(468g) Effects of Scale-up on the Mechanism and Kinetics of Crystal Nucleation

Crystal nucleation is an important step in solution crystallization of pharmaceutical compounds as it determines the crystal shape, polymorphic form, chiral form and the crystal size distribution of the product. Understanding this fundamental phase separation of molecules from the liquid to the solid phase is essential to exhibit process control. Nucleation is a stochastic process for unseeded crystallization processes involving low supersaturation levels or for small crystallizer volumes. Stirred small volume solution methods can be utilized to measure the stochastic nature of nucleation and to estimate nucleation kinetics. This way, probability distributions of induction times can be obtained from isothermal induction time experiments which can be used in conjunction with the classical nucleation theory (CNT) to estimate the nucleation rate and other important nucleation parameters. However, it remains unclear how scale-up affects the nucleation mechanism and nucleation kinetics in solution crystallization processes as the isothermal induction time method is currently limited to small scale experiments.

In this work, the effects of scale-up on the mechanism and kinetics of crystal nucleation will be presented. Four different crystallization setups were investigated, ranging from small magnetically-stirred 10 mL solutions to overhead-stirred solutions of 680 mL. The onset of crystallization was characterized using a Mettler Toledo Focussed Beam Reflectance Measurement (FBRM) probe. The crystallizer was programmed to start a new experiment after the FBRM probe detected crystallization. This way, many induction times were recorded in an automated fashion. The nucleation kinetics were estimated from the probability distributions of the induction times.

Nucleation kinetics were estimated for paracetamol in 2-propanol in the different crystallization platforms and at different supersaturation ratios. The nucleation rate was an order of magnitude faster in the magnetically-stirred crystallizer as compared to the crystallizers involving overhead stirring. The thermodynamic part of nucleation did not significantly change the nucleation rate whereas the kinetic nucleation parameter was found to be the rate-determining process when the crystallization process was scaled-up. In particular, the shear rate was rationalized to be the part of the kinetic parameter that changes most significantly when the crystallization process was scaled-up. The shear rate increases with volume and this effect becomes stable after the experiment has sufficiently increased in volume. In addition to these kinetic results, experiments involving the chiral sodium chlorate showed that the single nucleus mechanism is the underlying nucleation mechanism in all four tested crystallization setups. On the other hand it was found that cooling crystallization experiments were driven by a multi-nucleus mechanism. The presence of the single nucleus mechanism provides support for the use of probability distributions of induction times to estimate nucleation kinetics across all four tested crystallization setups.

In conclusion, a robust methodology was used to automatically record induction times using an FBRM probe. This approach is easily transferable and allows to capture induction times in different crystallizer types and volumes. The results presented herein on paracetamol and sodium chlorate contribute to a better understanding of the scale-up effects on the mechanism and kinetics of crystal nucleation of fine chemicals.