)°, respectively) (Fig. 2), with SpAc-DtButHCzB showing additional edge-to-face packing between adjacent spiro-acridine moieties (Fig. 2(a)) (centroid⋯H 2.6758(8) Å). Edge-to-face packing is prevalent in the packing of SpAc-tButHCzB with the spiro-acridine moieties forming interactions to both the fused ring core and adjacent spiro-acridines (centroid⋯H 2.6938(11) and 3.0051(12) Å, respectively), while the toluene solvents occupy the space (plane⋯plane 7.238(6) Å) between coplanar molecules (Fig. 2(b)). Face-to-face packing arrangements are generally considered unfavourable for light-emitting materials, as π–π interactions can promote non-radiative decay pathways, thereby reducing the photoluminescence quantum yield (Φ_PL). The observed packing in tButHCzB and SpAc-DtButHCzB indicates potentially strong π–π interactions, which are expected to contribute to aggregation-caused quenching. While the structure of SpAc-tButHCzB does not adopt a face-to-face alignment, this may be a result of inclusion of the toluene solvate between parallel-aligned SpAc-tButHCzB molecules. As such the tButHCz moieties could still facilitate aggregation and increase non-radiative decay in the solid state.²¹

Theoretical calculations

We undertook a computational study to understand the impact of saturation of the carbazole donor on the optoelectronic properties, first evaluating the ground-state geometry and orbital energies at the PBE0/6-31G(d,p) level of theory²² using a toluene solvent polarity field. The predicted HOMO and LUMO energy levels for DtButHCzB and tButHCzB are −5.28/−5.29 eV and −1.66/−1.67 eV and for SpAc-DtButHCzB and SpAc-tButHCzB are −5.36/−5.35 eV and −1.81/−1.80 eV (Fig. 3). The predicted stabilization of the HOMO and especially the LUMO levels for the donor-substituted derivatives is an unexpected anomaly, since the presence of the donor should mesomerically interact with the boron acceptor and lead to a destabilization of both the HOMO and especially the LUMO; this behavior was also predicted at both the PBE0/6-31G(d,p) and M06-2X/6-31G(d,p) levels in the gas phase (Fig. S5 and S6). This is in contrast to DtBuCzB (HOMO = −5.06 eV; LUMO = −1.71 eV)²³ and tDPA-DtCzB (HOMO = −4.93 eV; LUMO = −1.55 eV)²³ where there is a predicted destabilization of both the HOMO and LUMO with donor substitution para to the boron atom. Both DtButHCzB and tButHCzB are thus predicted to have larger HOMO–LUMO gaps (ΔE) of 3.62 eV, while SpAc-DtButHCzB and SpAc-tButHCzB are predicted to have smaller ΔE of 3.45 eV each. In comparison to the structurally related compounds DtBuCzB and TMInBN (both having ΔE = 3.66 eV), DtButHCzB and tButHCzB have a modestly smaller ΔE.

Fig. 3 Theoretical calculations (a) HOMO and LUMO energy level and orbital density plots (iso value 0.02) at the PBE0/6-31G(d,p) (toluene) level of theory. (b) Excited-state energy levels calculated at the RI-wPBEPP86/cc-pVDZ level of theory in the gas phase and corresponding difference density plots (yellow = positive; blue = negative; isovalue 0.001).

