Chemosphere 88 (2012) 1196–1201

Contents lists available at SciVerse ScienceDirect

Chemosphere

journal homepage: www.elsevier .com/locate /chemosphere

Arsenic and chromium speciation in an urban contaminated soil

Gautier Landrot a,⇑, Ryan Tappero b, Samuel M. Webb c, Donald L. Sparks a

a Plant and Soil Sciences, University of Delaware, 152 Townsend Hall, Newark, DE 19716, USAb National Synchrotron Light Source (NSLS), Brookhaven National Laboratory, Upton, NY 11973, USAc Stanford Synchrotron Radiation Light Source, SLAC National Accelerator Laboratory, Menlo Park, CA 94025, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 21 June 2011Received in revised form 19 March 2012Accepted 22 March 2012Available online 18 April 2012

Keywords:ChromiumArsenicContaminationSoilSpeciationXAFS

0045-6535/$ - see front matter � 2012 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.chemosphere.2012.03.069

⇑ Corresponding author. Present address: EarthBerkeley National Laboratory, Berkeley, CA 94720, US+1 510 486 5686.

E-mail addresses: gjlandrot@lbl.gov (G. Landrot), rtsamwebb@slac.stanford.edu (S.M. Webb), dlsparks@u

The distribution and speciation of As and Cr in a contaminated soil were studied by synchrotron-basedX-ray microfluorescence (l-XRF), microfocused X-ray absorption spectroscopy (l-XAS), and bulk extendedX-ray absorption fine structure spectroscopy (EXAFS). The soil was taken from a park in Wilmington, DE,which had been an important center for the leather tanning industry along the Atlantic seaboard of theUnited States, until the early 20th century. Soil concentrations of As, Cr, and Pb measured at certain loca-tions in the park greatly exceeded the background levels of these heavy metals in the State of Delaware.Results show that Cr(III) and As(V) species are mainly present in the soil, with insignificant amounts ofCr(VI) and As(III). Micro-XRF maps show that Cr and Fe are distributed together in regions where theirconcentrations are diffuse, and at local spots where their concentrations are high. Iron oxides, whichcan reduce Cr(VI) to Cr(III), are present at some of these hot spots where Cr and Fe are highly concen-trated. Arsenic is mainly associated with Al in the soil, and to a minor extent with Fe. Arsenate may besorbed to aluminum oxides, which might have transformed after a long period of time into an As–Al pre-cipitate phase, having a structure and chemical composition similar to mansfieldite (AlAsO4�2H2O). Thelatter hypothesis is supported by the fact that only a small amount of As present in the soil was desorbedusing the characteristic toxicity leaching procedure tests. This suggests that As is immobilized in the soil.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

The city of Wilmington, in New Castle County, Delaware, wasone of the capitals of the leather tanning industry on the AmericanEast Coast, from the late 1800s to the early 1900s. In 2001, aninvestigation by the Delaware Department of Natural Resourcesand Environmental Control (DNREC) identified 128 tannery sitesin the Wilmington area, which were subsequently clustered into52 sites (link of investigation report in Supporting information).Although many of these sites, which are covered by buildingsand parking lots, have been cleaned up after the DNREC investiga-tion, some of them remain untreated due to their close proximityto residential areas. Since arsenic (As) compounds were used atvarious stages of the tanning process, this element is often foundat high concentrations in soils at former tanning factories, espe-cially the ‘‘beam houses’’, where the hides were de-haired in largevats of As solution, using realgar (a-As4S4). The tanning processalso employed As-based coloring agents such as Scheele’s green(CuHAsO3) and Paris green (Cu(AsO2)2Cu(C2H3O2)2), as well as

ll rights reserved.

Sciences Division, LawrenceA. Tel.: +1 510 486 4077; fax:

appero@bnl.gov (R. Tappero),del.edu (D.L. Sparks).

sodium arsenate leather preservatives. Additionally, the wastegenerated at the tanning factories often contained chromium(Cr), since the chrome-tanned leather technique was one of theleather tanning methods used in the second part of the 19th century,which employed Cr salts, including Cr(III) sulfate (Sreeram andRamasami, 2003). The latter chemical was often synthesized atthe tanning factories, by reducing sodium dichromate with sulfurdioxide.

