Geochemistry of an acidic chromium sulfate plume

Geochemistry of an acidic chromium sulfate plume

Applied Geochemistry Applied Geochemistry 22 (2007) 357–369 www.elsevier.com/locate/apgeochem Geochemistry of an acidic chromium sulfate plume L. Edm...

1MB Sizes 6 Downloads 34 Views

Applied Geochemistry Applied Geochemistry 22 (2007) 357–369 www.elsevier.com/locate/apgeochem

Geochemistry of an acidic chromium sulfate plume L. Edmond Eary 1, Andy Davis

*

Geomega, 2995 Baseline Road, Suite 202, Boulder, CO 80303, United States Received 26 September 2005; accepted 30 October 2006 Editorial handling by R. Wanty Available online 23 December 2006

Abstract The historical disposal of acidic chromium sulfate solutions into unlined lagoons between 1953 and 1970 at an industrial site resulted in formation of a dense aqueous phase liquid (DAPL) plume [specific gravity 1.11 g/cm3, pH 3, up to 4700 mg/L Cr(III), and up to 90,000 mg/L SO4]. The DAPL sank through the shallow glacial till aquifer to an underlying impermeable gneissic bedrock from where it migrated downgradient along buried channels incised in the bedrock. Because of its high density, the plume chemistry is sharply stratified vertically. Chromium(III) predominates in the DAPL because excess Cr(VI) not reduced in the original process has been reduced by Fe(II) derived from silicates, while Cr(OH)3(am) occurs as surface coatings on silicate minerals and as discrete particles mixed with Fe(OH)3(am) and Al(OH)3(am). The solubility of Cr(OH)3(am) accurately describes Cr(III) concentrations in the plume and nearby groundwater, while Al and Fe in solution are also consistent with solubility-controlling oxyhydroxides. Because of these solubility controls, metal cations are attenuated relative to more mobile Cl and SO4, resulting in a chromatographic separation of solutes downgradient from the plume origin. The good agreement between predicted and observed solution concentrations illustrates the utility of equilibrium modeling when interpreting metal transport characteristics and in determining the efficacy of natural attenuation in subsurface systems.  2007 Elsevier Ltd. All rights reserved.

1. Introduction Experimental studies have established that the aqueous mobility of Cr in the environment depends primarily on its oxidation state and solution pH (Fendorf et al., 2000; Palmer and Wittbrodt, 1991; Rai et al., 1989). There are two stable oxidation states, Cr(VI) and Cr(III), under surface conditions. Chromium(VI) forms anionic species (i.e., HCrO 4 *

Corresponding author. Fax.: +1 303 938 8123. E-mail address: [email protected] (A. Davis). 1 Present address: MFG, Inc., 3801 Automation Way, Suite 100, Fort Collins, CO 80525, United States.

and CrO2 4 ) that are mobile in most oxidizing aqueous environments because significant solubility controlling phases are lacking, although adsorption may limit transport under acidic conditions (Davis and Olsen, 1995; Stollenwerk and Grove, 1985; Zachara et al., 1987, 1989). Exceptions to this generalization include the formation of Ba(S,Cr)O4, that may limit Cr(VI) concentrations in the presence of Ba2+ (Rai et al., 1989), chromate jarosites (KFe3(CrO4)2(OH)6) that may limit Cr(VI) concentrations under moderately acidic conditions (Baron et al., 1996a,b; Baron and Palmer, 1998), and Cr(VI)– Fe(OH)3 mixed precipitates that can occur in acidic solutions (Olazabal et al., 1997). Chromium(VI)

0883-2927/$ - see front matter  2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.apgeochem.2006.10.002

358

L.E. Eary, A. Davis / Applied Geochemistry 22 (2007) 357–369

species are strong oxidants that are rapidly reduced by Fe(II) aqueous species and solids (Buerge and Hug, 1997; Eary and Rai, 1988, 1989, 1991; Fendorf and Li, 1996; Fendorf et al., 2000; Loyaux-Lawniczak et al., 1999; Peterson et al., 1997; Saleh et al., 1989) and organic compounds (Deng and Stone, 1996; Hug et al., 1997; Tseng and Bielefeldt, 2002; Wielenga et al., 2001; Wittbrodt and Palmer, 1996) effectively limiting their longevity in most subsurface environments where these reductants are commonly present. In contrast to Cr(VI), Cr(III) forms hydrolyzed cationic species that rapidly precipitate as hydroxide, such as Cr(OH)3(am) and (Fe,Cr)(OH)3(am) (Rai et al., 1987; Sass and Rai, 1987). Similar to Fe and Al hydroxides, Cr(III) hydroxides have amphoteric solubilities that probably limit Cr(III) mobility in aqueous systems except under strongly acidic or strongly alkaline conditions (Rai et al., 1989). Chromium(III) may be oxidized to Cr(VI) by Mn(IV) oxides but this process may only be important in oxidized systems that contain high amounts of Mn(IV) oxides (Banerjee and Nesbitt, 1999; Eary and Rai, 1987; Pettine et al., 1991; Bartlett and James, 1979). Indeed, the characteristic rapid reduction kinetics for Cr(VI) and low solubility of Cr(III) hydroxide have led to the development of methodologies for in situ immobilization of Cr in aqueous systems by the introduction of reductants such as metallic Fe, aqueous Fe(II), Fe sulfides, and organic chemicals into aquifers (Blowes et al., 1997; James, 1996; Makos and Hrncir, 1995; Pratt et al., 1997; Puls et al., 1999). The purpose of this investigation was to understand the geochemical processes controlling the mobility of Cr in an acidic plume located downgradient from a former industrial site. Information from the various studies cited above is used, in combination with field data, laboratory data, and geochemical modeling to relate theory to empirical site conditions. 2. Site description At the industrial site examined in this study (Fig. 1), H2SO4, HCl, NaCl, NaClO4, Na2SO4, NH4Cl and Na2Cr2O7 were used in the manufacture of chemical blowing agents, stabilizers, antioxidants, and other specialty chemicals for the rubber and plastics industry, while Cr2(SO4)3 was a waste product. Between 1953 and 1970, acidic waste solutions with high concentrations of Na, Ca, Cr(III),

