Phase equilibria thermodynamics affects all facets of our daily lives. Of particular interest to the present study is the role of phase equilibria in the drug design process, more specifically, the solubility of pharmaceutical solids. This is crucial not only to predict bioavailability and toxicity, but is essential for production and formulation as well. Given the importance of the problem, many of the great scientific minds of the past century have attempted to develop theories to predict solubility. However, the current state-of the-art involves correlations of large sets of experimental solubility data.
A promising alternative design tool is molecular simulation. Not only would this allow for solubility predictions to be made, but molecular level insight into why a drug is soluble or insoluble can be obtained. However, molecular simulation has not yet evolved to a level in which it may be used as a practical design tool by novice modelers in industry. The focus of this dissertation is on developing molecular simulation tools useful for design purposes by a novice.
First, a formalism in which the solution phase residual chemical potential of a solute may be computed in an efficient, automated and reliable fashion is demonstrated. The method requires performing only a single simulation at each concentration, and may be used to reliably sample molecules of complex topologies. These calculations form the basis of novel methods developed to predict solubility. A general scheme is first presented wherein the chemical potential of the solid phase and the concentration dependent chemical potential of the solution phase are computed relative to a common reference. Next, it is demonstrated how solution theory may complement molecular simulation, and reduce the computational demand required to make solubility predictions. A novel method is demonstrated wherein the solution phase properties at finite concentration may be estimated from a single molecular simulation at infinite dilution combined with established solution theory. The method is applied to both phenanthrene and acetaminophen in pure and mixed solvents, with promising results comparable to the current state-of-the-art. Additionally, a hydrophobicity scale is developed to understand the phase behavior of proteins in ionic liquids.