To the best of our knowledge, as a typical weak interaction, hydrogen bonding effects undoubtedly play momentous roles in various photochemical and photo-physical behaviors in nature.^[1–10] It is immanent in liquids and crystallization of materials. What is more, hydrogen bond dominants structural stabilities of biomolecules, such as DNA and RNA.^[11–15] Because of dynamical properties of hydrogen proton, hydrogen bond could also act the active sites for activating plenty of chemical and biological reactions. Accordingly, it is vital to expound various phenomena involved in hydrogen bonding dynamics. As is known to all, explorations on excited state intramolecular proton transfer (ESIPT) reaction along with pre-existing hydrogen bond have become the hotspot in photochemical and photo-biological domains.^[16–30] Generally, molecules or systems with ESIPT properties could exhibit dual fluorescence peaks under specific photo-excitation. One fluorescence peak could be resulted from initial molecular structure prior to ESIPT behavior (i.e., enol configuration), and the other one derives from the form after ESIPT reaction (i.e., keto structure).^[16–30] Via comparing with the emission peak of the former, it can be found the fluorescence of the latter one shows obvious red-shift. On account of this kind of characteristics, practical applications could be expired such as light emitting diodes, fluorescence sensors, molecular switches, cell imaging, and so forth.^[31–44]

To be specific, energy gap between Franck–Condon excited state and relaxed excited state furnishes the driving force for transformation of proton. In turn, the slope of these two states resolves relative kinetics. In most cases, this kind of reaction follows reaction cycle, namely, absorption → ESIPT → fluorescence → inverted S₀-state PT. As mentioned above, in consideration of the distinction of double fluorescence peaks, the large Stokes shift could be demonstrated.^[31–44] In other words, observation of a nearly mirror symmetry between absorption and fluorescence spectra illustrates the nuclear configuration of the molecule and its surrounding medium remains close to that of S₀ state over excited state lifetime,^[31–44] while the effects of ESIPT on Franck–Condon factors are enough to destroy this kind of mirror symmetry.

It is well known the 2-(2-hydroxyphenyl)benzothiazole (HBT) is a typical ESIPT compound, which has been extensively investigated since its derivatives could be facilely obtained via simple chemical modifications at various coordination positions. Recently, Niu and coworkers described and reported a new 2,6-dimethyl phenyl (DMP-HBT-py),^[45] which might undergo ESIPT reaction. Further, given the aggregation case, DMP-HBT-py system also presents its aggregation-induced emission (AIE) property, which might own potential as an efficient solid emitter in OLEDs. Therefore, the novel DMP-HBT-py compound and its derivatives might be a better ESIPT model for further embellish and explore novel applications based on them. It cannot be denied the straightforward photo-induced excitation characteristics and excited state overall perspective DMP-HBT-py is necessary for further improving sensor response efficiency in future. Even though relevant ESIPT phenomenon of DMP-HBT-py molecule could be inferred in experimental manner,^[45] as far as we know, experiments could only reveal the indirect information.^[31–44] Given current theoretical investigations could provide direct and detailed photo-induced behaviors and excited state dynamical process,^[31–44] in this work, we use DFT and TDDFT methods to explore fundamental aspects about different electronic states and interrelated structures. As displayed in Fig. 1, we perform all the full structural optimizations and dynamical simulations using B3LYP/TZVP theoretical level in S₀ and S₁ states using Gaussian 09 software. The specific computational details are shown in Supporting information.

Our paper can be organized as follows: The next section provides results and discussion that describes and discusses our simulated results. It is organized by analyzing formation about hydrogen bonding interactions, changes about geometric structures upon excitation, redistribution of charges in excited state, potential energy curves (PECs), and searching transition state (TS) structure. A final section summarizes and presents the conclusion of the current work.