Cr(VI), NH4, Cl and SO4 were generated during chemical processing and disposed untreated into nearby unlined pits. The disposal pits were excavated in unconsolidated outwash sands and gravels typical of glacial till deposits. These surficial deposits characteristically have high hydraulic conductivities (K = up to 1 · 103 cm s1) in the shallow aquifer surrounding the industrial site. The glacial deposits unconformably overlie a comparatively impermeable gneissic bedrock. The concentrated waste solutions rapidly sank through less dense dilute water in the shallow aquifer and then migrated downgradient along the bedrock surface as a coherent dense aqueous phase liquid (DAPL) plume. The DAPL has a specific gravity of 1.11 g/cm3 and a chemical composition distinctly different from the overlying groundwater through which the plume sank. Because of its high density, remnant DAPL is present in bedrock depressions and is an ongoing source of solutes to the overlying groundwater through diffusion and mixing. 3. Methods 3.1. DAPL characterization Three multilevel piezometers (MP-1, MP-2 and MP-3) were installed in the late 1990s to determine the chemical stratigraphy of the DAPL (Fig. 1). The boreholes for MP-1 and MP-2 were advanced to bedrock, while MP-3 was terminated at 20.4 m (3 m into the DAPL). The piezometers were constructed from 5-cm diameter Schedule 80 PVC pipe and contained from 13 to 16 sampling ports at various depths in the boreholes. The sampling ports were connected to the surface with 0.6 cm diameter tubing, the space around the piezometer backfilled with silica sand, and individual sampling ports packed off with bentonite to preclude mixing between ports. After MP installation, each sampling port was developed by pumping with a peristaltic pump until the water was free of turbidity. Initial clogging of some ports was remedied by backflushing. When free flow was achieved, each port was pumped for 10–20 min to remove injected water. The ports were then allowed to equilibrate with the aquifer formation for three weeks. 3.2. Water sampling and analytical methods An array of 118 monitoring wells exists at the site in addition to the multilevel piezometers (Fig. 1).

L.E. Eary, A. Davis / Applied Geochemistry 22 (2007) 357–369

359

Fig. 1. Site map showing monitoring wells and areal spread of plume.

Subsets of these wells were sampled on a quarterly basis and samples field-filtered through 0.45 lm nitrocellulose filters. Samples for cation analyses were acidified to pH <2 with HNO3 and for anions preserved on ice at 4 C. Cations were analyzed by inductively-coupled plasma atomic emission spectroscopy (ICP/AES) and anions by ion chromatography. Specific conductance, pH, pE, temperature, and dissolved O2 were determined in the field with appropriate sensors inserted into a flow-through cell attached to each sampling port. Some water samples were also analyzed in the field by colorimetry for Fe(II) and total Fe (Fe(III) by difference) with Hach AccuVac ampules and field spectrophotometer (Hach, 1994). Chromium(VI) was determined colorimetrically within 24 h of sample collection using diphenylcarbazide (Skougstad et al., 1979). Total dissolved Cr was determined by ICP/AES [Cr(III) by difference]. The total acidity of a DAPL sample from monitoring well GW-42D, located adjacent to multilevel piezometer MP-2 (depth 10–12 m), was determined

by titration of a 1.5 L sample with 50% NaOH while continuously stirring and sparging with N2(g) to exclude air (Table 1). The titration was conducted slowly over the course of 8 h to allow equilibration with the metal hydroxide precipitates that formed as the solution was neutralized. 3.3. Laboratory determinations of DAPL precipitate solubility Samples of the precipitate slurry formed by the titration of the DAPL sample to determine total acidity were also used for solubility experiments. Splits of 50 mL of the slurry were placed in 50-mL polycarbonate centrifuge tubes containing 0.1 M NaCl previously purged of air with N2(g). Aliquots of 1 M HCl or 1 M NaOH were added to adjust the initial pH, after which the tubes were completely filled to minimize contact with air, capped, and placed in a constant temperature bath at 25 ± 2 C. Eleven tubes with initial pH values from 3.7 to 8.2 were prepared.

360

Table 1 Water compositions in monitoring well 42D and multilevel piezometer MP-2 at different depths Analyte

Well

Concentrations (mg/L)

