Chromium Isotopes and the Fate of
Hexavalent Chromium in the Environment

Science v.295, n.5562, 15mar02

Andre S. Ellis,1 Thomas M. Johnson,1* Thomas D. Bullen2

Measurements of chromium (Cr) stable-isotope fractionation in laboratory experiments and natural waters show that lighter isotopes reacted preferentially during Cr(VI) reduction by magnetite and sediments. The 53Cr/52Cr ratio of the product was 3.4 ± 0.1 per mil less than that of the reactant. 53Cr/52Cr shifts in water samples indicate the extent of reduction, a critical process that renders toxic Cr(VI) in the environment immobile and less toxic.

1 Department of Geology, University of Illinois at Urbana-Champaign, 245 Natural History Building, Urbana, IL 61801, USA.

2 Water Resources Division, U.S. Geological Survey, 345 Middlefield Road, Menlo Park, CA 94025, USA.

* To whom correspondence should be addressed. E-mail: tmjohnsn@uiuc.edu

Chromium is a common contaminant in surface water and groundwater (1, 2) because it is used widely in electroplating and other industries and occurs naturally at high concentration in ultramafic rocks. Under oxidizing conditions, Cr is highly soluble and mobile as the Cr(VI) anions chromate (CrO42-) and bichromate (HCrO4-). Cr(VI) is a suspected carcinogen (3). Under reducing conditions, Cr(VI) may convert to Cr(III), which is insoluble, strongly adsorbed onto solid surfaces (4), and less toxic. Cr(VI) can be removed from solution artificially by in situ reduction (5, 6), or naturally by reductants such as aqueous Fe(II), dissolved humic acids, and Fe(II)-bearing minerals (7, 8). For these reasons, knowledge of reduction rates is essential in many Cr(VI) contamination cases. For example, sufficient rates of natural reduction at a given site allow use of the "monitored natural attenuation" approach, which is much less expensive and disruptive than active remediation. Determining reduction rates can be difficult; multiple rounds of groundwater sampling and analysis over several years has been required in some contaminant plumes. Reduction reactions tend to enrich products in the lighter isotopes because they preferentially react (9), and the residual reactants become progressively enriched in the heavier isotopes as reduction proceeds (10-12). Here we show that Cr stable isotopes can be used to estimate the extent of reduction, which in turn can be used to estimate long-term reduction rates.

Cr has four stable isotopes of masses 50 (4.35%), 52 (83.8%), 53 (9.50%), and 54 (2.37%) (13). We developed a double-isotope spike method for measuring mass-dependent fractionation of Cr isotopes (i.e., variations in the relative abundances of light and heavy isotopes) (14). We measured Cr isotope fractionation during reduction of Cr(VI) by slurries of magnetite and two sediment samples (15). Because magnetite is a likely reducing agent in some aquifer sediments (16), the magnetite experiment provides a simple analog for a natural aquifer. The sediment slurries were microbially rich and chemically complex, and were collected from an intertidal mud flat in the northern reach of the San Francisco Estuary and from a pond in Urbana, Illinois. The water in the estuary is about 50% seawater, and the sediment was silt with a total organic carbon (TOC) content of 0.92%. The Urbana pond is freshwater runoff, and its sediment was clayey silt with a TOC of 0.29%.

