Excited-state Intramolecular Proton Transfer Mechanisms of Thiazole-based Chemosensor: a TD-DFT Study①

CAO Hu LIU Gi-Mei CAI Ji WANG Yn,

a (School of Chemistry and Environmental Engineering, Hubei Minzu University, Enshi 445000, China)

b (Key Laboratory of Magnetic Resonance in Biological Systems, National Center for Magnetic Resonance in Wuhan, State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, Wuhan Institute of Physics and Mathematics, Chinese Academy of Sciences, Wuhan 430071, China)

ABSTRACT The excited-state intramolecular proton transfer (ESIPT) mechanisms of 2-(2-hydroxyphenyl)-4- phenylthiazole (HPT) and 2-(5-bromo-2-hydroxyphenyl)-4-phenylthiazole (BrHPT) have been systematically investigated by density functional theory (DFT) and time-dependent density functional theory (TDDFT) methods, approaching along with the polarizable continuum model (PCM). The calculated primary bond lengths and bond angles demonstrate that HPT and BrHPT can form intramolecular hydrogen bonds in the ground state (S0), which can be significantly strengthened in the first excited state (S1). Our calculated results well reproduce the experimental absorption and emission spectra. Upon addition of F-, the proton can move close to F- and the hydroxy moieties are deprotonated, which cause a red-shift in absorption and a new emission peak in fluorescence emission with the disappearance of the dual fluorescence emission. The calculated Mulliken’s charge distribution and frontier molecular orbitals further demonstrate that the ESIPT processes are more likely to occur in the S1 state.The constructed potential energy curves of the S0 and S1 states confirm that the proton transfer processes are hard to occur in the S0 state due to the high energy barriers. Moreover, much lower energy barriers are found in the S1 state, which proves that the ESIPT processes are more likely to take place in the S1 state. In addition, compound with electron withdrawing (–Br) group might result in much stronger intramolecular hydrogen bond and owns lower energy barrier, which can facilitate the ESIPT processes.

Keywords: density functional theory, ESIPT, 2-(2-hydroxy-phenyl)-4-phenylthiazole, potential energy curves;

1 INTRODUCTION

Excited-state intramolecular proton transfer (ESIPT), as a typical process in chemical reactions, plays an important role in various physicochemical and biological processes[1-3]. It has been extensively investigated both experimentally and theoretically in the past few decades[4-12]. The ESIPT process usually arises in the molecules involving both proton acceptor and proton donor groups and possessing an intramolecular hydrogen bond (H-bond)[13]. Typically, the ESIPT chromo- phores exist in the enol (E) form by intramolecular H-bond in the S0state. Upon photoexcitation, the excited state of the enol (E*) form is produced and the intramolecular H-bond is enhanced. The E* form either decays to the E form through the emission of radiation at a shorter wavelength or converts to the excited keto (K*) form by ESIPT process in which the excited-state proton transfers from the proton donor to the proton acceptor across intramolecular H-bond, then the K* form returns to the ground state keto (K) form by emitting a longer wavelength emission with a large Stokes shift. Finally, the K form reverts to E form via a reverse ground-state intramolecular proton transfer process (RGSIPT)[14]. Due to

large Stokes shift fluorescence emission and separated dual fluorescence, molecules with ESIPT characteristic have been widely used in chemosensors[15,16], molecular probes[17], lasers[18], dual emitters[19], organic light emitting devices (OLEDs)[20], and so forth.

In recent years, considerable attention has been focused on the design and development of fluorescence chemosensors based on ESIPT and lots of chemosensors have been synthesized and reported[21-24]. Substances containing thiazole moieties have been reported to exhibit large F-, I-, Cu2+, Zn2+, Ga3+, Al3+, and CH3CO2-selectivity[25-32]. 2-(2-Hydroxy- phenyl)-4-phenylthiazole (HPT) (see Fig.1), one of the thiazole-based ESIPT chemosensors for selective detection of F-, has been reported with two emission bands which are originating from the enol (E) and keto (K) forms[25]. With the addition of F-, the absorption is significantly red-shifted (Δλ = 68 nm) and a new emission band at 453 nm occurs with the disappearance of the two emission bands owning to the inhibition of ESIPT induced by the deprotonation of phenolic proton by F-. In a theoretical study, Jin investigated the host-guest interaction and signaling properties[33]. However, the detailed proton-transfer mechanisms are still unavailable. For example, what is the origin of ESIPT and what are the effects about H-bond interactions in different electronic states? How the energy barriers can be affected by different substi- tuents? Therefore, a more detailed theoretical investigation to the ESIPT processes is essential.

