Ultrafast photophysics of an orange–red thermally activated delayed fluorescence emitter: the role of external structural restraint

Abstract

The application of thermally activated delay fluorescence (TADF) emitters in the orange–red regime usually suffers from the fast non-radiative decay of emissive singlet states (k^S_NR), leading to low emitting efficiency in corresponding organic light-emitting diode (OLED) devices. Although k^S_NR has been quantitatively described by energy gap law, how ultrafast molecular motions are associated with the k^S_NR of TADF emitters remains largely unknown, which limits the development of new strategies for improving the emitting efficiency of corresponding OLED devices. In this work, we employed two commercial TADF emitters (TDBA-Ac and PzTDBA) as a model system and attempted to clarify the relationship between ultrafast excited-state structural relaxation (ES-SR) and k^S_NR. Spectroscopic and theoretical investigations indicated that S₁/S₀ ES-SR is directly associated with promoting vibrational modes, which are considerably involved in electronic–vibrational coupling through the Huang–Rhys factor, while k^S_NR is largely affected by the reorganization energy of the promoting modes. By restraining S₁/S₀ ES-SR in doping films, the k^S_NR of TADF emitters can be greatly reduced, resulting in high emitting efficiency. Therefore, by establishing the connection among S₁/S₀ ES-SR, promoting modes and k^S_NR of TADF emitters, our work clarified the key role of external structural restraint for achieving high emitting efficiency in TADF-based OLED devices.

---

Introduction

Organic light-emitting diode (OLED) technology is becoming increasingly attractive due to its great potential for ultra-high definition (UHD) displays.^1–5 As a key factor in the performance of OLED devices, external quantum efficiency (η_EQE) is described as the ratio between the number of emitted photons and number of injected carriers. In most cases, the η_EQE of OLED devices can be expressed as the product of several contributing terms:^6,7

η_EQE = γ × η_EUE × Φ_F × η_out (1)

where the carrier balance factor (γ) and output coupling factor (η_out) are associated with the device design and fabrication, respectively, while exciton utilization efficiency (η_EUE) and fluorescent quantum yield (Φ_F) are regarded as intrinsic properties of the emitters and are associated with corresponding excited-state processes. Spin statistics show that the recombination of injected electrons and holes leads to 25% singlet and 75% triplet states in OLED emitting layers, respectively,^8,9 while η_EUE gives the fraction of excited states that can decay to the ground state radiatively. Since T₁ → S₀ decay is spin-forbidden, η_EUE is largely determined by the quantum yield (Φ_RISC) of the reverse intersystem crossing (RISC, T₁ → S₀, with rate constant of k_RISC), which competes with other T₁ relaxation routes (radiative k^T_R and non-radiative k^T_NR, Fig. S1†):^10,11

Enormous efforts have been made towards designing thermally activated delayed fluorescence (TADF) emitters with minimized singlet–triplet energy gap (ΔE_ST) and enhanced spin–orbit coupling (SOC) to enable thermal conversion of T₁ → S₁, which can potentially push η_EUE to ∼100%.^12–15 Since the k_RISC of TADF emitters is usually on a time scale of 10³–10⁵ s⁻¹, η_EUE is generally associated with slow dynamics of TADF emitters, which has been intensively investigated.^14,16–18

In addition to the efficient utilization of current-generated triplet states, the Φ_F of TADF emitters greatly affects the η_EQE of the corresponding OLED devices, which describes the ratio of singlet (S₁) states that can be radiatively relaxed to the S₀ state, i.e., competing plausible relaxation channels for the S₁ state (Fig. S1†):^19–21

In addition, Φ_F includes a contribution from the S₁ state directly generated by injected carriers (Φ_PF, prompt fluorescence) and converted from the T₁ state (Φ_DF, delayed fluorescence):

Φ_F = Φ_PF + Φ_DF (4)

For organic TADF emitters, k_ISC is usually slow (10⁷–10⁸ s⁻¹),^16,22 for which the presence of rapid non-radiative decay is a key factor that can significantly reduce the Φ_F of TADF emitters. In a weak coupling regime, the rate constant of the S₁ state non-radiative decay (k^S_NR) with energy gap (ΔE_S1–S0) and electronic coupling (C) can be described as^23–25

where ω_M and λ_M represent the frequency and reorganization energy contribution of the promoting vibrational mode, while l is the number of involved vibrational modes. By assuming l = 1, eqn (5) can be simplified to the famous energy gap law with the molecule-specific parameter γ as^21,25–27which indicates that k^S_NR increases exponentially with the reduction of the S₁ → S₀ energy gap (ΔE_S1–S0), leading to challenges in the design of long-wavelength TADF emitters with high Φ_F.^28–31 As a result, reported TADF emitters in the blue to green regime largely exhibit considerably high Φ_F,^32–34 while the η_EQE of the corresponding OLED devices are primarily affected by RISC. In contrast, the small ΔE_S1–S0 of orange–red TADF emitters leads to fast k^S_NR competing with the radiative decay of the S₁ state, resulting in a low Φ_F and subsequently an unsatisfactory η_EQE for OLED devices, although up to 100% η_EUE can be expected via efficient RISC.^{19,28,35–37}

Recently, enormous efforts have been made towards developing long-wavelength TADF-based OLED devices with a high η_EQE,^28,35 in which the low Φ_F of TADF emitters is usually regarded as the main obstacle.^28,38–40 Therefore, one can expect opportunities to further improve the η_EQE of TADF-based long-wavelength OLED devices by somehow effectively slowing down the k^S_NR rate.^41–44 Recently, Kwon and co-workers reported an acceptor–donor–acceptor (ADA) type TADF emitter (PzTDBA, Scheme 1) and realized >30% of the η_EQE of orange–red OLED devices.⁴⁵ Although the excited-state mechanism that is responsible for the high η_EQE is still ambiguous, the reported ∼100% Φ_F of PzTDBA in doping films indicates that non-radiative decay of the S₁ state is nearly terminated, which seems to violate the well-known energy gap law (eqn (6)). Therefore, understanding the underlying ultrafast photophysics that is associated with the slow k^S_NR of PzTDBA might provide inspiration for designing high-performance TADF emitters in the long-wavelength regime.

Scheme 1 Illustrated chemical structure of investigated TADF emitters: (a) TDBA-Ac, D–A type; (b) PzTDBA, A–D–A type. The charge donors and acceptors are displayed in blue and red, respectively.