GW-42D

1.68 m

3.20 m

4.12 m

4.57 m

5.34 m

6.10 m

6.86 m

7.32 m

7.77 m

8.23 m

9.15 m

9.91 m

10.67 m

11.43 m

2200 0.069 0.087 635 4050 1590 600 1300 640 82 175 23,500 23,500 18 9.1 78,000 3.36 <0.5 NA

<0.1 <0.008 0.023 11 0.028 36 <0.01 7 2.5 3.3 NA 250 190 0.058 23 390 5.93 <0.5 1719

0.15 <0.008 0.037 13 0.15 6.9 3.7 5.5 0.80 2.4 NA 270 140 0.2 20 570 5.43 <0.5 1730

0.21 <0.008 0.051 11 0.21 3.7 0.5 2.9 0.76 1.1 NA 89 75 0.19 9.7 160 5.11 <0.5 699

0.15 <0.008 0.077 10 0.12 2.9 0.5 2.2 1.1 2.4 NA 36 81 0.11 7.9 33 5.18 NA 330

0.28 <0.008 0.046 12 0.24 3.8 0.3 3 1.2 <1.0 NA 37 86 0.10 6.8 26 4.88 <0.5 387

0.29 <0.008 0.017 12 0.17 1.8 0.2 3.3 0.72 <1.0 NA 39 79 0.21 7.2 44 4.91 <0.5 402

0.26 <0.008 0.018 8.6 0.23 1.9 0.1 1.9 0.97 <1.0 NA 45 49 NA 12 100 4.90 <0.5 423

1.4 <0.008 0.012 8.6 1.5 1.4 0.2 2.5 0.81 1.0 NA 53 50 0.17 13 100 4.87 <0.5 443

0.76 0.013 0.022 7.1 1.2 1.6 0.05 2.4 14 8.9 NA 74 55 0.13 28 190 5.02 <0.5 693

39 NA NA 120 18 36 24 48 NA NA NA 700 500 0.066 320 2500 4.30 <0.5 6040

1200 0.56 <0.10 390 1600 2930 270 770 290 100 NA 15,000 11,000 1.8 160 52,000 3.69 <0.5 74,000

1100 0.60 <0.10 340 2400 1000 410 850 460 98 NA 18,000 14,000 2.7 210 62,000 3.50 <0.5 87,500

1500 0.77 <0.10 410 3300 1190 70 1100 620 120 NA 22,000 16,000 3.3 390 71,000 3.46 <0.5 96,300

1600 1.2 <0.10 440 4700 900 1200 940 700 45 NA 24,000 17,000 2.1 140 77,000 3.39 <0.5 102,600

NA 17.1

13 11.9

52 11.2

92 11

77 10.7

116 11.7

177 11.8

194 13.1

214 12.2

199 12.4

108 12.6

209 13.6

262 12.6

263 12.3

248 12.5

Multilevel piezometer MP-2 depths (m)

NA – not analyzed. a Dissolved O2.

L.E. Eary, A. Davis / Applied Geochemistry 22 (2007) 357–369

Al As Ba Ca Cr(III) Fe(II) Fe(III) Mg Mn K Si Na Cl NO3 NH4 SO4 pH D.O.a SC (lS/ cm) Eh (mV) T (C)

L.E. Eary, A. Davis / Applied Geochemistry 22 (2007) 357–369

The tubes were each sampled after 2 and 16 d by opening them, inserting a N2(g) bubbler to minimize the incursion of air, and removing a 10-mL aliquot which was filtered (0.45 lm) and acidified to pH <2 with 0.1 M HCl. The pH of the remnant solution in the tube was measured at the time of sampling. The extracted solution volume was replaced with de-aerated 0.1 M NaCl and the tubes recapped and placed in the constant temperature bath until the next sampling period. Total Cr concentration in each sample was determined by ICP/AES. 3.4. Aquifer material characterization Ten samples of aquifer material were obtained from a borehole located about 3 m from multilevel piezometer MP-2 at depths between 2.6 m and 10 m below ground surface. This depth profile extends from clean, sandy aquifer material (i.e., unexposed to DAPL) into the DAPL zone. Samples were analyzed for total Cr and Fe by X-ray fluorescence (XRF). Sample splits were also prepared for electron microprobe analysis (EMPA) by embedding grains in epoxy, polishing the surface under kerosene, and C coating (Davis et al., 1993). 4. Results and discussion 4.1. DAPL chemical stratigraphy The DAPL chemical stratigraphy is well illustrated by concentration profiles obtained from the multilevel piezometers (Fig. 2). The specific conduc-

361

tance ranges from 400–1000 lS/cm to 7.8 m below ground surface (bgs) and then increases sharply to 74,000–102,000 lS/cm (characteristic of DAPL) below 9 m bgs. Over the same depth interval, the pH decreases sharply from 5.2 to 3.4, while Cr(III) increases from 0.1 to 4700 mg/L. Other metal (Al, Fe and Mn) profiles (Table 1) are similar to Cr(III). The increase in specific conductance at the DAPL boundary reflects increased solute concentrations, including SO4 from 100 to 77,000 mg/L, Cl from 50 to 17,000 mg/L, and Na from 50 to 24,000 mg/L (Table 1). These solutes were the major components in the acidic waste solutions historically disposed in the pits at the site. The large increases in solute concentrations occur over a small depth interval, the diffuse layer (Fig. 2), indicating that the DAPL has retained a distinct boundary over time. 4.2. Redox conditions Redox conditions in the DAPL are moderately reducing with predominant Fe(II) and S present as SO4 (Table 1) with no detectable H2S(g) odor. Nitrogen occurs primarily as NH4 in the DAPL zone and as NO3 in the ambient groundwater above the DAPL. Dissolved O2 was <0.5 mg/L (Table 1), the reliable detection level for the field meters. The Eh was variable but consistent with the presence of dissolved Fe(II). The original waste solutions contained Cr2(SO4)3 and Na2Cr207, hence it was initially expected that both Cr(III) and Cr(VI) would be found in the

Fig. 2. Vertical profiles of water quality indicators from multilevel piezometer MP-2.

362

L.E. Eary, A. Davis / Applied Geochemistry 22 (2007) 357–369

groundwater, however only Cr(III) was detected. EMPA of aquifer samples identified ubiquitous layer silicates, mostly hydrobiotite, muscovite, illite and chlorite. The octahedral sites of these minerals typically contain Fe(II), which facilitates rapid reduction of Cr(VI) under acidic conditions (Brigatti et al., 2000; Eary and Rai, 1989). The DAPL also contains high Fe(II), presumably from the dissolution of these layer silicates present in the aquifer under the acidic conditions of the DAPL. Ferrous iron reduces Cr(VI) to Cr(III) quickly and completely by the generalized reaction: 3FeðIIÞ þ CrðVIÞ ! 3FeðIIIÞ þ CrðIIIÞ;

