Numerical Discrimination of Thermodynamic Monte Carlo Simulations in All Eight Statistical Ensembles

Abstract

Generalized expressions for thermodynamic properties in terms of ensemble averages are discussed for adiabatic and isothermal ensembles. They are implemented in the simulation code ms2 and validated by Monte Carlo simulations for the Lennard-Jones fluid. A comparison of the eight statistical ensembles regarding size scaling behavior, convergence, and stability is provided for state points throughout the homogeneous fluid region. The resulting data are in good agreement but differ in their statistical distributions. In closed systems, the statistical quality of the data is better than in open systems. Overall, the microcanonical ensemble performs best.

Extension of the MolMod Database to Transferable Force Fields

“…Molecular simulations with force fields have become an important tool for predicting thermophysical properties – as well as for studying nanoscopic processes – in many scientific disciplines. The link of the simulations to the real world is established by the employed substance models, i.e., by the force fields.…”