(730c) Optimization of a Transferable Shifted Force Field for Interfaces and Inhomogeneous Fluids | AIChE

(730c) Optimization of a Transferable Shifted Force Field for Interfaces and Inhomogeneous Fluids


Elliott, J. R. - Presenter, The University of Akron
Razavi, S. M., The University of Akron
Balçik, M., Istanbul Technical University
Ahunbay, M. G., Istanbul Technical University
Many commonly applied force fields have been developed based on the assumption of spherical symmetry and tail corrections to compute the thermodynamic properties. Unfortunately, many common applications involve inhomogeneous fluids where this assumption breaks down. For example, symmetry breaks down at the vapor-liquid interface because the liquid phase is roughly 500 times denser than the vapor phase, leading to very different forces for sites at various positions in the interface. In our experience, this necessitates a cutoff distance of roughly triple what is usually assumed in order to achieve self-consistent results for vapor pressure and surface tension. As another example, virtually all biological simulations involve some degree of inhomogeneity. Furthermore, a fundamental issue arises in the context of molecular dynamics because the discontinuity in the force field undermines the conservation of energy, leading to a range of possible compensating treatments, and poor consistency from one implementation to the next.

In this work, we review our progress in the development of a Transferable Shifted Force Field based on the United Atom perspective (TranSFF-UA). We consider a number of candidate potential models, taking the TraPPE-UA as a starting point. We universally apply a cut and shift of 3Ï?, where Ï? is the position where the potential initially crosses zero. Candidates to date include the LJ, 14-12-8-6, and Mie potential models. Point charges of polar molecules are adapted directly from the TraPPE-UA model, except in the case of water. For water, we consider the TIP4P(2005) model. This modification permits accentuation of the steepness of the potential, while considering the efficiency and parallelization of the Lennard-Jones (LJ) model, particularly in GPU systems. The simulation methodology involves LAMMPS in combination with GOMC, a relatively new open source Monte Carlo simulation platform. Particular attention is paid to consistency between the two methods and regions of phase space where each method performs best. All tabulated results are based on direct simulation, whereas thermodynamic perturbation theory is used as a meta-model to suggest prospective optimal potential parameters.

To demonstrate this methodology, vapor pressure, critical properties, saturated liquid density, and compressed fluid density are used to characterize potential models for n-alkanes, branched alkanes, ethers, olefins, naphthenics and aromatic compounds, and water. Vapor pressure deviations average near 9% for hydrocarbons. Saturated liquid density deviations average near 1% below a reduced temperature of 0.9. Compressed fluid densities exhibit deviations near 0.5% for the better potential models. Simply applying the shifted force to water significantly improves the vapor pressure correlation for the TIP4P model.

This work is supported in part by the Scientific and Technological Research Council of Turkey (TUBITAK) through the grant no. 114M178.