Environ. Sci. Technol. 2009, 43, 6989–6995
Natural Attenuation Potential of Phenylarsenicals in Anoxic Groundwaters M I C H A E L H E M P E L , † B I R G I T D A U S , * ,† CARSTEN VOGT,‡ AND HOLGER WEISS† Department Groundwater Remediation and Department Isotope Biogeochemistry, UFZ, Helmholtz Centre for Environmental Research - UFZ, Permoserstrasse 15, 04318 Leipzig, Germany
Received March 4, 2009. Revised manuscript received June 16, 2009. Accepted July 27, 2009.
The extensive production of chemical warfare agents in the 20th century has led to serious contamination of soil and groundwater with phenyl arsenicals at former ammunition depots or warfare agent production sites worldwide. Most phenyl arsenicals are highly toxic for humans. The microbial degradation of phenylarsonic acid (PAA) and diphenylarsinic acid (DPAA) was investigated in microcosms made of anoxic groundwater/sediment mixtures taken from different depths of an anoxic, phenyl arsenical contaminated aquifer in Central Germany. DPAA was not transformed within 91 days incubation time in any of the microcosms. The removal of PAA can be described by a first order kinetics without a lag-phase (rate: 0.037 d-1). In sterilized microcosms, PAA concentrations always remained stable, demonstrating that PAA transformation was a biologically mediated process. PAA transformation occurred under sulfate-reducing conditions due to sulfate consumption and production of sulfide. The addition of lactate (1 mM), a typical substrate of sulfate-reducing bacteria, increased the transformation rate of PAA significantly up to 0.134 d-1. The content of total arsenic was considerably reduced (>75%). Intermediates of PAA transformation were detected by high performance liquid chromatography-inductively coupled plasma mass spectrometry (HPLC-ICP-MS). Experiments with a pure strain and sterile controls of Desulfovibrio gigas spiked with PAA showed that the elimination process is linked to the presence of sulfide formed through bacterial activity. Phenyl arsenicals were likely immobilized in the sediment through sulfur substitution and a subsequent sulfur bond under the prevailing sulfate reducing condition. The results of this study indicate that PAA can undergo microbiologically mediated transformation in anoxic aquifers, leading to reduced concentrations in groundwater, which indicate a (enhancend) natural attenuation potential.
Introduction Arsenic is a ubiquitous element in the environment, and it appears naturally in a wide range of different compounds that can be highly toxic for humans (1). Most arsenic compounds contain hydrogen, oxygen, or sulfur. In recent * Corresponding author phone: +(49) 341 235 1769; fax: +(49) 341 235 1837; e-mail:
[email protected]. † Department Groundwater Remediation. ‡ Department Isotope Biogeochemistry. 10.1021/es9006788 CCC: $40.75
Published on Web 08/12/2009
2009 American Chemical Society
years, organic arsenicals such as methylated arsenic species or arsenosugars have also been discovered (2, 3). In addition to naturally occurring arsenic species, anthropogenic arsenic compounds also exist and act as contaminants in the environment. Phenyl arsenicals are such anthropogenic compounds, best known as the chemical warfare agents diphenylarsine chloride (Clark I), diphenylarsine cyanide (Clark II) and phenylarsine dichloride (Pfiffikus) during the First and Second World Wars (4). Even today these sternutators and their derivatives can be found, especially in the vicinity of abandoned ammunition depots, e.g., in Belgium (5), Japan (6, 7), China (8), and the Baltic Sea (9). It is still an unsolved question how to remediate groundwater which is highly contaminated with phenyl arsenicals. The compounds seem to be recalcitrant due to their persistent appearance for decades after their use. In Germany, five sites with a similar history and contamination profile are known. One of these former filling stations for chemical warfare agents is located near the city of Dessau in Central Germany (10). In spite of remediation efforts in 2005 through soil excavation, high concentrations of organic arsenicals are still being detected in the groundwater. The main contaminants are phenylarsonic acid (PAA) and diphenylarsinic acid (DPAA), both oxidation and hydrolysis products of chemical warfare agents, which cover up to 80% of total arsenic content (11). These compounds seem to be again persistent in light of their stability for over 60 years in storage tanks and, for an unknown period of time, in groundwater. Aside from some hot spots, the groundwater at the site is contaminated in a high volume with rather low concentrations. An ex