Exploration of high-performance triptycene-based thermally activated delayed fluorescence materials via structural alteration of donor fragment

Abstract

The utilization of thermally activated delayed fluorescence (TADF) materials in highly proficient organic light-emitting diodes (OLEDs) has attracted much attention. Based on TADF material TPA-QNX(CN)2, a series of three-dimensional donor-acceptor (D-A) triptycenes have been designed via structural modification of D-fragment. The influences of different D-fragments with various electron-donating strengths on the singlet-triplet energy gap (ΔE_ST), emission wavelength (λ_em), and electron/hole reorganization energy (λ_e/λ_h) are extensively studied by applying density functional theory (DFT) coupled with time-dependent density functional theory (TD-DFT). The computed results imply that as the electron-donating strength of the D-fragments increases, the ΔE_ST value decreases and λ_em is red-shifted for the molecules using the same acceptor units. Analogously, the ¹CT‒³CT state splitting (ΔE_ST (CT)) is also decreased by enlarging the twist angle (β) between the phenyl ring and alternative D-fragment. Therefore, efficient color tuning within a broad emission range (434–610 nm), as well as small ΔE_ST (CT) values (0.01–0.05 eV), has been accomplished by structural modification of the D-fragments. The greater electron-donating strength, the smaller ΔE_ST, and the smaller λ_h for PPXZ-QNX(CN)2 make it the best candidate among all the designed molecules.

1. Introduction

Since the first report by Chihaya Adachi and coworkers in 2009 [1,2], the utilization of thermally activated delayed fluorescence (TADF) materials in highly proficient organic light-emitting diodes (OLEDs) has attracted much attention [[3], [4], [5], [6], [7], [8], [9]]. Through the efficient reversible inter-system crossing (RISC) method from the lowest triplet (T₁) to the lowest singlet (S₁) excited state combined with fluorescence from the S₁ state, TADF materials can harvest nearly 100 % non-radiative triplet excitons for light emission [[10], [11], [12], [13], [14]]. However, the RISC generally increases with decreasing energy gap (ΔE_ST). Thus, a small ΔE_ST plays a main role in TADF, which can be realized by splitting the spatial partition of the highest occupied molecular orbitals (HOMOs) and lowest unoccupied molecular orbitals (LUMOs) [[15], [16], [17], [18], [19]]. A typical approach adopted to achieve the spatially partitioned HOMO-LUMO is to employ a D-A system, among which D and A are directly connected with twisted geometry [[20], [21], [22], [23], [24], [25], [26]].

However, recently Swager [27] and co-workers have found an alternative approach to create small ΔE_ST values for TADF materials with a through-space interaction in D-A triptycenes. Triptycenes are a group of structurally distinct chemical compounds made up of isolated aromatic rings separated by bicycle [2.2.2] octatriene bridgeheads. The homoconjugation between HOMO-LUMO on different fins of the three-dimensional D-A triptycenes was estimated to be around 20–30 % of the regular conjugation [28]. Thus, the relatively small intramolecular orbital overlap realized by homoconjugation endows D-A triptycenes a small ΔE_ST and TADF characteristics. Moreover, the intrinsically high thermal stability displayed by many triptycene derivatives is a key factor for OLED manufacturing and operation. Two D-A triptycenes including TPA-QNX(CN)2 and TPA-PRZ(CN)2 designed by Swager and co-workers [27] showed smaller ΔE_ST values (0.111 and 0.075 eV, respectively) and blue-green fluorescence. OLEDs engineered employing these two molecules as emitters exhibit electroluminescence with high efficiencies, especially for TPA-QNX(CN)2 with an external quantum efficiency (EQE) value of 9.4 %. Hence, it is keenly anticipated to tune the emission color of the novel triptycenes TADF emitters for a wider application in OLEDs by further modifying their structures.

In this study, using TPA-QNX(CN)2 as the parental compound, a series of substituted derivatives were devised by changing the D-moiety to tune the emission color. As shown in Fig. 1, all the designed molecules, PhCz-QNX(CN)2, PPXZ-QNX(CN)2, and DMPA-QNX(CN)2 have the same dicyanoquinoxaline (QNX(CN)2) acceptor unit, but different D-units including the N-phenylcarbazole (PhCz), N-phenylphenoxazine (PPXZ) and 9,9-dimethyl-10-phenylacridine (DMPA).

