Sampling configurations in Monte Carlo simulations for quantum mechanical studies of solvent effects

Discrete models obtained from computer simulation are in increase use to study solvent effects. The approach consists of generating supermolecular structures for quantum mechanical calculations. The properties of the solute are calculated as an ensemble average over configurations generated by the simulation. An analysis of the efficiency of the simulation shows that the number of configurations necessary for the ensemble average can be reduced drastically. As an application to solvatochromism, the calculated spectral shift of the ¹B_2u(π − π*) transition of benzene in water is shown to be the same whether it is calculated with many but correlated configurations or with just a few uncorrelated configurations. © 1998 John Wiley & Sons, Inc. Int J Quant Chem 66 : 249–253, 1998