Similarities between Inorganic Sulfide and the Strong Hg(II

Perrine Dranguet , Vera I. Slaveykova , Séverine Le Faucheur ... Scott C. Brooks , Carrie L. Miller , Jennifer J. Mosher , Xiangping L. Yin , Meghan ...
0 downloads 0 Views 150KB Size
Download PDF
Environ. Sci. Technol. 2005, 39, 4035-4041

Similarities between Inorganic Sulfide and the Strong Hg(II)-Complexing Ligands in Municipal Wastewater Effluent HEILEEN HSU-KIM† AND DAVID L. SEDLAK* Department of Civil & Environmental Engineering, University of CaliforniasBerkeley, Berkeley, California 94720

Municipal wastewater effluent contains ligands that form Hg(II) complexes that are inert in the presence of glutathione (GSH) during competitive ligand exchange experiments. In this study, the strong ligands in wastewater effluent were further characterized by comparing their behavior with sulfide-containing ligands in model solutions and by measuring their concentration after exposing them to oxidants. The strong Hg(II) complexes in wastewater effluent and the complexes formed when Hg(II) was added to S(-II) were retained during C18 solid-phase extraction (SPE) and did not dissociate in the presence of up to 100 µM GSH. In contrast, Hg(II) complexes with dissolved humic acid were hydrophilic and dissociated in the presence of GSH. The combination of sulfide and humic acid resulted in formation of Hg(II) complexes that were inert to GSH and were only partially retained by C18SPE, indicating that NOM interacted with the Hg-sulfide complexes. When wastewater effluent samples and model solutions of free sulfide, Zn-sulfide, and Fe-sulfide were exposed to 0.14 mM NaOCl for 1 h (to mimic conditions encountered during chlorine disinfection), the strong Hg(II)-complexing ligands were completely removed. Exposure of the wastewater effluent and the model ligands to oxygen for 2 weeks resulted in approximately 60% to 75% loss of strong ligands. The strong ligands that remained in the oxygen-oxidized samples were resistant to further oxidation by chlorine, indicating that oxidation of S(-II) results in the formation of other sulfur-containing ligands such as S8 that form strong complexes with Hg(II).

Introduction Previous studies of Hg(II) and monomethylmercury (MMHg) speciation in surface waters have focused on complexation of Hg by reduced sulfur-containing ligands in natural organic matter (NOM) (1-5). Humic substances isolated from surface waters contain a relatively small amount of thiol functional groups that should form complexes with Hg(II) that are stronger than those with carboxylate, phenolate, and nitrogencontaining ligands (2-4). Other sulfur-containing ligands such as HS- and polysulfides also may play a role in dissolved Hg(II) speciation, but their importance has not been established because they are difficult to study in oxic waters. In a previous study (6), we employed a competitive ligand exchange (CLE) method to detect strong Hg(II)-complexing * Corresponding author telephone: (510) 643-0256; fax: (510) 642-7483; e-mail: [email protected]. † Current address: University of Delaware, College of Marine Studies, 700 Pilottown Rd., Lewes, DE 19958. 10.1021/es050013i CCC: $30.25 Published on Web 04/23/2005

 2005 American Chemical Society

ligands in municipal wastewater effluent. The strongest Hg(II) complexes that formed in wastewater effluent were inert to the competing ligand glutathione (GSH). Conditional stability constants, cKHgL, for these complexes were greater than 1030 (for the reaction Hg2+ + L ) HgL). This relatively high value for cKHgL suggests that the ligands were not sulfurcontaining functional groups on humic substances, which have lower cKHgL values (1018.4-1025.5 (2, 3)). The ligands could contain inorganic sulfide, which forms Hg(II) complexes with conditional stability constants that are much higher than those for thiols (3, 7). Inorganic sulfide is unstable in the presence of oxygen (8) and on the basis of thermodynamics should not have been present in the wastewater effluent. Nonetheless, inorganic sulfide has been detected in wastewater effluent and oxic surface waters at concentrations ranging from 1 to 100 nM (9-14), which is greater than typical Hg(II) and MMHg concentrations (0.001-0.1 nM) (15). Research with other trace metals has suggested that at these concentrations, inorganic sulfide should control the speciation of soft metals such as Ag and Cu (9, 14, 16). In oxic surface waters, inorganic sulfide is expected to persist by complexation with metals, which slows oxidation by oxygen and prevents volatilization of H2S (16, 17). Therefore, inorganic sulfide in oxic surface waters and wastewater effluent is likely to consist of complexes with metals such as Fe, Cu, and Zn rather than free sulfide (i.e., HS- and H2S) (14). The contribution of Hg to contaminated watersheds from municipal wastewater treatment plants (WWTPs) is typically small compared to other surface water sources (18, 19). Nevertheless, WWTPs often are pressured to decrease Hg concentrations in their effluents, particularly in locations where the effluent discharge occurs upstream of Hgcontaminated ecosystems. A better understanding of Hg speciation in wastewater effluent may lead to modifications in the treatment process that improve Hg removal efficiencies. Wastewater effluent also serves as a source of strong Hg(II)-complexing ligands, such as inorganic sulfide and thiols (6, 9, 13), which may alter Hg(II) speciation in surface waters downstream of effluent discharge points. As a result, the complexation of Hg(II) by these strong ligands can greatly affect the mobility of Hg in the downstream watershed and the formation of monomethylmercury (7, 20-25). In this study, the strong Hg(II)-complexing ligands in wastewater effluent were characterized by comparing the complexation of Hg(II) in wastewater effluent to model solutions containing inorganic sulfide and dissolved humic acid. The CLE speciation method was employed to measure the concentration of strong Hg(II) complexes in wastewater effluent and model solutions before and after oxygen and chlorine oxidation. Results of this study provide insight into the identity of the strong ligands and their behavior in wastewater and surface water systems.