It has been shown that TD-DFT calculations do not provide accurate predictions for the energies of the low-lying excited states nor for the energy gap between S₁ and T₁ (ΔE_ST) for compounds that possess an SRCT emissive S₁ state.²⁴ Therefore, we computed the energies of the excited states at the RI-wPBEPP86/cc-pVDZ level of theory (Fig. 3(b)), which has been demonstrated to provide comparable accuracy to high-level coupled cluster calculations.^25,26 Double-hybrid density functional methods were employed in this study, as coupled cluster calculations were found to be computationally too demanding due to the molecular flexibility and the number of heavy atoms in these compounds. These calculations predict similar S₁ and T₁ levels of 3.18/3.19 eV and 2.93/2.93 eV for DtButHCzB and tButHCzB, respectively. This results in moderate ΔE_ST values of 0.25 and 0.26 eV, respectively. A high oscillator strength, f, of 0.51 and 0.49 is predicted for DtButHCzB and tButHCzB, respectively, suggesting that these two compounds will have high Φ_PL. The S₁ and T₁ energy levels of SpAc-DtButHCzB and SpAc-tButHCzB show similarly close values to each other at 3.15/3.15 eV and 2.88/2.89 eV, respectively. Both S₁ and T₁ are stabilized in SpAc-DtButHCzB and SpAc-tButHCzB as compared to DtButHCzB and tButHCzB, and their ΔE_ST and f values are comparable at 0.26 and 0.27 eV, and 0.48 and 0.50, respectively. All four emitters possess T₂ and T₃ states that are slightly higher in energy than their corresponding S₁ states. It has been shown that higher-lying triplet states that are close in energy to S₁ can participate in the reverse intersystem crossing (RISC) process and promote TADF in compounds with seemingly large ΔE_ST.²⁷ Difference density plots show the expected alternating pattern between density gain (yellow) and density loss (blue) for MR-TADF emitters that indicate the SRCT character of these emitters (Fig. 3(b)). For SpAc-tButHCzB and SpAc-DtButHCzB, the higher lying excited states S₂, S₃ and T₂ and T₃ are not predicted to possess LRCT character between the spiro-acridine (donor) and the MR-TADF core (acceptor) (Table S3). When compared to the unsaturated reference DtBuCzB¹⁸ (S₁ = 3.14 eV; T₁ = 3.02 eV; ΔE_ST = 0.12 eV calculated at the same level of theory), DtButHCzB and tButHCzB are predicted to show a blue-shifted emission and larger ΔE_ST (Fig. S7). By contrast, a decrease in S₁ energy is predicted compared to the structurally related TMInBN¹⁶ (S₁ = 3.26 eV; T₁ = 2.94 eV; ΔE_ST = 0.32 eV calculated at the same level of theory, Fig. S7). This stabilization can be assigned to the relatively stronger donating strength (shallower HOMO level) of the tetrahydrocarbazole derivatives as compared to the dimethylindole donor used in TMInBN (Table S2). However, the T₁ energy levels for tButHCzB and DtButHCzB (both 2.93 eV) are comparable to that of TMInBN (T₁ = 2.94 eV), resulting in a smaller ΔE_ST, which is expected to lead to more efficient TADF (Fig. S7). A similar trend is observed for the donor-decorated congeners SpAc-DtButHCzB and SpAc-tButHCzB as compared to their unsaturated counterpart SpAc-tCzBN^19,20 (S₁ = 3.11 eV; T₁ = 3.07 eV; ΔE_ST = 0.04 eV calculated at the same level of theory, Fig. 1) where the S₁ states in the former two are destabilized while their T₁ states are stabilized. Therefore, SpAc-DtButHCzB and SpAc-tButHCzB are predicted to possess much larger ΔE_ST of 260 and 270 meV, respectively, as compared to SpAc-tCzBN (Fig. S8).

Photophysical properties

We first investigated the photophysical properties of all four emitters in dilute toluene solutions (∼10⁻⁶ M) (Table 1). Their absorption spectra possess a sharp, intense, lowest energy absorption band associated with the S₀–S₁ SRCT transition peaking at λ_abs of 467, 468, 464 and 465 nm, with associated molar extinction coefficients, ε, of 44, 44, 47, 47 × 10⁴ M⁻¹ cm⁻¹ for tButHCzB, DtButHCzB, SpAc-tButHCzB and SpAc-DtButHCzB, respectively (Fig. 4(a)). There is a small hypsochromic shift of this absorption band for the donor decorated derivatives, SpAc-tButHCzB and SpAc-DtButHCzB that is consistent with their slightly larger ΔE (Table 3). The SRCT bands of these four saturated analogues are of similar energy compared to their respective bands in the unsaturated counterparts, DtBuCzB and SpAc-tCzBN. The four compounds have small Stokes shifts of around 12 nm and narrowband emission (FWHM, of 21, 21, 19, and 18 nm) that are reflective of their rigid structures. The PL peaks at λ_PL of 479, 480, 476, and 477 nm for tButHCzB, DtButHCzB, SpAc-tButHCzB, and SpAc-DtButHCzB, respectively (Fig. 4(b)). Solvatochromism studies of the absorption and PL confirmed the SRCT character of the emissive excited state in all four derivatives (Fig. S11 and S12).