In this work, the ultrafast photophysics of an orange–red TADF emitter (PzTDBA, ADA-type) and its DA-type analogue (TDBA-Ac, deep-blue) was investigated using femtosecond transient absorption (fs-TA), time-resolved fluorescence (tr-FL) and theoretical vibrational analysis.⁴⁶ Compared with the one-step D–A twisting of the TDBA-Ac emitter, PzTDBA exhibits two-step S₀/S₁ excited-state structural relaxation (ES-SR), i.e. fast D–A twisting and slow planarization of the Pz group. The promoting vibrational modes associated with the S₀/S₁ ES-SR motions of the TDBA-Ac and PzTDBA emitters were theoretically identified, which dominate the electronic-vibrational coupling (EVC) of the S₁ state. Meanwhile, promoting modes contribute to fast k^S_NR through the corresponding reorganization energy contribution (λ_M). In doping films, the S₀/S₁ ES-SR motions of PzTDBA are suppressed by external structural restraint, which greatly slows down k^S_NR and leads to ∼100% Φ_F. Our work established the connection among the S₀/S₁ ES-SR, promoting modes and S₁ state non-radiative decay and indicated the key role of medium rigidity in improving the emitting efficiency of TADF emitters, which should provide inspiration for the future development of TADF emitters.

Results and discussion

Low-lying excited states and steady state spectra

Fundamental photophysics of TDBA-Ac and PzTDBA emitters were first investigated in solvents with different polarities (Table S1†). In OLED devices, TADF emitters are usually doped in a wide-bandgap organic semiconductor as the host. To avoid multi-photon excitation of the host in fs-TA experiments, we prepared doping films (2 wt%) of TADF emitters by employing polystyrene (PS) as the host, and the polarity of the PS medium (ε = 2.6–2.7, Δf = 0.03) was determined to be comparable with toluene (Fig. S2†).^47,48 The UV/Vis absorption and fluorescence spectra of the TDBA-Ac and PzTDBA emitters in various solutions and films (2 wt%) can be seen in Fig. 1 and S3.† The TD-DFT (M06-2X/6-311G**, PCM = toluene) calculated vertical low-lying singlet and triplet states (S₁–S₃, T₁–T₃) are listed in Table S2† with the visualized distribution of frontier orbitals (HOMO−2 to LUMO+2) shown in Table S3.† The lowest-lying S₁ states of TDBA-Ac and PzTDBA are dominated by the H → L transition, corresponding to the charge transfer state (noted as ¹CT_DA) with low oscillator strength, which can be seen as weak absorption of TDBA-Ac (400–450 nm) and PzTDBA (420–540 nm). Upon UV excitation, solvatochromism was observed for TDBA-Ac and PzTDBA, indicating that the fluorescence emission originated from the lowest-lying ¹CT_DA state.^49,50 Intriguingly, ∼1000 cm⁻¹ (TDBA-Ac) and ∼1400 cm⁻¹ (PzTDBA) blue-shifted emissions were observed in PS doping films compared to in toluene (comparable polarity), indicating the presence of multifaceted interactions, which are discussed below.

Fig. 1 The steady UV/vis absorption and fluorescence spectra of TDBA-Ac (a) and PzTDBA (b) in various solutions and PS doping films. Measured total fluorescence quantum yield (Φ_F), corresponding prompt (Φ_PF) and delayed (Φ_DF) contribution, as well as the related contribution of the delayed component (Φ_DF/Φ_F) of TDBA-Ac (c) and PzTDBA (d).

The T₁ state of TDBA-Ac was identified as a local excited state on the acceptor (noted as ³LE_A) while the T₁ excitation of PzTDBA was localized on the donor (³LE_D), which allows direct T₁ → S₁ RISC (³LE_D/A → ¹CT_DA).^51,52 Meanwhile, the T₂ states of TDBA-Ac and PzTDBA were recognized as charge transfer states (³CT_DA). With identical orbital wavefunction to the ¹CT_DA state, the corresponding T₂ → S₁ RISC (³CT_DA → ¹CT_DA) is forbidden.^51,52 By further confirming the CT/LE nature of low-lying singlet/triplet states through natural transition orbital (NTO) analysis (Fig. S4 and S5†) and hole–electron analysis (Table S4†),^53,54 the T₁ → S₁ (³LE_D/A → ¹CT_DA) transition was assigned as the accessible RISC channel that is responsible for the delayed fluorescence of TDBA-Ac and PzTDBA.

The thermally accessible ΔE_ST is critical for RISC, while estimating ΔE_ST through the vertical excitation energy of S₁ and T₁ states was reported to be unreliable.⁵⁵ Thus, we further optimized the S₁ and T₁ geometry of the TDBA-Ac and PzTDBA emitters to estimate adiabatic . As listed in Table 1, TDBA-Ac and PzTDBA exhibit of ∼0.3 eV and ∼0.26 eV in low-polarity media (CHX and TOL, Δf < 0.02), respectively, facilitating RISC for harvesting T₁ states with moderate 〈S₁|Ĥ_SO|T₁〉. However, increasing to medium polarity (in DCM, Δf = 0.22) leads to enlarged (>0.33 eV) for TDBA-Ac and PzTDBA with nearly unchanged 〈T₁|Ĥ_SO|S₁〉, which might explain the reduced RISC rate (k_RISC). Note that the vertical (ΔE_ST) and adiabatic singlet-triplet energy gap does not have a certain magnitude relationship as it depends on multiple factors, such as the different ES-SR in the S₁/T₁ states and steepness of the corresponding potential energy surface (PES), which are simplified as schematic diagrams (Fig. S6†). The total Φ_F and corresponding prompt/delayed (Φ_PF/Φ_DF) contribution of TDBA-Ac and PzTDBA were further measured in different solutions and PS doping films. As illustrated in Fig. 1c, TDBA-Ac exhibits nearly solvent-independent Φ_F (30–50%). With increasing solvent polarity, the solvation of the ¹CT_DA (S₁) state and unchanged energy level of the ³LE_A (T₁) state resulted in reduced ΔE_ST,^50,56 which is consistent with the observed increasing of the Φ_DF contribution (Φ_DF/Φ_F, Fig. 1c). Upon optical excitation, S₁ → T₁ → S₁ is the only feasible channel for delayed fluorescence,^17,38,57 and increased Φ_DF/Φ_F might correspond to higher Φ_ISC and Φ_RISC. In contrast, the Φ_F of PzTDBA rapidly decreased from 37% in non-polar CHX to undetectably low in polar mediums, which implies different relaxation of the photo-generated S₁ state. Intriguingly, TDBA-Ac and PzTDBA exhibit nearly 100% Φ_F and an evenly divided contribution of Φ_PF and Φ_DF (Φ_DF/Φ_F ≈ 50%) in PS doping films, indicating that the S₁ state decay is dominantly radiative, while all ISC-generated T₁ states can be harvested through subsequent RISC. Considering the importance of Φ_F, resolving the photophysics behind the fully radiative relaxation of orange–red PzTDBA in doping film would be of high interest.

