Range-Separated Density Functionals in Predicting Correct Excitation Energies in Gas and Solvent Continuum: A Benchmark Investigation on a Large Set of Molecules

Abstract

Computationally cost-effective methods with high accuracy are indispensable in the field of quantum chemistry. Recently, descriptor-based tuning methods of range-separated (RS) functionals have attracted theoreticians because of their improved performance in computing various chemical properties. In this article, we have assessed the performance of our newly developed electron localization function (ELF) tuned [J. Comput. Chem. 2017, 38, 2258] and solvent (Sol) tuned [J. Comput. Chem. 2020, 41, 295] RS functionals in the calculation of lowest singlet vertical excitation energies of a large set of molecules in gas and solvent continuum. Moreover, EOM-CCSD benchmark values of excitation energies have been generated in gas and solvents. Notably, the benchmark values under the influence of the solvent continuum have been computed using perturbation theory and density approach (PTED) to take care of solvent effects in EOM-CCSD calculations. This study envisages that our ELF and Sol-tuned functionals can accurately reproduce EOM-CCSD benchmark values. Furthermore, our Sol-tuned functionals can predict the decrease of excitation energies with solvent polarity, which is consistent with EOM-CCSD results.