Determination of hexavalent chromium concentrations in matrix porewater from a contaminated aquifer in fractured sedimentary bedrock

Chemical Geology 419 (2015) 142–148

Contents lists available at ScienceDirect

Chemical Geology journal homepage: www.elsevier.com/locate/chemgeo

Determination of hexavalent chromium concentrations in matrix porewater from a contaminated aquifer in fractured sedimentary bedrock Jiujiang Zhao a,⁎, Tom Al a,1, Steven W. Chapman b, Beth Parker b, Katherine R. Mishkin c, Diana Cutt d, Richard T. Wilkin e a

Department of Earth Sciences, University of New Brunswick, New Brunswick E3B 5A3, Canada University of Guelph, School of Engineering, Ontario N1G 2W1, Canada c Superfund Technical Support Section U.S. EPA Region 2, U.S. Environmental Protection Agency, United States d Superfund Technology Liaison, Regional Science Program, Office of Research and Development, EPA Region 2, U.S. Environmental Protection Agency, United States e National Risk Management Research Laboratory, Ground Water and Ecosystems Restoration Division, U.S. Environmental Protection Agency, United States b

a r t i c l e

i n f o

Article history: Received 8 June 2015 Received in revised form 22 October 2015 Accepted 23 October 2015 Available online 25 October 2015 Keywords: Hexavalent chromium Bedrock Porewater Alkaline extraction Cation exchange ICP-MS

a b s t r a c t A new method for quantification of hexavalent chromium (Cr(VI)) in the porewater of rock core samples from contaminated sedimentary bedrock has been developed here. The method combines alkaline extraction with cation exchange column separation followed by determination of Cr concentrations by inductively coupled plasma mass spectrometry (ICP-MS). A porewater detection limit of 45 μg/L was determined by performing extractions on uncontaminated samples, and accounts for dilution of porewater volumes by the extraction solution. Recoveries of Cr(VI) in quality control (QC) samples were greater than 90% and there was no significant interference from Cr(III). Relative standard deviations (RSD) were less than 10% for QC samples spiked with Cr(VI), and 2 to 47% (average of 21%) for replicate analyses of core samples. Cr(VI) analyses were conducted on depth-discrete core samples collected at intervals of b 0.3 m from sandstone and siltstone bedrock within a contaminated groundwater plume. Groundwater samples were collected using multilevel well ports and were also analyzed for Cr(VI) concentrations. Significant Cr(VI) anomalies were observed in the rock matrix of the core samples. Overall, we observe general agreement in the Cr(VI) concentrations between the samples of immobile rock-matrix porewater and the samples of groundwater which is mobile in rock fractures. This method provides a viable procedure for determination of Cr(VI) concentration in bedrock porewater, and these datasets are valuable for developing conceptual models, assessing plume transport and fate, and for considering remedial options. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Chromium has two stable oxidation states, trivalent Cr(III) and hexavalent Cr(VI). At low concentration levels, Cr(III) is an essential micronutrient (Anderson, 1997), whereas Cr(VI) is considered a toxin and carcinogen (Barceloux, 1999). Cr(VI) is a contaminant of concern at many contaminated sites, including many Superfund sites (N 10%) in the U.S., where in 2011 it was #17 on the Priority List of Hazardous Substances for the Agency for Toxic Substances and Disease Registry (ATSDR, 2013). The US EPA drinking water standard for total chromium is 100 μg/L (USEPA, 2012), and according to the Canadian Drinking ⁎ Corresponding author at: Department of Earth and Environmental Sciences, University of Ottawa, Ontario K1N 6N5, Canada. E-mail address: [email protected] (J. Zhao). 1 Present address: Department of Earth and Environmental Sciences, University of Ottawa, ON, K1N 6N5, Canada.

http://dx.doi.org/10.1016/j.chemgeo.2015.10.034 0009-2541/© 2015 Elsevier B.V. All rights reserved.

Water Guidelines, the maximum acceptable concentration (MAC) for total chromium is 50 μg/L (Health Canada, 2012). In most naturally occurring minerals chromium is trivalent, such as in the mineral chromite (FeCr2O4). In groundwater, Cr(III) is effectively immobile due to its low solubility at near-neutral pH. In contrast, the mobility of Cr(VI) is relatively high when pH is neutral or above, but in neutral to acidic groundwater its mobility is limited by adsorption and reduction to Cr(III) (Nriagu, 1988). Anthropogenic Cr(VI) sources from tank leakage at chrome plating operations, or improper disposal practices are common causes of Cr(VI) contamination in groundwater (Palmer and Wittbrodt, 1991). Knowledge of the physical and chemical-reaction processes affecting Cr(VI) migration in groundwater is required to develop conceptual models and consider possible remediation options (Kent et al., 2007; Palmer and Wittbrodt, 1991). Many studies of Cr(VI) in groundwater indicate that its fate is controlled by the reduction– oxidation of Cr(VI)–Cr(III) (Fantoni et al., 2002; Gray, 2003; Oze et al., 2007), and the transport of Cr(VI) with mobile groundwater in fractured

