(346ba) Virial Coefficients As a Route from Molecular Models to Accurate Thermodynamic Models

Virial coefficients can provide a thermodynamic model that is exactly correct for a given molecular model. It is limited in state space (non-condensed, low density), but its accuracy can be self-assessed by examining convergence. Recent developments in the quality of molecular models, both semi-empirical and ab initio, provide a foundation to formulate thermodynamic models this way. Additionally, improvements in methodology allow for calculation of virial coefficients and their derivatives, particularly for non-pairwise molecular models (which are needed for highly accurate modeling).

We demonstrate with two types of applications: (a) Using a semi-empirical model, we generate an equation of state for a multicomponent mixtures, with exact description of composition dependence. This won’t yield results better than experiment, but it is preferable to other thermodynamic models at conditions where it is converged. We pursue a systematic, comprehensive effort to apply these methods to TraPPE potentials for a collection of 4 species - nitrogen, oxygen, carbon dioxide and ammonia and their mixtures. (b) Using highly accurate ab initio potentials, we generate an equation of state that exceeds the accuracy of what is possible experimentally. We examine the accuracy of ab intio virial coefficients for He-4 which are based on accurate first-principles 2- and 3-body molecular models from the literature and are generated using path-integral Monte Carlo (PIMC) calculations. We show that these coefficients can be used make predictions of pressure which are consistent with, and more precise than those obtained from experiment. We also compare the predictions of speed of sound against those from experiment.