Modelling and Simulation Applied to Sonocrystallization in Continuous-Flow Millichannel Contactors

Luca Mazzei
Department of Chemical Engineering, University College London, UK

Crystallization is widely employed in pharmaceuticals manufacturing; over 90% of all pharmaceutical products contain drugs in particulate, and usually crystalline, form. Typically, crystallizers are standard agitated vessels, and seeding is used to direct and control the nucleation process. Major challenges persist in conventional crystallizers regarding process controllability and product reproducibility owing to the complexity of the crystallization process, which involves several rapid, concurrent and closely interacting phenomena (e.g. nucleation and growth); this leads to wide particle size distributions and substantial batch-to-batch product variability, resulting in pharmaceutical formulation problems, such as bioavailability and drug stability. To overcome these challenges, one should design new-generation crystallizers able to deliver a step-change in particle production technology.

Continuous-flow processing and ultrasound have the potential to overcome these issues. Continuous crystallizers can operate in steady-state conditions and, if their size is small, intensify heat and mass transfer, thus improving process control and product reproducibility; ultrasound induce cavitation and acoustic streaming, allowing us to trigger nucleation on demand and alter the flow patterns and mixing within the unit. But to harness this potential, one should fully understand the effect of each additional degree of freedom introduced by ultrasound and be able to characterize the ultrasonic field and flow patterns in the crystallizer. Modelling and simulations can help understand better sonocrystallization and develop, design and optimize sonocrystallizers.

In this presentation, we investigate continuous-flow sonocrystallization of adipic acid in milli-channels. We use modelling and simulations to characterize cavitation and acoustic streaming in the crystallizer, relating them to nucleation, flow pattern, residence time distribution (RTD), mixing and crystallization performance. We show that inertial cavitation is closely related to nucleation and acoustic streaming can strongly affect the performance and behaviour of the crystallizer. In particular, a small increase in the capillary diameter can invert important trends, such as the mean crystal size against the sonication time: while in a 1.55 mm I.D. capillary the former increases with the latter, in a 3.2 mm I.D. capillary the opposite happens. This difference highlights the importance of acoustic streaming and shows that modelling and simulations are important design and optimization tools.