(528a) Finite-Size Effects on the Chemical Potentials of Molecular Solids | AIChE

(528a) Finite-Size Effects on the Chemical Potentials of Molecular Solids

Authors 

Dybeck, E. - Presenter, Pfizer Inc.
Wood, G., Pfizer
Numerous important properties of solid oral therapeutics are associated with the chemical potential of the active pharmaceutical ingredient (API) in the solid form. These properties include aqueous solubility, dissolution rate, and physical stability in its crystalline form among others. Molecular modeling of the crystal structures of APIs to estimate the chemical potentials has become an increasingly popular approach in recent years for confirming the suitability of API solid forms for further development in silico. However, due to the computational expense of these approaches, the models are limited to time and length scales far smaller than macroscopic materials, and rely on simplifications such as microscopic cells with periodic boundary conditions, as well as cut-offs for long-range interactions. These approximations can have a significant effect on the chemical potential of organic small molecule crystals, particularly at small system sizes.

Here we systematically examine the effect of supercell size and long-range cutoff distance on the computed chemical potentials of pharmaceutically relevant molecular crystals using classical molecular dynamics simulations. The Gibbs free energies of the crystal polymorphs and liquid phases of carbamazepine and sulfamerazine were computed from 0 K to 600 K for various system sizes and long-range cutoff radii. We compare the estimated chemical potentials with small systems and short cutoffs to those of the corresponding larger systems and show how these differences change with temperature. Furthermore, we show the effects that these small perturbations to the chemical potentials have on the macroscopic observables of enantiotropic transition temperature and solid-liquid melting temperature. Finally, we make recommendations for future molecular simulations of API crystal structures to balance the trade-off of chemical accuracy and computational efficiency.