Thermal conductivity of dense fluids and mixtures via nonequilibrium molecular dynamics

The thermal conductivity of dense fluids and fluid mixtures has been examined in detail. A methodology based on the computer simulation method of nonequilibrium molecular dynamics has been developed to evaluate the various mixing rules and n-fluid characterizations used in the chemical industry for obtaining thermal conductivities of dense nonpolar fluid mixtures. The present study is based on pure fluids and mixtures modeled by the Lennard-Jones (LJ) intermolecular potential functions. For representing the molecular parameters of mixtures, the results show that Kay's simple mixing rules generally do as well as the more sophisticated Enskog mixing rules. In addition, it is shown that the 2-fluid characterization is significantly better than the 1-fluid characterization for predicting the thermal conductivities of continuous, semi-continuous, and discrete mixtures, and that going to higher order approximations (or characterizations) does not improve results significantly. This is in contrast to what has previously been found for thermodynamic properties.; The internal rotational contributions to the thermal conductivity of nonspherical fluids and mixtures has also been examined using a nonequilibrium molecular dynamics technique that has been developed recently. Internal rotational contributions are needed for predicting thermal conductivities of polyatomic fluids and mixtures. It is shown that the usual approximation that these rotational contributions are independent of density is invalid, and that they can be quite significant. An approximate method has been developed for including the density dependence of these internal contributions to thermal conductivity in the semi-empirical methods currently in use.; Finally, the results of a preliminary investigation of the thermal conductivity enhancement of monatomic Lennard-Jones fluids in the critical region are reported. The results show that this anomalous behavior can be studied using computer simulations. The enhancements have been explained using the radial distribution functions in terms of the long-range correlation effects.