基于~1H NMR指纹图谱结合多变量分析的地沟油检测方法

Identification of illegal cooking oils by using ^1H NMR fingerprintscombined with multivariate analysis

地沟油事件是中国近期发生的严重食品安全事故,而针对地沟油的检测目前尚无非常有效的方法报道.本研究对60种市售植物油和地沟油样品分别进行了核磁H谱测定,建立了12个鉴别指标,采用聚类分析技术对样品进行聚类分析,并将样品分为8大类,建立了判别函数.将未知样品的1H NMR数据代入判别函数,可检测和判别未知油的来源和品质.将该方法应用于两次盲测共69个样品,结果显示,两次盲测的正确率分别为91.9%和93.8%,可作为食用油品质检测的重要参考.

Illegal cooking oil incident is a serious food safety problem happened in China recently. There are still no efficient methods for illegal cooking oil identification. In this paper, 60 kinds of oils were detected by ^1HNMR, and 12 identification indicators were found. Based on those 12 fingerprintings, 60 kinds of oils could be divided into 8 categories by using cluster analysis. According to the 8 categories, discriminant function was established. When the unknown sample＇s 1H NMR data were put into those functions, the discriminant analysis of the unknown samples could be done to identify the origin of the unknown oil samples. This method was applied twice in blind tests, and the results showed that the correct rate of the two blind tests were 91.9% and 93.8% respectively.