In the magnetite suspensions, a sharp initial decrease in Cr(VI) was followed by a slower, linear decrease (Fig. 1). In the sediment slurry experiments, Cr(VI) concentrations followed roughly a first-order trend (Fig. 1). As expected at pH 6 to 7, sorption of Cr(VI) was negligible. We confirmed the lack of reversible sorption by adding competing anions like sulfate and phosphate to desorb any adsorbed Cr(VI); no effects were observed. As reduction proceeded and the Cr(VI) concentrations decreased, 53Cr values of the remaining unreduced Cr(VI) increased, indicating preferential reduction of the lighter isotopes (Fig. 2). To calculate the size of this kinetic isotope effect, we used a Rayleigh fractionation model. The instantaneous isotope fractionation factor, , is defined by (1) where Rprod and Rreact are the 53Cr/52Cr ratios of the Cr reduced at an instant in time and in the reactant pool, respectively. The values were calculated using the relation (2) where 53Cr and 53Crini (14) refer to the unreacted Cr(VI) pool at the time of sampling and at the start of the experiment, respectively, and f is the fraction of Cr(VI) remaining. In all cases, models fit the data within the uncertainties (Fig. 2 and Table 1). Autoclaved control duplicates of the two sediment experiments reduced Cr(VI) as quickly as the unautoclaved experiments did, hence the reduction was abiotic.

Fig. 1. Cr(VI) concentration versus time in the experiments. SFE, San Francisco Estuary.

Fig. 2. 53Cr of unreduced dissolved Cr(VI) versus f, the fraction of Cr(VI) remaining. Dark line, = 0.9965 best-fit curve for Urbana pond and magnetite experiments. Dotted line, = 0.9967 best-fit curve for the San Francisco Estuary experiment.

Table 1. Isotopic fractionation in Cr(VI) reduction experiments.

Reducing agent		a
Magnetite		0.9965 
Estuarine sediment	0.9967 
Pond sediment		0.9965

These experiments show that reduction of Cr(VI) results in Cr stable-isotope fractionation. The similarity of the values determined for the three contrasting experiments suggests that the reduction mechanisms are similar, despite the differences in the reducing agents. Given that the reactions were abiotic, the similarity of the isotope ratio shifts likely reflects the kinetic isotope effect accompanying the breakage of Cr-O bonds of the Cr(VI) oxyanions. If so, these results should be applicable to aquifers where abiotic reduction dominates.

We obtained groundwater samples from sites in Putnam, Connecticut, and Berkeley, California (17). Both are contaminated by Cr plating wastes. At the California site, groundwater flows through natural alluvium and artificial fill. On the basis of several years of data, previous workers concluded that Cr(VI) has been partially removed from solution, probably by reduction (18). At the Connecticut site, groundwater flows through glacial deposits that likely contain magnetite and other Fe(II)-bearing minerals.

To estimate the initial isotopic compositions of the contaminants, we measured 53Cr in samples of three plating baths in active use, two at the Connecticut site and a third from another location with a different plating process (Table 2). We also analyzed the chromic acid supply used to make up plating baths at the Connecticut site, two laboratory Cr reagents, and three basaltic rock standards (19) representing diverse geologic environments (Table 2). The rock analyses provided information about the bulk 53Cr of Earth. All of these samples yielded 53Cr values close to zero. Cr ores are of igneous origin; because little isotopic fractionation is expected at high temperatures, these ores should (and apparently do) inherit 53Cr values from Earth's mantle. We expect that Cr isotopes are not fractionated during purification (20), and thus most or all supplies of industrial Cr should have 53Cr values close to 0, like the samples measured here.

Table 2. Cr isotope ratio determinations.

			[Cr(VI)]
Sample 			(mg/liter)	d53Cr %
Reagent Cr, rocks 
K2Cr2O7 reagent 	-- 		0.35  
Cr(NO3)2- reagent 	-- 		0.32 
BIR basalt (Iceland) 	-- 		-0.04 
BHVO basalt (Kilauea) 	-- 		0.05 
JB basalt (Japan) 	-- 		-0.04 
Plating bath 		~105 		0.37 
Connecticut site  
Plating bath 1 		~105 		0.36 
Plating bath 2 		~105 		0.29 
CrO3 supply 		-- 		-0.07 
MW-8 groundwater 	4.5 		2.23 
MW-9 groundwater 	8.61 		1.28 
MW-11s groundwater  	16.1 		1.93 
MW-11 groundwater 	0.63 		3.62 
MW-12 groundwater 	1.63 		3.96 
California site 
MW-3 groundwater 	0.98 		1.08 
MW-10 groundwater 	0.13 		5.79 

