二维近红外-中红外相关谱在掺假芝麻油判别中的应用  

Application of Two-Dimensional NIR-MIR Correlation Spectroscopy in Identifying Adulterated Sesame Oil

作　　者：王宝贺 杨仁杰[1] 杨延荣[1] 孙雪杉 刘海学[2] 张伟玉[1] 

机构地区：[1]天津农学院工程技术学院,天津300384 [2]天津农学院农业分析测试中心,天津300384

出　　处：《理化检验（化学分册）》2017年第2期134-138,共5页Physical Testing and Chemical Analysis(Part B:Chemical Analysis)

基　　金：国家自然科学基金(31201359;81471698);天津市自然科学基金(14JCYBJC30400);国家大学生创新训练项目(201510061055)

摘　　要：提出了基于二维近红外-中红外相关谱判别掺假芝麻油的方法。分别配制了40个纯芝麻油样品和40个掺假芝麻油(掺入的玉米油的体积分数在3%~60%之间)样品,并采集了所有样品的近红外光谱和中红外光谱。在4 540~6 000cm^(-1)对650~1 800cm^(-1)内进行同步二维近红外-中红外相关谱计算,建立了掺假芝麻油的多维偏最小二乘判别模型,并将其预测性能与二维近红外相关谱和二维中红外相关谱判别模型的预测性能进行了比较。结果表明:上述3个模型对预测集未知样品的判别正确率分别为96.3%,92.6%,96.3%。A method for identifying adulterated sesame oil using two-dimensional NIRMIR correlation spectroscopy was proposed. 40 pure sesame oil samples and 40 adulterated sesame oil samples （with volume fraction of adulterating corn oil varying from 3% to 60%）were prepared. And NIR and MIR spectra of all these samples were collected. The synchronous 2D NIR-MIR （4540-6 000 cm ^-1 vs. 650-1 800 cm^-1） correlation spectra of these samples were calculated to construct N-way partial least squares discriminant analysis （NPLS-DA） model to identify adulterated sesame oil. The NPLS-DA models based on normalized synchronous 2D NIR and 2D MIR correlation spectra, were also constructed and their performance of discrimination to identify adulterated sesame oil were compared. It was found that the rates of accuracy of discrimination of the above mentioned 3 NPLS-DA models for the prediction set were 96. 3%, 92.6%, 96.3%, respectively.

关 键 词：二维近红外-中红外相关谱 多维偏最小二乘判别 掺假芝麻油 

分 类 号：O657.33[理学—分析化学]