The Optical Design And Implementation Of Multicolor Fluorescence Spectrum Detection For The Microchip Electrophoresis System

Microchip Electrophoresis (ME) has long been regarded as a powerful analytical separation technique that is an alternative to more traditional methods such as capillary electrophoresis (CE), gel electrophoresis (GE) and liquid chromatography (LC). In the electrophoresis, samples binding with fluorescence are separated and detected, making the optical fluorescence detecting unit a vital part of the whole system.During the past decades, a varity of systems have been developed for the ME, such as the system employing the dichroscopes, optical filters and photomultiplier tubes (PMTs). However, PMTs present some drawbacks that prevent their use in compact genotyping systems:they cannot be integrated, so detection of the four fluorescences can occur only using four different PMTs, increasing its size and cost. Some other systems use gratings as the dispersion component along with custom-made APDs. But the low efficient of grating will lead to low signal noise ratio (SNR) and the custom-designed APDs will also limit their applications due to the high cost and unchangeable structure. Nowadays, CCDs are more widely applied in biology, chemistry, as well as the electrophoresis. Fluorescence detection systems which make use of CCDs have shown up in recent years. However, the applications of CCDs have been limited in ME because of the low sensitivity, compared with PMTs or APDs. Thus we developed a fluorescence spectrum detection system using prism as the dispersion component, in order to increase the sensitivity of the CCD-based detection system. In detail, research work in this thesis includes:(1) This thesis has done the simulation in Zemax, investigating the linear dispersion with respect to apex angle of the prism and the incident angle. The simulation results indicate that using prism as the dispersion component is feasible for the multicolor fluorescence spectrum detection system.(2) This thesis has designed, built and debugged the laser-induced fluorescence detection system for ME, using the prism and CCD. Besides, the transformation algorithm from spectrum image to spectrum is introduced and SNR, as well as the signal background ratio (SBR), of fluorescein solution with different concentrations is tested on this system. The results show that the fluorescein solution at concentration of 10-9 mol/L still holds the SNR of 20.961 dB and SBR of 0.438. Moreover, the photobleaching experiment demonstrates that low-concentration solution will be photobleached more easily, compared to the high one.(3) This thesis utilizes the multiple-wavelength scanning method to conduct the wavelength calibration and studies different peek-locating algorithms and different fitting functions. The final calibration function is three order and the centroid algorithm is used to locate the peek. The resolution of the fluorescence spectrum detection system is then measured, with the resolution of 546.074 nm spectral line to be 4.597 nm and that of 576.960 nm and 579.066 nm spectral lines to be 6.119 nm.(4) The static fluorescence detection experiments measure the SNR and SBR of four fluorochromes (6-FAM, HEX, TAMRA, ROX) at concentration of 10-7 mol/L and 10-8 mol/L, which will be further used in the ME. The results indicate that 6-FAM is not suitable for this system and that the SNR and SBR of the other 3 fluorochromes have the following trend:HEX>TAMRA>ROX.(5) This thesis proposes the fluorescent correction algorithm to extract the fluorescence intensity of different fluorochromes, based on the least-square solution of overdermined equations. Using the mixure of fluorochromes solution as the sample, the static experiments prove that the proposed algorithm is stable and well-functioned.(6) The ME experiments have been conducted on this system. The results demonstrate that during the electrophoresis process, a fourth fluorochrome is separated, which is later proved to be Syber Green. After revising the algorithm to adapt to the fouth fluorochrome, our system as well as the fluorescent correction algorithm is properly-worked and the results show that we have finally built up a stable, sensitive and low-cost multicolor fluorescence spectrum detection system for ME.