(563b) The Fourier Correlation Method for Predicting Diffusivities in Multicomponent Mixtures Simulated by Molecular Dynamics
AIChE Annual Meeting
2009
2009 Annual Meeting
Computational Molecular Science and Engineering Forum
Industrial Applications of Computational Chemistry and Molecular Simulation II
Thursday, November 12, 2009 - 12:48pm to 1:06pm
As has been recognized for several decades, molecular dynamics simulations can be an elegant way to calculate and predict transport properties, including diffusion coefficients. The active development in the field over the years has led to a number of simulation techniques for diffusivity, most notably the Green-Kubo relation for equilibrium molecular dynamics as well as nonequilibrium molecular dynamics (NEMD) techniques. In this work we report on a new method for predicting bulk diffusivities in multicomponent mixtures. The technique, which we call Fourier correlation (FC), can be implemented through a simple algorithm within existing equilibrium MD computer simulations. The FC method conceptually combines macroscopic equations of change, finite Fourier transforms, and microscopic fluctuations. We briefly outline the rationale and advantages of the method and show preliminary results for obtaining diffusivities in nonideal binary and ternary liquid mixtures.