Loratadine Crystallization: Population Balance Model (PBM)

Crystallization CFD Simulation of Loratadine Using the PBM in ANSYS Fluent

Description

This project simulates the crystallization process of Loratadine within a dual-inlet nozzle, using the Population Balance Model (PBM) in ANSYS Fluent. The primary objective is to study particle size distribution and how mixing affects crystal formation as ethanol and water streams interact — these fluids mix to create supersaturation conditions that drive nucleation and subsequent growth of Loratadine crystals. The simulation offers useful insight into the relationship between hydrodynamics, turbulence, and crystallization kinetics, with direct relevance to designing and controlling similar industrial crystallization processes.

The computational domain represents a dual-inlet nozzle, with ethanol and water entering through separate inlets and mixing before exiting through a single outlet. The geometry was built in Design Modeler, and the domain was meshed in ANSYS Meshing using a tetrahedral mesh, with refinement concentrated in the mixing region to improve accuracy. The total mesh contained 506,677 elements, balancing accuracy against computational cost, with mesh quality verified to maintain low skewness and support stable convergence.

Methodology

The simulation combined the Eulerian multiphase model with the Realizable k-epsilon turbulence model to capture both phase interaction and turbulent mixing behavior. The Discrete Population Balance Model tracked Loratadine crystal sizes across 10 discrete bins, ranging from 1×10⁻⁸ m to 8.1681×10⁻⁸ m. A custom UDF defined the nucleation and growth kinetics governing crystal formation, allowing these rates to depend on local supersaturation and flow conditions throughout the domain.

The primary phase — an ethanol-water mixture — was modeled using the Species model, while the secondary phase represented the dispersed Loratadine particles themselves. The solver was pressure-based and transient, using a time step of 1×10⁻⁷ s with second-order upwind discretization applied for improved accuracy.

Conclusion

The simulation captured effective mixing and crystallization of Loratadine throughout the nozzle. Particle size distribution across the 10 defined bins showed smaller particles dominating the early stages of the process, consistent with active nucleation, while larger particles emerged later as growth progressed:

The UDF successfully captured crystal growth's dependence on local flow conditions, with flow results revealing strong mixing zones near the inlets and a progressively more uniform crystal distribution toward the outlet. Together, these results confirm that the combined PBM-UDF approach reliably represents the coupled interaction between flow dynamics and particle growth throughout the crystallization process.