The goal of this study is to determine the concentrations, distri-butions, and potential toxicity of Cr and As present in the soil ofChristiana Park, located on Church Street, adjacent to the ChristianaRiver, in Wilmington, using geochemical and synchrotron-basedtechniques. According to a report issued by DNREC (link in Sup-porting information), this park, about 6.6 acres in size, was createdin the early 20th century by landfill operations. Miscellaneousmaterials, including soil, brick, glass, garbage, coal ash, and slag,were used to cover up a native marsh and alluvial deposits fromthe Christiana River. Several industrial and commercial sites,including tanneries, surrounded the park until the early 20th cen-tury. According to DNREC, the 52 tannery sites identified in theWilmington area were located for the most part along Walnut,Tatnall, and Monroe Streets, near the Christiana River. Five of thesesites are less than two blocks from Christiana Park. Therefore, it islikely that some of the fill materials used to create the park werewastes generated from the surrounding tanning factories.

Fig. 1. Spatial distribution of Cr, As, Fe, Mn, Zn, and Ni in Soil 1 and 2 at depths of 0–20 cm and 20–40 cm below the surface. The relative concentrations of the elements areshown on the right in the colored bar scales, for each element (in number of fluorescence count). In the box on the left: spatial distribution of As and Zn in Soil 1 at depths 20–40 cm below the surface, with maximum of fluorescence counts multiplied by a factor of five.

G. Landrot et al. / Chemosphere 88 (2012) 1196–1201 1197

Chemical extraction procedures have been extensively em-ployed to determine elemental distribution of As and Cr in contam-inated solid waste (Cancès et al. 2008; Hopp et al. 2008; La Forceet al. 2000; Yang and Donahoe, 2007). However, the accuracy ofthese techniques is limited, since re-adsorption and redistributionof the target element can occur during the extraction procedures.Therefore, the concentration of the target element can be over orunder estimated (Ostergren et al., 1999; Van Herreweghe et al.,2003). Previous studies have reported that As and Cr in contami-nated soils are mainly associated with Fe and/or Al phases (Arconet al., 2005; Beaulieu and Savage, 2005; Bhattacharya et al., 2002;Cancès et al., 2008; Gao and Schulze, 2010; Hopp et al., 2008; Lundand Fobian, 1991). Iron oxides and iron(III) oxyhydroxides areimportant As and Cr scavengers because of their abundance inthe environment, high surface area, and chemical affinity for As(III)and As(V) compounds (Sparks, 2003; Vodyanitskii, 2009). Addi-tionally, Cr(III) can sorb to organic materials (Hopp et al., 2008),or co-precipitate with goethite (a-FeOOH) to form an a-(Fe,Cr)OOH phase, due to structural similarities between the hostFe(III) mineral and the pure Cr surface precipitate phase (a-CrOOH)(Charlet and Manceau, 1992; Hansel et al., 2003). Aluminum oxidescan sorb As(V), and this surface species can transform after a longperiod to a three dimensional As–Al precipitate phase (Arai andSparks, 2002). It has also been shown that As can sorb to calcite,sulfides, kaolinite, and montmorillonite (La Force et al., 2000).Lastly, manganese oxides can effectively sorb and oxidize As(III)or Cr(III), to produce As(V) or Cr(VI), the latter being the most toxicform of chromium (Fendorf and Zasoski, 1992; Parikh et al. 2008;Post, 1999; Weaver and Hochella, 2003). Conversely, Fe oxides, sul-fides, or dissolved organic substances can reduce Cr(VI) to Cr(III)(Fendorf and Guangchao, 1996; Peterson et al., 1996). Therefore,to assess the potential toxicity of Cr and As in the soil of Christiana

