Parameters in the SAFT formalism are almost exclusively determined by fitting all parameters simultaneously to saturated vapor pressure and liquid density data. As a result, unavoidable lumping of physical effects occurs in the parameter estimation process. That is, parameters for dispersion do not only reflect dispersion, chain formation parameters do not only account for covalent bond energies, and parameters for association do not only measure hydrogen bonding. Moreover, the current methodology for determining parameters can result in multiple parameter sets for the same thermodynamic state of a fluid.

This presentation describes a novel and general global optimization–quantum chemistry framework for determining SAFT parameters in which each physical effect (dispersion, chain formation, association, others) is separated from all others and calculated by matching as closely as possible, and in a relative sense, the same physical effect predicted by quantum chemistry. To our knowledge, this is the first time a methodology for separating and subsequently determining SAFT parameters independently and without the use of experimental vapor pressure and liquid density data has been proposed. This new global optimization-quantum chemistry methodology

(1) Uses a novel all-atom approach to molecular segments (i.e., where the segments of a molecule are its individual atoms).

(2) Uses C6 dispersion coefficients from quantum mechanics are determined to quantify parameters for the dispersion term in SAFT.

(3) Determines chain formation and hydrogen bonding parameters (i.e., new covalent bond energy parameters, and unequal cross association strengths) by matching, as closely as possible, relative covalent bond energies and relative hydrogen bonding energies for binary mixtures computed from quantum chemistry.

Our global optimization-quantum chemistry approach is validated by comparing phase equilibrium predicted, not correlated, by SAFT using parameters fit to quantum chemistry results with experimental data. Several challenging examples involving associating mixtures are presented.