situ remediation technology (e.g., pump and treat) would be very expensive due to the high volume of water that would need to be handled. Only a few methods have been described so far regarding the remediation of groundwater contaminated by organic arsenical compounds. The treatment methods published to date either refer primarily to the original chemical warfare agent, or are not applicable for in situ remediation of contaminated groundwater due to the chemical (e.g., addition of toluene) and/or high temperature requirements (12, 13). With the enzyme manganese peroxidase from wooden destructive basidiomycetes Haas et al. (14) were able to achieve 100% degradation of Clark I. After 24 h of treatment in the presence of glutathione and a subsequent thiol derivatization, no physiological relevant arsenic compounds could be determined. The mold fungus Trichoderma harzianum could oxidize triphenylarsine, a precursor for Clark synthesis, to triphenylarsine oxide. Another fungus, the germ of white rot Phanerochaete chrysosporium, oxidized phenylarsine oxide (PAO) to PAA (15). These biotechnological approaches are dependent on the availability of oxygen as the oxidizing agent and electron acceptor for fungal enzymes and are therefore not applicable for in situ remediation techniques due to the typically anoxic conditions in contaminated groundwater. Nakamiya et al. (16) isolated the bacteria Kytococcus sedentarius from DPAA contaminated soil by the use of toluene as sole source for carbon. This strain was able to utilize DPAA under aerobic conditions within 3 days. The degradation product was arsenic acid. To the best of our knowledge, the degradation of phenylarsonic compounds under anoxic groundwater conditions has, so far, not been described. Thus, the aim of this study was to investigate the potential of the autochthonous bacterial community in an organo-arsenic contaminated site for degrading the phenyl arsenicals PAA and DPAA under anoxic in situ conditions. VOL. 43, NO. 18, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6989
TABLE 1. Characteristics and Anion Concentrations of the Groundwater Sample pH
Eh (mV)
6.3 -285.4
conductivity (µS/cm) 431
NO3SO42PO43Fetotal (mg L-1) (mg L-1) (mg L-1) (mg/L) 3.1
177.7
0.1
9.6
Materials and Methods Materials. All chemicals used in this study were of analyticalgrade or higher purity. Deionized water (18.2 MΩ cm) was obtained from a Direct-QTM 5 system (Millipore) and was used to prepare all necessary dilutions. A 1 M phosphoric acid was prepared from 85% H3PO4 (p.a., Merck). Glassware and solutions were always sterilized by autoclaving (20 min, 121 °C) or filtration before use. Sampling. Water samples were taken at the site of an abandoned ammunition depot and filling station for chemical warfare agents in Central Germany. The characteristics of the groundwater and concentrations of important ions are shown in Table 1. The iron content is dominated by Fe2+ ions. For sampling, the direct push-technology with the Geoprobe-System “Sampling Point 16” was used (17). The samples containing groundwater mixed with fine grained aquifer material were taken downgradient from the assumed contamination source at a depth of 4.4-5.4 m below the surface, this representing an anaerobic and highly polluted region (up to 1.8 mg L-1 As) of the local aquifer. The static groundwater level was 2.2 m below the surface. The samples were stored in completely filled sterile 1000 mL glass bottles sealed with airtight screw caps. The bottles were stored statically in the dark at 12 °C in an incubator (WTC Binder) until further treatment. Set-up of Groundwater Microcosms. For the set up of groundwater microcosms, the samples were treated in an anaerobic glovebox (gas atmosphere: 95% nitrogen, 5% hydrogen; Coy Laboratory Products Inc., U.S.) to replicate groundwater conditions and to avoid any oxygen contamination. 100 mL of the groundwater/sediment-mixture (approximately 1.5 g solids) were filled into 118 mL serum bottles (Glasgera¨tebau Ochs, Bovenden Lengern, Germany) and sealed airtight with inert Teflon-coated butyl rubber stoppers and aluminum crimp caps (ESWE Analysentechnik, Gera, Germany). For each groundwater/sediment mixture tested, three replicates were prepared. The treatment variants were: abiotic; without additional electron donors; with lactate as the additional electron donor. Lactate was added by means of an anoxic 1 M stock solution (Carl Roth GmbH & Co. KG, Karlsruhe, Germany) getting a final lactate concentration of 2 mM. For the setup of abiotic controls, microcosms were autoclaved (20 min, 121 °C) three times on three consecutive days to obtain sterile control