Environ. Sci. Technol. 2006, 40, 6943-6948
Experimentally Determined Uranium Isotope Fractionation During Reduction of Hexavalent U by Bacteria and Zero Valent Iron L A U R A K . R A D E M A C H E R , * ,†,‡ CRAIG C. LUNDSTROM,‡ THOMAS M. JOHNSON,‡ ROBERT A. SANFORD,‡ JUANZHO ZHAO,‡ AND ZHAOFENG ZHANG‡ Department of Geosciences, University of the Pacific; Department of Geology, University of Illinois
Variations in stable isotope ratios of redox sensitive elements are often used to understand redox processes occurring near the Earth’s surface. Presented here are measurements of mass-dependent U isotope fractionation induced by U(VI) reduction by zerovalent iron (Fe0) and bacteria under controlled pH and HCO3- conditions. In abiotic experiments, Fe0 reduced U(VI), but the reaction failed to induce an analytically significant isotopic fractionation. Bacterial reduction experiments using Geobacter sulfurreducens and Anaeromyxobacter dehalogenans reduced dissolved U(VI) and caused enrichment of 238U relative to 235U in the remaining U(VI). Enrichment factors (ε) calculated using a Rayleigh distillation model are -0.31‰ and -0.34‰ for G. sulfurreducens and A. dehalogenans, respectively, under identical experimental conditions. Further studies are required to determine the range of possible values for 238U/235U fractionation factors under a variety of experimental conditions before broad application of these results is possible. However, the measurable variations in δ238U show promise as indicators of reduction for future studies of groundwater contamination, geochronology, U ore deposit formation, and U biogeochemical cycling.
Introduction Uranium contamination is currently a topic of concern due to the legacy of nuclear weapons and fuel production (e.g., ref 1). Uranium exists in groundwater systems in two stable valence states. The more oxidized valence, U(VI), forms the UO22+ uranyl ion, which, in turn, forms complexed species such as UO2(CO3)22-. These complexes are typically thought to be unsusceptible to adsorption, therefore, mobile and bioavailable in the environment (2). However, recent studies indicate these complexes may adsorb to bacteria cell walls (3). The more reduced valence, U(IV), is highly insoluble, readily precipitates out of solution as UO2, and is immobile and non-bioavailable. Accordingly, reduction of U(VI) to U(IV) is a critically important reaction and is the focus of numerous U bioremediation efforts (i.e., refs 4-7). * Corresponding author phone: (209) 946-7351; fax: (209) 0462362; e-mail:
[email protected]. † University of the Pacific. ‡ University of Illinois. 10.1021/es0604360 CCC: $33.50 Published on Web 10/20/2006
2006 American Chemical Society
Mass-dependent isotope fractionations are useful for detecting and potentially quantifying reduction reactions. In general, bonds involving lighter isotopes are easier to break than bonds with heavier isotopes. Thus, bonds involving lighter isotopes react more rapidly and are concentrated in solid products (8). These differing reaction rates result in kinetic isotope fractionation between reactants and products if back-reaction rates are small. The remaining unreacted pool becomes progressively enriched in heavier isotopes as reduction proceeds. Thus, the degree of isotopic enrichment is a useful means for detecting and quantifying the extent of reduction. This approach has been used with stable isotope ratios of nitrate, sulfate, Se(VI), and Cr(VI) as indicators of reduction (i.e., refs 9-14). No stable U isotopes exist, but the half-lives of 238U and 235 U (4.5 × 109 and 0.7 × 109 years, respectively) are sufficiently long that mass-dependent fractionation of these isotopes is analogous to that of stable isotopes, provided the time scales of the geochemical processes addressed are sufficiently short. Few studies have examined mass-dependent fractionation of 238U and 235U. Chen and Wasserburg (15) detected no variations in the 238U/235U ratio in terrestrial samples with a precision of approximately 3‰. Cowan and Adler (16) compiled data from U ore deposits measured by gas source mass spectrometry (UF6 prepared for 235U enrichment by gas diffusion) with a precision of approximately 0.3‰ and observed a 2‰ range in 238U/235U. This observation suggested the possible existence of small variations in 238U/235U in earth materials. A recent study indicates that precise measurements of 238U/235U are possible on natural materials (23), and knowledge of natural 238U/235U variations will impact studies using U isotopes and their decay products to date and interpret the geologic record. Incidentally, 234U/238U ratios are known to vary in nature at the percent level; however, these probably reflect alpha recoil effects (17-22) and are not useful as indicators of mass-dependent isotope fractionation. The present contribution reports results from controlled laboratory experiments designed to determine the extent of 238U/235U fractionation during chemical reduction by bacteria and Fe0. We observed systematic changes in 238U/235U with progressing bacterial reduction but no isotopic fractionation during Fe0 reduction. These results are significant because variations in 238U/235U are potentially a valuable tool for studying the formation of U ore deposits, groundwater contamination, and biogeochemical cycling.
