(M1430-08-43) Impact of Freezing Rate on Lyophilization Primary Drying Efficiency and Porous Structure: Comparison of Three Measurement Techniques

Purpose: The morphology of the ice in the frozen product influences primary drying efficiency during freeze-drying. In the absence of collapse, the ice morphology, influenced by freezing rate, is controlled by ice nucleation temperature and post nucleation cooling rate. As ice undergoes sublimation, it leaves behind a porous structure. Currently, the specific surface area (SSA) measured by gas adsorption (GA) is used as a surrogate measure of the average pore size in the lyophilized cake. Measurement of pore size and size distribution by low pressure mercury porosimetry (LPMIP) is a less commonly used method. Characterization of pore features by x-ray micro computer-aided tomography (XRμCT) is an emerging method of understanding the internal dry cake structure. The goal of this project was to determine freezing rate effects on pore features characterized by three measurements and dry product resistance to sublimation, with a long-term goal of designing more efficient lyophilization cycles for manufacturing.

Methods: Sucrose (5% w/v solution) was prepared in 2ml cut serum vials with 7 mm fill depth. Four freezing protocols with freezing rates from fastest to slowest were achieved by combing two ice nucleation temperatures (−10°C and −5°C) and two post nucleation cooling rates (5°C/min and 0.2°C/min) in a freeze-dryer (MicroFD® ,Millrock Technology). Primary drying shelf temperature of −35°C, −33°C, −30°C and −25°C were employed to achieve comparable product temperatures (39.0±0.5°C) subsequent to freezing rates ranging from fastest to slowest. LyoSim® blocks were used to reduce the edge effect. Specific surface area (SSA) of freeze-dried cake was determined by krypton adsorption. Pore size and size distribution were measured using low pressure mercury porosimetry (LPMIP). A 3D non-destructive image of pore structure in the cake was acquired by XRμCT. Dry product resistance, Rp, was calculated based on the thermocouple data in primary drying.

Results: As shown in Table 1, using primary drying conditions that yield the same product temperature, the drying time was reduced from 43 hrs (using the fastest freezing protocol) to 14 hrs (using the slowest freezing protocol). At each ice nucleation temperature, faster post nucleation cooling rate (5°C/min) cycle produced cakes with higher SSA compared to slower post nucleation cooling rate (0.2°C/min). At each post nucleation cooling rate, ice nucleation temperature did not influence SSA. Median pore diameters by LPMIP revealed a narrow pore size distribution for cakes produced by controlling ice nucleation temperature at −10°C and wider for −5°C. XRμCT images from the center of cakes produced from the four freezing protocols showed no meaningful difference in median pore diameter. A superior correlation was showed between the slope of the dry product resistance and the reciprocal of medium pore size from LPMIP (Figure 1) than either SSA or XRμCT.

Conclusion: Fast freezing produced lyophilized cakes with fine pore structure (measured by LPMIP), producing high resistance to sublimation and slow freeze-drying. The results from SSA and XRμCT did not show a strong correlation with dry product resistance. The strong correlation of dry product resistance with the reciprocal of median pore size (by LPMIP) agrees well with the theoretical model¹ of pore resistance to vapor flow at low pressure.

References: Pikal, M.J., Shah, S., Senior, D. and Lang, J.E., 1983. Physical chemistry of freeze-drying: measurement of sublimation rates for frozen aqueous solutions by a microbalance technique. Journal of pharmaceutical sciences, 72(6), pp.635-650.

Acknowledgements:Funding from the Dane Kildsig Center for Pharmaceutical Processing Research (CPPR) is greatly appreciated. The authors also thank Millrock Technology for the use of the research freeze-dryer (MicroFD).


Table1: Characterization of pore structures from lyophilized cakes and primary drying results from four freezing protocols


Figure 1: Linear correlation between product resistance (Rp) slope and 1/median pore diameter by LPMIP