基于荧光猝灭原理的光纤化学传感器定量分析模 型的建立与应用

根据光纤化学传感器的结构与信号传输特征，推导并建立了适合于定量描述基于荧光猝灭原理的光纤化学传感器对待测物响应的非线性数学模型。以此为基础，为更加简捷、直观地反映响应信号与待测物浓度的函数关系，建立可线性化的回归方程，准确预报待测物的浓度，提出了多模型回归技术建立此类光纤化学传感器定量分析模型的方法。为难定量分析模型的适用性，将该法应用于芘丁酸光纤化学传感器对甲硝唑、呋喃妥因和氧氟沙星等待测物响应的最佳数学模型的选择之中。结果表明，所选的数学模型对各待测物在一定范围内的浓度都能进行准确的预报。

Quantitative analysis model for fiber-optic chemical sensors based on fluorescence quenching for analytes and its application

Because the fiber-optic chemical sensors based on the fluorescence quenching has an adjustable short distance from the membrane to the end of fiber optics, forming a space of microcell, a diminution of fluorescence is induced by inner filter effects, involving absorption of both excited light from the light source and emitted light from the membrane. If the absorption spectra of analytes overlap the fluorescence excitation and/or emission spectra of the fluoroprobe in the membrane, the quenching signal of the fiber-optic chemical sensors would be produced by the resonance energy transfer. In addition, dynamic quenching happens to some haloid, heavy metal compounds and aromatic nitro organic compounds. According to the structural characterization of the fiber-optic chemical sensors based on fluorescence quenching and the mechanism of fluorescence multiple quenching, a non-linear mathematical model was deduced and described for the quantitative analysis model for the fiber-optic chemical sensors. A multiple model regression technique for the quantitative analysis model was reported to provide rapidly and directly the relations between the response signal and the concentration of analytes and establish linear regression equation for predicting the concentration of analytes. The technique was applied to fit best mathematical model from the control samples of the therapeutic drugs based on the response of fiber-optic chemical sensors, such as metronidazole, nitrofurantoin and ofloxacin. The pyrenebutyric acid was chosen as a fluoroprobe for constructing the fiber-optic chemical sensor to response the samples. The experimental results showed that these models had some good characteristics and gave an alternative method for establishing quantitative analysis models for the fiber-optic chemical sensors.