(435e) Experimental Confirmation of Scale-Down Mixing Models from Commercial to Lab Scale with CFD for Sterile Drug Products

The ability to model a large, commercial-scale process at a smaller scale in the laboratory is a valuable tool to increase the efficiency of process development. Characterizing the mixing process at commercial scale can be particularly challenging with sterile products where process monitoring tools and sampling can be a risk to the integrity of the sterile boundary. Having a lab-scale mixing model predictive of the large scale operation will enable more effective small scale batches to decrease the number of commercial scale batches needed during process development, consequently saving time, materials, and other costs. Additionally, the use of computational fluid dynamics (CFD) to create a virtual model can further reduce the number of small scale batches needed as well.

A scale-down model is being developed using CFD between a commercial scale 700L tank and a development scale 10L tank. These tanks are of stainless steel construction with bottom-mounted agitators used for mixing and formulation of sterile drug products. The primary responses investigated in this liquid-liquid mixing study was time to homogeneity as impacted by varying viscosities and densities of two equal-volume constituent components, mixing speed, and tank fill level. Raman and conductivity probes were utilized to collect real-time data throughout the mixing process, providing value in characterizing liquids that mix too quickly to be conducive to offline sampling. Experiments were designed considering statistical principles in conjunction with ranges of Richardson and Reynolds numbers to interrogate the boundaries between safe operating spaces and regions of failure. These experiments are executed physically in the tank and compared with corresponding CFD simulations to validate the model.