(284f) Quality By Control (QbC) Guided Quality By Design (QbD) for the Optimal Operation of Integrated Crystallizer-Wet Mill Systems

The simultaneous control of crystal shape and size is particularly important in fine chemical and, especially, pharmaceutical crystallization. These two quantities not only influence significantly the dissolution rate and bioavailability of final drug products, but also contribute to the manufacturability and efficiency of downstream operations. The manipulation of crystal shape, however, is difficult since it requires the decoupled growth rate control of individual crystal faces. The supersaturation and solvent system dependency of these rates are often not strong enough to enable impactful crystal shape control through temperature and/or solvent composition variation. Promising results have been achieved by using tailored growth rate modifier additives in the case of model systems, however, this does not provide a general framework for controlling the crystal shape.¹

If the crystal size or shape cannot be achieved in the crystallizer - because the target is out of the achievable domain of the crystallization process – downstream correcting steps are involved. The most often used processes are granulation and/or milling. Recently, the wet-mill process was integrated with the crystallizer in different ways to increase the achievable crystal size domain. It was observed that the integrated operation leads to significant improvements in the produced crystal size.² However, similar studies have not been carried out yet for crystal shape control problems.

The aim of this work is to analyze and evaluate the possibilities of controlling the bivariate size distribution (which by definition involves crystal size and shape information) of high-aspect ratio crystals in an integrated crystallizer-external wet mill system. It is considered that in the crystallizer primary and secondary nucleation, crystal growth and dissolution along the length and width axes occur, whereas in the wet mill – in addition to the aforementioned mechanisms - fragmentation and attrition is also considered. The advanced hydrodynamic description enables the direct implication of stirrer revolution speed in the model. The generated system of hyperbolic partial differential and ordinary differential equations is solved by the fully discretized high resolution finite volume method, involving GPU acceleration, which was showed to bring significant computational time reduction for bivariate population balance equations.³ The dynamic optimization of wet-mill rotation speed, crystallizer temperature and recirculation flowrate revealed that the target bivariate crystal size distribution can be much better achieved with the integrated system than manipulating only the crystallization conditions.

The optimum operating profiles obtained with respect to various constraints, objectives and initial conditions show clear similarities, which suggests that the general operating pattern is system independent. After this discovery enabled by the modelling and process optimization, or, in broader concept, by the QbC approach, the methods of experimental design, or, on broader concept, the QbD, can be applied to determine the exact values of heating/cooling rates as well as the stirring rate of the internal/external wet mill for a particular crystallization system. The experimental validation of the method is in progress.

References

1. Klapwijk, A. R., Simone, E., Nagy, Z. K. & Wilson, C. C., Cryst. Growth Des. 16, 4349–4359 (2016).
2. Yang, Y., Song, L., Gao, T. & Nagy, Z. K., Cryst. Growth Des. 15, 5879–5885 (2015).
3. Szilágyi, B. & Nagy, Z. K., Comput. Chem. Eng. 91, 167–181 (2016).