Park, one not only has to determine the oxidation states of theseelements, but also their compartmentalization in the soil and theirassociation with soil components that can potentially change theirchemical properties. This can be achieved using a combined l-XRF/l-XAS/ l-XRD approach. Accordingly, the co-localization of Cr/Mn,and Cr/Fe, and the distribution of Cr oxidation states are deter-mined in this study with l-XRF and l-XANES. Additionally, sinceFe and Mn can have opposing effects on Cr oxidation states, andalso due to the drastic variability in mineral oxidation capacityamong the different types of manganese oxides, the nature of theMn and Fe mineral phases at hot spots featuring Cr, Fe, and Mn,in the fluorescence maps can be determined with l-XRD. Lastly,the oxidation states of As and its local environment at the molec-ular scale are determined by bulk EXAFS spectroscopy. The infor-mation provided in this study about Cr and As speciation in thesoil of Christiana Park could be useful in deciding on the best reme-dial action strategy for other As and Cr contaminated sites.

2. Material and methods

Soil samples were collected at eight locations randomly chosenin the park. Samples were taken at two depths per location,between 0–20 cm, and 20–40 cm. Soil samples were digested withan aqua regia mixture, and elemental concentrations weremeasured with ICP-AES. Soil samples from two locations, labeled‘‘Soil 1’’ and ‘‘Soil 2’’, with the highest As and Cr concentrationswere selected for macroscopic and synchrotron-based analyses.The pH and Eh of each sample was determined by mixing �0.5 gsoil with �10 mL DDI water, settling for 20 min, and measuringthe supernatant using a combination electrode. The Synthetic Pre-cipitation Leaching Procedure (SPLP) and the Toxicity Characteris-tic Leaching Procedure (TCLP) for arsenic were applied to Soil 1 and

Fig. 2. (A) Cr and Fe correlation plots: Cr Ka intensities over the Fe Ka intensities per pixel in l-XRF maps of Soil 1 and 2 at 0–20 cm and 20–40 cm; and (B) Cr and Fe l-XRFmaps used to make the correlation plot of Soil 2 at 0–20 cm depicted in (A); and (C) Reconstructed Cr l-XRF maps after selecting with SMAK three sets of points – labeled (1),(2), and (3) – from the correlation plot of Soil 2 at 0–20 cm depicted in (A).

1198 G. Landrot et al. / Chemosphere 88 (2012) 1196–1201

2, according to EPA methods 1311 and 1312 (SI). The samples werethen centrifuged and analyzed by GF-AAS for total As concentra-tion in solution. Soil samples 1 and 2 were also sieved to 25 lmfor bulk EXAFS spectroscopy and micro-XAS analysis (details ofBulk Extended X-Ray Absorption Fine Structure Spectroscopy (EX-AFS) analyses and micro XAS analyses in SI).

3. Results and discussion

3.1. Soil composition

The soil samples taken at eight random locations in the parkcontained in average 33 mg Kg�1 Cr, 40 mg Kg�1 As, and 67 mgKg�1 Pb at a depth of 0–20 cm below the surface, and 56 mg Kg�1

Cr, 153 mg Kg�1 As, and 296 mg Kg�1 Pb at a depth of 20–40 cmbelow the surface (Table S1). According to a document releasedby the DNREC (link in SI) typical concentrations in Delaware soils

range from 5 to 30 mg Kg�1 Cr, 1 to 10 mg Kg�1 As, and 30 to100 mg Kg�1 Pb. Therefore, the average concentrations of thesethree elements measured in the Christiana Park soils were allabove the state average. Additional information on the soil compo-sition of the soil samples is provided in SI.

3.2. Elemental distribution

Arsenic desorption from Soil 1 and 2 was conducted based onthe EPA’s SPLP method 1311 and TCLP method 1312. Both des-orbed only a very small amount (less than 3%) of As from Soil 1and 2 at each depth (0–20 cm to 20–40 cm), which suggests thatAs is strongly sorbed to the soil.