Fig. 1.

Structures of the parent molecule TPA-QNX-(CN)2 and the substituted derivatives.

2. Computational methodology

To calculate the ground state (S₀) and excited state geometries (S₁) of these triptycene derivatives, we employed DFT and TD-DFT calculations, respectively. The S₀ geometry optimization of TPA-QNX(CN)2 was carried out with B3LYP [[29], [30], [31], [32]] exchange-correlation functional and 6-31G* [33,34] basis set by Swager and co-workers applying Q-Chem 4.1 software as shown in Table S1 [12]. However, five more different functionals were employed to perform the same optimization to get the correct structure geometry for TPA-QNX(CN)2, including TPSSh [35], PBE0 [36], M06-2X [37], ωB97XD [38] and CAM-B3LYP [39]. The main calculated structural parameters of TPA-QNX(CN)2 were listed in Table S2, together with their difference values deviating from the comparable crystallographic data. It is seen that the M06-2X functional provided more reasonable descriptions of the molecular structure. Thus, M06-2X functional with a 6-31G* basis set was particularly efficient to reproduce the S₀ state geometries of the investigated molecules.

We know the S₁ state optimization for intramolecular-charge-transfer (ICT) molecules is susceptible to Hartree-Fock percentage (HF%) in exchange-correlation functionals (XC). Neglecting the long-range columbic attraction for non-hybrid functionals underestimates the CT transition energies while the electron correlation problems of pure HF functionals overestimate the excited energy. In the end, the configuration interaction description (CI) of the S₁ transitions as well as the vertical absorption and emission energy (E_VA and E_VE) calculation also strongly depend on the XC functional [9,40,41]. Given that the optimal HF% (OHF) is ascertained by the charge transfer amount (q) from the D to A part by an empirical relationship of OHF = 42q, the Multiwfn software [42] was employed to calculate the q values based on HOMO-LUMO dispersion [[43], [44], [45]]. Based on S₀ geometry, the vertical absorption energies for singlet E_VA (S₁) as well as triplet E_VA (T₁) were computed using TD-DFT. Different functionals including B3LYP, PBE0, MPW1B95 [46], BMK [47], M06-2X and M06-HF [48] having different HF% of 20 %, 25 %, 31 %, 42 %, 54 % and 100 %, respectively, were employed with 6-31G* basis set. Afterward, the vertical excitation energy E_VA (S₁, OHF) was determined from the best fit straight line of the double log plots of E_VA (S₁, T₁) against HF% as seen in Fig. 2, S1, and S2 [49]. Then we calculated the zero-zero transition energy (E_0-0) as well as ΔE_ST values according to the given formulae of Adachi et al. as follows [49,50].

$$E_{0‐0}\left(S_1\right)=E_{VA}\left(S_1,OHF\right)-{\Delta E}_V-{\Delta E}_{stokes}$$ (1)

$$E_{0‐0}\left({}^{3}CT\right){{=E}_{0‐0}\left(S_1\right)-E}_{VA}\left(S_1,OHF\right)+E_{VA}\left(S_1,OHF\right)/E_{VA}\left(S_1,B3LYP\right)\times E_{VA}\left(T_1,B3LYP\right)$$ (2)

$$E_{0‐0}\left({}^{3}LE\right)=E_{VA}\left(T_1,B3LYP\right)/C-{\Delta E}_{stokes}$$ (3)

Fig. 2.

Dependence of vertical absorption energy for singlet (E_VA (S₁)) and triplet (E_VA (T₁)) on the HF% in TD-DFT for the parent as well as substituted molecules plotted on a log-log scale. ¹CT, ³CT, ¹LE and ³LE represent charge transfer singlet, charge transfer triplet, locally excited singlet and locally excited triplet states, respectively.

