Mercury Reduction and Oxidation by Reduced ... - ACS Publications

Nov 22, 2011 - Peng Liao , Wenlu Li , Yi Jiang , Jiewei Wu , Songhu Yuan , John D. Fortner , and ..... Amy B. Gruss , Regina Rodriguez , Christine O. ...
27 downloads 0 Views 893KB Size
ARTICLE pubs.acs.org/est

Mercury Reduction and Oxidation by Reduced Natural Organic Matter in Anoxic Environments Wang Zheng,* Liyuan Liang, and Baohua Gu* Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, United States

bS Supporting Information ABSTRACT: Natural organic matter (NOM)-mediated redox cycling of elemental mercury Hg(0) and mercuric Hg(II) is critically important in affecting inorganic mercury transformation and bioavailability. However, these processes are not well understood, particularly in anoxic water and sediments where NOM can be reduced and toxic methylmercury is formed. We show that under dark anoxic conditions reduced organic matter (NOMre) simultaneously reduces and oxidizes Hg via different reaction mechanisms. Reduction of Hg(II) is primarily caused by reduced quinones. However, Hg(0) oxidation is controlled by thiol functional groups via oxidative complexation, which is demonstrated by the oxidation of Hg(0) by low-molecular-weight thiol compounds, glutathione, and mercaptoacetic acid, under reducing conditions. Depending on the NOM source, oxidation state, and NOM:Hg ratio, NOM reduces Hg(II) at initial rates ranging from 0.4 to 5.5 h1, which are about 2 to 6 times higher than those observed for photochemical reduction of Hg(II) in open surface waters. However, rapid reduction of Hg(II) by NOMre can be offset by oxidation of Hg(0) with an estimated initial rate as high as 5.4 h1. This dual role of NOMre is expected to strongly influence the availability of reactive Hg and thus to have important implications for microbial uptake and methylation in anoxic environments.

’ INTRODUCTION Redox cycling between elemental mercury [Hg(0)] and mercuric species [Hg(II)] is a key process affecting the fate of Hg in natural ecosystems. Aqueous Hg(II) can be transformed to neurotoxic methylmercury (MeHg) by anaerobic microorganisms, which accumulates in the aquatic food web.1,2 Hg(II) can also be reduced to the volatile Hg(0), thereby being removed from solution and becoming unavailable for microbial uptake and methylation.36 Reduction of Hg(II) has been shown to occur via photochemical reduction in oxic surface water,3,7 and by microorganisms 1,4,5,8,9 and particulate minerals 6,10 under suboxic and anoxic conditions. Hg(II) reduction has also been observed in the presence of naturally occurring humic substances (or natural organic matter, NOM) in soil and water.1113 NOM consists of redox reactive but chemically heterogeneous natural organic substances that exist ubiquitously in aquatic and terrestrial environments.1417 In addition to its widely recognized metal binding capabilities (e.g., with Hg(II), Fe(III), and radioactive U(VI)),1821 NOM can mediate electron transfer 2226 and is one of the most important electron shuttles in natural environments.22,27 For example, NOM has been shown to accept electrons from microorganisms and then transfer them to terminal electron acceptors such as Fe(III) and iron oxide minerals.22,28 Among the various functional groups in NOM, quinone/hydroquinone pairs are the dominant redox active moieties,23,29 acting as both electron acceptors and r 2011 American Chemical Society

donors in the oxidationreduction cycle. NOM also forms strong complexes with Hg(II) through reduced sulfur (S) or thiolate functional groups in NOM,1820 and this reaction makes Hg(II) (as Hg(II)-NOM complexes) difficult to be reduced, for example, by stannous chloride (SnCl2).30,31 Of particular interest is the recent finding that reduced NOM can mediate both the reduction and oxidation of Hg in the same anoxic environment,26 although the exact mechanisms by which these reactions occur are not fully understood. It has been suggested that the oxidation involves reactive thiol/thiolate interactions with Hg(0), rather than involving redox active quinone functional groups.26 However, it remains uncertain if and at what rate the dissolved Hg(0) forms adducts directly with the thiol functional groups in NOM. In this study, detailed reaction rates and mechanisms between NOM and Hg(II) or Hg(0) were investigated under controlled dark, anoxic conditions using chemically reduced NOM of different origins. Our primary objective was to provide an in-depth understanding of the NOM-mediated reduction and oxidation of Hg in anoxic environments where Hg speciation and availability are of particular concern to biological Received: September 27, 2011 Accepted: November 22, 2011 Revised: November 21, 2011 Published: November 22, 2011 292

