A potential model for the study of ices and amorphous water: TIP4P/Ice

TLDR

Several water models are evaluated for their ability to predict ice properties, noting that improved melting temperature predictions do not compromise other properties. The authors aim to develop a new water model that accurately reproduces ice densities and coexistence curves. Parameters were fitted to the equation of state and key melting and coexistence points, and the resulting phase diagram was computed. The new TIP4P/Ice model predicts a melting temperature of 272.2 K for hexagonal ice and yields densities and coexistence curves that outperform the previously best TIP4P model, indicating that common rigid models are inadequate.

Abstract

The ability of several water models to predict the properties of ices is discussed. The emphasis is put on the results for the densities and the coexistence curves between the different ice forms. It is concluded that none of the most commonly used rigid models is satisfactory. A new model specifically designed to cope with solid-phase properties is proposed. The parameters have been obtained by fitting the equation of state and selected points of the melting lines and of the coexistence lines involving different ice forms. The phase diagram is then calculated for the new potential. The predicted melting temperature of hexagonal ice (Ih) at 1bar is 272.2K. This excellent value does not imply a deterioration of the rest of the properties. In fact, the predictions for both the densities and the coexistence curves are better than for TIP4P, which previously yielded the best estimations of the ice properties.