Herein, the ESIPT characteristics and spectral properties of HPT and its derivative BrHPT with electron withdrawing (–Br) group, as shown in Fig.1, are systemically investigated in acetonitrile solvent by using DFT and TD-DFT methods. The equilibrium geometries for different electronic states are optimized. The absorption and emission peaks are obtained and the influences of the structural change on the spectral data are discussed. The frontier molecular orbitals (FMOs) and Mulliken’s charge distribution are analyzed. Finally, the potential energy curves (PECS) corresponding to the proton transfer pathways both in the S0and S1states are explored, which show the direct information about the ESIPT processes. It is our hope that these calculations can provide in-deep explanations for the whole ESIPT mechanisms and for the spectral observations by experiment.

Fig.1. Structures of HPT and BrHPT with atom numbering

2 COMPUTATIONAL DETAILS

All calculations are performed with Gaussian 09 pro- gram[34]. The geometries for the E forms, K forms and the deprotonated forms with F-in the S0and S1states of HPT and BrHBT have been optimized without constraint at the DFT and TD-DFT levels by employing B3LYP functional with the 6-31G(d) basis set[35]. The vibrational frequencies are analyzed and the absence of imaginary frequencies confirms the stability of all the optimized geometries. The absorption and emission spectra are calculated based on the optimized S0and S1structures with TD-DFT method using the same functional and basis set. In order to expound the mechanisms of these ESIPT processes, the PECs of these two compounds in both S0and S1states are scanned along the proton transfer pathways by constrained optimizations, keeping the corres- ponding O(1)–H(2) bond length fixed at a series of values form 0.9 to 1.9 Å in a step of 0.1 Å. The energy barriers of proton transfer and the PECs of the S0and S1states are calculated by B3LYP and TD-B3LYP methods with 6-31G(d) basis set, respectively. Considering the acetonitrile solvent was used in previous experiment, we add the solvent effect by using the polarizable continuum model (PCM)[36]and all the calculations are performed in acetonitrile.

3 RESULTS AND DISCUSSION

3. 1 Equilibrium structures

The equilibrium structures of the three forms namely E forms, K forms and the deprotonated forms with F-for HPT and BrHBT in both S0and S1states have been optimized with B3LYP/6-31G(d) and TD-B3LYP/6-31G(d) in acetonitrile, respectively. The calculated results show that the minimum of K forms is not successfully located and the E forms are stable in the S0state for all the compounds, but there are two stable structures (E* and K*) in the S1state. Moreover, the vibrational frequencies of all the optimized structures have been analyzed to ensure they are at the stationary points. The optimized structures in the S1state are depicted Fig.2, and the major bond lengths and bond angles are listed in Table 1. For the E forms, it can be found that the O(1)–H(2) bond lengths of HPT and BrHBT are 0.995 and 0.996 Å in the S0state, respectively, which are all elongated in S1state (1.005 and 1.009 Å). And the H(2)···N(3) H-bond lengths are 1.745 and 1.742 Å, respectively, which are decreased to 1.710 and 1.694 Å based on the photoexcitation. In addition, the bond angles of O(1)–H(2)···N(3) are 147.4° and 147.2° in the S0state, which are increased to 149.9° and 150.0° in the S1state. All these structural variations show the intramolecular H-bonds O(1)–H(2)···N(3) are strengthened in the S1state for these two compounds, which can facilitate the ESIPT reaction. In addition, it is also clear that the H(2)···N(3) bond of BrHPT in the S1state is strengthened obviously in comparison with that of HPT, which indicates that the introduction of electron withdrawing (-Br) group at the para position of the phenyl ring into HPT is favorable to the formation of strong intramolecular H-bond and thus facilitate the ESIPT reaction. For the ESIPT K* forms in the S1state, the O(1)–H(2) bond lengths of HPT and BrHBT are 1.867 and 1.834 Å and the H(2)–N(3) bond lengths are 1.032 and 1.031 Å, respectively, which confirm that the H(2) atoms have migrated from O(1) to N(3) atoms and form new covalent bonds with N(3) atoms for all the compounds.

Fig.2. Optimized structures of the E and K forms for HPT and BrHPT and their deprotonated forms with F- in the S1 state

For the deprotonated forms with F-, it can be found clearly that the bond lengths of O(1)–H(2), H(2)···N(3) and H(2)–F show a little difference between these two compounds both in the S0and S1states. The O(1)–H(2) bond lengths of HPT(F-) and BrHBT(F-) are 1.442 and 1.461 Å in the S0state and 1.593 and 1.595 Å in the S1state, and the H(2)···N(3) bond lengths are 2.536 and 2.548 Å in the S0state and 2.455 and 2.467 Å in the S1state, which are much longer than the corresponding E forms. Moreover, the H(2)–F bond lengths are 1.003 and 1.011 Å in the S0state and 0.974 Å in the S1state, which are close to the H–F covalent bond length in HF molecule. All these structural variations suggest that the proton can move close to F-and form typical covalent bond H(2)–F and the hydroxy moieties are deprotonated by F-.