microcosms. All microcosms were cultivated statically in the dark at a temperature of 20 °C for at least 13 weeks. Samples for chemical analysis of sulfide and different arsenic species were taken by means of sterile plastic syringes which were flushed with nitrogen before use to avoid oxygen contamination inside the microcosms. Experiments with Desulfovibrio gigas. Desulfovibrio gigas strain DSM 1382 was ordered from the Deutsche Sammlung fu ¨ r Mikroorganismen and Zellkulturen (DSMZ), Braunschweig, Germany. Cells were grown in a mineral salt medium described elsewhere (18) amended with 20 mM sulfate (Na2SO4, Merck) and 10 mM lactate (sodium lactate, J.T. Baker). For examining the effect of sulfide on PAA different microcosms were set up (118 mL serum flasks with 95 mL mineral salt medium amended with 20 mM sulfate and 10 mM lactate). The first microcosm contains only mineral salt medium with amendments (A). In three other bottles 5 mL freshly grown D. gigas cells were injected. One of these 6990
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 18, 2009
bottles served as living microcosm (B). An abiotic, sulfidecontaining control (C) was set up by autoclaving one serum flask (20 min; 121 °C). In a further bottle (D) 5 mL sterile filtered (0.2 µm, acetate membrane, NALGENE, U.S.) D. gigas cells were injected (second abiotic, sulfide-containing control). All microcosms (A-D) were spiked immediately after complete setup with PAA (final concentration of 1.5 mg L-1) from a stock solution (1 g L-1). All microcosms were sampled for sulfide and arsenic species analysis within one hour after PAA addition for the first time. Analytical Methods. Different organic and inorganic arsenic species were detected and quantified by a speciation method described by Daus et al. (11) via high performance liquid chromatography (HPLC) with a RSpak NN-614 column (150 × 6 mm, Shodex Japan). The chromatographic system (BECKMAN, System Gold, Fullerton, U.S.) was combined with an inductively coupled plasma mass spectrometry (ICP-MS; PQ ExCell, THERMO). For arsenic species measurements, analytical standards were produced for calibration from stock solutions of arsenate (1000 mg L-1 As, Titrisol, Merck), PAA (1000 mg L-1 As, Fluka Analytical) and PAO (1000 mg L-1 As, Sigma, Switzerland). For quantification of not identified arsenic species (named by their retention time in the chromatogram, e.g., 1650 s) and diphenylarsinic acid in the samples, mean calibration parameters were used. The total arsenic and iron content of the liquid phase in the microcosms was determined by ICP-atomic emission spectrometry (ICP-AES; CIROS, Spectro Analytical Instruments GmbH & Co. KG) after filtrating through a 45 µm membrane filter (cellulose acetate, Sartorius AG, Göttingen, Germany) at the beginning and the end of each degradation test series. Likewise, the arsenic species were analyzed by HPLC-ICP-MS. For both analyses the samples were acidified with phosphoric acid to a final concentration of 10 mM (19). The nonacidified filtrate was used for the detection of the anions nitrate and sulfate using the ion-chromatograph DX 500 (Dionex) with an IonPack AS12A/AG12A column. Subsequently, the microcosm’s sediment matrix on the filter was washed two times with 10 mL phosphoric acid (0.1 M) and two times with 10 mL deionized water. The liquid phase was filled up to a volume of 50 mL with deionized water. All samples were diluted to get a final concentration of phosphoric acid of 10 mM and analyzed on arsenic species. The soil material was dried and total arsenic and iron content was determined by wavelength dispersive X-ray fluorescence spectrometry (WDXRF, SRS 3000, Siemens). Sulfide was determined weekly by the photometric method of Cline (20) using modifications described elsewhere (21). Investigations of Unknown Organo-Arsenic Species. To elucidate the structure of occurring unknown species which accumulated in the biotic microcosms, the HPLC was coupled with a pressure electro-spray ionization (ESI) and a MSD G 1946 B (Agilent) (11). Samples from sulfate-reducing microcosms, under which the unknown organo-arsenic species were formed, were used for oxidation experiments with hydrogen peroxide (H2O2; suprapur, 30% v/v, Merck). The test samples contained an unknown organo-arsenic species with a retention time of approximately 1650 s in concentrations above 500 µg L-1 As. Two subsamples were diluted 1:5 with deionized water, transferred to 1.8 mL vials and treated with phosphoric acid (final concentration 10 mM). One of the subsamples was spiked with 10 µL H2O2 (1%) and subsequently both were analyzed for arsenic species by HPLC-ICP-MS.
