Screening Crystallization Solvents for Yield and Purity Upgrade Using High-Throughput Solubility

Abstract Crystallization is ubiquitous as a method to purify chemical intermediates and drug substances across the pharmaceutical industry. A significant challenge in the development of a crystallization process is the identification of an optimal solvent system that provides high crystallization yield and purity. In this work, we describe a screening approach leveraging high-throughput experimentation and modeling to estimate impurity rejection and crystallization yield across a range of solvents and antisolvents. Models to compute impurity rejection were informed on the basis of the experimentally determined impurity retention mechanisms. Despite greater error associated with predicting yield and purity upgrade for solid solutions as compared to systems with immiscible impurities, our proposed method accurately estimated the yield (R2 = 0.88) and purity (R2 = 0.73) achieved by crystallizations in a variety of solvents and antisolvents for three industrially relevant compounds. These results demonstrate a practical workflow where high-throughput solubility can inform the crystallization process design space, offering a data-rich approach for crystallization development across the pharmaceutical industry.