Table 1 TDDFT calculated vertical (E) and adiabatic (E*) excitation energy of the lowest-lying (S₁, T₁) states and corresponding energy gaps (ΔE_ST, ) of TDBA-Ac and PzTDBA in different mediums using PCM as the model; SOC matrix elements for ISC (〈S₁|Ĥ_SO|T₁〉) and RISC (〈T₁|Ĥ_SO|S₁〉) were calculated by using the linear response approach

	TDBA-Ac			PzTDBA
	CHX	TOL	DCM	CHX	TOL	DCM
E(S1)/eV	3.617	3.626	3.667	3.130	3.142	3.207
E(T1)/eV	3.250	3.251	3.252	3.015	3.014	3.010
ΔEST/eV	0.367	0.375	0.415	0.115	0.128	0.197
E*(S1)/eV	3.474	3.482	3.509	2.792	2.804	2.861
E*(T1)/eV	3.173	3.173	3.174	2.539	2.538	2.531
/eV	0.301	0.309	0.335	0.253	0.266	0.330
〈S1|ĤSO|T1〉/cm^−1	0.338	0.339	0.294	0.016	0.018	0.058
〈T1|ĤSO|S1〉/cm^−1	0.341	0.341	0.341	0.053	0.052	0.058

---

tr-FL and excited-state relaxation

The time-resolved fluorescence (tr-FL) measurements of TDBA-Ac and PzTDBA were further performed by time-correlated single-photon counting (TCSPC) upon excitation at λ_ex = 400 nm.

The resulting traces were fitted as the sum of prompt (τ_PF) and delayed (τ_DF) components (Fig. S7, S8† and Table 2). Combining with measured Φ_PF and Φ_DF, the rate constants for ISC (S₁ → T₁, k_ISC) and RISC (T₁ → S₁, k_RISC) as well as corresponding Φ_ISC and Φ_RISC were estimated (Table 2). With calculated SOC matrix elements (〈S₁|Ĥ_SO|T₁〉 and 〈T₁|Ĥ_SO|S₁〉), we attempted to reproduce the experimentally extracted values of k_ISC and k_RISC using the thermal vibration correlation function (TVCF)^16,54,58 and semi-classical Marcus^59,60 approaches. However, as shown in Table S5,† TVCF and Marcus approaches failed to describe k_ISC and k_RISC, which might be attributed to the ignorance of non-Condon effects, such as Herzberg–Teller coupling and spin–vibronic coupling.^61–63 Meanwhile, we noticed that calculation errors for the k_ISC of PzTDBA are much more significant than for TDBA-Ac, which might imply higher structural flexibility of PzTDBA and will be discussed in detail below. By further estimating the experimental values of k^S_R, k^S_NR and k^T_NR (Table 2), the quantitative contribution of the plausible relaxation channels was calculated for the S₁ and T₁ states.

Table 2 Measured photophysical data for the TDBA-Ac and PzTDBA emitters in N₂-saturated solutions and PS doping films

	TDBA-Ac				PzTDBA
	CHX	TOL	DCM	PS	CHX	TOL	DCM	PS
a Concentration of 10^−5 M. b Doping concentration of 2 wt%. c Calculated by kPF = 1/τPF and kDF = 1/τDF. d Rate constants k^SR, kISC, kRISC and k^TNR were calculated using the method described by Adachi et al.^10,19 e Calculated using k^SNR = kPF − k^SR − kISC. f Calculated using ΦISC = kISC/(kISC + k^SR + k^SNR). g Calculated using ΦRISC = ΦDF/(1 − ΦPF).
τ PF/ns	6.72	21.35	36.15	14.12	8.70	23.95	13.62	38.25
τ DF/μs	0.15	0.14	0.48	1.21	0.10	0.29	0.23	1.10
Φ F	0.53	0.49	0.47	0.90	0.37	0.13	<0.01	0.95
Φ PF	0.23	0.16	0.11	0.44	0.03	0.03	<0.01	0.50
Φ DF	0.30	0.33	0.36	0.46	0.34	0.10	0	0.45
Φ ISC	0.57	0.67	0.78	0.51	0.93	0.78	0	0.47
Φ RISC	0.39	0.39	0.43	0.82	0.35	0.10	0	0.90
k PF/10^7 s^−1	14.88	4.69	2.77	7.14	11.49	4.17	7.35	2.63
k ^SR/10^6 s^−1	33.48	7.70	3.05	31.43	3.10	1.13	—	13.16
k ^SNR/10^6 s^−1	30.29	7.82	3.16	3.49	5.28	8.02	73.46	0.69
k ISC/10^7 s^−1	8.50	3.14	2.15	3.65	10.66	3.25	—	1.25
k DF/10^6 s^−1	3.57	3.54	1.01	0.74	3.63	0.43	—	0.86
k RISC/10^5 s^−1	18.75	17.57	4.98	6.69	13.42	0.52	—	8.20
k ^TNR/10^5 s^−1	31.50	32.55	9.60	4.49	35.91	4.24	—	4.53

---

As visualized in Fig. 2, the S₁ relaxation of TDBA-Ac is dominated by the slightly increased ISC with increased solvent polarity, while the ISC-generated T₁ states can be converted to S₁ with nearly identical Φ_RISC, which leads to the nearly unchanged Φ_F of TDBA-Ac in CHX (0.53), TOL (0.49) and DCM (0.47). In contrast, although ISC still dominates the S₁ decay of PzTDBA in CHX and TOL, the generated T₁ states largely decay non-radiatively to S₀ rather than thermally converting to S₁via RISC, for which a low Φ_F was observed in CHX (0.37) and TOL (0.13). In high-polarity DCM, the S₁ decay of PzTDBA is predominately occupied by the non-radiative path to S₀, leading to an undetectable Φ_F (<0.01), which might be attributed to the exponentially increased k^S_NR described by band-gap law with the reducing of the S₁ → S₀ energy gap (ΔE_S1–S0).

Fig. 2 Visualized contribution of ISC (S₁ → T₁), radiative (R, S₁ → S₀) and non-radiative (NR, S₁ → S₀) decay of S₁ state for TDBA-Ac (a) and PzTDBA (b) in various solutions and PS doping films. Contribution of RISC (T₁ → S₁) and non-radiative (NR, T₁ → S₀) decay of T₁ state for TDBA-Ac (c) and PzTDBA (d) in various solutions and PS doping films, radiative decay of T₁ state was ignored due to low contribution.

Intriguingly, TDBA-Ac and PzTDBA exhibit nearly identical pattern of S₁ and T₁ decay in PS doping films, which is largely differed with observation in solutions. The S₁ decay of TDBA-Ac and PzTDBA in PS doping films are dominated by equally divided radiative decay (S₁ → S₀) and ISC (S₁ → T₁), while ISC-generated T₁ states can be further converted to S₁ through efficient RISC (Φ_RISC > 0.9). Due to negligible role of non-radiative path in both S₁ (<5%) and T₁ (<10%) state decay, TDBA-Ac and PzTDBA exhibit unexpectedly high Φ_F in PS doping films, which agrees with the reported high η_EQE of corresponding OLED devices.^45,46 However, such high Φ_F (0.95) of PzTDBA observed in PS doping film seems to be inconsistent with the energy gap law, implying unique photophysics by which non-radiative decay of S₁ state can be greatly suppressed.

fs-TA and structural relaxation

The fs-TA signal upon UV excitation (λ_ex = 320 nm) of TDBA-Ac and PzTDBA in CHX solution and PS doping film were recorded by broadband probe (λ_probe = 350–750 nm) with delay times up to 7.0 ns, and the temporal resolution of fs-TA was measured to be ∼90 fs. As shown in Fig. S9 and S10,† the measured fs-TA signals of TDBA-Ac and PzTDBA are contributed by negative ground-state bleaching (GSB) in the λ_probe = 350–400 nm regime, and the positive excited-state absorption (ESA) band extended in the longer wavelength region up to λ_probe = 750 nm.

