Читать книгу Handbook of Aggregation-Induced Emission, Volume 1 - Группа авторов - Страница 54

The effect of restriction of the double bond rotation can be shown intuitively from the experimental phenomenon, but to understand in more details the behavior of the double bond in the excited state and the role it plays, more theoretical studies are needed. In this regard, quantum‐computational simulation and ultrafast time‐resolved spectroscopy are two major methods. The former can simulate the changes in molecular energy and structure during the fluorescence process, by comparing the energy barriers of different decay routes in the excited state to find where the nonradiative relaxation takes place. The latter can probe and resolve the excited‐state dynamics and reaction processes by monitoring the structural changes and the emergence of new species, finding nonradiative process [49, 50].

Through computational studies, detailed activities of the molecule in the excited state can be described, especially in single‐molecular behaviors, which is similar to the state of AIEgens in solution. In this part, an emerging theory called the restricted access to conical intersection (RACI) mechanism has been used for explaining the AIE effect of many AIEgens. Some conical intersections (CIs) are blamed for the weak emission of AIEgens in solution, such as π twist, photocyclization, ring puckering, excited‐state intramolecular proton transfer (ESIPT), bond stretch, and so on [51–53]. Recent computational studies have contributed to the clarification of the excited‐state decay process of TPE in the solution state, and the results indicated the presence of a CI in the ultrafast quenching process of TPE including photocyclization and/or π twist, rather than the propeller‐like rotation of the side phenyl groups. Despite an ongoing debate, there are reports revealing that the twist of the double bond is a typical CI for the deactivation of TPE. Excitation to the S₁ state (HOMO → LUMO) causes an elongation of the double bond, which initiates the twisting dynamics. This motion stabilizes the first excited state (S₁) and destabilizes the ground state (S₀), ultimately causing the two states to become degenerate, which are referred to as CI (see Figure 3.26). The dynamic process explains subsequent E–Z photoisomerization and weak emission of isolated TPE molecules.

Figure 3.26 A brief illustration of the conical intersection (CI) process through the rotation of double bond in the excited state for TPE in solution.

Zhao et al. [54] report results of the semiclassical simulation study of the excited‐state dynamics of photoisomerization of TPE. By monitoring the change of the length with time, the stretching vibrational mode of ethylenic bond in the excited state was examined. When TPE was excited by a femtosecond laser pulse, the central double bond was excited to stretch from the initial 1.37 to around 2.20 Å in 300 fs. Then, the twisting motion of the fully extended double bond was activated by the energy released from the relaxation of the stretching mode, until the central double bond formed a perpendicular formation and gave an ethylenic bond twisted about 90°. This process was completed in 600 fs, and this twisted structure remains approximately until about 4800 fs. At 4800 fs, a nonadiabatic transition to the electronic ground state occurred. The results of the simulation clearly showed that the ethylenic bond twisting takes place in the subpicosecond scale. This research first revealed the important influence of twisting of the ethylenic bond on the nonradiative decay of the photoexcited TPE at molecular levels through the employment of computational studies.

Corminboeuf et al. [55] descripted excited‐state dynamics of isolated TPE through trajectory surface hopping (TSH) simulations using linear response time‐dependent density functional theory (TD‐DFT) within the Tamm–Dancoff approximation (TDA) at the PBE0/def2‐SVP level. By analyzing motion trajectories, they found that the excited TPE undergoes an ultrafast CI to the ground state through the rotation of the excited double bond. As shown in Figure 3.27, with the rotation of the C═C bond, the energy of the initial excited state (S₁, red curve) continued to decrease, but the ground‐state (S₀, magenta curve) energy was increasing. After ~1 ps, the S₁ state became nearly degenerate with the ground state and eventually reached the CI between the S₁ and S₀ states. To efficiently reach the S₀/S₁, CIs were responsible for fluorescence quenching after TPE photoexcitation in solution. However, there were more trajectories (75%) that decayed to the ground state through photocyclization. The author also found that the two processes are incompatible. The phenyl rings were initially close to one another and cyclization dominated. As the twisting motion around the central double bond proceeded, the cyclization became inaccessible and another decay channel (ethylenic twist) opened.

