Novel method based on inherent connection between absorption and fluorescence spectra for water contaminant analysis

Multi-source spectroscopy is increasingly applied in water contaminant analysis, and general existing spectral features are based on direct mathematical statistics rather than revealing inherent connection between multisource spectra, which has restricted the accuracy and robustness performance. Here in, a novel method is proposed and it is based on the inherent connection between fluorescence and absorption spectra, which can reflect deeper information than conventional methods. The relevant theory was analyzed based on energy level transition and the symmetry between absorption and fluorescence spectra, and then three features were extracted related to internal molecular properties and dependent on two sources of spectral information simultaneously. The three features include the width of the fluorescence emission peak, Stokes shift and symmetry axis between absorption peak and fluorescent emission peak, that correspond to bandwidth of the ground state, vibrational relaxations and conjugate systems respectively. Their significant change of values in monitoring can reflect richer and deeper information in pollution events for contamination tracing and subsequent processing, such as the category, common properties and functional groups of contaminants. The effectiveness of this feature extraction method was assessed by conducting experiments with sample mixtures of typical chemicals and four real water samples. The results highlight the potential of these features in water pollution early warning and contaminant analysis with richer information and stronger robustness.