Metal Ion Speciation and Dissolved Organic Matter Composition in Soil Solutions

Abstract:

Knowledge of the speciation of heavy metals and the role of dissolved organic matter (DOM) in soil solution is a key to understand metal mobility and ecotoxicity. In this study, soil column-Donnan membrane technique (SC-DMT) was used to measure metal speciation of Cd, Cu, Ni, Pb, and Zn in eighteen soil solutions, covering a wide range of metal sources and concentrations. DOM composition in these soil solutions was also determined. Our results show that in soil solution Pb and Cu are dominant in complex form, whereas Cd, Ni and Zn mainly exist as free ions; for the whole range of soil solutions, only 26.2% of DOM is reactive and consists mainly of fulvic acid (FA). The metal speciation measured by SC-DMT was compared to the predicted ones obtained via the NICA-Donnan model using the measured FA concentrations. The free ion concentrations predicted by speciation modelling were in good agreement with the measurements. Diffusive gradients in thin-films gels (DGT) were also performed to quantify the labile metal species in the fluxes from solid phase to solution in fourteen soils. The concentrations of metal species detected by DGT were compared with the free ion concentrations measured by DMT and the maximum concentrations calculated based on the predicted metal speciation in SC-DMT soil solutions. It is concluded that both inorganic species and a fraction of FA bound species account for the amount of labile metals measured by DGT, consistent with the dynamic features of this technique. The comparisons between measurements using analytical techniques and mechanistic model predictions provided mutual validation in their performance. Moreover, we show that to make accurate modelling of metal speciation in soil solutions, the knowledge of DOM composition is the crucial information, especially for Cu; like in previous studies the modelling of Pb speciation is not optimal and an updated of Pb generic binding parameters is required to reduce model prediction uncertainties.