(42a) Importance of Cocrystal Discovery Workflow for API Purification and Development | AIChE

(42a) Importance of Cocrystal Discovery Workflow for API Purification and Development


Wood, G., Pfizer
Pickard, F. IV, Pfizer
Jones, K., Pfizer Inc
Chekal, B. P., Pfizer Inc
Importance of cocrystal discovery workflow for API purification and development

Kulkarni, Samir A., Wood, Geoffrey, Pickard, Frank, Kris Jones and Chekal Brian P.

Pfizer World Research & Development, Groton Laboratories, Groton, CT 06340, United States

In the current work, we describe a systematic and effective method to discover new co-crystals. The method consists of comparison of different in silico methods to find the list of co-formers suitable to form co-crystal with the given target compound. The experimental techniques for synthesizing co-crystals are time consuming and expensive, our goal is to develop an in silico method to rationalize co-crystal formation. Using the computational and experimental data generated by a researcher at Pfizer, we critically evaluated the predictive prowess of several co-crystal screening tools in use at Pfizer today, including: gFIT1, multicomponent hydrogen bond propensity2 with shape matching3, and quantum chemistry-based clustering analyses. Our computational screens ranked coformers based on their predicted likelihood to form co-crystals. A computational assay was considered “successful” if it able to correctly predict co-crystal synthesis amongst the top ranked coformers. This blind analysis of our current computational methods also allowed us to better understand where they succeed, and more importantly where they fall short. We ultimately used these results to refine our computational workflows, and improve their predictive prowess.

Co-crystal development is important to the pharmaceutical industry because it offers the opportunity to modify the physiochemical properties of crystalline API without altering its covalent structure. While there are many computational assays that promise to rationally predict co-crystals, including: hydrogen bonding propensities, solubility parameters, crystal database screening, thermodynamic characteristics, etc., in reality the most reliable method to discover and prepare new co-crystals is still based on empirical screening.3-4

Co-crystal formation is mediated by strong intermolecular interactions, such as: hydrogen bonding, halogen bonding and π-π interactions. This project will study the influence of individual solvents on the formation of co-crystals, specifically what is an API’s affinity for self-association relative to its affinity to the coformer. Also, the role of solvent in co-crystal formation remains poorly understood and solvent can be critical in obtaining particular co-crystals from solution. Once we determined the solvents and crystallization conditions, then Crystal16 (a small scale multiple reactor system) was used to determine the change in saturation temperature due to their ability to form stable co-crystals. If the co-crystal is stable, then its solubility will be lower than the solubility of its pure components. We then generated a two-phase diagram by mixing the pure components together in a selected solvent system as well as the solubility of pure components in the given solvent system. This method provides an experimental assay to efficiently screen many different coformers in a high throughput fashion.3 The solvent drop grinding method will also be used as an initial screening tool for co-crystal screening.

This combined computational and experimental work helped us to evaluate the possible platform to discover new co-crystal systems in various solvent environments, and this high quality experimental reference data will ultimately allow us to critically evaluate available computational methods for predicting co-crystal formation.


  1. Abramov, Y. A.; Loschen, C.; Klamt, A., Rational coformer or solvent selection for pharmaceutical cocrystallization or desolvation. Journal of pharmaceutical sciences 2012, 101 (10), 3687-97.
  2. Galek, P. T. A.; Allen, F. H.; Fabian, L.; Feeder, N., Knowledge-based H-bond prediction to aid experimental polymorph screening. CrystEngComm 2009, 11 (12), 2634-2639.
  3. ter Horst, J. H.; Deij, M. A.; Cains, P. W., Discovering New Co-Crystals. Crystal Growth & Design 2009, 9 (3), 1531-1537.
  4. Gadade, D. D.; Pekamwar, S. S., Pharmaceutical Cocrystals: Regulatory and Strategic Aspects, Design and Development. Advanced Pharmaceutical Bulletin 2016, 6 (4), 479-494.
  5. Kulkarni, S. A.; McGarrity, E. S.; Meekes, H.; ter Horst, J. H., Isonicotinamide self-association: the link between solvent and polymorph nucleation. Chemical communications (Cambridge, England) 2012, 48 (41), 4983-5.