Quantitative detection of buffalo milk adulteration with cow milk using Fourier transform near infrared spectroscopy

A near infrared (NIR) spectroscopy model was used to quantitatively detect buffalo milk adulteration with cow milk. Pasteurized buffalo milk samples were purchased from a dairy farm and from a local supermarket. Adulterated milk samples were prepared with ratio of cow milk to buffalo milk at 9 levels of 10:90, 20:80, 30:70, 40:60, 50:50, 60:40, 70:30, 80:20 and 90:10 wt%. Spectra of pure buffalo milk, pure cow milk and adulterated milk samples were recorded by a Fourier transform NIR spectrometer in the wavenumber range of 12500-4000 cm^-1 with resolution of 8 cm^-1. A NIR spectroscopy quantitative model was developed with partial least square (PLS) regression. The NIR spectroscopy model showed ability to detect adulterated milk as follows: Rval² = 0.998, RMSEP = 2.121 wt%, Bias =-0.396 wt% and RPD = 18.1. NIR spectroscopy coupled with PLS algorithm was shown to be an alternative technique to detect buffalo milk adulteration with cow milk in the global dairy industry.