Use of NIR hyperspectral imaging and multivariate curve resolution (MCR) for detection and quantification of adulterants in milk powder

Abstract

In this work, a methodology was proposed for detection and quantification of milk powder adulterants based on the combination of near infrared hyperspectral imaging and multivariate curve resolution method (MCR). No priori information about the adulterant present in the sample was necessary and only five sample calibrations were used for development of the calibration model. Mixtures of milk powder with only one adulterant (whey powder, starch, urea and melamine) were studied in concentrations ranged from 5 to 30% (w/w). For melamine, lower concentrations (up to 0.05%) were tested to evaluate de detection ability of the proposed methodology. Also, mixtures of milk powder with 2 (starch/urea) and 3 adulterants (starch/urea/whey powder) were studied in concentrations in the range of 1–10% (w/w). MCR was able to recover the adulterant spectra providing its identification and quantification with absolute errors lower than 5 percentage points.