(349e) Population Balance Modeling of Cooling-Antisolvent Crystallization to Derisk Tech-Transfer | AIChE

(349e) Population Balance Modeling of Cooling-Antisolvent Crystallization to Derisk Tech-Transfer


Tan, L. - Presenter, Bristol-Myers Squibb
Mitchell, N., Process Systems Enterprise
Dummeldinger, M., Bristol-Myers Squibb Co.
Engstrom, J., Bristol-Myers Squibb
Rosenbaum, T., Bristol-Myers Squibb
We have developed a population balance model (PBM) for BMS Compound A crystallization process, using gPROMS FormulatedProducts (gFP), and apply the model to predict the PSD for four ~35 kg scale plant batches, with good agreement to actual data. The PBM is initially constructed from lab scale (~10 g) crystallizations on the basis of particle size distribution (PSD) measurements and solution concentration measurements of the active pharmaceutical ingredient (API) throughout the crystallization process. The kinetic parameters associated with secondary nucleation are then refined using PSD data from four initial plant runs during process-scaleup, keeping other parameters identical to those from the laboratory experiments. The updated PBM is used to predict the PSD for another set of four plant batches, run with a different agitator type, and accurately predicts an increase in particle size. Analysis of the crystallization kinetics, simulated using gFP, indicates that secondary nucleation due to attrition is the predominant mechanism influencing the PSD in the crystallization process for compound A, thus explaining why agitator type has a strong influence on PSD. The PBM is used to predict PSD from crystallizations run under different conditions, different scales, and in different reactors using the global sensitivity analysis (GSA) module in the software. The model helps to provide a more robust particle size control strategy by employing fundamental crystallization kinetics to predict PSD, which is useful to anticipate changes upon tech transfer or future scale-up.