【作者】 李晓婷；

【导师】 王纪华；

【作者基本信息】 上海交通大学 ， 园艺学， 2013， 博士

【摘要】 在农业生产中，随着农药的超标和滥用，农药对生态的破坏以及农药残留对人类身体健康的危害越演越烈，这一问题已成为全球的焦点，越来越受到社会的广泛关注和高度重视。因此，各个国家持续制定要求越来越严格的农药残留限量标准，并研究与发展相应的农药残留检测技术。红外显微成像技术是一种快速、无损、绿色的检测技术，具有高精度、高灵敏度、图谱合一、微区化和可视化等优点，是了解复杂物质的空间分布和分子组成的强有力方法。本文以红外显微成像系统为检测工具，以氯氰菊酯、毒死蜱以及阿维菌素为研究对象，对果蔬样品表面的农药残留以及生物农药掺假识别开展了定性及定量研究。本文分析了毒死蜱、氯氰菊酯等农药的红外显微图像特征，研究结果得到了氯氰菊酯和毒死蜱主要分子结构及其在红外谱区的特征吸收峰，为该技术用于表面滴加氯氰菊酯和毒死蜱溶液苹果皮的研究提供了依据。对表面滴加单一组分农药苹果皮的研究结果表明，随着氯氰菊酯与毒死蜱两种农药浓度的降低归属于两者的特征吸收峰的数量依次减少，且发生位移的特征吸收峰的数量依次增加。说明随着两种农药溶液浓度的降低，苹果自身的组成成分以及含水量对检测的干扰程度增强，以致红外显微成像技术的检测灵敏度下降。相关图像的分析结果可以方便、快速地获取氯氰菊酯、毒死蜱在苹果皮上分布信息及差异。在单一组分农药研究的基础上进行苹果皮表面滴加复配农药的研究。在滴加3种浓度以及不同比例氯氰菊酯和毒死蜱混合溶液苹果皮的红外谱图及二阶导数谱上均检测到了氯氰菊酯和毒死蜱的共同存在。且随着农药复配溶液浓度的降低，特征吸收峰的总数量依次减少；并且随着农药复配溶液中毒死蜱比例的增加，毒死蜱特征吸收峰的数量逐渐增多，氯氰菊酯特征吸收峰的数量逐渐减少。上述分析证实了红外显微成像技术在农药残留的应用领域中较为灵敏。在果蔬样品内部品质的不同对检测产生的影响的探讨中可知：苹果皮含水量的不同对检测有显著影响，当含水量过高或者过低时检测效果都不好，当相对含水量为50%左右时检测效果较好；苹果皮色素对检测有一定的影响，但具体影响不确定，尚有待于深入的研究；蔬菜叶片表面形态的差异对检测有一定的影响，在较为平整的叶片上应用效果较好，相反地，在褶皱较多的叶片上应用效果相对较差。本文分别基于红外显微成像技术及衰减全反射红外光谱技术建立了一种生物农药掺入化学农药的定性与定量检测技术。定性结果表明：随着阿维菌素中掺入毒死蜱比例的增加，归属于阿维菌素的特征吸收峰的数量不断减少，峰强度不断减弱；相反地，归属于毒死蜱的特征吸收峰的数量逐渐增加，峰强度逐渐加强。利用偏最小二乘法建立阿维菌素乳油制剂掺入毒死蜱的定量预测模型并进行优化，结合外部检验集对模型的性能进行了验证。定量结果表明：衰减全反射红外技术可以准确测定阿维菌素乳油制剂中掺假毒死蜱的含量。通过异常值诊断、光谱预处理及建模参数的优化，提高了模型的预测精度。模型决定系数R2(%)为99.88，校正集均方根误差RMSEC为0.44，交互验证均方根误差RMSECV为0.79，预测集均方根误差RMSEP为0.70。本研究为果蔬表面的农药残留快速检测及生物农药掺假化学农药的快速识别提供了一种新方法，为应用红外显微成像技术及衰减全反射红外光谱技术定性及定量检测农药残留及生物农药掺假奠定了基础。更多还原

