Session 1: Granulation - The Big Picture Agglomeration at the Sharp End - Industrial Practice and Needs

Not Yet Available Detecting and Measuring Nucleation in High Shear Granulation M. Oullion , G.K. Reynolds 2 and M.J. Hounslow 1 Session 4: High Shear Processes 2 23 1) University of Sheffield, UK 2) AstraZeneca, UK Particulate processes are of great importance in industries like chemical, agricultural or pharmaceutical industries [1]. High-shear granulation is commonly part of the solid product design chain. Consequently, there is a growing interest for this process. A wide range of end-use properties can be modified thanks to this method (flow-ability, dissolution rate, composition, particle size etc.). However, the number of granule properties changing with time makes it difficult to control. In addition, it involves several phases (solid, liquid, gas). The properties analysed during the process are usually limited (size, strength, porosity), the evolution of the liquid phase distribution being rarely measured. Consequently the nucleation mechanism resulting from the interactions between the liquid binder and the primary powder is poorly understood. This observation has led us to focus our attention on the liquid phase behaviour during the early stage of the process. A study has been performed on industrial compounds provided by AstraZeneca. Several samples were withdrawn during each granulation batch. The particle size distribution (PSD) and the liquid content were measured. Optical microscopy coupled with Image Analysis was used to determine the PSD, because of its accuracy on small size particles. A liquid tracing procedure was implemented in order to quantify the liquid content. The impacts of the water addition technique, of the drop size distribution [2] and of the impeller speed [3] were investigated. The results reveal the occurrence of a layering phenomenon inducing a binder content distribution with size. The competition between the different mechanisms (nucleation, aggregation, breakage) is also discussed. Finally, it is planned to use a 2-D model and simulate this experimental data in order to estimate the mechanistic kinetics of the process.