Process modelling of protein crystallisation: A case study of lysozyme

Abstract

With the rise in interest of protein crystallisation as a purification step in downstream processing, there is significant interest in the process modelling of these crystallisation steps. Herein, we demonstrate and compare the applicability of “traditional” nucleation and growth models, commonly used to model small molecule crystallisation, for the successful population balance modelling of lysozyme crystallisation at the 100 mL and 1 L scales. Results show that both empirical power-law and first-principles models for nucleation and growth provide good fits to experimental data. Results from parameter estimation highlight a high degree of model sensitivity to initial guesses and stress the importance of providing particle size estimates in order to extract sensible data from the models. Estimates obtained for the 100 mL scale provided suitable initial guesses for the 1 L scale, despite significant differences in the final values obtained at each scale. For future work, further investigation into model validation upon scale-up is recommended. The work performed demonstrates the effectiveness of population balance modelling in the prediction of protein crystallisation behaviour, regardless of the underlying physical phenomena.