Table 1 Photophysical properties

Compound	Medium	λabs(ε)^c/nm (10^4 M^−1 cm^−1)	λPL^d/nm	FWHM^e/nm	ES1^f/eV	ET1^g/eV	ΔEST^h/eV	ΦPL^i/%	τp^j/ns	τd^k/ms
a In dilute toluene solution (∼10^−6 M). λexc = 380 nm.b Spin-coated thin films of 1 wt% emitter in mCP. λexc = 340 nm.c Lowest energy absorbance peak; molar extinction coefficient ε in parentheses.d PL maximum.e Full-width half maximum of the emission band.f S1 energy level determined from the onset of the steady-state PL spectrum in 2-MeTHF glass at 77 K. λexc = 380 nm.g T1 energy level determined from the onset of the time-gated emission spectra (10–85 ms) in 2-MeTHF glass at 77 K. λexc = 380 nm.h ΔEST = ES1 − ET1.i Photoluminescence quantum yield; absolute ΦPL of the thin films was measured using an integrating sphere; relative ΦPL in solutions was measured by a comparative method using quinine sulfate as the reference (Φr = 54.6% in 1 N H2SO4).^35j Prompt lifetimes were measured by TCSPC with a 200 ns time window. λexc = 375 nm.k Delayed lifetimes were measured by MCS with a 1 s time window. λexc = 340 nm.
tButHCzB	Sol.^a	467 (35)	479	21	2.62	2.30	0.32	91/76	7.5	—
	Film^b		491	45	2.65	2.34	0.31	61/42	5.3	95
DtButHCzB	Sol.^a	468 (39)	480	21	2.61	2.30	0.31	87/69	7.1	—
	Film^b		490	34	2.63	2.31	0.32	57/45	5.5	154
SpAc-tButHCzB	Sol.^a	464 (44)	476	19	2.66	2.29	0.37	86/69	7.2	—
	Film^b		483	27	2.67	2.32	0.35	53/43	5.2	108
SpAc-DtButHCzB	Sol.^a	465 (47)	477	18	2.64	2.26	0.35	90/77	6.2	—
	Film^b		483	26	2.66	2.31	0.35	53/44	6.3	148

---

Fig. 4 (a) Molar absorptivity in dilute toluene solution (10⁻⁶ M). (b) Steady-state PL spectra in dilute toluene solution (10⁻⁶ M; λ_exc = 380 nm). (c) Steady-state PL spectra of 1 wt% doped films in mCP (λ_exc = 340 nm). (d) Photos of spin-coated films under UV-irradiation of SpAc-DtButHCzB, SpAc-tButHCzB, DtButHCzB and tButHCzB.

Time-resolved photoluminescence (TRPL) measurements in degassed toluene solution (10⁻⁶ M) reveal no delayed emission, which is not unusual for MR-TADF compounds possessing moderately large ΔE_ST.^28–31 The PL lifetimes, τ_PL, are 7.2, 6.2, 7.5, and 7.1 ns for SpAc-tButHCzB, SpAc-DtButHCzB tButHCzB, and DtButHCzB, respectively. In degassed toluene solution tButHCzB, DtButHCzB, SpAc-tButHCzB and SpAc-DtButHCzB have high Φ_PL of 91, 87, 86, and 90%, which decrease to 76, 69, 69, and 77% in air, respectively. This decrease under aerated conditions points to accessible triplet states involved in the emission mechanism. However, delayed emission is too weak to detect due to competing non-radiative decay and the inefficient population of the triplet excited states.

