(337a) Optimization of Continuous Spin Freezing in Single Vial Unit By Implementing Computational Fluid Dynamics Modeling and Simulation

Freeze-drying is a common method for increasing the stability and shelf life of heat-sensitive and water-labile (bio)pharmaceuticals. The pharmaceutical sector is currently transitioning from batch to continuous manufacturing. According to the recent FDA guideline regarding continuous manufacturing of drug substances and drug products (Q13), maintaining a level of control in CM necessitates an understanding of process dynamics. Three consecutive phases are involved in the pharmaceutical freeze-drying process: freezing, primary and secondary drying. Using cold sterile gas and rapid rotation of glass vials along their longitudinal axis, the temperature of the aqueous solution is lowered during the spin freezing stage until ice starts to nucleate, followed by ice growth. The gas flow and temperature are controlled during the spin freezing process to manage the ice crystal size distribution. During the primary drying, the chamber pressure is lowered considerably below the vapor pressure of the ice. Further, an infrared (IR) heater supplies energy for sublimation during the primary drying process. Desorption of the unfrozen water takes place during secondary drying. Being the first step in the dehydration process, the freezing phase in the lyophilization process is crucial as it determines the ice shape, the pore size of the final dried product and hence the drying conditions. To understand the dynamics during the spin freezing step as part of a recently developed continuous freeze-drying process, for the first time, a detailed model based on computational fluid dynamics (CFD) is developed (CFX v.19.2, Ansys Inc.) for the single vial unit (SVU) freeze dryer. The CFD simulation provides the vial temperature and a local gas flow pattern inside the cooling chamber of the investigated system.

The rotational speed of the vial during spin freezing and the velocity of the cooling gas which both - next to the gas temperature - determine the cooling rate of the product varies from 2000 rpm to 5000 rpm and 2 to 20 m/s, respectively. Multiple reference frame (MRF) and moving mesh modeling methodologies are being studied for simulating the gas flow inside the chamber. To simulate a steady-state flow, MRF is employed. The MRF simulation results are utilized as the initial condition for the moving mesh technique to model flow in the transient state. Further, the heat equation is solved to simulate the vial temperature. The temperature along the vial is measured using the Thermal IR camera (Fig. 1). The experimental data support the temperature profile obtained through simulation. The validity of the gas flow behavior in the chamber was analyzed using the tracer simulation. Smoke test is used to confirm the tracer simulation. The smoke flow in the chamber is recorded using a high speed camera. The findings of the simulation and the experiments are in good agreement.

Finally, the design will be further optimized using the verified and validated CFD simulation. To optimize the cooling rate, several geometric factors such as nozzle size and placement, gas outlet location, and dimension will be studied. Thus, detailed mechanistic modeling improves understanding of process dynamics and aids in the optimization of the spin freezing set up.