(367e) Predicting the Impact of Particle Size and Shape Distributions on Filtration Performance of Crystallized Products

Filtration is a crucial downstream step in pharmaceutical, fine chemical, and agrochemical industries, where crystallized products are common. The filter cakes formed during this process have characteristics such as structure, porosity, and permeability that influence subsequent steps including washing and drying. The impact of particle size and shape distributions (PSSDs) for a case of needle-like crystals on cake structure and resistance has recently been demonstrated experimentally, with results validating a model for cake structure and average porosity.¹ However, metrics such as cake resistance or filtration time, crucial for industrial process optimization, are yet to be modelled and predicted.

This study uses a combined experimental and modelling approach to predict key filtration metrics for needle-like crystallized products based on their PSSD. Such models can significantly reduce the time and cost associated with resource-intensive experimental trials. In the modelling phase, a two-step approach is proposed. First, a non-equilibrium Monte Carlo (MC) model simulates the cake’s formation, using in-house software. Second, fluid flow through the compacted cake’s final structure at steady-state conditions is modelled with computational fluid dynamics (CFD), in OpenFOAM. Compared to discrete element method (DEM)-CFD, which performs iterative CFD simulations on multiple packings, this framework requires less computational effort. Additionally, compared to forced-based DEM, MC requires fewer input parameters. This MC-CFD framework was applied to several case studies, examining PSSD characteristics including shape (spherical and needle-like), polydispersity, and operating conditions such as pressure drop. In the experimental phase, populations with varied PSSDs were created and characterized using novel imaging-based techniques.² Constant pressure filtration experiments were performed using a purpose-built filtration rig, with the results compared to simulations.

This work introduces a predictive tool that minimizes experimental reliance, enabling faster and more cost-effective filterability assessments in research and development. It also lays the groundwork for expanding into modelling washing and drying processes.

References

1. Perini, G. Predictive Design of Filtration Processes - The Pharmaceutical Industry and the Impact of Crystal Size and Shape, Ph.D. Thesis, The University of Manchester, 2020.
2. Rajagopalan, A. K. A Dual Projection Imaging System to Characterize Crystallization Processes: Design and Applications, Ph.D. Thesis, ETH Zurich, 2019.