(743e) Scale up of Heat Transfer for Dry Granular Material in a Bladed Mixer

Drying is an important manufacturing step in the production of active pharmaceutical ingredients (APIs) following the crystallization and filtration of the material. Several approaches exist to carry out the drying procedure, but a common method used by the pharmaceutical industry is agitated drying. During the process, a wet bed of API is heated in a jacketed cylindrical vessel while being mixed by a rotating impeller until the moisture content is reduced to a desired level. The complexity of the process stems from the fact that heat transfer, mass transfer, and changes in physicochemical properties can occur simultaneously throughout drying. Complications often plague the procedure, including issues such as lengthy drying times, over-drying, nonuniform drying, agglomeration, attrition, and form changes. These circumstances make agitated drying a complicated process to understand and control. When considering scale up, these challenges are coupled with the difficulties typically associated with transferring knowledge from lab scale to pilot or manufacturing scale. As a result, it can be difficult to design an appropriate drying protocol that optimizes heat transfer and can be translated from scale to scale while minimizing the risk for adverse conditions.

In this work, we decouple the problem and focus on studying the heat transfer aspect of agitated drying using dry glass beads. More specifically, we investigate the influence of bed fill height and geometric scaling on heat transfer by studying both the average and standard deviation of the bed temperature. Our approach consists of employing a combination of experiments and discrete element method (DEM) modeling techniques to analyze how heat transfer scales with material fill level and equipment sizing. We carry out a design of experiments and compute heating times as well as heat transfer coefficients for the different conditions. We find that fill height and geometric scaling influence the flow and compressibility of the bed and therefore creates a balance between conduction and granular convection as the dominant mode of heat transfer. We leverage the discrete nature of DEM modeling to evaluate features such as the coordination number and normal overlap between particles. We find that the relationship between fill height, geometric scaling, and heating time is complex and depends on the geometry area available for heat transfer and the particle contacts. Finally, we relate simulation findings to experimental results for validation. This work provides better fundamental insights into how heat transfer scales in a bladed mixer and could greatly aid in the development of drying protocols.