dx.doi.org/10.1021/es203402p | Environ. Sci. Technol. 2012, 46, 292–299

Environmental Science & Technology

ARTICLE

methylation. Model thiol compounds with known chemical structures were used to provide direct evidence of oxidative reactions between thiols and Hg(0) leading to the formation of Hg(II)-thiol complexes. The occurrence of these reactions in the presence of reduced NOM underscores their importance for mercury speciation, microbial uptake, and methylation under anaerobic environments.

’ MATERIALS AND METHODS Four NOM isolates from both aquatic and terrestrial origins were selected for the study. Elliott soil humic acid (ES-HA) and Suwannee River NOM isolates (SR-NOM) were obtained from the International Humic Substances Society (IHSS). Local surface soil at the U.S. Department of Energy’s Integrated Field Research Center (IFRC) site in Oak Ridge, Tennessee, was used to isolate the IFRC-humic acid (IFRC-HA) and fulvic acid (IFRC-FA). General characteristics of these NOM samples (elemental compositions and 13C NMR) are provided in the Supporting Information (SI, Table S1, and Figure S1). For all NOM solution preparations, dissolved organic carbon (DOC) concentration was determined using a Shimazu TOC-5000A total organic carbon analyzer.20,31 Additional details of extraction, purification, and characterization of these NOM samples are given elsewhere.20,25 The methods of Kappler et al. 32 and Peretyazhko and Sposito 33 were followed in the preparation of chemically reduced NOM (NOMre). Briefly, NOM powder samples (as received or prepared) were first dissolved in a phosphate buffer (10 mM, pH 7) at the concentration of about 83 mM C. An aliquot of each NOM suspension was transferred into a large serum bottle. Samples were then continuously purged for 15 min with H2 in the presence of Pd catalysts (5% Pd on alumina powder, 1 g/L, Acros Organics) and allowed to react on a rotary shaker for 24 h at ∼22 °C. Samples were subsequently purged with ultrapure Ar to remove excess H2 and filtered through 0.2-μm Supor membrane filters to remove catalysts and particulate matters or microbial cells if any. The untreated NOM was prepared similarly but without H2/Pd treatment. The untreated NOM hereto is referred to as oxidized NOM (NOMox) since its reducing capacity is typically 1 order of magnitude lower than the chemically reduced NOM.33 All NOMre samples were kept in the dark in an anoxic glovebox (98% N2 and 2% H2) to avoid potential oxidation and photochemical reactions. Reactions between Hg(II) and NOMre were subsequently examined at different DOC:Hg ratios. To ensure that a strictly anoxic condition was maintained during the experiment, all sample preparations including filtration and centrifugation were conducted in the glovebox, and all reagent solutions were deoxygenated. To minimize potential photochemical reactions, samples were prepared and kept in amber vials and protected from light during reactions. A series of NOM working solutions ranging from 0.004 to 0.83 mM C was prepared in the phosphate buffer, and an aliquot of Hg(II) (as HgCl2) added to these working solutions to reach a final concentration of ∼20 nM. All samples were then equilibrated for 4 h to compare the extent of Hg(II) reduction within the reaction time. Our preliminary experiments indicate that Hg(II) reduction by NOM is rapid and usually reaches a pseudoequilibrium within ∼12 h. After equilibration, an aliquot of the reactant solution was taken for total Hg (HgT) analysis, and the rest of the sample purged by ultrapure N2 for 5 min to remove purgeable gaseous

