Accurate prediction of solution-phase chemistry requires proper treatment of solvation effects, yet most quantum calculations either neglect solvent or employ continuum models ignoring its molecular nature. We present a self-consistent reaction density functional theory (sc-RxDFT) that couples electronic structure calculations of solute with molecular-level solvent description through bidirectional iterative optimization. Unlike sequential approaches, both solute electronic structure and solvent density are minimized self-consistently, enabling mutual polarization adaptation while maintaining computational tractability. We validate sc-RxDFT on three benchmarks: (1) solvation free energies of 15 amino acid side chain analogs, where sc-RxDFT outperforms continuum models; (2) geometry optimization of aqueous water, demonstrating stable convergence and correct polarization-induced structural changes; and (3) the SN2 reaction between chloride and chloromethane in water, where sc-RxDFT accurately predicts the reaction barrier and reproduces the single-barrier mechanism observed experimentally, while continuum models and nonself-consistent approaches fail. This framework provides a practical alternative between expensive explicit solvent simulations and approximate continuum models.
Liu et al. (Sat,) studied this question.