Optimization of particle transfers in the Gibbs ensemble for systems with strong and directional interactions using CBMC, CFCMC, and CB/CFCMC

Abstract

Computing accurate vapor–liquid equilibrium (VLE) diagrams using Monte Carlo simulations is not always straightforward because of the difficulty in inserting and deleting molecules. The acceptance probability of insertions/deletions is sharply reduced at low temperatures and in systems with strong and/or directional interactions, like water and ionic liquids. In this work, we highlight the efficiency of configurational bias continuous fractional component Monte Carlo (CB/CFCMC) for the computation of VLE diagrams of strongly interacting systems. We show that CB/CFCMC improves the particle exchange probability in the Gibbs ensemble about two times at low temperature and about one-third at high temperature compared to CFCMC and about 2 orders of magnitude compared to CBMC. The (CB-)CFCMC methods therefore allow for routine and efficient computation of VLE diagrams in systems with strongly interacting molecules. We demonstrate the successful application of the method in the simulation of VLE diagrams of water and DMF using the Gibbs ensemble and study the adsorption isotherm of water in a DMOF variant using grand-canonical Monte Carlo. Although CFCMC methods drastically improve single-particle insertion, additional trial moves and biasing are needed to study the collective behavior such as growth and decay of water clusters. As a downside of the current CFCMC methods in the Gibbs ensemble, we note that the use of two fractional molecules and the resulting coupling between the gas and the liquid phase leads to less than optimal biasing and efficiency.