Chemometrics applied to quantitative analysis of ternary mixtures by terahertz spectroscopy.

Chemometrics was applied to qualitative and quantitative analyses of terahertz spectra obtained in transmission mode. A series of mixtures of three pure analytes, namely, citric acid, D-(-)fructose, and α-lactose monohydrate under various concentrations, was prepared as pressed pellets with polyethylene as binder. Then, terahertz absorbance spectra were recorded by terahertz time domain spectroscopy and analyzed. First, principal component analysis allowed one to correctly locate the samples into a ternary diagram. Second, quantitative analysis was achieved by partial least-squares (PLS) regression and artificial neural networks (ANN). The concentrations were predicted with values of relative mean square error lower than 0.9% for the three constituents. As a conclusion, chemometrics was demonstrated to be very efficient for the analysis of the ternary mixtures prepared for this study.