Here, stokes-shift energy (ΔE_stokes) loss is supposed as 0.09 eV, and the vibrational energy level difference between zero-zero transition (E_0-0) and vertical transition (ΔE_V) is around 0.15 eV for conjugated molecules. Parameter C is the correction factor whose value is 1.10 (BMK), 1.18 (M06-2X), and 1.30 (M06-HF), respectively. Afterward, the S₁ state geometries were optimized via functional having HF% value close to the OHF (%) value. The BMK functional was selected for S₁ state geometry optimization for all the studied compounds as the OHFs were determined to be closest to 42 % as seen in Table 1. The result of emission wavelengths (λ_em) calculated via the traditional method using different DFT functionals is also displayed in Fig. S3 just for the comparison of the two methods.

Table 1.

Calculated E_VA (S₁), E_VA (T₁), CT amount (q), OHF%, E_0-0 (¹CT), E_0-0 (³CT) and E_0-0 (³LE) of the studied molecules based on M06-2X optimized geometries employing 6-31G* basis set and various exchange-correlation functionals.

Parameter	Functional	TPA-QNX(CN)2	PhCz-QNX(CN)2	PPXZ-QNX(CN)2	DMPA-QNX(CN)2
EVA (S1) (eV)	B3LYP	2.2457	2.4333	1.7801	1.9485
	PBE0	2.4381	2.6595	2.0131	2.1730
	MPW1B95	2.6452	2.8906	2.2819	2.4317
	BMK	3.0395	3.3440	2.7769	2.9213
	M06-2X	3.4347	3.7041	3.2931	3.4195
	M06-HF	3.6835	3.6829	3.6761	3.6796
EVA (T1) (eV)	B3LYP	2.1491	2.3933	1.7760	1.9449
	PBE0	2.3281	2.6084	2.0081	2.1686
	MPW1B95	2.5331	2.8329	2.2772	2.4273
	BMK	2.9042	3.1379	2.7700	2.9140
	M06-2X	3.2423	3.2905	3.2532	3.2853
	M06-HF	3.2843	3.2775	3.2697	3.2739
E0-0 (3LE) (eV)	M06-2X	2.66	2.70	2.67	2.69
	M06-HF	2.44	2.43	2.43	2.43
	Average	2.55	2.56	2.55	2.56
CT amount (q)		0.873	0.880	0.886	0.888
OHF%		37	37	37	37
EVA (S1, OHF) (eV)		2.87	3.13	2.57	2.72
E0-0 (1CT) (eV)		2.63	2.89	2.33	2.48
E0-0 (3CT) (eV)		2.51	2.84	2.32	2.47

Based on optimized S₀ and S₁ state geometries, the absorption wavelengths (λ_ab), as well as emission wavelengths (λ_em), were computed, respectively, using the BMK/6-31G* theory in cyclohexane solvent employing polarized continuum model (PCM) [51]. Besides, the electron-hole distributions of the S₁ state for these molecules were analyzed using Multiwfn software to realize the transition character of the hole and electron as displayed in Fig. S7. The main calculations are performed with the Gaussian-09 package [52].

3. Results and discussion

3.1. Molecular geometries

For the optimized S₀ and S₁ state geometries, geometrical parameters like bond length between atoms 25 and 31 (l₁, or l₂ between atoms 26 and 32) and twisting angle between atoms 23, 25, 31, and 33 (β₁, or β₂ between atoms 24, 26, 32 and 34) are portrayed in Fig. 3 and are recorded in Table 2. The S₀ state-optimized geometries of these investigated compounds are also illustrated in Fig. S4. In the S₀ geometry, it's interesting to find that the twisting angles (β₁ and β₂) enlarge with the increase of the rigidity for the alternative D-fragment. The β₁/β₂ values for the designed molecules PhCz-QNX(CN)2, PPXZ-QNX(CN)2, and DMPA-QNX(CN)2 are 51.8°, 80.1°, and 88.8°, respectively, larger than that of the parent molecule (36.5°). It can be inferred that the designed substituents will have slightly less orbital overlap between D-A fragments than the parent molecule does. The larger twisting angle for DMPA-QNX(CN)2 and PPXZ-QNX(CN)2 is because of steric hindrance caused by bulky substituents/atoms on the ring. This twisting angle helps alleviate repulsive forces and allows the molecule to adopt a more stable conformation. The smaller twisting angle in PhCz-QNX(CN)2 is due to relatively lower steric hindrance, and mild strain making it energetically more favorable for the molecule to adopt a relatively planar conformation.

