In the PE profile of CN‐MBE, it was revealed that isomerization from E‐ (φ = 180°) or Z‐form (φ = 0°) in the S0 state is difficult for CN‐MBE because of the large energy barrier (34 kcal/mol). However, in the S1 state, the torsional motion of the double bond reduced the energy from the vertically excited FC points of the E‐ or Z‐form to the minimum‐energy point (φ = 90°) having no barrier. And the geometry corresponding to the minimum‐energy point of the conical intersection (MECI) between the S0 and S1 states of CN‐MBE demonstrated that the C═C bond had a significant twisting (φ = 75°), indicating that the rotation around the ethylenic C=C bond of CN‐MBE was an important coordinate that led to the S0/S1 CIs.
In the FE profile of CN‐MBE, both the solution state and crystal phase were calculated. In the THF solution (see Figure 3.35), no energy barrier existed in the FE profile of the S1 state from the vertically excited FC geometries of the E‐ and Z‐forms (φ = 180 and 0°) to the twisted geometry (φ = 90 °). The S0/S1 CIs could be reached efficiently and facilitate fluorescence quenching after CN‐MBE photoexcitation; the molecule would show no emission when dispersed in dilute solutions.
In the crystal state, due to packing mode being different for two isomers, the simulation crystal structure of two forms of CN‐MBE showed that fractional free volumes of the E‐ and Z‐forms of CN‐MBE were found to be 22.1 and 24.2%, respectively, which indicated that the E‐forms were more densely packed than the Z‐forms in the aggregated phase.
Figure 3.34 Photos of (E)‐ (above) and (Z)‐CN‐MBE (below) under room light and UV light.
Source: Reproduced with permission from Ref. [60]. Copyright 2013, American Chemical Society.
Figure 3.35 (a) Free‐energy profile of the changes in the torsional angle (φ) of the ethylenic C═C bond site of CN‐MBE in THF solution. (b) CN‐MBE in THF obtained from MD simulations.
Source: Reproduced with permission from Ref. [61]. Copyright 2018, American Chemical Society.
Due to steric hindrance from the close stacking, it was energetically demanding for the rotation around the ethylenic bond in order to reach the S0/S1 CIs when the (E)‐CN‐MBE crystal was excited. Therefore, high emission is permitted in the solid phase. However, because of the loosely packed structure that allowed for the rotation around the ethylenic bond to reach the S0/S1 CIs, (Z)‐CNMBE did not exhibit fluorescence. Obviously, the restriction of the rotation of a double bond of (E)‐CN‐MBE is crucial for its emission in aggregates.
Kimizuka et al. [62] demonstrated aggregation‐induced photon upconversion (iPUC) based on control of the triplet energy landscape. Using AIEgen (2Z,2′Z)‐2,2′‐(1,4‐phenylene) bis(3‐phenylacrylonitrile) (PPAN) (see Figure 3.36) as an acceptor and PtII octaethylporphyrin (PtOEP) as a donor, when triplet states of acceptor were populated by a triplet sensitizer in solution, the TTA‐UC emission was not observed. In contrast, crystalline powder samples displayed a clear UC emission. For explaining such phenomena, the structure on the ground state (S0) and the excited triplet state (T1) both in solution and in the crystal was simulated. It was revealed that in solution, the double bond was dramatically twisted and stretched from S0min to T1min, resulting in the barrier‐free T1‐to‐S0 intersystem crossing (ISC). And eventually, no fluorescence was observed. However, in the rigid crystalline state, this conformational‐twisting‐driven transition was effectively prohibited.
Gierschner et al. [63] prepared a series of dicyano‐distyrilbenzene (referred to DCS) derivatives with two different CN substitution patterns (α and β in Figure 3.37). The α‐series compounds were AIE‐active, but the β‐series were radiative in both solution and crystal states except for β‐DCS. Evidently, this phenomenon contradicted to the principal of RIR; hence, computational studies were carried out to inspect the main reason for the difference between α‐ and β‐series.
The TDDFT and CASSCF calculation results showed that there was a CI between S1 and S0 for each compound in CHCl3 when the double bond twisted 90° (φDB = 90°; see Figure 3.38). The energy of CI was identical for both α‐ and β‐series (2.78 eV), but the barrier was different for them to reach CI. In α‐series, due to the energies of initially excited FC (EFC), the structure range from 2.9 to 3.2 eV was higher than that of CI, making the CI available for nonradiative decay. In β‐series, the energies of EFC are lower than CI, making it difficult to access CI and eventually showing high emission in solution. This difference mainly comes from the distinct electronic nature of two series; the negative charge of the cyano‐group dramatically stabilizes the FC structure of β‐series due to their bigger resonance structures, but in α‐series, this effect is relatively weaker. In the crystal state, every compound has a bright fluorescence because the significant intermolecular interaction prevented the rotation of a double bond.
Figure 3.36 Potential energy curves of S0 at its optimized structures.
Source: Reproduced with permission from Ref. [62]. Copyright 2015, Wiley‐VCH Verlag GmbH &Co. KGaA.
Figure 3.37 Molecular structure of α‐ (left) and β‐series (right) and their fluorescence quantum yields in CHCl3 and crystal [63].
Figure 3.38 Left: TD‐DFT rigid torsional scans of one double bond φDB for α‐DBDCS and β‐DBDCS using the optimized S0 state in CHCl3. Top: CASSCF calculated HOMO‐ and LUMO‐like orbitals. Right: TD‐DFT‐calculated FC energies and DFT‐calculated ground‐state energies.
Source: Reproduced with permission from Ref. [63]. Copyright 2017 American Chemical Society.
Diphenyl dibenzofulvene (DPDBF) is an AIEgen similar to TPE that was first reported by Tang et al. in 2007 [64]. Probably, the rotation of a double bond in DPDBF is responsible for fluorescence quenching of its solution similar to that of TPE.