(321s) Effect of Torsional Potential on the Phase Behavior of Organic Compounds

Authors: 
Bernard-Brunel, D. - Presenter, Department of Chemical Engineering
Potoff, J. J. - Presenter, Wayne State University


In molecular mechanics force fields, the total energy of the system is determined through the summation of all inter- and intra molecular interactions in the system. Intermolecular interactions typically consist of dispersive and Coulombic terms, while intramolecular interactions contain terms for bond stretching, bond bending and rotation around various dihedral angles. These intramolecular potentials are used to control the conformational behavior of the molecule of interest. Although significant research has been performed in the generation of potentials for intramolecular interactions, it is unknown how sensitive thermophysical properties, including phase behavior, are to the parameters used to control the conformation behavior of the molecules. In this work, configurational-bias Monte Carlo simulations in the grand canonical ensemble are combined with histogram-reweighting techniques to assess the effect of torsional potential parameters on the predicted vapor-liquid equilibria for a variety of organic compounds. Molecules investigated in this work include linear and branched alkanes, alcohols, ketones and ethers. The Transferable Potentials for Phase Equilibria (TraPPE) force field was used as a reference calculation. Using the same non-bonded parameters as defined by the TraPPE force field, new parameters were derived for torsional potentials such that the energy barrier between the various dihedral conformations was lower than the temperature at which the simulations were performed, allowing for a nearly flat distribution of sampled dihedral angles. Additional calculations are performed where dihedral angles are sample from a Gaussian distribution. Simulations where dihedral angles were sampled from a flat distribution were found to yield results identical to that of the original TraPPE force field. Simulations that sampled from a single Gaussian distribution of dihedral angles resulted in an underprediction of the saturated liquid densities.