Results and Discussion Degradation of Organo-Arsenic Species in Groundwater Microcosms. Freshly sampled mixtures of anoxic groundwater and fine grained sediment were anoxically incubated, hence simulating the in situ conditions at the contaminated
FIGURE 1. Time development of the arsenic species PAA (∆), PAO (x) and DPAA (]), an unknown arsenic compound having a retention time of 1650 s ([) and total arsenic (0) in the sterile microcosms (A), the untreated living microcosms (B) and the living microcosms spiked with 2 mM lactate (C). The curves are mean values of three replicates; error bars demonstrate the standard deviation. site. The anoxic groundwater contained sulfate in concentrations up to 178 mg L-1 as the main electron acceptor. The dominant arsenic species in the original sample were As(V) (175 µg L-1 As), PAA (1214 µg L-1 As) and DPAA (139 µg L-1 As), resulting in a total arsenic concentration of 1755 µg L-1. By using sediment and groundwater from the site, the bacterial community inhabiting the investigated aquifer zone was represented within the microcosms. Most groundwater bacteria are attached to particles, forming a biofilm (22).
Figure 1 shows the time development for the different arsenic species in the tested microcosms. The curves are mean values of three replicates with standard deviation shown in the error bars. For the determination of total arsenic, the concentrations of the arsenic species arsenate and arsenite were also included; these species were low but stable over the observed period of time (data not shown). Under sterile conditions, the concentrations of the different arsenic species remained constant (Figure 1A), VOL. 43, NO. 18, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6991
FIGURE 2. HPLC-ICP-MS chromatogram for arsenic detection in living microcosms after 2 weeks of incubation (biotic microcosm (-) and lactate spiked biotic microcosm (---)). demonstrating that the compounds were chemically stable under the given incubation conditions. In the living microcosms, the PAA concentration decreased at a rate of 0.329 µM L-1 d-1 within the first two weeks; the rate slowed to 0.048 µM L-1 d-1 after two weeks of incubation (Figure 1B). After 13 weeks of incubation only a concentration of 71 µg L-1 of PAA was detectable, hence more than 95% PAA had been removed from solution. The total arsenic content showed a similar trend with an overall arsenic elimination of approximately 77%. During the first and fourth week of cultivation an unknown species occurred in concentrations up to 235 µg L-1 As as an intermediate, and at a retention time of 1650 s (Figure 1B). Figure 2 illustrates a chromatogram for arsenic species analysis by HPLC-ICP-MS for the two nonsterile microcosms after two weeks of cultivation. The unknown compound, which is named 1650 s due to the retention time in the chromatogram, is observed as a clear peak. In contrast to PAA, the concentration of DPAA was stable during the whole cultivation period. In the living microcosms spiked with lactate (Figure 1C), the PAA concentration decreased within five weeks to values of approximately 20 µg L-1 As, at a rate of 0.245 µM L-1 d-1. Afterward no changes in the concentration were observed. The unknown arsenic compound with a retention time of 1650 s could not be detected. With the elimination of PAA from the liquid phase, PAO was formed with highest concentrations (316 µg L-1 As) after 14 days (see Figure 2). A reduction in the pentavalent arsenic of PAA takes place:
Within seven weeks of further cultivation PAO was also removed from the liquid phase. Like the unknown arsenic species in the untreated living microcosms, PAO was a metabolite in the lactate-spiked microcosms during the transformation process of PAA. In contrast, DPAA remained stable for the whole cultivation period. In both biotic microcosm series, sulfate was used as the electron acceptor, as indicated through several points. Black precipitates formed in the microcosms in the course of incubation, which is typical for iron sulfide. The groundwater contained dissolved iron in a concentration of around 10 mg 6992
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 18, 2009
TABLE 2. Final Concentration (Mean Value of Three Replicates) and Standard Deviation of Total Arsenic and Sulfate in Liquid and Solid Phase of the Filtrated Samples at the End of Each Test Period liquid phase -1
microcosms
As (µg L )
sterile living living with lactate
2321 ( 125 758 ( 36 725 ( 140
2-
SO4
solid phase -1
(mg L )
163 ( 1 123 ( 1 13 ( 8
As (µg g-1) 147 ( 2 302 ( 13 309 ( 13
