The study of carbamazepine solid forms continues to be a forum for advancing fundamental knowledge about polymorphism, hydrate formation, crystal engineering of pharmaceuticals, and structure–property relationships of pharmaceutical relevance. The bioavailability of carbamazepine, an antiepileptic drug with low water solubility, is dissolution-rate limited and is thus affected by the crystalline form used in the formulation. The study of cocrystals of carbamazepine has been a particularly active area of research, including a recent report that a cocrystal of carbamazepine and saccharin was shown to be bioequivalent to carbamazepine form III in dogs.
Much of the current interest in cocrystals of active pharmaceutical ingredients (APIs) is focused on structure–property relationships; however, there still exists a need for improved methods of screening APIs and potential coformers for cocrystal formation. Determining if a particular combination of an API and a coformer will result in a cocrystal is an empirical exercise that beneﬁts from a broad experimental search space and effective methods. We have approached the design of screening experiments with the intent of using solution based methods that have the ability to generate supersaturation with respect to the cocrystal in a liquid phase that is ideally saturated or undersaturated with respect to the reactants. The screening methods reported here operate under these conditions although supersaturation is generated in a unique manner in each method.
Without a priori knowledge of the cocrystal and its solubility, some screening methods are more likely to lead to supersaturation of reactants if the solubility of one of the reactants is much lower than the other, for example crystallization by cooling or evaporation. Other methods that rely on dissolution of reactants, or addition of reactants at or below their solubilities, generate supersaturation of cocrystal without the risk of crystallizing reactants. This behavior has been explained by the dependence of cocrystal solubility on coformer solution concentration. Dissolution of reactants (API and coformer) or dissolution of API in solutions saturated with coformer can lead to cocrystal formation in aqueous or organic solutions and by vapor sorption onto solid mixtures of reactants.
Cocrystal screening studies are generally carried out in ternary systems (API, coformer, and solvent) and phase diagrams that describe the conditions for thermodynamic stability provide insight about the paths that may lead to cocrystal formation. In practice, screening and crystallization protocols follow one of two strategies (1) use of solvents or solvent mixtures where the cocrystal congruently saturates and thus the components have similar solubility, or (2) use of nonequivalent reactant concentrations in order to reach the cocrystal stability region in noncongruently saturating solvents. In this article we report the use of these strategies with four screening methods, evaporative high throughput screening (HTS), SonicSlurry, wet cogrinding, and reaction crystallization (RC).
In order to design successful cocrystal screening experiments, it is helpful to consider reactant solubilities, as well as solution and solid phase compositions in the context of phase solubility diagrams (PSDs) and triangular phase diagrams (TPDs). Using these tools, the concepts that can lead to successful cocrystallization in each of the four screening methods reported here will be discussed and compared.
Crystallization methods and conditions for screening were generally based on one of two strategies. In Strategy 1, solvents or solvent mixtures with similar solubilities for reactants were selected and stoichiometric solution concentrations of reactants were used. In strategy 2, solvents or solvent mixtures of different solubilities for reactants were selected, and nonstoichiometric reactant solution concentrations were used. The main point in the design of experiments was to allow the solution to develop sufﬁcient supersaturation for cocrystal nucleation while avoiding supersaturation with respect to reactants.
Scott L. Childs,* Naır Rodrıguez-Hornedo,* L. Sreenivas Reddy, Adivaraha Jayasankar, Chinmay Maheshwari, Linda McCausland, Rex Shipplett and Barbara C. Stahlya. CrystEngComm, 2008, 10, 856–864