Upon UV excitation, high-lying S_n states can be populated, followed by rapid internal conversion (IC, S_n → S₁) to the long-lived S₁ state, which explains the dramatically reshaped ESA in the 1 ps delay time.^64,65 For the subsequent delay times of 1 ps to 1 ns, fs-TA evolution visualizes wavepacket motion that highly depends on the topology of S₁ PES.^41,66–68 For instance, the role of the excited-state structural relaxation (ES-SR) in the S₁ state decay of DA- and DAD/ADA-type TADF emitters has been intensively reported.^69–73 Suffering from the poor structural sensitivity of UV/Vis fs-TA, S₁/S₀ ES-SR usually leads to minimized alternation of the ESA shape, which is consistent with the observed fs-TA of TDBA-Ac and PzTDBA for 1 ps to 1 ns delay times.^41,74

For characterizing S₁/S₀ ES-SR, the geometries of TDBA-Ac and PzTDBA in the S₀, S₁ and T₁ states were optimized using DFT and TD-DFT approaches. The resulting structures are illustrated in Table S6,† while total reorganization energies (λ_S0→S1 and λ_S1→S0) were calculated for evaluating the ES-SR of the TDBA-Ac and PzTDBA emitters together with the root of the mean squared displacement (RMSD_S1/S0) between the S₀ and S₁ states (Fig. S11†). As listed in Table 3, the calculated values of λ_S0→S1, λ_S1→S0 and RMSD_S1/S0 indicate a more pronounced ES-ER of PzTDBA in the S₁ state than TDBA-Ac, which obviously cannot be attributed to structural extension (DA to ADA structure) as the total reorganization energy and RMSD are not additive. To evaluate the contribution of each of the molecular fragments of the TDBA-Ac and PzTDBA emitters to the ES-SR, several critical structural parameters were defined, as illustrated in Fig. S12† and measured (Table 3) for the S₀ and S₁ states. Upon vertical excitation, TDBA-Ac and PzTDBA initially remain in S₀ geometry in the Franck–Condon (FC) region, following by ES-SR until reaching the global minimum of S₁ PES, i.e. optimized S₁ geometry. For TDBA-Ac, S₁/S₀ ES-SR featured a slight increase in the D–A twisting angle β from 89.82° (S^FC₁) to 92.00° (S^T₁), while the dihedral angles of the Ac (α) and TDBA (γ) framework bending remained nearly unchanged in the S₁ state decay. In contrast, ADA-type PzTDBA exhibits higher flexibility than that of TDBA-Ac. In addition to the fast motion of the D–A twisting angle (β) from 79.72°/79.06° (S^FC₁) to 89.02°/91.43° (S^T₁), framework planarization of the center donor (Pz) was observed as the dihedral angle (α) increased from 164.57° (S^T₁) to 180.00° (S^TP₁). The simultaneous S₀–S₁ state changing of α and β angles may correspond to a two-step S₁/S₀ ES-SR (S^FC₁ → S^T₁ → S^TP₁), which has been widely reported,^75,76i.e. fast D–A twisting (S^FC₁ → S^T₁) followed by slow framework planarization (S^T₁ → S^TP₁). Target analysis was further performed to acquire quantitative information on S₁/S₀ ES-SR that may play a key role in the S₁ state relaxation. By including three or four sequential decay processes, measured fs-TA data can be well-reproduced by the displayed decay-associated spectra (DAS) of each decay components and concentration evolution of each transient species (Fig. 3), while species-associated spectra (SAS) can be seen in Fig. S13 and S14.†

Table 3 DFT and TD-DFT (M06-2X, 6-311G**, PCM = toluene) calculated optimal geometric parameters of TDBA-Ac and PzTDBA in the S₀, S₁ and T₁ states, as well as the calculated total reorganization energy and RMSD associated with the S₁ state ES-SR^a

	TDBA-Ac			PzTDBA
	*α (°)	**β (°)	***γ (°)	*α (°)	**β1/β2 (°)	***γ1/γ2 (°)
a *Bending dihedral angle of donor (Ac/Pz); **donor–acceptor twisting angle; ***bending dihedral angle of acceptor (TDBA).
S0 geometry	176.50	89.82	11.36	164.57	79.72/79.06	11.56/10.99
S1 geometry	175.45	92.00	12.12	180.00	89.02/91.43	11.65/11.37
T1 geometry	176.57	90.82	4.77	180.00	88.60/91.33	11.40/11.40
λ S0→S1 (cm^−1)	1161			2790
λ S1→S0 (cm^−1)	1273			2552
RMSDS1/S0 (Å)	0.053			0.241

---

Fig. 3 The decay-associated spectra (DAS) extracted from the fs-TA data of TDBA-Ac in CHX solution (a) and PS doping film (b) as well as PzTDBA in CHX solution (c) and PS doping film (d). Fitted concentration evolution of TDBA-Ac (e) and PzTDBA (g) in CHX solution (dashed lines) and PS doping film (solid lines). Promoting vibrational modes of TDBA-Ac (f) and PzTDBA (h) with large contribution to the HR factors. Calculated HR factors and reorganization energy contribution of each vibrational mode of TDBA-Ac (i) and PzTDBA (j) for the S₁ → S₀ transition in CHX solution. The dominant modes displayed in (f) and (h) are highlighted correspondingly.

The initial decay (A → B) with 300–400 fs time constant was accompanied by considerable ESA reshaping, corresponding to rapid IC (S_n → S₁). The fs-TA was subsequently dominated by S₁/S₀ ES-SR until the wavepacket reached the global minimum of S₁ PES. For PzTDBA, highlighted two-step S₁/S₀ ES-SR (B → C → D, i.e. S^FC₁ → S^T₁ → S^TP₁) predicted by calculated S₀ and S₁ geometries was observed, featuring a slight ESA reshaping. The observed fast (4.8 ps in CHX, S^FC₁ → S^T₁) and slow (334 ps in CHX, S^T₁ → S^TP₁) steps correspond to D–A twisting (β₁/β₂ angles) and Pz planarization (α angle), respectively.

For TDBA-Ac, one-step S₁/S₀ ES-SR was observed (7.2 ps, S^FC₁ → S^T₁) in the PS doping film, corresponding to the D–A twisting (β angle), which is comparable with the fast S₁/S₀ ES-SR step of PzTDBA. However, S₁/S₀ ES-SR of TDBA-Ac was unobservable in CHX, which might be attributed to a less pronounced S₀ → S₁ structural change than for PzTDBA, i.e. nearly unchanged β angle in the S₀ (89.82°) and S₁ (92.00°) states.

