Detection and Quantification of Adulterants in Roasted and Ground Coffee by NIR Hyperspectral Imaging and Multivariate Curve Resolution

In this work, a methodology was proposed for detection and quantification of the four main roasted and ground coffee adulterants: coffee husks, roasted and powdered corn kernels, wood sticks, and soil, based on the combination of near-infrared hyperspectral imaging and multivariate curve resolution (MCR). By using this procedure, it was possible to detect the adulterants and their confirmation was accomplished comparing the reference spectra of each adulterant with the spectra recovered by the MCR model. The methodology was suitable for the quantification of adulterants in the mixtures which ranged from 1 to 40% (w/w), with errors lower than 4%. Also, a chart control card based on the scores of the MCR was developed, in order to detect the adulterated samples in a qualitative way, without the need of concentration values.