ð1Þ

hence, Cr(VI) and Fe(II) do not co-exist in solution (Buerge and Hug, 1997; Eary and Rai, 1988; Fendorf and Li, 1996; Saleh et al., 1989). 4.3. DAPL acidity The DAPL sample used to determine total acidity was initially dark green due to 4050 mg/L of Cr(III). The addition of NaOH immediately resulted in the formation of a light green precipitate, characteristic of Cr(OH)3(am), although it is also possible that some of the green color was from Fe(OH)2 precipitation with SO4 to form metastable Type II green rust (O’Loughlin and Burris, 2004). As more NaOH was added, a brown precipitate formed, probably Fe(OH)3(am), producing an olive-green slurry at the end of the titration. The total acidity of the DAPL to pH 7.0 was 0.43 M. In addition to strong acid anions (e.g., Cl and SO4), the DAPL contained mineral acidity in the form of dissolved metals, i.e., Fe, Cr, Al and Mn (Table 1), precipitation of which as hydroxides strongly buffers the solution pH, according to the reaction:

KMgFe2 AlSi3 O10 ðOHÞ2 þ 3H2 O þ 7Hþ ! Kþ þ Mg2þ þ 2Fe2þ þ AlðOHÞ3 ðamÞ þ H4 SiO04 ð3Þ Evidence for such reactions occurring in the aquifer is indicated by the predominance of Fe2+ over Fe3+ in the DAPL and equilibria with Al(OH)3(am). Additionally, silicic acid substantially increases at low pH, exceeding the solubility of SiO2(am) at pH <4, whereas in the more neutral pH portions of the aquifer, conditions are near equilibrium with quartz or chalcedony (Fig. 3). These results indicate that increased silicate mineral dissolution is occurring in the acidified parts of the aquifer, which may partially neutralize DAPL acidity. However, this role is probably limited in that silicate dissolution reactions (e.g., Eq. (3)) are slow, hence the DAPL will remain acidic for an extended period of time. 4.4. Solubility relationships To depict DAPL chemistry relative to equilibrium conditions, activities of individual aqueous þ species (e.g., Cr3+, CrOH2+, CrðOHÞþ 2 ; CrSO4 ) in groundwater samples were calculated using the MINTEQ4 geochemical code (Eary and Jenne, 1992). The primary source of thermodynamic data used with MINTEQ4 was from Krupka et al. (1988), Rai et al. (1987) and Sass and Rai (1987) for the Cr aqueous species and solids (Table 2). These activities were then summed to give a total

Menþ þ nH2 O ! MeðOHÞn þ nHþ ðn ¼ 2 or 3Þ ð2Þ The high mineral acidity indicates that DAPL neutralization would require reaction with strong bases or alternately, long periods of reaction with weak bases. The aquifer system at the site contains some weak bases in the form of common layer silicates e.g., biotite, muscovite, and chlorite, that may consume a small fraction of the DAPL acidity through hydrolysis reactions. Dissolution of layer silicates, e.g., biotite, under the moderately reducing, acidic conditions of the DAPL can occur by:

Fig. 3. Comparison of dissolved silica concentrations to the solubilities of different forms of silica (SiO2) as calculated with MINTEQ4 for site groundwaters.

L.E. Eary, A. Davis / Applied Geochemistry 22 (2007) 357–369

363

Table 2 Thermodynamic data for aqueous Cr species and mineral forms used for geochemical modeling Reaction

log K25C

Aqueous Species CrOH2+ + H+ = Cr3+ + H2O þ CrOH2þ þ H2 O ¼ CrðOHÞþ 2 þH 0 2þ CrOH þ 2H2 O ¼ CrðOHÞ3 þ 2Hþ þ CrOH2þ þ 3H2 O ¼ CrðOHÞ 4 þ 3H 2þ 2 þ CrOH þ SO4 ¼ CrSO4 þ H2 O 0 CrOH2þ þ SO2 4 ¼ CrOHSO4 2þ 2 2CrOH þ SO4 ¼ ðCrOHÞ2 SO2þ 4 0 2CrOH2þ þ 2SO2 4 ¼ ðCrOHÞ2 ðSO4 Þ2 2  þ CrO4 þ H ¼ HCrO4 0 þ CrO2 4 þ 2H ¼ H2 CrO4

3.57 6.21 12.57 24.06 6.18 3.50 6.58 4.97 6.49 5.18

Solids Cr(OH)3(am): CrOH2+ + 2H2O = Cr(OH)3(am) + 2H+ Gibbsite (lc): Al3+ + 3 H2O = Al(OH)3 (lc) + 3H+ Al(OH)3(am): Al3+ + 3H2O = Al(OH)3(am) + 3H+ þ Alunite: Kþ þ 3Al3þ þ 2SO2 4 þ 6H2 O ¼ KAl3 ðSO4 Þ2 ðOHÞ6 þ 6H 3þ 2 þ Jurbanite: Al þ SO4 þ 6H2 O ¼ AlSO4 OH  5H2 O þ H Basaluminite: 4Al3þ þ SO2 4 þ 10H2 O ¼ Al4 ðSO4 ÞðOHÞ10 Fe(OH)3(am): Fe3+ + 3H2O = Fe(OH)3(am) + 3H+ Manganite: Mn2+ + 2H2O = MnOOH + 3H+ + e þ Schwertmannitea : 8Fe3þ þ 1:6SO2 4 þ 12:8H2 O ¼ Fe8 O8 ðOHÞ4:8 ðSO4 Þ1:6 þ 20:8H Birnessite: Mn2+ + 2H2O = MnO2 + 4 H+ + 2e Gypsum: Ca2þ þ 2SO2 4 þ 2H2 O ¼ CaSO4  2H2 O

5.75 9.35 10.8 1.4 3.8 22.5 4.89 25.34 18.0 43.6 4.58

a

Average composition and log K for schwertmannite from Bigham et al. (1996).