Intriguingly, the extracted time constants of D–A twisting (S^FC₁ → S^T₁, 14.0 ps) and Pz planarization (S^T₁ → S^TP₁, 636 ps) of PzTDBA in the PS doping film are 2–3 times slower than corresponding S₁/S₀ ES-SR steps in CHX (Fig. 3g and S15†), implying a higher potential barrier for S₁ PES, which is consistent with our fs-TA observations for multiple-resonance emitters.⁷⁷ Moreover, slow S₁ isomerization of azo-benzene embedded in polymer films was reported previously, attributing to external structural restraint from polymer micro-networks.⁷⁸

As discussed above, the fluorescence emission of PzTDBA was strongly quenched (Φ_F = 0.37 in CHX) in solutions due to the presence of fast non-radiative decay, while a high Φ_F (>0.95) in the PS doping film may be feasible if non-radiative S₁ → S₀ decay can be greatly suppressed. Our fs-TA data revealed higher S₁/S₀ ES-SR barriers of PzTDBA in PS doping films than in CHX, which might imply the presence of an underlying association between S₁/S₀ ES-SR and the non-radiative decay of the S₁ state, i.e. restrained S₁/S₀ ES-SR leads to suppressed non-radiative decay. Conversely, k^S_NR was described to be highly correlated with promoting vibrational modes by the energy gap law (eqn (5)).^21,23 Thus, we attempted to unveil the underlying relationship among the S₁/S₀ ES-SR, promoting modes and non-radiative decay of the S₁ state.

Vibrational analysis and non-radiative decay

The energy gap law (eqn (5)) clearly describes how promoting vibrational modes affect k^S_NR through their vibrational frequency (ω_M) and reorganization energy (λ_M), while promoting modes are vibrational modes that are considerably involved in the EVC of the S₁ state through a pronounced Huang-Rhys (HR) factor (S_k).^24,79,80 We further calculated the S_k and λ_k contribution of each of the vibrational modes involved in the S₁ → S₀ transition (Fig. 3i and j), in which the promoting modes associated with the k^S_NR of TDBA-Ac and PzTDBA are illustrated in Fig. 3f and h, respectively.

As shown in Fig. 3j and h, three promoting modes were identified for the S₁ → S₀ transition of PzTDBA due to considerable S_k. The modes at 410 cm⁻¹ (ω^DA-1_M) and 421 cm⁻¹ (ω^DA-2_M) correspond to symmetric and asymmetric D–A twisting and exhibit considerable reorganization energy contributions (λ^DA-1_M = 174 cm⁻¹ and λ^DA-2_M = 350 cm⁻¹), which are associated with the fast S₁/S₀ ES-SR step observed on fs-TA of PzTDBA, i.e. S^FC₁ → S^T₁ with the D–A twisting motion. For TDBA-Ac (Fig. 3i and f), the corresponding S₁/S₀ ES-SR of D–A twisting is associated with promoting modes at ω^DA-1_M = 247 cm⁻¹ and ω^DA-2_M = 387 cm⁻¹ with substantially lower λ^DA-1_M (45.9 cm⁻¹) and λ^DA-2_M (132 cm⁻¹), which is consistent with its rigid structure predicted theoretically, i.e. less pronounced S₁/S₀ ES-SR than PzTDBA.

Meanwhile, S₁ → S₀ decay of PzTDBA features a promoting mode at ω^D_M = 32.8 cm⁻¹ with a surprisingly high HR factor (S_k = 6.12) and considerable reorganization energy (λ^D_M = 201.1 cm⁻¹), corresponding to the bending motion of the Pz framework, which is clearly associated with the observed slow S₁/S₀ ES-SR step of PzTDBA, i.e. S^T₁ → S^TP₁. Intriguingly, this particular mode was also observed for TDBA-Ac at ω^D_M = 11.6 cm⁻¹ with a much lower HR factor (S_k = 0.15) and a two orders of magnitude lower reorganization energy contribution (λ^D_M = 1.70 cm⁻¹) than that of PzTDBA, indicating that it was excluded from the S₁ → S₀ decay of TDBA-Ac. As a result, the second S₁/S₀ ES-SR step (S^T₁ → S^TP₁) was absent from the fs-TA signal of TDBA-Ac.

Furthermore, we investigated the influence of medium polarity on the promoting modes of the TDBA-Ac and PzTDBA emitters. As shown in Fig. S16,† for PzTDBA in DCM solution, an extra promoting mode at 15.4 cm⁻¹ was observed with considerable S_k but ignorable λ_k, which might be less associated with the non-radiative decay of S₁ state PzTDBA. In contrast, a promoting mode at 12.4 cm⁻¹ was observed for TDBA-Ac with considerable S_k (8.71) and λ_k (108.23 cm⁻¹) in DCM solution, which is very different from the case of TDBA-Ac in low-polarity solvents, indicating that the S₁/S₀ ES-SR of TDBA-Ac (DA-type) is more evidentially coupled with charge transfer than PzTDBA (ADA-type).

As discussed above, the one- (TDBA-Ac, S^FC₁ → S^T₁) and two-step (PzTDBA, S^T₁ → S^TP₁) S₁/S₀ ES-SR are directly associated with promoting vibrational modes that are considerably involved in the EVC of the S₁ state through their HR factor. Meanwhile, promoting modes contribute to the k^S_NR of the S₁ state through corresponding λ_M, which implies that S₁/S₀ ES-SR can significantly affect k^S_NR (and subsequently Φ_F) through the promoting vibrational modes (Fig. 4). For instance, the one-step S₁/S₀ ES-SR of TDBA-Ac is associated with vibrational modes with low S_k and λ_M, while the promoting modes associated with the two-step S₁/S₀ ES-SR of PzTDBA have much higher S_k and λ_M. As a result, the non-radiative channel plays a minor role in the S₁ decay of structurally rigid TDBA-Ac, while the emission of structurally flexible PzTDBA is severely hampered by the fast non-radiative decay of the S₁ state. Furthermore, the two-step S₁/S₀ ES-SR motions of PzTDBA are greatly suppressed in the PS doping films due to the external structural restraint, for which the S₁ non-radiative decay associated with the promoting modes might be correspondingly weakened, leading to the greatly improved Φ_F of PzTDBA in the doping films. In this sense, the external structural restraint for S₁/S₀ ES-SR motions might be critical for achieving a high η_EQE for TADF-based OLED devices.

Fig. 4 Simplified energetic diagram illustrating two-step ES-SR (S^FC₁ → S^T₁ and S^T₁ → S^TP₁) of PzTDBA in low-polarity solution (orange line, TOL), high-polarity solution (cyan line, DCM) and PS doping film (red line). The double arrows represent the potential barriers of D–A twisting (S^FC₁ → S^T₁) and Pz planarization (S^T₁ → S^TP₁).

