Time-Dependent Long-Range-Corrected Density-Functional Tight-Binding Method Combined with the Polarizable Continuum Model

In this study, excited-state free energies and geometries were efficiently evaluated using a linear-response time-dependent long-range-corrected density-functional tight-binding method integrated with the polarizable continuum model (TD-LC-DFTB2/PCM). Although the LC-DFTB method required the evaluation of the exchange-type term, which was moderately computationally expensive, a single evaluation of the excited-state gradient for a system consisting of more than 1000 atoms in a vacuum was completed within 30 min using one CPU core. Benchmark calculations were conducted for 3-hydroxyflavone, which exhibits dual emission: the absorption and enol-form emission wavelengths calculated by TD-LC-DFTB2/PCM agreed well with those predicted based on the density functional theory using a long-range corrected functional; however, there was a large error in the predicted keto-form emission wavelength. Further benchmark calculations for more than 20 molecules indicated that the conventional TD-DFTB method underestimated the absorption and 0-0 transition energies compared with those which were measured experimentally, whereas the TD-LC-DFTB2 method systematically overestimated these metrics. Nevertheless, the agreement of the results of the TD-LC-DFTB2 method with those obtained by the CAM-B3LYP method demonstrates the potential of the TD-LC-DFTB2/PCM method. Moreover, changing the range separation parameter to 0.15 minimized this deviation.