The S₁ and T₁ levels were determined from the onsets of the steady-state PL and time-gated emission spectra, respectively, in 2-methyltetrahydrofuran (2-MeTHF) glass at 77 K (Table 2 and Fig. 4). The S₁ levels increase narrowly from DtButHCzB (2.61 eV) to SpAc-tButHCzB (2.66 eV). There is a clear structure–property trend observed, where both decreasing the number of tert-butyl groups from four to two and introducing the acridine donor moiety introduces a small destabilization of the S₁ energy level, while the T₁ level remains unaffected by the number of tert-butyl groups (T₁ = 2.30 eV for DtButHCzB and tButHCzB). There is only a small stabilization of the T₁ energy in the derivatives containing the spiro-acridine donor (T₁ = 2.29 eV for SpAc-DtButHCzB and SpAc-tButHCzB). The trend in T₁ energies is well reproduced by the double hybrid calculations, while the trend in S₁ levels is only partially captured by the computations, which incorrectly predict a stabilization of the S₁ state in the donor-decorated derivatives. When compared to the structurally related emitter DtBuCzB (S₁ = 2.66 eV; T₁ = 2.53 eV in frozen toluene),¹⁸ the measured S₁ and T₁ energies for DtButHCzB and tButHCzB are both stabilized. This red-shift can be reasonably attributed to the reduced polarity of the toluene host matrix compared to the 2-MeTHF host matrix employed in this study. Notably, the S₁ energies of SpAc-DtButHCzB and SpAc-tButHCzB are comparable with their unsaturated analogue SpAc-tCzBN (S₁ = 2.64 eV; T₁ = 2.49 eV in frozen toluene),¹⁹ even though the latter was measured in a less polar environment. Interestingly, the T₁ energy of the reference compound SpAc-tCzBN is approximately 0.2 eV higher than those measured for SpAc-DtButHCzB and SpAc-tButHCzB. The ΔE_ST values are 320, 310, 370, and 350 eV for tButHCzB, DtButHCzB, SpAc-tButHCzB, and SpAc-DtButHCzB, respectively. These ΔE_ST represent a large barrier for direct RISC from T₁ to S₁, but one that is not impossible to overcome, given literature examples with similarly large ΔE_ST values yet nonetheless show TADF behavior.^32–34

Table 2 Rate constants of tButHCzB, DtButHCzB, SpAc-tButHCzB and SpAc-DtButHCzB

Compound	kr/×10^7 s^−1	knr/×10^7 s^−1	kISC/×10^8 s^−1	kRISC/×10 s^−1
tButHCzB	5.94	3.80	3.60	1.44
DtButHCzB	1.77	1.33	1.10	2.95
SpAc-tButHCzB	4.31	3.38	1.15	2.31
SpAc-DtButHCzB	3.00	2.66	1.02	1.89

---

As 1 wt% doped films in mCP, a host with a suitably high triplet energy (2.90 eV),³⁶ the PL spectra of SpAc-tButHCzB and SpAc-DtButHCzB are bathochromically shifted and broadened at λ_PL of 483 nm for both (FWHM of 27 and 26 nm; Fig. 4(c)), respectively, compared to those in toluene. The PL spectra of tButHCzB and DtButHCzB show an even larger red-shift and broadening to λ_PL of 491 and 490 nm (FWHMs of 45 and 34 nm), respectively. This observed red-shift and broadening of the emission band indicates that these molecules have a strong tendency to aggregate, which becomes even more pronounced at higher doping concentrations (Fig. S17). The observed aggregation is likely caused by the formation of π–π stacked dimers, as observed in the crystal structures, even for doping concentrations as low as 1 wt%. Additionally, the Φ_PL values of 53–61% represent a significant decrease compared to the solution-state measurements (Φ_PL of 86–91%, Table 2), this is consistent with these compounds aggregating. This behavior is consistent across a range of host materials of varying steric bulk and polarity (Fig. S18 and Table S4), underlining the intrinsic tendency to form aggregates in the solid state. This aggregation and the subsequent ACQ may be caused by the formation of H-bonds in the solid state, increasing the reorganisation energy due to the contribution from C–H stretching vibrational modes.^37,38 The S₁ and T₁ energy levels were determined from the onset of the steady-state and time-gated PL spectra at 77 K. For SpAc-DtButHCzB and SpAc-tButHCzB the S₁ levels were found to be 2.66 and 2.67 eV, respectively. The T₁ levels were 2.31 and 2.32 eV for SpAc-DtButHCzB and SpAc-tButHCzB, respectively. These values result in ΔE_ST of 350 meV for both SpAc-DtButHCzB and SpAc-tButHCzB and slightly smaller values of 320 and 310 meV for DtButHCzB and tButHCzB, respectively. As observed in frozen 2-MeTHF, the time-gated PL (10–85 ms) spectra of the films show residual delayed fluorescence in addition to phosphorescence (Fig. S14). The TRPL measurements under vacuum reveal that the PL of all four emitters decays with prompt lifetimes in the range of 5.2–6.3 ns, which are of similar magnitude to the solution-state measurements. In TRPL multichannel scaling experiment long delayed components were found with lifetimes, τ_d, of 148, 108, 154 and 95 ms for SpAc-DtButHCzB, SpAc-tButHCzB, DtButHCzB and tButHCzB, respectively. Compared to their unsaturated analogues DtBuCzB (69 μs; in mCBP)¹⁸ and SpAc-tCzBN (45 μs; 10 wt% in PhCzBCz)¹⁹ the PL of the tetrahydrocarbazole based emitters decays with a much longer lifetime, which is due to the comparably larger ΔE_ST in these compounds. Temperature-dependent TRPL measurements (Fig. 5) confirmed the TADF character of the four compounds. For the SpAc-extended emitters, SpAc-DtButHCzB and SpAc-tButHCzB, there is a very long lived emission at low temperatures (Fig. 5(a) and (c)), which is assigned to phosphorescence, while for DtButHCzB and tButHCzB (Fig. 5(b) and (d)) the delayed emission component of the TRPL significantly decreases at temperatures below room temperature, due to insufficient thermal energy to overcome the large ΔE_ST.

