(148a) Crystal Morphology Prediction: An Alternative Framework Via Kmc

Accurate and rapid modelling of crystal morphologies is essential for various industrial applications, ranging from catalysis to manufacture of pharmaceuticals, OLEDs, and many others. Morphology prediction requires precise characterization of the growth conditions for determining the relative growth rates of the crystals faces and the ensuing morphology. This work strives to inform real-time morphology predictions for molecules of interest. To do so, we analyze growth rate predictions following step velocity model theory as well as those generated via Kinetic Monte Carlo (KMC) simulations and use existing visualization frameworks designed within the Doherty group to compare the predicted crystal morphologies.

In this work we investigate the potential of KMC simulation techniques to both inform step velocity models for non-ideal systems and to predict accurate crystal morphologies. We first demonstrate the validity of this approach by comparing simulation results to step velocity model predictions for the simpler systems of Kossel and centrosymmetric growth units without impurities present. We then expand to more complex systems including non-centrosymmetric growth units as well as systems containing impurities in solution. Herein, an impurity, or imposter, typically refers to a structurally similar molecule to the solute, or one which alters or inhibits the growth of, the primary crystallizing species. Our approach is demonstrated for real molecules grown from solution, both centrosymmetric (e.g. naphthalene, adipic acid) and non-centrosymmetric (e.g. paracetamol). Our results demonstrate the applicability of informing crystal morphology predictions for molecules of interest, including but not limited to active pharmaceutical ingredients. This work also discusses potential techniques utilized to reduce simulation time such as temporal coarse-graining methodologies and investigates the limitations of such approaches.