Ultraviolet-visible and fluorescence spectra indicate the binding and transformation properties of hexavalent chromium in DOM solution

Abstract

Dissolved organic matter (DOM) with complex functional groups affects the trace of binding and transformation of hexavalent chromium (Cr(VI)) in the environment. However, the mechanisms and quantitative indicators are still ambiguous. In this study, the perspectives of spectroscopy with synchronous and excitation-emission matrix (EEM) fluorescence, and ultraviolet-visible (UV–vis) were employed to probe the interaction of DOM and Cr(VI). The spectroscopic parameters with Dslope240–280 (the slope of the log-transformed absorbance in the range of 240–280 nm), DSlope325–375 (the slope of the log-transformed absorbance in the range of 325–375 nm), SR (the ratio of the slope of the log-transformed absorption coefficient with 275–295 nm to 350–400 nm), SFI (specific fluorescence intensities), AvgFI/AvgUVA (the ratio of average fluorescence to average UV), and SS (Stokes shift) of UV–vis and fluorescence were extracted to characterize the DOM properties. The UV–vis results showed carboxylic-like and total phenolic/carboxylic-like groups were contributed to the complexation of Cr, which resulted in forming macromolecular aggregates, however, the degree of participation of different functional groups was uneven and the binding sequence depended on the types of fractions especially the ratio of protein. Meanwhile, the EEM results proved interaction between DOM and Cr also affected the hydrophobicity and quantum yield of solutions. In addition, according to the change orders of substances and functional groups by Two-dimensional correlation spectroscopy (2DCOS) analysis, aromatic amino acids substances were much more sensitive binding onto Cr than fulvic-like fractions and humic-like fractions, and the mechanisms for Cr(VI) retention were developed. Statistical analysis results showed that SR and SFI were significantly correlated with the concentration of Cr(VI), and multiple linear regression models were developed to predict concentrations of Cr(VI). Moreover, it was found the goodness of fit of SFI was better than that of SR at low Cr concentrations, but opposite at high concentrations.