Distillation Process Understanding from Analytical Monitoring and Computational Modeling

Aspen is a modeling tool used across multiple industries to simulate batch and continuous processes. When utilized during pharmaceutical small molecule development, an Aspen model can reduce the amount of laboratory testing needed before scale up by accurately predicting process parameters. For this particular study, an Aspen model was developed to understand solvent composition during a batch distillation process. Specifically, following the synthesis of an intermediate grade API (active pharmaceutical ingredient), a vacuum distillation is eventually performed to solvent swap from the extraction solvent (tert-butyl methyl ether, TBME) to the crystallization solvent (acetic acid). The goal for this work was to understand the molar ratio of TBME to acetic acid as the distillation progressed to determine the end point of the distillation without offline solvent composition analysis. In addition to Aspen modeling, real time composition information was obtained by monitoring the composition of the vapor condensation on the tip of an IR probe inserted into the reaction vessel. This study will illustrate how analytical monitoring by online IR and offline NMR has been combined with computational modeling in ASPEN Batch Modeler to generate process knowledge for the vacuum distillation, a key unit operation that needs to reduce the TBME:acetic acid molar ratio below a specified target to meet process quality goals. Knowledge around the distillation profile generated as a result of this study contributes to general process knowledge and can potentially eliminate the need to confirm the distillation endpoint via offline solvent composition analysis.