Numerical scheme looks to pave the way for universal method to predict solubilities

Creating solutions is central to studying chemistry. Predicting solubility of crystalline solids by computer simulations is particularly challenging. Direct simulations that wait for solid-liquid equilibrium to be established can be prohibitively time-consuming. Alternative approaches, using free-energy calculations, are faster but require specific expert knowledge. New work has expanded on an emerging technique that looks to offer an easier, more robust approach that might be used universally to predict solubilities.

Li et al. demonstrate more capabilities of a numerical scheme they developed to predict the solubility of crystalline solids. Computer simulation can determine the solubilities of simple systems such as noble gases or alkaline halides. Here the authors take one step further, and investigate for the first time the solubilities of more complex materials, such as phenanthrene, caffeine, calcite and aragonite, over ranges of temperatures and pressures.

The group determined solubility by equating the chemical potential of the solute in solution and in the solid phase. In computing the chemical potentials of solid molecules, they use a molecular adaptation of the well-tested Einstein crystal method, which was proposed by one of the authors 35 years ago.

The calculation of the chemical potential of the solute in solution is more challenging and the authors refine a recently proposed methodology where a cavity is formed in the solution first, and the solute is inserted into the cavity in a second step. An important advantage of the proposed methodology is that it can be easily implemented in standard open-source programs of molecular dynamics, such as LAMMPS.

While the solubilities obtained were in reasonable agreement with experimental results, most of the force fields proposed were developed to reproduce the properties of the liquid phase, so that the solubility was not a target property. The work, however, opens the door for determining new force fields that attempt to reproduce the solubilities.

Source: “Computational methodology for solubility prediction: Application to sparingly soluble organic/inorganic materials,” by Lunna Li, Tim Totton, and Daan Frenkel, The Journal of Chemical Physics (2018). The article can be accessed at https://doi.org/10.1063/1.5040366 .