中国激光, 2020, 47 (10): 1011002, 网络出版: 2020-10-16
非平滑非负矩阵分解解析土壤多环芳烃三维荧光光谱 下载: 708次
Analysis on Three-Dimensional Fluorescence Spectra of PAHs in Soil Using Nonsmooth Non-Negative Matrix Factorization
光谱学 三维荧光光谱 非负矩阵分解 土壤 多环芳烃 组分识别 spectroscopy three-dimensional fluorescence spectra nonnegative matrix factorization soil polycyclic aromatic hydrocarbons component recognition
摘要
采用荧光分光光度计直接获取土壤中多环芳烃(PAHs)的三维荧光光谱,并利用非平滑非负矩阵分解(nsNMF)对其进行解析,结果表明,非负矩阵分解(NMF)能够从混叠光谱中提取出单一多环芳烃的荧光光谱信号。在随机初始值下,nsNMF优于基于交替式非负最小二乘的标准非负矩阵分解(NMF/ANLS),解析光谱与参考光谱的相似系数均在0.824以上。特别是在农田土壤中,菲和蒽的解析光谱与标准参考光谱的相似系数分别由0.758、0.845(NMF/ANLS)提高到0.907、0.913(nsNMF)。三维荧光光谱结合nsNMF能够实现土壤多环芳烃组分的快速识别。
Abstract
Three-dimensional fluorescence spectra of polycyclic aromatic hydrocarbons (PAHs) in soil are directly recorded using a fluorescence spectrophotometer. To identify the components of PAHs in soil, nonsmooth non-negative matrix factorization (nsNMF) are used. Results show that NMF can effectively extract the fluorescence spectrum signal of a single PAH from the mixture spectrum. The similarity coefficient between the analytical spectra and corresponding reference spectra obtained by nsNMF under random initial values is all above 0.824, which is higher than that of the standard NMF based on alternating non-negative least squares (NMF/ANLS). In farmland soil, the similarity coefficients of phenanthrene and anthracene between the analytical spectra and corresponding reference spectra increased from 0.758 and 0.845 (NMF/ANLS) to 0.907 and 0.913 (nsNMF), respectively. The combination of three-dimensional fluorescence spectra and nsNMF can facilitate rapid identification of components of PAHs in soil.
黄尧, 赵南京, 孟德硕, 左兆陆, 程钊, 陈宇男, 陈晓伟. 非平滑非负矩阵分解解析土壤多环芳烃三维荧光光谱[J]. 中国激光, 2020, 47(10): 1011002. Huang Yao, Zhao Nanjing, Meng Deshuo, Zuo Zhaolu, Cheng Zhao, Chen Yunan, Chen Xiaowei. Analysis on Three-Dimensional Fluorescence Spectra of PAHs in Soil Using Nonsmooth Non-Negative Matrix Factorization[J]. Chinese Journal of Lasers, 2020, 47(10): 1011002.