L-1 (Table 1), which evidently precipitated during the experiment. Sulfide was determined in low concentrations (maximum values up to 1.8 mg L-1) and the samples had the typical odor of hydrogen sulfide. Correspondingly, the biotic microcosms contained fewer sulfates than abiotic controls after 13 weeks of incubation (Table 2). In sterile microcosms, the sulfate content remained stable. Since the arsenic species were chemically stable under sterile conditions, the observed transformations of arsenic compounds in the living microcosms are considered to be the effect of microbial activity and the resulting reductive conditions. Due to the anoxic incubation conditions and the production of sulfide in the course of incubation, the transformations were independent of molecular oxygen. Whereas in the untreated living microcosms the elimination of PAA worked through an unknown arsenic species, the elimination pathway in the lactate spiked microcosm is assumed to work through the direct reduction of the pentavalent arsenic in PAA to the trivalent form in PAO (Figure 1). The PAA concentration curves have the shape of first-order kinetics. The living microcosms showed generally no lag-phase. The sulfide formation and the PAA transformation respectively occurred within the first week after setup, indicating that the microorganisms were already adapted to the prevailing in situ conditions. Comparing the kinetics of the PAA removal in both living microcosms, a clear enhancement of the reaction rate through the addition of lactate is obvious. Assuming first order kinetics, the correlation coefficients for the PAA concentration data of the first 35 days are significant (0.92 and 0.99). The rate constants are 0.037 and 0.134 d-1 for the microcosms without and with additional lactate, respectively. The reaction was clearly accelerated by the addition of lactate by a factor of 3.7. The total arsenic content was reduced up
TABLE 3. Total Arsenic Balance with Standard Deviation for the Whole Test Media (Based on Used Masses/Volumes)a
microcosms sterile
living
solid phase leachable by liquid phase (XRF values) 0.1 M H3PO4 recovery (µg) (µg) (µg) (%)
start end
147 ( 0 232 ( 12
419 ( 9 257 ( 7
162 ( 8
115
start end
147 ( 0 76 ( 4
440 ( 32 554 ( 17
55 ( 3
117
147 ( 0 72 ( 14
415 ( 91 531 ( 103
32 ( 6
113
living with lactate start end
Recovery is calculated as the sum of liquid phase + solid phase + leachable by 0.1 M H3PO4 at the end divided by the sum of liquid phase + solid phase at the beginning. a
to 81% of the initial concentration. In the end, the total concentration was dominated by DPAA, which was not degradable under the given conditions. Further investigations with anoxic groundwater/sediment microcosms from other depths of the drilling profile showed similar trends (data not shown), i.e., the presented results could be reproduced. Total Arsenic Content and Mass Balance. Table 2 shows the final concentration, after 13 weeks of incubation, for total arsenic in solid and liquid phase and for sulfate in liquid phase. The total arsenic concentrations, compared in liquid and solid phase, substantiate the fact that the removed PAA was immobilized at the sediments. The elimination of arsenic, especially PAA from the liquid phase, occurred for both living cultures in similar amounts, seen in the degradation curves and the arsenic concentrations in the sediments. Through the addition of lactate this process could be accelerated (Figure 1C). The leaching of sediments with phosphoric acid, a soft extraction by ion exchange and slight decrease in pH, at the end of each test series should provide information about the bond strength of the eliminated arsenic species (23). The leachable fraction demonstrates what quantity of arsenic is easy to mobilize from the sediment. In the extracts, the main arsenic compounds were arsenate, arsenite, and PAA. DPAA was detected only in very small amounts with no significant differences between the variants of microcosms (data not shown). This points to the stability of DPAA in solution. From the living microcosms, some amounts of the unknown arsenic compound (1650s) could also be mobilized. No other species were analyzed in substantial concentrations and the sum of the species was in agreement with the total As concentration. The highest leachable fraction of arsenic was found in the sediments of the sterile microcosms. The untreated living microcosms showed higher values of the leachable fraction than the lactate-spiked living microcosms (Table 3). Hence, arsenic species were bonded stronger to the sediment through microbial activity, which can be tightened by an extra nutrient (lactate). The treatment of the samples at the end of each test series (shaking before filtering) leads also to a mobilization of arsenic from the sediment. Comparing the total arsenic concentration, calculated through accumulation of the arsenic species from the concentration curves (Figure 1), with the total arsenic content of the filtrated samples (Table 2) at the end of the test period, this fact can be substantiated. Considering the volume of the test series and the mass of the containing sediment a balance for total arsenic in the whole test medium was calculated (Table 3). Comparing the arsenic content at the beginning and the end of each test series, recovery rates between 113% and 117% were achieved. The end values for arsenic are slightly higher due to the propagation of the different analytical errors in each analysis