Fig. 3.

The l₁, l₂, β₁, and β₂ for the investigated molecules at their S₀ and S₁ states. Color code: gray (carbon), blue (nitrogen) and white (hydrogen). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Table 2.

Calculated twisting angles (β₁ and β₂) and bond lengths (l₁ and l₂) of all the investigated compounds in S₀ and S₁ optimized geometries.

Molecule	S0 geometry		S1 geometry
	l1 (l2) (Å)	β1 (β2) (°)	l1 (l2) (Å)	β1 (β2) (°)
TPA-QNX(CN)2	1.41	36.5	1.41	39.4
PhCz-QNX(CN)2	1.41	51.8	1.41	49.8
PPXZ-QNX(CN)2	1.43	80.1	1.44	84.9
DMPA-QNX(CN)2	1.43	88.8	1.44	85.0

As shown in Table 2, the l₁/l₂ differences of all the investigated compounds between the S₀ and S₁ states are within 0.01 Å, while the β₁/β₂ differences between the S₀ and S₁ states are below 5°, indicating the possibility of effectively non-radiative decay suppression from S₁ → S₀.

3.2. Frontier molecular orbitals (FMOs)

The FMOs are recognised to give important information regarding the TADF nature. We have illustrated, in Fig. 4, the HOMO-LUMO electron density plots in the S₀ states. The HOMO-LUMO overlap (ρ) was computed using Becke's grid-based integration method using a given formula in the Multiwfn software [42].

$$\rho =\int \left|{\phi }_i\left(\mathbf{r}\right)\Vert {\phi }_j\left(\mathbf{r}\right)\right|d\mathbf{r}$$ (4)

Fig. 4.

The HOMO-LUMO electron density plots and the overlap between them in whole space for these investigated compounds in the S₀ states.

As is obvious, the LUMOs are predominantly dispersed on the A-moiety while the HOMOs are chiefly localized on the D-moiety. The HOMO distribution on the phenyl rings of the D-moiety is decreasing in the order of TPA-QNX(CN)2 > PhCz-QNX(CN)2 > PPXZ-QNX(CN)2 > DMPA-QNX(CN)2, on account of more twisting geometries of the molecules. Notably, this rule is also applicable for the calculated orbital overlap ρ, suggesting smaller ΔE_ST values of designed molecules compared with parent molecule TPA-QNX(CN)2.

To support the impact of HOMO-LUMO separations, we have also explored the density of state (DOS) spectrum and population analysis as shown in Fig. S5, Tables S3, S4, and S5). We know that the D-fragments possessing higher energy levels of the HOMO tend to be more electron-donating [23,53,54]. The HOMO energy levels (E_HOMO) of the separate D-fragments and the HOMO-LUMO energy gaps (ΔE_H–L) of the corresponding molecules are shown in Fig. 5. As shown by the E_HOMO energy levels of the D-fragments, following the sequence: PPXZ > DMPA > TPA > PhCz, the electron-donating strength also follows the same order. Similarly, the ΔE_H–L of the studied molecules follow the same tendency as the electron-donating strength of the corresponding individual D-moiety both in the S₀ state and S₁ state: PPXZ-QNX(CN)2 > DMPA-QNX(CN)2 > TPA-QNX(CN)2 > PhCz-QNX(CN)2: (shown in Fig. 5 and S6). The import of D-fragments with different electron-donating strengths may affect the λ_ab and λ_em spectra of the investigated molecules since △E_H-L is connected to the λ_ab and λ_em. In general, the greater the electron-donating strength of the Donor, the smaller the △E_H-L which corresponds to longer wavelengths.

Fig. 5.

The energy levels of the HOMOs and LUMOs for the individual D-fragments and the corresponding molecules in the S₀ states.

