Simulating solvent effects in organic chemistry: combining quantum and molecular mechanics

As computational methods improve, the need for accuracy must still be tempered with practicality. When calculating how molecules interact in solution, treating solute molecules quantum mechanically and the surrounding solvent molecules classically combines accuracy with computational efficiency. Computational methods for modeling solute-solvent interactions generally fall into two categories. A macroscopic approach treats the solvent as a continuous (unstructured) medium characterized by a bulk dielectric constant. Macroscopic methods, while generally faster because solvent molecules are not explicitly represented, are unable to provide specific details of the solute-solvent interaction. The microscopic method, which is the subject of this article, treats the solvent in its discrete molecular form. Most molecular dynamics or Monte Carlo simulations use molecular-mechanical (MM) potentials for both solute and solvent. However, hybrid quantum-mechanical and molecular-mechanical potential functions have recently emerged. We describe this combined approach for simulating chemical reactions in solution