Rapid detection of pure coconut oil adulteration with fried coconut oil using ATR-FTIR spectroscopy coupled with multivariate regression modelling

Abstract

Attenuated total reflection- Fourier transform infrared (ATR-FTIR) spectroscopy along with multivariate regression modelling was utilized to develop the methodology for classification and quantification of refined, bleached and deodorized (RBD) pure coconut oil (PCO) from its adulterant RBD fried coconut oil (FCO). Principal component analysis (PCA) was applied on 3000-2800 cm−1, and 1800-500 cm−1 and linear discriminant analysis (LDA) was used on selected 13 wavenumbers. Principal components regression (PCR) and Partial least squares regression (PLS-R) models were constructed and compared for normal, 1st, and 2nd derivatives. PLS-R model for the 1st derivative of 1800-500 cm−1 showed the best results for prediction with high precision and accuracy (RPD: 20.03, RE %: 4.288). The lowest limit of detection of FCO in PCO was predicted as 0.5% v/v. This detected concentration is not the constraint of the experiment pursued; instead, it is the lowest concentration of FCO adulteration used in our work. Our findings are only applicable to detection of FCO adulteration at same frying degree while adulteration detection of FCO at different frying degrees is altogether a different adulteration issue. This study provides valuable information to oil industries and the regulatory authorities to build standard guidelines for the detection of oil adulteration.