species activity [RCr(III)] for comparison to solubility relationships. Individual ion activities were calculated within MINTEQ4 using the extended Debye-Hu¨ckel with B-dot model, which is accurate up to ionic strengths of about 0.7 M (Parkhurst, 1990). This range in ionic strength covers most of the DAPL and associated groundwater range, although a few samples had ionic strengths up to 1.2 M. Geochemical modeling results show that the upper limit for RCr(III) in the DAPL is consistent with Cr(OH)3(am) solubility (Fig. 4). A linear regression of RCr(III) versus pH for pH <6 and RCr(III) > 106.5 M (which eliminates the nondetect samples), gives a slope of 2.15 (r2 = 0.87), that is close to the theoretical value of 2.0 expected for the hydrolysis reaction: ð4Þ

Fig. 4. Comparison of RCr(III) to the solubilities of Cr(III) hydroxides calculated by MINTEQ4 for site groundwaters.

that would be expected to dominate Cr(III) speciation under acidic conditions (Rai et al., 1987). The lower bound for RCr(III) is less well defined but can be represented by the solubility of (Cr0.69Fe0.31)(OH)3(am) (Fig. 4). This hydroxide has the highest proportion of Fe(III) in solid solution with Cr(III) for which solubility data are available (Sass and Rai, 1987). Sass and Rai (1987) were

unable to synthesize hydroxides by precipitation with Cr(III) to Fe(III) molar ratios greater than 0.69–0.31, indicating that solutions with higher proportions of Cr(III)–Fe(III) may preferentially form end-member solids. The DAPL contains substantially higher Cr(III) than Fe(III), hence forms

CrðOHÞ3 ðamÞ þ 2Hþ ¼ CrOH2þ þ 2H2 O

364

L.E. Eary, A. Davis / Applied Geochemistry 22 (2007) 357–369

hydroxides with high ratios of Cr(III)–Fe(III). Also, the Fe(III) produced by reaction (1) may be converted to Fe(II) by coupled electron-cation transfer reactions at the surfaces of Fe-containing silicates such as biotite (Eary and Rai, 1989). These reactions are known to be nonstoichiometric and effectively reduce aqueous Fe(III) to Fe(II) under acidic conditions (White and Yee, 1985). Solubility control by Cr(OH)3(am) is also indicated for precipitates created during DAPL titration. Chromium(III) concentrations determined after two days of equilibration (Fig. 5) usually exceeded the solubility of Cr(OH)3(am). However, after 16 days, Cr(III) concentrations approached Cr(OH)3(am) solubility constraints, indicating that as the DAPL is slowly neutralized Cr(III) is removed from solution as Cr(OH)3(am). Because Cr(OH)3(am) tends to remain noncrystalline (Ammonette and Rai, 1990), its solubility as a function of pH will define Cr(III) evolution in DAPL as it is slowly neutralized by water–rock interactions. The formation of Cr(OH)3(am) and Fe(OH)3(am) in the DAPL-affected portion of the aquifer was also indicated by increases in Cr and Fe solid concentrations in the aquifer matrix with depth. For example, Cr increased from 10 to 20 mg/kg between 0 and 4.6 m bgs to >1000 mg/kg at 9 m bgs, while Fe increased from about 0.5–3.6 wt.% over the same interval. Samples from >8.5 m bgs are within the DAPL (Fig. 2). EMPA of the aquifer samples from the DAPL zone showed that Cr was commonly associated with

Fig. 5. Comparison of RCr(III) concentrations determined in the DAPL precipitate solubility experiments to the solubility of Cr(OH)3(am) calculated by MINTEQ4.

Fe(III) and Al hydroxides (commonly present as coatings on layer silicates), typically containing from a few tenths to a few wt.% Cr (Fig. 6a) and less commonly as discrete particles Fig. 6b. Other types of secondary precipitates found in the DAPL zone include barite (BaSO4), gypsum ðCaSO4  2H2 OÞ (not shown) and Al-sulfates of indeterminate type (Fig. 6c), possibly jurbanite (AlSO4OH Æ 5H2O), basaluminite ðAl4 SO4 ðOHÞ10  5H2 OÞ, or alunite (KAl3(SO4)2(OH)6).

Fig. 6. EMPA photos of aquifer solids from borehole CPT-2; (a) chlorite grain rimmed by mixed Fe–Al–Cr hydroxides from 4.9-m bgs; (b) discrete particle of Fe–Al hydroxide with 1.1 wt.% Cr from 10-m bgs; (c) Al-silicate grain with an inner rim of an indeterminate Al-sulfate and outer rim of Fe oxyhydroxide with scattered blebs of Al-sufate and barite from 10-m bgs.

L.E. Eary, A. Davis / Applied Geochemistry 22 (2007) 357–369

Based on the modeling results, it seems unlikely that the small Cr percentages within the Fe–Al hydroxides determined by EMPA represent true solid solutions because much lower RCr(III) levels than observed would be expected if (Cr0.01Fe0.99)(OH)3(am) type solids were controlling the solubility of dissolved Cr (Fig. 4). Instead, the hydroxides are probably comprised of fine-grained Fe–Al hydroxide mixed with small amounts of relatively pure Cr(OH)3(am) particles, although this could not be clearly distinguished from EMPA. The typical occurrence of the hydroxides as coatings on layer silicates indicates that hydrolysis reac-

365

tions have caused local neutralization of the DAPL and precipitation of secondary hydroxides which also serve as solubility controls for other cationic metals besides Cr(III). For example, with a few exceptions at pH > 5.5, RAl(III) levels are bounded by the solubilities of Al(OH)3(am), microcrystalline gibbsite [Al(OH)3(lc)] phases (Fig. 7a) which have been postulated as solubility controls for Al in surface waters (Driscoll et al., 1984). In acid mine drainage areas, Al concentrations are typically constrained by the solubilities of Al(OH)3(am), Al(OH)3(lc) and a variety of Al-sulfate phases (Nordstrom, 1982; Nordstrom and Ball, 1986;

Fig. 7. Levels of (a) RAl(III); (b) RFe(III) (SO4 concentrations in mg/L); and (c) RMn(II) (SO4 concentrations in mg/L) in site groundwater compared to solubilities of common secondary minerals.

