Phase separation in freeze-dried amorphous solids: Detection and controlling factors

Date of Completion

January 2009


Health Sciences, Pharmacy




Lyophilization is a process often used to stabilize labile molecules in the solid-state. The process of lyophilization presents many obstacles for formulation development due to the nature of the process, which involves a freezing step, followed by sublimation of ice, and removal of unfrozen water at a higher temperature. During the freezing process, nucleation and growth of ice with simultaneous concentration of the solute in the non-ice phase occurs. The impact of the concentration effect and cold temperatures on amorphous phase behavior during the lyophilization process has not been extensively studied and is therefore not well understood. This has mainly been due to limitations in methods of detection. Current methods of detection have been mainly limited to the detection of multiple glass transitions using DSC and visual observation of the freeze-dried cake using SEM. The occurrence of amorphous phase separation in freeze-dried solids and methodology for detection is reviewed. Evaluation of a novel Raman mapping technique and a simple analysis of the data using model PVP/dextran systems revealed the ability of the technique to identify phase separation in freeze-dried samples. The results of the analysis were evaluated based on DSC, i.e. detection of multiple Tg's in the freeze-concentrate, and were found to be comparable where limitations were acknowledged for both techniques on the boundaries of phase separation. Estimates of phase compositions were determined and were found to be qualitatively in agreement between DSC and Raman. Pair distribution function analysis (PDF) of x-ray powder diffraction (XRPD) data confirmed the results observed by Raman and detected phase separation in some samples below the limit of detection of either Raman or DSC. The Raman technique was applied to model protein-stabilizer systems. Phase separation was identified in all mixtures of Ficoll/BSA and in 1:1 and 3:1 trehalose:lysozyme samples, thus demonstrating the potential of Raman to be used in screening for phase separation in protein formulations. Thermally stimulate current spectroscopy (TSC) was also evaluated as a potential tool for amorphous phase separation. TSC analysis confirmed to presence of phase separation in polymer systems with the surprising observation of largely different glass transition temperatures than DSC. Speculations of the nature of artifacts presented in TSC spectra are discussed. Lastly, in an effort to elucidate some of the parameters which control the process of phase separation during freezing, nucleation temperature and cooling rate were studied in the absence of thermal history effects and compared to a sample mimicking the product temperature profile of a freeze-dried vial. Cooling rate and nucleation temperature, alone, were found to have a minimal effect on phase separation. The largely important factor of thermal history dominated any other effects. Thus, control of the process of phase separation can be achieved through sample volume and nucleation temperature while considering the presence of a thermal gradient in a vial.^