(sum of up to six species in three extraction steps). However, taking into account the analytical errors, the eliminated amounts of PAA in the living microcosms were found in equal ratios in the sediments. Experiments for Identification of Unknown Arsenic Species. The elimination of PAA proceeded over partial unidentified intermediates under sulfate-reducing conditions. An identified intermediate was trivalent PAO in the lactate-spiked microcosms (Figures 1 and 2), which was probably formed by PAA-reducing bacteria. It is known that the pentavalent inorganic arsenic can be used by anaerobic bacteria as the terminal electron acceptor (3). In the nonspiked test series the arsenic species at a retention time of 1650 s occur as a temporal metabolite. Because of the relatively late retention time of the unknown species (1650 s), this compound is hypothesized to have a dimeric structure with two PAA molecules possibly combined, e.g., by a sulfur bridge. Possible structures of this unknown arsenic species are
For inorganic arsenic the formation of sulfur complexes in an anoxic sulfidic environment is known. In this case the dominant forms of inorganic sulfur complexes are thioarsenates, i.e., arsenic in the pentavalent form (24). Thioarsenite species, i.e., inorganic arsenic in the reduced trivalent form, can be formed as well, but those compounds are not persistent in iron containing environments (25). The formation of dimeric or even polymeric structures of phenyl arsenicals with S-As-bridges was shown before in organic solvents (26, 27). However, their formation or stability in water samples is not described yet. In what valences the unknown arsenic species in the living microcosms occurred is not yet known. It is assumed that the substitution of oxygen with sulfur in PAA is the trigger of the elimination of PAA from the groundwater, which is supposed to work over a sulfur-based immobilization onto the sediments under the sulfate reducing conditions. The sorption of As to Fe(II)sulfides such as mackinawhite or pyrite is described for inorganic As-species as well (28). To get further information on the structure of the unknown arsenic species with a retention time of 1650 s, investigations with HPLC-ESI-MS, oxidation experiments and experiments with a sulfate reducing bacterial strain were applied. No structural information about the unknown arsenic species could be gained by linking the HPLC with an ESI-MS (11). Neither a positive nor a negative ionization of the compound could be achieved. For control purposes the peaks of PAA and DPAA were also analyzed and the corresponding masses could be verified (11). Under the applied conditions, the unknown arsenic species seems to be nonionizable. Considering first-order kinetics, a chemical or biochemical reaction might be responsible for the decrease in PAA concentration. The prevailing sulfate-reducing conditions might lead to favorable conditions for a chemical transformation of PAA. The higher rate in the presence of lactate could be explained by a lower redox potential in these samples and/or the higher sulfide concentration in the water. To verify these results, we set up different biotic and abiotic sulfidecontaining microcosms and spiked them with PAA. The microcosms also contained living, killed, or no cells of a strain of the sulfate-reducing species D. gigas. The initial concentration of sulfide was between 5.7 and 7.2 mg L-1 (179-226 µM), i.e., more than 4 times higher than the maximal sulfide concentration in the groundwater/sediment microcosms. An abiotic, sulfide-free control contained only mineral salt medium with sulfate, lactate and PAA. Samples VOL. 43, NO. 18, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6993
FIGURE 3. Chromatograms of HPLC-ICP-MS measurements for arsenic detection, taken from microcosms inoculated with D. gigas and spiked with PAA. The microcosms contained sulfide in concentrations between 179 and 226 µM. The samples for analysis were taken within an hour after addition of PAA. (A): control without D. gigas and sulfide (B): inoculated with D. gigas (C): inoculated with D. gigas and autoclaved (D): amended with a sterile-filtered solution of D. gigas. for PAA analysis were taken 1 h after addition of PAA. No significant change in PAA concentration occurred in the sulfide-free, sterile control bottle as seen in the chromatogram in Figure 3A. In all sulfide-containing microcosms, PAA was transformed within 1 h to the unknown arsenic species, regardless whether living D. gigas cells were