The distribution of Cr, As, Mn, Fe, Ni, and Zn in Soil 1 and 2 atdepth 0–20 cm and 20–40 cm is depicted in Fig. 1. Micro-XRFimages show two distinct As and Cr environments; diffuse Cr andAs regions, as well as bright, highly concentrated ‘‘hot-spots’’

Fig. 3. (A) Normalized l-XANES at the Cr K-edge taken at several Cr hot spot in Soil 1 at 20–40 cm below the surface and Soil 2 at 0–20 cm and 20–40 cm below the surface;and (B) Normalized l-XANES at the As K-edge taken at several As hot spots in Soil 1 at 20–40 cm below the surface and Soil 2 at 0–20 cm and 20–40 cm below the surface.Gray stripe in A: highlight of the pre-white line region or the spectra; dashed line in A: position of the Cr(VI) pre-edge feature at 5993.5 eV; gray stripe in B: highlight of thepost-white line region of the spectra; dashed line: position of the As(V) white line standard at 11874 eV; and circle: highlight of a post-white line feature.

G. Landrot et al. / Chemosphere 88 (2012) 1196–1201 1199

displayed in red color in Fig. 1. Although the diffuse As in the soilseems to be associated with Fe, it does not seem to associate withthe other elements measured by l-XRF at As hot spots, except at afew hot spots (in some l-XRF maps of Soil 1 and 2, not shown inFig. 1), and for zinc, as indicated by As and Zn bright hot spots inthe maps of Soil 1 at 0–20 cm and Soil 2 at 20–40 cm (Fig. 1).The association between As and Zn is well known. For instance,the two elements can sorb to the same mineral phases, e.g., gibb-site or goethite (Gräfe et al., 2004), or As can sorb to a Zn-bearingmineral, like hemimorphite (Zn4Si2O7(OH)2�2H2O) (Mao et al.,2010). The distribution of Cr is correlated with the distribution ofFe in all l-XRF maps of Soil 1 and 2, for both diffuse Cr and Cr athot spots. Fig. 2A shows the Cr Ka intensities per pixel in eachl-XRF map of Soil 1 and 2 at 0–20 cm and 20–40 cm, over the FeKa intensities. A feature in the MicroAnalysis Toolkit (http://smak.-sams-xrays.com) for data analysis of microprobe data enables theselection of points in the correlation plots shown in Fig. 2A, toreconstruct a Cr l-XRF map with each selected set of points. Thiscommand thus enables one to locate the different Cr phases inthe l-XRF maps. Fig. 2A shows that two Cr phases, a diffuse Crphase ubiquitously present in the l-XRF maps, and a Cr phase lo-cated at Cr hot spots, that also contain Fe, are present in the l-XRFmaps used to make the correlation plots of Soil 1 at the two depthprofiles, and Soil 2 at 20–40 cm in Fig. 2A. The l-XRF map used tomake a correlation plot of Soil 2 at 0–20 cm in Fig. 2A featuresthree Cr phases; a diffuse Cr phase ubiquitously distributed inthe l-XRF map, and two phases that are located at Cr hot spots.

One is a Cr hot spot that is also a Fe hot spot (Fig. 2B), and the sec-ond phase is located at two Cr hot spots that contain lower concen-trations of Fe. Therefore, these results suggest that Cr in the soil ispresent in several forms, including those containing Cr and Fe.These results support findings reported in previous studies thataddressed the distribution of Cr in soils, which showed that Cr(III)can sorb to Fe oxides (Hopp et al., 2008; Vodyanitskii, 2009), orco-precipitate with iron (Charlet and Manceau, 1992; Hanselet al., 2003).

3.3. Cr and As oxidation states

The oxidation states of As and Cr were measured by l-XANES atseveral Cr and As hot spots found in the l-XRF maps of Soil 1 at 20–40 cm and Soil 2 at 0–20 cm and 20–40 cm below surface (Fig. 3).No l-XANES spectrum was collected in Soil 1 at 0–20 cm, since theconcentrations of As and Cr at this depth are less than 10 mg Kg�1

(Table S1 in SI). Since all l-XANES shown in Fig. 3A do not feature apre-edge feature at 5993.5 eV characteristic of Cr(VI) (Petersonet al., 1997), Cr(III) is present at each hot spot analyzed. The shapeof the l-XANES spectra can give information about the nature ofthe Cr(III) phase. Several l-XANES spectra shown in Fig. 3A, for in-stance Spot 2 in Soil 2 (0–20 cm), feature a sharp white line, and ahump at half the height of the white line, at 6002 eV (gray stripe inFig. 3A). These features are characteristic of a chromite (FeCr2O4)XANES standard spectrum (Peterson et al., 1997; Werner et al.,2007). Since chromite is a Cr mineral, which can be naturally found

