近红外光谱法快速检测饮料中的食用合成色素

以光谱扫描范围为1?100～1?650?nm，分辨率为0.2?nm，通过对不同质量浓度的日落黄、柠檬黄、胭脂红和苋菜红溶液扫描，得到的原始光谱图经过一阶导数、二阶导数等方法进行预处理。然后通过修正的最小二乘（modified partial least square，MPLS）法结合变量标准化且去散射处理的光谱预处理手段建立相关的MPLS模型。最后利用所建模型，对饮料中的食用合成色素含量进行预测。预测相关系数最高可达0.991，预测标准差最低为0.009%，相对分析误差最大为3.51。结果表明，近红外光谱可作为一种简单、快速、无损、可靠的方法用于检测饮料中的食用合成色素。

Rapid Determination of Synthetic Pigments in Fruit-Flavored Beverage by Near Infrared Spectroscopy

A rapid method based on near infrared (NIR) spectroscopy was developed for the determination of synthetic pigments in fruit-flavored beverage. NIR spectra of sunset yellow, tartrazine, ponceau 4R and amaranth solutions at different concentrations were recorded in the wavenumber range of 1 100–1 650 nm at a 0.2 nm interval, respectively and preprocessed by first derivative and second derivative. The modified partial least square (MPLS) regression models were established after spectral preprocessing by standard normal variate (SNV) + detrending and applied to predict the synthetic pigment contents in beverage. The results showed that the maximum correlation coefficient of prediction (Rp) was 0.991, the minimum standard error of prediction (SEP) was 0.009%, and the maximum standard deviation ratio (SDR) was 3.51. Accordingly, NIR spectroscopy can be used as a simple, rapid, nondestructive and reliable method to determine the contents of synthetic pigments in beverage.