J. Zhao et al. / Chemical Geology 419 (2015) 142–148

porous bedrock is subject to attenuation by diffusion, adsorption and reduction in the rock matrix adjacent to fractures (Fantoni et al., 2002; Friedly et al., 1995). As with other dissolved contaminants, diffusion of Cr(VI) from groundwater flowing in fractures into the immobile porewater within the rock matrix can strongly attenuate plumes in fractured sedimentary rock, but also may cause long-term persistence due to back diffusion and hinder remediation efforts (Parker et al., 2010). Standard groundwater sampling and analytical methods are available to monitor contamination via conventional wells or multilevel monitoring systems. These methods sample groundwater primarily from the fracture networks. To understand the fate and transport of Cr(VI) in fractured porous bedrock, an appropriate sampling and analytical procedure is required for the determination of Cr(VI) concentrations in the matrix porewater where much of the contaminant mass resides at many contaminated sites.

143

In conducting research into the fate and transport of chlorinated solvents in fractured porous bedrock, Parker et al. (2012) have developed methods using the Discrete Fracture Network (DFN) approach. Utilizing the DFN method, it is possible to quantify the porewater concentrations of volatile organic compounds (VOCs) at contaminated sites, through drilling and extraction of VOCs from rock core samples, allowing for the creation of detailed profiles of contaminant concentrations. Such measurements are integrated into an overall methodology combining matrix sampling and analysis for contaminant concentrations with several other types of borehole measurements and core sample analyses. Our study was motivated by the expectation that application of the DFN approach will improve the understanding of matrix diffusion and reaction processes for Cr(VI) in bedrock porewater, and represents the first DFN application for this contaminant. The conceptual model for Cr(VI) contamination in fractured porous bedrock is shown in Fig. 1,

Fig. 1. Schematics of (a) cored hole collected from plume zone for rock core sub-sampling and the diffusion of the Cr(VI) from contaminated groundwater into bedrock matrix, (b) comparison between porewater Cr(VI) from subsampled rock core samples and groundwater Cr(VI) from the sample collected from MLS port and (c) hypothetical Cr(VI) matrix profiles off a fracture for a scenario with a finite source term (declining after 10 years).

144

J. Zhao et al. / Chemical Geology 419 (2015) 142–148

where the contaminant initially resides in the mobile groundwater contained in fractures but gradually diffuses into the adjacent matrix (Fig. 1a). Fig. 1b demonstrates that Cr(VI) concentrations in groundwater samples are collected over relatively large intervals (e.g. 3 m) using a FLUTe™ multilevel monitoring system (MLS), and are likely to reflect the concentrations in the mobile fracture water which may be higher or lower than in the matrix porewater in cases of expanding or shrinking plumes, respectively. Over time, diffusion leads to elevated concentrations in the matrix porewater (Fig. 1a, b), and if the source is removed during remediation efforts or becomes depleted, the contaminated porewater represents a long-term persistent source of contamination to the groundwater in fractures via back diffusion (Fig. 1c). Several methods have been developed for extracting Cr(VI) from a variety of solid materials such as soils and sediments (Unceta et al., 2010), but extraction methods for bedrock samples are not reported in the literature. A method using heated 0.28 mol/L Na2CO3 and 0.5 mol/L NaOH (EPA 3060A) has been shown to be an effective method for extracting both aqueous and solid-phase Cr(VI) species in soils (USEPA, 1996a), but this method is intended for colorimetric analysis which has relatively high detection limits. For example, EPA Method 7196A has a suggested Cr(VI) concentration range of 0.5 to 50 mg/L (USEPA, 1992). Alternatively, EPA 3060A could be used with the intention of quantifying Cr concentrations by ICP-MS, but high total dissolved solids (TDS) from the extraction reagents necessitates large dilutions prior to ICP-MS analysis, which leads to detection limits on the order of 1 mg/L. Other extraction methods have been applied, including unheated Na2CO3 − NaOH extraction, phosphate buffer extraction, and sonication with 0.1 mol/L NaOH (James et al., 1995), but they also result in high TDS and unsatisfactory detection limits. In alkaline conditions, Cr(VI) is retained in solution because reduction to Cr(III) is inhibited. However, Cr(III) may form soluble species at high pH, potentially contributing an artifact to reported Cr(VI) concentrations (Johnson, 1990; Richard and Bourg, 1991). Other artifacts affecting the final reported Cr(VI) recovery may arise from adsorption of Cr(III) to colloidal solids, particularly Fe oxyhydroxides, and from reduction of Cr(VI) to Cr(III). Considering these constraints, extraction solutions must use a high pH to prevent reduction of Cr(VI) and retain it in solution. Filtration is required to remove fine particulates with adsorbed Cr(III), and any remaining Cr(III) can be removed by a cation-exchange column (Ball and McCleskey, 2003; Johnson, 1990). The cation exchange column can also retain residual colloids and major cations (Na, K, Ma and Ca) to reduce TDS. Ultimately, the amount of Cr(VI) that can be extracted from bedrock porewater is limited by the porewater content or porosity, and the Cr(VI) concentration in the porewater. The objectives of this research were to: 1) develop a new analytical procedure to determine the concentration of Cr(VI) in bedrock porewater, and 2) compare Cr(VI) concentrations in porewater to those in groundwater sampled from standard multilevel wells. The goals of the new procedure were to: 1) avoid biases resulting from Cr(VI) reaction (e.g., negative bias resulting from Cr(VI) reduction to Cr(III)), and 2) provide minimum detection limits for Cr(VI) in porewater that are sufficiently low to allow for a mechanistic understanding of processes controlling its fate and transport. 2. Study site and material The study site is a former electroplating facility in New Jersey, U.S.A. where chromic acid was released approximately three decades ago from a ruptured storage tank to a groundwater system comprising a shallow unconsolidated aquifer with underlying fractured porous sandstone and siltstone of the Passaic Formation (Herman, 2001). A plume of Cr(VI)-contaminated groundwater moving toward a major river over a distance of approximately 800 m has since developed down gradient in overburden and bedrock. Bedrock core samples were collected from within the plume about midway between the site and river. Continuous core samples were

