A CELLULAR-AUTOMATA MODEL OF WATER

A model of liquid water has been created using cellular automata. In particular we use a kinematic, asynchronous, stochastic system to generate, iteratively, configurations simulating water emerging from initial conditions. We vary the cluster-breaking probability to generate variability in the network analysis. The breaking probability is found to have a close relationship to the temperature of liquid water when comparisons are made using the average number of hydrogen bonds, the average cluster size, and the number of free water molecules. A good correlation exists between the viscosity and the average cluster size as well as the number of free water molecules and the vapor pressure. The cellular automata model that we have generated gives a picture of large, extended clusters with very few small clusters or single water molecules. The graphic picture is very similar to the results described by Geiger, Stillinger, and Rahman and also Mezei and Beveridge. We identify this with a random network model of liquid water.