1200 G. Landrot et al. / Chemosphere 88 (2012) 1196–1201

in Delaware, Pennsylvania, and Maryland (Pearre and Heyl, 1960),the sources of Cr(III) at these hot spots may not be anthropogenic.However, some l-XANES spectra in Fig. 3A look different than thespectrum of chromite, i.e. the l-XANES spectra of spot 1 and 2 inSoil 2 at 20–40 cm and spot 3 in Soil 1 at 20–40 cm, and featurea hump at about a fifth the height of the white line. Additionally,the l-XANES spectrum of spot 4 in Soil 1 at 20–40 cm does not fea-ture a sharp white line like in the other spectra, and looks similarto the XANES spectrum of Cr-bearing goethite (Frommer et al.,2009). Therefore, our results suggest that different mineral phasescontaining Cr(III) are present in the soil, and some of them mayalso contain Fe.

The position of the inflection point of the first derivative of eachAs l-XANES spectra shown in Fig. 3B is located at about 11874 eV,which means that As(V) is present at each hot spot analyzed. Theregion located at the high-energy side of the white line (gray areain Fig. 3B) in the l-XANES spectra taken at the As K-edge can giveinsight into the mineral phase in which As is present (Cancès et al.,2008). If this region exhibits significant features, As is included inthe structure of crystalline phases. Conversely, if the region doesnot possess any feature, As is associated with less-ordered phases.The latter case corresponds to what we observe in our data sincemost l-XANES spectra shown in Fig. 3B do not exhibit any featureat the high-energy side of the white line (region in the gray area inFig. 3B). However, since the l-XANES spectrum of spot 1 in Soil 2 at0–20 cm features a hump at about 11885 eV (encircled in Fig. 3B),As may be included in crystalline mineral phases at this location.Therefore, our results suggest that, similar to Cr, As is present inthe soil in various chemical environments, but unlike Cr, As existsin the oxidized form (i.e., arsenate).

3.4. Molecular environment of As

The concentrations of Cr in all soil samples collected in Christi-ana Park were not high enough for bulk EXAFS spectroscopy atbeamline 11-2, which requires a few hundred of mg Kg�1 of Cr,since the highest concentration found in our samples was115 mg Kg�1. Additionally, the concentrations of As in Soil 1 at0–20 cm and Soil 2 at 20–40 cm (14 and 21 mg Kg�1, respectively)were not high enough for bulk EXAFS measurements. Therefore,bulk EXAFS spectroscopy analyses were conducted only on Soil 1at 20–40 cm below the surface and Soil 2 at 0–20 cm that had aconcentration of 985 and 50 mg Kg�1 As, respectively. The energypositions of the XANES first derivative’s inflection points of thetwo soil samples and the As(V) standard were similar to each other(Fig. S6), meaning that arsenate is the main form of As in the twosoil samples. To identify the As standards needed for linear combi-nation fitting of the two soil sample EXAFS spectra, Principal Com-ponent Analysis (PCA) and target transform analysis could not beapplied, since a minimum of three components (i.e. number ofsamples) is needed for PCA analysis. Therefore, the choice of stan-dards used for linear combination fitting was based on a knowl-edge of Christiana Park’s history, observations made from l-XRFmaps, and the fact that these standards must contain As(V), dueto the results from bulk and l-XANES spectroscopy. The selectionof standards used for linear combination fitting and resultsare shown in Fig. S5A. The EXAFS spectra of mansfieldite(AlAsO4�2H2O) and As/Zn-goethite were the best set of standardsto do least squares fitting of the EXAFS spectrum of the two soilsamples. The mansfieldite standard contributed to a major partof the fits (at least 65%). The presence of this mineral, which israrely found in nature (Allen et al., 1948), has never been reportedin the soil of Delaware. Therefore, the mineral present in ChristianaPark could be a phase that contains both As and Al, whose EXAFSspectrum is similar to the one of mansfieldite. A former studyshowed that a three dimensional alumino-arsenate precipitate,

whose EXAFS spectrum is also similar to the one of mansfieldite,can form when As(V) is sorbed to c-Al2O3 for a long period of time(a few months) (Arai and Sparks, 1992). Therefore, the precipitatereported in this former study could be present in our two soil sam-ples, since c-Al2O3 is ubiquitously found in the environment. Thecontribution of the As–Zn goethite in the linear combination fitsis consistent with our observations made from the l-XRF mapsand the fluorescence emission analyses at As hot spots, whichshowed that As, Zn, and Fe can be present at the same location.

