Computer-Aided Solvent Selection for the Design of Integrated Synthesis, Crystallization, and Isolation Processes

The extensive use of solvents in pharmaceutical manufacturing has major impacts on overall process performance and final product quality. Solvent selection often requires tedious and resource-intensive experimental investigations. Nevertheless, the growing interest in embedding sustainable practices in process development workflows has motivated pharmaceutical manufacturers and researchers to deploy digital tools for enhancing process understanding and guiding the design of greener materials and processes. In this work, a computer-aided mixture/blend design (CAM ^b D) methodology is developed to identify simultaneously the optimal solvent or solvent/antisolvent mixtures, mixture compositions, and process conditions in end-to-end processes, considering integrated synthesis, crystallization, and isolation. Within the proposed approach, thermodynamic models are used to couple property prediction with process modeling to describe pharmaceutical process systems, while mathematical optimization techniques are applied to search the large design space of potential solvents and process conditions and identify the most promising designs among thousands of possible options. Further consideration is also given to the integrated design of reaction, crystallization, and wash solvents, where design decisions across the entire process are interlinked. The methodology is applied to the synthesis, crystallization, and isolation of mefenamic acid as a case study, while considering a broad range of process-wide key performance indicators (KPIs) that quantify product quality and resource efficiency, such as product purity and process energy consumption. In order to explore the trade-offs between competing KPIs, the proposed approach is extended to solve multiobjective design problems and generate a list of Pareto-optimal designs. The results highlight that the optimal designs are strongly dependent on the criteria considered and motivate the need for an integrated approach to solvent selection and process design. The findings of this study are expected to contribute to the digitalization of solvent selection for the accelerated development of sustainable pharmaceutical processes.