366

L.E. Eary, A. Davis / Applied Geochemistry 22 (2007) 357–369

Sullivan et al., 1988), which are also important Al solubility controls in the DAPL-affected aquifer. Iron(III) concentrations in acidic systems are often constrained by Fe(OH)3(am) solubility (Langmuir and Whittemore, 1971). A plot of RFe(III) against pH for site groundwater indicates that Fe(OH)3(am) controls Fe solubility at pH <4.5, but there is substantial deviance to higher values at higher pH (Fig. 7b). Solutions were also oversaturated with schwertmannite [ideally Fe8O8(OH)6SO4] indicating that it may also control Fe(III) solubility (Bigham et al., 1996). The low pH solutions also illustrate the increase in Fe(OH)3(am)

solubility caused by the formation of FeSOþ 4 and  FeðSO4 Þ2 under the acidic conditions of the DAPL, which contains up to 90,000 mg/L SO4. The deviances from the predicted solubility curve at pH > 4.5 may reflect disequilibrium with Fe(OH)3(am) or variability in Fe(OH)3(am) solubility with pH (Langmuir and Whittemore, 1971). Such discrepancies may also be caused by passage of colloidal particles through filters during sample collection. The behavior of total Mn(II) is similar to Fe(III) with a linear decrease up to about pH 5.5 and increasing scatter at higher pH (Fig. 7c). Under

Fig. 8. Saturation indices calculated by MINTEQ4 for site groundwaters for different solids, including (a) Gypsum, (b) Jurbanite, and (c) Basaluminite.

L.E. Eary, A. Davis / Applied Geochemistry 22 (2007) 357–369

acidic conditions, RMn(II) shows an upper limit consistent with manganite (c-MnOOH) solubility and a lower limit approximated by birnessite (dMnO2) solubility. While manganite is known to form from the oxidation of Mn(II) and is favored over feitknechtite (b-MnOOH) in SO4-containing solutions (Hem and Lind, 1983; Kessick and Morgan, 1975), it is unlikely that there could be much oxidation of Mn(II) under the DAPL redox conditions. Given that Mn oxyhydroxides form only under oxidizing conditions (Hem and Lind, 1983), it is more likely that the Mn concentrations in the DAPL reflect only dissolution of oxide, hydroxide, and silicates because of the acidic conditions, with an absence of Mn re-precipitation. The major anion in the DAPL is SO4, hence metal sulfates may be important solubility controls in the most acidic portions of the plume. Saturation indices for gypsum show a trend from undersaturation at pH > 4.5 to slight oversaturation at pH <4.5 (Fig. 8a). Consistent with the observed solubility equilibria, a CaSO4 phase, presumably gypsum (a common precipitate in disposal pits during operation of the facility) was found by EMPA in a sample from the DAPL zone. Hydrated Al–SO4 phases were also observed in EMPA of samples from the DAPL zone. Solutions with pH <4.8 were commonly oversaturated with jurbanite (Fig. 8b), while those with pH > 4.0 were also oversaturated with basaluminite (Fig. 8c), resulting in an overlapping formation region for these two Al–SO4 solids. The modeling results also indicated that conditions of oversaturation with alunite solubility existed at pH > 4. Basaluminite is metastable

367

compared to alunite over its probable stability range but alunite usually requires temperatures >25 C for formation, whereas basaluminite precipitates directly from low temperature solutions (Nordstrom, 1982). Consequently, in the acidic DAPL-aquifer system, basaluminite may be a more likely secondary solid than alunite. Jurbanite, basaluminite, and an unnamed AlSO4OH phase have been postulated to be solubility controls for dissolved Al in acidic systems (Nordstrom, 1982; Karathanasis et al., 1988; Van Breemen, 1973) comparable to those found in the DAPL-affected aquifer. 5. Conclusions The practical effects of solubility controls on metal mobilities are placed in perspective by comparing SO4 and Cl–Cr(III) and Al concentrations with distance downgradient from the DAPL origin (Fig. 9). Sulfate and Cl decrease by a factor of 102–103 over a distance of 600–1150 m in the downgradient direction. These anions migrate conservatively, hence the decrease represents dispersion and dilution of DAPL-derived solutes within the shallow aquifer. In contrast, Cr(III) and Al decrease by a factor of 104–105, considerably more than the anions because of the strong pH solubility control of their respective hydroxide solids. The net result is a chromatographic separation of unretarded solutes in the distal portion of the plume from less mobile metals in the DAPL proximal to the site. These findings have implications for the efficacy of natural attenuation (NA) of acidic groundwater systems, in that as the pH is neutralized, NA reactions effectively constrain the downgradient transport of reactive constituents. For Cr, the conceptual chemical model based on empirical studies of solubility and redox kinetics accurately portrays Cr mobility in the aquifer at and downgradient from the original disposal site. Acknowledgements

Fig. 9. Changes in groundwater solute concentrations with horizontal distance from the origin of the DAPL plume.

The authors gratefully acknowledge the assistance of Dr. John Drexler for solid EMPA, Fred Luizer, Elizabeth Sopher, and Laura Triplet for ICP/AES and IC analyses and maintenance of experimental studies; all at the Department of Geological Sciences, University of Colorado, Boulder, Colorado. Reviews by J.W. Ball and G. Lee are gratefully acknowledged as resulting in improvements to the paper.

