(224e) Automated Low Solubility Conglomerate/Solid Interaction Impurity Rejection Workflow Via Front to End API Synthesis to Crystallization in-Silico Optimization

Low solubility conglomerate and solid interaction impurities are challenging to purge in crystallization process hence require extensive synthesis conditions re-optimization to suppress those impurities from formation. This work introduces an automated in-silico optimization framework that allow quick identification of optimal reaction condition setpoint that ensure critical quality attributes (CQAs) meet specifications across design space edge. A front to end mechanistic reaction kinetics model and solubility limited impurity purge (SLIP) crystallization model is built that can predict post crystallization impurity profile under different synthesis conditions. An optimization formulation is proposed applying CQA/yield penalty function at design space edge using BigM method. Finally, we present a classical case study that demonstrates optimal reaction conditions are found for different impurity solubility scenarios. This work can be applied to accelerate existing process acceptable range (PAR) experimental program when there is a need to quickly search for new reaction conditions setpoint across entire API synthesis step.