Fig. 5 Temperature-dependent TRPL decays of 1 wt% doped films in mCP of (a) SpAc-DtButHCzB, (b) DtButHCzB, (c) SpAc-tButHCzB and (d) tButHCzB; (λ_exc = 340 nm).

A comparison of the rate constants reveals that SpAc-DtButHCzB and DtButHCzB exhibit slower reverse intersystem crossing rates (k_RISC) compared to those with only two tert-butyl groups (tButHCzB and SpAc-tButHCzB). This suggests that the additional bulky substituents may hinder RISC. At the same time SpAc-DtButHCzB and DtButHCzB also display reduced non-radiative decay rates. Radiative decay rates (k_r) are generally comparable across the four emitters, indicating that prompt fluorescence is not significantly affected by the substitution.

Electrochemistry

The electrochemical properties were investigated in DCM with 0.1 M [ⁿBu₄N]PF₆ as the supporting electrolyte using ferrocene/ferrocenium (Fc/Fc⁺) as an internal standard. The potentials are referenced to SCE (E_ox/red [V vs. SCE] = E_ox/red [V vs. Fc/Fc⁺] + 0.46).³⁹ All four emitters possess similar oxidation (E_ox) and reduction (E_red) potentials in DCM, ranging from 0.96 to 1.08 V and −1.81 to −1.88 V, respectively (Fig. 6 and Table 3). All measured electrochemical processes are irreversible. Compounds SpAc-DtButHCzB and SpAc-tButHCzB have anodically shifted E_ox and E_red values compared to those of DtButHCzB and tButHCzB, in line with the calculated values. The HOMO and LUMO levels were inferred from the E_ox/E_red values using the relationship E_HOMO/LUMO = −(E_ox/red + 4.8 eV).⁴⁰ The corresponding HOMO levels of SpAc-DtButHCzB and SpAc-tButHCzB are −5.42 and −5.39 eV, respectively. DtButHCzB and tButHCzB have shallower HOMO levels at −5.31 and −5.30 eV, respectively. A similar trend is observed for the LUMO energy levels, with a value of −2.47 eV for both DtButHCzB and tButHCzB, while SpAc-DtButHCzB and SpAc-tButHCzB have stabilized values of −2.53 and −2.52 eV, respectively. It is evident that the presence of the additional tert-butyl groups in SpAc-DtButHCzB and DtButHCzB stabilizes the HOMO, while the LUMO remains unaffected. The corresponding ΔE values are larger for SpAc-DtButHCzB and SpAc-tButHCzB than for DtButHCzB and tButHCzB, which results from the stronger stabilization of the HOMO compared to the LUMO in the former two compounds. This highlights again the above-mentioned theoretical calculation anomaly, since the predicted ΔE (Table 3) is smaller for SpAc-tButHCzB and SpAc-DtButHCzB, than for their unsaturated parent emitters, tButHCzB and DtButHCzB.

