Synchronous front-face fluorescence spectroscopy for authentication of the adulteration of edible vegetable oil with refined used frying oil

• Published in: Food chemistry 2017-02, Vol.217, p.274-280
• Main Authors: Tan, Jin, Li, Rong, Jiang, Zi-Tao, Tang, Shu-Hua, Wang, Ying, Shi, Meng, Xiao, Yi-Qian, Jia, Bin, Lu, Tian-Xiang, Wang, Hao
• Format: Article
• Language: English
• Subjects: Adulteration; Edible vegetable oil; Food Contamination - analysis; Partial least squares regression; Plant Oils - chemistry; Spectrometry, Fluorescence - methods; Synchronous front-face fluorescence; Used frying oil
• ISSN: 0308-8146; 1873-7072
• DOI: 10.1016/j.foodchem.2016.08.053

•Front-face fluorescence spectroscopy was utilized for authentication of used frying oil.•Discrimination of used frying oil from fresh vegetable oil was achieved by PCA–LDA.•The using time of used frying oil was estimated by PLS.•The adulteration of fresh vegetable oil with used frying oil was accurately determined.•Both laboratory prepared and real-world used frying oil samples were investigated. Synchronous front-face fluorescence spectroscopy has been developed for the discrimination of used frying oil (UFO) from edible vegetable oil (EVO), the estimation of the using time of UFO, and the determination of the adulteration of EVO with UFO. Both the heating time of laboratory prepared UFO and the adulteration of EVO with UFO could be determined by partial least squares regression (PLSR). To simulate the EVO adulteration with UFO, for each kind of oil, fifty adulterated samples at the adulterant amounts range of 1–50% were prepared. PLSR was then adopted to build the model and both full (leave-one-out) cross-validation and external validation were performed to evaluate the predictive ability. Under the optimum condition, the plots of observed versus predicted values exhibited high linearity (R2>0.96). The root mean square error of cross-validation (RMSECV) and root mean square error of prediction (RMSEP) were both lower than 3%.