Excited state intramolecular proton transfer fluorescence emission of o-hydroxyphenyl-triazine derivatives

Abstracto-hydroxyphenyl-triazine derivatives with formyl substituents were surveyed for the excited state intramolecular proton transfer (ESIPT). The occurrence of ESIPT was confirmed by well-separated emission bands for the derivatives. A low energy change from enol to keto in the excited state explains ESIPT for the derivatives.

Excited state intramolecular proton transfer (ESIPT) is a very important reaction in chemical and biological systems. It has received considerable interest in the areas of photochemistry, photophysics and photobiology since the investigation of the fluorescence of methyl salicylate [1,2]. Generally, ESIPT is achieved by an ultrafast photoinduced enol-keto tautomerization process during a four-level reaction (E→E * →K * →K, E: enol, K: keto) via five- [3,4], six- [5], or rarely seven-membered quasi rings [6]. In this process, absorption occurs from E→E * and emission from K * →K, which produces well-separated dual emission bands and an abnormally large stokes shift without self-absorption in the second emission band. o-hydroxyphenyl-triazine derivatives are well-known ultraviolet absorbers, that ESIPT is generally thought to occur in despite the lack of solid experimental evidence. Recent theoretical and experimental investigations of o-hydroxyphenyl-triazine derivatives have drawn opposite conclusions on this for the occurrence or absence of ESIPT, respectively [7][8][9][10]. Consequently, new o-hydroxyphenyl-triazine derivatives need to be developed and their spectral data analyzed to determine if ESIPT can occur in these compounds. In general, the reported o-hydro-*Corresponding authors (email: hongruli1972@gmail.com; fanggao1971@gmail.com) xyphenyl-triazine derivatives have alkyl or aryl substituents on the hydroxyphenyl ring [7][8][9][10]. The electron-donating nature of alkyl or aryl groups could reduce the acidity of the hydroxy group and lower the possibility of ESIPT occurrence. In contrast, the acidity of the hydroxy group could be enhanced by electron withdrawing group, which would facilitate ESIPT. We recently investigated the occurrence of ESIPT in novel o-hydroxyphenyl-triazine derivatives (  The derivatives C1 and C2 were synthesized by an established route (see Appendix). The final products were purified by column chromatography and the structures were characterized by 1 H-NMR and elemental analysis (see Appendix). Figure 2 depicts the actual and normalized absorption spectra of C1 and C2 in cyclohexane, ethyl acetate, and acetonitrile from 260-400 nm. In these spectra, C1 displayed double absorption bands, while C2 exhibited a single absorption band. The maximum absorption wavelength of the first absorption band of C1 was similar to that of the absorption spectroscopy of C2 ( Figure 2, Table 1). The second absorption band of C2 was red-shifted by approximately 60 nm with respect to the first absorption band (Table 1).
The data in Table 1 suggest clearly that the maximal absorption wavelength of first absorption band C1 is almost the same as the maximal absorption wavelength of C2, and the second absorption wavelength of C1 exhibits an approximately 60 nm red-shift with respect to them in various solvents. Both C1 absorption bands were observed in various solvents, including the aprotic and protic solvents cyclohexane, ethyl acetate, tetrahydrofuran, and methanol (Table 1). In contrast, C2 had only one absorption band in all the solvents. These results indicate that the longwavelength absorption band of C1 is not greatly affected by the polarity or protic nature of solvent. Furthermore, the double absorption bands of C1 exhibited almost no change on the addition of triethylamine or acetic acid to the solvents (see Appendix) while maintaining the concentration of C1. Even if in pure acetic acid or triethylamine, C1 still exhibits obvious long-wavelength absorption (see Appendix). Therefore the long-wavelength absorption of C1 probably does not arise from internal proton transfer, but it could be ascribed to the intramolecular hydrogen bond at the ground state. Figure 3 shows the actual and normalized fluorescence spectra from 320-600 nm of C1 and C2 in cyclohexane, ethyl acetate and acetonitrile. Interestingly, C1 displayed well-separated double emission bands in ethyl acetate and acetonitrile, but only one emission band in cyclohexane. In

C1
C2 Solvents  Consequently, it is reasonable that the second emission of C1 disappeared completely in methanol (see Appendix).
The long-wavelength emission of C1 could also be reduced by adding acetic acid and triethylamine while maintaining the concentration of C1 (see Appendix). Further addition of acetic acid or triethylamine while maintaining the concentration of C1 resulted in recovery of the second emission band of C1 (see Appendix). The long-wavelength emission of C1 could not be detected in pure acetic acid or triethylamine (see Appendix). The second emission band of C1 exhibited an abnormally large Stokes shift of approximately 220 nm (Table 1). These results demonstrate that while the first emission band of C1 is from normal emission, the second emission band can be attributed to ESIPT. This means that the four-level reaction (E→E * →K * →K) via a six-membered quasi ring could take place in ethyl acetate and acetonitrile when C1 is excited (Figure 4). We also observed ESIPT emission of C1 in other strong aprotic polar solvents, such as DMF and DMSO (see Appendix). ESIPT emission of C1 in aprotic polar solvents could be suppressed by the addition of non-polar solvents, such as cyclohexane and benzene. On incremental addition of cyclohexane, the long-wavelength emission of C1 in ethyl acetate is gradually reduced and eventually completely disappeared ( Figure 5). This indicates that the long-wavelength emission band of C1 is closely related to the polarity, acidity/basicity, and protic nature of the solvents, which is quite different to the long-wavelength absorption band. This observation is reasonable because the long-wavelength emission band of C1 is produced from an internal proton transfer process. The emission intensity of C1 was lower than that of C2 in all solvents except for cyclohexane under the same excitation conditions (Figure 3). As a result, the fluorescence emission quantum yields of C1 are much lower than those of C2 in all solvents except for cyclohexane (Table 1). It is accepted generally that fluorescence quantum yields of organic compounds are diminished by ESIPT. This means that the enol-keto tautomerization in the excited state weakens the emission of C1.
To understand the occurrence of ESIPT in C1, the energy changes from E * →K * for internal proton transfer were calculated for the S 0 and S 1 states (see Appendix), and plot-   ted as a function of O-H distance ( Figure 6). These energy changes will determine if ESIPT occurs, because it is a real chemical reaction. The energy plot ( Figure 6) clearly shows that E * is more stable than K * , and K is much more unstable than E. A small energy barrier (16 kJ/mol) existed for the internal proton transfer from E * →K * , which means molecular thermal motion is sufficient to overcome this barrier at room temperature. In contrast, the energy keeps increasing with O-H distance from E→K. This indicates that it is difficult for ground state intramolecular proton transfer (E→K) to occur, while the reverse proton transfer process (K→E) can take place easily. This is favorable for the occurrence of ESIPT. The calculations also confirmed that a large dipole moment (4.19 D, 1 D=6.695×10 -30 C m) change occurs for E * →K * . This in turn explains why the solvent polarity is favorable for ESIPT occurrence, as observed in our experiments. This is mainly due to the occurrence of a huge structural reorganization for C1 during ESIPT.
In summary, these results illustrate the satisfactory wellseparated ESIPT emission bands for a novel o-hydroxyphenyl-triazine derivative. This provides strong experimental and theoretical evidence that o-hydroxyphenyl-triazine derivatives are able to undergo ESIPT. In consideration of its structural simplicity and modifiability, C1 could be an ideal model for the development of novel o-hydroxyphenyl-triazine derivatives that can undergo ESIPT. These derivatives could also be used to probe the ESIPT dynamics of o-hydroxyphenyl-triazine systems.