Hg(0) (HgP). The remaining nonpurgeable Hg (HgNP) was analyzed following oxidation by BrCl (0.2 M) at 2.5% (v/v).34 HgT and HgNP were measured following EPA method 1631E 34 using cold-vapor atomic fluorescence spectroscopy (CVAFS) (Tekran 2600, Tekran Instruments, Inc.). The detection limit for Hg is e2.5 pM. HgP was determined as the difference between HgT and HgNP. Reactions between dissolved Hg(0) and NOM were performed similarly as described above using a Hg(0) stock solution prepared according to established procedures.26,35 The concentrations of Hg(0) and Hg(II) in the stock solution were measured before each use; typically dissolved Hg(0) was ∼150200 nM and Hg(II) was lower than 1 nM after 48-h equilibration. Kinetic experiments were also performed between Hg(II) and NOM at selected NOM concentrations of 0.008, 0.042, and 0.42 mM C (or DOC:Hg ratios of 400, 2000, and 20000, respectively) in phosphate buffer in 40-mL amber glass vials. The final Hg concentration was ∼20 nM, and the final solution volume was 40 mL. A subsample of 12 mL was taken from the reactant solution with a syringe (without opening the vial) at desired time intervals and immediately purged by ultrapure N2 for 5 min to remove purgeable Hg(0). The remaining HgNP was stabilized and analyzed as described above. To estimate the mass balance, we monitored the HgT concentration by reacting a subsample with 2.5% (v/v) BrCl (0.2 M), and the recovery of HgT was generally >88% after 48 h. Loss of Hg onto container wall was determined by washing the reaction vessel with 20 mL of BrCl (0.2 M) at 0.5% (v/v) and usually 80%) in the presence of NOMox and in the blank. HgNP was positively correlated with DOC:Hg ratios; the higher NOMre concentration led to increased HgNP in solution. These results provided direct evidence that Hg(0) strongly interacts with NOMre, and that this interaction is likely responsible for the apparent inhibition of Hg(II) reduction at high DOC:Hg ratios (Figure 1a). Hg Reduction and Oxidation Kinetics. Consistent with the extent of Hg(II) reduction in the fixed-time experiments (Figure 1a), the kinetics of Hg(II) reduction also depends on DOC:Hg ratios (Figure 2). For reduced ES-HA, IFRC-HA and SR-NOM, the highest HgP was produced in the presence of 0.042 mM C (DOC:Hg = 2000:1), For IFRC-FA, the highest Hg(II) reduction occurred at ∼0.42 mM C (Figure 2). In all experiments, Hg(II) reduction proceeded rapidly and dominated within the first 15 min (Figure 2). Thereafter, Hg(II) reduction continued with notably slower rates in the presence of SR-NOM and IFRC-FA. For ES-HA and IFRC-HA, the reduction stopped, and the trend was reversed, indicating that the competing, oxidative reactions have led to lower amounts of Hg(0) production. For example, in the presence of 0.042 mM C IFRC-HA (Figure 2d), nonpurgeable Hg(II) decreased from 100% to 30% in 15 min, but subsequently increased to 45% over the next 5-h reaction period. Therefore, the results obtained after 4 h in the fixed-time experiments (Figure 1a) reflect the net balance between Hg(II) reduction and the competing reactions. This competing reaction is attributed to the interaction between dissolved Hg(0) and NOMre as shown in Figure 1b.

Figure 1. (a) Reduction of Hg(II) (as HgCl2 at ∼20 nM) and (b) oxidation of dissolved Hg(0) (∼15 nM) by NOM as a function of DOC: Hg molar ratio after 4-h equilibration under dark, anoxic conditions. CHgNP/C0 is the ratio of nonpurgeable Hg (HgNP) to the initial total Hg. In the legend, “re” denotes chemically reduced NOM by H2/Pd treatment, and “ox” is untreated or oxidized NOM. Blank is the phosphate buffer (10 mM) at pH 7 treated the same way as the reduced NOM. Legends apply to both figures. Error bars represent 1 SD calculated based on eq 1.

HgNP by reaction with NOMre indicates that Hg(II) was reduced to Hg(0) and this was confirmed by direct purging and analysis of dissolved elemental Hg(0) in solution. However, further increase in NOMre concentration inhibited production of HgP, suggesting that reduction of Hg(II) occurs only when the DOC:Hg ratio is lower than a certain threshold for each NOM used (i.e., DOC:Hg ratio is e20 000 for IFRC-FA, and it is e5000 for other NOM samples) (Figure 1a). In general, humic acids (ES-HA and IFRC-HA) exhibited higher reducing capacities than the IFRC-FA. Approximately 70% of Hg(II) was reduced in the presence of 0.1 mM C of each humid acid (DOC: Hg = 5000:1). IFRC-FA showed the lowest reducing capacity, with a relatively large amount (∼0.42 mM C) required to reduce 60% of the Hg(II). As expected from our previous study,26