present or not (Figure 3B-D). These results indicate that the unknown arsenic species was chemically formed by a reaction of PAA with sulfide. The oxidation experiment showed that the unknown arsenic species was immediately eliminated with increasing concentrations of PAA after adding H2O2, whereas the concentration in the control sample, which received no H2O2, remained stable (data not shown). Changes in other species were not significant; the main structure of PAAsthe phenyl-arsenic-bondswas persistent. The addition of H2O2 is a common technique to prove the presence of sulfur-arsenic-species, so-called thioarsenates, by their destruction; these species are easily oxidized to the oxoanions of arsenic under these conditions (29). Thus, the results of the oxidation experiment indicate that the unknown arsenic species is a thioarsenate with intact phenyl-arsenic-structure. Environmental Implications. In our microcosm study, we demonstrate that the concentration of organic arsenicals in anoxic groundwater samples can be significantly reduced by a bacterial-mediated process. The removal mechanism seems to be a bonding of a phenyl arsenic compound onto the sediment. This process is induced by sulfide which is formed by the activity of sulfate-reducing bacteria, verified with sterile control microcosms in which no sulfate reduction and PAA transformation occurred, respectively. The extent to which the microbial community is important for the fixation onto the sediment is unknown at this time and will be investigated in further experiments. Noticeably, spiking of lactate as an additional nutrient resulted in elimination of PAA via the trivalent PAO as metabolite, indicating that microcosms may use the pentavalent arsenic compound (PAA) as a terminal electron acceptor (3). However, the PAO is also removed from the solution by an additional reaction step which might be a chemical substitution of oxygen with sulfur. Hence, our results suggest that both the pentavalent PAA as well as the trivalent PAO can react with sulfide, leading to phenyl-arsenic thioarsenates which are immobilized in sediments. The stability of the immobilization products is an important question for the evaluation of the process for a natural attenuation strategy based on the stimulation 6994
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 18, 2009
of sulfate reduction within the contaminated aquifer, e.g., by addition of sulfate and appropriate electron donors. Under sulfate-reducing conditions a remobilization was not observed in our experiments. However, the instability of inorganic thioarsenates under oxidizing conditions is well-known (24). This has to be taken into consideration for a remediation approach. Nevertheless, the total arsenic content could be considerably reduced by more than 75% in our microcosm experiments. These results strongly indicate a natural attenuation potential for phenyl arsenicals at the investigated field site by the autochthonous microorganisms under sulfate-reducing conditions which can be enhanced. To our knowledge, we have hereby verified with this work, for the first time, a natural attenuation potential for the heretofore persistent phenyl arsenicals in anaerobic groundwater.
Acknowledgments We are grateful to Dr. J. Mattusch for his support with the ESI-MS analyses, and we thank the Department of Analytical Chemistry for their assistance. We thank the Deutsche Bundesstiftung Umwelt (grant 20007/919) and the Landesanstalt fu ¨ r Altlastenfreistellung (LAF) of the state Saxony-Anhalt for financial support.
Literature Cited (1) Roy, P.; Saha, A. Metabolism and toxicity of arsenic: A human carcinogen. Curr. Sci. 2002, 82 (1), 38–45. (2) Edmonds, J. S.; Francesconi, K. A. The origin of arsenobetaine in marine animals. Appl. Organometal. Chem. 1988, 2, 297– 302. (3) Oremland, R. S.; Stolz, J. F. The ecology of arsenic. Science 2003, 300, 939–944. (4) Martinetz, D. Der Gaskrieg 1914-1918. Entwicklung, Herstellung und Einsatz chemischer Kampfstoffe; Bernard & Graefe: Bonn, 1996. (5) Bausinger, T.; Preuβ, J. Environmental remnants of the first world war: soil contamination of a burning ground for arsenical ammunition. Bull. Environ. Contam. Toxicol. 2005, 74 (6), 1045– 1053. (6) Hanaoka, S.; Nagasawa, E.; Nomura, K.; Yamazawa, M.; Ishizaki, M. Determination of diphenylarsenic compounds related to abandoned chemical warfare agents in environmental samples. Appl. Organometal. Chem. 2005, 19, 265–275. (7) Ishizaki, M.; Yanaoka, T.; Nakamura, M.; Hakuta, T.; Ueno, S.; Komuro, M.; Shibata, M.; Kitamura, T.; Honda, A.; Doy, M.; Ishii, K.; Tamaoka, A.; Shimojo, N.; Ogata, T.; Nagasawa, E.; Hanaoka, S. Detection of bis(diphenylearsine)oxide, diphenylarsinic acid and phenylarsonic acid, compounds probably derived from chemical warfare agents, in drinking well water. J. Health Sci. 2005, 51 (2), 130–137.
