Aflatoxin B1 Detected by Terahertz Time-domain Spectroscopy

摘要

Developing physical methods to detect aflatoxin is a significant area of research in food safety. Because the current detection methods that use near-infrared spectroscopy are not sufficiently accurate, we focus here on using Terahertz Time-Domain Spectroscopy (THz-TDS) to detect aflatoxin B1 in food. We first divided the concentration of aflatoxin B1 preserved in acetonitrile into ten classifications which are 5ppb, 10ppb, 25ppb, 50ppb, 100ppb, 500ppb,1ppm and 10ppm respectively. Each classification corresponds to a certain THz spectroscopy. For the samples with unknown concentration, we respectively used Least-Squares Support Vector Machine (LS-SVM), Naive Bayes and Back-Propagation Artificial Neural Network (BP-ANN) to identify their concentrations based on THz spectroscopy. Verification results from High Performance Liquid Chromatography (HPLC) show that, compared to the other two methods, LS-SVM has more precision for identifying the concentration of alatoxin B1. The proposed method may provide another effective way for rapid detection of aflatoxin in food.