Diagnosis of the unexpected fluorescent contaminants in quantifying dissolved organic matter using excitation-emission matrix fluorescence spectroscopy

Abstract

Dissolved organic matter (DOM) is widely present in aqueous environments and plays a significant role in pollutant mitigation and transformation. So far, excitation-emission matrix (EEM) fluorescence spectroscopy coupled with parallel factor analysis (PARAFAC) has been widely applied to quantify fluorescent DOM. However, this approach fails to provide accurate concentration of DOM when fluorescent contaminants exist. In this work, a new method, prior linear decomposition (PLD), is developed to solve this problem by introducing prior information, i.e., EEMs of DOM, into data decomposition. First, EEM of humic acid (HA) with different numbers of random Gaussian peaks are tested to confirm the robustness of PLD. The percentages for the relative errors within 5% are found to be 97.7% and 69% using PLD and PARAFAC, respectively. Then, the determination of mixture of HA with several contaminants is performed, validating the feasibility of DOM quantification and capability of contaminant diagnosis using PLD for synthetic water samples. Finally, DOM-containing natural water samples collected from a polluted lake, river and wastewater treatment plant (WWTP) are measured. The testing results confirm that PLD provides an accurate result with less evaluated error than PARAFAC and the EEMs of the contaminants can be inferred precisely. This work clearly demonstrates that PLD offers a robust approach for quantifying fluorescent DOM, which is of great significance in both natural and engineered aqueous environments.