食用植物油中农药残留及苯并(a)芘含量的共线双脉冲LIBS快速定量检测方法研究

In this research, a new method that using collinear dual-pulse laser induced breakdown spectroscopy (LIBS) combined with faint signal extraction method is proposed to detect pesticide residues and benzo(a)pyrene contents in edible vegetable oils. Experimental setup of LIBS will be improved and optimized acorrding to research contents and object, also effect of parameters of spectrum acquisition will be studied. After spectrum acquisition, several methods including non-orthogonal amplitude frequency analysis of the smoothed squences(NAFASS), similarity measurement, statistical detection of the hidden distortions in spectra and probability circles analysis are used to process spectral data, remove noise signal, decrease background singal of edible vegetable oils, and obtain characteristic spectra of element and radical group of pesticide residues and benzo(a)pyrene. Then characteristic spectra of element and radical group are combined suitably, and related to the corresponding real values of pesticide residues and benzo(a)pyrene contents to develop prediction models by chemometrics. After that, pesticide residues and benzo(a)pyrene contents in edible vegetable oils can be predicted by prediction models..The detection method for detecting pesticide residues and benzo(a)pyrene in edible vegetable oils will be obtained after the research project is accomplished. The detection method also can be used in other farm products/foods, and this can monitor farm product/food safety in China effectively and ensure people's health. This is meaningful for society.

本项目在课题组多年农药残留光谱检测的基础上，提出利用共线双脉冲激光诱导击穿光谱技术结合微弱信号提取快速定量检测食用植物油中农药残留及苯并(a)芘含量的新方法。针对研究内容和检测对象，改进及优化光谱检测实验装置，并探讨光谱采集参数等影响；采用平滑序列非正交幅频、相似性测度、光谱微弱差异计算及概率圆分析等方法处理光谱数据，去除光谱噪声并降低本体信号，藉此获取痕量农药残留及苯并(a)芘的特征光谱(元素特征谱线和基团特征谱带)；将元素特征谱线与基团特征谱带优化组合，采用化学计量学方法将其与农药残留及苯并(a)芘真实含量进行关联，建立预测模型，实现农药残留及苯并(a)芘含量快速检测。. 项目的完成将获得食用植物油中农药残留及苯并(a)芘含量的快速检测方法。该方法可用于其他农产品/食品中的农药残留及苯并(a)芘快速检测，可以有效监测我国农产品/食品安全，保障人们身体健康，社会意义重大。

农药残留和苯并(a)芘为食用植物油的质量安全问题之一，严重危害人们的身体健康。实现农药残留和苯并(a)芘痕量成分快速检测是保证食用植物油质量安全的有效方法。. 研究项目利用共线双脉冲激光诱导击穿光谱技术快速定量检测食用植物油中的痕量农药残留及苯并(a)芘含量。针对项目研究内容和检测对象，改进及优化LIBS 检测实验装置，探讨4种不同光谱采集方式（原始样品、滤纸富集、木片富集、石墨富集）的优劣并优化激光能量、光谱采集延时及两脉冲延时等光谱采集参数；采用平滑、多元散射校正等方法进行光谱预处理，利用无信息变量消除、竞争性自适应重加权采样等算法提取4种农药（乐果、倍硫磷、敌草快、腐霉利）和苯并(a)芘的特征光谱，应用内标法、多元线性回归、偏最小二乘回归及最小二乘支持向量机方法建立食用植物油中4种农药残留和苯并(a)芘含量的定量预测模型。研究结果表明，4种光谱采集方式中，石墨富集方式的LIBS光谱信号最强，效果最好，木片和滤纸富集方式的效果次之，而原始样品的效果最差；对比常温和低温富集，低温富集的效果更优，LIBS光谱信号更强。此外，竞争性自适应重加权采样及无信息变量消除均为有效的特征光谱选择方法，内标法有效校正激光能量波动等因素影响，上述方法结合偏最小二乘回归及最小二乘支持向量机等方法可有效提高农药残留及苯并(a)芘定量预测模型的精度和稳定性。在项目的支持下，发表及录用相关论文6篇，其中SCI论文3篇，授权发明和实用新型专利各1项，公开发明专利1项。. 研究项目获得了食用植物油中农药残留及苯并(a)芘含量的快速检测方法。该方法可用于其他农产品/食品中的农药残留及苯并(a)芘快速检测，可以有效监测我国农产品/食品安全，保障人们身体健康，社会意义重大。