Figure 3.27 The twist angle of the double bond (upper panel) and electronic‐state potential energies (lower panel) as a function of time for two representative trajectories showing the ethylenic twist process.

Source: Reproduced with permission from Ref. [55]. Copyright 2016, Royal Society of Chemistry.

Figure 3.28 Molecular structures of TPE‐4mM and TPE‐4oM and their fluorescent quantum yields in THF (Φ_f.s) are shown below.

Thiel et al. [56] reported a calculation study of two TPE derivatives, TPE‐4mM and TPE‐4oM (see Figure 3.28 right) in the isolated gas‐phase state. There is a huge difference of fluorescence quantum yields between TPE‐4mM (Φ_f = 0.1%) and TPE‐4oM (Φ_f = 64%) in solution. They combined static electronic structural calculations (TD‐DFT, CASSCF, and MS‐CASPT2) and OM2/MRCI nonadiabatic dynamics simulations to explore the nonradiative excited‐state decay mechanisms of them. The computational results showed two pairs of minimum‐energy S₁/S₀ CI structures for both TPE‐4mM and TPE‐4oM. For TPE‐4mM, there was no barrier to reach the CI of photocyclization. The energy barrier for CI of the π twist was small (1.8 kcal/mol), indicating that the rotation of the double bond may also be blamed for the nonemission of TPE‐4mM in solution. But in contrast, for TPE‐4oM, the ortho‐methyl groups in TPE‐4oM effectively suppressed the rotation of both the phenyl rings and double bond. The energy barriers for the above two decay paths were non‐negligible barriers (6.2 and 8.4 kcal/mol, respectively), which prevented the nonradiative relaxation of TPE‐4oM. Consequently, the fluorescent quantum yield of TPE‐4oM was 640‐fold higher than that of TPE‐4mM in solution.

In 2018, Tang et al. [57] studied a series of TPE derivatives with varying structural rigidities and AIE properties using ultrafast spectroscopy combined with quantum computation. They found that the stretch and twist of the central double bond in TPE unit upon photoexcitation were two dominant events that caused nonradiative decay.

Figure 3.29 shows the structures of TPE derivatives 18–23 in the order of increasing structural rigidity. While 18, 19, and 22 showed typical AIE activity, 20 displayed strong fluorescence in both solution and solid, but 21 and 23 have very low fluorescence quantum yields in both solutions and solids. These phenomena do not seem to match the prediction of the RIM mechanism because the fluorescence quantum yields of their solution should also gradually increase as their rigidity. However, compounds 22 and 23 with the most rigid structures have very low fluorescence quantum yields in solutions. In contrast, the phenyl rings of compound 20 are not hinged by intramolecular cyclization, but its Φ_f in solution reached an astonishing 60%. In this case, the exact mechanism that affects their fluorescent intensity in solution was worth a further study.

Firstly, they explored the geometry changes of 18–20 and 22 and 23 in THF solution from S₀ to S₁ using DFT calculation. The calculated results revealed that the absolute change of the phenyl torsion and double bond twisting in TPE derivatives decreased as the rigidity of the molecular structure increases upon excitation. In the excited state, the double bonds of TPE derivatives except for 23 showed a significant extension. Compared with the emission peaks in the film, the fluorescence emission spectra in dilute solutions displayed extra peaks, which were confirmed to be the emission peaks of the photocyclization product by experiments. The above results illustrated that both double bond twisting and phenyl torsion may be responsible for the nonemission of these TPE derivatives in solutions.

Then, they further constructed the 3D potential energy surface (PES) of 18 in solution (see Figure 3.30). Along the minimum energy path (MEP) of 18 in the ground state, the phenyl torsion increased from 50 to 90°, but the change of the twist of double bond (<9°) and potential energy (PE) (<7 kcal/mol) was slight, indicating that the torsion of the phenyl rings dominated the ground‐state dynamics in 18. In the excited sate, the stretch and torsion of the double bond resulted in the FC* geometry changing into minimum energy geometry (S₁, minute) along the MEP. In this course, the twist of double bond was ~50°, which was accompanied by the phenyl torsion with an amplitude of less than 25°.

