三维荧光光谱分析法以其灵敏度高、 选择性好、 操作简单和可用于多组分混合物分析等优点成为诸多研究者在海面溢油鉴别中的热点选择。 但三维荧光光谱中存在的瑞利散射会对光谱的准确检测产生较大的影响, 因此有效地消除瑞利散射对后续光谱的定性鉴别和定量分析具有重要意义。 采用仪器校正法、 空白扣除法、 Delaunay三角形内插值法和缺损数据重构(MDR)法对海面溢油三维荧光光谱中的瑞利散射进行校正。 首先以海水的SDS胶束溶液作为溶剂, 将航空煤油和润滑油按不同相对体积分数比配制8个校正样本和3个测试样本; 然后利用FS920稳态荧光光谱仪采集11个样本的三维荧光光谱数据, 并分别采用仪器校正法、 空白扣除法、 Delaunay三角形内插值法和缺损数据重构(MDR)法消除瑞利散射的干扰; 再利用核一致诊断法估计出最佳的组分数; 最后利用平行因子分析(PARAFAC)对混合油样本的三维荧光光谱数据进行定性鉴别和定量分析。 研究结果表明: 采用发射波长滞后激发波长以消除瑞利散射的仪器校正法会丢失部分有效光谱信息; 采用空白扣除法无法彻底消除瑞利散射, 在光谱中仍然存在散射干扰, 利用PARAFAC解析后得到的激发、 发射光谱会出现失真, 且预测的浓度值偏差较大; 采用Delaunay三角形内插值法消除瑞利散射后, 利用PARAFAC解析所得到的激发、 发射光谱与真实光谱吻合度较高, 且预测的浓度值偏差较小; 而采用MDR消除瑞利散射后, 利用PARAFAC解析所获得的激发、 发射光谱与真实光谱吻合度最高, 且相较于其他几种方法预测的浓度值偏差最小, 得到的样本回收率为98.9%和100%, 预测均方根误差均小于等于0.130。 根据定性鉴别、 定量分析的结果, MDR能够在保证原有特征光谱不失真的基础上有效消除瑞利散射带来的影响, 是一种消除三维荧光光谱数据中瑞利散射较为理想的方法。

Three-dimensional fluorescence spectroscopy has become a hot topic in the identification of oil spills by many researchers because of its high sensitivity, good selectivity, simple operation and analysis for multi-component mixtures. However, the Rayleigh scattering in the three-dimensional fluorescence spectrum will have a great influence on the accurate detection of the spectrum, so it is of great significance to eliminate the influence of the Rayleigh scattering effectively for the qualitative identification and quantitative analysis of the spectrum. In this paper, the instrument calibration method, background subtraction method, Delaunay triangle interpolation method and Missing Data Recovery (MDR) method were used to correct the Rayleigh scattering in the three-dimensional fluorescence spectrum of the oil spill. Firstly, the seawater SDS micelle solution was used as a solvent, the jet fuel and the lube were mixed according to different relative volume fraction ratios to prepare eight calibration samples and three test samples. Then, the three-dimensional fluorescence spectra of 11 samples were determined by FS920 steady-state fluorescence spectrometer. Moreover, the interference of Rayleigh scattering was eliminated by instrument calibration method, background subtraction method, the Delaunay triangle interpolation method and missing data recovery (MDR) method respectively. Then the kernel consensus diagnosis method was used to estimate the optimal number of components. Finally, the PARAFAC was used to qualitatively identify and quantify the three-dimensional fluorescence spectrum data of the mixed oil samples. The results show that the instrument calibration method using the emission wavelength lag excitation wavelength to eliminate Rayleigh scattering will lose part of the effective spectral information. The background subtraction method cannot completely eliminate the Rayleigh scattering, and there is still scattering interference in the spectrum. The excitation and emission spectra obtained by PARAFAC will be distorted, and the predicted concentration value deviation is large. After the Rayleigh scattering is eliminated by Delaunay triangle interpolation method, the excitation and emission spectra obtained by PARAFAC have a higher agreement with the real spectrum, and the predicted concentration value deviation is small. However,after the Rayleigh scattering is eliminated by MDR, the excitation and emission spectra obtained by PARAFAC analysis have the highest agreement with the real spectrum, and the predicted concentration value deviation is the smallest of these methods, and the sample recovery rate is 98.9% and 100% respectively, the RMSEP is limited to less than 0.130. According to the results of qualitative identification and quantitative analysis, MDR can effectively eliminate the influence of Rayleigh scattering on the basis of ensuring that the original characteristic spectrum is not distorted. It is an ideal method to eliminate Rayleigh scattering in the three-dimensional fluorescence spectrum.