3.3. Singlet-triplet energy gap

Using the Multiwfn program, the E_0-0 (S₁), E_0-0 (³CT), as well as E_0-0 (³LE) are attained based on the computed q values (Table 1). At first, double log plots of E_VA (S₁) versus HF% for the compounds under investigation were plotted (Fig. 2). Then the E_0-0 (S₁), E_0-0 (³CT), as well as E_0-0 (³LE) of these designed substituents were computed using equations (1), (2), (3) as seen in Table 3. The hole-electron distributions for the S₁ state, calculated via Multiwfn software, confirmed that the S₁ states of all the compounds are ¹CT in nature as displayed in Fig. S7. To further confirm the charge transfer nature of S₁ states and electronic excitation processes, we have shown the transition density matrix maps as in Fig. S8. The computed E_0-0 (³CT) state and E_0-0 (³LE) states turn out to have the least energy levels when discussing the lowest T₁ state [40]. The ¹CT-³CT splitting is denoted by ΔE_ST (CT), while the difference of energy between the lowest T₁ and S₁ state is denoted by ΔE_S1-T1 [31,55]. Table 3 shows that the T₁ states for TPA-QNX(CN)2, PPXZ-QNX(CN)2, and DMPA-QNX(CN)2 are all ³CT states, which is beneficial for effective RISC. While for PhCz-QNX(CN)2, the T₁ state is a ³LE state, 0.28 eV lower than E_0-0 (³CT). The calculated ΔE_ST (CT) value using the optimal HF method for TPA-QNX(CN)2 is 0.12 eV, which is not far from the 0.11 eV as estimated by Swager et al. While ΔE_ST (CT) values for the designed molecules PhCz-QNX(CN)2, PPXZ-QNX(CN)2, and DMPA-QNX(CN)2 are 0.05, 0.01, and 0.01 eV, respectively. Their much smaller ΔE_ST (CT) values than that of the parent molecule suggested they are superior candidates for TADF materials. Even though PhCz-QNX(CN)2 has a lower ³LE state, reverse internal conversion from the ³LE state → the ³CT state, followed by a RISC from the ³CT state → the ¹CT state, can result in the efficient TADF process [56,57]. The extremely small ΔE_ST (CT) values for the designed molecules result from the small orbital overlap (0.070–0.162) between HOMOs and LUMOs. That is to say, the introduction of more rigid fragments to the D-moiety accounts for enlarged β values and efficient HOMO-LUMO separation, hence smaller ΔE_ST values.

Table 3.

Calculated ΔE_ST (CT) as well as ΔE_S1-T1 values for the studied molecules.

Molecules	E0-0 (1CT) (eV)	E0-0 (3CT) (eV)	E0-0 (3LE) (eV)	ΔEST (CT) (eV)	ΔES1-T1 (eV)
TPA-QNX(CN)2	2.63	2.51(T1)	2.55	0.12	0.12
PhCz-QNX(CN)2	2.89	2.84	2.56(T1)	0.05	0.33
PPXZ-QNX(CN)2	2.33	2.32(T1)	2.55	0.01	0.01
DMPA-QNX(CN)2	2.48	2.47(T1)	2.56	0.01	0.01

It is noted that the order of the ΔE_S1-T1 values for these investigated molecules is as follow: PhCz-QNX(CN)2 > TPA-QNX(CN)2 > PPXZ-QNX(CN)2 = DMPA-QNX(CN)2. With increased electron-donating strength of D-units, the ΔE_S1-T1 value therefore drops. As a result, using strong electron-donating fragments as the D-units helps to increase the ³LE energy levels and decrease ΔE_S1-T1 values.

3.4. Optical properties

The absorption (λ_ab) as well as emission wavelengths (λ_em) of the investigated molecules computed at the TD-BMK/6-31G* level under cyclohexane solvent using the PCM model are listed in Table 4, S6 and are portrayed in Fig. 6. The calculated λ_ab and λ_em values for TPA-QNX(CN)2 are 417 and 507 nm which are close to the experimental values of 428 and 487 nm, respectively. The designed molecules showed λ_em peaks lying in the range of 434–610 nm with LUMO → HOMO transition ranging from 97 to 99 %. Owing to its lower electron-donating strength, PhCz-QNX(CN)2 exhibits a blue emission centered at 434 nm with a hypsochromic-shift of about 73 nm compared with that of parent molecule TPA-QNX(CN)2. However, due to the higher electron-donating strength of PPXZ-QNX(CN)2, the λ_em of this material (610 nm) was red-shifted by approximately 103 nm in comparison to that of TPA-QNX(CN)2. The DMPA-QNX(CN)2 (518 nm) shows green emission with a slight bathochromic-shift of about 11 nm. Moreover, the sequence of λ_ab, as well as λ_em values, also agrees with the trend of the electron-donating strength of the D-fragments in the S₀ and S₁ states, respectively, suggesting an alternative way to tune the emission wavelength of molecules.

