Quantitative analysis of film coating in a fluidized bed process by in line NIR spectrometry and multivariate batch calibration

A method is described which enables real-time analysis of film coating on p harmaceutical pellets during an industrial manufacturing process. Measureme nts were conducted on the solid particulate material by near-infrared (Nm) spectrometry utilizing a diffuse reflectance fiber-optic probe positioned i nside a fluidized bed process vessel. Time series of NIR spectra from 11 ba tches generated a three-way data matrix that was unfolded and modeled by pa rtial least squares (PLS) in a multivariate batch calibration. The process conditions were deliberately varied according to an experimental design, Th is yielded good predictability of the coating thickness with a best model f it, R-2 = 0.97, for one PLS-projection, and a root-mean-square error of cal ibration = 2.2 mu m (range tested 0-50 mu m). The regression vector was sho wn to be highly influenced by responses that are both direct (aliphatic C-H stretch overtones) and indirect (aromatic C-H stretch overtones), from fil m component and core material, respectively. The impact of different data p retreatment methods on the normalization of the regression vector is report ed. Justifcation of the process calibration approach is emphasized by good correlation between values predicted from NIR data and reference image anal ysis data on dissected pellets and a theoretical nonlinear coating thicknes s growth model. General aspects of in-line NIR on solids and multivariate b atch calibration are discussed.