Process design and modeling to achieve high yield and optimum purity for continuous synthesis of carbamazepine

Abstract

Chemical synthesis is an essential step in the drug substance manufacturing process as it affects both the yield and purity of the product. This paper focuses on the use of reaction network kinetic modeling to design a high yielding continuous synthesis of an anti-epileptic drug substance carbamazepine (CBZ) that minimizes impurity formation. The reactor system was constructed with two continuously stirred tank reactors (CSTRs) in series. The synthesis was carried out by reacting potassium cyanate (KOCN) with iminostilbene (ISB) in acetic acid at concentration levels of ISB above the room temperature solubility to improve reaction kinetics and yield. The outlet of the synthesis reactor was followed by an integrated continuous precipitation step to obtain CBZ precipitate. The reaction system itself was studied first in batch reactions to determine kinetic parameters, including reactant orders, of four primary reactions. Using the reactant order kinetic parameters obtained by the batch kinetic study, causes of discrepancies between the batch and continuous systems were theorized and mitigated by adjusting (1) the KOCN addition method, (2) the KOCN addition split ratio (between the 2 CSTRs), and (3) the ISB dissolution method (above room temperature solubility). After synthesis optimization was completed, a single continuous cooling crystallization in ethanol was performed on the CBZ precipitate to obtain the target polymorphic form (CBZ form III) within the impurity limits of USP. Overall, this paper supports the design and optimization of a continuous drug substance synthesis process of CBZ by demonstrating the capability of reaction network modeling for maximizing yield and minimizing impurities.