Population Balance Modeling in the Salt–Cocrystal Continuum: from Kinetic Identification to a Model-Based Design Space

“Many drugs in the new chemical entity pipelines have bioavailability issues related to poor solubility and permeability. A solution to these issues is to modify the solid form of the active pharmaceutical ingredient (API). Solid-form engineering, such as the generation of different polymorphs, hydrates, solvates, and other multicomponent systems, such as salts and cocrystals, can improve the physicochemical properties of an API. Most approaches have drawbacks, such as salt formation not being viable when molecules don’t have ionizable groups, or hydrates/solvates that lack stability due to the loss of water/solvent from the crystal structure (Guo et al., 2021). Over the past two decades, cocrystals have emerged as a promising approach to optimize drug performance, offering improved solubility, stability, and bioavailability (Karimi-Jafari et al., 2018, Yousef, 2019). The cocrystallization process offers many advantages for the drug’s physicochemical properties. However, the scaling-up process from a few milligrams to kilograms can prove challenging and time consuming.”

In this paper in Chemical Engineering Science, scientists, Ankur Kapil^a, Darpandeep Aulakh^a, Ruth Lunt^a, Almos Botond Orosz^cd, Botond Szilagi^de, Lorrayne Clague^a, Zoltan K. Nagy^cd and Michael D. Hamlin^a, present a practical framework for scaling up the crystallization of the model multicomponent hemi system of pyrimethanil (an anilinopyrimidine fungicide) and L-tartaric acid (hemi-PYRTAR) using PBM.

“Growth kinetics were first established from seeded desupersaturation experiments at the 100 mL scale, and the model was subsequently refined using seeded cooling crystallization experiments to account for nucleation and agglomeration contributions. Two model variants were calibrated and evaluated: first, based on 100 mL data alone, and second, incorporating an additional 1 L experiment. Scale-up predictions at 2 L scale were validated against experimental PSD data. The calibrated model was further applied to sensitivity analysis and model-based screening region calculations to support rational selection of operating conditions.

“This approach reduced reliance on DoE and provided a mechanistic, predictive path to technology transfer. For large-scale manufacturing, our framework provides a path to predict particle attributes, improve downstream processing, and ensure reproducibility at scale. This model supports faster and more reliable transfer of multicomponent crystallization processes from development to manufacturing. To the best of authors knowledge, this is the first study to combined PBM model development, scale-up validation, SHAP based global sensitivity analysis and model-based screening region mapping for a pharmaceutical multicomponent crystallization system within the salt-cocrystal continuum.”

^a Veranova, Devens, MA, USA
^b Veranova, Cambridge, UK
^c Davidson School of Chemical Engineering, Purdue University, 47907 West Lafayette, IN, USA
^d Crysyst INC, 47907 West Lafayette, IN, USA
^e Department of Chemical and Environmental Process Engineering, Faculty of Chemical Technology and Biotechnology, Budapest University of Technology and Economics, Muegyetem rkp. 3., Budapest H-1111, Hungary