Building a Crystallization Kinetics Database through Automated Screenings

Despite the large number of studies demonstrating the application of crystallization kinetic models for process development, reported kinetic parameters across literature often differ by several orders of magnitude, even for the same compound, temperature, and solvent composition. This is due to a large number of inconsistencies in common methodologies, including different simplifications to the supersaturation expression, different assumptions in mass balance and kinetic models, different analytical tools used to measure crystal growth, and the use of different seeding strategies and mixing approaches during kinetic screening.

Methods:

This work leverages our previously designed methods for the automated screening of nucleation and growth kinetics using a Technobis crystalline. Appr. 20 solute-solvent pairs have been screened for kinetics using consistent scales, mixing, crystal size assessment techniques, and models. The pairs include small molecule pharmaceuticals with different molecular sizes and complexities, as well as some inorganic salts that serve as a point of comparison.

Results:

The resulting database is presented as a kinetic log-log map of nucleation constants against crystal growth constants. This map is used to place new compounds in context with existing literature, and to study the relationship between solute structure, solvent composition, and crystallization kinetics.

Conclusions:

The developed kinetics database serves to standardize early crystallization screenings, presenting a point of comparison between new and existing solutes, filling a role similar to that of a Robeson plot for gas permeation. As the library keeps expanding, it will serve to facilitate early process development efforts as well as to answer open academic questions on what are the main factors driving a solute’s tendency to nucleate and grow in a given solvent.

Keywords:

Crystallization, automation, kinetics, process design, modeling