Assessment of Chemometric Methods for the Non-Invasive Monitoring of Solid Blending Processes Using Wireless near Infrared Spectroscopy

Authors:

Abstract
The US Food and Drug Administration's process analytical technology (PAT) initiative has encouraged the pharmaceutical industry to strengthen research with a view to developing new analytical technologies for improving blending processes. Near infrared (NIR) spectroscopy, which is sensitive to both physical and chemical attributes of substances, has proved a powerful non-invasive analytical technique for this purpose. In this paper, we propose a criterion to assess blending uniformity and accurately detect the end-point of a blending process. The criterion was established by using partial least squares (PLS) regression and calibration-free methods including moving block standard deviation of spectra, principal component scores distance analysis and the mixing index ( M). The mixing index was found to provide accurate, robust results with a view to assessing blending uniformity, as was PLS for detecting blending end-points. This novel criterion allowed us to assess the homogeneity of a mixture, without the need of a reference method, a golden batch or a calibration set. In fact, the M index was found to be an easy method and it is a highly suitable tool for PAT as it allows the accurate, precise real-time monitoring of blending processes with minimal intervention of the analyst.

This publication has 18 references indexed in Scilit: