Prediction of Agglomerate Type during Scale-Up of a Batch Crystallization Using Computational Fluid Dynamics Models

Abstract

The impact of hydrodynamics on agglomeration during the crystallization of an active pharmaceutical ingredient (API) was investigated. The type of agglomerate formed was experimentally observed to correlate with agitation level at the laboratory and kilo-lab scales. It was hypothesized that differences in agglomerate type were related to differences in the collision rate of primary crystals, caused by differences in the local degree of agitation (e.g., the local values of fluid turbulence dissipation rate, ε). Spatial distributions of ε were determined from computational fluid dynamics (CFD) models at process scales ranging from laboratory (200 mL) to commercial scale (875 L). Higher values of ε were calculated for conditions shown to result in the formation of rounded, compact agglomerates, while at the lower values of ε, looser agglomerates of flakelike particles were observed. Predictions for pilot-plant- and commercial-scale crystallization operating conditions were made using local ε values as the scaling parameter.