Fig. 6 Cyclic voltammetry (CV) and different pulse voltammetry (DPV) measurements performed in DCM solution (0.1 M [ⁿBu₄N]PF₆) at a scan rate of 0.1 V s⁻¹ and referenced to SCE (E_ox/red [V vs. SCE] = E_ox/red [V vs. Fc/Fc⁺] + 0.46).³⁹

Table 3 Electrochemistry data obtained by DPV measurements in DCM solution (0.1 M [ⁿBu₄N]PF₆)

Compound	Eox^a/eV	HOMO exp.^b/eV	HOMO calc.^c/eV	Ered^d/eV	LUMO exp.^e/eV	LUMO calc.^f/eV	ΔE exp^g/eV	ΔE calc.^g/eV
a Oxidation potential vs. SCE obtained by DPV measurements with a scan rate of 0.1 V s^−1 and referenced to SCE(Eox/red [V vs. SCE] = Eox/red [V vs. Fc/Fc^+] + 0.46).^39b HOMO = −(Eox,Fc/Fc^+ + 4.8 eV).c Calculated HOMO energy level on PBE0/6-31G(d,p) (toluene) level.d Ered vs. SCE obtained by DPV measurements with a scan rate of 0.1 V s^−1 and referenced to SCE (Eox/red [V vs. SCE] = Eox/red [V vs. Fc/Fc^+] + 0.46).^39e LUMO = −(Ered,Fc/Fc^+ + 4.8 eV).f Calculated LUMO energy level at the PBE0/6-31G** (toluene) level.g ΔE = |(HOMO–LUMO)|.
tButHCzB	0.96	−5.30	−5.29	−1.87	−2.47	−1.67	2.83	3.62
DtButHCzB	0.97	−5.31	−5.28	−1.88	−2.47	−1.66	2.84	3.62
SpAc-tButHCzB	1.05	−5.39	−5.35	−1.82	−2.52	−1.80	2.87	3.54
SpAc-DtButHCzB	1.08	−5.42	−5.36	−1.81	−2.53	−1.81	2.89	3.55

---

Conclusions

In this study we explored an approach to decrease the conjugation length and planarity of the donor fragment used in MR-TADF emitters in order to blue-shift the emission. Partially saturating the carbazole donor also positively improved the solubility of these compounds. The four emitters in this study all have very high Φ_PL in solution and emit in the sky-blue region, showing very narrowband PL spectra. Unfortunately, in the solid state, these molecules showed a strong tendency to aggregate, even at low doping concentrations, and this resulted in a significant drop in their Φ_PL values, which rendered their use as emitters in OLEDs not desirable. Nonetheless, this work highlights a robust structure–property study that explores new chemical space in MR-TADF emitter design. The design concept was successful in that the emission was blue-shifted as a result of partial saturation of the carbazole donors. Current efforts are focused on derivatizing these compounds to mitigate the strong quenching in the solid state and recover the high Φ_PL observed in solution.

Author contributions

M. F.: conceptualization, data curation, formal analysis, investigation, methodology, visualization, writing – original draft preparation, writing – review & editing. A. P. M.: data curation, formal analysis, investigation, visualization, writing – review & editing. D. B. C.: writing – review & editing. E. Z.-C.: conceptualization, funding acquisition, methodology, project administration, resources, supervision, writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Data availability

The research data supporting this publication can be accessed at https://doi.org/10.17630/ba2c5474-135e-4b6e-84d4-74980f549178.

Supplementary information: ¹H NMR and ¹³C NMR spectra, HRMS, and elemental analysis; details of X-ray crystallography; supplementary computational data and coordinates; supplemental photophysical data. see DOI: https://doi.org/10.1039/d5tc02496j.

CCDC 2465301, 2465303 and 2465304 contain the supplementary crystallographic data for this paper.^41a–c

Acknowledgements

This project has been funded by the European Union Horizon 2021 research and innovation programme under grant agreement no. 101073045 (TADFsolutions). We thank the EPSRC for support (EP/X026175/1, EP/W015137/1, EP/R00188X/1, EP/W007517/1, and EP/Z535291/1). We thank Aminata Mariko for her help with the synthesis of the tetrahydrocarbazole derivatives.