collected in 1.5 m runs using conventional wireline drilling techniques and a triple-tube core barrel, which minimizes mechanical breaks and provides cores that reflect in-situ fracture distributions. Cores were collected over an interval of 21 to 108.4 m below the ground surface (bgs) for a total cored interval of 87.4 m. A total of 400 samples were collected from the cores over this interval for an average sample spacing of about 0.22 m. Immediately after obtaining each drill core, they were photographed and logged in detail, and sampled at discrete interval lengths of 0.03 to 0.06 m using a hammer and chisel. The core subsamples were immediately wrapped in aluminum foil and parafilm, placed in Mil-Spec film-foil bags, and vacuum sealed to preserve moisture and redox conditions. Supporting datasets were collected from the borehole samples, including geophysical logs and transmissivity profiles (Keller et al., 2013), followed by MLS installation (Cherry et al., 2007) for depth-discrete groundwater sampling and head measurements. The MLS system has 7 ports, each spanning a 3.0 m depth interval. Of the 400 rock samples collected, 147 were selected for assessment of Cr(VI) distribution in the rock matrix, with a focus on zones overlapping with MLS ports, but also filling data gaps between and above the shallower MLS ports where Cr(VI) concentrations were highest. Prior to analysis, the rock samples were removed from the packaging material, the outer rinds trimmed using a hammer and chisel, and a smaller subsample from the inner part of the core sample was crushed using a hydraulic press with stainless steel plates. The crushed material was sieved (b 2 mm mesh) to remove coarse fragments and samples were stored in a sealed glass bottle in a refrigerator prior to extraction. 3. Methods 3.1. Extraction method For extraction of Cr(VI), approximately 1.5 g of crushed sample was added to 10 mL of 0.01 mol/L NaOH (99.99%, Sigma-Aldrich) in 15 mL centrifuge tubes that were sealed and placed on a rotator. Samples were extracted for 30 min, and then centrifuged for a minimum of 10 min, or until visibly clear. The supernatant was transferred to another 15 mL centrifuge tube. The crushed rock samples were then rinsed with 5 mL of 0.01 mol/L NaOH to remove any remaining Cr(VI), centrifuged, and the supernatant solutions were combined with the initial 10 mL of extract solution. The extract solutions were filtered sequentially with 0.45 and 0.1 μm filters and then acidified to pH 2–5 using 1 mol/L HNO3 (Plasma Pure Plus, SCP Sciences). Extractions were also conducted following the same procedure but using deionized water instead of NaOH to evaluate the importance of maintaining high pH during the extraction procedure. 3.2. Cation exchange column separation procedure During the method development experiments, it was found that filtration alone, even to b 0.1 μm, did not remove traces of Cr(III) associated with Fe-rich particulate matter. Several ion exchange methods have been used for chromium speciation studies in aqueous solutions (Ball and McCleskey, 2003; Frenzel, 1998; Johnson, 1990), and in this study, the method was further developed to remove trace amounts of Cr(III) using cation exchange. The cation exchange resin (AG50W-x8 100-200, Eichrom Technologies, LLC) was pre-cleaned using three repeat flushes with ethanol to remove organic impurities, and then rinsed with DI water to remove the ethanol. The resin was then further cleaned with three repeat washes using 50% (v/v) HNO3 followed by a DI water flush. The cleaned resin was stored in 1% HNO3 solution. The columns (0.8 cm diameter × 4 cm length, polypropylene; Bio-Rad Laboratories, Canada, Ltd.) were prepared by pipetting 1 mL of AG50W-x8 resin, cleaning with 15 mL 50% HNO3 and conditioning with 15 mL 0.01% HNO3. The extract solutions (15 mL) were loaded into the columns and the columns were eluted with 10 mL of 0.01% HNO3. The final eluent (25 mL) was collected and acidified to a final concentration of 1% HNO3.

