Microscopic study and modeling of thermodiffusion in binary associating mixtures.

Thermodiffusion in associating mixtures is a complex phenomenon, owing to the strong dependence of the molecular structure of such mixtures on concentration. In this paper, we attempt to elucidate this phenomenon and propose a qualitative mechanism for the separation of species in binary associating mixtures. A correlation between the sign change in the thermal diffusion factor and a change in the molecular structure, mixture viscosity, and the excess entropy of mixing in such mixtures is established. To quantify this correlation, we modify our recently developed dynamic model based on the Drickamer nonequilibrium thermodynamic approach [M. Eslamian and M. Z. Saghir, Phys. Rev. E 80, 011201 (2009)] and propose expressions for the estimation of thermal diffusion factor in binary associating mixtures. The prediction power of the proposed expressions, as well as other widely used models, are examined against the experimental data. The proposed theoretical expressions are self-contained and only rely on the viscosity data as input and predict a sign change in the thermal diffusion factor in associating mixtures.