三维荧光光谱结合HGA-RBF神经网络在多环芳烃浓度检测中的应用

采用FS920荧光光谱仪分析了苯并[k]荧蒽(BkF)、苯并[b]荧蒽(BbF)和两者混合物的荧光特性.结果表明BkF的两个荧光峰分别位于306 nm/405 nm和306 nm/430 nm,BbF的两个荧光峰分别位于306 nm/410nm和306 nm/435 nm.BkF和BbF不同浓度配比及其相互间的荧光干扰,使得混合物荧光特性差异较大,荧光强度和浓度间关系变得复杂.为准确测定混合物中BkF和BbF的浓度,采用递阶算法优化的径向基神经网络对其进行检测,结果表明BkF和BbF的平均回收率分别为98.45％和97.71％.该方法能够实现多环芳烃类污染物共存成分的识别和浓度预测.

Concentration Detection of Polycyclic Aromatic Hydrocarbon Combining Three-dimensional Fluorescence Spectroscopy with HGA-RBF Neural Network

Three-dimensional excitation-emission matrix fluorescence spectroscopy of Benzo f[k]Fluoranthene (BkF),Benzo [b] Fluoranthene (BbF),and a mixture of these two substances were analyzed with FS920 fluorescence spectrometer.The results show that the fluorescence peaks of BkF can be observed at 306 nm/405 nm and 306 nm/430 nm,and the fluorescence peaks of BbF locate at 306 nm/410 nm and 306 nm/435 nm.In the mixture of BkF and BbF,concentration ratio and fluorescence interferences make excitation-emission matrix spectra of mixture change largely.Hence,the relationship between fluorescence intensity and concentration is complicated.In order to determine the concentration of BkF and BbF in mixture,radial basis function neural network optimized by hierarchical genetic algorithm was applied and the average recovery of BkF and BbF are 98.45％ and 97.71％,respectively.The results showed that the possibility of the identification and concentration prediction of different components in mixed sample of polycyclic aromatic hydrocarbons.