Experimental Methods Materials. All chemicals and materials were purchased from Fisher Scientific and were ACS-reagent grade, unless otherwise noted. Stock solutions were prepared by dissolving chemicals in 18 MΩ cm ultrapure water (Milli-Q, Millipore). Trace-metal-grade acids were used for pH adjustments and for acid-cleaning of equipment. Sulfide stock solutions were prepared daily by dissolving washed crystals of Na2S‚9H2O in Milli-Q water that had been sparged for at least 60 min with N2 (ultrahigh purity). Stock solutions of 4.0-8.0 mM GSH (Sigma-Aldrich) were prepared for GSH CLE experiments and stored for less than 1 week at 4 °C. VOL. 39, NO. 11, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4035

Reagents for total Hg analysis were prepared according to procedures outlined in EPA method 1631 (26). Hg(II) calibration standards were made by diluting a Sigma-Aldrich ICP/AA Hg(II) standard (997 mg Hg(II) per liter in 10% HNO3) to 5.0 µg/L Hg in 0.5% (v/v) bromine monochloride. Wastewater effluent samples and synthetic samples were amended with Hg(II) using a stock solution of 10 µM Hg(NO3)2 in 0.1 N HNO3. All samples were collected and stored in glass bottles with Teflon-lined screwcaps. Bottles, tubing, and sampling equipment were cleaned first by an overnight soak in diluted Micro detergent (Cole Parmer) followed by an overnight soak in 1 N HCl. Wastewater Effluent Samples. Wastewater effluent samples were collected from two municipal wastewater treatment plants, referred as WWTP 1 and WWTP 2, located in the San Francisco metropolitan area. WWTP 1 is a 167 MGD (7.3 m3 s-1) treatment plant that utilizes activated sludge for secondary treatment followed by disinfection with chloramines. Samples from WWTP 1 were collected after secondary clarification (referred to as predisinfection) and after chlorination/dechlorination (referred to as final effluent). WWTP 2 is a 1.7 MGD (0.07 m3 s-1) treatment plant that employs trickling biofilters for secondary treatment followed by ammonia removal and UV disinfection. Samples from WWTP 2 were collected only at the plant discharge point, after UV disinfection. All samples were filtered within 2 h of collection with an acid-cleaned 0.2-µm nylon capsule filter (Whatman Polycap AS) fitted with a peristaltic pump and polyethylene tubing. The pH of the filtered wastewater samples ranged from 7.4 to 7.6 at both treatment plants. Competitive Ligand Exchange-Solid-Phase Extraction (CLE-SPE). The formation of strong Hg(II) complexes in the filtered wastewater effluent samples was measured using competitive ligand exchange with GSH as the competing ligand for Hg(II) (6). Effluent samples were amended with approximately 1 nM Hg(II) and allowed to equilibrate for 5 h to 1 week. After equilibration, GSH was added to aliquots of the sample and equilibrated for approximately 1 h prior to separation by C18 solid-phase extraction (SPE). In wastewater effluent, most of the added Hg(II) forms hydrophobic complexes that are retained by C18 SPE (6). Upon addition of GSH at pH 7-8, Hg(II) forms HgH2(GSH)22- complexes, which pass through the C18 column and are measured in the hydrophilic fraction. To quantify conditional stability constants cKHgL, GSH titrations were conducted by varying the GSH concentration from 0.5 to 100 µM. Hg(II) complexes in samples that did not dissociate in the presence of 20-100 µM GSH had cKHgL values greater than 1030 (6). The concentration of strong Hg(II)-complexing ligands was quantified by conducting Hg(II) titrations at a fixed GSH concentration. This method is similar to previously described CLE methods for the quantification of ligands of other trace metals (e.g., refs 27-30). Aliquots of the wastewater effluent samples were amended with 0.8-500 nM Hg(II) and equilibrated for 40-48 h. After equilibration, 14 µM GSH was added to the sample and the mixture equilibrated for an additional hour prior to separation of the hydrophilic HgH2(GSH)22complexes by C18-SPE. The Hg titration provides a means for quantifying the strong ligand concentration because Hg(II) added to the samples preferentially forms hydrophobic complexes with the strong ligands. When all of the strong ligands are saturated with Hg(II), the excess Hg(II) forms hydrophilic complexes with GSH. The concentration of hydrophilic Hg(II) (i.e., HgH2(GSH)22-) increases by an equal amount with added Hg(II), resulting in a slope of unity (when the data are plotted in the format shown in Figure 2). A titration resulting in 1:1 slope starting at the origin indicates the absence of strong Hg(II) complexing ligands (solid line in Figure 2). Previous researchers have estimated the 4036

