Laser induced fluorescence spectroscopy for detection of Aflatoxin B1 contamination in peanut oil

Abstract

The Aflatoxin B1 (AFB1) contamination in peanut oil is harmful and widespread. In this study, a laser induced fluorescence spectroscopy (LIF) system was developed for rapid and non-invasive screening of four varieties of peanut oils contaminated with different levels of AFB1. Discriminant analysis was conducted to classify the four varieties of peanut oils into acceptable (AFB1 < 20 μg kg−1) and contaminated (AFB1 ≥ 20 μg kg−1) levels. The results of single-step modelling, which directly classify the whole 540 peanut oils into acceptable and contaminated levels, could achieve an accuracy of 95% and 93% for calibration and validation sets respectively. Considering the fluorescence spectra could easily be influenced by the difference of physicochemical properties of different varieties of peanut oils, two-steps modelling method was conducted. In the first step, the total 540 peanut oils were classified into four varieties with all accuracy of 100% based on 600–800 nm, by using LDA algorithm. In the second step, the AFB1 contamination levels of each variety were discriminated. The classification accuracy was increased by at least 5% by conducting two-steps modelling. The results of quantitative analysis showed that for single variety, PLSR based on 400–600 nm could achieve accurate quantitative prediction. The feasibility of LIF technique for the rapid and nondestructive detection of AFB1 contamination in different varieties of edible oils was proved in this study. Further research will focus on the detection of naturally infected oils.