Rapid Crystallization Process Development Strategy from Lab to Industrial Scale with PAT Tools in Skid Configuration

Abstract

Batch cooling crystallization is a commonly used separation and purification step in the pharmaceutical industry. Various properties of the crystalline product from a batch crystallizer can have a strong impact on the efficiency of downstream processes such as filtration and drying, on the formulation process and on the dissolution behaviour of the drug. Development of the crystallization processes presents a major challenge in the process development of an active pharmaceutical ingredient (API). Therefore, it is beneficial to develop a rapid crystallization process development strategy to industrial scale. In this paper we present a strategy for rapid process development and apply this strategy for androsta-1,4-diene-3,17-dione, cyclic 17-(2,2-dimethyltrimethylene acetal), a pharmaceutical intermediate produced by Merck Sharp and Dohme. The major advantages of the strategy are that there is no requirement of the crystallizer design modification, the calibration of the process analytical technology (PAT) tools can be performed at industrial scale, and the determination of the operating window can be done directly at the industrial scale. This strategy allows for process optimization directly at the industrial scale, thus eliminating the need for time-intensive scale-dependent study. The implementation of this strategy at industrial scale was performed with the help of PAT tools arranged in a unique skid-based configuration. The skid which contains both the concentration sensors and the crystal size distribution (CSD) sensors can be connected to the existing crystallizers, thereby avoiding the time and cost-intensive modifications in the crystallizer design. The modular nature of the skid offers opportunities to choose the PAT tools which complement the solute–solvent model system. The skid makes it possible to gather the relevant information concerning the thermodynamics and kinetics of the model system in situ during the crystallization runs at the industrial scale. A strategy for process development based on a sensor skid is beneficial for the industry as it is intrinsically rapid and can be combined with the development of control strategies which lead to consistent product quality.