Antimony(V) reaction with particulate natural organic matter: Sorption behavior, binding mechanism, and environmental implications

Abstract

Antimony (Sb) is a metalloid extensively used in industrial products with a high potential for accumulation in organic-rich soils. Under oxic conditions, dissolved Sb occurs in its pentavalent oxidation state, Sb(V), but little is known about sorptive interactions between Sb(V) and natural organic matter (OM). Therefore, we conducted batch experiments with Sb(V) and particulate OM (POM) of peat as a model sorbent for particulate soil OM. We varied reaction time (1–21 d), concentrations of Sb(V) (1–218 µM) and POM (1–5 g L⁻¹), pH (1.5–10.5), as well as type (NaCl and CaCl₂) and concentration of the background electrolyte (no salt, 0.01 and 0.1 M). Additionally, we explored the molecular mechanism of Sb(V) binding to POM at pH ≤ 4 using X-ray absorption spectroscopy. Based on this information, we modelled Sb(V) sorption data with the Stockholm Humic Model (SHM). Sorption of Sb(V) to POM did not cause Sb reduction and decreased from 109 mmol kg⁻¹ at pH 3 to 72.4 mmol kg⁻¹ at pH 5 (1 g L⁻¹ POM, 0.01 M NaCl). Although sorption maxima were found at pH 1.8–2.8, up to ∼ 10 % of total Sb(V) was still removed from solution at pH 6. An increase in POM concentration, ionic strength, and the presence of Ca²⁺ promoted Sb(V) sorption. The high Sb(V) sorption capacity of POM was associated with a low binding affinity. Sorption kinetics of Sb(V) were slow and generally showed bi-phasic patterns. Apparent half-life times of the fast and slow sorption process at pH 4 and 5 were on average 12.7 and 117 h, respectively. The slowly sorbing Sb(V) was ascribed to diffusion of Sb(V) into net-negatively charged POM particles. X-ray absorption spectroscopy analyses showed 6.3 ± 0.4 O atoms at 1.98 ± 0.01 Å (mean ± SD), followed by 2.1 ± 0.4 C atoms at 2.82 ± 0.05 Å, implying pentavalent Sb in octahedral coordination and bidentate complexation by polycarboxylic, hydroxy-carboxylic, and/or polyol ligands in 5- or 6-membered ring structures. The SHM accurately described Sb(V) sorption edges and sorption isotherm data collected at pH 3–5. SHM calculations implied that Sb(V) sorption to POM may be quantitatively relevant even at pH > 7 and impairs the precipitation of Ca-antimonates at acidic pH. We also predicted competition effects showing that Al, Pb, and MoO₄²⁻ can cause substantial desorption of Sb(V) when present in excess over Sb(V). Model predictions also indicated that Sb(V) complexation by POM in organic soils becomes negligible in presence of > 1 wt% metal oxides. Overall, our results show that although POM has an enormous potential for Sb(V) sequestration over a broad pH range, competing ions and mineral sorbents may strongly decrease Sb(V) binding by POM. This implies that Sb(V) binding to POM is only relevant in oxic metal-poor soil environments such as ombrotrophic peats, organic surface layers, and OM-rich microenvironments of mineral soils.