(561g) Implementation of a Dynamic Mechanistic Model Towards Optimal Design and Scale-up of an Atypical Freeze-Drying Process

Authors: 
Sen, M. - Presenter, Eli Lilly and Company
Wegiel, L., Eli Lilly and Company
Moreira, J., Process Systems Enterprise
Rajniak, P., Process Systems Enterprise
Freeze-drying is widely used by pharmaceutical companies for manufacturing materials that are otherwise sensitive to moisture or high temperatures. A freeze-drying cycle is energy intensive since it happens over several days with the primary drying step being the longest. The energy and time required to maintain the desired shelf temperature and support the vacuum for several days at a time contributes towards a significant cost factor. In an optimal freeze-drying cycle, the drying time must be designed to be energy efficient as well as meet the product critical quality attributes (CQAs).

This work illustrates process optimization of a previously designed sub-optimal cycle, following a model-based design of experiment approach to minimize the cycle time while meeting the product CQAs. This cycle optimization was unique in that the formulation contained acetic acid and no bulking agent. These atypical formulation components led to a rare freeze-drying cycle optimization opportunity.

A mechanistic, non-steady state process model with a moving boundary (sublimation interface) was used in this work [1]. The cycle time was reduced by 50% (~2.5 days). The process map obtained from the model was used to establish the proven acceptable ranges (PARs) on the critical process parameters and identify the design space. The process was then scaled from lab to pilot plant based on the identified design space. This entire process optimization workflow required less than ~20 grams of the active pharmaceutical ingredient (API) in experimentation.

References

[1] Liapis and Bruttini, 1994, A theory for the primary and secondary drying stages of the freeze-drying of pharmaceutical crystalline and amorphous solutes: comparison between experimental data and theory. Sep. Technol., 4, 144-155