368

L.E. Eary, A. Davis / Applied Geochemistry 22 (2007) 357–369

References Ammonette, J.E., Rai, D., 1990. Identification of noncrystalline (Fe, Cr)(OH)3 by infrared spectroscopy. Clays Clay Min. 38, 129–136. Banerjee, D., Nesbitt, H.W., 1999. Oxidation of aqueous Cr(III) at birnessite surfaces: constraints on reaction mechanisms. Geochim. Cosmochim. Acta 63, 1671–1687. Baron, D., Palmer, C.D., 1998. Solubility of KFe(CrO4)2 Æ 2H2O at 4–75 C. Appl. Geochem. 13, 961–973. Baron, D., Palmer, C.D., Stanley, J.T., 1996a. Identification of two, iron-chromate precipitates in a Cr (VI) contaminated soil. Environ. Sci. Technol. 30, 964–968. Baron, D., Palmer, C.D., Stanley, J.T., 1996b. Solubility of KFe3(CrO4)2(OH)6 at 4 to 35 C. Geochim. Cosmochim. Acta 60, 3815–3824. Bartlett, R.L., James, B., 1979. Behavior of chromium in soils. (III). Oxidation. J. Environ. Qual 8, 31–35. Bigham, J.M., Schwertmann, U., Traina, S.J., Winland, R.L., Wolf, M., 1996. Schwertmannite and the chemical modeling of iron in acid sulfate waters. Geochim. Cosmochim. Acta 60, 2111–2121. Blowes, D.W., Ptacek, C.J., Jambor, J.L., 1997. In-situ remediation of Cr(VI)-contaminated groundwater using permeable reactive wells: laboratory studies. Environ. Sci. Technol. 31, 3348–3357. Brigatti, M.F., Franchini, G., Lugli, C., Medici, L., Poppi, L., Turci, E., 2000. Interaction between aqueous chromium solutions and layer silicates. Appl. Geochem. 15, 1307–1316. Buerge, I.J., Hug, S.J., 1997. Kinetics and pH dependence of chromium(VI) reduction by iron(II). Environ. Sci. Technol. 31, 1426–1432. Davis, A., Olsen, R.L., 1995. The geochemistry of chromium migration and remediation in the subsurface. Ground Water 33, 759–768. Davis, A., Drexler, J., Ruby, M., Nicholson, A., 1993. Micromineralogy of mine wastes in relation to lead bioavailability, Butte, Montana. Environ. Sci. Technol. 27, 1415–1425. Deng, B., Stone, A.T., 1996. Surface catalyzed chromium(VI) reduction: reactivity comparisons of different organic reductants and different oxide surfaces. Environ. Sci. Technol. 30, 2484–2494. Driscoll, C.T., Baker, J.P., Bisogni, J.J., Schofield, C.L., 1984. Aluminum speciation and equilibria in dilute acidic surface waters of the Adirondack Region of New York State. In: Bricker, O.P. (Ed.), Geological Aspects of Acid Deposition. Butterworth Publishers, Boston, pp. 5–75. Eary, L.E., Jenne, E.A., 1992. Version 4.00 of the MINTEQ Geochemical Code, PNL-8190/UC-204, Pacific Northwest Laboratory, Richland, Washington. Eary, L.E., Rai, D., 1987. Kinetics of chromium(III) oxidation to chromium(VI) by reaction with manganese dioxide. Environ. Sci. Technol. 21, 1187–1193. Eary, L.E., Rai, D., 1988. Chromate removal from aqueous wastes by reduction with ferrous ion. Environ. Sci. Technol. 22, 972–977. Eary, L.E., Rai, D., 1989. Kinetics of chromate reduction by ferrous ions derived from hematite and biotite at 25 C. Am. J. Sci. 289, 180–213. Eary, L.E., Rai, D., 1991. Chromate reduction by subsurface soils under acidic conditions. Soil Sci. Soc. Am. J. 55, 676– 683.

