Kinetics of Polymorphic Phase Transformations of o-Aminobenzoic Acid: Derivation and Application of Exact and Approximate Dispersive Kinetic Models Plus Molecular Dynamics Simulation of Pre-Nucleation Aggregates

The specific rate at which one crystalline phase converts to another can vary as a function of time under isothermal conditions. This behavior gives rise to sigmoidal kinetic transients that are characteristic of nucleation and growth. Such curves are commonly fitted using the Johnson-Mehl-Avrami-Erofe'ev-Kolmogorov (JMAEK) equation. However, due to the ambiguity surrounding time exponent in the JMAEK model, we present an alternative two-parameter model and apply it to the study of solution-mediated polymorphic conversions of the prototypical molecule, o-aminobenzoic acid (o-ABA) [Jiang, S.; Jansens, P. J.; ter Horst, J. H. Control over polymorph formation of o-aminobenzoic acid. Cryst. Growth Des. 2010, 10, 2541-2547]. Using our dispersive kinetic model (DKM) a distribution of activation energies, D(E), is reconstructed from each experimental transient. Then, using each D(E), a particle size distribution (PSD) of the critical nuclei formed during each phase transformation is predicted. Lastly, molecular dynamics (MD) simulations are performed to study the pre-nucleation behavior of o-ABA in solution, under experimentally relevant conditions, to complement the kinetic modeling of the phase conversion in the solid-state. We observe that o-ABA molecules weakly associate with each other to form a variety of "loose" aggregates. These aggregates are mostly dimers and trimers exhibiting H-bonding and π-π interactions in different configurations that generally do not conform to the known crystal packing arrangements of the common o-ABA polymorphs. Therefore, the observed pre-nucleation self-association is more consistent with a non-classical nucleation pathway whereby monomer densification occurs ahead of cluster formation and, eventually, structural ordering. Our molecular-level simulations in solution complement the earlier study performed using experimental measurements on bulk crystals, with the DKM serving to bridge the scale gap between the two approaches by providing a window into the nano-scale species controlling the overall rate of conversion.