Chromium is mainly associated with Fe in Soil 1 and 2, sincethese two elements were systematically distributed together inthe regions where the concentrations of Cr and Fe were diffuse,and at hot spots where these two elements were present at highconcentrations. The l-XRD and fluorescence emission analysesperformed at hot spots (Figs. S2, S3, and S4A) suggest that Crmay be naturally part of Fe-bearing minerals, or sorbed to Fe oxi-des in Soil 1 and 2, which is in agreement with several previousstudies that showed that Fe oxides in soils are strong Cr scaveng-ers. Therefore, if chromate, the most toxic form of Cr, is presentin the soil, it is likely to undergo reduction, since Fe oxides can re-duce Cr(VI) to Cr(III). Additionally, macroscopic observations fromSoil 1 and 2 suggest that Cr(III), which is less toxic and mobile thanCr(VI), is not likely to be oxidized to Cr(VI) by Mn oxides, since theassociation between Mn and Cr was only observed in the l-XRFmaps at regions where the concentrations of these two elementswere diffuse. Therefore, if tannery waste materials containingCr(VI) were introduced in Christiana Park during the last century,a major part of chromate present in those wastes has been proba-bly reduced to Cr(III) by Fe oxides over the years, and perhaps alsoby other Cr(VI) reducers, including organic matter. This hypothesisis supported by the l-XANES spectra taken at several Cr hot spotsin Soil 1 and 2, showing that only Cr(III) is present in these loca-tions. Therefore, the apparent absence of Cr(VI) in the soil, the lackof high concentrations of Cr measured at different locations inChristiana Park, the distribution of Cr in the soil, and its associationwith other elements that can potentially affect its chemical proper-ties, suggest that the presence of Cr in the recreation park does notpose a potential threat to the public.

However, high concentrations of As were measured in the park.These concentrations exceeded by far the background level of As inDelaware soils (i.e. 11 mg Kg�1). Results obtained from synchro-tron-based spectroscopy suggest that As is sorbed to Al oxides, oroccluded in Al–As precipitate phases similar to mansfieldite. Thelatter hypothesis is supported by the results from EPA’s TCLP andSPLP tests, in which little As was desorbed from Soil 1 and 2. Addi-tionally, the l-XRD and fluorescence emission analyses conductedat hot spots suggest that As might be also sorbed to Fe oxides to aminor extent. Therefore, our results are in good agreement withprevious studies that showed that Al and Fe phases control the dis-tribution of As in contaminated soils. The environmental factorsaffecting the stability of these phases need to be determined, tobetter assess the potential risk to humans of As present at highconcentrations in the soil. Even if the results from the TCLP andSPLP tests indicate that As is immobilized in the soil, As couldpotentially desorb from Fe(III) oxides if these phases undergoreduction in anoxic conditions (La Force et al., 2000; Morin andCalas, 2006; Wang et al. 2010). This scenario might occur, forexample, if the nearby Christiana River flooded the park anddepleted O2 in the soil. Therefore, although remedial action inChristiana Park was finally completed in 2008, which consistedof installation of a one-foot clean protective barrier on top of ageo-textile demarcation fabric across the entire site, the presenceof As in the soil still poses an environmental threat. If the Fe(III)phases (e.g. goethite) containing As became unstable, this metal(loid) could diffuse to other areas near the park, or leach to thegroundwater. Additionally, the speciation and mobility of Pb,

G. Landrot et al. / Chemosphere 88 (2012) 1196–1201 1201

which is present in the park at high concentrations, should be alsoassessed.