J. Zhao et al. / Chemical Geology 419 (2015) 142–148

145

The efficiency of column separation was tested by loading 10 μg/L of K2Cr(VI)2O7 and Cr(III)(NO3)3 in 10 mL 0.01% HNO3 in the columns and then eluting with 15 mL of 0.01% HNO3 (5 × 3 mL). To optimize the sample pH range for column separation, the pH dependence of cation exchange separation efficiency was also tested by elution of a series of standard Cr(VI) and Cr(III) solutions over the pH range 0–12. 3.3. Quality control Reagent blanks and silica-sand blanks were subjected to the entire extraction procedure to obtain the method detection limit (MDL) and extraction detection limit (EDL) respectively (details are provided in Section 4.4). An in-house laboratory reference sample (LRS) was prepared by homogenizing a mixture of crushed material from five bedrock samples (sandstone/siltstone), which were collected from 37 to 48 m bgs. One LRS was extracted for each batch (total three batches) and analyzed to ensure extraction and measurement consistency. Cr(VI) and Cr(III) were spiked into blanks and the LRS to monitor the recovery of Cr(VI) and interference from Cr(III). Eight samples from 21 to 103 m bgs were extracted and analyzed in triplicate to evaluate method precision, and splits of the same samples were spiked with Cr(VI) to monitor the recovery of Cr(VI). 3.4. Determination of Cr using ICP-MS Analyses for total Cr were conducted by ICP-MS (Agilent 7700x) in He-gas mode, which can remove polyatomic interferences (40Ar12C+, 35 17 + Cl O and 35Cl16OH+) on 52Cr and provides high signal to noise ratio (Rakhunde et al., 2012; Unceta et al., 2010). The most abundant Cr isotope, 52Cr, was selected to quantify the concentration of Cr. An internal standard of 45Sc (0.5 mg/L Sc) was used to monitor instrument stability. Matrix-matched calibration standards were used to account for possible matrix effects. 4. Results and discussion 4.1. Method development for Cr(VI) extraction In previous studies, the Cr(VI) extraction time for solid samples ranges from 5 min to 24 h using different extraction reagents (James et al., 1995; Mandiwana, 2008; Narukawa et al., 2007; USEPA, 1996a). Because of their high solubility (Barceloux, 1999), Cr(VI) compounds can be extracted by water or alkaline solutions, but it is possible that Cr(VI) in solution could react with Fe(II) on fresh surfaces of Fe(II)containing minerals created during the crushing process. This reaction is strongly pH dependent and high pH inhibits reduction of Cr(VI) as shown in Eq. (1): þ 3ΞFe–O°þ CrO2 4 þ 2H þ 5H2 O→3FeðOHÞ3 ðsÞ þ CrðOHÞ3 ðsÞ

Fig. 2. Comparison of DI water extraction and 0.01 mol/L NaOH extraction for the 1.5 g laboratory-made reference samples, with extraction times from 2 to 30 min. ▲: 0.01 mol/L NaOH ♦: DI water (error bars represent the standard deviation of triplicate measurements).

were higher than DI water, and reached a plateau after 10 min. No loss of Cr was observed over a period of 10 to 30 min, indicating an extraction time of 30 min was sufficient. The pH of the NaOH solution was monitored during the experiment and did not drop below 12. 4.2. Optimization of the column separation procedure Tests of the efficiency of column separation (Fig. 3) involved measurements of Cr(VI) in the effluent from five sequential 3-mL eluent additions (0.01% HNO3). Following sample loading, but prior to flushing, analyses of Cr(VI) in the effluent indicated that 90% had passed through the column. After flushing with 9 mL of 0.01% HNO3, 96% was eluted. No Cr was detected in the effluent for blanks spiked with Cr(III), indicating that Cr(III) was retained in the cation exchange column. These results demonstrate the effectiveness of our procedure for separating Cr(III) from Cr(VI). At low pH, cations are the major form of Cr(III) (i.e., Cr3+, Cr(OH)2+ and Cr(OH)+ 2 ) (Rai et al., 1987), and as such, they were retained on the cation exchange resin (AG50W-x8). Ball and McCleskey (2003) report good recovery of Cr(VI) for cation exchange separation in the pH range 2 to 11, and low recovery of Cr(VI) at pH b 2. At low pH, the reduction of Cr(VI) may occur. At high pH (N9), Cr(III) forms the relatively soluble species Cr(OH)− 4 (Rai et al., 1987) that may pass through the column. To test for these artifacts, column separations were conducted

