Multiscale Computational Model for Polymorphism and Growth of Crystals in Solution Crystallization

Understanding the self-assembly of molecules during crystallization is critical for the precise synthesis of crystalline materials. Diverse types of materials such nanomaterials, active pharmaceutical ingredients (APIs), proteins, zeolites, metal- and covalent-organic frameworks are manufactured using crystallization. Although crystallization is widely used, the relationship between molecular events during crystallization and the outcome of crystallization is not fully known. Such understanding is limited because of a large number of entities, long time and length scales, highly stochastic nature, presence of multiple energy minima in the crystal energy landscape, and complex interaction of solute-solvent molecules involved in the process of crystallization. To gain mechanistic insights and relate the outcome of crystallization with molecular events, I derive a multi-scale model. The multi-scale model combines molecular simulations, a semi-classical double-well approach, non-equilibrium sampling techniques, and continuous mathematical models to relate the solute-solvent interactions with the experimentally observable properties such as nucleation and growth rates. The multi-scale model is tested with the help of small organic molecules such as glutamic acid and histidine, as well as porous crystalline frameworks such as UiO-66 and COF-5. The model can reproduce the experimental results and link the smaller time-scale events, such as the exchange of solvent molecules in the solvation shell, with the experimentally observed crystal structure and morphology. In the case of crystallization of organic frameworks, the model can also predict the rate of formation of crystals due to oriented attachment quantitatively. Furthermore, the mechanistic insights derived from such analysis unify the theoretical and empirical observations laid down since the inception of crystallization research.