Acknowledgements

Portions of this work were performed at Beamline X27A,National Synchrotron Light Source (NSLS), Brookhaven NationalLaboratory (BNL). Beamline X27A is supported by the US Depart-ment of Energy (DOE) – Geosciences (DE-FG02-92ER14244 toThe University of Chicago – CARS). Use of the NSLS was supportedby DOE, Office of Science, Office of Basic Energy Sciences, underContract No. DE-AC02-98CH10886. This research was also carriedout at the Stanford Synchrotron Radiation Lightsource, a Director-ate of SLAC National Accelerator Laboratory and an Office ofScience User Facility operated for the US Department of Energy Of-fice of Science by Stanford University. The SSRL Structural Molecu-lar Biology Program is supported by the DOE Office of Biologicaland Environmental Research, and by the National Institutes ofHealth, National Center for Research Resources, Biomedical Tech-nology Program (P41RR001209). The authors would like to thankBrian McCandless (Institute of Energy Conversion, U. of Delaware)for assistance in XRD data collection; Tiffany Thomas, JenniferSeiter, Matthew Siebecker, and Gerald Hendricks for assistance insample collection and geochemical analyses.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.chemosphere.2012.03.069.

References

Allen, V.T., Fahey, J.J., Axelrod, J.M., 1948. Mansfieldite, a new arsenate, thealuminum analogue of scorodite, and the mansfieldite scorodite series. Am.Mineral. 33, 122–132.

Arai, Y., Sparks, D.L., 2002. Residence time effects on arsenate surface speciation atthe aluminum oxide–water interface. Soil Sci. 167, 303–314.

Arcon, I., Van Elteren, J.T., Glass, H.J., Kodre, A., Slejkovec, Z., 2005. EXAFS and XANESstudy of arsenic in contaminated soil. X-ray Spectrom. 34, 435–438.

Beaulieu, B.T., Savage, K.S., 2005. Arsenate adsorption structures on aluminum oxideand phyllosilicate mineral surfaces in smelter-impacted soils. Environ. Sci.Technol. 39, 3571–3579.

Bhattacharya, P., Mukherjee, A.B., Jacks, G., Nordqvist, S., 2002. Metal contaminationat a wood preservation site: characterisation and experimental studies onremediation. Sci. Total Environ. 290, 165–180.

Cancès, B., Juillot, F., Morin, G., Laperche, V., Polya, D., Vaughan, D.J., Hazemann, J.-L.,Proux, O., Brown Jr., G.E., Calas, G., 2008. Changes in arsenic speciation througha contaminated soil profile: a XAS based study. Sci. Total Environ. 397, 178–189.

Charlet, L., Manceau, A., 1992. X-ray adsorption spectroscopy study of the sorptionof Cr(III) at the oxide–water interface. II. Adsorption, coprecipitation, andsurface precipitation on hydrous ferric interface. J. Colloid Interface Sci. 148,443–458.

Fendorf, S., Guangchao, L., 1996. Kinetics of chromate reduction by ferrous iron.Environ. Sci. Technol. 30, 1614–1617.

Frommer, J., Nachtegaal, M., Czekaj, I., Weng, T.-C., Kretzschmar, R., 2009. X-rayabsorption and emission spectroscopy of Cr(III) (Hydr) oxides: analysis of the K-pre-edge region. J. Phys. Chem. A. 113, 12171–12178.

Fendorf, S.E., Zasoski, R.J., 1992. Chromium(III) oxidation by d-MnO2. 1.Characterization. Environ. Sci. Technol. 26, 79–85.

Gao, X., Schulze, D.G., 2010. Chemical and mineralogical characterization of arsenic,lead, chromium, and cadmium in a metal-contaminated histosol. Geoderma156, 278–286.

Gräfe, M., Nachtegaal, M., Sparks, D.L., 2004. Formation of metal-arsenateprecipitates at the goethite–water interface. Environ. Sci. Technol. 38, 6561–6570.

Hansel, C.M., Wielinga, B.W., Fendorf, S.E., 2003. Stuctural and compositionalevolution of Cr/Fe solids after indirect chromate reduction by dissimilatoryiron-reducing bacteria. Geochim. Cosmochim. Acta 67 (3), 401–412.