The plating bath samples also had 53Cr values close to 0, with a mean of 0.34. This is somewhat surprising, because Cr(VI) reduction is essential to the plating process and large amounts of Cr are removed from the plating baths by this reaction. Roughly one-tenth of the Cr(VI) is consumed in a typical week, after which the lost Cr(VI) is replenished. This cycle has been repeated for more than 5 years in the two Connecticut baths. A simple model of this process reveals that an isotopic fractionation factor of 0.9997 for Cr(VI) removed by the plating reaction would result in a steady-state value of +0.3 for the plating bath relative to the Cr(VI) supply after about 20 cycles.

In light of these observations, we suggest that Cr released as plating waste generally has an initial 53Cr value slightly greater than zero. If so, detection of Cr(VI) reduction in groundwater systems would be relatively simple, as the initial 53Cr value would be known and groundwater values greater than that would directly indicate the extent of reduction. If, on the other hand, plating wastes have variable 53Cr values, then it may be possible to distinguish different contamination sources via their 53Cr values.

Our measured groundwater 53Cr values (Table 2) ranged from 1.1 to 5.8 in the samples from the California site, and from 1.3 to 4.0 in the samples from the Connecticut site. All of the groundwater Cr(VI) analyses show enrichment in the heavy isotope relative to the plating baths. Apparently, Cr(VI) reduction has preferentially removed lighter isotopes from the groundwater. The variation in 53Cr values at each site suggests that reduction of Cr(VI) is occurring and has progressed to different degrees in different parts of the contaminant plumes. The highest 53Cr values are found in the samples with lowest Cr(VI) concentration at both sites. This result is expected because the fringe areas of the contaminant plumes likely have greater degrees of reduction than the plume cores, where Cr concentrations are high and the reducing power of the aquifer materials has been depleted.

Stable Cr isotope ratios can thus serve as indicators of the extent of Cr(VI) reduction in groundwater. For example, if we assume an value of 0.9966 and a value of 0.34 (the mean of the plating bath analyses) for the initial 53Cr of the contaminant Cr, we can use Eq. 1 to determine the extent of reduction for two of the Connecticut groundwater samples, MW-9 and MW-12; for these samples, 31% and 68% of the Cr(VI) initially present in the wells was reduced, respectively.

Cr(VI) reduction by bacteria or reducing agents other than those studied here could induce greater or lesser isotopic fractionation than we observed. Processes other than reduction, such as sorption, precipitation, and uptake by plants and algae, can remove Cr from solution (1, 21-23). If these processes and/or Cr(III) oxidation induce isotopic fractionation, this could complicate the interpretation of 53Cr measurements. However, as with S and Se isotopes (12), we expect that the dominant cause of Cr isotope fractionation is oxyanion reduction. Cr isotope studies may also be useful in assessing redox conditions in modern or ancient oceans.

REFERENCES AND NOTES

1. R. J. Bartlett, B. R. James, in Chromium in the Natural and Human Environments, J. O. Nriagu, E. Niebor, Eds. (Wiley, New York, 1988), pp. 267-304.

