Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Freeze-dried biopharmaceutical formulations represent an important share of the current biopharmaceutical market. However, the process is costly and time-consuming thus the need for optimization is high. The process essentially consists of three main stages: freezing, primary drying and secondary drying. During freezing, the solution is cooled to below its glass transition temperature resulting in a mix of glassy, amorphous components and ice crystals. Primary drying starts with lowering the pressure and increasing the temperature until ice sublimates to vapour while a dried product remains. Trapped water enclosed in the cake structure is removed during secondary drying when the temperature is further increased. Freezing has long been considered as a trivial part of the lyophilization process, but it is recently considered inherently linked to primary drying. This can be attributed to the size of ice crystals. Bigger crystals will leave larger pores and facilitate sublimation and vice versa. The freezing profile determines the size of the crystals and according to Kasper et al (2011) a fast cooling rate in traditional shelf ramp freezing will result in larger ice crystals, while a slower cooling rate leads to smaller ones. Moreover, a fast cooling rate should facilitate sublimation of ice during primary drying and consequently reduce the drying time and limit process costs. However, freezing is a variable process due to supercooling, nucleation and crystal growth. To confirm the hypothesis from Kasper et al (2011), the dried product mass transfer resistance (Rp) of a certain lyophilization cycle can be analysed in function of the height of the dried product layer (Ldried). In this work, Rp versus Ldried profiles are obtained using previously established process modelling techniques based on the product temperature during the lyophilization process. Consequently, multiple full-length lyophilization cycles were carried out with different freezing profiles and different antibody concentrations to determine their influence on the Rp versus Ldried profile. The cooling rates used in this thesis ranged from 0,1°C/min to 1°C/min (with and without annealing) and extra experiments were conducted with a shelf that was pre-cooled at -40°C prior to loading the vials. The different antibody concentrations used were 10 mg/mL, 20 mg/mL, 50 mg/mL and 100 mg/mL. Visual and microscopic characterisation of the cakes showed that an antibody concentration of 20 mg/mL or lower lead to shrunken cakes and microscopic collapse. Adjustment of the sublimation parameters to less severe conditions did not avoid the micro collapse. All cakes of 50 mg/mL or higher did not show shrinkage and the electron microscope pictures were considered successful. Additionally, it was observed that the concentration influences the Rp vs Ldried profile proportionally and higher concentrations therefore lead to higher Rp profiles and an increased drying time. Determination of the influence of the freezing profile was not successful. Either no differences were observed or the results were contradictory to prior research. It is therefore hypothesized that the inherent variability of freezing, the lyophilization process or the process modelling used in this thesis did not allow to draw decisive conclusions. Consequently, further research regarding cooling rate should focus on controlling the freezing profile by controlling supercooling and nucleation temperature.