Development of a new water-water interaction potential and application to molecular processes in ice
A new potential energy function modeling the interaction among water molecules is presented. The dominant attractive interaction among the molecules, the electrostatic interaction, is represented with a multipolar expansion around the center of mass. The dipoles and quadrupoles are polarizable while the octopoles and hexadecapoles are fixed. The parameters in the multipole expansion are taken from experimental measurements on isolated water molecules or ab initio calculations. The molecules are treated as rigid objects. In addition to the electrostatics, the potential energy function includes dispersion energy and repulsive interaction at short range. The repulsive interaction includes many-body effects through density dependent parameters.To test the multipolar representation of the electric field, a comparison was made with ab initio calculations of the field near water clusters (Moller-Plesset MP2 calculations) and in the interior of an ice crystal (gradient corrected Density Functional Theory calculations). Very good agreement was found even for distances that are significantly shorter than typical intermolecular distances. Free parameters in the repulsive energy term were fitted to reproduce ab initio calculations of the energy of the water dimer as a function of distance, and the energy of ice Ih as a function of lattice constant.This new water potential energy function is, therefore, transferable to different types of environments. We have interpolated the potential energy surface from the gas phase (clusters) to the condensed phase (ice).We have applied the new potential function in a study of molecular processes on the surface of ice Ih, an environment which is in between gas phase and solid phase. The proton disorder in ice Ih leads to interesting and complex phenomena at the ice surface. The diffusion mechanism and the activation barriers for diffusion of water molecules in ice were evaluated by finding minimum energy paths for various diffusion hops. The results were used in a kinetic Monte Carlo simulation of surface diffusion in order to determine the overall diffusivity and an effective activation energy barrier. Satisfactory agreement with experimental measurements is found, but the experimental data is quite limited.
- Physics