(357p) Computational Investigation of the Kinetics and Thermodynamics of Crystal Nucleation

Abstract: Crystal nucleation is of fundamental importance for the modeling and control of many processes. Examples can be given from atmospheric sciences, biology, materials research, and, of course, the pharmaceutical industry. Despite its ubiquity, we do not have a complete understanding of crystal nucleation due to challenges in experiments that arise from the stochasticity of nucleation happening at the nanoscale. Direct observation of nucleation in experiments is still challenging and rare, which makes molecular simulations an invaluable asset in understanding nucleation. Modeling of industrial crystallizers requires knowledge of nucleation rates and solubilities of multiple crystal phases, which can be hard to get experimentally. A complete in silico access to these kinetic and thermodynamics properties can significantly reduce the time and money that goes into experiments.

My research in the Ramkrishna Group at Purdue University focuses on the use of molecular simulations to gather kinetic and thermodynamic information of crystal nucleation. In this poster, I will provide an overview of my past and present work on obtaining free energy surfaces, calculating nucleation rates, and solubilities for different systems. More specifically, I will share details on multidimensional free energy surfaces which provide information about nucleation mechanisms at changing supersaturations, use of appropriate free energy theories to extract thermodynamic properties, combining first-principles calculations and machine learning for improved prediction of solubilities, and dependence of nucleation rates on polymorphic composition of the nuclei. Moreover, I will introduce a multiscale modeling approach for predicting the outcome of a crystallization process solely from first principles.

Research Interests: I am interested in working towards creating automated computational tools to expedite molecular discovery and production. I am particularly interested in pharmaceutical applications, where computer aided property prediction can benefit the drug production pipeline in multiple stages from drug discovery to formulation development.