Rapid Detection of Milk Fat Adulteration in Yoghurts Using Near and Mid-Infrared Spectroscopy

Abstract

Both Fourier-transform near-infrared (FT-NIR) and mid-infrared (FT-MIR) spectroscopy with chemometrics were used for the fast detection of milk fat adulteration in yoghurts without sample preparation. Soft independent modelling of class analogy models of both NIR and MIR spectra showed successful detection of milk fat adulteration and identification of the type of adulterant oils. Partial least squares regression models of a representative adulterant yielded high correlation coefficients (above 0.98), low standard error of prediction (lower than 7.12%) and high residual predictive deviation values (above 4.35) for both NIR and MIR spectra. Additionally, regardless of the source of adulterant oils used, separate NIR and MIR PLSR models were developed for quantification of milk fat ratio (%) in yoghurts as an alternative approach to measuring the level of adulterant oil (%); NIR spectroscopy was found to be superior in these models. (C) 2020 Elsevier Ltd. All rights reserved.