A quasiclassical trajectory study of reactivity and product energy disposal in H+H2O, H+D2O, and H+HOD

The dynamics of the H + H2O -->H-2 + OH, and some isotopic counterpart reactions has been investigated by quasiclassical trajectory (QCT) calculations, and using a recently developed potential energy surface [Wu , J. Chem. Phys. 113, 3150 (2000)] that was derived from high quality ab initio calculations. We make an extensive comparison with QCT and quantum scattering results based on other surfaces, particularly that from Ochoa and Clary, as well as with experimental results. Our results show that, in agreement with earlier theoretical results, the cross sections for the reaction of translationally hot hydrogen atoms with ground state H2O (yielding H-2 + OH) and with ground state D2O (yielding HD + OD) are significantly smaller than experiment. Our results are in agreement with accurate quantum results on comparably accurate surfaces, thereby showing that the disagreement with experiment is not a problem with either the dynamics method or the potential surfaces. In contrast to this, other properties of the reaction dynamics are in generally excellent agreement with experiment. For example, the role of stretch excitation on the H + D2O cross sections follows the trends observed in the experiments. Bend excitation is found to be more active than was previously thought in enhancing reactivity, but is still within experimental uncertainty. Water rotation is found to play an important role in experiments that sample j(H2O) values of 5 or greater. Our studies of the H + D2O and H + HOD -->H-2 + OD,HD + OH reactions yield isotope branching ratios and product distributions (for both spectator and newly formed diatoms) that are generally in good agreement with experiment. The only exception to this arises with the HD rotational distributions in H + D2O, where the observed distributions show less excitation and broader distributions. The internal distributions of experimentally unresolved products are also discussed. We conclude that the new potential energy surface used here is very accurate for describing the H + H2O -->H-2 + OH reaction and isotopic counterparts, providing significant improvement over previously published results. (C) 2001 American Institute of Physics.