基于傅里叶变换中红外光谱识别正常食用植物油和精炼潲水油模型分析

潲水油回流餐桌等食品安全问题越来越受到社会关注，探寻准确、快速、高效的潲水油鉴别新方法成为食用油安全性检测的新要求。用傅里叶变换中红外光谱技术（Fourier transform mid-infrared spectroscopy，FT-MIR）对精炼潲水油（refining hogwash oils，RHOs）和4 种不同正常食用植物油（菜籽油、大豆油、花生油和玉米油）进行快速检测，结合偏最小二乘判别法（PLS-DA）建立了RHOs和4 种不同正常食用植物油的判别模型。结果表明，在全光谱范围（4 000～450 cm–1）内，经二阶求导（Savitzky-Golay，5 点）后，RHOs和4 种不同正常食用植物油FT-MIR有显著差异。PLS-DA模型对22 个未知样品预测发现，判别模型的整体正确判别率均为100%。此结果表明FT-MIR结合化学计量学方法可以作为RHOs和4 种不同正常食用植物油（菜籽油、大豆油、花生油和玉米油）区分的一种有效技术手段。

Discrimination of Refined Hogwash Oils from Edible Vegetable Oils by FT-MIR Spectroscopy

In this study, Fourier transform-infrared spectroscopy (FT-MIR）was applied to rapidly distinguish refined
hogwash oils (RHOs) from four different edible vegetable oils, rapeseed oil, soybean oil, peanut oil and corn oil. A
multivariate statistical procedure based on cluster analysis (CA) coupled to partial least squares-discriminant analysis (PSLDA),
was elaborated, providing an effective classification method. It was shown that there were significant differences
between RHOs and different edible vegetable oils based on FT-MIR spectra after second derivative (Savitzky-Golay,
5 point) transformation in the whole wavelength range (4 000–450 cm–1). The PLS-DA procedure was then applied to
classify twenty-two unknown oil samples with a correction rate of 100%. These results demonstrate that FT-MIR combined with
chemometric analysis can be used as an effective method to discriminate RHOs from these four different edible vegetable oils.