人工神经网络—化学发光法在多组分同时测定中的应用

【作者】 李金梅；

【导师】 李保新；

【作者基本信息】 陕西师范大学 ， 分析化学， 2008， 硕士

【摘要】 化学发光分析法具有灵敏度高、线性范围宽、分析速度快以及仪器设备相对简单等诸多优点，近年来在无机及有机痕量和超痕量分析领域得到了广泛应用。然而，由于其选择性差，限制了该方法在复杂样品分析中的应用。当一种或多种干扰物质共存时，采用该方法很难直接对分析物进行检测。由于人工神经网络具有良好的自适应性，适合于线性和非线性系统，能够模拟多输入输出间的复杂关系，已成为近年来化学计量学中的热门研究领域。将化学计量学中的人工神经网络方法应用到化学发光分析中，从一定程度上解决了传统化学发光选择性差的问题，扩大了化学发光的应用范围。本文的主要目的是建立人工神经网络化学发光多元校正模型，并研究其在实际分析中的应用。本论文包括两个部分。第一部分叙述了人工神经网络的原理及结构，重点介绍了最常用的基于误差反向传播算法神经网络的构建，并就人工神经网络在分析化学方面的最新进展进行了综述。第二部分是研究报告，具体内容如下：一、人工神经网络辅助停流化学发光法同时测定敌敌畏和氧乐果将人工神经网络校正方法应用于化学发光分析，采用停流混合技术，建立了一种同时测定敌敌畏和氧乐果的新方法。在鲁米诺一过氧化氢化学发光体系中，过氧化氢首先氧化敌敌畏和氧乐果生成过氧化磷酸盐，该产物具有比过氧化氢更强的氧化能力，可氧化鲁米诺产生更强的化学发光。本法根据敌敌畏和氧乐果在该化学发光体系中反应的动力学曲线有显著的差异，通过测定和记录整个停流过程的化学发光强度，运用人工神经网络建立校正模型，实现了不经分离对敌敌畏和氧乐果的同时测定。该法的采样、停流和进样都有计算机自动控制，因而具有操作简便、灵敏度高、精密度好等优点。并已成功地用于蔬菜表面这两种有机磷农药残留的测定。二、化学发光法结合人工神经网络同时测定卡托普利和氢氯噻嗪提出了一种静态化学发光体系结合人工神经网络多元校正方法同时测定卡托普利和氢氯噻嗪的新方法。实验发现，在罗丹明6G存在下，卡托普利和氢氯噻嗪均能被Ce（Ⅳ）氧化产生强的化学发光，但二者的化学发光动力学特性有显著性差异。通过测定和记录整个过程中的化学发光信号，运用人工神经网络建立校正模型并进行预测，实现了不经分离对卡托普利和氢氯噻嗪的同时测定。本法具有灵敏度高、线性范围宽等优点。并用于药物中这两组分的同时测定，结果令人满意。三、鲁米诺-铁氰化钾化学发光体系测定孔雀石绿基于鲁米诺在碱性条件下可以被铁氰化钾催化氧化产生化学发光，孔雀石绿对此化学发光具有增敏作用这一现象，结合流动注射技术建立了一种直接测定孔雀石绿的流动注射化学发光新方法。结果表明，该方法的线性范围为1．0×10^-6-9．0×10^-5mol／L，检出限为4×10^-7mol/L。对1．0×10^-5mol／L的孔雀石绿连续进行七次平行测定，其相对标准偏差为2．3％。更多还原

【Abstract】 Chemiluminescence （CL） detection has been widely used in many fields due to its attracting features including high sensitivity, low detection limit, wide linear dynamic range and inexpensive instruments. However, the relatively poor selectivity of the CL method itself limits its direct application to the analysis of complicated sample. The use of some analytical methods combined with multivariate calibration can be considered a promising, faster, direct and relatively less expensive alternative for the multicomponent analysis of mixture. This paper focused on the multivariate calibration of chemiluminescence combined with artificial neural network method.This thesis includes two parts. In part one, the principle and structure of artificial neural network （ANN） were introduced, and the development of ANN in analytical chemistry was also summarized. Part two is research report, and the obtained results are listed as follows:1. Simultaneous determination of dichlorvos and omethoate using stopped-flow chemiluminescence with the aid of artificial neural network calibrationIn this paper, artificial neural network and stopped-flow CL measurement was combined for the simultaneous determination dichlorvos and omethoate. The method is based on the different kinetics between the two analytes in Luminol-H₂O₂ CL system.The mixed CL intensity was monitored and recorded on the whole process of stopped-flow, and the data obtained were processed chemometrically by use of an artificial neural network.. The proposed method was successfully applied to the simultaneous determination of the organophosphorous pesticides in some vegetable samples.2 Simultaneous determination of captopril and hydrochlorothiazide using chemiluminescence with the aid of artificial neural networksA chemiluminescence system combining artificial neural network multivariate calibration for simultaneous determination of captopril and hydrochlorothiazide was proposed. In the presence of rhodamine 6 G, Ce（Ⅳ）could oxidize captopril and hydrochlorothiazide, respectively, to produce strong CL emission, and the two CL dynamic characteristics was significantly different. The CL intensity was measured and recorded on the whole process, and the obtained data were processed by the chemometric approach of artificial neural network.The proposed method was applied to the simultaneous determination of CPL and HCT in pharmaceutical formulation with satisfactory results. 3. Determination of malachite green based on chemiluminescence reaction of Luminol with ferricyanideA novel chemiluminescence method coupled with flow injection technique for the determination of malachite green is developed. It is based on the enhancement of malachite green on the chemiluminescence reaction of luminol with ferricyanide in sodium hydroxide medium. The linear calibration range of the chemiluminescence intensity to the malachite green concentration is 1.0×10^-6 to 1.0×10^-4 mol/L. The relative standard deviation for 1.0×10^-5mol/L malachite is 2.3% （n=7）, and the detection limit is 4×10^-7 mol/L.更多还原