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 11, 2005

concentration of strong ligand by extrapolating the 1:1 portion of the titration to the x-axis intercept (30). In this study, the strong ligand concentration was calculated by subtracting the hydrophilic Hg concentration from total Hg concentration at each data point in the 1:1 portion of the titration. The strong ligand concentration is reported as the mean of the data in each titration. Hg(II) Complexation in Model Solutions. The Hg(II)complexing ligands in the wastewater effluent were compared to model solutions containing either inorganic sulfide, dissolved peat humic acid (International Humic Substances Society standard), or a combination of both inorganic sulfide and peat humic acid. All solutions were prepared in airsaturated Milli-Q water containing 20 mM KNO3 and 3 mM 3-N-morpholinopropansulfonic acids buffer (MOPS, Sigma Aldrich). Ammonium hydroxide (trace metal grade) was used to adjust the pH to 7.3. The three model solutions contained the following components (added to the pH 7.3 water in the order as listed): 1 nM Hg(II) + 1 nM S(-II), 1 nM Hg(II) + 15 mg/L humic acid, and 8 nM Hg(II) + 8 nM S(-II) + 15 mg/L humic acid. After allowing Hg(II) to equilibrate for at least 12 h, Hg(II) complexation was measured by conducting GSH CLE titrations on aliquots of each sample followed by extraction of HgH2(GSH)22- complexes with C18-SPE. Oxidation Studies. Oxidation of the strong ligands by oxygen and chlorine was studied in the wastewater effluent samples and in model solutions containing inorganic sulfide. O2-oxidation experiments were conducted by storing filtered wastewater effluent samples for up to 2 weeks in 4-L amber glass screwcap bottles with approximately 1 L of air-filled headspace. After storage of the sample (i.e., the oxygenoxidation period), the concentration of Hg(II)-complexing ligands was quantified by Hg(II) titration and GSH CLE. Chlorine oxidation of the wastewater effluent was conducted by adding 10 mg/L Cl2 (0.14 mM) and 100 mg/L Cl2 (1.4 mM) (added as NaOCl) to freshly collected wastewater effluent samples. Aliquots of oxygen-oxidized effluent samples also were chlorinated for 1 h with 10 mg/L Cl2. After 1 h, all chlorinated effluent samples were dechlorinated with 0.4 mM or 4 mM ascorbic acid prior to addition of Hg(II). Control samples of unchlorinated wastewater effluent indicated that the addition of 4 mM ascorbic acid did not alter Hg(II) speciation. Oxidation studies also were conducted with model sulfide solutions containing HS-/H2S, Zn-sulfide, and Fe-sulfide complexes. These solutions were prepared by dissolving 4-8 nM Na2S‚9H2O in air-saturated Milli-Q water containing 20 mM KNO3, buffered to pH 7.4 by 3 mM MOPS. In the chlorination experiments, the model solutions were buffered to pH 7.6 with 10 mM phosphate, rather than MOPS buffer. S(-II) speciation in the wastewater effluent samples was expected to be dominated by metals such as Fe and Zn (14). Therefore, model S(-II) solutions containing Fe-sulfide and Zn-sulfide complexes were prepared for the oxidation studies. To produce the Fe-sulfide and Zn-sulfide solutions, Fe(III) (as FeCl3) and Zn(II) (as Zn(NO3)2) were added to aliquots of the model S(-II) solutions at concentrations equal to the initial S(-II) concentration. After the preparation of the model sulfide solutions, aliquots were decanted into individual 60-mL glass bottles and capped with Teflon-lined screwcaps. Bottles were filled to minimize headspace (to prevent off-gassing of H2S). Oxidation by oxygen was conducted by storing the airsaturated S(-II) solutions in the dark for up to 4 weeks until the addition of Hg(II). To test the stability of S(-II) during chlorination, separate aliquots of the model S(-II) solutions were chlorinated with 10 mg/L Cl2 (0.14 mM Cl2 added as NaOCl) for 1 h and dechlorinated with 0.4 mM ascorbic acid. The concentration of strong ligands remaining in the oxidized model solutions was measured by converting the

ligands to hydrophobic Hg-sulfide complexes. The exchange was initiated by addition of 9-12 nM Hg(II), which was a slight excess relative to the initial S(-II) concentration. Stability constants for HgS complexes are orders of magnitude larger than those for Fe- and Zn-sulfide complexes (3133). Therefore, Hg(II) was expected to replace Fe and Zn to form HgS complexes, which are retained by C18-SPE, even in the presence of 1-20 µM GSH (shown in Figure 1b). After allowing the Hg(II) to equilibrate in the model solutions for 40-48 h, 6 µM GSH was added and equilibrated for 1 h. Hg(II) that was in excess of S(-II) formed hydrophilic HgH2(GSH)22- complexes that were separated from the HgS complexes by C18-SPE. The concentration of Hg-sulfide complexes in the sample was estimated by taking the difference between the total Hg(II) concentration and the hydrophilic Hg (i.e., HgH2(GSH)22-) concentration. The time for S(-II) oxidation was recorded starting at the initial addition of S(-II) and ending at the time of Hg(II) addition. Preliminary experiments indicated that the formation of HgS complexes required at least 40 h. Therefore, to ensure equilibration of HgS complexes, the GSH CLE and extraction by C18-SPE were performed at least 40 h after the addition of Hg(II). Sample Analysis. The concentration of Hg(II) was measured in the C18-extracted samples (referred as hydrophilic Hg) by cold vapor atomic fluorescence spectrophotometry using SnCl2 reduction and dual-stage gold amalgamation (34, 35). The mean daily detection limits ((standard deviation) for Hg(II) was 85 pg ((61 pg, n ) 33 days), which corresponds to 0.004 nM (0.85 ng/L) for a 100-mL sample volume. Unless otherwise noted, error bars reported with the data were calculated according to (1 standard deviation of the Hg(II) calibration curve (36). Prior to extraction by CLE-SPE, 92.5% ((9.1%, n ) 135) of the total Hg was recovered in aliquots of the wastewater effluent samples and model solutions, indicating that loss of Hg to container walls was not significant during the oxidation studies. Chlorine demand in the wastewater was monitored by measuring the total chlorine concentration after 1 h of chlorine contact time using the DPD colorimetric method (37). The average 1-h chlorine demand for the wastewater effluent was 5.8 mg/L Cl2 ((0.4 mg/L) at WWTP 1 and 0.7 mg/L Cl2 ((0.7 mg/L) at WWTP 2. The detection limit for total chlorine residual (based on 3 times the standard deviation of the blank) was 0.13 mg/L Cl2 (1.9 µM). The addition of chlorine concentration of up to 100 mg/L Cl2 (1.4 mM Cl2) to the effluent samples from WWTP 1 and WWTP 2 did not result in measurable loss of chlorine, indicating that breakpoint chlorination was not important and that free chlorine rather than combined chlorine was formed upon addition of NaOCl to the effluent samples.

