On the use of a modified square-well model for prediction and correlation of thermal diffusion factors in binary liquid mixtures

Abstract

Recently reported thermal diffusion factors obtained from heats of transport have been used to test the ability of existing theories to predict thermal diffusivities using model parameters regressed from pure component transport data. None of the theories tested were able to provide even qualitative agreement with experiment. However, predictions using a square-well Enskog theory, modified here for consistency with normal experimental definitions, were found to be approximately correct using pure-component square-well parameters regressed from viscosity and thermal conductivity and an assumed mixing rule for cross terms. Predicted values are very sensitive to the energy mixing rule, and correlation of experimental data is possible for some systems using this theory with one adjustable parameter in the mixing rule.