To directly verify the influence of the external structural restraint on the S₁ non-radiative decay of PzTDBA, we measured temperature-dependent emission spectra of PzTDBA in polyethylene oxide (PEO) doping films (Fig. S18†). With a glass transition temperature (T_g) of 220 K, PEO provides a “softer” microenvironment than PS (T_g = 373 K) and can be further “softened” by increasing the temperature.^81,82 As a TADF emitter, PzTDBA was expected to be more fluorescent at higher temperatures due to faster RISC. However, we observed clear fluorescence quenching with increasing temperature in the 297–347 K range, indicating that non-radiative decay was enhanced in the medium with reduced external structural restraint, which is consistent with our analysis described above.

Note that we cannot perform vibrational analysis for TADF emitters with the presence of the external structural restraint, but we speculate that the λ_M of the promoting modes of PzTDBA might be greatly reduced due to the external structural restraint, in agreement with slowed S₁/S₀ ES-SR motions in the doping films, especially the Pz bending mode (ω^D_M) associated with the slow S₁/S₀ ES-SR step (S^T₁ → S^TP₁) might be terminated. As a result, S^T₁ instead of S^TP₁ might dominate the emission of PzTDBA in the PS doping films (see Fig. 4), which explains the observed blue-shifted emission in the PS doping film compared with the case in toluene solution, which has comparable polarity to the PS medium. Meanwhile, the emission of PzTDBA in solutions mainly originates from S^TP₁ due to the accessible barrier of S^T₁ → S^TP₁, which thus suffers from the plague of non-radiative decay (Fig. 4).

Last but not least, in addition to the S₁ state, we noticed that TDBA-Ac and PzTDBA exhibit a suppressed non-radiative decay channel of the T₁ state in the PS doping films compared to in solutions (Fig. 2c and d), which inspired us to consider the role of T₁/S₀ ES-SR in the non-radiative decay of the T₁ state. As shown in Fig. S17,†TDBA-Ac exhibits several vibrational modes with considerable S_k in the region <100 cm⁻¹, but the low λ_M implies that ES-SR may not be strongly associated with non-radiative T₁ → S₀ decay. In contrast, the promoting modes of PzTDBA at 32.9 cm⁻¹ and 421/410 cm⁻¹ contribute considerable λ_M, while pronounced T₁/S₀ ES-SR was indicated (Table 3), including Pz bending and D–A twisting. Therefore, it might be plausible that the non-radiative T₁ → S₀ decay is similarly associated with two T₁/S₀ ES-SR motions, because the non-radiative T₁ → S₀ decay of PzTDBA was strongly suppressed in the PS doping film due to the external structural restraint. However, unlike S₁ relaxation, without direct spectroscopic evidence of T₁/S₀ ES-SR in different media, verifying this hypothesis requires further investigation.

Conclusions

To summarize, we performed a comprehensive investigation of the photophysics of two TADF emitters, TDBA-Ac (DA-type, blue light) and PzTDBA (ADA-type, orange–red light), using fs-TA, tr-FL and theoretical approaches. Compared with the one-step S₁/S₀ ES-SR (S^FC₁ → S^T₁) of TDBA-Ac, the S₁ state decay of PzTDBA is dominated by two steps, S₁/S₀ ES-SR (S^FC₁ → S^T₁ → S^TP₁), while greatly slowed S₁/S₀ ES-SR motions of PzTDBA were observed in the PS doping films due to the external structural restraints. Vibrational analysis indicated that the S₁/S₀ ES-SR motions are directly associated with the promoting modes that are considerably involved in the EVC of the S₁ state through their own S_k, while the promoting modes contribute to fast k^S_NRvia λ_M. In doping films, the external structural restraint leads to suppressed S₁/S₀ ES-SR and reorganization energy contribution of the promoting modes, resulting in slowed k^S_NR and a highly fluorescent S₁ state that is favorable for OLED application. With TDBA-Ac and PzTDBA as model systems, we established the connection among the S₀/S₁ ES-SR, promoting modes and k^S_NR of TADF emitters, which indicated the key role of the external structural restraint in obtaining high Φ_F TADF emitters. Our work also provides a new direction for OLED design, and it might be necessary to take the rigidity of host materials used for the emitting layer into consideration to avoid emission quenching through fast non-radiative decay associated with the S₁/S₀ ES-SR of TADF emitters, especially for TADF emitters with a relatively small band gap, which might be a useful reference for future OLED application.

Data availability

All experimental and calculational data are available from the corresponding author upon reasonable request.

Author contributions

Yixuan Gao: conceptualization, methodology, investigation, data curation, formal analysis, visualization, and writing – original draft; Yaxin Wang: methodology, data curation, formal analysis, and validation; Zilong Guo: investigation, visualization, project administration, and supervision; Yan Wan: methodology and resources; Zheng Xue and Yandong Han: resources and project administration; Wensheng Yang: supervision, resources, and funding acquisition; Xiaonan Ma: conceptualization, formal analysis, funding acquisition, supervision, and writing – review and editing.

Conflicts of interest

The authors declare no competing financial interest.

Acknowledgements

We acknowledge the financial support from the National Key Research & Development Program of China (Grant No. 2020YFA0714603 and 2020YFA0714604). We thank Dr Yingli Niu (Beijing Jiaotong University) for his useful tutoring on MOMAP calculation and Dr Lei Zhou (HORIBA Scientific, Beijing) for valuable discussion on tr-FL measurements.

