(342g) A Coupled Modeling Framework As a Soft Sensor for Monitoring the Drying of Active Pharmaceutical Ingredient Powders in Drug Substance Manufacturing Process Trains

Active pharmaceutical ingredients (APIs) are the key components in pharmaceutical formulations that give the desired therapeutic effects. Several products are manufactured as oral solid dosage products in the form of either tablet or capsules. Prior to blending, granulation and milling of the final blend of APIs and other filler products, the APIs themselves are taken through several size engineering steps post synthesis in the form of crystallization, filtration, washing and drying. APIs being particulate products have several metrics associated with them such as a mean size, span, modalities and spread of the particle size distribution (PSD). The steps from filtration to the drying of the APIs can be carried out in agitated filter dryers (AFDs) which are versatile and widely used equipment in the industry.

Population balance models (PBMs) are the go-to tools for building simple yet reasonably mechanistic models for tracking the changes in a PSD of a particulate system during an engineering unit operation. Bulk heat transfer (HT) equation models are used to predict the heating and drying rates of the particulate API material. In this work, a PBM was coupled with a bulk HT model to calculate the drying rate and a real-time ‘wetness’ measurement of the particles. The coupled model was employed to give the PSD and wetness tracking of the bed during drying operations in an AFD for API development. The PBM consisted of various agglomeration and attrition phenomena occuring during the process which have been encoded as simple kinetic rate expressions as a function of important material properties of the components and the process variables relevant to an AFD operation. The coupled PBM-HT model framework has been proposed as a soft-sensor tool in manufacturing settings physical process analytical technology (PAT) tools cannot be employed due to safety regulations. The model would detect the endpoint for the drying and advise the engineer to end the operation a priori thereby leading to model-guided experimental studies.

The in silico model results generated have been compared to the various data generated using the in-house novel laser based PAT tool in our lab. The drying has been carried out for generic model API particles and anti-solvent combinations. The effect on the changes in PSDs have been reported along with the drying profiles. It is hoped that these model guided experimental studies would serve as building blocks for further development of AFD operations of APIs in pharmaceutical industry.

[1]: Pandit, Ajinkya, Qihang Zhang, Moo Sun Hong, Wenlong Tang, Charles D. Papageorgiou, Neda Nazemifard, Yihui Yang et al. "Laser Speckle Probe for Monitoring Pharmaceutical Drying." In 2023 AIChE Annual Meeting. AIChE, 2023.

[2]: Zhang, Q., Gamekkanda, J.C., Pandit, A. et al. Extracting particle size distribution from laser speckle with a physics-enhanced autocorrelation-based estimator (PEACE). Nat Commun 14, 1159 (2023). https://doi.org/10.1038/s41467-023-36816-2