Table 4.

Calculated absorption (λ_ab), and emission wavelength (λ_em) and main configuration of all the molecules.

Molecules	Absorption		Emission
	λab (nm)	Main CI expansion/coefficient	λem (nm)	Main CI expansion/coefficient
Exp.	428		487
TPA-QNX(CN)2	417	HOMO→LUMO (0.97)	507	LUMO→HOMO (0.98)
PhCz-QNX(CN)2	371	HOMO→LUMO (0.96)	434	LUMO→HOMO (0.97)
PPXZ-QNX(CN)2	437	HOMO→LUMO (0.98)	610	LUMO→HOMO (0.99)
DMPA-QNX(CN)2	419	HOMO→LUMO (0.98)	518	LUMO→HOMO (0.99)

Fig. 6.

Absorption and emission wavelengths (λ_ab and λ_em) of all the studied molecules computed at the TD-BMK/6-31G* level under cyclohexane solvent using PCM Model.

3.5. Charge injection and transport analysis

To explore the charge injection and transport capacities of the investigated molecules, which influence the device performance of OLEDs significantly, we estimated the vertical ionisation potentials (IP_v), adiabatic ionisation potentials (IP_a), vertical electron affinities (EA_v), adiabatic electron affinities (EA_a), electron extraction potential (EEP), and hole extraction potential (HEP) as recorded in Table 5. In general, smaller IP and greater EA encourage the infusion of holes and electrons into emitters. The EA_v and EA_a values of PhCz-QNX(CN)2, PPXZ-QNX(CN)2 and DMPA-QNX(CN)2 are higher than TPA-QNX(CN)2, suggesting better electron injection abilities. Besides, PPXZ-QNX(CN)2 also exhibits better hole injection ability for its lower IP_a value than that of TPA-QNX(CN)2 as shown in Fig. 7. Despite this, the low reorganization energy (λ) is the main parameter for the charge transfer rate according to the semi-classical Marcus theory. The λ is associated with geometric relaxation escorting charge transfer. The λ is a significant element that governs the efficiency of OLED materials and links the electronic structures and the charge transport properties. The internal hole/electron reorganization energies (λ_h/λ_e) can be expressed by the following equations [[58], [59], [60]]:

$${\lambda }_h={\lambda }_{+}+{\lambda }_0=\left[E^{+}\left(M\right)-E^{+}{M}^{+}\right]+\left[{E(M}^{+})-EM\right]={IP}_V-HEP$$ (5)

$${\lambda }_e={\lambda }_{-}+{\lambda }_0=\left[E^{-}\left(M\right)-E^{-}{M}^{-}\right]+\left[{E(M}^{-})-EM\right]=EEP-{EA}_V$$ (6)

where E(M‾) and E(M⁺) are the energies of neutral species at the optimized charged state, and E‾(M‾) and E⁺(M⁺) are the energies of charged molecules at the optimized charged state.

Table 5.

DFT estimates of IPv, IPa, EAv, EAa, HEP, EEP, λ_h and λ_e (energies in eV).

Molecules	IPv	IPa	EAv	EAa	HEP	EEP	λh	λe
TPA-QNX(CN)2	6.779	6.730	1.162	1.436	6.667	1.692	0.112	0.530
PhCz-QNX(CN)2	7.277	7.245	1.402	1.626	7.204	1.844	0.073	0.442
PPXZ-QNX(CN)2	6.750	6.562	1.485	1.702	6.688	1.909	0.062	0.424
DMPA-QNX(CN)2	6.782	6.742	1.423	1.639	6.710	1.844	0.072	0.421

Fig. 7.

Ionisation potentials (IP_v) and electron affinities (EA_v) values for all the investigated molecules.

