Void-induced dissolution in molecular dynamics simulations of NaCl and water.

To gain a better understanding of the interaction of water and NaCl at the surface during dissolution, we have used molecular dynamics to simulate the interface with two equal-sized slabs of solid NaCl and liquid water in contact. The introduction of voids in the bulk of the salt, as well as steps or pits on the surface of the NaCl slab results in a qualitative change of system structure, as defined by radial distribution functions (RDFs). As an example, the characteristic Na-Na RDF for the system changes from regularly spaced narrow peaks (corresponding to an ordered crystalline structure), to a broad primary and smaller secondary peak (corresponding to a disordered structure). The change is observed at computationally short time scales of 100 ps, in contrast with a much longer time scale of 1 mus expected for complete mixing in the absence of defects. The void fraction (which combines both bulk and surface defects) required to trigger dissolution varies between 15%-20% at 300 K and 1 atm, and has distinct characteristics for the physical breakdown of the crystal lattice. The void fraction required decreases with temperature. Sensitivity studies show a strong dependence of the critical void fraction on the quantity and distribution of voids on the surface, with systems containing a balanced number of surface defects and a rough surface showing a maximum tendency to dissolve. There is a moderate dependence on temperature, with a 5% decrease in required void fraction with a 100 K increase in temperature, and a weak dependence on water potential model used, with the SPC, SPC/E, TIP4P, and RPOL models giving qualitatively identical results. The results were insensitive to the total quantity of water available for dissolution and the duration of the simulation.