Abstract

This paper describes the combined application of crystal structure prediction and targeted experimental crystallization to diflunisal, a fuorinated, aromatic carboxylic acid. Graph set analysis has been used to analyze and classify the structural predictions, and crystallization solvents have been selected to promote the formation of crystals containing the most commonly predicted hydrogen-bonding motifs. Accordingly, four new crystal structures of this material were solved, two solvates and two new polymorphs. This example further highlights both the insight offered by crystal structure prediction software and the limitations of current solution crystallization strategies.