摘要
以大豆油(33种)、花生油(39种)、葵花籽油(17种)、玉米油(10种)、棕榈油(28种)、芝麻油(37种)和菜籽油(58种)为样品,采用反相高效液相色谱法(RP-HPLC)分别测定其甘三酯组成.以乙腈和二氯甲烷为流动相,蒸发光散射检测器检测,通过JUPITER C18色谱柱将油脂分离为多簇峰(38个);随机取2/3样品数量作为定标集,使用NIRCal-5.2软件的数据处理向导,以甘三酯相对含量为变量,对其进行log(logarithm)对数处理和ncl(Normalization by Closure)标准化处理,使用聚类分析-主成分分析(CLU-PCA)方法,经过优化识别模型,7种植物纯油被正确识别为7类;其余1/3样品为验证集,除玉米油和菜籽油的识别率为90%和92%外,其他识别模型的准确率和验证准确率均为100%.
We determined triacylglycerols of 222 oil samples, including 33 soybean oil samples, 39 peanut oil samples, 17 sunflower seed oil samples, 10 maize oil samples,28 palm oil samples,27 sesame oil samples and 58 rapeseed oil samples, by reversed phase high-performance liquid (RP-HPLC). The samples were detected by an evaporative light-scattering detector and were separated to 38 peak clusters through a JUPITER C18 chromatographic colume using acetonitrile and methyl alcohol as mobile phases. We selected 2/3 samples randomly as the calibration set, and conducted logarithm treatment and ncl (Normalization by Closure) treatment by using NIRCal 5.2 software taking the relative content of triacylglycerols as variable. We analyzed the data by cluster--principal component analysis (CLU-PCA), and optimized the identification model;and the seven kinds of pure vegetable oils were identified as seven types correctly. The other 1/3 samples were used as the validation set, and the results showed that the identification rates of the model for the vegetable oils were 100% ,except maize oil and rapeseed oil, which were 90% and 92%.
出处
《河南工业大学学报(自然科学版)》
CAS
北大核心
2014年第1期1-5,共5页
Journal of Henan University of Technology:Natural Science Edition