(309b) Thermodynamically Constrained Inverse Monte Carlo: Potentials Optimized to Reproduce Structure and Thermodynamics

We propose a new inverse Monte Carlo methodology, based on optimization theory, which offers the crucial advantage that the model can be constrained to reproduce any desired thermodynamic property of the parent fluid, while simultaneously reproducing its structure. We argue that this methodology permits the systematic development of models for conducting quantitative multiscale simulations of polymeric and biologic systems. To benchmark this methodology for the known case of the Lennard Jones liquid. We subsequently consider the development of an effective pair interaction for argon from the experimental structure and equation-of-state data. Finally, we examine liquid methane and show that a united atom model can simultaneously reproduce the structure, the internal energy and the pressure obtained from corresponding all-atom simulations.