Development of a sensor-based fluorescent method for quality evaluation of used frying oils

Abstract

The sensor-based fluorescent spectroscopy was developed for the rapid evaluation of oil quality at different frying stages. Manganese tetraphenylporphyrin was selected as the fluorescent sensor according to its sensitivity and selectivity to the oxidation products in frying oil. The performance of the selected fluorescent sensor was calculated based on density functional theory. The fluorescent sensor was then made to detect frying oil quality every 24 h within eighteen days. A fluorescence profile for each oil sample was obtained from the fluorescent sensor reacting with the oil sample. The feature data of fluorescence spectra were extracted using the parallel factor analysis algorithm. All oil samples were classified into three oxidation groups using principal component analysis and support vector machine algorithms. Finally, the classification algorithm of the support vector machine coupled with the genetic algorithm achieved the best classification accuracy of 96.67% and 93.33% in the training set and test set, respectively. The overall results demonstrated that this fluorescent sensor could successfully be applied in the discrimination of frying oil with different oxidation states.