基于傅里叶近红外光谱的食用油过氧化值间接检测研究  

Indirect Determination of Peroxide Value of Edible Oils by Fourier Transform Near Infrared Spectrum Spectroscopy

作　　者：高媛[1] 王宁[1] 于修烛[1] 陈秀梅[1] 

机构地区：[1]西北农林科技大学食品科学与工程学院,杨凌712100

出　　处：《中国粮油学报》2015年第5期138-142,共5页Journal of the Chinese Cereals and Oils Association

基　　金：陕西省科技攻关(2012K02-11)

摘　　要：为了建立食用油过氧化值的快速测定方法,以食用油为材料,基于氢过氧化物与三苯基磷（TPP）快速反应生成三苯基氧磷（TPPO）,利用TPP和TPPO在近红外光谱的吸收差异性。在光谱范围为4 650～4 500 cm^-1以一阶求导及Norris平滑（NF）后采用PLS建立食用油过氧化值近红外测定模型,通过校正样品及盲样对模型进行验证,并与国标碘量法进行比较分析。结果表明：校正集预测相关系数为0.999 4,校正均方根误差为0.143 9,验证预测相关系数为0.992 6,预测均方根误差为0.555 6,盲样预测相关系数为0.997 1,标准偏差为0.400 1,所建方法与国标法测定过氧化值结果非常接近。表明傅里叶近红外光谱间接测定食用油过氧化值是可行的。In order to develop a high efficient method for the determination of peroxide value （PV） in edible oils, PV determination calibration has been established by edible oils as experiment materials. The stoiehiometric reactions of hydroperoxides with the triphenylphosphine （TPP） in oil have been utilized to produce triphenyl - phosphineoxide （TPPO） basing on TPP as well as TPPO absorption differences in Near Infrared Spectrum. The first derivation and Norris smoothing in 465 0 ～ 450 0 cm^-1 spectra have been applied to build PLS models for Near Infrared Spectrum determination of peroxide values in edible oils, the model was validated by standard samples and blinded samples. The results showed ： correlation coefficient of calibration was 0.992 6, the root mean square error of calibration （ RMSEC ） was 0. 143 9, the validation model with a correlation coefficient of 0.992 6, the root mean square error of prediction （RMSEP） was 0. 555 6, the correlation coefficient of blind validation was 0. 997 1, and SD was 0. 400 1. The result also showed that the proposed method is correlated to national standard method, and there are no significant differences between these methods. Therefore, the proposed method can be accessible to provide a rapid determination of peroxide value in the edible oils.

关 键 词：食用油 过氧化值 傅里叶近红外光谱 检测 

分 类 号：TQ646[化学工程—精细化工]