Fendorf, S.E., Li, G., 1996. Kinetics of chromate reduction by ferrous iron. Environ. Sci. Technol. 30, 1614–1617. Fendorf, S., Wielenga, B.W., Hansel, C.M., 2000. Chromium transformations in natural environments: The role of biological and abiological processes in chromium (VI) reduction. Internat. Geol. Rev. 42, 691–701. Hach, 1994. DR/2000 Spectrophotometer Procedures Manual. Hach Co., Loveland, CO. Hem, J.D., Lind, C.J., 1983. Non-equilibrium models for predicting forms of precipitated manganese oxides. Geochim. Cosmochim. Acta 47, 2037–2046. Hug, S.J., Laubscher, H.U., James, B.R., 1997. Iron(II) catalyzed photochemical reduction of chromium(VI) by oxalate and citrate in aqueous solutions. Environ. Sci. Technol. 31, 160– 170. James, B.R., 1996. The challenge of remediating chromiumcontaminated soil. Environ. Sci. Technol. 30, 248A–251A. Karathanasis, A.D., Evangelou, V.P., Thompson, Y.L., 1988. Aluminum and iron equilibria in soil solutions and surface waters of acid mine watersheds. J. Environ. Qual. 17, 534– 543. Kessick, M.A., Morgan, J.J., 1975. Mechanism of autoxidation of manganese in aqueous solution. Environ. Sci. Technol. 9, 157–159. Krupka, K.M., Erikson, R.L., Mattigod, S.V., Schramke, J.A., Cowan, C.E., Eary, L.E., Morrey, J.R., Schmidt, R.L., Zachara, J.M., 1988. Thermochemical data used by the FASTCHEM package, Final Report EA-5872, Electric Power Research Institute, Palo Alto, CA. Langmuir, D., Whittemore, D. O., 1971. Variations in the stability of precipitated ferric oxyhydroxides. In: Hem, J.D. (Ed.), Nonequilibrium Systems in Natural Water Chemistry, Adv. Chem. Ser., No. 106, Am. Chem. Soc., Washington, DC, pp. 209–233. Loyaux-Lawniczak, S., Refait, P., Lecomte, P., Ehrhardt, J.J., Genin, J.M., 1999. The reduction of chromate ions by Fe (II) layered hydroxides. Hydrol. Earth Systems Sci. 3, 593– 599. Makos, J.D., Hrncir, D.C., 1995. Chemistry of Cr(VI) in a constructed wetland. Environ. Sci. Technol. 29, 2414–2419. Nordstrom, D.K., 1982. The effect of sulfate on aluminum concentrations in natural waters: some stability relations in the system Al2O3–SO4–H2O at 298 K. Geochim. Cosmochim. Acta 46, 681–692. Nordstrom, D.K., Ball, J.W., 1986. The geochemical behavior of aluminum in acidified surface waters. Science 232, 54–56. Olazabal, M.A., Nikolaidis, N.P., Suib, S.A., Madariaga, J.M., 1997. Precipitation equilibria of the chromium(VI)/Fe(III) system and spectroscopic characterization of the precipitates. Environ. Sci. Technol. 31, 2898–2902. O’Loughlin, E.J., Burris, D.R., 2004. Reduction of halogenated ethanes by green rust. Environ. Toxicol. Chem. 23, 41–48. Palmer, C.D., Wittbrodt, P.R., 1991. Processes affecting the remediation of chromium-contaminated sites. Environ. Health Perspectives 92, 25–40. Parkhurst, D. L., 1990. Ion-association models and mean activity coefficients of various salts. In: Melchior, D. C., Bassett, R. L. (Eds.), Chemical Modeling of Aqueous Systems, II; ACS Symp. Ser. 416, Am. Chem. Soc., Washington, DC, pp. 30– 43. Peterson, M.L., Brown, G.E., Parks, G.A., Stein, C.L., 1997. Differential redox and sorption of Cr(III/VI) on natural

L.E. Eary, A. Davis / Applied Geochemistry 22 (2007) 357–369 silicate and oxide minerals: EXAFS and XANES results. Geochim. Cosmochim. Acta 61, 3399–3412. Pettine, M., Millero, F.J., La Noce, T., 1991. Chromium(III) interactions in seawater through its oxidation kinetics. Mar. Chem. 34, 29–46. Pratt, A.R., Blowes, D.W., Ptacek, C.J., 1997. Products of chromate reduction on proposed subsurface remediation material. Environ. Sci. Technol. 31, 2492–2498. Puls, R.W., Paul, C.J., Powell, R.M., 1999. The application of in situ permeable reactive (zero-valent iron) barrier technology for the remediation of chromatecontaminated groundwater: a field test. Appl. Geochem. 14, 989–1000. Rai, D., Sass, B.M., Moore, D.C., 1987. Chromium(III) hydrolysis constants and solubility of chromium(III) hydroxide. Inorg. Chem. 26, 345–349. Rai, D., Eary, L.E., Zachara, J.M., 1989. Environmental chemistry of chromium. Sci. Total Environ. 86, 15–23. Saleh, F.R., Parkerton, T.F., Lewis, R.V., Huang, J.H., Dickson, K.L., 1989. Kinetics of chromium transformations in the environment. Sci. Total Environ. 86, 25–41. Sass, B.M., Rai, D., 1987. Solubility of amorphous chromium(III)–iron(III) hydroxide solid solution. Inorg. Chem. 26, 2228–2232. Skougstad, M .W., Fishman, M.J., Friedman, L.C., Erdman, D.E., Duncan, S.S., 1979. Methods for determination of inorganic substances in water and fluvial sediments Book 5, Chap. A1. Techniques of Water-Resources Investigations of the US Geol. Surv., US Government Printing Office, Washington, DC, 339–340.

369

Stollenwerk, K.G., Grove, D.B., 1985. Adsorption and desorption of hexavalent chromium in an alluvial aquifer near Telluride, Colorado. J. Environ. Qual. 14, 150–155. Sullivan, P.J., Yelton, J.L., Reddy, K.J., 1988. Solubility relationships of aluminum and iron minerals associated with acid mine drainage. Environ. Geol. Water Sci. 11, 283–287. Tseng, J.K., Bielefeldt, A.R., 2002. Low-temperature chromium(VI) biotransformation in soil with varying electron acceptors. J. Environ. Qual. 31, 1831–1841. Van Breemen, N., 1973. Dissolved aluminum in acid sulfate soils and in acid mine water. Soil Sci. Soc. Amer. Proc. 37, 694– 697. White, A.F., Yee, A., 1985. Aqueous oxidation-reduction kinetics associated with coupled electron-cation transfer from ironcontaining silicates at 25C. Geochim. Cosmochim. Acta 49, 1263–1275. Wielenga, B., Mizuba, M.M., Hansel, C.M., Fendorf, S., 2001. Iron promoted reduction of chromate by dissimilatory ironreducing bacteria. Environ. Sci. Technol. 35, 522–527. Wittbrodt, P.R., Palmer, C.D., 1996. Effect of temperature, ionic strength, background electrolytes, and Fe(III) on the reduction of hexavalent chromium by soil humic substances. Environ. Sci. Technol. 30, 2470–2477. Zachara, J.M., Girvin, D.C., Schmidt, R.L., Resch, C.T., 1987. Chromate adsorption on amorphous iron oxyhydroxide in the presence of major groundwater ions. Environ. Sci. Technol. 21, 589–594. Zachara, J.M., Ainsworth, C.C., Cowan, C.E., Resch, C.T., 1989. Adsorption of chromate by subsurface soil horizons. Soil Sci. Soc. Am. J. 53, 418–428.