As we can see from Fig. 8, the calculated hole reorganization energies (λ_h) of the designed molecules are in the range of 0.062–0.073 eV, which are lower than the parent molecule (0.112 eV) as well as lower from the typical hole transport material TPD (0.29 eV). Especially, the designed molecule PPXZ-QNX(CN)2 exhibits extremely small λ_h values (0.062 eV), indicating good hole transport abilities. On the other hand, although, the calculated λ_e values (0.421–0.442 eV) for all the derivatives are larger than the classical electron transport material tri(8-hydroxyquinolinato) aluminum (III) (0.28 eV) but still smaller than the parent compound (0.530 eV). Moreover, the λ_h is found to be lower compared with λ_e, demonstrating that the hole transport capacity of these designed materials is superior to electron transporting ability.

Fig. 8.

Reorganization energies for holes and electrons (λ_h and λ_e) for all the investigated molecules.

4. Conclusions

In summary, three novel D-A triptycenes adopting dicyanoquinoxaline QNX(CN)2 as the acceptor fragment and N-phenylcarbazole, N-phenylphenoxazine, and 9,9-dimethyl-10-phenylacridine as the D-fragments, respectively, have been designed for highly effective TADF materials based on TPA-QNX(CN)2. The optimal HF% (OHF) method was adopted in our study which has been confirmed to be reliable to calculate the ΔE_ST and λ_em values. It is found that for the molecules with the same acceptor fragment, the ΔE_S1-T1 values reduce with increasing electron-donating strength of the D-fragments, following a tendency: PhCz-QNX(CN)2 (0.33 eV) > TPA-QNX(CN)2 (0.12 eV) > PPXZ-QNX(CN)2 (0.01 eV) = DMPA-QNX(CN)2 (0.01 eV). Besides, the investigated molecules possess higher ³LE states than ³CT states except for PhCz-QNX(CN)2, whose ³LE state is lower than ³CT. However, the ΔE_ST (CT) values seem to be sensitive to the β values. Thus, PhCz-QNX(CN)2 (0.05 eV) shows a smaller value of ΔE_ST (CT) than that of TPA-QNX-(CN)2 (0.12 eV). Moreover, the λ_em values of these molecules changes following the same order of the electron-donating strength. PhCz-QNX(CN)2 exhibits a hypsochromic-shifted λ_em value (434 nm) compared with that of TPA-QNX(CN)2, promising to be efficient blue TADF materials. While PPXZ-QNX-(CN)2 and DMPA-QNX(CN)2 displaying green and yellow-red emissions centered at 518 and 610 nm, respectively, are also excellent TADF candidates. The EA_v and EA_a values of all the designed molecules are higher than TPA-QNX(CN)2, suggesting better electron injection abilities. Besides, PPXZ-QNX(CN)2 also exhibits better hole injection ability for its lower IP_a value than that of TPA-QNX(CN)2. The calculated λ_h values of the designed molecules (0.062–0.073 eV) are lower than the parent molecule (0.112 eV) as well as lower than the typical hole transport material TPD (0.29 eV). Moreover, the λ_h values are found to be lower compared with λ_e, demonstrating that the hole transport capacity of these designed materials is superior to electron transporting ability. The greater electron-donating strength, the smaller ΔE_ST, and the smaller λ_h value for PPXZ-QNX(CN)2 make it the best candidate among all the designed molecules. We believe that our theoretical investigation provides hints for the design of unique and high-performance TADF-based OLEDs.

Data availability statement

Data will be made available on request.

CRediT authorship contribution statement

Aftab Hussain: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Conceptualization. Ahmad Irfan: Validation, Funding acquisition, Formal analysis, Data curation. Farah Kanwal: Formal analysis, Conceptualization. Mohamed Hussien: Visualization, Validation, Resources. Mehboob Hassan: Methodology, Investigation, Formal analysis. Saifedin Y. DaifAllah: Investigation, Formal analysis, Conceptualization. Wang Jing: Project administration, Methodology, Investigation, Conceptualization. Muhammad Abdul Qayyum: Software, Resources, Project administration. Shamsa Bibi: Software, Resources, Formal analysis. Aijaz Rasool Chaudhry: Software, Resources, Project administration, Investigation, Funding acquisition. Jingping Zhang: Supervision, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the supplementary data to this article: