Environ. Sci. Technol. 2008, 42, 228–234
Determination of Methylated Arsenic-Sulfur Compounds in Groundwater DIRK WALLSCHLÄGER* AND JACQUELINE LONDON Environmental & Resource Sciences Program, Trent University, 1600 West Bank Dr, Peterborough, ON K9J 7B8, Canada
Received April 02, 2007. Revised manuscript received September 13, 2007. Accepted September 14, 2007.
Arsenic speciation was determined by anion-exchange chromatography-inductively coupled plasma-mass spectrometry (AEC-ICP-MS) in groundwater samples collected from an aquifer impacted by methylated As pesticides. Besides the four expected arsenic species AsO33-, AsO43-, (CH3)AsO32- and (CH3)2AsO2-, up to nine other arsenic species were encountered, which constituted a major fraction of the total arsenic concentration in most samples. We then synthesized the thioderivatives of (CH3)AsO32- and (CH3)2AsO2-, and characterized the formed products by electrospray-tandem mass spectrometry (ES-MS-MS). The presence of (CH3)AsO2S2-, (CH3)AsOS22-, (CH3)2AsOS- and (CH3)2AsS2-, was confirmed in the groundwater by retention time matching plus ES-MS-MS in collected AEC fraction, and the presence of the trivalent methylated arsenic species (CH3)AsO22- was suggested based on retention time matching only. These arsenic species have not been observed in ambient waters before, and are likely to occur in many environments containing methylated arsenic species and reduced sulfur compounds. They can persist in some of these particular samples for periods of up to six months without preservation, but tend to convert into the corresponding pentavalent oxyspecies. Acidification with HCl shifted speciation equilibria rapidly, and is thus unsuitable for stabilizing samples containing these novel arsenic species; cryofreezing or no sample preservation avoided this artifact.
Introduction Preface. The following abbreviations are used for the discussed arsenic species throughout this manuscript to improve clarity of presentation: AsO33- ) arsenite ) As(III); AsO43- ) arsenate ) As(V); AsO3S3- ) monothioarsenate ) MTAs(V); AsO2S23- ) dithioarsenate ) DTAs(V); AsOS33- ) trithioarsenate ) TTAs(V); (CH3)AsO22- ) monomethylarsenite ) MMAs(III); (CH3)AsO32- ) monomethylarsenate ) MMAs(V); (CH3)AsO2S2- ) monomethylmonothioarsenate ) MMMTAs(V); (CH3)AsOS22- ) monomethyldithioarsenate ) MMDTAs(V); (CH3)AsS32- ) monomethyltrithioarsenate ) MMTTAs(V); (CH3)2AsO- ) dimethylarsenite ) DMAs(III); (CH3)2AsO2- ) dimethylarsenate ) DMAs(V); (CH3)2AsOS) dimethylmonothioarsenate ) DMMTAs(V); (CH3)2AsS2) dimethyldithioarsenate ) DMDTAs(V); regardless of actual protonation state, only the fully deprotonated forms of all species are shown here for simplicity. * Corresponding author phone: (705) 748-1011 × 7378; fax: (705) 748-1569; e-mail:
[email protected]. 228
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 1, 2008
Arsenic (As) contamination of ground waters in contact with drinking water sources is a concern in many areas of the world, due to the carcinogenicity of the inorganic As species As(III) and As(V) (1). Although the pentavalent methylated As species, MMAs(V) and DMAs(V), are routinely measured in speciation studies, they are typically not paid much attention in the associated human risk assessments because they are not carcinogenic, much less toxic than As(III) and As(V), and usually only found in minor quantities as the result of biological methylation (2). However, these two species were also used as biocides (3), so in ecosystems where substantial quantities of these chemicals were brought out, their relative importance in comparison to As(III) and As(V) may increase. It was established long ago that the metabolism of As oxyanions in organisms involves a sequence of alternating reductions and oxidative methylations, leading to the formation of the trivalent methylated As compounds MMAs(III) and DMAs(III) as intermediates (2). Since recent studies discovered that these species exhibit in vitro genotoxicity (4), the original hypothesis that methylation of inorganic As species is a detoxifying process has come under scrutiny. Furthermore, it was shown that some sulfur analogs of methylated As oxyanions are also formed during this metabolism (5–7); the toxicological properties of these compounds are still unclear to date, but there is some evidence that methylated As–S compounds are significantly more toxic than their oxyanion counterparts (8). If such previously ignored methylated As species of known or potential toxicological concern exist in organisms, then it appears warranted to examine if they also occur in environmental waters and soils/sediments, and what their roles and associated risks are. The discovery of these novel methylated As species in organisms was made possible by using liquid chromatography separations (coupled to atomic spectrometry detection); MMAs(III) and DMAs(III) have been separated from the “conventional As species” MMAs(V), DMAs(V), As(III), and As(V) by ion-pairing (9) and anion-exchange chromatography (AEC) (10), while methylated As–S compounds were separated from As oxyanions by gel filtration (5), reversedphase HPLC (11), and AEC (7). In many previous studies of As speciation in environmental samples, hydride-generation– gas-chromatography was used as the separation technique (12), so the fact that neither trivalent methylated As oxyanions nor methylated As-thioanions have been observed before by this approach suggests that it may be incapable of differentiating between such species and the pentavalent methylated As oxyanions. Recent studies have demonstrated that thio-derivatives of the inorganic As oxyanions exist in ambient waters containing reduced sulfur compounds, such as sulfide (13, 14). However, due to the typical quantitative irrelevance of methylated As oxyanions in ambient waters and soils/ sediments, there is to date no report of the existence of trivalent methylated As compounds or methylated As-S compounds in those matrices. In this study, we examined a reducing aquifer impacted directly by methylated As biocides by AEC-ICP-MS, and present the first evidence of the presence of both trivalent methylated As compounds and methylated As–S compounds in an ambient water.
Experimental Section Analytical Methods. Arsenic species were separated by AEC using NaOH as the eluant (see Table 1), and detected online by ICP-MS. Figure 1 shows the baseline separation of 13 As 10.1021/es0707815 CCC: $40.75
2008 American Chemical Society
Published on Web 11/27/2007
TABLE 1. Optimized Parameters of the AEC-ICP-MS Method separation columns eluant gradient flow rate sample volume
typical retention times [min]
Ion Pack AG-16 + AS-16 4-mm (12–32) (Dionex, Sunnyvale, CA) NaOH (0.1 mol L-1) 0–3 min 2.5 mmol L-1 3–5 min 2.5 f 20 mmol L-1 5–10 min 20 mmol L-1 10–20 min 20 f 100 mmol L-1 20–24 min 2.5 mmol L-1 1.2 mL min-1 25 µL US1 ) 2.6 (dead volume); MMAs(III) ) 3.4; DMAs(V) ) 4.5; DMMTAs(V) ) 6.2; As(III) ) 7.0; DMDTAs(V) ) 8.7; MMAs(V) ) 9.8; MMMTAs(V) ) 11.5; MMDTAs(V) ) 15.2; US2 ) 16.0; As(V) ) 16.5; US3 ) 17.4; MTAs(V) ) 18.0 detection standard mode (As only)
instrument plasma RF power nebulizer flow rate lens Voltage dwell times
Elan DRCII (Perkin-Elmer, Shelton, CT) 1,300 W 1.05 - 1.1 mL min-1 (optimized daily) static lens, around 8 V (optimized daily) m/z ) 75 (for quantification) 700 ms, m/z ) 77 and m/z ) 82 (for interference monitoring) 100 ms each detection DRC mode (As + S)
instrument plasma RF power nebulizer flow rate lens Voltage dwell times reaction gas cell parameters
Elan DRCII 1,300 W 1.05 - 1.1 mL min-1 (optimized daily) autolens, (optimized for AsO+ and SO+ daily) m/z ) 91 (AsO+) and m/z ) 48 (32S16O+) 500 ms each, m/z ) 50 (34S16O+)100 ms O2, 0.75 mL min-1 Rpq ) 0.6 (AsO+), 0.3 (SO+); Rpa ) 0 ES-MS-MS
instrument sample flow rate declustering/entrance potentials source/curtain gas ion spray voltage collision gas collision energy
API 2000 (Applied Biosystems, Mississauga, ON) 25 µL min-1 -90 V/-7 V 40/10 -3,000 V “high” 22 MMMTAs(V), 25 MMDTAs(V), 10 MMTTAs(V), 18 DMMTAs(V), 20 DMDTAs(V)
species achieved under these conditions; since not all encountered As species were available as reference compounds, the overlaid chromatograms for two real samples (after appropriate dilution) are shown instead. This column/ eluant combination evidently avoids problems with irreversible retention of some methylated As-S species on the popular PRP-X100 AEC column (Hamilton, Reno, NV) reported in other studies (7, 11). The chromatographic recovery () sum of all measured species/total As concentration in the identical sample) for 18 selected samples during the preservation/ stability study was 102.7 ( 7.1%, which demonstrates that no As species were bound irreversibly to the AEC column during analysis. Retention times for DMAs(III) and MMAs(III) were determined (see Supporting Information (SI), Figure S1a) using (CH3)AsI2 and (CH3)2AsI (courtesy of Chris Le), which were hydrolyzed in degassed water just prior to analysis, and retention times for DMMTAs(V), DMDTAs(V), MMMTAs(V), MMDTAs(V), and MTAs(V) were confirmed using standards synthesized in our laboratory (see below and ref 14). Retention times varied slightly in the analyzed groundwater samples, due to interferences from the complex matrix, but peak identities could always be assigned by using
DMAs(V), As(III), MMAs(V), and As(V) as “internal retention time markers” (after spiking, if necessary). Calibration was performed with mixed standard solutions of the four commercially available As species DMAs(V) (Sigma-Aldrich, Oakville, ON), As(III) (JT Baker, Phillipsburg, NJ), MMAs(V) (ChemService, West Chester, PA), and As(V) (Sigma-Aldrich) (see SI, Figure S1b); all other As species encountered were quantified by peak area comparison to the closest of these species. The validity of this approach is confirmed by the good match between the sum of all As species in a sample and its total As (TAs) concentration, which was independently