Notes and references

1. Y.-Z. Shi, H. Wu, K. Wang, J. Yu, X.-M. Ou and X.-H. Zhang, Chem. Sci., 2022, 13, 3625–3651.
2. Y. Huang, E.-L. Hsiang, M.-Y. Deng and S.-T. Wu, Light: Sci. Appl., 2020, 9, 105.
3. Z. Yang, Z. Mao, Z. Xie, Y. Zhang, S. Liu, J. Zhao, J. Xu, Z. Chi and M. P. Aldred, Chem. Soc. Rev., 2017, 46, 915–1016.
4. G. Hong, X. Gan, C. Leonhardt, Z. Zhang, J. Seibert, J. M. Busch and S. Bräse, Adv. Mater., 2021, 33, 2005630.
5. H.-W. Chen, J.-H. Lee, B.-Y. Lin, S. Chen and S.-T. Wu, Light: Sci. Appl., 2017, 7, 17168.
6. M. A. Baldo, D. F. O'Brien, Y. You, A. Shoustikov, S. Sibley, M. E. Thompson and S. R. Forrest, Nature, 1998, 395, 151–154.
7. K. Goushi, K. Yoshida, K. Sato and C. Adachi, Nat. Photonics, 2012, 6, 253–258.
8. L. J. Rothberg and A. J. Lovinger, J. Mater. Res., 1996, 11, 3174–3187.
9. M. Pope, H. P. Kallmann and P. Magnante, J. Chem. Phys., 1963, 38, 2042–2043.
10. Q. Zhang, B. Li, S. Huang, H. Nomura, H. Tanaka and C. Adachi, Nat. Photonics, 2014, 8, 326–332.
11. H. Noda, H. Nakanotani and C. Adachi, Sci. Adv., 2018, 4, eaao6910.
12. H. Uoyama, K. Goushi, K. Shizu, H. Nomura and C. Adachi, Nature, 2012, 492, 234–238.
13. J. M. Kaminski, A. Rodríguez-Serrano, F. Dinkelbach, H. Miranda-Salinas, A. P. Monkman and C. M. Marian, Chem. Sci., 2022, 13, 7057–7066.
14. 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.
15. M. Cai, M. Auffray, D. Zhang, Y. Zhang, R. Nagata, Z. Lin, X. Tang, C.-Y. Chan, Y.-T. Lee, T. Huang, X. Song, Y. Tsuchiya, C. Adachi and L. Duan, Chem. Eng. J., 2021, 420, 127591.
16. Q. Peng, D. Fan, R. Duan, Y. Yi, Y. Niu, D. Wang and Z. Shuai, J. Phys. Chem. C, 2017, 121, 13448–13456.
17. T. Hosokai, H. Matsuzaki, H. Nakanotani, K. Tokumaru, T. Tsutsui, A. Furube, K. Nasu, H. Nomura, M. Yahiro and C. Adachi, Sci. Adv., 2017, 3, e1603282.
18. K. R. Naveen, P. Palanisamy, M. Y. Chae and J. H. Kwon, Chem. Commun., 2023, 59, 3685–3702.
19. Q. Zhang, H. Kuwabara, W. J. Potscavage, S. Huang, Y. Hatae, T. Shibata and C. Adachi, J. Am. Chem. Soc., 2014, 136, 18070–18081.
20. Z. Kuang, G. He, H. Song, X. Wang, Z. Hu, H. Sun, Y. Wan, Q. Guo and A. Xia, J. Phys. Chem. C, 2018, 122, 3727–3737.
21. J. A. Treadway, B. Loeb, R. Lopez, P. A. Anderson, F. R. Keene and T. J. Meyer, Inorg. Chem., 1996, 35, 2242–2246.
22. T. Ogiwara, Y. Wakikawa and T. Ikoma, J. Phys. Chem. A, 2015, 119, 3415–3418.
23. R. Englman and J. Jortner, Mol. Phys., 1970, 18, 145–164.
24. Y.-C. Wei, S. F. Wang, Y. Hu, L.-S. Liao, D.-G. Chen, K.-H. Chang, C.-W. Wang, S.-H. Liu, W.-H. Chan, J.-L. Liao, W.-Y. Hung, T.-H. Wang, P.-T. Chen, H.-F. Hsu, Y. Chi and P.-T. Chou, Nat. Photonics, 2020, 14, 570–577.
25. C. E. Whittle, J. A. Weinstein, M. W. George and K. S. Schanze, Inorg. Chem., 2001, 40, 4053–4062.
26. J. S. Wilson, N. Chawdhury, M. R. A. Al-Mandhary, M. Younus, M. S. Khan, P. R. Raithby, A. Köhler and R. H. Friend, J. Am. Chem. Soc., 2001, 123, 9412–9417.
27. J. V. Caspar and T. J. Meyer, J. Phys. Chem., 1983, 87, 952–957.
28. Y. Xiao, H. Wang, Z. Xie, M. Shen, R. Huang, Y. Miao, G. Liu, T. Yu and W. Huang, Chem. Sci., 2022, 13, 8906–8923.
29. Y. Liu, X. Xiao, Y. Ran, Z. Bin and J. You, Chem. Sci., 2021, 12, 9408–9412.
30. C. Wang, X. Li, Y. Gao, L. Wang, S. Zhang, L. Zhao, P. Lu, B. Yang, S. Su and Y. Ma, Adv. Opt. Mater., 2017, 5, 1700441.
31. A. Abdurahman, Y. Chen, X. Ai, O. Ablikim, Y. Gao, S. Dong, B. Li, B. Yang, M. Zhang and F. Li, J. Mater. Chem. C, 2018, 6, 11248–11254.
32. T.-L. Wu, M.-J. Huang, C.-C. Lin, P.-Y. Huang, T.-Y. Chou, R.-W. Chen-Cheng, H.-W. Lin, R.-S. Liu and C.-H. Cheng, Nat. Photonics, 2018, 12, 235–240.
33. H. J. Kim, H. Kang, J. Jeong, S. H. Park, C. W. Koh, C. W. Kim, H. Y. Woo, M. J. Cho, S. Park and D. H. Choi, Adv. Funct. Mater., 2021, 31, 2102588.
34. X. Tang, Q. Bai, T. Shan, J. Li, Y. Gao, F. Liu, H. Liu, Q. Peng, B. Yang, F. Li and P. Lu, Adv. Funct. Mater., 2018, 28, 1705813.
35. S. Sharma and A. K. Pal, J. Mater. Chem. C, 2022, 10, 15681–15707.
36. G. Chen, J. R. Swartzfager and J. B. Asbury, J. Am. Chem. Soc., 2023, 145, 25495–25504.
37. Y. Liu, J. Yang, Z. Mao, Y. Wang, J. Zhao, S.-J. Su and Z. Chi, Chem. Sci., 2023, 14, 1551–1556.
38. Y. Liu, C. Li, Z. Ren, S. Yan and M. R. Bryce, Nat. Rev. Mater., 2018, 3, 18020.
39. J. Xu, Y. Dai, J. Zhang, Z. Jia, Q. Meng and J. Qiao, Adv. Opt. Mater., 2023, 2300989.
40. Z. Li, J.-R. Zhang, X.-K. Tian, S. Yang, S. Chen, H. Zhou and X.-G. Yang, Chem. Sci., 2022, 13, 9381–9386.
41. W. Zhang, H. Song, J. Kong, Z. Kuang, M. Li, Q. Guo, C. Chen and A. Xia, J. Phys. Chem. C, 2019, 123, 19322–19332.
42. H. Miranda-Salinas, A. Rodriguez-Serrano, J. M. Kaminski, F. Dinkelbach, N. Hiromichi, Y. Kusakabe, H. Kaji, C. M. Marian and A. P. Monkman, J. Phys. Chem. C, 2023, 127, 8607–8617.