Results Hg(II) Complexation in Wastewater Effluent. Strong Hg(II)-complexing ligands in the wastewater effluent samples were detected by CLE-SPE using two different titrations. In the first titration, the GSH concentration was varied between 0 and 140 µM in aliquots of final effluent from WWTP 1 that were amended with 1.0 nM Hg(II). The concentration of hydrophilic Hg (i.e., HgH2(GSH)22-) was measured after allowing the Hg(II) to equilibrate in the effluent sample for 5 h, 2 days, and 9 days. This type of GSH titration can be used to estimate the lower limits of cKHgL for the ligands that form strong complexes at this total Hg(II) concentration (6). In the WWTP 1 final effluent samples (collected after chlorine disinfection), the concentration of hydrophilic Hg initially increased with increasing GSH, demonstrating the formation of hydrophilic HgH2(GSH)22- complexes (Figure 1a). As previously described (6), strong (or inert) Hg(II)

FIGURE 1. Hydrophilic Hg(II) measured by C18-SPE after GSH CLE titration: (a) filtered final effluent from WWTP 1 that was amended with 1.0 nM Hg(II) and equilibrated for 5 h ([), 2 days (9), and 9 days (O) prior to addition of GSH and (b) model solutions containing Hg(II) + humic acid (]), Hg(II) + S(-II) (×), and Hg(II) + S(-II) + humic acid (4), pH 7.3. complexes that have cKHgL greater than 1030 are indicated by the persistence of hydrophobic Hg species at higher GSH concentrations (i.e., 20-100 µM GSH). In the WWTP 1 effluent sample (Figure 1a), the presence of the strong Hg(II) complexes was difficult to detect when Hg(II) was equilibrated in the sample for 5 h (GSH concentrations greater than 50 µM were not included in the titration); less than 25% of the Hg(II) was inert to GSH after 5 h. As the equilibration time was increased to 2 and 9 days, the percentage of GSH-inert Hg(II) increased to 63% and 78%, respectively. In the second type of titration performed on the wastewater effluent samples, the GSH concentration was held constant at 14 µM while the Hg(II) concentration was varied between 0.8 and 225 nM (Figure 2). The CLE titration with Hg(II) was used to estimate the total concentration of strong ligands in a manner similar to that employed with other trace metals (e.g., refs 27-30). In freshly collected final effluent samples from WWTP 1 (inset in Figure 2a), hydrophilic Hg increased slightly when total Hg(II) was varied from 0.8 to 10 nM. In this range of added Hg concentration, the slope was less than 0.2, indicating that most of the added Hg(II) was converted to strong Hg(II) complexes and that the concentration of strong ligands was greater than 10 nM. Upon extending the Hg(II) titration to 225 nM (Figure 2), the slope of the titration increased to approximately 0.7, indicating saturation of the strong hydrophobic ligands. On the basis of the data points greater than 50 nM total Hg, the average strong ligand concentration was 38 nM ((2 nM, n ) 3). The slope of the titration did not achieve unity at concentrations of up to 225 nM total Hg, which can be caused by the presence of multiple Hg(II)-complexing ligands. At high Hg(II) concentrations, weaker hydrophobic ligands may be able to compete with 14 µM GSH that was added to the sample, resulting in a fraction of Hg(II) that is hydrophobic. VOL. 39, NO. 11, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4037

FIGURE 3. Removal of strong Hg(II)-complexing ligands in wastewater final effluent samples and in model sulfide solutions by 1-h chlorination with 10 mg/L Cl2, by O2-oxidation for at least 2 weeks, and by chlorination of O2-oxidized samples. Error bars represent 1 standard error (n ) 2-4). Data with no error bars reflect single measurements.

FIGURE 2. Hg(II) CLE titrations of wastewater effluent from WWTP 1. (a) Freshly collected predisinfection wastewater effluent (×) and final effluent ([) were immediately amended with Hg(II) without further treatment. Separate aliquots of final effluent were chlorinated for 1 h with 10 mg/L Cl2 (4) and 100 mg/L Cl2 (O) prior to addition of Hg(II). (b) Final effluent stored in the presence of oxygen for 2 weeks prior to addition of Hg(II) (]) and freshly collected final effluent ([). Separate aliquots of the stored samples also were subsequently chlorinated for 1 h by 10 mg/L Cl2 (0). At ambient Hg(II) concentrations (i.e., generally less than 0.1 nM), Hg(II) should be complexed predominantly by the strong hydrophobic ligands. To gain insight into the identity of the strong ligands, the concentration of strong ligand was quantified in wastewater collected before the disinfection process (i.e., after secondary treatment) at WWTP 1. Results of the Hg CLE titration indicated that the concentration of strong ligand was greater before disinfection than in the final effluent (Figure 2a). In the portion of the titration where total Hg(II) was greater than 50 nM, the slope was approximately 0.2 in the predisinfection sample, whereas the slope was approximately 0.7 in the final effluent sample (Figure 2a). This result suggested that the strong ligands were partially removed during disinfection. The final effluent sample from WWTP 1 was further treated in the laboratory by chlorination for 1 h. Hg(II) CLE titrations of samples that were chlorinated with 10 and 100 mg/L Cl2 resulted in higher hydrophilic Hg concentrations (open symbols in Figure 2a). Furthermore, the concentration of strong ligand decreased to less 10% of the initial concentration, indicating that the strong ligands were almost completely removed with additional chlorination (Figure 3). When aliquots of the WWTP 1 effluent samples were stored for 2 weeks in the presence of oxygen, the concentration of strong ligands decreased to approximately 16 nM ((3.8 nM), which is approximately 43% of the initial strong ligand concentration (Figure 3). Separate aliquots of the O2-oxidized samples were further treated with 10 mg/L Cl2 for 1 h, resulting in a strong ligand concentration of 9.1 nM ((1.2 nM) (Figure 2b), which was approximately 24% of the initial concentration (Figure 3). Chlorination of the O2-oxidized sample partially 4038