ð1Þ

where ΞFe–O° represents a neutral Fe(II) oxide surface species. A solution of 0.01 mol/L NaOH (pH 12) was used as the extractant in this study. Compounds of Cr(III), such as Cr(OH)3, have low solubility in neutral solutions but alkaline conditions could lead to an analytical interference from Cr(III) (Rai et al., 1987). Accordingly, the method includes a cation exchange step to remove traces of Cr(III) and the QC protocol involved Cr(III) spikes. A higher concentration of NaOH is effective for Cr(VI) extraction (James et al., 1995), however, the higher TDS is problematic for ICP-MS analysis and higher pH increases the potential for interference from the Cr(OH)− 4 species. Comparison of DI water and 0.01 mol/L NaOH extraction is shown in Fig. 2. Extracted Cr increased in the first 5 min for both DI water and NaOH, but a decrease of Cr was observed in the DI water extraction after 10 min, indicating loss of Cr(VI). The decrease likely reflects loss of Cr from solution due to Cr(VI) reduction and precipitation of insoluble Cr(III) hydroxide. Cr concentrations in 0.01 mol/L NaOH extractant

Fig. 3. Elution profile of Cr(VI) after sample loading in the AG50w-x8 cation exchange columns. Cr(VI) sample: 10 mL of K2Cr2O7 (10 μg/L).

146

J. Zhao et al. / Chemical Geology 419 (2015) 142–148

at variable pH (0–12) on samples spiked with either Cr(VI) or Cr(III) (10 μg/L). Consistent with Ball and McCleskey (2003), recoveries indicate complete elution of Cr(VI) over the pH range of 2–12. For samples spiked with Cr(III), no Cr was detected in the effluent from columns operated between pH 2 and 10, while a slight increase in Cr (up to 2% of the total spike) was eluted from the columns above pH 10. Consequently, during application of the method, the pH of the samples was adjusted to 2–5 before proceeding with cation exchange separation and 0.01% (v/v) HNO3 was used to elute Cr(VI) from the column. 4.3. Calculation of porewater Cr(VI) concentration Rock matrix porosity and rock particle density measurements were conducted with a gravimetric method (Emerson, 1990) on 38 of the samples that were extracted and analyzed for Cr(VI). The average matrix porosity (ϕm) was 0.10 with a range from 0.03 to 0.16 and a standard deviation (SD) of 0.03. Dry bulk densities (ρb) were calculated using the porosity values according to Eq. (2): ρb ¼ ð1−ϕm Þ  ρp

ð2Þ

where ρp is the rock particle density. The average dry bulk density (±SD) was 2.46 ± 0.10 g/cm3. The concentrations of Cr(VI) in the porewater (Fig. 4a) were calculated according to Eq. (3), which assumes the rock is fully saturated and that total porosity is equal to effective porosity:

C pw ¼

C ext  V ext  ρb Mr  ϕm

ð3Þ

4.4. Quality control The instrument detection limit (IDL) of 0.001 μg/L (N = 10) for Cr was calculated as 3 times the SD of the Cr signal in 1% HNO3. A MDL of 0.02 μg/L (N = 10) for Cr in the reagent blanks (0.01 mol/L NaOH) was calculated as 3 times the SD of the reagent blanks. Similarly, the EDL of 0.03 μg/L was calculated as 3 times the SD for extractions conducted on silica sand blanks. To account for the background Cr concentrations in the site samples, a higher EDL of 0.1 μg/L was calculated as 3 times the SD of extraction results from bedrock samples collected below 70 m bgs, where Cr(VI) contamination was absent. The absence of contamination at this depth is supported by groundwater analyses from the MLS ports (Fig. 4b), which were conducted by CH2M HILL Applied Sciences Laboratory (ASL). Methods SW7199 (USEPA, 1996b) and EPA 218.6 (USEPA, 1991) have been used for determination of Cr(VI) by ASL (http://sccr.ch2m.com/corporate/services/asl/assets/ services/ASL_Clean_Trace.pdf). The MDL for Cr(VI) detection in groundwater was 0.02 μg/L. The results of extractions on spiked QC samples are shown in Table 1. The Cr(VI) spike recoveries in all blanks and LRS were higher than 90%, which demonstrates complete Cr(VI) extraction and recovery. For reference, USEPA method 3060A for extraction of Cr(VI) in contaminated soils defines complete recovery in the range of 80–120% for laboratory control samples and 75–125% for matrix spike samples (USEPA, 1996a). All recoveries of spiked Cr(III) were lower than 5%, which indicates the Cr(III) interference was negligible. Results of three replicate analyses of the LRS yielded an analytical precision of 13% (294 ± 39 μg/L). Triplicate analyses of eight core samples provide an average RSD of 21%, with a range of 2% to 47%. For seven of the eight samples, recoveries of Cr(VI) spikes were N90% (Table 2). 4.5. Porewater Cr (VI) concentrations

where Cpw is the estimated pore water Cr(VI) concentration, Cext is the Cr(VI) concentration in the extract solution, Vext is the final extract solution volume (25 mL), ρb is the dry bulk density, M.r is the rock sample mass and ϕm is the matrix porosity.