Experimental Section Starting Materials. The 238U/235U (ratio) of natural U is 137.85 ( 0.4 (24). The precision of 238U/235U measurement by mass spectrometry is limited by the weaker signal intensity produced by the less abundant 235U. Therefore, we used an enriched U standard, SRM U-500 from the National Institute of Standards and Technology (NIST) (25) as the U source for the experiments. The NBS SRM U-500 contains nearly equal quantities of 238U and 235U and has a 238U/235U of 1.0003 ( 0.001. Stock solution of U containing approximately 1000 mg l-1 U of uranyl nitrate was prepared by repeatedly dissolving and drying NBS SRM U-500 in concentrated HNO3-. The final uranyl nitrate solution was redissolved in weak HNO3-. Drying down the uranyl nitrate solution repeatedly with concentrated hydrochloric acid and redissolving it in a weak hydrochloric acid produced the uranyl chloride stock solution. Preliminary experiments suggest that the presence of nitrate in the U starting solution inhibited cell growth and/ VOL. 40, NO. 22, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6943
or reduction. Cell mass and precipitate accumulation occurred in reaction vessels with uranyl chloride but not in uranyl nitrate experiments. In addition, A. dehalogenans is known to not tolerate high concentrations (>1 mM) of nitrate (26). Therefore, subsequent bacterial experiments were performed using a UO2Cl2 solution, which does not appear to interfere with biotic reduction processes. Using a UO2Cl2 solution also minimized concerns that the presence of nitrate may facilitate reoxidation (27). Results are only reported for uranyl chloride solutions used in the biotic experiments. Uranyl nitrate was used during abiotic (Fe0) experiments. These stock solutions were stored in 250 mL serum bottles with crimp-sealed gray butyl stoppers. The stock solutions were flushed for 30 min with N2 to create an O2-free headspace and solution prior to use in each reduction experiment. Zerovalent Iron Experiments. Abiotic reduction experiments were modified after Gu et al. (28), who evaluated the effectiveness of Fe0 at reducing U(VI). Fluorescence spectroscopic measurements of the reaction products indicated that reduction, rather than sorption, was the dominant process removing U from solution in these experiments (28). Fe0 experiments were conducted in 25 mL serum bottles with 18 mL of 0.1 M NaHCO3 and 2 mL of 1300 mgL-1 U (650 mgL-1 238U + 650 mgL-1 235U) for a starting solution of 130 mgL-1. These experiments began with the addition of 2 g of iron filings (325 mesh) to serum bottles, which were then crimp-sealed and placed on a shaker at approximately 200 rpm to maintain well-mixed solutions. The headspaces of the bottles were not flushed with N2 gas prior to beginning the experiments because reduction with Fe0 proceeds rapidly, and the rapid reduction reaction consumes the O2 present in the reaction vessel. The H2 generated by these reactions created positive pressure in the serum bottle. Samples for U(VI) concentration and isotope analyses were extracted from the reaction vessels in approximately 20 min intervals using N2 flushed syringes after injecting N2 gas into the serum bottle to compensate for the removed sample volumes. The pH was monitored as samples were extracted, and samples maintained a pH of 7.5 ( 0.5 throughout experiments. Extracted samples were filtered with 0.20 µm filters. All experiments were performed in duplicate. A control experiment, identical to the reduction experiments except that Fe0 was absent, was also conducted simultaneously. Bacterial Experiments. Two microbes, Anaeromyxobacter dehalogenans (strain 2CP-C) and Geobacter sulfurreducens, were utilized in the bacterial reduction experiments. G. sulfurreducens was selected for its success at anaerobically reducing metals (e.g., ref 29). G. sulfurreducens and other Geobacteriacea are currently being evaluated for their ability to reduce U in situ in contaminated groundwater (30). A. dehalogenans was selected because it also grows anaerobically by metal reduction and is unrelated to the Geobacter species (26, 31). A. dehalogenans was recently shown to reduce U(VI) to U(IV) (32). Cultures (100 mL) of A. dehalogenans and G. sulfurreducens were grown with 4 mM acetate and 5 mM fumarate in a bicarbonate-buffered (10 mM), low phosphate basal salts medium (pH 7-7.5) (31, 32). Fumarate was reduced after approximately 48 h, and the cells multiplied to make the medium visibly turbid. These dense cultures were transferred in 15 mL aliquots to replicate 25 mL anaerobic culture tubes. The tubes were loaded in an anaerobic chamber containing a gas atmosphere of 10% CO2 and up to 5% H2. Each culture tube was amended with 1 mM acetate. Rezazurin (0.5 mg l-1) was added to the medium to detect any infiltration of oxygen; no free oxygen was detected during the experiments. Approximately 0.25 mL ((25%) of 1360 mgL-1 U (as UO2Cl2) was added to each 25 mL culture tube to make a final solution strength of approximately 14 mgL-1 total U in the serum bottle. Sample volumes between 0.5 and 1 mL were extracted 6944
