【作者】 刘丽丽；

【导师】 孙裕晶；

【作者基本信息】 吉林大学 ， 农业机械化工程， 2009， 硕士

【摘要】 鉴于目前蔬菜农药残留分析检测耗时长,制备后的样品稳定性差而影响分析的准确性等问题,本文以白菜乐果残留为代表性研究对象,研究基于超高压预处理的蔬菜农药残留近红外光谱技术。通过试验选择适用于近红外光谱检测的超高压预处理工艺参数,确定近红外光谱检测的技术参数;从直观观察白菜乐果残留样品近红外光谱和偏最小二乘法近红外光谱模型两方面分析超高压预处理对近红外光谱检测的影响。旨在寻找一种新的蔬菜中农药残留检测方法,提高农药残留分析的速度和样品的稳定性。主要研究工作如下:1、理论分析了近红外光谱检测的技术、影响近红外光谱检测的因素、超高压对近红外光谱检测的影响以及乐果和蔬菜的近红外光谱特征,提出基于超高压预处理的蔬菜农药残留近红外光谱检测的研究方案。2、根据样品中固体悬浮物的大小和量的多少和气相色谱检测试验分析,选取了适用于近红外光谱检测白菜乐果残留的超高压预处理的参数(保压时间为3min、压强为300MPa、添加提取试剂50mL、白菜25g);根据近红外光谱效果选取近红外光谱仪器的技术参数(分辨率为4cm-1、扫描次数为32次、光程为1mm)。3、在选取超高压预处理的参数时发现,随着放置时间的增加,压强大的样品首先出现沉淀,且蔬菜原液中的沉淀随压强的增大而减少。4、使用德国Bruker公司的VECTOR 22/N近红外光谱仪扫描样品,直观观察了近红外光谱图,超高压预处理对试剂、白菜原液、含有农药的提取溶液近红外光谱的吸收度、峰位、平滑度没有影响。5、建立了两组白菜乐果残留浓度的近红外光谱偏最小二乘法定量分析模型(简称光谱模型),一组未进行超高压预处理(简称无压),另一组样品进行超高压预处理(简称有压)。无压样品的光谱模型的相关系数为75.37%、内部交叉验证均方差为0.192,检出限无法达到10-4(质量体积浓度);有压样品的光谱模型的相关系数为87.14%,内部交叉验证均方差为0.139,检出限可达10-5,较无压样品光谱模型的检出限提高了近2个数量级;在整个建模过程中,有压样品光谱数据稳定,出现的奇异点少。以全天工作计算,基于超高压预处理的白菜乐果残留近红外光谱检测速度至少是NY/T761-2004有机磷检测速度的20倍。更多还原