Fig. 4a shows the vertical distribution of porewater Cr(VI) concentrations for 147 bedrock samples. Values of ϕm and ρb were determined for 38 of the 147 samples. For these samples (Δ symbols in Fig. 4a) the

Fig. 4. Graphs showing (a) porewater Cr(VI) from rock core extractions (coring conducted April 26–May 1, 2012), (b) bar graph of groundwater Cr(VI) concentrations from groundwater sampling (MLS), (c) hydraulic head (3 temporal snapshots) and (d) transmissivity derived from FLUTe liner transmissivity testing (Keller et al., 2013). The unit “m bgs” is meters below the ground surface. The horizontal axis of (a) and (b) is on a logarithmic scale. The triangle symbols Δ represent concentrations calculated (Eq. (3)) from measured dry bulk density and porosity values. For these, the error bars represent the relative standard deviation (±21%) calculated from triplicated samples. The diamond symbols ♦ represent concentrations calculated with the averages of measured dry bulk density and porosity values. In this case the error bars represent the maximum and minimum Cpw calculated (Eqs. (4) and (5)) using the average ± 1SD for the dry bulk density and porosity. The horizontal solid black bars in (b) represent the depth intervals of the sampling ports in the MLS and the respective Cr(VI) concentrations in groundwater collected on August 2, 2012; whereas the dash bars in (b) represent the Cr(VI) concentration in groundwater collected on September 10, 2012 (analyzed by ASL). NGVD in (c) is National Geodetic Vertical Datum.

J. Zhao et al. / Chemical Geology 419 (2015) 142–148 Table 1 Recoveries of Cr(VI) and Cr(III) in spiked QC samples (crushed rock sample mass = 1.5 g). QC sample

Spiked Cr(VI) recovery (%)

Spiked Cr(III) recovery (%)

Reagent blank Silica sand LRS

94 ± 8 (N = 9) 94 ± 3 (N = 6) 92 ± 6 (N = 5)

4 ± 2 (N = 6) 4 ± 3 (N = 6) 4 ± 2 (N = 3)

measured values were used to estimate Cpw and the error bars represent the average RSD of ± 21% calculated from the triplicate samples (Table 2). For samples without measured ϕm and ρb values (♦ symbols in Fig. 4a), Cpw was determined using the average ϕm (0.10) and ρb (2.46 g/cm3) values. In this case the error bars represent the maximum and minimum Cpw calculated using ±1SD of ϕm and ρb, as Eqs. (4) and (5): max C pw ¼

C ext  V ext  ρb þ SDρb
Mr  ϕm −SDϕm

min C pw ¼

C ext  V ext  ρb þ SDρb
M r  ϕm −SDϕm


ð4Þ


ð5Þ

where max Cpw is maximum Cpw, min Cpw is minimum Cpw, SDρb is the standard deviation of ρb and SDϕm is the standard deviation of ϕm. This approach results in error bars ranging from − 28% to + 47% relative to the average Cpw; a range that reflects the potential variability of concentrations and is comparable to the RSD of the triplicate samples (±21%). The detection limit for Cr(VI) in porewater (PWDL) depends on both the EDL and the sample properties (mass of sample extracted, rock matrix density and porosity). The PWDL can be estimated from analyses of the samples collected from the deep zone, where groundwater is unaffected by Cr(VI) contamination. The average and SD for porewater Cr(VI) concentration in samples from these zones is 26 ± 15 μg/L (N = 21), and the effective PWDL is 45 μg/L (three times the SD). This elevated PWDL is a consequence of the low porewater volume and the resulting high dilution factor (Vext ∙ ρb / Mr ∙ ϕm) in the calculation of porewater concentrations. Fig. 4 also provides Cr(VI) concentrations in groundwater samples from the MLS ports (Fig. 4b), measurements of hydraulic head versus depth at three points in time (Fig. 4c) and estimates of transmissivity versus depth (Fig. 4d). Both the porewater and the groundwater data define a trend of decreasing Cr(VI) concentrations with depth, and groundwater concentrations are below the MDL (b0.02 μg/L) in the bottom two ports. This suggests that the contaminated plume is contained in the shallow (b 50 m bgs) portion of the aquifer and is consistent with elevated transmissivity values above 50 m depth, and with the upward hydraulic gradients that are observed across the section (with some variability in August 2012 shortly after installation of the MLS, but consistent thereafter). There is not always good correspondence between the porewater and groundwater data, for example, at the depth interval of MLS port

Table 2 Reproducibility of measurements and recoveries of Cr(VI) in rock samples. (Sample mass: 1.5 g. Standard deviations were calculated from triplicate measurements). Sample#

Porewater Cr(VI) concentration (μg/L)

RSD (%)

