Environ. Sci. Technol. 2010, 44, 1043–1048
Cr Stable Isotopes As Indicators of Cr(VI) Reduction in Groundwater: A Detailed Time-Series Study of a Point-Source Plume EMILY C. BERNA,† T H O M A S M . J O H N S O N , * ,† RICHARD S. MAKDISI,‡ AND ANIRBAN BASU† Department of Geology, 253 Natural History Building, University of Illinois, Urbana, Illinois 61801, and Stellar Environmental Solutions, Inc., 2198 Sixth Street, Suite 201, Berkeley, California
Received July 28, 2009. Revised manuscript received December 7, 2009. Accepted December 11, 2009.
Chromium stable isotope ratios show promise as indicators of Cr(VI) reduction in groundwater, but no published study has yet demonstrated that expected relationships between 53Cr/ 52 Cr and Cr(VI) concentration, position, and time occur in an actual groundwater plume. We present an extensive data set from a point-source plume in Berkeley, CA; data extend over 5 years and 14 locations covering the entire plume. We interpret the data using a Rayleigh distillation model with an effective fractionation factor that incorporates an intrinsic fractionation factor determined from incubations of site sediments and accounts for reservoir effects in the restricted subsurface zones where Cr(VI) reduction is thought to occur. The groundwater 53Cr/52Cr and Cr(VI) concentration data are consistent with a scenario where the system has reached a steady state: Cr(VI) reduction continues, the extent of reduction at any point is constant over time, reduction proceeds to completion at the downgradient edge of the plume, and the plume is no longer advancing. The overall consistency of the results with a reasonable model for the site supports the use of Cr isotope-based estimates of reduction, but we discuss current uncertainties and limitations of the approach as well.
Introduction Hexavalent chromium [Cr(VI)] is a contaminant of concern in many groundwater systems. Cr(VI) contamination can occur naturally (1), but most Cr(VI) contamination in water is anthropogenic, resulting from releases from electroplating, leather tanning, lumber treatment, and other industries (2). In groundwater systems, Cr(VI) is potentially toxic, highly soluble, and mobile, existing as HCrO4-, CrO42-, and Cr2O72oxoanions (3). Cr(III) is much less toxic and is immobile, as it is insoluble and adsorbs strongly onto solid surfaces. Reduction of Cr(VI) to Cr(III) is a key process in remediation schemes (4). Reduction can occur naturally through reactions involving Fe(II)-bearing minerals, microbial action, and/or other mechanisms (3), and it can be induced * Corresponding author e-mail:
[email protected]; phone: 217-244-2002; fax: 217-244-4996. † University of Illinois. ‡ Stellar Environmental Solutions, Inc. 10.1021/es902280s
2010 American Chemical Society
Published on Web 12/29/2009
artificially by abiotic reductants (5, 6) or biostimulation (7). Quantifying Cr(VI) reduction rates can be difficult and expensive. Usually, this is done by measuring Cr(VI) concentrations at an array of points over an extended period of time and determining loss through reduction by mass balance. This approach suffers from the effects of dilution, advection, and adsorption. Dilution through the influx of uncontaminated water at the water table, advection of a highconcentration groundwater mass away from a well, and adsorption can all cause decreases in Cr(VI) concentration without decreasing the contaminant mass. Measurements of Cr stable isotope abundances can be used to indicate Cr(VI) reduction and perhaps to quantify it. Cr has four stable isotopes, 50Cr, 52Cr, 53Cr, and 54Cr, with abundances of 4.35%, 83.8%, 9.5%, and 2.37%, respectively. Isotopic fractionation occurs during reduction of Cr(VI) (8, 9) as Cr-O bonds are rearranged: The reaction product is enriched in lighter isotopes relative to the reactant, and as reduction progresses, the reactant becomes steadily enriched in heavier isotopes. This shift in the proportions of heavy and light isotopes is quantified by measuring the 53Cr/52Cr abundance ratio. The 53Cr/52Cr ratio of dissolved Cr(VI) increases progressively as the Cr(VI) is partially removed by reduction, and thus, the extent of reduction can be estimated from 53Cr/52Cr data. Cr isotope fractionation occurs during both abiotic and microbial reduction, and