(119a) Advancements in the Prediction and Control of Batch High Shear Granulation and Continuous Twin Screw Granulation

Batch high shear granulation and continuous twin screw granulation can both be successfully predicted and controlled using equations relating to the granulation work done per unit mass of dry powder.   Models based on work per unit mass can and have been used to successfully control granulation processes in real time, and was used as part of a QbD NDA Submission for a drug product where the granulation process needed to be controlled within narrow limits.   These models have an additional advantage in that they can be developed based upon work per unit mass, which allows a sampling, substantially reducing the raw material requirements needed to create a response surface.

  Material surface area has a significant impact on the water and work requirements for granulation.   If a Work model is used to control a granulation process, then material surface area will need to be controlled within limits, or an adjustment will need to be made to the target Work value for the batch.   Methodologies to do this are discussed.         

  If an even finer degree of control or potentially more precise modeling is desired, then it’s also useful to measure the minimum amount of water needed to cause a significant change in impeller power during water addition, referred to as “Xsat” in this discussion.   For high shear granulation, observation of Xsat as it varied with scale allow the generation of models that can predict how the water requirements will change with scale, and allows construction of granulation impeller blades designed to ensure similar water requirements across scale.  

  For high shear granulation, a differential equation that integrates both granulation work and the amount of water in excess of Xsat is found to be useful in providing better models and reducing the number of dimensions that need to be considered when creating a linear regression response surface model to predict granulation outcome.   It appears likely this differential equation is a good fundamental equation for modeling most granulation processes over moderate changes in water excess, and will reduce the resulting predictions to a single characteristic line that requires very few experiments to define.   The equation has been evaluated for three very different drug products of differing particle size ranges and physical properties and shown to be consistent.