Mercury sorption to sediments: Dependence on grain size, dissolved organic carbon, and suspended bacteria

Abstract

A combination of laboratory scale derived correlations and measurements of grain size distribution, DOC (dissolved organic carbon) concentration, and density of suspended bacteria promises to be useful in estimating Hg(II) sorption in heterogeneous streambeds and groundwater environments. This was found by shaking intact sediment and fractions thereof (<63–2000 μm) with solutions of HgCl₂ (1.0–10.0 ng ml⁻¹). The intact sediment was also shaken with the Hg(II) solutions separately in presence of DOC (6.5–90.2 μg ml⁻¹) or brought in contact with suspensions of a strain of groundwater bacteria (2 × 10⁴–2 × 10⁶ cells ml⁻¹). Hg(II) sorption was rather weak and positively correlated with the grain size, and the sorption coefficient (K_d) varied between about 300 and 600 ml g⁻¹. By using the relative surface areas of the fractions, K_d for the intact sediment was back calculated with 2% deviation. K_d was negatively correlated with the concentration of DOC and positively correlated with the number of bacteria. A multiple regression showed that K_d was significantly more influenced by the number of bacteria than by the grain size. The findings imply that common DOC concentrations in groundwater and streambeds, 5–20 μg ml⁻¹, will halve the K_d obtained from standard sorption assays of Hg(II), and that K_d will almost double when the cell numbers are doubled at densities that are common in aquifers. The findings suggest that simultaneous measurements of surface areas of sediment particles, DOC concentrations, and bacterial numbers are useful to predict spatial variation of Hg(II) sorption in aquifers and sandy sediments.

Introduction

Concern about the distribution and fate of mercury (Hg) and the likelihood that it reaches health risk levels in food and drinking water has directed research towards understanding Hg transport in soil and water (DeMarco et al., 2006, Kongchum et al., 2006). Models can describe Hg transport by advective–dispersive equations modified to account for retention processes, such as sorption (Yin et al., 1997a, Sarkar et al., 1999, Carrol et al., 2000, Kim and Corapcioglu, 2002) and sedimentation (Carrol et al., 2000, DiLeonardo et al., 2006). One of the major challenges for this approach is the spatial variability in hydraulic, geologic, and geochemical parameters. Of particular importance is the heterogeneity of the sorption capacity of the geological material, which may vary by an order of magnitude at the centimetre scale for nonionic organic compounds in sediments (Elabd et al., 1986).

Characteristics of soil and sediment particles, such as mineral composition, clay and organic matter content, and metal coatings (Schlüter, 1997), influence sorption of Hg. Minerals, such as quartz and feldspar, which are predominantly found in medium to coarse particle sizes (Barber et al., 1992), have a permanent negative charge and may retain cationic Hg by electrostatic forces. Likewise, temporary and patchy cover of mineral surfaces by anions, such as hydroxyl and sulphate ions, clay minerals, oxyhydroxides of Al, Fe, and Mn (Coston et al., 1995, Brown et al., 1999), and soil organic carbon (SOC) (Wen et al., 1998, Brown et al., 1999), retains Hg by electrostatic forces and complexation (Yin et al., 1996). Furthermore, the size of the solid particle may affect sorption and concentration of Hg (Fukue et al., 2006), since smaller sized particles are mainly composed of weathering-resistant, net positively charged minerals of Fe and Al (Coston et al., 1995).

The sorption of mercury to natural soils and mineral surfaces is affected by particulate (POC) and dissolved organic carbon (DOC) due to their strong adsorption affinity for Hg (Yin et al., 1996, Ravichandran, 2004, Jing et al., 2007). DOC sorbs to mineral surfaces of Al and Fe oxides-hydroxides by either electrostatic attraction or ligand exchange between carboxyl and hydroxyl groups of DOC and OH and OH₂ groups of the surface minerals (Kaiser and Zech, 1998), and when SOC dissolves, mainly at higher pH, it forms DOC–Hg complexes (Yin et al., 1996, Schlüter, 1997).

Likewise, the presence of bacteria suspended in the aqueous phase or sorbed to the solid can influence the sorption of Hg. Most bacteria have an overall anionic cell surface at common environmental pH’s as a result of the pK_a’s of carboxyl (4–6) and phosphoryl (∼7) groups of peptidoglycans, lipopolysaccharides, and phospholipids in the cell wall (Schiewer and Volesky, 2000). These ligands form complexes with metals (Daughney et al., 2002). The high sorption capacity is one of the characteristics that make bacteria a valuable sorbent for removal of metals from wastewater and polluted areas (Chen and Wilson, 1997, Schiewer and Volesky, 2000), e.g. where mercury is released from gold extraction.

One of the challenges in the development of methodologies to account for spatially variable flow and sorption parameters is to identify correlations between flow and sorption parameters and between sorption coefficients and geologic and geochemical parameters. With the latter, laboratory scale derived correlations at hand, laborious preparations of sorption isotherms may be replaced by indirect methods allowing for more rapid screening of field samples. With this in mind, a study was designed to establish quantitative relationships between sorption of Hg(II) to a saturated sediment and the size distribution of the sediment particles, DOC concentration, and cell density of bacteria.

We hypothesized that sorption of Hg(II) to the solid phase would be (i) negatively related to the particle size since cation exchange was assumed to be an important sorption mechanism, with smaller particles sizes having greater capacity than larger particles; (ii) negatively related to DOC concentrations since Hg(II) was assumed to have high affinity for DOC; (iii) negatively related to the density of suspended bacteria as a result of Hg(II) complexation at net negatively charged cell surfaces of suspended bacteria.