基于改进标准加入法的碳酸饮料中多种色素含量的同时测定

Determination Multiple Food Colorants in Carbonate Beverages by Generalized Standard Addition Method

采用正交设计设计法设计包含柠檬黄、日落黄、诱惑红、苋菜红、胭脂红及亮蓝6种色素水性混合样品，分别测试其紫外-可见光谱（UV）并用支持向量回归（SVR）得到6种色素同时测定的UV-SVR模型．饮料样品经超声去除CO2后分别添加7水平不同含量的色素标准品并分别测定其UV光谱，利用UV-SVR模型预测其中色素含量，线性回归分析后得到多种色素在饮料中的含量．结果表明，基于UV-SVR-标准加入法多种色素含量同时测定，6种色素含量预测值与实际值的回归分析相关系数在0．99以上，相对标准偏差（RSD）为0．5％～2．5％；用于实际饮料样品测定时，能够同时预测其中多种色素的含量，测定结果与高效液相色谱（HPLC）法相比绝对误差小于1mg·L^-1．

84 mixed sample solutions that contain 6 kinds of food colorant, lemon yellow, sunset yellow, red temptation, amaranth, carmine and brilliant blue were prepared using orthogonal design method. The UV-Vis spectra （UV） of these pre- pared samples were measured from 300--800 nm, and support vector regression （SVR） were used to build UV-SVR model for simultaneous estimation of the 6 kinds of food colorants. The UV spectra of the beverage samples that were added 7 different level food colorants after ultrasonic removing of COz were measured, the predicted contents of the colorants in the spiked bever- age samples were obtained using the previously built UV-SVR model, the predicted values were linearly regressed with the spiked contents of the colorants, and the contents of the food colorants in raw beverage samples can be estimated as the predic- ted values when the spiked contents are zeros. The results show that the simultaneous determination of contents of multiple col- orants based on UV-SVR-generalized standard addition method （GSAM）, the correlation coefficient of regression analysis of the estimated values and the actual values of are higher than 0.99, the relative standard deviations （RSDs） are located with in the range 0.5 % -- 2.5 %. The proposed UV-SVR-GSAM approach can be used for determination of multiple colorants in beverages with the absolute errors less than 1 mg . L^-1 compared to that of high performance liquid chromatography （HPLC）.