Evaluation of the Moisture Prediction Capability of Near-Infrared and Attenuated Total Reflectance Fourier Transform Infrared Spectroscopy Using Superdisintegrants as Model Compounds

Abstract

The superdisintegrants (SDs) moisture content measurement by near-infrared (NIR) spectroscopy and attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy has been evaluated against thermogravimetric analysis as a reference method. SDs with varying moisture content were used to build calibration and independent model verification data sets. Calibration models were developed based on the water-specific NIR and ATR-FTIR spectral regions using partial least-square regression methods. Because of the NIR water low molar absorptivity, NIR spectroscopy handled higher moisture content (∼81%, w/w) than ATR-FTIR (∼25%, w/w). A two-way ANOVA test was performed to compare R(2) values obtained from measured and predicted moisture content (5%-25%, w/w) of SDs. No statistically significant difference was observed between the predictability of NIR and ATR-FTIR methods (p = 0.3504). However, the interactions between the two independent variables, SDs, and analytical methods were statistically significant (p = 0.0002), indicating that the predictability of the analytical method is material dependent. Thus, it would be important to recognize this highly dependent material and analytical method interaction when using NIR moisture analysis in process analytical technology to analyze and control critical quality and performance attributes of raw materials during processing with the goal of ensuring final product quality attributes.