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 11, 2005

removed the strong Hg(II)-complexing ligands but did not completely remove the strong ligands as was the case for the freshly collected samples. In final effluent from WWTP 2, the Hg(II) CLE titrations of the samples indicated that the strong ligand concentration was approximately 26 nM ((3 nM) (CLE titration data not shown). Treatment of the WWTP 2 samples with chlorine or oxygen resulted in the removal of the strong ligands, similar to results obtained for samples collected from WWTP 1 (Figure 3). Greater than 85% of the strong Hg(II)-complexing ligands were removed when the samples were chlorinated with 10 and 100 mg/L Cl2. Approximately 54% ((12%) of the initial strong ligands remained after storing the samples in the presence of oxygen for 2 weeks. Upon chlorination of the O2-oxidized samples, the strong ligand concentration slightly decreased to 40% ((5%) of the initial concentration. Hg(II) Complexation in Model Solutions. CLE titrations of the model solutions containing sulfide and humic acid indicated that S(-II) formed strong hydrophobic complexes with Hg(II), whereas humic acid did not (Figure 1b). In the Hg-sulfide model solutions, approximately 95% of Hg(II) was retained on the C18 column in the absence of GSH. Upon addition of up to 100 µM GSH, hydrophilic Hg increased to approximately 14% of the total Hg(II). The GSH titration of the humic acid solution (] in Figure 1b) indicated that the Hg-humic acid complexes were almost completely hydrophilic. Only about 10% of the Hg was retained on the C18 column, in the absence of GSH. When GSH was added, none of the Hg was retained by C18. When 15 mg/L humic acid was added to a solution containing 8 nM Hg(II)-sulfide, the hydrophilic Hg accounted for approximately 40% of the total Hg, at GSH concentrations up to 100 µM. Additional studies with the model solutions were conducted to evaluate the stability of S(-II) in the form of free sulfide (HS-/H2S), Zn-sulfide, and Fe-sulfide complexes. S(-II) was quantified by converting S(-II) in the sample into hydrophobic HgS complexes and separating the HgS complexes by CLE-SPE. The Hg(II) titration of a freshly prepared solution containing 4 nM S(-II) (Figure 4) exhibited a slope of unity only when the total Hg(II) exceeded the concentration of S(-II). Results of the titration indicated that Hg(II) forms 1:1 complexes with S(-II) in the presence of 6 µM GSH. By taking the difference between total Hg and hydrophilic Hg for the data points greater than 5 nM total Hg, the

FIGURE 4. Hg(II) CLE titration of a model solution containing 4 nM S(-II) (pH 7.3). The dotted line represents the linear regression of data points greater than 5 nM total Hg.

FIGURE 5. Formation of HgS in air-saturated solutions containing 7.0-8.0 nM (a) Zn-sulfide and (b) Fe-sulfide. Hg(II) was added to Zn-sulfide and Fe-sulfide solutions that were freshly prepared (]) or to solutions that had been stored for 2 days (2), 6 days (b), and 8 days (×). Data points represent the average ((1 standard error) of replicate samples (n ) 2-4). concentration of S(-II) in the model solution was estimated to be 3.8 nM ((0.4 nM, n ) 5), which was 94% ((8.9%) of the initial S(-II) added. In the ZnS solution, 89.0% ((10.6%, n ) 7) of initial S(-II) was recovered by complexation with Hg(II). In the solution containing Fe-sulfide complexes, 82.7% ((6.3%, n ) 5) of the initial S(-II) was recovered. The equilibration time required for Hg(II) to form HgS complexes was monitored in newly formed and aged Znsulfide and Fe-sulfide solutions (Figure 5). In solutions of newly formed Zn-sulfide complexes (i.e., less than 30 min), approximately 80% of the Zn was displaced after 20 min, and the complete replacement occurred within 24 h (Figure 5a). In addition, HgS complexes that formed during this process were stable in these solutions for more than 100 h. When the

FIGURE 6. Stability of S(-II) in air-saturated model solutions containing HS-/H2S (×), Fe-sulfide (b), and Zn-sulfide (]). Error bars represent 1 standard error of replicate samples (n ) 2-4). Zn-sulfide was stored (in air-saturated solutions) for 2 days prior to the initiation of the Hg exchange experiment, the complete replacement of Hg(II) required more than 50 h. In the 2-day aged Zn-sulfide solution, the percentage of S(-II) as HgS increased from 10% at 5 h to 80% at 50 h (triangle symbol in Figure 5a). Similar results were obtained with Fe-sulfide complexes (Figure 5b). In newly formed Fe-sulfide complexes, 80100% of the initial S(-II) was measured as HgS within 20 h of the addition of Hg(II). However, when the Fe-sulfide complexes were aged for 6 days, the formation of HgS upon addition of Hg(II) occurred at a slower rate. These Hg(II) exchange experiments indicated that the replacement of ZnS and FeS complexes with HgS occurred over 1-2 days. Therefore, in subsequent experiments, Hg(II) was added to the solutions and equilibrated for 44-48 h prior to measurement of HgS complexes. The Hg(II)-complexing capability of S(-II) was further characterized by exposing the sulfide model solutions to oxygen and chlorine. When the solutions were stored in the presence of oxygen for 3-5 days, 50% of the initial S(-II) was available to form strong Hg(II) complexes (Figure 6). After 10 days in the presence of oxygen, approximately 30% of the initial concentration of Hg(II)-complexing ligands remained and was stable for at least 30 days. The effect of chlorine oxidation was measured by exposing the model sulfide solutions to 10 mg/L Cl2 (0.14 mM) for 1 h (Figure 3). Less than 10% of initial sulfide (measured by the Hg CLE titration) was recovered after chlorination. Aliquots of model solutions that had been exposed to oxygen for 30 days were subsequently exposed to 10 mg/L Cl2 for 1 h. Approximately 25%-40% of initial S(-II) was recovered as HgS complexes after exposure to O2 and did not significantly change upon additional chlorination (Figure 3).