43. T. Yang, B. Liang, Z. Cheng, C. Li, G. Lu and Y. Wang, J. Phys. Chem. C, 2019, 123, 18585–18592.
44. J.-L. He, Y. Tang, K. Zhang, Y. Zhao, Y.-C. Lin, C.-K. Hsu, C.-H. Chen, T.-L. Chiu, J.-H. Lee, C.-K. Wang, C.-C. Wu and J. Fan, Mater. Horiz., 2022, 9, 772–779.
45. D. Karthik, Y. H. Jung, H. Lee, S. Hwang, B. Seo, J. Kim, C. W. Han and J. H. Kwon, Adv. Mater., 2021, 33, 2007724.
46. D. H. Ahn, S. W. Kim, H. Lee, I. J. Ko, D. Karthik, J. Y. Lee and J. H. Kwon, Nat. Photonics, 2019, 13, 540–546.
47. P. J. B. Clarricoats, Proc. IEE-Part C Monogr., 1962, 109, 401.
48. Z. Tang, C. Chang, F. Bao, L. Tian, H. Liu, M. Wang, C. Zhu and J. Xu, Polymers, 2021, 13, 284.
49. X. Niu, P. Gautam, Z. Kuang, C. P. Yu, Y. Guo, H. Song, Q. Guo, J. M. W. Chan and A. Xia, Phys. Chem. Chem. Phys., 2019, 21, 17323–17331.
50. W. Zhang, J. Kong, D. Hu, M. Tao, X. Niu, S. Vdović, D. Aumiler, Y. Ma and A. Xia, J. Phys. Chem. C, 2020, 124, 5574–5582.
51. M. K. Etherington, J. Gibson, H. F. Higginbotham, T. J. Penfold and A. P. Monkman, Nat. Commun., 2016, 7, 13680.
52. J. Gibson, A. P. Monkman and T. J. Penfold, ChemPhysChem, 2016, 17, 2956–2961.
53. Y. Wang, Z. Guo, Y. Gao, Y. Tian, Y. Deng, X. Ma and W. Yang, J. Phys. Chem. Lett., 2022, 13, 6664–6673.
54. Y. Wang, Y. Tian, Y. Gao, Z. Guo, Z. Xue, Y. Han, W. Yang and X. Ma, J. Phys. Chem. Lett., 2023, 14, 9665–9676.
55. H. Sun, C. Zhong and J.-L. Brédas, J. Chem. Theory Comput., 2015, 11, 3851–3858.
56. X. Zhao and J. Zhao, Chem. Commun., 2022, 58, 7666–7669.
57. H. Noda, H. Nakanotani and C. Adachi, Chem. Lett., 2019, 48, 126–129.
58. Q. Ou, Y. Shao and Z. Shuai, J. Am. Chem. Soc., 2021, 143, 17786–17792.
59. N. Aizawa, Y. Harabuchi, S. Maeda and Y.-J. Pu, Nat. Commun., 2020, 11, 3909.
60. Y. Olivier, B. Yurash, L. Muccioli, G. D'Avino, O. Mikhnenko, J. C. Sancho-García, C. Adachi, T.-Q. Nguyen and D. Beljonne, Phys. Rev. Mater., 2017, 1, 075602.
61. S. Lin, Z. Pei, B. Zhang, H. Ma and W. Liang, J. Phys. Chem. A, 2022, 126, 239–248.
62. J. Tatchen, N. Gilka and C. M. Marian, Phys. Chem. Chem. Phys., 2007, 9, 5209.
63. B. R. Henry and W. Siebrand, J. Chem. Phys., 1971, 54, 1072–1085.
64. K. A. Zachariasse, M. Grobys, I. Riickert and W. Kiihnle, J. Photochem. Photobiol., A, 1997, 105, 373–383.
65. M. C. Drummer, R. B. Weerasooriya, N. Gupta, B. T. Phelan, A. J. S. Valentine, A. A. Cordones, X. Li, L. X. Chen and K. D. Glusac, J. Phys. Chem. C, 2022, 126, 1946–1957.
66. T. Bolze, J. Wree, F. Kanal, D. Schleier and P. Nuernberger, ChemPhysChem, 2018, 19, 138–147.
67. H. Song, H. Zhao, Y. Guo, A. M. Philip, Q. Guo, M. Hariharan and A. Xia, J. Phys. Chem. C, 2020, 124, 237–245.
68. J. Köhler, T. Quast, J. Buback, I. Fischer, T. Brixner, P. Nuernberger, B. Geiß, J. Mager and C. Lambert, Phys. Chem. Chem. Phys., 2012, 14, 11081.
69. Y. Rout, C. Montanari, E. Pasciucco, R. Misra and B. Carlotti, J. Am. Chem. Soc., 2021, 143, 9933–9943.
70. Z. Zhang, Y.-S. Wu, K.-C. Tang, C.-L. Chen, J.-W. Ho, J. Su, H. Tian and P.-T. Chou, J. Am. Chem. Soc., 2015, 137, 8509–8520.
71. Y. Chen, K.-H. Chang, F.-Y. Meng, S.-M. Tseng and P.-T. Chou, Angew. Chem., Int. Ed., 2021, 60, 7205–7212.
72. Z. Ma, Z. Yang, L. Mu, L. Deng, L. Chen, B. Wang, X. Qiao, D. Hu, B. Yang, D. Ma, J. Peng and Y. Ma, Chem. Sci., 2021, 12, 14808–14814.
73. Y. Wen, S. Xiao, H. Liu, X. Tian, J. De, T. Lu, Z. Yang, D. Zou, Y. Lv, S.-T. Zhang, Q. Su and B. Yang, J. Mater. Chem. C, 2021, 9, 17511–17517.
74. V. Karunakaran and S. Das, J. Phys. Chem. B, 2016, 120, 7016–7023.
75. Y. Gao, Y. Wang, Z. Guo, Y. Wan, C. Li, B. Yang, W. Yang and X. Ma, J. Phys. Chem. B, 2022, 126, 2729–2739.
76. C.-W. Ju, B. Li, L. Li, W. Yan, C. Cui, X. Ma and D. Zhao, J. Am. Chem. Soc., 2021, 143, 5903–5916.
77. H. Gao, S. Shen, Y. Qin, G. Liu, T. Gao, X. Dong, Z. Pang, X. Xie, P. Wang and Y. Wang, J. Phys. Chem. Lett., 2022, 7561–7567.
78. J. Bahrenburg, F. Renth, F. Temps, F. Plamper and W. Richtering, Phys. Chem. Chem. Phys., 2014, 16, 11549.
79. H.-M. Pan, C.-C. Wu, C.-Y. Lin, C.-S. Hsu, Y.-C. Tsai, P. Chowdhury, C.-H. Wang, K.-H. Chang, C.-H. Yang, M.-H. Liu, Y.-C. Chen, S.-P. Su, Y.-J. Lee, H. K. Chiang, Y.-H. Chan and P.-T. Chou, J. Am. Chem. Soc., 2023, 145, 516–526.
80. Y.-C. Wei, B.-H. Chen, R.-S. Ye, H.-W. Huang, J.-X. Su, C.-Y. Lin, J. Hodgkiss, L.-Y. Hsu, Y. Chi, K. Chen, C.-H. Lu, S.-D. Yang and P.-T. Chou, Angew. Chem., Int. Ed., 2023, 62, e202300815.
81. C. Fanggao, G. A. Saunders, E. F. Lambson, R. N. Hampton, G. Carini, G. Di Marco and M. Lanza, J. Polym. Sci., Part B: Polym. Phys., 1996, 34, 425–433.
82. M. Hénot, A. Chennevière, E. Drockenmuller, K. Shull, L. Léger and F. Restagno, Eur. Phys. J. E, 2017, 40, 11.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4sc00460d

---