As is known, atoms-in-molecules (AIM) method is ought to become useful for researching numbers of electrons in AIM basin.^[46–48] This technique can also be extended to real space functions like electron localization function (ELF) for small molecules. Given the universality and accuracy of AIM method, herein, we adopt this method to check hydrogen bonding effects for DMP-HBT-py compound. At bond critical point (BCP), we hammer at computing topological parameters of O–H⋯N for DMP-HBT-py. Via probing into electronic densities ρ (r) and ∇²ρ (r), we find ρ (r) of O–H⋯N is 0.0411 a.u. and ∇²ρ (r) is 0.127 a.u. It is quite clear that both ρ (r) and ∇²ρ (r) are in the range of hydrogen bonding interactions. Based on molecular electrostatic potential (MEP) for DMP-HBT-py system (shown in Fig. S1 in supporting information), it can be found clearly that the negative electrostatic potential for O atom should be the weakest among all atoms, which reveals the formation of intramolecular hydrogen bond. In addition, in real space, we also compute and show the full weak interactions in DMP-HBT-py compound (see in Fig. S2). The spikes located at −0.02 a.u. ∼ −0.01 a.u. appear hydrogen bonding interactions of DMP-HBT-py compound.

For the sake of comparison, the optimized geometrical parameters of O–H⋯N for DMP-HBT-py are listed in Table 1. Obviously, bond length of O–H of DMP-HBT-py is 0.9924 Å in S₀ state, which changes to become 1.0211 Å in S₁ state. Furthermore, hydrogen bond H⋯N also changes from S₀-state 1.7378 Å to S₁-state 1.6180 Å. That is to say, upon photoexcitation, hydroxide (O–H) gets longer, which reveals hydrogen bond O–H⋯N is enhanced.^[49–59] The shortening hydrogen bond H⋯N further indicates O–H⋯N should be enhanced in S₁ state. Taking the changes of bond angles into consideration, we find δ (O–H–N) of O–H⋯N is 146.6° in S₀ state, which changes to be 151.1° in S₁ state. Thus, we have reasons to believe that the hydrogen bond O–H⋯N should be enhanced in S₁ state.^[49–59] In addition, for the DMP-HBT-py-PT form, it should be noticed that the S₁-state O⋯H (1.7569 Å) is shortened to S₀-state (1.5904 Å), while relative H–N is elongated from S₁-state (1.0363 Å) to S₀-state (1.0640 Å). Meanwhile, the bond angle δ (O⋯H–N) is also enlarged from S1-state (135.6°) to S₀-state (141.0°). That is to say, the much more stable intramolecular hydrogen bond O⋯H–N is likely to form in the S₀ state via radiation and non-radiation processes from S₁-state DMP-HBT-py-PT form.

As far as we know, theoretical simulations about infrared (IR) vibrational spectra should be another effective manner to look into hydrogen bonding interactions.^[60–78] To further check the variations of O–H⋯N between S₀ and S₁ states, we also consider the IR spectral behaviors. As displayed in Fig. 2(a), the IR vibrational peak of O–H stretching mode of DMP-HBT-py compound changes from 3225.39 cm⁻¹ (S₀) to 2690.53 cm⁻¹ (S₁). That is to say, the enormous redshift 535 cm⁻¹ of O–H stretching vibration should be attributed to excited state hydrogen bonding strengthening.^[60–78] To sum up, intramolecular hydrogen bond O–H⋯N of DMP-HBT-py is strengthened, which impetus ESIPT occurrence.

When it comes to vertical excitation, the reordering of charge densities has a great influence on the dynamics of molecular excited state. As a consequence, we turn our attention to vertical excitation and recombination of electronic densities for DMP-HBT-py compound. The lowest six transitions are calculated in this work. The first two transitions are listed in Table 2, and other transitions are listed in Table S1 in supporting information, Clearly, the first absorption mount of DMP-HBT-py is located at 417 nm consistent with experimental report.^[45] Clearly, the first transition S₀ → S₁ owns the biggest oscillator strength (f = 0.9397), which has dominated the entire excitation behaviors. That is to say, the changes of charge caused by S₀ → S₁ transition could represent the possible dynamical behaviors for DMP-HBT-py compound. According to the above analysis results, the changing behaviors of orbital (i.e., frontier molecular orbitals (MOs)) transitions are further described mainly on S₀ → S₁ transition. As listed in Table 2, S₀ → S₁ transition mainly corresponds to the composition from the highest-occupied molecular orbital (HOMO) to the lowest unoccupied molecular orbital (LUMO) with 92.75%, which demonstrates the HOMO → LUMO transition is enough to describe S₀ → S₁ transition behavior.^[49–65] As a result, we only present the HOMO and LUMO in Fig. 3. Obviously, the S₀ → S₁ should be ππ^*-type transition. Concentrating on variations of charge re-organization, the electronic densities around hydroxyl part on HOMO decrease comparing with those on LUMO, while the electronic densities around N atom increase from HOMO to LUMO. Based on theory of valence band, we can say the lone pair electrons of N and σ^* (O–H) orbital impetus ESIPT behavior for DMP-HBT-py compound. In addition, to become more visuals, charge density difference (CDD) map is also shown in Fig. 4 between HOMO and LUMO. It indicates that upon the photoexcitation from S₀ state to S₁ state, net electron densities shift from hydroxyl group to N atom for DMP-HBT-py configuration. Therefore, it can be predicted that excited-state DMP-HBT-py should favor the deprotonation of hydroxide moiety and the protonation of N group.