Spiked Cr(VI) recovery (%)

5 60 80 86 89 140 212 377

297 ± 26 20 ± 4 31 ± 15 2310 ± 733 374 ± 77 20 ± 2 38 ± 11 43 ± 1

9 20 47 33 20 11 29 2

94 ± 22 102 ± 16 98 ± 4 91 ± 28 97 ± 6 106 ± 19 95 ± 0.1 89 ± 2

147

1 porewater samples contain N 1000 μg/L Cr while the groundwater concentrations in MLS port 1 are b 200 μg/L Cr. This illustrates the value of the DFN approach to characterize contaminant distributions in fractured porous bedrock. Concentration data from borehole devices such as the FLUTe MLS ports represent depth-averaged values that are weighted toward concentrations in the mobile water from fractures while the DFN approach provides porewater data from discrete intervals (0.03–0.06 m). The discrete-depth porewater data reveal the heterogeneous nature of the Cr(VI) concentrations within the low permeability rock matrix, including maxima that likely reflect the local proximity to fractures (e.g., 6400 μg/L at 22.5 m bgs and 3300 μg/L at 39.8 m bgs), and even within the zones of the highest Cr(VI) concentrations some samples are below the PWDL. A FLUTe liner was deployed temporarily to minimize potential for open-hole mixing between geophysical logging and hydraulic measurement activities, and before installation of the MLS system. Despite that effort, comparison of the porewater and groundwater data suggests that mixing may have occurred prior to installation of the MLS. There are elevated Cr(VI) concentrations in groundwater samples from ports 3 to 5, while porewater Cr(VI) concentrations in the same interval are near the PWDL. This could reflect open-hole mixing and downward migration of Cr(VI)-contaminated groundwater. Downward migration in the open borehole is supported by the observation that Cr(VI) concentrations in MLS ports 3 to 5 decreased after MLS installation by 4 to 77% between August 2, 2012 (solid horizontal bar in Fig. 4b) to September 10, 2012 (dashed horizontal bar in Fig. 4b). 5. Implications The method developed in this work provides a reliable procedure to analyze Cr(VI) in the porewater of bedrock samples with detection limits in the range of drinking water standards (about 50 μg/L). This method will be useful to support efforts to evaluate Cr(VI) contaminated sites for their natural attenuation potential (Kent et al., 2007). Compared to the measurements of Cr(VI) in groundwater, the porewater data reveal the complex and heterogeneous nature of contamination, which results from fracture-controlled flow and diffusive mass transfer between the mobile water in fractures and the relatively immobile porewater in the rock matrix. Although porewater is less mobile, solute readily moves in and out by diffusion. Accumulation of Cr(VI) in the rock matrix retards growth of the plume and attenuates concentrations in fractures, but at a later stage when the plume source is removed or diminished, the mass in the matrix provides a persistent source of Cr(VI) to the fractures via back diffusion to sustain the plume. Initial recovery efforts by pumping at this site removed approximately 30% of the total mass released and it is likely that the plume is now being sustained largely by back diffusion from the rock matrix. Thus efforts focused on risk assessment and groundwater restoration should consider the current mass distribution, recognizing that the majority of the mass is now in the rock matrix, where transport is diffusion controlled. A key issue at this site, which requires further study, is whether reduction from Cr(VI) to Cr(III) is occurring in the rock matrix, and if so, what effect it will have on the longevity of Cr(VI) releases to groundwater by back-diffusion. Acknowledgments Funding for site activities and support was provided by the U.S. EPA and Army Corps of Engineers. CH2M Hill was the primary site consultant and provided field and logistical support for the core subsampling activities and also provided the groundwater data from the MLS and other site data. Supplemental funding for development of the laboratory methods was provided by the University Consortium for Field-Focused Groundwater Contamination Research (http://g360.uoguelph.ca/ consortium).