2. J. Robles-Camacho and M. A. Armienta, J. Geochem. Explor. 68, 167 (2000) [ISI].

3. A. Kortenkamp, et al., Arch. Biochem. Biophys. 329, 199 (1996) [CrossRef][ISI][Medline] .

4. S. Loyaux-Lawniczak, P. Lecomte, J.-J. Ehrhardt, Environ. Sci. Technol. 35, 1350 (2001) [CrossRef][ISI][Medline].

5. D. W. Blowes, C. J. Ptacek, J. L. Jambor, Environ. Sci. Technol. 31, 3348 (1997) [CrossRef][ISI].

6. C. M. Lytle, et al., Environ. Sci. Technol. 32, 3087 (1998) [CrossRef][ISI].

7. I. J. Buerge and S. J. Hug, Environ. Sci. Technol. 32, 2092 (1998) [CrossRef][ISI].

8. P. R. Wittbrodt and C. D. Palmer, Environ. Sci. Technol. 30, 2470 (1996) [CrossRef][ISI].

9. J. Hoefs, Stable Isotope Geochemistry (Springer-Verlag, New York, ed. 3, 1987).

10. J. Boettcher, O. Strebel, S. Voerkelius, H. L. Schmidt, J. Hydrol. 114, 413 (1990) [ISI].

11. H. G. Thode, J. Monster, in Fluids in Subsurface Environments--a Symposium (American Association of Petroleum Geologists, Tulsa, OK, 1965), pp. 367-377.

12. T. M. Johnson, M. J. Herbel, T. D. Bullen, P. T. Zawislanski, Geochim. Cosmochim. Acta 63, 2775 (1999) [ISI].

13. M. Rotaru, J. L. Birck, C. J. Allegre, Nature 358, 465 (1992) [ISI] .

14. This method is fundamentally different from that used to measure radiogenic 53Cr variations that occur in ancient meteorites. Radiogenic variations in 53Cr/52Cr occur in some meteorites as a result of 53Mn decay during the early history of the solar system, but essentially all Earth materials have identical radiogenic contents (24). Specifically, we determine variations in the 53Cr/52Cr ratio (15) and express results as per mil deviations from a standard: 53Cr () = [(53Cr/52Cr)sam - (53Cr/52Cr)std]/(53Cr/52Cr)std × 1000, where sam and std refer to sample and National Institute of Standards and Technology Standard Reference Material 979, respectively. External measurement precision is ±0.2.

15. Supplementary details of experimental procedures and mass spectrometry are available on Science Online at www.sciencemag.org/cgi/content/full/295/5562/2060/DC1.

16. L. D. Anderson, D. B. Kent, J. A. Davis, Environ. Sci. Technol. 28, 178 (1994) [ISI].

17. Groundwater samples were taken according to EPA protocols (25).

18. R. Makdisi, in Second International Conference on Remediation of Chlorinated and Recalcitrant Compounds, Monterey, CA, G. B. Wickramanayake, A. R. Gavaskar, M. E. Kelley, Eds. (KelleyBattelle, Columbus, Ohio, 2000), pp. 185-192.

19. K. Govindaraju, special issue, Geostand. Newsl. 13 (1989).

20. Roasting of chromite ore to recover Cr should not significantly fractionate Cr isotopes because fractionation is minimal at high temperatures. Subsequent purification steps at lower temperatures may have the potential to fractionate isotopes, but unless there are large losses of Cr, little overall fractionation results (by mass balance).

21. B. R. James and R. J. Bartlett, J. Environ. Qual. 13, 67 (1984) [ISI].

22. E. E. Cary, Ed., Chromium in Air, Soil, and Natural Waters (Elsevier Biomedical, Amsterdam, 1982).

23. B. R. James and R. J. Bartlett, J. Environ. Qual. 12, 177 (1983) [ISI].

24. A. Shukolyukov and G. W. Lugmair, Science 282, 927 (1998) [Abstract/Full Text] .

25. SW-846 Test Methods for Evaluating Solid Waste, Physical/Chemical Methods (U.S. Environmental Protection Agency, Washington, DC, 1992).

26. Supported by NSF grant EAR 00-01153. Initial work was supported by Geological Society of America graduate student grant 6421-99. We thank N. Nikolaidis and L. Hellerich (University of Connecticut), J. Miller (National Chromium), and R. Makdisi (Stellar Environmental Solutions) for groundwater and plating bath samples, and D. Schrag, C. Bethke, and C. Lundstrom for comments. 26 November 2001; accepted 28 January 2002 10.1126/science.1068368