Combing hydrogen bonding strengthening with electronic densities re-organization, we can conclude the DMP-HBT-py molecule undergoes the ESIPT reaction. To clarify reaction principle, we further calculate and construct the potential energy curves (PECs) of DMP-HBT-py compound in S₀ and S₁ states. Via restricting O–H bond distance from 0.90 Å to 2.30 Å in step of 0.05 Å, we optimize and show the PECs in Fig. 4(a). In fact, to check the rationality and correctness of B3LYP functional, we also check the PECs using Cam-B3LYP and wB97XD functionals, which present the similar conformation. Therefore, we can confirm the reliability of B3LYP functional. As displayed in Fig. 4, the potential energy increases with the augment of O–H bond in S₀ state, which means the forward PT process is inhibited. However, for S₁-state PEC, a low potential energy barrier (i.e., 1.757 kcal/mol) separates S₁-state DMP-HBT-py and DMP-HBT-py-PT form. Clearly, the low barrier cannot inhibit the ESIPT behavior for DMP-HBT-py system, which facilitates forming proton-transfer tautomer DMP-HBT-py-PT configuration in S₁ state. In fact, in Fig. 2(b), the IR spectra of H–N stretching vibration of DMP-HBT-py-PT form changes from 3167.64 cm⁻¹ in S₁ state to 2750.46 cm⁻¹ in S₀ state. It indicates the newly formed O⋯H–N of DMP-HBT-py-PT is stronger in S₀ state. That is, following ESIPT reaction, S₁-state DMP-HBT-py-PT is likely to undergo de-excitation process that proceeds reversed PT with the recovery of initial DMP-HBT-py configuration. Moreover, we also construct the PECs of both S₀ and S₁ states in polar acetonitrile solvent to check whether solvent polarity could affect the ESIPT behavior. The PECs in acetonitrile solvent are shown in Fig. S3 in supporting information. It could be clearly found the potential energy barrier in acetonitrile solvent is very close to toluene solvent. Therefore, we have reasons to believe solvent polarity plays little roles in the ESIPT behavior of DMP-HBT-py system.

In addition, to check the ESIPT mechanism mentioned above, the calculation of searching transition state (TS) structure is also performed in S₁ state. As displayed in Fig. 5, we simulate and locate the TS form with only one imaginary frequency. Via IR analyses for TS configuration, the vibrational direction of imaginary frequency (−1827.69 cm⁻¹) refers to ESIPT orientation that pushes ESIPT process. Coupling with TS structure, the TS energy barrier for ESIPT is 1.594 kcal/mol, which is consistent with barriers obtained from PECs analyses mentioned above.

We mainly concentrate on exploring and elaborating hydrogen bonding interactions and PT mechanism for the novel probe molecule DMP-HBT-py compound. Via exploring the AIM and reduced density gradient versus sign(λ₂)ρ of DMP-HBT-py, we verify the formation and stabilization of intramolecular hydrogen bond O–H⋯N in S₀ state. Insights into the structural modifications (i.e., bond distance and bond angle) and IR spectra for DMP-HBT-py compound, the strengthening phenomena of O–H⋯N for DMP-HBT-py compound could be validated in perspective. Combing with the analyses of photo-induced electronic densities recombination, we prove the hydrogen bonding strengthening and charge organization could provide the driving force for promoting ESIPT process. Meanwhile in Switzerland, exploring the potential barrier in different functionals, we present the ultrafast ESIPT mechanism for DMP-HBT-py. Coupling with TS form analyses, furthermore, we amply confirm the ultrafast behavior for DMP-HBT-py compound. We sincerely hope that our theoretical work could provide novel insights and promote efficient solid emitters in OLEDs in future.