(361e) A Multiscale Model for the Granulation of Powders in a Spheronizer

The spheronizer is a rotor-based equipment used in the pharmaceutical industry as a mixing and pelletizing process to obtain uniformly sized spherical pellets from a mixture of powder extrudates, the form of which is cylindrical. It consists of a rotating disk that spins at high speed at the bottom of a cylindrical bowl, thereby increasing friction with the cylindrical particles. Powder flow in a spheronizer is characterized by a toroidal motion that promotes the roundness of the pellets and increases their density. One important issue related to this process is the likely occurrence of segregation within the bed of particles, which can lead to a lack of uniformity of the pellets and, subsequently, to a poor quality of the end-use product. Granulation is a process whereby primary particles mix and adhere to one another to form larger entities called granules. A standard spheronizer can be easily transformed into a granulator by adding to it an atomizer that sprays water droplets to the surface of the bed of particles. Despite the fact that the mechanisms governing the formation of these granules are very complex, their understanding is crucial to the development of efficient granulation processes yielding granules with the desired properties. Granulation processes can be modeled with the help of population balance models. Along this line, numerous studies have considered mechanisms such as coalescence or breakage to predict the evolution of granulation with respect to time. Although these mechanisms play a fundamental role in representing the real behavior of such systems, it has become evident that particle flow also needs to be modeled. This presentation will focus on the development of a multiscale model involving a population balance model at the process scale, a compartmental approach that takes into account specific flow behaviors in different zones of the equipment, and a discrete element based model at the particle scale. After introducing briefly this model, simulation results will be presented and compared to experimental data obtained in our pilot-scale rig.