Hopp, L., Nico, P.S., Marcus, M.A., Peiffer, S., 2008. Arsenic and chromiumpartitioning in a podzolic soil contaminated by chromated copper arsenate.Environ. Sci. Technol. 42, 6481–6486.

La Force, M.J., Hansel, C.M., Fendorf, S., 2000. Arsenic speciation, seasonaltransformations, and co-distribution with iron in a mine waste-influencedpalustrine emergent wetland. Environ. Sci. Technol. 34, 3937–3943.

Lund, U., Fobian, A., 1991. Pollution of two soils by arsenic, chromium and copper,Denmark. Geoderma 49, 83–103.

Mao, M., Lin, J., Pan, Y., 2010. Hemimorphite as a natural sink for arsenic in zincdeposits and related mine tailings: evidence from single-crystal EPRspectroscopy and hydrothermal synthesis. Geochim. Cosmochim. Acta 74,2943–2956.

Morin, G., Calas, G., 2006. Arsenic in soils, mine tailings, and former industrial sites.Elements 2, 97–101.

Ostergren, J., Brown Jr., G.E., Parks, G., Tingle, T., 1999. Quantitative speciation oflead in selected mine tailings from Leadville. CO. Environ. Sci. Technol. 33,1627–1636.

Parikh, S.J., Lafferty, B.J., Sparks, D.L., 2008. An ATR-FTIR spectroscopic approach formeasuring rapid kinetics at the mineral/water interface. J. Colloid Interface Sci.320, 177–185.

Pearre, N.C., Heyl Jr., A.V., 1960. Chromite and other mineral deposits in serpentinerocks of the Piedmont Upland, Maryland, Pennsylvania and Delaware. US Geol.Survey Bull. 1082-K, 707–833.

Peterson, M.L., Brown, G.E., Parks, G.A., 1996. Direct evidence for heterogeneousredox reaction at the aqueous chromium/magnetite interface colloid. SurfacesA. 107, 77–88.

Peterson, M.L., Brown, G.E., Parks, G.A., Stein, C.L., 1997. Differential redox andsorption of Cr(III/VI) on natural silicate and oxide minerals: EXAFS and XANESresults. Geochim. Cosmochim. Acta 61, 3399–3412.

Post, J.E., 1999. Manganese oxide minerals: crystal stuctures and economic andenvironmental significance. Proc. Natl. Acad. Sci. 96 (7), 3447–3454.

Sparks, D.L., 2003. Environmental Soil Chemistry. Academic Press, San Diego, CA.Sreeram, K.J., Ramasami, T., 2003. Sustaining tanning process through conservation,

recovery, and better utilization of chromium. Resour. Conserv. Recy. 38 (3),185–212.

Van Herreweghe, S., Swennen, R., Vandecasteele, C., Cappuyns, V., 2003. Solid phasespeciation of arsenic by sequential extraction in standard reference materialsand industrially contaminated soil samples. Environ. Pollut. 122, 323–342.

Vodyanitskii, Y.N., 2009. Chromium and arsenic in contaminated soils (review ofpublications). Eurasian Soil Sci. 42, 507–515.

Wang, Y., Morin, G., Ona-Nguema, G., Juillot, F., Guyot, F., Calas, G., Brown, G.E.,2010. Evidence for different surface speciation of arsenite and arsenate on greenrust: an EXAFS and XANES study. Environ. Sci. Technol. 44, 109–115.

Weaver, R.M., Hochella, M.F., 2003. The reactivity of seven Mn-oxides with Cr3þaq : a

comparative analysis of a complex, environmentally important redox reaction.Am. Mineral. 88, 2016–2027.

Werner, M.L., Nico, P.S., Marcus, M.A., Anastasio, C., 2007. Use of micro-XANES tospeciate chromium in airborne fine particles in the sacramento valley. Environ.Sci. Technol. 41, 4919–4924.

Yang, L., Donahoe, R., 2007. The form, distribution and mobility of arsenic in soilscontaminated by arsenic trioxide, at sites in southeast USA. Appl. Geochem. 22,320–341.