Discussion The CLE titrations of wastewater effluent provide insight into the ligands responsible for the formation of strongly complexed Hg(II) and the fate of the complexes in the aquatic environment. The titration results of the wastewater effluent samples indicated that, upon addition of Hg(II), strong Hg(II) complexes were formed over a period of 2-9 days. These strong Hg(II) complexes were stable for several days in the presence of oxygen (Figure 1a). However, if the strong ligands were exposed to oxygen prior to addition of Hg(II), the concentration of strong ligands decreased by approximately 60% (Figure 3). Furthermore, the strong Hg(II)-complexing ligands were almost completely removed after exposure to chlorine doses comparable to those employed for wastewater disinfection (generally less than 10 mg/L Cl2 (38)). VOL. 39, NO. 11, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4039

The conditional stability constants of these complexes (cKHgL > 1030) suggests that the strong ligands are not humic substances or thiol-containing organic ligands (6). On the basis of published stability constants and previous reports indicating the presence of 50-100 nM sulfide in municipal wastewater effluent (9), it is possible that the ligands consist of inorganic sulfide species. Furthermore, the titrations of model solutions (Figure 1b) indicated that Hg-sulfide complexes are hydrophobic and inert to the addition of GSH, which is consistent with data from titrations of wastewater effluent. In the wastewater effluent samples, most of the added Hg(II) formed hydrophobic complexes. However, the Hg(II) was not completely removed by C18-SPE (i.e., approximately 20% of the Hg(II) passed through the C18 column at zero GSH in Figure 1a). As indicated by the results depicted in Figure 1b, HgS complexes are completely retained on the C18 column. If Hg(II) was complexed entirely by sulfide in the wastewater effluent and if there was an excess of sulfide relative to added Hg(II) as our data indicate, then all of the Hg(II) should have been retained by the C18 column. It is possible that the presence of humic substances or similar types of organic matter in the wastewater effluent accounted for the incomplete removal of the Hg(II) complexes in the wastewater effluent titrations. The strongly complexed metal may have been associating with the organic matter (4), or the organic matter may have altered the properties of the SPE resin. The potential for humic substances to cause breakthrough of complexed Hg(II) is indicated by the GSH titration performed in the model solution containing both humic acid and S(-II) (Figure 1b). Further research is needed to evaluate the possibility that HgS complexes interact with humic substances. The removal of model S(-II) ligands by oxidants was consistent with oxidation studies of the wastewater effluent samples. In both the wastewater effluent and model S(-II) solutions, the strong ligands were completely removed after exposure to chlorine and were partially removed after 2 weeks of exposure to oxygen (Figures 3). Furthermore, the strong ligands that remained after O2-oxidation were resistant to further loss by reaction with chlorine, indicating that sulfide was transformed to a more stable sulfur-containing ligand. The O2-oxidation rate for free sulfide at neutral pH varies widely, with half-times ranging from 6 to 50 h (8, 39-41). The presence of trace concentrations of Fe(III) accelerates the oxidation rate because Fe(III) catalyzes the oxidation of S(-II) by O2 (17). Therefore, oxygen was expected to completely oxidize S(-II) in the HS-/H2S and Fe-sulfide solutions within 2 weeks, and S(-II) was not likely to be the ligand responsible for the strong Hg(II) complexes after 2-weeks exposure to O2 (Figure 3). Products of S(-II) oxidation by O2 include SO42-, SO32-, and S2O32- (8, 40, 41). Stability constants for Hg2+ complexes with SO42-, SO32-, and S2O32(31) indicate that these ligands form weak complexes that should dissociate in the presence of GSH (42). The oxidation of S(-II) also produced other species that can form strong complexes with Hg(II). For example, polysulfides Sx2- (8) form complexes with Hg(II) with stability similar to that of Hg-sulfide complexes (43, 44). However, frontier molecular orbital calculations indicate that polysulfides are more reactive than HS- with chlorine. The highest occupied molecular orbital (HOMO) energy of polysulfides such as S42- is +1.4 eV, greater than the HOMO energy for HS- (-2.5 eV) (45). From a thermodynamic perspective, polysulfides should be as reactive as HS- with NaOCl, if not more reactive. Therefore, polysulfides are not expected to account for the strong Hg(II) complexes in the O2-oxidized HS-/H2S solutions; however, in the Fe- and Zn-sulfide solutions, polysulfides may be protected from chlorine oxidation by complexation with Fe and Zn. 4040

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 11, 2005

Another intermediate of sulfide oxidation is elemental sulfur, S8 (46), which has a lower HOMO energy level (-9.0 eV (45)) than HS-. Therefore, the oxidation of S8 by chlorine is expected to be slower than that of HS- or Sx2-. S8 may be present in the O2-oxidized HS-/H2S and Fe-sulfide solutions and may be forming stable complexes with Hg(II). For example, the S8 ring may be able to accommodate the Hg2+ ion; however, further research is needed to understand the complexation of Hg(II) by S8. Similar to the HS-/H2S and Fe-sulfide solutions, exposure of the Zn-sulfide solutions to O2 led to a decrease in the concentration of strong ligands (Figure 3). However, the oxidation of S(-II) by O2 was not expected to be fast under these conditions, because complexation by Zn inhibits the oxidation of S(-II) (16, 17, 47). Zn-sulfide (and Fe(II)-sulfide) has been detected as metal-sulfide clusters in oxic surface waters (14), and Zn-sulfide clusters can persist for weeks under oxic conditions (16). Hg(II) replacement experiments (Figure 5) demonstrated that longer Hg(II) equilibration times (at least 40 h) were required for 6-day-aged ZnS and FeS. In the wastewater effluent, the formation of the strong Hg(II) complexes also occurred over a period of several days (Figure 1a). These results suggest that Zn-sulfide clusters undergo an aging process during the O2-oxidation experiments so that less of the S(-II) in aged clusters is available for Hg(II) complexation. Additional studies are needed to determine the rate of transformation of Zn-sulfide complexes when exposed to oxygen. The oxidation studies indicated that the strong Hg(II) complexes in wastewater effluent are stable over weeks in the presence of oxygen. Similarly, sulfur-containing ligands such as S8 or metal-sulfide clusters also are stable under oxic conditions. Therefore, sulfide species may play an important role in Hg(II) speciation in wastewater effluent. Other sources of S(-II) also may serve as a source of strong Hg(II)-complexing ligands to surface waters. For example, anoxic sediments are a source of acid volatile sulfides (in the form of free sulfide and metal sulfides) that are stable for hours upon suspension of the sediments in oxic waters (47). Further research is needed to assess the occurrence of strong Hg(II)-complexing ligands in natural waters and their role in Hg biogeochemistry.