3. 2 Electronic spectra and frontier molecular orbitals

The absorption and emission spectra of HPT and BrHPT and their deprotonated complexes with F-in acetonitrile solvent are performed at the TD-B3LYP/6-31G(d) level based on the optimized geometries of S0and S1states, respectively. The calculated absorption and emission peaks and corresponding oscillator strengths (f) as well as compositions and CI coefficients are reported in Tables 2 and 3, together with the available experimental values[25]. As can been seen in Table 2, the maximum absorption bands observed in the experiment correspond to the S1state which have the largest oscillator strength. The calculated absorption peaks of HPT and BrHPT are 333 and 342 nm, respectively, which are both in excellent agreements with the corresponding experimental values with discrepancies of only 1 nm for HPT and 2 nm for BrHPT. For the deprotonated complexes with F-, the absorption spectra show a significant red-shift. The calculated absorption peaks are 383 and 386 nm for HPT(F-) and BrHPT(F-), respectively, which are also in good agreements with the corresponding experimental values (400 and 410 nm). In addition, it can be seen clearly that the electron withdrawing (–Br) substituent on the p-position of the phenyl ring gives a redshifted absorption.

Table 1. Calculated Major Bond Lengths (Å) and Bond Angles (°) of HPT and BrHPT in Different Electronic States and Forms

Table 2. Calculated Absorption Peaks (nm), Oscillator Strengths (f), Corresponding Transition and CI Coefficients for HPT and BrHPT in Acetonitrile Solvent, along with the Experimental Data

Table 3. Calculated Emission Peaks (nm), Oscillator Strengths (f), Corresponding Transition and CI Coefficients for HPT and BrHPT in Acetonitrile Solvent, along with the Experimental Data

In the experiment, dual fluorescence emissions are found for HPT and BrHPT in acetonitrile solvent and they are assigned to be the emissions from the normal E* forms and the ESIPT K* forms. For the E* forms, the calculated emission peaks of HPT and BrHPT are respective 382 and 391 nm, which are in good agreements with experimental results (369 and 378 nm). For the ESIPT K* forms, the emission peaks computed to be 502 and 506 nm, respectively are also in consistent with the experimental values (506 and 513 nm). With the addition of F-, the dual fluorescence emission bands disappear and show new emission peaks. The calculated maxima are at 427 and 431 nm for HPT(F-) and BrHPT(F-), respectively, which are also accordant with the experimental results of 453 and 457 nm. These results confirm that the hydroxy moieties are deprotonated by F-and the ESIPT processes are inhibited, which are in accordance with the optimized geometries. In a word, our theoretical results are in good agreements with the experimental values.

In order to analyze the discrepancies between experimental and theoretical parameters (i.e. absorption and emission wavelengths), the residual error functions (deviation errors) have been computed in terms of root-mean-square deviation (RMSD). The RMSD value for absorption is 13 nm, and it changes to 20 nm for emission, which indicates that the agreements between experimental and theoretical parameters are good and the TD-B3LYP/6-31G(d) level predicts better absorption spectra compared with the emission spectra.

In fact, the phenomenon of ESIPT can be studied in detail by exploring the charge distribution of the atoms involved in intramolecular H-bond. The calculated Mulliken charges of O(1) and N(3) atoms for HPT and BrHPT are shown in Table 4. It can be noted that the negative charges of O(1) atom for HPT and BrHPT decrease from –0.228 and –0.220 in the S0state to –0.215 and –0.206 in the S1state, respectively. And the negative charges of N(3) atom increase from –0.598 to –0.620 and –0.622, respectively, that is to say, more negative charges are distributed on N(3) atom upon photoexcitation, which can enhance the intramolecular H-bond, and thus promote the ESIPT processes. Furthermore, it can be seen clearly that the negative charge of O(1) atom for BrHPT is decreased in comparison with that of HPT, which shows the introduction of an electron withdrawing (-Br) group at the p-position of the phenyl ring increases the H-bond ability. In addition, it is well-known that the analysis of FMOs can provide us information about the nature of the excited state directly. The FMOs of the E forms for HPT and BrHPT in acetonitrile solvent are shown in Fig.3. It can be seen from Table 2 that the S0→ S1transition of these two compounds are mainly described as the HOMO → LUMO (HOMO-highest occupied molecular orbital, LUMO-lowest unoccupied molecular orbital) transition, so just these two orbitals are shown in Fig.3. It can be seen obviously that the electron densities of HOMO and LUMO are different. Herein, we mainly focus on the differences about the moiety involved in intramolecular H-bond O(1)–H(2)···N(3). It can be found that the electron densities of N(3) atom increase and those of O(1) atom decrease after transition from HOMO to LUMO. In other words, the charges transfer from O(1) to N(3) atom based on HOMO → LUMO transition, which can facilitate the ESIPT processes.

