Modeling of chemical transformations in water clusters by parallel Monte-Carlo and ab initio quantum chemistry

A. A. Moskovsky ( moskov-AT-lcc-DOT-chem-DOT-msu-DOT-ru.gif ), S.S. Konyukhov, D.A. Firsov, A.V. Nemukhin
Laboratory of Chemical Cybernetics, Chemistry Department , M. V. Lomonosov Moscow State University, 119899 Moscow, Russian Federation.

The technique is being developed to model chemical reactions in water by using cluster approaches. The parallel Monte-Carlo algorithm has been developed and utilized to search the lowest-energy structures of doped water clusters. Ab-initio methods of quantum chemistry have been used to calculate the energy.

The parallel implementation of Monte-Carlo basin-hopping algorithm has been worked out for an efficient search of the points of global minima on ab-initio potential energy surfaces. The basin hopping technique combines a gradient optimization with the Metropolis walk in the configuration space. Usually, the solute species are described by the Hartree-Fock type approaches. In order to reduce computational cost, water molecules are described by effective fragment potential method. To perform quantum chemistry calculations we use the PC GAMESS software package. In practice, the systems with up to 20 water molecules from the solvation shells can be treated with this approach.

By using the Pentium III based computational clusters, the following reactions have been simulated: OH-+CO2--> HCO3- , HXeOH --> H2O +Xe and H2OO-> HOOH. The computed energy profiles have been analyzed, and the role of the solvent water molecules was studied. A special attention was paid to the changes in solvation shells along reaction paths.