its magnitude varies according to reaction mechanism and rate (8-10). Fractionation induced by adsorption of Cr(VI) (11) or precipitation of chromate salts (12) is very weak, as these processes do not alter the Cr bonding greatly. Two studies have reported Cr isotope data from groundwater systems. The first (8) presented eight reconnaissance analyses to show that elevated 53Cr/52Cr ratios occur in Cr(VI) of groundwater systems in which Cr(VI) reduction is thought to occur. The second (1) presented a more extensive data set for a Mojave desert groundwater system containing Cr(VI) derived from industrial and natural sources. Cr isotope data were used to distinguish Cr(VI) from the two sources and to provide evidence for Cr(VI) reduction in an anoxic zone. However, because of the complexity of the system and the wide 53Cr/52Cr range of the natural Cr(VI), only a tentative correlation between 53Cr/52Cr shifts and Cr(VI) reduction was obtained. To date, no published study has shown straightforward relationships between 53Cr/52Cr ratio and position, time, and Cr(VI) concentration consistent with the hypothesis that 53Cr/52Cr indicates Cr(VI) reduction in a real groundwater system. Here, we present a detailed 53Cr/52Cr data set for dissolved Cr(VI) in a groundwater plume undergoing natural reduction. The data provide good spatial coverage over a period of several years. The system is relatively simple, with a single point source of Cr(VI). We interpret the 53Cr/52Cr data using a modified Rayleigh distillation model for Cr(VI) in groundwater as it flows and interacts with restricted reducing zones in the aquifer. Using this model, we construct a scenario that is consistent with the observed spatial and temporal patterns in the 53Cr/52Cr and concentration data. We discuss ways to apply the Cr isotope method, provide insight into the interpretation of time-series data, and discuss uncertainty of the results and important limitations in current knowledge. We also present results of experiments done to obtain estimated fractionation factors for reduction by sediments from this site. These are the first such results for a contaminated aquifer, and they are used as a preliminary site-specific calibration for the Rayleigh calculations. VOL. 44, NO. 3, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1043
FIGURE 1. Site map. Numbers give locations of permanent monitoring wells; 2006 direct-push samples are indicated “HP”. Rectangle around MW-18 gives approximate location of injected biostimulant zone. Contour lines give approximate Cr(VI) concentrations in mg/L.
Site Description, Sampling, and Methods The site is located in an industrial area in northwestern Berkeley, California, on sandy alluvial plain deposits near San Francisco Bay. In the 1970s, Cr(VI) contamination resulted from Cr plating operations; contaminated soil was removed from the source area in the late 1990s (13). The plume is contained in an unconfined, sandy aquifer with westward flow. Depth to water ranges from 0.5 to 3 m below the ground surface; the bottom of the water-bearing unit is greater than 8 m below ground surface. Sediment cores exhibit moderate heterogeneity, including lenses of clayrich sediment. In 2008, 15 permanent monitoring wells existed; additional samples were obtained in 2006 through temporary sampling points created by direct-push coring. Locations of sampling points are given in Figure 1. In situ remediation has been conducted at the site, with injection of biostimulants in 2004 and 2006. The resulting zones of Cr(VI) reduction apparently did not affect any of the sampling points used in the present study, except for monitoring well MW-18, which has had no detectable Cr(VI) since shortly after the first injection, and MW-19, which had a decrease in concentration after 2004 (see below). Sampling. Sediments used in the slurry incubation experiments were obtained using direct-push coring methods described in the Supporting Information. Three different sediment types were observed: a sandy/gravelly type, a reddish brown clay-rich type, and a green clay-rich type. The sandy/gravelly material is poorly sorted; grain size ranged from clay size to >25-mm gravel with the mode being medium sand (1 mm). Both clay-rich sediments are dominated by clay-size particles interspersed with
