PII S0016-7037(99)00285-9 Speciation, reactivity, and cycling of Fe and Pb in a meromictic lake MARTIAL TAILLEFERT, 1,3, *CHARLES-PHILIPPE LIENEMANN, 2,4 JEAN-FRAN¸ COIS GAILLARD, 1 and DIDIER PERRET 2 1 Department of Civil Engineering, Northwestern University, Evanston, Illinois 60208-3109, USA 2 Institut de Chimie Mine ´rale et Analytique, Universite ´ de Lausanne, CH-1015 Lausanne, Switzerland 3 College of Marine Studies, University of Delaware, Lewes, Delaware 19958, USA 4 Laboratoires WOLFF Environnement, F-69500, Bron, France (Received December 23, 1998; accepted in revised form June 21, 1999) Abstract—A suite of analytical techniques were combined to study the chemical speciation of Fe and Pb in the water column of a lake characterized by a biogenic meromixis (Paul Lake, MI). Depth profiles of Fe 2+ and “dissolved” Pb display significant concentration gradients below the chemocline, i.e., they increase from below detection limit to ca. 100 M for Fe 2+ and 2 nM for Pb d . Significant correlations between particulate organic matter, hydrous iron oxides, and particulate Pb suggest that Pb is scavenged by Fe-rich particles formed at the oxic-anoxic transition. Transmission electron microscopy shows that particles of hydrous iron oxides form complex aggregates with natural organic matter at and below the oxic-anoxic transition. Experiments with batch reactors show that these organo-mineral moieties remove Pb rapidly during their formation. Thermodynamic calculations predict that FeS and PbS are respectively saturated and oversaturated in the monimolimnion, although the presence of neither FeS nor PbS was observed. This suggests that the solubilities of Fe and Pb are influenced by complexation. Voltammetric experiments on filtered samples show that Pb is weakly complexed in the mixolimnion and strongly complexed in the monimolimnion. A conditional stability constant for Pb complexation is determined using metal titration curves assuming a simple 1:1 stoichiometry and gives logK cond = 9.4  0.8 M -1 in the monimolimnion. These speciation results are confirmed by ion exchange chromatography, which demonstrates that more than 98% of Pb is complexed by natural organic matter. Copyright © 2000 Elsevier Science Ltd 1. INTRODUCTION Lead is a toxic pollutant (Nriagu, 1990; Needelman, 1990; Goyer, 1993; Nriagu et al., 1996a), and its presence in aquatic systems primarily results from anthropogenic activities (Schaule and Patterson, 1981; Boyle et al., 1986; Nriagu and Pacyna, 1988; Ritson et al., 1994; Veron et al., 1994; Erel and Patterson, 1994). Lead enters lakes from rivers, atmospheric deposition, soil leachates, and groundwater seepage. The con- stant presence of oxygen in oligotrophic lakes and the high reactivity of lead with oxide particles ensures its uptake from the water column, resulting ultimately in its burial in the sediment (Sigg et al., 1987; Ritson et al., 1994; Nriagu et al., 1996b). In contrast, the cycling of lead in the anoxic bottom waters of eutrophic lakes is controversial (Hamilton-Taylor and Davison, 1995). Lead has been shown to be remobilized during the dissolution of redox sensitive particles (mainly hydrous iron oxides) in the water column (Balistrieri et al., 1992a) and at the sediment-water interface (SWI) (Hamilton-Taylor et al., 1984; White and Driscoll, 1985; Benoit and Hemond, 1990; Taillefert et al., 1997; Viollier, 1995). Simultaneously, it can be removed from anoxic waters by precipitation with sulfides (Sigg, 1985; Frevert, 1987; Balistrieri et al., 1994) or by adsorption onto FeS (Davison et al., 1992; Morse and Arakaki, 1993). However, lead removal was not evident in many systems where the ionic activity product (IAP) exceeds the solubility product of PbS. Benoit and Hemond (1990) summarize the possible explana- tions for such behavior: (i) Pb precipitates in all environments but in some waters the particles formed are too small to be retained on filters; (ii) Pb is complexed by dissolved ligands that outcompete sulfide complexation and precipitation; (iii) freshly precipitated PbS may have a higher K sp than well crystallized minerals; and (iv) the precipitation may be hin- dered kinetically. Because of this behavior and because Pb is toxic, it is necessary to determine its chemical speciation in the dissolved phase and to assess its biogeochemical cycle in lakes. It is generally accepted that trace elements are most effi- ciently removed at oxic-anoxic transitions by organic or inor- ganic colloidal particles, which have high specific surface areas (Tipping, 1981; Davis, 1984). It has also been shown that natural organic matter (NOM) and hydrous oxides can interact; however, the nature of this interaction has been subject to conflicting interpretations. Some authors believe that in natural systems hydrous iron oxides are coated by low molecular weight NOM, thus modifying their surface reactivity (Stumm et al., 1980; Tipping, 1981; Davis, 1984; Zhou et al., 1994; Gu et al., 1995). This idea is supported by laboratory experiments, which demonstrated that adsorptive properties are changed in the presence of dissolved NOM. In contrast, microscopic in- vestigations (Fortin et al., 1993; Perret et al., 1994; Tessier et al., 1996; Lienemann et al., 1999) have showed that under natural conditions hydrous oxides and high molecular weight NOM form intimate structures. The aggregation of these enti- ties leads to a more complicated structure in which trace elements may not only be scavenged at the surface, but also embedded in the mineral-organic moieties (Laxen and Sholko- vitz, 1981; Laxen, 1984, 1985). The chemical speciation of iron and lead in the water column and the scavenging of lead by hydrous iron oxides can be calculated by thermodynamic equilibrium models that use sur- *Author to whom correspondence should be addressed (mtaillef@udel.edu). Pergamon Geochimica et Cosmochimica Acta, Vol. 64, No. 2, pp. 169 –183, 2000 Copyright © 2000 Elsevier Science Ltd Printed in the USA. All rights reserved 0016-7037/00 $20.00 + .00 169