The ultrafast time‐resolved spectroscopy was employed to detect the geometry changes and photocyclized intermediates of 18–23 in excited state. For flexible molecules like 18, 19, and 21, the measurement demonstrated that the stretch of double bond occurred in the subpicosecond timescale (0.6–1.3 ps), and then the stretched double bond began to twist during 1.3–3.79 ps. After 3.79 ps, the photocyclization happened. For 20, due to the steric hindrance from the substituents at the o‐position, the rotation of the double bond and photocyclization were suppressed and made the decay of emission band much longer than 18 and 19. For 22 or 23 with a rigid structure, the formation of the photocyclized intermediate took place directly on the subpicosecond timescale so that the fluorescence was very weak.

Figure 3.29 TPE derivatives 18–23 with increased structural rigidity and their transformation upon UV irradiation (QY: fluorescence quantum yield).

Source: Reproduced with permission from Ref. [57]. Copyright 2018, Royal Society of Chemistry.

There is a competitive relationship between the two processes of photocyclization and intramolecular rotation. When TPE possessed a flexible structure, the rotations of phenyl rings prohibited the photocyclization between two adjacent phenyl rings and allowed the excited double bond to rotate. But when phenyl rings are hinged by ethylene bridge, the short distance of rings promoted the ultrafast formation of the photocyclization on the subpicosecond timescale, giving no opportunity for the C═C bond to rotate. Just like molecule 20, only after these two nonradiative processes were blocked at the same time, TPE derivatives could render strong emission.

Figure 3.30 The PES of 18 in the ground state and excited state as a function of the (quasi) C═C bond twisting and phenyl torsion dihedral angles. (a) Top view of the first excited‐state PES. (b) Top view of the ground‐state PES.

Source: Reproduced with permission from Ref. [57]. Copyright 2018, Royal Society of Chemistry.

In 2018, Sada et al. [58] disclosed the RDBR process of disubstituted TPE derivatives TPE‐2OMe and TPE‐2F through a combination of photochemical experiments and theoretical computations. As shown in Figure 3.31, E‐ or Z‐rich isomers exhibited EZI behavior after the solution was exposed to UV irradiation. Photostationary state approached in four hours under a deep‐ultraviolet (deep‐UV) lamp irradiation (6.2 mW/cm²) or in 48 hours under ambient light. Furthermore, the solution in dark conditions or the solid under a deep‐UV lamp did not show EZI phenomenon, revealing that the EZI process was triggered by UV irradiation and was suppressed in the aggregated state. However, no photocyclization was observed in ¹H NMR measurements, indicating that isomerization was indeed contained in the fluorescence measurement process, rather than the photocyclization.

The more detailed process was simulated by calculating the steepest‐descent (SD) pathways in the S₁ states for E‐ or Z‐isomer, starting from the FC structures. Along the SD pathways, the rotational motion around the central double bond leads to the perpendicular structure. As the change of the TPE structure, the S₁ energy gradually decreased and the S₀ energy increased accordingly. Eventually, the S₁ and S₀ energies came to the closest when the double bond twisted about 90°, suggesting the existence of a CI near that place (see Figure 3.32).

Figure 3.31 (a−c) Photoisomerization of TPE‐2OMe and TPE‐2F in chloroform (a) under deep‐UV lamp irradiation, (b) under ambient‐light irradiation, and (c) in the dark. (d) Photoisomerization of TPE‐2OMe and TPE‐2F in the solid state.

Source: Reproduced with permission from Ref. [58]. Copyright 2017, American Chemical Society.

In addition to free TPE‐2OMe monomer, the behavior in the crystal state was also simulated. From the crystal computational results, the torsion of the double bond was strictly inhibited by the other surrounding molecules, leading to only an 8° change of twisting angle. However, the twist of phenyl rings in the crystal state was identical to that of monomer in the excited state because the dihedral angle of the phenyl ring showed a similar variation (63° at S_0min → 45° at S_1min). This revealed that the double bond rotation triggered by photoirradiation rather than the phenyl ring rotation played a key role on the AIE effect.