Detecting Solution Phase Behavior in Particle Simulations

Abstract

The availability of accurate liquid phase diagrams are of importance to the formulation industry where it is necessary to formulate for specific physical behaviour to ensure product performance (e.g. to avoid stability issues such as phase separation etc.). It takes significant levels of effort to generate these diagrams experimentally which requires repeating each time a chemical ingredient is replaced. Alternative theoretical and computational methods can speed up turn around and reduce costs. One such technique is to run molecular simulations of the mixtures to sample the phase state and use these to interpolated across to obtain the full phase diagram. Additionally, molecular simulations can provide physical insights, such as topological and behavioural mechanisms, not accessible by other methods. In this work, we present a methodology for efficiently analyzing molecular simulations and elucidating the phase behaviour across a liquid ternary phase diagram using the examples of toluene + methanol + water and xylene + methanol + water. We present atomistic molecular dynamics and coarse-grain dissipative particle dynamics simulations and show the utility of a several metrics to profile the local mass concentrations to determine the solutions phase behaviour.