References

1. J. M. Dos Santos, D. Hall, B. Basumatary, M. Bryden, D. Chen, P. Choudhary, T. Comerford, E. Crovini, A. Danos, J. De, S. Diesing, M. Fatahi, M. Griffin, A. K. Gupta, H. Hafeez, L. Hammerling, E. Hanover, J. Haug, T. Heil, D. Karthik, S. Kumar, O. Lee, H. Li, F. Lucas, C. F. R. Mackenzie, A. Mariko, T. Matulaitis, F. Millward, Y. Olivier, Q. Qi, I. D. W. Samuel, N. Sharma, C. Si, L. Spierling, P. Sudhakar, D. Sun, E. Tankeleviciu Te, M. Duarte Tonet, J. Wang, T. Wang, S. Wu, Y. Xu, L. Zhang and E. Zysman-Colman, Chem. Rev., 2024, 124(24), 13736.
2. S. Madayanad Suresh, D. Hall, D. Beljonne, Y. Olivier and E. Zysman-Colman, Adv. Funct. Mater., 2020, 30(33), 1616.
3. Q. He, M. Li and S. J. Su, ChemPhysChem, 2025, 26(7), e202400955.
4. M. Du, J. Zhou, X. Luo, L. Duan and D. Zhang, Moore and More, 2024, 1(1), 3004.
5. E. Tankeleviciute, I. D. W. Samuel and E. Zysman-Colman, J. Phys. Chem. Lett., 2024, 15(4), 1034.
6. B. Cui, J. Zhou, C. Duan and H. Xu, Adv. Opt. Mater., 2025, 2403295.
7. J. Y. Woo, M. H. Park, S. H. Jeong, Y. H. Kim, B. Kim, T. W. Lee and T. H. Han, Adv. Mater., 2023, 35(43), e2207454.
8. T. W. Lee, T. Noh, H. W. Shin, O. Kwon, J. J. Park, B. K. Choi, M. S. Kim, D. W. Shin and Y. R. Kim, Adv. Funct. Mater., 2009, 19(10), 1625.
9. L. Duan, L. Hou, T.-W. Lee, J. Qiao, D. Zhang, G. Dong, L. Wang and Y. Qiu, J. Mater. Chem., 2010, 20(31), 6392.
10. Nat. Photonics, 2009, 38, 441, https://www.nature.com/articles/nphoton.2009.133.pdf .
11. C. M. Cole and S. D. Yambem, Adv. Opt. Mater., 2024, 13(1), 2402019.
12. D. Hall, S. M. Suresh, P. L. dos Santos, E. Duda, S. Bagnich, A. Pershin, P. Rajamalli, D. B. Cordes, A. M. Z. Slawin, D. Beljonne, A. Köhler, I. D. W. Samuel, Y. Olivier and E. Zysman-Colman, Adv. Opt. Mater., 2019, 8(2), 1901627.
13. Y. Yuan, X. Tang, X. Y. Du, Y. Hu, Y. J. Yu, Z. Q. Jiang, L. S. Liao and S. T. Lee, Adv. Opt. Mater., 2019, 7(7), 1801536.
14. T. Hatakeyama, K. Shiren, K. Nakajima, S. Nomura, S. Nakatsuka, K. Kinoshita, J. Ni, Y. Ono and T. Ikuta, Adv. Mater., 2016, 28(14), 2777.
15. L. Wu, X. Mu, D. Liu, W. Li, D. Li, J. Zhang, C. Liu, T. Feng, Y. Wu and J. Li, et al., Angew. Chem., Int. Ed., 2024, 63(38), e202409580.
16. C.-Z. Du, Y. Lv, H. Dai, X. Hong, J. Zhou, J.-K. Li, R.-R. Gao, D. Zhang, L. Duan and X.-Y. Wang, J. Mater. Chem. C, 2023, 11(7), 2469.
17. Y. T. Lee, C. Y. Chan, M. Tanaka, M. Mamada, U. Balijapalli, Y. Tsuchiya, H. Nakanotani, T. Hatakeyama and C. Adachi, Adv. Electron. Mater., 2021, 7(4), 2001090.
18. Y. Xu, Z. Cheng, Z. Li, B. Liang, J. Wang, J. Wei, Z. Zhang and Y. Wang, Adv. Opt. Mater., 2020, 8(9), 1902142.
19. X. Song, S. Shen, S. Zou, F. Guo, Y. Wang, S. Gao and Y. Zhang, Chem. Eng. J., 2023, 467, 143557.
20. Q. Wu, J. Li, D. Liu, Y. Mei, B. Liu, J. Wang, M. Xu and Y. Li, Dyes Pigm., 2023, 217, 111421.