determined in a parallel HNO3-preserved sample by inductively coupled plasma-dynamic reaction cellmass spectrometry (ICP-DRC-MS), using O2 as the reaction gas. This As speciation mass balance was 102.6 ( 7.8% for the samples analyzed in the site survey (SI, Table S1), showing that no major As species was missed, in accordance with our studies of inorganic As-S species (14). Contrary to our usual approach (15), we did not use eluant suppression prior to ICP-MS detection, because DMAs(V) is lost during the suppression step (14), so it is likely that some of the unusual As species encountered here would also VOL. 42, NO. 1, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
229
rather than providing actual concentrations. Samples were split immediately into one sample for total As determination, preserved with 1% (V/V) HNO3, and one for As speciation analysis, left unpreserved, but without a headspace. Samples were transported and stored cool and dark and analyzed within a week of collection.
Results and Discussion
FIGURE 1. Separation of commercially available (labeled in black) and unusual (labeled in gray;) As species by AEC-ICP-MS; chromatograms of two real diluted ground waters are overlaid and scaled for clarity; elution order (see Table 1 for retention times): 1 ) US1, 2 ) MMAs(III), 3 ) DMAs(III)/ DMAs(V), 4 ) DMMTAs(V), 5 ) As(III), 6 ) DMDTAs(V), 7 ) MMAs(V), 8 ) MMMTAs(V), 9 ) MMDTAs(V), 10 ) US2, 11 ) As(V), 12 ) US3, 13 ) MTAs(V) [“US” ) unidentified species”]. interact with the suppressor in an undesired manner. Arsenic was normally detected in the standard mode (m/z ) 75), which yielded instrumental detection limits around 0.2 µg L-1, slightly better than the dynamic reaction cell (DRC) mode, using O2 as the reaction gas, where As is detected as AsO+ (m/z ) 91) (16). Chromatographic signals were accepted to be As species if they showed signals at m/z ) 75 and 91 (in DRC mode) and no signal at m/z ) 77 (used to check for chlorine-based interferences, such chloride; cf. Figure 1). Additionally, sulfur was measured in the DRC mode as SO+ at m/z ) 48 and 50 (corresponding to 32S and 34S) (17) to check for the presence of sulfur in As species (see SI, Figure S2). Molecular mass spectra of the methylated arsenic-sulfur compounds were recorded by electrospray-tandem mass spectrometry (ES-MS-MS) in the negative ion mode (cf. Table 1), using a linear ion trap instrument (API 2000, Applied Biosystems, Toronto, ON) for sensitivity enhancement. ESMS-MS was used only offline with sample introduction by a syringe pump. Electrospray conditions were optimized for each species individually such that maximum signal of the molecular ion was obtained. Fragmentation conditions in the collision cell were chosen such that at least two fragment ions (where possible) had comparable intensity to the molecular ion. Synthesis of Methylated Arsenic-Sulfur Compounds as Analytical Standards. A saturated sulfide solution was prepared by bubbling H2S gas through deionized water for one hour. Then, MMAs(V) or DMAs(V) were added as their sodium salts (10 mg L-1 As), and after 30 min reaction (at pH 4), the solution was made alkaline with NaOH (20 mmol L-1; pH 12.3) to ensure that any formed MeAs-S species were present as anions for the ES-MS-MS analysis. The reaction mixtures were analyzed by ES-MS-MS at various times for up to 24 h to monitor the formation of the various observed reaction products. No attempts were made to isolate or purify the standards further. Ground Water Samples. Twenty groundwater samples were collected initially from wells in an aquifer impacted severely by methylated As pesticides, where MMAs(V) was the main component and DMAs(V) was a minor component, using a peristaltic pump with inline filtration (
