Aggregation-volume-bias Monte Carlo simulations of vapor-liquid nucleationbarriers for Lennard-Jonesium

A combination of the aggregation-volume-bias Monte Carlo algorithm and the umbrella sampling technique is applied to investigate homogeneous vapor-liq uid nucleation. This combined approach is simple, general, and robust. Its efficiency is demonstrated for nucleation of Lennard-Jonesium, for which th e precise calculation of the nucleation barriers takes only a few minutes a t higher supersaturations to a few hours at lower supersaturations. Compari son of the simulation results to the classical nucleation theory (CNT) show s that CNT overestimates the barrier heights by a value nearly independent of the supersaturation, but provides a reasonable description of the critic al cluster sizes. (C) 2001 American Institute of Physics.