Solid–liquid coexistence in ionic systems and the properties of the interface

Abstract

The crystal–melt interfaces of pure KCl and LiCl crystals with (Li,K)Cl melts of various compositions have been investigated by direct simulation of coexisting crystalline and molten regions. This system forms no solid compounds involving both lithium and potassium and exhibits a eutectic close to the equimolar composition. Simulations at various melt compositions allow the liquidus to be determined and compared with experiment. In order to test the precision achievable by a direct simulation of coexisting liquid and solid phases for the thermodynamic properties, we make preliminary calculations on the melting of a model of NaCl, which has recently been studied by Anwar et al. [2003, J. chem. Phys., 118, 728] using thermodynamic integration. Including anionic polarization, effects in the interionic interaction model lower the melting temperature, despite the fact that there is no discernible effect on the liquid or solid structure in these simple monovalent melts. The method of direct simulation of coexistence also allows for the study of the interfacial structure and dynamics. These can be compared with previous studies [Davidchack, R. L., and Laird, B. B., 1998, J. chem. Phys., 108, 9452; Sibug-Aga, R., and Laird, B. B., 2002, J. chem. Phys., 116, 3410] of the solid–liquid interfaces in binary mixtures of hard spheres of sufficiently different size ratio to prevent solid compound formation. Despite the additional Coulombic interactions the ionic interfaces resemble the hard spheres in many regards, but there are differences. For example, when traversing the interface the perpendicular diffusion rises more rapidly than the parallel diffusion, most likely due to charge ordering.

Notes

The experimental molar volume for the crystal has been extrapolated from the results in [26] from 873 to 1071.3 K.

It should be noted that in layers E and F it appears that there are more K⁺ ions (red) than Cl⁻ ions (black). This is an artificial effect of printing the figure, as the black points are plotted first. In fact the opposite is true, as some of the K⁺ ions have left the layer and been replaced with Li⁺ ions, which are plotted separately in figure 14.