In recent decades, Statistical Associating Fluid Theory (SAFT) has advanced both in the accuracy of the underlying perturbation theory and in the representation of increasingly realistic molecular interactions. In this study, we extend the SAFT-VR Mie equation of state to a general sum of Sutherland potentials with temperature- and density-dependent prefactors, termed SAFT-VR Sum. A generalized Sutherland sum offers extensive flexibility in representing molecular interactions. We demonstrate this ability by regressing a sum of Sutherland potentials to reproduce the binary interactions of argon and neon from ab initio simulations. With three-particle effects approximated via the Axilrod–Teller–Muto potential and quantum nuclear effects incorporated via Feynman–Hibbs corrections, the resulting molecular model accurately captures thermophysical properties such as the viscosity and phase-equilibrium densities within 0.2% and the second virial coefficient within 1% above 15 K. The SAFT-VR Sum equation of state, with monomer interactions built solely from molecular information, predicts most thermodynamic properties of argon and neon within 2%, although properties such as the isochoric and isobaric heat capacities and the second virial coefficient are less accurately described by the perturbation theory. To probe the limits of the SAFT framework, we examine a potential with two force minima. For this potential, the α-function hypothesis commonly used in the development of perturbation theories fails to capture the second order perturbation term. Overall, a sum of Sutherland potentials provides sufficient flexibility for effective molecular descriptions to reproduce thermophysical properties of simple molecules within the experimental uncertainty. Advancing the thermodynamic perturbation theory thus offers a promising path to further drive SAFT toward experimental accuracy.
Jervell et al. (Mon,) studied this question.