【Abstract】 Issues of food safety are worldwide highlights. Pesticide residues in vegetables are among the most sensitive problems in daily life, which not only results in the loss of finance、credibility, even the lives, but also causes the economic barriers to the international trade and serious environmental pollution. Only by enhangcing the monitoring efforts, exploring more convenient, faster, cost-effective detection method, can the harmful effects of the pesticide residues be kept within limits effectively.The current analysis methods of the pesticide residues have some drawbacks. For example, the complication of the operation process, time-noneffective and poor stability of the prepared samples which often makes the accuracy of the analysis below acceptable standard. So, a novel method, near infrared detection of the pesticide residues in vegetables based on ultra-high pressure pretreatment, is put forward.Near infrared detection is fast, efficient, simple and no sample pretreatment required. Ultra-high pressure provides high rate of extraction, low impurities, short time of extraction, less use of solvents and pollution-free. The active ingredients are more stable at room temperature conditions.In this paper, near infrared detection technologies of the pesticide residues in vegetables based on ultra-high pressure pretreatment are studied. The organophosphorus pesticide residue-rogor in Chinese cabbage is taken as representative study object. Research programs are identified according to theoretical analysis. Pretreatment process parameters are selected for fitting near infrared detection. The technical parameters of near infrared detection instrument are determined. The impacts of ultra-high pressure pretreatment on near infrared detection are analyzed by visual observation of near infrared spectrum and analysis of partial least squares model of near infrared spectrum. It is aimed at finding a new way to improve the speed and stability of the analysis of pesticide residues in vegetables. The main research contents of this thesis are as follows:1、The theories of near infrared detection, factors and parameters in near infrared detection, impacts of ultra-high pressure pretreatment on near infrared detection and near infrared spectrum characteristics of rogor and vegetables are analyzed. It is found that the changes of some chemical bonds, some composition, particle size and amount of suspended solids affect the material characteristic peak of near infrared spectrum, optical path and absorbance of near infrared spectrum, in turn affect the wave. So near infrared detection of the pesticide-residue in vegetables based on ultra-high pressure pretreatment is raised through theoretical analysis.2、According to the gas chromatography tests and the analysis of particle size and amount of suspended solids, pretreatment process parameters are selected for fitting near infrared detection of rogor residue in Chinese cabbage, immersion time is 3 minutes, volume of reagent is 50mL, quality of vegetables is 25mg, pressure is 300MPa. The nominal resolution is 4cm-1 selected through the numbers of data points that the near infrared spectra transforms into, the accumulating scans is 32 chosen according to the effect of reducing the noise after averaging the spectrum, the optical path of near infrared detection annex is 1mm selected by absorbance and energy situation of the spectrum. T est is operated in about 20℃; the concentration of rogor is from 10-3 to 10-6. The total number of samples is 28, which is divided into validation set , 24 ,and prediction set , 4.During choosing the ultra-high pressure pretreatment parameters, it is found that vegetable juice treated by different pressures (no pressure, 100MPa, 200MPa, 300MPa, 400MPa) deposits form the high pressure distinguished from 300MPa after standing in a short time. Over time, all samples deposit, the juice is clearer in turns along with increasing pressures. The less suspended solids are contained in samples, the less seriously the scattering is impacted, which is good for improving spectral quantity. The phenomena indicates that affecting scattering made by suspended solids in samples become less with increasing pressure.3、The impacts of ultra-high pressure pretreatment on near infrared detection are studed initially. The near infrared spectrum are observed from intuition. The absorbance, position of peak and smooth of the spectra are not affected by the ultra-high pressure. There are no remarkable differences in spectrum treated by different pressures.4、the PLS models are used to analyze impact of ultra-high pressure pretreatment on near infrared detection. In the process of data preprocessing and modeling, data of samples treated by pressure are more stable than that of samples without ultra-high pressure. There are less outliers. By comparing results obtained by different spectral pretreatments, the appropriate wave-length range and methods of preprocessing of samples with pressure and without are got. Spectrum of samples without ultra-high pressure are dealt with in the range between 4500 and 10000 cm-1, treated by VN, smooth (25 points), after optimized, the wave-length are reduced in the 6101.9-5448.1 cm-1 and 4601.5-4499.3 cm-1, the PLS model is established by FD+MSC. Spectrum of samples with ultra-high pressure are dealt with in the range between 4100 and 11000 cm-1, treated by smooth (21 points), after optimized, the wave-length are reduced in the 6800-6099.9 cm-1 and 4601.5-4248.5cm-1, the PLS model is built by VN.The model’s quantity of samples with ultra-high pressure is better than that of samples without ultra-high pressure. Analyzing the PLS model and the results of prediction processed by OPUS, the related coefficient of models of samples with pressure and without are 75.93% and 56.8% respectively, the RMSECV are 0.139 and 0.192, the errors of the tow group samples’prediction are 1.557%, 8.17%, -86.88% and 26.57%, -49.39% -93.06%.The limit of near infrared detection of model of samples with ultra-high pressure is improved, 2 times higher than no pressed samples model. Model of samples with ultra-high pressure can detect the 10-5 (quality-volum concentration), Model of samples without ultra-high pressure does not finish detecting at 10-4(quality-volum concentration). The limit of detection of the two models can’t meet the prescription of rogor pesticide residue ruled by the ministry of agriculture, 10-6.5、Calculating in terms of the full-time work, the detection speed of near infrared detection of the pesticide residue in vegetables based on ultra-high pressure pretreatment, is 20 times higher than that obtained by NY/T761-2004 detection of organic phosphorus. Full operation of near infrared detecting pesticide residues based on ultra-high pressure pretreatment, a FT-NIR instrument can detect 30 samples in 1 hour; while a gas chromatogragh instrument can detect 30 samples in 24 hours.Limited by the test time and personal experience,the near infrared detection of rogor residue in Chinese cabbage based on ultra-high pressure pretreatment is investigated in this pater. In terms of the results of the preliminary study, the crafts of pretreatment and the methods of data pretreatments and modeling need to be consummated further. The near infrared detection of pesticide residues in various vegetables should be researched deeply.更多还原