(253at) Recent Advances in the Development of Cassandra: An Open Source Monte Carlo Framework for Phase Equilibria Calculations

The chemical sciences rely on a set of fundamental mathematical theories that serve as a basis to various computational modeling techniques. These methodologies have proven to be powerful tools to understand matter at the microscopic level. Among these techniques, quantum mechanical (QM) calculations provide the most fundamental description of matter by numerically solving the Schrödinger wave equation. In principle, any chemical system can be described by this equation. However, obtaining its solution is a formidable task even for relatively simple systems. Consequently, approximations must be made to address problems whose length and time scales are inaccessible by QM methods. One path for working toward this objective is classical molecular simulations. These techniques approximate atomic interactions using continuous interaction potentials, which can be developed empirically or through quantum theory. Within classical simulations, two main set of methodologies are used: molecular dynamics (MD) and Monte Carlo (MC). The principle of the first method lies on the numerical solution of Newton's equations of motion to propagate the system, whereas the second stochastically generates states of a system according to a predetermined statistical mechanical probability distribution.

Our group has developed Cassandra, an open source MC framework that allow the simulation of complex chemical systems within different statistical mechanical ensembles. Cassandra has been used to simulate fluids that involve a wide range of molecules such as water, alcohols, carbon dioxide, ionic liquids or mixtures of these. In this poster, recent advances in the development of this software package will be described. New capabilities include the implementation of advanced methodologies based on expanded ensembles to study dense systems, incorporation of polarization through fluctuating charges and Drude oscillators, implementation of non-Boltzmann sampling methodologies, the addition of fast pairwise alternatives to Ewald summation, the ability to simulate adsorption in porous materials and the implementation of neighbor lists. In addition, a series of scripts and utilities have been developed to help the user set up calculations faster and with fewer errors.