Interpretation of heavy metal speciation in sequential extraction using geochemical modelling

Abstract

Environmental context Heavy metal pollution is a worldwide environmental concern, and the risk depends not only on their total concentration, but also on their chemical speciation. Based on state-of-the-art geochemical modelling, we pinpoint the heavy metal pools approached by the widely used sequential extraction method. The finding of this paper can help users of sequential extraction methods to better interpret their results. Abstract In this study, the metal (Cd, Cu, Zn and Pb) fractionation determined by selective sequential extraction (SSE) was compared with metal speciation calculated using a geochemical model, the Multi-Surface Model (MSM). In addition, the sources of Cd, Cu and Zn extracted in the SSE were identified with the help of the modelling. The results showed that the SSE-based Cd fractionation contradicted the modelled results, with the organic-bound Cd as respectively the least and the most important species. This contradiction was explained by the model and was attributed to the weak specific adsorption of Cd to organic matter; For Cu, a good agreement was found between SSE and model fractionation, both recognising organic-bound Cu as the most dominant fraction. The high affinity of organic matter for Cu reduced the degree of Cu extracted in steps preceding the oxidation step. The SSE measured a larger exchangeable Zn fraction than the model predicted, which could be explained by Zn extracted from organic-bound, oxide-bound forms, and certain rapidly dissolvable Zn-minerals if present. Zinc in the micropores of minerals was probably not extracted in 0.43M HNO3, thus was not included in the modelling for adsorption calculation, which could explain to a certain extent the larger amount of oxide-bound Zn determined in the SSE than calculated in the model. The modelling results for Pb were less reliable than for other metals because of a poor accuracy of Pb concentration in solution predicted. The findings of this paper can help users of the sequential extraction methods to better interpret their results.