Active pharmaceutical ingredients (APIs) and intermediates are purified by crystallization and isolated from the mother liquor by filtration and drying. Pressure filtration on an agitated filter dryer (AFD) is widely used for the APIs and intermediates due to its ease of operation. However, the filtration typically requires long cycle times due to the widely different cake properties of APIs, posing to be bottleneck in manufacturing. We present design and development of automated data-rich leaf filtration platform to efficiently evaluate filtration performance and compressibility risk of APIs, while requiring fewer experiments and using less material.
The newly designed filtration system integrates in-house designed automation software controls for the automated slurry feeding, measurement of cake height via laser/acoustic sensors, solvent pumps, balance, and valves. Leaf filter consists of a transparent glass body for potential use of camera vision-based analysis of cake properties. We enabled automated dynamic pressure modulation (DPM), allowing a single data-rich filtration experiment with ascending pressures for advancing the fidelity of filtration predictions. We demonstrate several case studies of this platform for predicting filtration performance on larger scales, while significantly reducing the cycle time and human interventions.
