Efficient ab initio sampling methods in rate constant calculations for proton-transfer reactions

In this article, the classical potential based importance Monte Carlo sampling method of Iftimie [J. Chem. Phys. 113, 4852, (2000)] is applied to an ab initio simulation of the proton transfer tautomerization reaction of malonaldehyde in an aprotic, nonpolar solvent. It is demonstrated that ad hoc bond-energy bond-order relations derived from bond evolution theory combined with Pauling's valence bond ideas can be used to construct a molecular mechanics guidance potential for the simulation of the proton transfer reaction which improves the statistics of the calculation by three orders of magnitude. The sampling method is extended to simulations in which quantum effects are treated using the imaginary time path-integral representation. A new algorithm based on multiple Markov chain theory is introduced by which it is possible to obtain very short integrated correlation lengths in calculations of quantum static correlation functions. (C) 2001 American Institute of Physics.