(8) Wada, T.; Nagasawa, E.; Hanaoka, S. Simultaneous determination of degradation products related to chemical warfare agents by high-performance liquid chromatography/mass spectrometry. Appl. Organometal. Chem. 2006, 20, 573579. (9) Garnaga, G.; Wyse, E.; Azemard, S.; Stankevicˇius, A.; Mora, S. d. Arsenic in sediments from the southeastern Baltic Sea. Environ. Pollut. 2006, 144 (3), 855–861. (10) Deutscher Bundestag. Drucksache 13/2733, 1995. Available at http://dip.bundestag.de/btd/13/027/1302733.asc. (11) Daus, B.; Mattusch, J.; Wennrich, R.; Weiss, H. Analytical investigations of phenyl arsenicals in groundwater. Talanta 2008, 75 (2), 376–379. (12) Nakajima, T.; Kawabata, T.; Kawabata, H.; Takanashi, H.; Ohki, A.; Maeda, S. Degradation of phenylarsonic acid and its derivatives into arsenate by hydrothermal treatment and photocatalytic reaction. Appl. Organometal. Chem. 2005, 19, 254–259. (13) Sieke, R. W.; Lippke, G.; Krippendorf, A.; Haas, R.; Lu ¨ dtke, S. Destruction of Diphenylarsine Chloride (CLARK I) with Activated Ozone. Environ. Sci. Pollut. Res. 1998, 5 (4), 199–201. (14) Haas, R.; Scheibner, K.; Hofrichter, M. Enzymatische Umsetzung von Arsenkampfstoffen durch das Pilzenzym Mangan-Peroxi¨ kotox. 2003, 15 (4), 224–226. dase. UWSF - Z. Umweltchem. O (15) Hofmann, K.; Hammer, E.; Ko¨hler, M.; Bru ¨ ser, V. Oxidation of triphenylarsine to triphenylarsineoxid by Trichoderma harzianum and other fungi. Chemosphere 2001, 44 (4), 697–700. (16) Nakamiya, K.; Nakayama, T.; Ito, H.; Edmonds, J. S.; Shibata, Y.; Morita, M. Degradation of arylarsenic compounds by microorganisms. FEMS Microbiol. Lett. 2007, 274, 184–188. (17) Dietrich, P.; Leven, C. Direct push-technologies. In Groundwater Geophysics. A Tool for Hydrogeology; Kirsch, R. , Ed.; Springer: Berlin, 2006; Vol. XVIII, pp 321-340. (18) Vogt, C.; Go¨deke, S.; Treutler, H.-C.; Weiβ, H.; Schirmer, M.; Richnow, H.-H. Benzene oxidation under sulfate-reducing conditions in columns simulating in situ conditions. Biodegradation 2007, 18 (5), 625–636.
(19) Daus, B.; Mattusch, J.; Wennrich, R.; Weiss, H. Investigation on stability and preservation of arsenic species in iron rich water samples. Talanta 2002, 58, 57–65. (20) Cline, J. D. Spectrophotometric determination of hydrogen sulfide in natural waters. Limnol. Oceanogr. 1969, 14 (3), 454– 458. (21) Kleinsteuber, S.; Schleinitz, K. M.; Breitfeld, J.; Harms, H.; Richnow, H. H.; Vogt, C. Molecular characterization of bacterial communities mineralizing benzene under sulfate-reducing conditions. FEMS Microbiol. Ecol. 2008, 66, 143–157. (22) Goldscheider, N.; Hunkeler, D.; Rossi, P. Review: Microbial biocenoses in pristine aquifers and an assessment of investigative methods. Hydrogeol. J. 2006, 14 (6), 926–941. (23) Orero-Iserte, L.; Roig-Navarro, A. F.; Herna´ndez, F. Simultaneous determination of arsenic and selenium species in phosphoric acid extracts of sediment samples by HPLC-ICP-MS. Anal. Chim. Acta 2004, 527 (1), 97–104. (24) Stauder, S.; Raue, B.; Sacher, F. Thioarsenates in Sulfidic Waters. Environ. Sci. Technol. 2005, 39 (16), 5933–5939. (25) Wilkin, R. T.; Wallschla¨ger, D.; Ford, R. G. Speciation of arsenic in sulfidic waters. Geochem. Trans. 2003, 4 (1), 1–7. (26) Cordes, A. W.; Gwinup, P. D.; Malmstrom, M. C. The crystal and molecular structure of diphenyldiarsenic trisulfide. a fivemembered arsenic-sulfur ring compound. Inorg. Chem. 1972, 11 (4), 836–838. (27) Tani, K.; Hanabusa, S.; Kato, S.; Mutoh, S.; Suzuki, S.; Ishida, M. Thioacylsulfanylarsines (RCS2)xAsPh3-x, x ) 1-3: synthesis, structures, natural bond order analyses and reactions with piperidine. J. Chem. Soc., Dalton Trans. 2001, 518–527. (28) Wolthers, M.; Charlet, L.; Weijden, C. H. v. d.; Linde, P. R. v. d.; Rickard, D. Arsenic mobility in the ambient sulfidic environment: Sorption of arsenic(V) and arsenic(III) onto disordered mackinawite. Geochim. Cosmochim. Acta 2005, 69 (14), 3483–3492. (29) Beak, D. G.; Wilkin, R. T.; Ford, R. G.; Kelly, S. D. Examination of arsenic speciation in sulfidic solutions using X-ray absorption spectroscopy. Environ. Sci. Technol. 2008, 42 (5), 1643–1650.
ES9006788
VOL. 43, NO. 18, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6995