Acknowledgments This research was supported in part by the University of California Toxic Substances Research and Teaching Program Coastal Toxicology Component. H.H. was supported by a graduate fellowship from the National Physical Science Consortium. We also thank G.W. Luther (University of Delaware) and F. Black (UCsSanta Cruz) for their valuable comments on this work.

Literature Cited (1) Xia, K.; Skyllberg, U. L.; Bleam, W. F.; Bloom, P. R.; Nater, E. A.; Helmke, P. A. X-ray absorption spectroscopic evidence for the complexation of Hg(II) by reduced sulfur in soil humic substances. Environ. Sci. Technol. 1999, 33, 257-261. (2) Benoit, J. M.; Mason, R. P.; Gilmour, C. C.; Aiken, G. R. Constants for mercury binding by dissolved organic matter isolates from the Florida Everglades. Geochim. Cosmochim. Acta 2001, 65, 4445-4451. (3) Haitzer, M.; Aiken, G. R.; Ryan, J. N. Binding of mercury(II) to dissolved organic matter: The role of the mercury-to-DOM concentration ratio. Environ. Sci. Technol. 2002, 36, 3564-3570. (4) Haitzer, M.; Aiken, G. R.; Ryan, J. N. Binding of mercury(II) to aquatic humic substances: Influence of pH and source of humic substances. Environ. Sci. Technol. 2003, 37, 2436-2441. (5) Qian, J.; Skyllberg, U.; Frech, W.; Bleam, W. F.; Bloom, P. R.; Petit, P. E. Bonding of methyl mercury to reduced sulfur groups in soil and stream organic matter as determined by X-ray absorption spectroscopy and binding affinity sites. Geochim. Cosmochim. Acta 2002, 66, 3873-3885.

(6) Hsu, H.; Sedlak, D. L. Strong Hg(II) complexation in municipal wastewater effluent and surface waters. Environ. Sci. Technol. 2003, 37, 2743-2749. (7) Benoit, J. M.; Gilmour, C. C.; Mason, R. P.; Heyes A. Sulfide controls on mercury speciation and bioavailability to methylating bacteria in sediment pore waters. Environ. Sci. Technol. 1999, 33, 951-957. (8) Zhang, J.; Millero, F. J. The products from the oxidation of H2S in seawater. Geochim. Cosmochim. Acta 1993, 57, 1705-1718. (9) Adams, N. W. H.; Kramer, J. R. Silver speciation in wastewater effluent, surface waters and pore waters. Environ. Toxicol. Chem. 1999, 18, 2667-2673. (10) Bowles, K. C.; Ernste, M. J.; Kramer, J. R. Trace sulfide determination in oxic freshwaters. Anal. Chim. Acta 2003, 477, 113-124. (11) Luther, G. W.; Tsamakis, E. Concentration and form of dissolved sulfide in the oxic water column of the ocean. Mar. Chem. 1989, 27, 165-177. (12) Mylon, S. E.; Benoit, G. Subnanomolar detection of acid-labile sulfides by the classical methylene blue method coupled to HPLC. Environ. Sci. Technol. 2001, 35, 4544-4548. (13) Rozan, T. F.; Benoit, G.; Luther, G. W. Measuring metal sulfide complexes in oxic river waters with square wave voltammetry. Environ. Sci. Technol. 1999, 33, 3021-3026. (14) Rozan, T. F.; Lassman, M. E.; Ridge, D. P.; Luther, G. W. Evidence for iron, copper and zinc complexation as multinuclear sulphide clusters in oxic rivers. Nature 2000, 406, 879-882. (15) U.S. Environmental Protection Agency. Mercury Study Report to Congress; Volume III: Fate and Transport Study of Mercury in the Environment; U.S. Government Printing Office: Washington, DC, 1997. (16) Rozan, T. F.; Luther, G. W. Voltammetric evidence suggesting Ag speciation is dominated by sulfide complexation in river water. In Environmental Electrochemistry: Analyses of Trace Element Biogeochemistry; ACS Symposium Series 811; Taillefert, M., Rozan, T. F., Eds.; American Chemical Society: Washington, DC, 2002; pp 371-387. (17) Vazquez, F.; Zhang, J. Z.; Millero, F. J. Effects of metals on the rate of oxidation of H2S in seawater. Geophys. Res. Lett. 1989, 16, 1363-1366. (18) Mugan, T. J. Quantification of total mercury discharges from publicly owned treatment works to Wisconsin surface waters. Water Environ. Res. 1996, 68, 229-234. (19) Bodaly, R. A.; Rudd, J. W. M.; Flett, R. J. Effect of urban sewage treatment on total and methyl mercury concentrations in effluents. Biogeochemistry 1998, 40, 279-291. (20) Mierle, G.; Ingram, R. The role of humic substances in the mobilization of mercury from watersheds. Water, Air, Soil Pollut. 1991, 56, 349-357. (21) Watras, C. J.; Morrison, K. A.; Bloom, N. S. Chemical correlates of Hg and methyl-Hg in northern Wisonsin Lake waters under ice-cover. Water, Air, Soil Pollut. 1995, 84, 253-267. (22) Guentzel, J. L.; Powell, R. T.; Landing, W. M.; Mason, R. P. Mercury associated with colloidal material in an estuarine and an openocean environment. Mar. Chem. 1996, 55, 177-188. (23) Driscoll, C. T.; Hosapple, J.; Schofield, C. L.; Munson, R. The chemistry and transport of mercury in a small wetland in the Adirondack region of New York, USA. Biogeochemistry 1998, 40, 137-146. (24) Babiarz, C. L.; Hurley, J. P.; Benoit, J. M.; Shafer, M. M.; Andren, A. W.; Webb, D. A. Seasonal influences on partitioning and transport of total mercury and methylmercury in rivers from contrasting watersheds. Biogeochemistry 1998, 41, 237-257. (25) Barkay, T.; Gillman, M.; Turner, R. R. Effects of dissolved organic carbon and salinity on bioavailability of mercury. Appl. Environ. Microbiol. 1997, 63, 4267-4271. (26) U. S. Environmental Protection Agency. Method 1631, Revision D: Mercury in Water by Oxidation, Purge and Trap, and Cold Vapor Atomic Fluorescence Spectroscopy; U.S. Government Printing Office: Washington, DC, 2001. (27) van den Berg, C. M. G.; Donat, J. R. Determination and Data Evaluation of Copper Complexation by OrganicLigands in Seawater Using Cathodic Stripping Voltammetry at