21. N. J. Silva, F. B. Machado, H. Lischka and A. J. Aquino, Phys. Chem. Chem. Phys., 2016, 18(32), 22300.
22. D. Hall, J. C. Sancho-Garcia, A. Pershin, D. Beljonne, E. Zysman-Colman and Y. Olivier, J. Phys. Chem. A, 2023, 127(21), 4743.
23. X. Yan, Z. Li, Q. Wang, Y. Qu, Y. Xu and Y. Wang, J. Mater. Chem. C, 2022, 10(41), 15408.
24. D. Hall, J. C. Sancho-Garcia, A. Pershin, G. Ricci, D. Beljonne, E. Zysman-Colman and Y. Olivier, J. Chem. Theory Comput., 2022, 18(8), 4903.
25. T. Schwabe and S. Grimme, Phys. Chem. Chem. Phys., 2007, 9(26), 3397.
26. M. Kondo, Chem. Phys. Lett., 2022, 804, 139895.
27. Z. Huang, H. Xie, J. Miao, Y. Wei, Y. Zou, T. Hua, X. Cao and C. Yang, J. Am. Chem. Soc., 2023, 145(23), 12550.
28. L. Yang, L. Dong, D. Hall, M. Hesari, Y. Olivier, E. Zysman-Colman and Z. Ding, SmartMat, 2022, 4(2), 1149.
29. T. J. Penfold, F. B. Dias and A. P. Monkman, Chem. Commun., 2018, 54(32), 3926.
30. J. Wang, Y. Yasuda, Y. Ren, R. Kondo, D. B. Cordes, H. Kaji and E. Zysman-Colman, Org. Chem. Front., 2025, 12(10), 3279.
31. X. Wu, B.-K. Su, D.-G. Chen, D. Liu, C.-C. Wu, Z.-X. Huang, T.-C. Lin, C.-H. Wu, M. Zhu, E. Y. Li, W.-Y. Hung, W. Zhu and P.-T. Chou, Nat. Photonics, 2021, 15(10), 780.
32. D. Hall, K. Stavrou, E. Duda, A. Danos, S. Bagnich, S. Warriner, A. M. Z. Slawin, D. Beljonne, A. Kohler and A. Monkman, et al., Mater. Horiz., 2022, 9(3), 1068.
33. C. Cao, J. H. Tan, Z. L. Zhu, J. D. Lin, H. J. Tan, H. Chen, Y. Yuan, M. K. Tse, W. C. Chen and C. S. Lee, Angew. Chem., Int. Ed., 2023, 62(10), e202215226.
34. X. F. Luo, F. L. Li, J. W. Zou, Q. Zou, J. Su, M. X. Mao and Y. X. Zheng, Adv. Opt. Mater., 2021, 9(21), 2100784.
35. W. H. Melhuish, J. Phys. Chem., 1961, 65(2), 229.
36. R. J. Holmes, S. R. Forrest, Y.-J. Tung, R. C. Kwong, J. J. Brown, S. Graon and M. E. Thompson, Appl. Phys. Lett., 2003, 82(15), 2422.
37. K. Zhang, J. Liu, Y. Zhang, J. Fan, C.-K. Wang and L. Lin, J. Phys. Chem. C, 2019, 123(40), 24705.
38. Y.-Z. Shi, K. Wang, X.-C. Fan, H. Wu, X.-M. Ou, J. Yu, J.-S. Jie and X.-H. Zhang, J. Mater. Chem. C, 2022, 10(12), 4717.
39. V. V. Pavlishchuk and A. W. Addison, Inorg. Chim. Acta, 2000, 298(1), 97.
40. B. Dandrade, S. Datta, S. Forrest, P. Djurovich, E. Polikarpov and M. Thompson, Org. Electron., 2005, 6(1), 11.
41. (a) M. Fatahi, A. P. McKay, D. B. Cordes and E. Zysman-Colman, CCDC 2465301: Experimental Crystal Structure Determination, 2025 DOI:10.5517/ccdc.csd.cc2nrbvm; (b) M. Fatahi, A. P. McKay, D. B. Cordes and E. Zysman-Colman, CCDC 2465303: Experimental Crystal Structure Determination, 2025 DOI:10.5517/ccdc.csd.cc2nrbxp; (c) M. Fatahi, A. P. McKay, D. B. Cordes and E. Zysman-Colman, CCDC 2465304: Experimental Crystal Structure Determination, 2025 DOI:10.5517/ccdc.csd.cc2nrbxp.

---