Monitoring of a pharmaceutical blending process using near infrared chemical imaging

This study demonstrated a novel application of near-infrared chemical imaging (NIR-CI) for monitoring the blending process of Yinhuang powder. An experiment design was created including eight intermediates. The blending process was executed at different rotation velocities and ended at various time points. Baicalin (BAI) of Yinhuang powder could be spatially determined by the method of basic analysis of correlation between analytes (BACRA). However, starch (STA) and Lonicera japonica extract (LJE) could not be identified by the BACRA method due to high correlation coefficients of each other. Subsequently. characteristic wavenumbers were used to generate a RBG image, which indicated the distribution of the BAI (red), LJE (green) and STA (black). Furthermore, the homogeneity of BAI distribution in a ternary system during the blending process was measured by histogram analysis and moving block macropixel relative standard deviation (MBMRSTDEV). In the histogram analysis, the standard deviation decreased from 0.171 to 0.032 with increased blending time, which indicated that the Yinhuang powder gradually became homogeneous. However, other statistical parameters such as kurtosis and skewness were difficult to be used in understanding the blending process of Yinhuang powder. MBMRSTDEV was introduced as a suitable approach to evaluate the homogeneity. The result indicated that the blending process of the powder experienced different blending stages. The MBMRSTDEV parameter could provide an advantage in visualizing the trend of blending process. These results highlighted a promising technology to extract critical process information and provided essential process knowledge of the blending process of Yinhuang powder. (c) 2012 Elsevier B.V. All rights reserved.