148

J. Zhao et al. / Chemical Geology 419 (2015) 142–148

References Anderson, R.A., 1997. Chromium as an essential nutrient for humans. Regul. Toxicol. Pharmacol. 26, S35–S41. ATSDR, 2013. Summary Data for 2013 Priority List of Hazardous Substances. Agency for Toxic Substances and Disease Registry, Division of Toxicology and Environmental Medicine, Atlanta, GA. Ball, J.W., McCleskey, R.B., 2003. A new cation-exchange method for accurate field speciation of hexavalent chromium. Talanta 61, 305–313. Barceloux, D.G., 1999. Chromium. J. Toxicol. Clin. Toxicol. 37, 173–194. Cherry, J.A., Parker, B.L., Keller, C., 2007. A new depth-discrete multilevel monitoring approach for fractured rock. Ground Water Monit. Remediat. 27, 57–70. Emerson, D.W., 1990. Notes on mass properties of rocks? Density, porosity, permeability. Explor. Geophys. 21, 209–216. Fantoni, D., Brozzo, G., Canepa, M., Cipolli, F., Marini, L., Ottonello, G., Vetuschi Zuccolini, M., 2002. Natural hexavalent chromium in groundwaters interacting with ophiolitic rocks. Environ. Geol. 42, 871–882. Frenzel, W., 1998. Highly sensitive semi-quantitative field test for the determination of chromium (VI) in aqueous samples. Fresenius J. Anal. Chem. 361, 774–779. Friedly, J.C., Davis, J.A., Kent, D.B., 1995. Modeling hexavalent chromium reduction in groundwater in field-scale transport and laboratory batch experiments. Water Resour. Res. 31, 2783–2794. Gray, D.J., 2003. Naturally occurring Cr6+ in shallow groundwaters of the Yilgarn Craton, Western Australia. Geochem. Explor. Environ. Anal. 3, 359–368. Health Canada, 2012. Health Canada (2012) Guidelines for Canadian Drinking Water Quality—Summary Table. Water, Air and Climate Change Bureau, Healthy Environments and Consumer Safety Branch, Health, Canada, Ottawa, Ontario. Herman, G.C., 2001. Hydrogeological framework of bedrock aquifers in the Newark Basin, New Jersey, geology in service to public health. Eighteenth Annual Meeting of the Geological Association of New Jersey, South Brunswick, New Jersey, pp. 6–45. James, B.R., Petura, J.C., Vitale, R.J., Mussoline, G.R., 1995. Hexavalent chromium extraction from soils: a comparison of five methods. Environ. Sci. Technol. 2377-2381. Johnson, C.A., 1990. Rapid ion-exchange technique for the separation and preconcentration of chromium(VI) and chromium(III) in fresh waters. Anal. Chim. Acta 238, 273–278. Keller, C.E., Cherry, J.A., Parker, B.L., 2013. New method for continuous transmissivity profiling in fractured rock. Groundwater 52, 352–367. Kent, D.B., Puls, R.W., Ford, R.G., 2007. Monitored Natural Attenuation of Inorganic Contaminants in Ground Water Volume 2 Assessment for Non-Radionuclides Including Arsenic, Cadmium, Chromium, Copper, Lead, Nickel, Nitrate, Perchlorate,

and Selenium. National Risk Management Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, Cincinnati, Ohio. Mandiwana, K.L., 2008. Rapid leaching of Cr(VI) in soil with Na3PO4 in the determination of hexavalent chromium by electrothermal atomic absorption spectrometry. Talanta 74, 736–740. Narukawa, T., Riley, K.W., French, D.H., Chiba, K., 2007. Speciation of chromium in Australian fly ash. Talanta 73, 178–184. Nriagu, J.O.N.E., 1988. Chromium in the Natural and Human Environments. Wiley, New York. Oze, C., Bird, D.K., Fendorf, S., 2007. Genesis of hexavalent chromium from natural sources in soil and groundwater. Proc. Natl. Acad. Sci. U. S. A. 104, 6544–6549. Palmer, C.D., Wittbrodt, P.R., 1991. Processes affecting the remediation of chromiumcontaminated sites. Environ. Health Perspect. 92, 25–40. Parker, B.L., Chapman, S.W., Cherry, J.A., 2010. Plume persistence in fractured sedimentary rock after source zone removal. Groundwater 48, 799–803. Parker, B.L., Cherry, J.A., Chapman, S.W., 2012. Discrete fracture network approach for studying contamination in fractured rock. AQUA mundi 101–116. Rai, D., Sass, B.M., Moore, D.A., 1987. Chromium (III) hydrolysis constants and solubility of chromium(III) hydroxide. Inorg. Chem. 26, 345–349. Rakhunde, R., Deshpande, L., Juneja, H.D., 2012. Chemical speciation of chromium in water: a review. Crit. Rev. Environ. Sci. Technol. 42, 776–810. Richard, F.O.C., Bourg, A.C.M., 1991. Aqueous geochemistry of chromium: a review. Water Res. 25, 807–816. Unceta, N., Séby, F., Malherbe, J., Donard, O.F.X., 2010. Chromium speciation in solid matrices and regulation: a review. Anal. Bioanal. Chem. 397, 1097–1111. USEPA, 1991. Method 218.6: Determination of Dissolved Hexavalent Chromium in Drinking Water, Groundwater and Industrial Wastewater Effluents by Ion Chromatography. U.S. Environmental Protection Agency. USEPA, 1992. Method 7196A: Chromium, Hexavalent (Colorimetric). U.S. Environmental Protection Agency. USEPA, 1996a. Method 3060A: Alkaline Digestion for Hexavalent Chromium. U.S. Environmental Protection Agency. USEPA, 1996b. Method 7199: Determination of Hexavalent Chromium in Drinking Water, Groundwater. U.S. Environmental Protection Agency. USEPA, 2012. 2012 Edition of the Drinking Water Standards and Health Advisories. Office of Water, U.S. Environmental Protection Agency, Washington, DC.