(139e) Combining Simulation and Experiment to Understand Polymorph Selection of Drug Molecules in Different Solvents

The importance of polymorph selection in drug discovery and development has been well recognized. Effects of polymorphism on multiple properties of drugs, such as physical and chemical stability, bioavailability and manufacturability. Nucleation is the first stage of crystallization that could determine the formation of a specific crystalline structure. Studying nucleation is challenging, both experimentally and using computer simulations. Nucleation is a stochastic process and only involves the order of only 100 to 1000 molecules, making it challenging to design experiments to study. On the other hand, nucleation is also a rare event, occurring on time scales far beyond those accessible to simple molecular simulation methods.

In this work, we combined simulations and experiments to study the nucleation of small molecules in different solvents. As the Ostwald’s rule stated that the polymorph with a lower nucleation energy barrier always crystallizes first in general, our simulations could provide information on polymorph selection that are consistent with the experimental data. Two polymorphs of sulfamerazine (Form I and II), and three different solvents, water, methanol and acetonitrile, were selected for our study. By implementing the String Method in Collective Variables (SMCV) in combination with the construction of order parameters, we sketched the minimum free energy paths of nucleation and calculated the energy barriers for both of the two forms. Simulation results showed that Form I has lower energy barriers in all the solvents than Form II, indicating that Form I would be easier to nucleate. PXRD data from our crystallization experiments showed consistent results that Form I was the favorable form in these solvents.