(749a) Free Energy Perturbation Approach to Molecular Model Development
As computer simulations are increasingly used to complement or replace experiments, highly accurate descriptions of microscopic interactions are required to achieve realistic results. One way to achieve this goal, while maintaining computational efficiency, is to use effective classical force fields with highly optimized interaction potential parameters. Here we extend our recently introduced force field optimization technique,1 based on the free energy perturbation principles,2 and develop a full-fledged methodology for general molecular and coarse-grained model development.3 Using experimental and quantum chemical data for thermodynamic, structural, and dynamic properties as a reference, we show how the approach is capable of finding globally optimal model parameters with minimal computational expenses. The speed of the technique, which is given by the exhaustive use of statistical data from molecular simulations, makes it possible to explore many-dimensional parameter spaces and identify optimal models within minutes upon collecting data from a single simulation. Optimization examples to illustrate the proposed strategy include developments of molecular models of dilute and concentrated solutions, as well as bulk solids. The functional forms for the description of the interaction potentials include simple Lennard-Jones and point charge models as well as more complex models comprising exponential-6 pair potentials, Gaussian charges, and polarizability, which have been used to develop molecular models for the accurate description of aqueous electrolytes and fluid phase equilibria.
1) Vlcek L., Chialvo A.A., Cole D.R “Optimized Unlike-Pair Interactions for Water-Carbon Dioxide Mixtures Described by the SPC/E and EPM2 models.” J. Phys. Chem. B 2011, 115, 8775-8784.
2) Chialvo A.A., Excess Properties of Liquid-Mixtures from Computer-Simulation: a Coupling-Parameter Approach to the Determination of Their Dependence on Molecular Asymmetry. Mol. Phys. 1991, 73, 127-140.
3) Vlcek L., Moucka F., Nezbeda, I., Chialvo A.A. “Free energy perturbation approach to atomistic and coarse-grained model optimization via fast scanning of many-dimensional parameter spaces”, J. Chem. Theor. Comp. (in preparation)