Fig.3. View of the frontier molecular orbitals (HOMO and LUMO) for HPT and BrHPT

Table 4. Calculated Mulliken Charges (a.u.) of O(1) and N(3) Atoms for HPT and BrHPT

3. 3 Potential energy curves

Fig.4. Potential energy curves of the S0 and S1 states for HPT and BrHPT along with the proton transfer coordinate

In order to illustrate the ESIPT reaction mechanisms clearly, both the S0and S1states PECs of HPT and BrHPT are scanned along the proton transfer pathways with fixed O(1)–H(2) bond lengths in a series of values. As shown in Fig.4, for the S0state, the energy of HPT and BrHPT increases along with the elongations of O(1)–H(2) bond from its stable E forms and the potential barriers are 8.90 and 8.79 kcal·mol-1, respectively. It indicates that the stable E forms can barely convert into K forms for these two compounds because of their high energy barriers, that is to say, it is difficult for the GSIPT processes to occur. In comparison, the RGSIPT processes can easily occur, since the energy barriers are just 0.33 and 0.37 kcal·mol-1for HPT and BrHPT, respectively.

For the S1state, the energy barriers for the conversion of E* forms to K* forms are 2.68 and 1.24 kcal·mol-1for HPT and BrHPT, respectively, which are much lower than those in the S0state. It indicates that the proton transfer processes are more likely to occur in the S1state. In addition, the energy barriers for the K* forms to E* forms transformation are 7.75 and 7.47 kcal·mol-1, respectively, which are significantly higher than that of the proton transfer processes. Therefore, the forward ESIPT processes are much faster than the backward ones. Furthermore, it is easy to find that the energy barrier of BrHPT in the S1state is lower than that of HPT, which suggests that the electron withdrawing (–Br) group can facilitate the ESIPT processes. As is known, the ESIPT processes are usually interrupted to protic solvents because the intramolecular H-bond needed for the ESIPT process is interrupted by solvents. According to the emission spectra and potential energy curves, we believe that the presence of aprotic solvent acetonitrile will not perturb the ESIPT process. Herein, the proton transfer processes and mechanisms should be concluded as follows: Firstly, the N(3) atom can capture the proton of hydroxyl group to form an intramolecular H-bond in the S0state and the stable E forms cannot transfer the proton because of the increased potential energies; on the basis of the photoexcitation, the stable E forms are excited to the E* forms. Then, the ESIPT processes occur with proton spontaneously transferring from the O(1) atom to the adjacent N(3) atom and form the K* forms due to the relative low energy barriers. In turn, the K* forms decay back to the S0state via radiative transition, which induces the distinct redshifted emission as compared with those of the E* forms. Finally, the RGSIPT appears returning to the E forms by crossing almost negligible low barriers.

4 CONCLUSION

In summary, the ESIPT mechanisms of HPT and BrHPT have been investigated systematically based on DFT and TD-DFT methods using B3LYP functional at the basis set of 6-31G(d). Based on the analysis of the bond lengths and bond angles of these stable forms in both S0and S1states, we found that the intramolecular H-bonds O(1)–H(2)···N(3) are significantly strengthened in the S1state, which can facilitate the proton transfer process effectively. Upon the addition of F-, the proton can move close to F-and form typical covalent bond H(2)–F and the hydroxy moieties are deprotonated by F-. The vertical transition energies are calculated on the basis of optimized geometries of the S0and S1states, and good agreements are found between our theoretical computations and the experimental absorption spectra and fluorescence emission spectra. Electron withdrawing (–Br) substituent on the para-position of the phenyl ring gives redshifted absorption. The interactions of HPT and BrHPT with F-cause red-shift in absorption and new emission peak in fluorescence emission with the disappearance of dual fluorescence emission due to the inhibition of ESIPT. Further, via the FMOs and Mulliken’s charge distribution analysis, we find that the redistribution of charge would facilitate the ESIPT process. In addition, the constructed PECs of these two compounds in the S0and S1states further confirm that the proton transfer reactions can take place in the S1state easier than in the S0state. What's more, the presence of electron acceptor (–Br) group can enhance the strength of the intramolecular H-bond and facilitate the ESIPT processes for HPT systems.