Application of near-infrared spectroscopy for monitoring the formulation process of low-dose tablets

Abstract

To study the feasibility of near-infrared (NIR) spectroscopy on low-dose tablets, chemometric models are developed to analyze the content uniformity, tablet moisture, tablet hardness, compression force, mean particle size, and particle size distribution by using the same drug. The correlation coefficients (R) of content uniformity, tablet moisture, tablet hardness, compression force, and mean particle size were obtained with good results of 0.998, 0.997, 0.983, 0.972, and 0.999, respectively. The root mean square error of prediction and low root mean square error of calibration values of these five models and additional seven models of particle size distribution were all low and close to the root mean square error of cross validation. Both validation and reliability evaluation on the content model were conducted to show the accuracy and reliability for the determination of an active content with a range from 70% to 130% of the nominal tablets. This research shows that NIR spectroscopy combined with a chemometric technique, i.e., partial least squares, is a viable tool for the quality determination of low-dose tablets in laboratory-scale samples.