A procedure to optimize scale-up for the primary drying phase of lyophilization.

This article describes a procedure to facilitate scale-up for the primary drying phase of lyophilization using a combination of empirical testing and numerical modeling. Freeze dry microscopy is used to determine the temperature at which lyophile collapse occurs. A laboratory scale freeze-dryer equipped with manometric temperature measurement is utilized to characterize the formulation-dependent mass transfer resistance of the lyophile and develop an optimized laboratory scale primary drying phase of the freeze-drying cycle. Characterization of heat transfer at both lab and pilot scales has been ascertained from data collected during a lyophilization cycle involving surrogate material. Using the empirically derived mass transfer resistance and heat transfer data, a semi-empirical computational heat and mass transfer model originally developed by Mascarenhas et al. (Mascarenhas et al., 1997, Comput Methods Appl Mech Eng 148: 105-124) is demonstrated to provide predictive primary drying data at both the laboratory and pilot scale. Excellent agreement in both the sublimation interface temperature profiles and the time for completion of primary drying is obtained between the experimental cycles and the numerical model at both the laboratory and pilot scales. Further, the computational model predicts the optimum operational settings of the pilot scale lyophilizer, thus the procedure discussed here offers the potential to both reduce the time necessary to develop commercial freeze-drying cycles by eliminating experimentation and to minimize consumption of valuable pharmacologically active materials during process development.

Dehydration kinetics of prostaglandin E1 in a lipid emulsion.

The overall dehydration kinetics of prostaglandin E1 (PGE1) in a lipid emulsion at 35 degrees C were found to fit a model whereby the kapparent measured at each pH is simply the sum of the product of the fraction of the PGE1 at the interface, fi, and the rate constant at the interface, ki, plus the product of the fraction of the PGE1 in the aqueous phase, faq, and the rate constant in the aqueous phase, kaq. The values for fi and faq were reported earlier as a function of pH at 35 degrees C. The kaq and kapparent were experimentally determined as a function of pH at 35 degrees C. The ki was indirectly determined from the stability data in the emulsion. Microscopic rate constants for dehydration of PGE1 in the aqueous phase and interface at 35 degrees C were estimated from the experimental data. Based on the kinetic evaluation performed, it appears that the dehydration kinetics might be manipulated by the addition of charged surface active agents.

Determination of the pH-dependent phase distribution of prostaglandin E1 in a lipid emulsion by ultrafiltration.

The distribution of prostaglandin E1 (PGE1) in a lipid emulsion has been shown to be consistent with a three-phase model which assumes that solute may reside in the bulk aqueous and oil phases and at the oil/water interface. Calculations suggest that, in a lipid emulsion having an average particle size of 0.11 micron, it is theoretically possible for a surface active species such as PGE1 to exist predominantly at the interface. Aqueous phase concentrations of PGE1 versus pH were measured in an emulsion having an oil/water phase volume ratio of 0.1 by the use of an ultrafiltration technique in order to estimate the relative percentages of PGE1 in each phase. From bulk oil/water partition coefficient determinations, the amount of PGE1 present in the bulk oil phase of the emulsion was concluded to be insignificant. At emulsion pH values less than 5, PGE1 resides preferentially (greater than 97%) at the interface. With increasing pH's, the percentage of PGE1 in the aqueous phase increases, reaching 51% at high pH's. A model which assumes that both the nonionized and the ionized PGE1 species may be present at the interface, depending on the pH, was shown to be consistent with the data. Estimates were made of the distribution coefficients of the ionized and nonionized PGE1 between the interface and the aqueous phase and their concentration dependence. The apparent pKa of PGE1 at the interface derived from these data was 6.8. The distribution coefficients were used to generate a distribution profile of the various PGE1 species as a function of the pH. This distribution profile will be useful in explaining kinetic data of PGE1 in the emulsion as a function of pH.