Multivariate calibration of the degree of crystallinity in intact pellets by X-ray powder diffraction

XRPD is the method of choice to determine crystalline content in an amorphous environment. While several studies describe its use on powders, little information is available on its performance on finished products. The method’s use may be limited not only by the need of sample pretreatment and its validation but also by the propensity of some materials to recrystallize when exposed to heat or mechanical stress. In this work the authors describe an attempt at constructing a model based on the XRPD measurement of intact layered pellets using univariate methods based on peak heights and PLS regression. Results indicate that neither the goodness-of-fit (below 0.9 for all tested variables), nor the RMSEC values (above 5 for all tested variables) of any model based on peak height were good enough to consider them for everyday use. PLS regression however provided a model with improved characteristics (R2 = 0.9581, RMSEC = 3.04) despite the low API content and individual loading characteristics also reflected the validity of the model. PLS analysis also indicated that a specific sample may be different in some formulation characteristic that did not register on other examinations. This further indicates the method’s usefulness in the analysis of intact dosage forms