(28)

(29) (30)

(31) (32)

(33)

(34)

(35)

(36) (37)

(38) (39) (40) (41) (42) (43) (44) (45) (46) (47)

Varying Detection Windows. Anal. Chim. Acta 1992, 257, 281-291. van den Berg, C. M. G. Determination of Copper, Cadmium and Lead in Seawater by Cathodic Stripping Voltammetry of Complexes with 8-Hydroxyquinoline. J. Electroanal. Chem. 1986, 215, 111-121. Donat, J. R.; Lao, K. A.; Bruland, K. W. Speciation of Dissolved Copper and Nickel in South San-Francisco BaysA Multimethod Approach. Anal. Chim. Acta 1994, 284, 547-571. Miller, L. A.; Bruland, K. W. Competitive equilibration techniques for determining transition metal speciation in natural waters: Evaluation using model data. Anal. Chim. Acta 1997, 343, 161181. Smith, R. M.; Martell, A. E. Critical Stability Constants, Vol. 4 Inorganic ligands; Plenum Press: New York, 1976. Luther, G. W.; Ferdelman, T. G. Voltammetric characterization of iron(II) sulfide complexes in laboratory solutions and in marine waters and porewaters. Environ. Sci. Technol. 1993, 27, 1154-1163. Cutter, G. A.; Walsh, R. S.; Silva de Echols, C. Production and speciation of hydrogen sulfide in surface waters of the high latitude North Atlantic Ocean. Deep-Sea Res. II 1999, 46, 9911010. Gill, G. A.; Fitzgerald, W. F. Picomolar Mercury Measurements in Seawater and Other Materials Using Stannous Chloride Reduction and 2-Stage Gold Amalgamation with Gas-Phase Detection. Mar. Chem. 1987, 20, 227-243. Bloom, N.; Fitzgerald, W. F. Determination of Volatile Mercury Species at the Picogram Level by Low-Temperature GasChromatography with Cold-Vapor Atomic Fluorescence Detection. Anal. Chim. Acta 1988, 208, 151-161. Skoog, D. A.; Holler, F. J.; Nieman, T. A. Principles of Instrumental Analysis; 5th ed.; Harcourt Brace: Orlando, FL, 1998. Eaton, A. D.; Clesceri, L. S.; Greenberg, A. E., Eds. Standard Methods for the Examination of Water and Wastewater; 20th ed.; American Public Health Association, American Water Works Association, Water Environment Federation Publishers: Washington, DC, 1998. Tchobanoglous, G.; Burton, F. L.; Stensel, H. D. Wastewater Engineering, Treatment and Reuse; 4th ed.; McGraw-Hill: New York, 2003. Millero, F. J.; Hubinger, S.; Fernandez, M.; Garnett, S. Oxidation of H2S in seawater as a function of temperature, pH, and ionic strength. Environ. Sci. Technol. 1987, 21, 439-443. Cline, J. D.; Richards, F. A. Oxygenation of hydrogen sulfide in seawater at constant salinity, temperature, and pH. Environ. Sci. Technol. 1969, 3, 838-843. O’Brien, D. J.; Birkner, F. B. Kinetics of oxygenation of reduced sulfur species in aqueous solution. Environ. Sci. Technol. 1977, 12, 1114-1120. Oram, P. D.; Fang, X. J.; Fernando, Q.; Letkeman, P.; Letkeman, D. The formation constants of mercury(II)-glutathione complexes. Chem. Res. Toxicol. 1996, 9, 709-712. Jay, J. A.; Morel, F. M. M.; Hemond, H. F. Mercury speciation in the presence of polysulfides. Environ. Sci. Technol. 2000, 34, 2196-2200. Paquette, K. E.; Helz, G. R. Inorganic speciation of mercury in sulfidic waters: The importance of zerovalent sulfur. Environ. Sci. Technol. 1997, 31, 2148-2153. Luther, G. W. The frontier-molecular-orbital theory approach in geochemical processes. In Aquatic Chemical Kinetics; Stumm W., Ed.; John Wiley and Sons: New York, 1990; pp 173-198. Yao, W.; Millero, F. J. Oxidation of hydrogen sulfide by hydrous Fe(III) oxides in seawater. Mar. Chem. 1996, 52, 1-16. Simpson, S. L.; Apte, S. C.; Batley, G. E. Effect of short-term resuspension events on trace metal speciation in polluted anoxic sediments. Environ. Sci. Technol. 1998, 32, 620-625.

Received for review January 3, 2005. Revised manuscript received March 24, 2005. Accepted March 25, 2005. ES050013I

VOL. 39, NO. 11, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4041