拉曼光谱-聚类分析法快速鉴别掺伪花生油

Rapid detection of adulterated peanut oil by Raman spectrum-cluster analysis

中文关键词:  掺伪花生油  拉曼光谱  聚类分析  快速  鉴别

英文关键词:adulterated peanut oil  Raman spectrum  cluster analysis  rapid  identify

中文摘要:

      目的 建立快速鉴别掺伪花生油的拉曼光谱-聚类分析方法。方法 以不同产地、不同品牌的多批次花生油、大豆油、玉米油、菜籽油、葵花籽油、精炼棕榈油、精炼棉籽油及精炼地沟油为样品, 在780 nm和532 nm激光光源下, 扫描和比较其普通、扩展及导数拉曼光谱的形态。结果 在532 nm激光光源的扩展光谱及一阶导数光谱中, 花生油与低价植物油及精炼地沟油光谱的信息量最大, 样品间光谱形态的差异显著, 谱峰得到有效分离。基于此全波段光谱信息和形态建立的多步聚类分析模型及鉴别程序对36份不同花生油、105份不同低价植物油、30份仿冒花生油和38份不同精炼地沟油的判别正确率均为100%, 对180份 5%及以上的掺假花生油的判别正确率达86%以上, 对75份5%及以上的掺杂花生油的判别正确率为92%, 对72份5%及以上的掺杂植物油的判别正确率达92%以上。样品测量时无需制备样品及消耗化学试剂, 测量和分析一份样品仅耗时5 min左右。结论 所建立的拉曼光谱-聚类分析模型既可准确鉴定花生油, 还可准确鉴定各种类型的掺伪花生油, 可实现对掺伪花生油的快速、无损和准确鉴别。

英文摘要:

      Objective To establish a Raman spectrum-cluster analysis for rapid detection of adulterated peanut oil. Methods The shapes of ordinary, extended and the 1st derivative Raman spectra were scanned and compared, based on the multiple batches of peanut oil, soybean oil, corn oil, rape seed oil, sunflower seed oil, refined palm oil, refined cotton seed oil and refined bio-waste oil with various producing locations and brands, under both 780 nm and 532 nm laser sources. Results The extended and the 1st derivative Raman spectra under 532 nm showed the most abundant spectral information and significantly distinct patterns among peanut oil, low-price vegetable oil and bio-waste oil. A multi-step model and identification program of Raman spectrum--cluster analysis was developed on full-range spectra information and patterns. The discriminate rate was 100% for 36 peanut oils, 105 low-price vegetable oils, 30 counterfeit peanut oils and 38 refined bio-waste oils. The discriminate rates were 86%, 92% and 92%, respectively, for 180 peanut oils mixed with at least 5% low-price vegetable oil, 75 peanut oils mixed with at least 5% bio-waste oil, and 72 vegetable oils mixed with at least 5% bio-waste oil. The measurement process required neither sample preparation nor chemical reagent consumption, and the time cost was 5 min for each sample. Conclusion This Raman spectrum-cluster analysis model can identify the peanut oil as well as kinds of adulterated peanut oil in a rapid, non-destructive and accurate way.