(410g) Computational Modeling of the Residence Time Distribution of Twin Screw Extrusion for Continuous Manufacturing Applications

Twin screw extruders (TSE), where two co- (or counter-) rotating screws pump viscous material along the length of the barrel, while providing both distributive and dispersive mixing, have found many applications across multiple industries – polymer, food, pharma, etc. As an inherently continuous unit operation, TSEs can be seamlessly integrated into any continuous drug product manufacturing train to provide, e.g. high-shear wet granulation, melt granulation, amorphous solid dispersion hot-melt extrusion, etc. Optimization of the performance of the TSE relies on the control of the residence time distribution (RTD) of the material in the extruder.

In this work, the RTD of a TSE was captured by leveraging near infrared (NIR) process analytical technology (PAT) positioned at the die of the extruder. Following pulse additions of a NIR-detectable tracer to the material fed to the TSE, NIR spectra of the polymer extrudate was captured in-line and was translated to percent tracer content. In-line tracer concentration data was then used to model the RTD of the TSE using a tank-in-series model. Additionally, a plug flow time-delay component was convolved with the tank in series model to more closely match the measured distribution. Finally, the RTD was modeled for different screw speeds and feed rates to compare process attributes as a function of extruder fill level. Using the established models, we aim to predict mixing behavior in the TSE for varied total throughputs and leverage this information to engineer a control system for maintaining a uniform tracer concentration in the extrudate during scale up to a commercial level.