Computational evaluation of geometries, functionals, and solvation models for UV photolysis of micropollutants in water: Parameters comparison to calculate excited states by time-dependent density functional theory

Quantum chemical calculation is a useful tool to quasi-theoretically evaluate various routes of chemical reactions but considerably influenced by calculation manners and input parameters. This study compared geometries (local minimum and global minimum), calculation levels (CAM-B3LYP/6–31++G(d,p) and ω B97X-D/def2-TZVPPD), and solvation models (IEFPCM and SMD) to evaluate UV photolysis of micropollutants by time-dependent density functional theory (TD-DFT). Target compounds were photo-reactive micropollutants including sulfamethoxazole, diclofenac, ciprofloxacin, carbamazepine, 6PPD, and 6PPD-quinone. Input coordinates were optimised by DFT followed by excited states calculation by TD-DFT, in which the identical calculation level was applied. The global minimum exhibited lower Gibbs free energy at the optimised geometry than the local minimum, but excited states calculation was overall comparable when the same level of calculation was applied. Electronic configuration of each excited state indicated that UV-induced photoreaction was described as aggregation of various orbitals. ω B97X-D/def2-TZVPPD and CAM-B3LYP/6–31++G(d,p) exhibited different configurations, but excited states characteristics were qualitatively similar between two calculation levels. By applying SMD solvation, molecular orbital of sulfamethoxazole and carbamazepine illustrated clear comparability with experimental studies concerning photo-excited chemical structures. Consequently, CAM-B3LYP/6–31++G(d,p) was applicable to swiftly obtain oscillator strengths and absorbance spectra, and ω B97X-D/def2-TZVPPD with SMD solvation model would be recommended for detailed analysis including molecular orbitals. As future implications, the calculation strategies indicated in this study could be applicable for other micropollutants to computationally evaluate photo-induced degradation, and outcomes of this study would be beneficial for wider application such as creating a predictive model of UV photolysis.