基于FTIR技术快速测定食用油中叔丁基对苯醌含量的研究

Rapid Determination of 2-Tert-Butyl-1,4-Benzoquinone in Edible Oils Based on FTIR Technology

研究傅里叶红外光谱技术(FTIR)快速测定食用油中叔丁基对苯醌(TBBQ)含量的分析方法,系统地比较了FTIR结合朗伯比尔定律法及FTIR结合偏最小二乘法(PLS)的建模效果。结果表明,原始波谱经差谱和二阶导数预处理后在1656 cm^(-1)处采用朗伯比尔定律法建模,所得的线性回归方程为TBBQ(μg/mL)=1089.5Abs-11.35,相关系数R^(2)为0.9981,15组煎炸油和10组市售食用油的外部验证实验的平均相对偏差分别仅为4.89%和6.80%。采用FTIR结合PLS法对原始波谱建模,建模区间为1675~1645 cm^(-1),模型相关系数R^(2)为0.9956,校正均方差为0.57,预测均方差为3.32,15组煎炸油和10组市售食用油的外部验证实验的平均相对偏差分别为8.53%和8.40%,说明基于FTIR的朗伯比尔定律法和PLS法均能有效测定食用油中TBBQ含量,其中朗伯比尔定律法预测准确度较高、模型稳定性更好。

In order to rapidly analyzed the 2-tert-butyl-1,4-benzoquinone(TBBQ)content in edible oils,FTIR technology is studied in this paper,the modeling effects of FTIR combined with Lambert-Beer's Law method and FTIR combined with partial least squares(PLS)were systematically compared.The results showed that the original spectrum was pretreated by difference spectrum and second derivative,then modeled by Lambert Beer's law method at 1656 cm^(-1),the linear regression equation is TBBQ(μg/mL)=1089.5 Abs-11.35,the correlation coefficient R^(2) is 0.9981,and the mean relative deviation of the 15 groups of frying oil and 10 groups of commercial edible oil of external validation tests was only 4.89%and 6.80%,respectively.In addition,the original spectrum was modeled by FTIR combined with PLS,and the modeling interval was 1675~1645 cm^(-1),the correlation coefficient of the resulting model was 0.9956,the corrected mean square error was 0.57,the predicted mean square error was 3.32,and the average relative deviation of 15 groups of frying oil and 10 groups of commercial edible oil of external validation tests was 8.53%and 8.40%,respectively.It indicates that FTIR combined with Lambert Law method and FTIR-PLS method can effectively determine TBBQ content in frying oil,and FTIR combined with Lambert law method has higher prediction accuracy and better model stability.