Prediction of onset of crystallization in amorphous pharmaceutical systems: phenobarbital, nifedipine/PVP, and phenobarbital/PVP.

The aim of this work is to determine if a stability testing protocol based on the correlations between crystallization onset and relaxation time above the glass transition temperature (T(g)) can be used to predict the crystallization onsets in amorphous pharmaceutical systems well below their T(g). This procedure assumes that the coupling between crystallization onset and molecular mobility is the same above and below T(g). The stability testing protocol has been applied to phenobarbital, phenobarbital/polyvinylpyrrolidone (PVP) (95/5, w/w), and nifedipine/PVP (95/5, w/w). Crystallization onsets have been detected by polarized light microscopy examination of amorphous films; molecular mobility has been determined by dielectric relaxation spectroscopy above T(g) and by both isothermal calorimetry and modulated differential scanning calorimetry below T(g). We find that small amounts of PVP significantly retard re-crystallization. This dramatic effect of PVP is not related to mobility, so this approach applies, at best, to extrapolation of high temperature data on a given formulation to low temperatures. Variation in molecular mobility at these concentrations of PVP is not the dominant factor in determining variation in propensity for re-crystallization from glassy systems; we suggest surface interactions between PVP and nuclei and/or small crystals slowing growth control variation in crystallization kinetics between formulations.