DENSITY-SCALING - A NEW MONTE-CARLO TECHNIQUE IN STATISTICAL-MECHANICS

We demonstrate the feasibility of using "umbrella sampling" to do Monte Carlo Markov-sampling runs each covering a substantial range of density: "density-scaling Monte Carlo," or DSMC. One can obtain in this way not only the usual canonical averages but also the relative free energy as a function of density. To test this it has been applied to systems for which there are some previous reliable results: the hard-sphere system and the restricted primitive model of 1:1 and 2:2 electrolytes. The method proves to be startlingly powerful in that very extensive results can be obtained with very few DSMC runs. An important further motivation is the prospect of using the technique to study phase transition regions. © 1991.