Near-infrared microscopy and image analysis for pharmaceutical process control

Near-infrared microscopy (NIRM) provides low-level matrix information on the composition and distribution of ingredients within an intact pharmaceutical blend or tablet. The technique couples microscopy with spectroscopy, allowing spectral information to be obtained from small regions (~625 μm2) with sample movement provides information on a specified area. The combination of spatial and spectral information gives rise to a hyper-spectral data cube, whereby following a defined pixel though the data cube gives rise to a NIR spectrum. These cubes contain over 25,000 spectra and therefore data manipulation and image analysis are critical to producing an informative chemical image which provides a physical fingerprint of pharmaceutical solid dosage forms. To address the applicability of NIRM to pharmaceutical samples, it was necessary to evaluate the optimum data acquisition conditions, considering where spectral data actually came from. The sampling volume was explored by examination of the lateral resolution and depth of penetration of the NIR radiation. It was found for NIRM that 95% of the signal came from a minimum depth of 109 μm. Lateral resolution was found to be dependent on magnification and aperture size. Using NIRM, a wide range of pharmaceutical samples have been examined, ranging from pre-tabletting blends through to finished dosage forms. By use of multivariate and image analysis methods it has been possible to extract useful chemical information and create 'visual' images of the sample. These images provide information on the size, shape and distribution of each component within the formulation. This information has been utilized in studies of blend homogeneity, dissolution and manufacturing production issues. In each example, NIRM provides novel insight into the matrix changes induced upon processing, improving knowledge and understanding of pharmaceutical processes and potentially showing the applicability of this technology for pharmaceutical process control