Linear-Response and State-Specific Polarizable Continuum Models Yield Different ESIPT Tautomer Geometries of Flavonoid Anions

Abstract The linear-response polarizable continuum model (LR-PCM) is routinely used to optimize excited-state geometries in ESIPT systems, followed by single-point energy corrections with the state-specific PCM (SS-PCM). This mixed protocol implicitly assumes that the LR-PCM minimum coincides with the minimum on the SS-PCM potential energy surface (PES)─an assumption that has never been systematically scrutinized in ESIPT systems. Here, we investigate this question using two flavonoid anions (API-B4 and LUT-B4-α/β) with zero, one, or two explicit water molecules hydrogen-bonded to the C4′–O– group. By constructing one-dimensional potential energy curves along the O5–H proton-transfer coordinate, we compare LR-PCM and SS-PCM descriptions. The ESIPT barrier is found to be nearly identical between the two models (differences < 0.5 kcal/mol) and largely insensitive to explicit waters. Each explicit water molecule raises the S1 excitation energies by about 0.1 eV by stabilizing the HOMO more than the LUMO. More importantly, the keto minimum obtained with LR-PCM is not always a minimum on the SS-PCM PES. This discrepancy is correlated with the substantial increase in the excited-state dipole moment induced by SS-PCM and is consistently reproduced across three hybrid functionals (ωB97X-D, CAM-B3LYP, M062X). Overall, this work provides new insight into the role of the SS-PCM-enhanced excited-state dipole moment in emission energy calculations.