Carol A. Schwegel

Extraction and detection of a new arsine sulfide containing arsenosugar in molluscs by IC-ICP-MS and IC-ESI-MS/MS

Using IC-ICP-MS and IC-ESI-MS/MS, an unknown arsenical compound in molluscs has been identified as a new arsine sulfide containing analog of a known arsenosugar and is referred to as As(498). This species has been observed in four separate shellfish species following a mild methanol–water extraction. As(498) is unstable, especially in acid, and converts to the arsine oxide containing arsenosugar As(482) over time. Chromatographic retention of As(498) was observed on an anion exchanger ION-120 column but the species did not elute as a well defined peak from a PRP-X100. Mass spectrometric analysis of As(498) at pH 9.0 produced an [M–H]− species at a mass to charge of 497 in the negative-ion mode. A synthetic standard of As(498) was made by bubbling hydrogen sulfide into a stock solution of arsenosugar As(482). The retention time and ESI-MS/MS data were identical for the synthetic standard of As(498) and the unknown arsenical in shellfish extracts.

An evaluation of sample dispersion media used with accelerated solvent extraction for the extraction and recovery of arsenicals from LFB and DORM-2

An accelerated solvent extraction (ASE) device was evaluated as a semi-automated means for extracting arsenicals from quality control (QC) samples and DORM-2 [a standard reference material (SRM)]. Unlike conventional extraction procedures, the ASE requires that the sample be dispersed in an inert dispersion media prior to the extraction. The need to disperse the sample in a support matrix prior to extraction is demonstrated by a 58% reduction in the extraction efficiency of arsenic (AsExtraction Efficiency) from DORM-2 if the sample is not homogeneously suspended in the dispersion media. Three dispersion media (Filter Aid, Q-Beads, Teflon) were evaluated in terms of the arsenic extraction recoveries (AsExtraction Recovery) from laboratory fortified blanks (LFB) and matrices (LFM). The arsenicals investigated were arsenobetaine (AsB), arsenite [As(III)], dimethylarsinic acid (DMA), disodium methylarsenate (MMA) and arsenate [As(V)]. The first dispersion medium, Filter Aid (a high density glass bead, ∼0.04 mm particle size) produced near-quantitative removal of As(III) and As(V), while ∼90% of the MMA and ∼10% of the DMA was removed by the Filter Aid. This is contrary to the 99.6% ± 0.8 (±2σ) AsExtraction Recovery obtained for an AsB in a LFB. The lack of retention of AsB on the Filter Aid is consistent with the 94.8% ± 11.8 (±2σ) AsExtraction Efficiency for DORM-2 which contains predominately AsB. Fortifying DORM-2 with an AsB, As(III), As(V), MMA and DMA mixture produces a LFM AsExtraction Recovery of 70.0% ± 9.8 (±2σ). The low AsExtraction Recoveries are a result of the Filter Aid binding and retaining certain anionic arsenicals. The second dispersion media, Q-beads (a soda lime glass bead, 0.8 mm spherical), did not exhibit these binding interactions. The AsExtraction Recovery in a LFB for Q-Beads was 98.7% ± 5.6 (±2σ). The AsExtraction Efficiency and the LFM AsExtraction Recovery from DORM-2 using Q-beads were 88.2% ± 14.6 and 83.2% ± 11.8 (±2σ), respectively. Finally, two Teflon particle sizes were evaluated (250 µm and 1000 µm diameter) as dispersion media. The Teflon dispersion media produced LFB AsExtraction Recoveries of 98.7% ± 3.4 (±2σ) (for the 250 µm particle) and 95.2% ± 3.6 (±2σ) (for the 1000 µm particle). The AsExtraction Efficiencies for DORM-2 were 88.6% ± 2.2 and 84.8% ± 2.2 (±2σ) for the 250 µm and 1000 µm Teflon dispersion media, respectively. AsExtraction Recoveries of the LFMs for DORM-2 using the 250 µm and 1000 µm Teflon were 101.9% ± 2.6 and 101.5% ± 1.8 (±2σ), respectively. ASE cell components and the ASE solvent reduction vials were evaluated for potential analyte losses. The stainless steel frit located at the exit of the ASE cell was found to bind up to 13% of the arsenicals present in a LFB. The resolubilization of As(III) from the Pyrex ASE vials using 18 MΩ water required greater than 8 h. The slow resolubilization after the solvent reduction was not observed for the other analytes.