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 22, 2006
from the reaction vessels for U(VI) concentration and isotope analysis at designated time intervals (approximately 0, 2, 4, 8, 32, 48, and 96 h). Sample syringes were flushed with N2, and then N2 was injected into the reaction vessels to compensate for the removed sample volumes. Extracted samples were filtered with 0.20 µm filters. Wu et al. (32) established that U(VI) reduction by A. dehalogenans is biologically mediated and does not proceed in the absence of cells. Killed control experiments were performed in prior studies and not repeated here (32). Control experiments, without the addition of cells, were included during each experiment set. All experiments were performed in triplicate. Concentration Measurements. Concentrations of U were measured by isotope dilution mass spectrometry using a 233U + 236U double-spike either on separate small aliquots of the sample using thermal ionization mass spectrometry (TIMS) or directly on the fraction used for isotope analysis using multi-collector inductively coupled plasma mass spectrometer (MC-ICP-MS). Isotope Ratio Measurements. The double-spike approach is widely used for isotopic measurements of other elements (e.g., 33-35). A double-spike solution with a known 233U/ 236U (approximately 15 µgL-1 236U and 5 µgL-1 233U) and small contributions of 235U and 238U (236U/238U ) 820, 236U/235U ) 500) was added to each sample, which contained approximately 20 micrograms U. The minimal contributions of 235U and 238U from the double spike were accounted for and mathematically subtracted using an iterative approach to solve a nonlinear set of equations. A 233U + 236U doublespike technique enables precise control of instrumental mass discrimination during the mass spectrometric analysis (23). Ten to 50 µL aliquots of the samples collected from the experiments were added to 4 mL of a 2% nitric acid solution containing the double spike. These dilute solutions were then analyzed directly without further preparation. The dilute matrix solutes in these solutions did not affect the 238U/235U based on a comparison of these direct addition results with those obtained when the double-spiked U was purified using anion exchange (7N HNO3 and AG1-X8, 100-200 mesh, ref 36). Agreement between processed and unprocessed samples is expected, as the double-spike technique should account for the small mass bias changes that might result from the diluted sample matrix. Isobaric interferences are not generated by the Na concentrations (e.g., approximately 5 mgL-1) and other minor solute components in experimental samples. The measured 236U/233U provides a direct indication of the mass bias during an analysis. No significant changes in this ratio occurred as sample matrix concentration increased with size of aliquot indicating that instrumental mass bias did not significantly change with increasing matrix concentration for abiotic or biotic experiments. 235U/238U, 233U/238U, and 236U/238U isotope ratios were measured on a Micromass Isoprobe MC-ICP-MS located at the Field Museum of Natural History in Chicago, Illinois. Aqueous samples in 2% HNO3 were introduced into the plasma as dry aerosols generated by an Aridus desolvating nebulizer. Typical sensitivities were 100 V per mgL-1 U at an uptake rate of 0.1 mL per minute. Ion beams of 233U, 235U, 236U, and 238U were collected by a Faraday cup collector array. 235U/238U, 233U/238U, 236U/238U were measured simultaneously for a total of 20 10-second integrations. These data were input into an iterative data reduction technique similar to that used in previous double-spike analysis (35) to calculate the sample’s 238U/235U, corrected for discrimination and the small contributions of 235U and 238U from the double spike. Results are expressed in delta notation
δ238U )
(
)
Rmeas - 1 1000‰ Rstd
(1)
TABLE 1. Summary of Isotopic Results from Abiotic and Bacterial Reduction Experiments experiment Fe0
A. dehalogenans
G. sulfurreducens
b
∆δ238U (‰)a
errorb
0 41 64
-0.03 -0.08 -0.02
0.07 0.09 0.10
0 120 240 480 2160 4320 5760
-0.09
0.09
-0.10 -0.03 0.19 0.26
0.06 0.05 0.06 0.04
0 480 2160 4320
-0.02 -0.02 0.04 0.19
0.10 0.06 0.05 0.05
time (min)
a ∆δ238U represents the offset of the data point relative to the standard Standard analytical error
FIGURE 1. Results from a representative abiotic Fe0 reduction experiment indicating the total [U(VI)] ([238U] + [235U]) remaining in solution as a function of time. Initial [U(VI)] for Fe0 experiments was 130 mgL-1. Concentrations were measured using isotope dilution (error
