State-Specific Treatment of Solvent Effect on Excited States in Organic Photoredox Catalysis

Abstract

Excited state behavior in organic photocatalysis is strongly influenced by solvent polarity. We benchmark solvation treatments for predicting charge-transfer (CT) and locally excited (LE) states by combining time-dependent density functional theory (TDDFT) with implicit and explicit solvation. Within implicit models, linear-response (LR) and four state-specific (SS) schemes (cLR, ${\rm cLR}^2$ , VEM, IBSF) are assessed, and explicit solvation is modeled using QM/MM with electrostatic embedding. Across representative systems, we examine solvent-dependent excited state energetics to compare the solvation models. For CT states, SS approaches consistently capture solvent-induced stabilization, whereas LR underestimates solvent relaxation and misrepresents solvent trends. Among SS methods, cLR and ${\rm cLR}^2$ show nearly identical solvent dependence, and the additional transition density term in ${\rm cLR}^2$ vanishes for triplets and becomes relevant only when the singlet state contains appreciable LE character. VEM produces similar solvent dependence, while IBSF tends to over-stabilize CT states in polar solvents. In explicit solvation, QM/MM reproduces the SS trends with small, systematic shifts in absolute energies. By contrast, LE states exhibit weak solvent sensitivity and are consistently described by all methods. Separately, results across density functionals highlight the importance of long-range corrected hybrids for describing CT states. These results provide practical guidelines for selecting solvation models and exchange correlation functionals to reliably capture solvent effects in organic photocatalysis, particularly in systems with pronounced CT states.