An in vitro assessment of bioaccessibility of arsenicals in rice and the use of this estimate within a probabilistic exposure model

In this study, an in vitro synthetic gastrointestinal extraction protocol was used to estimate bioaccessibility of different arsenicals present in 17 rice samples of various grain types that were collected across the United States. The across matrix average for total arsenic was 209 ng/g±153 ([xmacr ]±2σ). The bioaccessibility estimate produced an across matrix average of 61%±19 ([xmacr ]±2σ). The across matrix average concentrations of inorganic arsenic (iAs) and dimethylarsinic acid (DMA) were 81 ng/g±67.7 and 41 ng/g±58.1 ([xmacr ]±2σ), respectively. This distribution of iAs concentrations in rice was combined with the distribution of consumption patterns (from WWEIA) in a Stochastic Human Exposure and Dose Simulator model to estimate population-based exposures. The mean consumption rate for the population as a whole was 15.7 g per day resulting in a 0.98 μg iAs per day exposure. The mean consumption rate for children 1–2 years old was 7 g per day resulting in a 0.48 μg iAs per day exposure. Presystemic biotransformation of DMA in rice was examined using an in vitro assay containing the anaerobic microbiota of mouse cecum. This assay indicated that DMA extracted from the rice was converted to dimethylthioarsinic acid, although a second oxygen–sulfur exchange to produce DMDTA was not observed.

Investigation of arsenic speciation on drinking water treatment media utilizing automated sequential continuous flow extraction with IC-ICP-MS detection

Three treatment media, used for the removal of arsenic from drinking water, were sequentially extracted using 10 mM MgCl2(pH 8), 10 mM NaH2PO4(pH 7) followed by 10 mM (NH4)2C2O4(pH 3). The media were extracted using an on-line automated continuous extraction system which allowed the arsenic in each of the extraction fluids to be speciated on-line using IC-ICP-MS. The 10 mM MgCl2 preferentially extracted As(III) from each of the media. The percentage of the arsenic extracted by the MgCl2, relative to a HNO3/H2O2 digestion of the media, ranged from 0.1-2.3% for the three solids. The next sequential extraction fluid, 10 mM NaH2PO4, extracted some of the residual As(III) remaining on each of the media but the predominant species extracted was As(V). The 10 mM NaH2PO4 extracted 15.3 to 42.8% of the total arsenic relative to a total digested concentration for each of the media. The As(III) and As(V) stability studies conducted in these two extraction fluids indicated that conversion between As(III) and As(V) was not significant for the short extraction fluid sample contact time associated with the on-line continuous flow extraction cell. Finally, the 10 mM (NH4)2C2O4 extraction fluid was utilized in an off-line analysis mode because the Fe and As concentrations extracted from the media were not compatible with direct ICP-MS detection. The (NH4)2C2O4 extracted 2.9-29% As(III) for all three media and caused an oxidation of As(III) to As(V) during the extraction period for one of the three media. The sum of the arsenic from each of the three extraction fluids represented 92%, 44% and 53% of the available total arsenic for the three media, respectively. The speciation results for each media were obtained by adding all the speciation results from all three extraction fluids together and the resulting distribution of As(III)/As(V) compared well with the speciation results obtained via XANES.

Investigation of sequential and enzymatic extraction of arsenic from drinking water distribution solids using ICP-MS

A sequential extraction approach was utilized to estimate the distribution of arsenite [As(iii)] and arsenate [As(v)] on iron oxide/hydroxide solids obtained from drinking water distribution systems. The arsenic (As) associated with these solids can be segregated into three operationally defined categories (exchangeable, amorphous and crystalline) according to the sequential extraction literature. The exchangeable As, for the six drinking water solids evaluated, was estimated using 10 mM MgCl(2) and 10 mM NaH(2)PO(4) and represented between 5-34% of the total As available from the solid. The amorphously bound As was estimated using 10 mM (NH(4))(2)C(2)O(4) and represented between 57-124% of the As available from the respective solid. Finally, the crystalline bound As was estimated using titanium citrate and this represented less than 1.5% of the As associated with the solids. A synthetic stomach/intestine extraction approach was also applied to the distribution solids. The stomach fluid was found to extract between 0.5-33.3 microg g(-1) As and 120-2,360 microg g(-1) iron (Fe). The As concentrations in the intestine fluid were between 0.02-0.04 microg g(-1) while the Fe concentration ranged from 0.06-0.7 microg g(-1) for the first six drinking water distribution solids. The elevated Fe levels associated with the stomach fluid were found to produce Fe based precipitates when the intestinal treatment was applied. Preliminary observations indicate that most of the aqueous Fe in the stomach fluid is ferric ion and the observed precipitate produced in the intestine fluid is consistent with the decreased solubility of ferric ion at the pH associated with the intestine.