Extension of the anisotropic united atoms intermolecular potential to amines, amides and alkanols: Application to the problems of the 2004 Fluid Simulation Challenge

Abstract

A combined study by quantum chemistry and Monte Carlo simulation is conducted to extend the anisotropic united atoms (AUA) intermolecular potential to alkylamines, alkylamides and alkanols. Ab initio simulations are used to derive electrostatic point charges of the investigated molecules. The amplitude and location of these charges are optimized to best reproduce the electrostatic field around the molecule. The Lennard–Jones parameters of the amine NH 2 and alcohol OH groups are regressed to match the vapor–liquid equilibrium properties of methylamine, propylamine, methanol and ethanol. The Lennard–Jones parameters of the CH 2 and CH 3 groups are taken from previously published work [P. Ungerer, C. Beauvais, J. Delhommelle, A. Boutin, B. Rousseau, A.H. Fuchs, Optimization of the anisotropic united atoms intermolecular potential for n-alkanes, J. Chem. Phys. 112 (2000) 5499–5510]. The three problems of the Fluid Simulation Challenge are investigated with a unique AUA parametrization for the organic solvents as well as intermolecular potentials from the literature to treat gases or water. In the first problem, AUA potential parameters for the carbonyl C O group [S. Kranias, D. Pattou, B. Levy, A. Boutin, An optimized potential for phase equilibria computation for ketone and aldehyde molecular fluids, Phys. Chem. Chem. Phys. 5 (2003) 4175] are used to simulate the vapor–liquid equilibrium of acetone and butyramide in the Gibbs ensemble. A good agreement is found with experimental vapor pressures and vaporization enthalpies. In the second problem, the Henry constants of four gases in ethanol are determined by biased Widom test insertions. This property is predicted within 15% on average, and its variations with temperature agree with experiments. In the third problem, the heats of mixing are computed through simulations in the NPT ensemble. Good agreement with experimental data is found for the heptane–butylamine system but not for the water–butylamine system. Possible ways to improve prediction are briefly discussed.