【作者】 许成才；

【导师】 闫卫平；

【作者基本信息】 大连理工大学 ， 测试计量技术及仪器， 2006， 硕士

【摘要】 芯片毛细管电泳具有微型、快速、低样品量等特点，被认为是生化分析领域最具发展潜力的一种技术。光纤嵌入式毛细管电泳芯片因其体积小、检测光路简单等诸多优点，正在引起人们的关注。但是，其检测系统中外加高压电场、高强度激光、仪器和芯片结构等诸多因素产生的噪声会混入检测信号，影响系统的检测灵敏度。因此，本文基于激光诱导荧光检测原理，开展了光纤嵌入式毛细管电泳芯片的信号检测及信号消噪处理的研究。 本文在大量实验的基础上，设计了光纤嵌入式“十”字型和弯曲型毛细管电泳芯片，采用MEMS技术制作了以玻璃为基体的毛细管电泳芯片。改进了以AT89C51单片机为核心的高压控制系统，实现了样品的多次进样分离操作及信号的自动检测功能。开发了基于VC++的系统应用软件，可实现数据通信、数据显示与保存等功能。在自行搭建的检测系统上，利用毛细管电泳芯片对罗丹明B、罗丹明6G进行了最低浓度限及分离等检测。针对毛细管电泳信号的噪声特性，本文首先利用Matlab软件模拟出毛细管电泳信号，通过选取评价消噪效果优劣的标准，比较研究了三族小波基函数、分解层数、阈值方法等对消噪的影响，确定出“小波基函数—sym4、分解层数—4、各级阈值方法”的最佳消噪策略。通过与传统的傅立叶变换消噪算法比较，证明小波消噪能更有效地消除电泳信号中的噪声，保留有用信号的特征，且消噪后峰形和峰高失真小，基线比较平稳。本文利用该策略对检测到的真实电泳信号进行消噪处理，获得了较好的结果。实验表明，该系统具有结构简单、操作方便、灵敏度高、可重复进样等优点，有利于微流控系统的微型化和集成化。更多还原

【Abstract】 Microchip capillary electrophoresis(CE) has been recognized as a powerful tool for biochemical analyses due to its micro size, fast separation and low sample requirement. CE microchip based on optical fiber embedded arouses people’s attention for its small volume, simple detection optical route. However, some noises which affect system’s detection sensitivity are introduced into the signal because of the influence of high voltage, instrument, chip structure and some other factors. So the study of signal detection of optical fiber embedded CE microchip and signal denoising disposal is discussed in this paper base on the theory of laser-induced fluorescence.Cross and curving CE micro chips were designed and manufactured using MEMS technics upon many experiments in this papaer. The high voltage system was improved which realized operations of repeated injection and separation and signal auto-detection. The system application software based on VC++ was developed which realized functions of data communication, display, saving and so on. The whole detection system was constructed and the lowest concentration detection and separation detection of Rhodamine B and Rhodamine 6G were completed. According to the character of CE signal and noise, the simulated CE signal was produced using Matlab software. Three typical wavelet sets, level of decomposition and threshholding method were evaluated by choosing denoising criterion. The sym4 wavelet with a level decomposition of 5 and the all levels thresholding method were conformed as the optimum denoising strategy. Compared with Fourier transform, wavelet denoising could remove the noise adequately, reserve the character of useful signal and achieve a better preservation of the shapes of signal. Using the optimum denoising strategy, a good result was obtained. The result shows that, this detection system has some characteristics, such as compact structure, high sensitivity and repetition, which are favour to the development of micrmation and integration of micro CE chip.更多还原