254871 Fick Diffusion Coefficients In Ternary Liquids From Equilibrium Molecular Dynamics

Tuesday, October 30, 2012: 10:00 AM
415 (Convention Center )
Thijs JH Vlugt, Process & Energy, Delft University of Technology, Delft, Netherlands, Xin Liu, RWTH Aachen University, Aachen, Germany, Erin McGarrity, Delft University of Technology, Delft, Netherlands, Ana Martin Calvo, University Pablo de Olavide, Seville, Spain, Sofia Calero, Department of Physical, Chemical, and Natural Systems, University Pablo de Olavide, Seville, Spain and Andre Bardow, Institute of Technical Thermodynamics, RWTH Aachen University, Aachen, Germany

An approach for computing ternary Fick diffusivities directly from equilibrium Molecular Dynamics (MD) simulations is presented and demonstrated for a liquid mixture of chloroform-acetone-methanol. In our approach, Fick diffusivities are obtained from Maxwell-Stefan (MS) diffusivities and the so-called matrix of thermodynamic factors. MS diffusivities describe the friction between different molecular species and can be directly computed from equilibrium MD simulations. The thermodynamic factor is the concentration derivative of the activity coefficient describing the deviation from ideal mixing behaviour. It is important to note that all mutual diffusion experiments measure Fick diffusion coefficients, while previous equilibrium molecular dynamics simulation did only provide MS diffusivities. The required thermodynamic factor to convert MS into Fick diffusivities and vice versa, however, has been difficult to extract from both simulations and experiments leaving a gap between theory and applications. Here, we extend an efficient computing scheme for the thermodynamic factor recently proposed by the authors and co-workers [1] for ternary complex mixtures. The novel computing scheme extracts the thermodynamic factor from density fluctuations in small subsystems embedded in a larger simulation box. The computed thermodynamic factor is in excellent agreement with experimental VLE data as well as calculations using the COSMO-SAC method. Our approach provides an efficient route to obtain multicomponent diffusion coefficients in liquids based on a consistent molecular picture and therefore bridges the gap between theory and experiment.

[1] X. Liu, S.K. Schnell, J-M Simon, D. Bedeaux, S. Kjelstrup, A. Bardow, T.J.H. Vlugt, J. Phys. Chem. B, 2011, 115, 12921-12929.


Extended Abstract: File Not Uploaded