【Abstract】 In agricultural production, the pesticides-caused environmental pollutionas well as the harm caused to human health by residual pesticides isconstantly aggravating due to the massive and unreasonable use of pesticides.This issue has attracted a great deal of attention from all circles of society andis coming on global focus. Therefore, world governments formulate theincreasingly strict pesticide residue limits, and develop the detectiontechnologies of pesticide residues accordingly.Infrared micro-imaging technique is a kind of rapid, nondestructive andnon-polluted detection technique, which also has the advantages of theimage-spectrum combined, micro-area partition, visualization, high-accuracy,high-sensitivity and so on. It’s a powerful method to understand the spatialdistribution and the molecular composition of complex materials. Infrared micro-imaging system was as a testing tool, and Cypermethrin, Chlorpyrifos,Abamectin were as research objects in this paper. It launched qualitative andquantitative studies for pesticide residues on the surface of fruit and vegetablesamples and biological pesticide adulteration and recognition. The infraredmicro-image features of Chlorpyrifos, Cypermethrin and other pesticideswere analyzed in the paper. The results got the main molecular structure ofCypermethrin, Chlorpyrifos and its characteristic absorption peaks in theinfrared spectral region. Therefore, it provided feasibility study for thetechnology which used in apple skin sprayed with Cypermethrin andChlorpyrifos solution.The apple skins that sprayed with single component pesticide werestudied. The results show that with the decrease of concentration of twopesticides, the numbers of the characteristic absorption peaks which attributedto both pesticides reduced in turn. And the numbers of characteristicabsorption peaks which shifted increased in turn. At the same time with thereduction of the concentration of two pesticides, the interference of apple’sown composition and water content to detection enhanced.So that thedetection sensitivity of IR micro-imaging technique decreased. Thedifferences of distribution information of Cypermethrin and Chlorpyrifos onapple skin were easily and quickly seen by compare correlation images.The apple skins sprayed with pesticide mixtures were studied on the basis of single component pesticides. The co-existence of Cypermethrin andChlorpyrifos was detected by IR spectra and second derivative spectra ofapple skin which sprayed with three different concentrations and differentproportions of Cypermethrin and Chlorpyrifos mixed solution. In addition,with the decrease of concentration of pesticide mixtures, the total number ofcharacteristic absorption peaks reduced in turn. With the increase ofproportions of Chlorpyrifos in pesticide mixtures, the numbers ofcharacteristic absorption peaks which belonged to Chlorpyrifos were moreand more. But the numbers of characteristic absorption peaks which belongedto Cypermethrin were fewer and fewer. It confirmed that the IRmicro-imaging technique was more sensitive in the areas of pesticideresidues.This study probed the different internal quality of fruit and vegetablesamples which impacted on the detection. The different water contents ofapple skin significantly affected the detection results. It was not good whenthe water contents were too high or too low. It was better when the watercontent was50%. The apple skin pigment certain affected the detectionresults. But specific impact was uncertain. It needed to be in-depth study. Thedifferences of vegetables leaf surface morphology certain impacted on thedetection. The application effect was better in relatively flat leaves. On thecontrary, the application effect was poor in relatively drape leaves. And the identification and detection technology of biological pesticidesadulterated with chemical pesticides was established based on IRmicro-imaging technique and ATR-FTIR technique for the first time. Thequalitative results showed with the increase in the proportions of Chlorpyrifoswhich was adulterated into Abamectin, the number of characteristicabsorption peaks belonged to Abamectin decreased gradually, and the peakintensity weakened. The number of characteristic absorption peaks attributedto Chlorpyrifos was just the opposite, and the peak intensity increasedgradually. The quantitative prediction models of Abamectin EC adulteratedwith Chlorpyrifos were established by partial least squares method andoptimized. And then the external validation sets was used to validation themodel performance. The quantitative results showed that ATR-FTIRtechnology can accurately determine the content of Abamectin EC adulteratedwith Chlorpyrifos. The model of prediction precision was improved throughspectrum pretreatment, outliers diagnosis and modeling parametersoptimization. The determination coefficient (R2(%)) was99.88. The rootmean square error of calibration (RMSEC), the root mean square error ofcross-validation (RMSECV) and the root mean square error of prediction(RMSEP) was0.44,0.79and0.70, respectively.In this study, it provided a new method for rapid detection of pesticideresidues of fruits and vegetables and rapid identification of biological pesticides adulterated with chemical pesticides. It laid a foundation for theapplication of IR micro-imaging technique and ATR-FTIR technology toqualitative and quantitative detection of pesticide residues and biologicalpesticides adulteration.更多还原