Julien Andrieu

To optimize the primary drying step of freeze‐drying processes, it is very important to control morphological parameters of the frozen matrix. A mathematical model that simulates temperature profiles during freezing process of standard pharmaceutical formulations was set up and the ice crystal mean sizes were semi‐empirically estimated from the simulated thermal profiles. Then, water vapor mass transfer kinetics during sublimation step was estimated from ice phase morphological parameters. All these numerical data were compared with experimental data and a quite good agreement was observed, which confirmed the adequacy of the present model calculations. It was confirmed that, for a given formulation, the mass transfer parameters during freeze‐drying were strongly dependent on morphological textural parameters, and consequently, on the nucleation temperatures that fix the ice phase morphology. The influence of freezing rates was also predicted from the simulations, proving that an increase of cooling rates led to slower primary drying rates. Copyright © 2010 Curtin University of Technology and John Wiley & Sons, Ltd.

This work concerns the experimental study in syringe configuration of the sublimation period for a model formulation under operating conditions commonly used for pharmaceutical proteins freeze-drying. Different methods for the detection of the sublimation end-point were compared. This exhaustive comparison has shown that the Lyotrack sensor (Annecy, France) had the best sensitivity and presented the best signal-to-noise ratio followed by the Pirani gauge; a significant advantage of Lyotrack sensor relies on the possibility of sterilization. The scale method, which is based on continuous weighing of a single syringe or a single vial, underestimated largely the sublimation times and, consequently, was not representative at all of the whole syringe (or vial) set behavior. Consequently, caution should be taken for its use for setting up laboratory data and for scaling up these data for the optimization of commercial freeze-drying cycles. Moreover, four standard freeze-drying cycles with different values of main operating parameters (shelf temperature and total gas pressure) were investigated in order to determine overall heat and mass transfer coefficient values in syringe configuration. As expected, higher shelf temperature resulted in higher sublimation rates, but higher total gas pressures led to lower sublimation rates; nevertheless, the total gas pressure influence was less pronounced than the shelf temperature influence. The influence of syringe location in the supporting rack on drying rate values was also investigated: no front/back location influence was observed, but a small side location effect was observed, as syringes located at the edge and at the center of the rack presented slightly higher mean sublimation rates. The two main sublimation parameters, namely, the water vapor mass transfer resistance, noted R_p, and the overall heat transfer coefficient, noted K_v, were determined by the pressure rise analysis method (PRA method) by using on-line and in situ pressure rise tests. The K_v values were higher when the gas temperature or the total gas pressure increased but the R_p values remained quite constant as a function of these two parameters. A comparison between syringe and vial configuration data proved that when glass vials were used instead of syringes, the values of the sublimation rates and of the permeability of the freeze-dried matrix did not change significantly.

The influence of total gas pressure and shelf temperature on sublimation kinetics of BSA-based formulation in glass vial as geometrical configuration was determined with a pilot freeze-dryer in standard operating freeze-drying conditions. The sublimation rate curves showed three different periods with a plateau corresponding to a stationary regime. These kinetics data were mainly dependent on the shelf temperature and slightly influenced by the total gas pressure. Thus, the sublimation process in our conditions was mainly governed by overall heat transfer rate from the plate and from the surroundings to the sublimation front. Moreover, it proved that the water vapor mass transfer mechanism through the dried layer occurs by molecular diffusion in Knudsen regime. Finally, these experimental sublimation kinetics data were found in a quite fair agreement with the set up results. They confirm the validation of previous modeling of mean product temperature profiles during the freeze-drying by using the finite element code FEMLAB in real vial geometry (2-D).

A mathematical model that simulates temperature profiles during freezing process of standard pharmaceutical formulations (mannitol and BSA based) was set up in two‐dimensional axsymmetric space, and the ice crystal mean sizes were semi empirically estimated from the resulting temperature profiles. Water vapor mass transfer permeability values during sublimation step were also estimated from ice phase morphological parameters. All these numerical data were compared with experimental data, and a quite good agreement was observed that confirmed the adequacy of the present model calculations. It was confirmed that, for a given formulation, the mass transfer parameters during freeze‐drying were strongly dependent on morphological textural parameters, and consequently, on the nucleation temperatures that fix the ice phase morphology. The influence of freezing rate was also predicted from the simulations, proving that an increase of cooling rates led to slower primary drying rates. © 2007 American Institute of Chemical Engineers AIChE J, 2007