Environ. Sci. Technol. 2006, 40, 3879-3885
Sulfur and Oxygen Isotope Fractionation during Benzene, Toluene, Ethyl Benzene, and Xylene Degradation by Sulfate-Reducing Bacteria KAY KNO ¨ L L E R , * ,† C A R S T E N V O G T , ‡ HANS-HERRMANN RICHNOW,‡ AND STEPHAN M. WEISE† Departments of Isotope Hydrology and Isotope Biogeochemistry, Umweltforschungszentrum (Centre for Environmental Research) Leipzig-Halle, Theodor-Lieser-Strasse 4, 06120 Halle/Saale, Germany, and Umweltforschungszentrum (Centre for Environmental Research) Leipzig-Halle, Permoserstrasse 15, 04318 Leipzig, Germany
We examined the oxygen and sulfur isotope fractionation of sulfate during anaerobic degradation of toluene by sulfate-reducing bacteria in culture experiments with Desulfobacula toluolica as a type strain and with an enrichment culture Zz5-7 obtained from a benzene, toluene, ethylbenzene, and xylene (BTEX)-contaminated aquifer. Sulfur isotope fractionation can show considerable variation upon sulfate reduction and may react extremely sensitively to changes in environmental conditions. In contrast, oxygen isotope fractionation seems to be less sensitive to environmental changes. Our results clearly indicate that oxygen isotope fractionation is dominated by isotope exchange with ambient water. To verify our experimental results and to test the applicability of oxygen and sulfur isotope investigations under realistic field conditions, we evaluated isotope data from two BTEX-contaminated aquifers presented in the recent literature. On a field scale, bacterial sulfate reduction may be superimposed by processes such as dispersion, adsorption, reoxidation, or mixing. The dual isotope approach enables the identification of such sulfur transformation processes. This identification is vital for a general qualitative evaluation of the natural attenuation potential of the contaminated aquifer.
Introduction Making use of intrinsic bioremediation processes in monitored natural attenuation approaches has become an interesting option to manage hydrocarbon-contaminated aquifers. Qualitative and quantitative assessment of the biodegradation is essential for risk assessment operations as well as to develop appropriate remediation measures. In recent years, the utilization of stable isotope tracers has considerably gained importance for evaluating in situ biodegradation in aquifers. One approach to demonstrate in situ degradation is to investigate the carbon and hydrogen * Corresponding author phone: ++49-345-5585433; fax: ++49345-5585559; e-mail:
[email protected]. † UFZ Leipzig-Halle, Halle/Saale, Germany. ‡ UFZ Leipzig-Halle, Leipzig, Germany. 10.1021/es052325r CCC: $33.50 Published on Web 05/17/2006
2006 American Chemical Society
isotope ratios of contaminants and their degradation products (e.g., refs 1 and 2). However, for a reliable prediction of in situ biodegradation, the characterization and quantification of electron-accepting processes, such as the reduction of nitrate, iron, and sulfate, is extremely valuable, especially to evaluate the active oxidation potential of contaminated aquifers and the sustainability of the degradation process over long periods of time typical for natural attenuation. Of those electron-accepting processes, bacterial sulfate reduction (BSR) plays a very important role since sulfatereducing organisms can contribute considerably to the mineralization of organic contaminants in an anoxic environment and dissolved sulfate is a common constituent of the groundwater at many contaminated sites (3-8). Quantification of BSR by concentration pattern of sulfate and coexisting sulfide in aquifers is often a challenge due to possible dilution, dispersion, mixing, matrix effects, and mineral precipitation. However, the combination of hydrochemical and stable isotope tools may provide an appropriate technique for a quantitative assessment of BSR in contaminated aquifers. The quantitative relationship between sulfate concentrations and sulfur isotope ratios can be expressed by a Rayleigh model (2). According to this model, a crucial parameter linking concentration and isotope data is the isotopic enrichment factor 34. Enrichment factors for sulfur during BSR observed in experimental setups and field studies cover a wide range between -2‰ and -46‰ (e.g., refs 9-13). Although the large variety in enrichment factors, which are attributable to a multitude of environmental parameters, is not well understood, several studies indicate an inverse correlation between the extent of sulfur isotope fractionation and cell-specific sulfate reduction rates (sSRR) (e.g., refs 12, 14, and 15). The sSRR itself is affected by ambient environmental factors such as temperature or substrate availability well as by the growth conditions of individual strains. Furthermore, sulfur isotope fractionation can depend on the type of substrate (e.g., refs 13 and 16) or on the sulfate supply (17). For the particular conditions of benzene, toluene, ethyl benzene, and xylene (BTEX) contamination, only very few enrichment factors for sulfur isotopes have been reported yet. Experimental data referring to toluene as the sole source of organic carbon showed fractionation factors between -20‰ and -47‰ (18). Field investigation in aquifers contaminated with a cocktail of different BTEX compounds yielded a range of enrichment factors for sulfur from -14‰ to -25‰ (19-21). Besides sulfur isotopes, the oxygen isotopic composition of the residual sulfate may contribute valuable information for assessing BSR. However, the general isotope fractionation mechanisms of sulfate oxygen in the residual sulfate during BSR remain a topic of controversial discussion. In recent literature, two different mechanisms are discussed. While the first hypothesis suggests the dominance of a kinetic enrichment of the heavy oxygen isotope in the residual sulfate (22-26), the second hypothesis favors the idea of an isotopic equilibration of the oxygen in the sulfate with the oxygen in the ambient water via sulfur compounds generated intracellularly as intermediates during BSR (27-30). A comprehensive mathematical model describing the oxygen isotope fractionation during BSR is provided by Brunner et al. (31). Only very few data have been reported on the behavior of oxygen isotopes in residual sulfate during BSR in contaminated aquifers (8, 20, 31). This limited number of even VOL. 40, NO. 12, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3879
somewhat contradictory data does not allow general assumptions valid for the specific conditions of an organic contamination. Therefore, our motivation for this study is to close this gap and to provide systematic data on the BSRrelated oxygen and sulfur isotope fractionation in contaminated groundwater obtained during laboratory experiments. Furthermore, we are investigating the use of oxygen and sulfur isotopes of residual sulfate to serve as a tool for assessing sulfate reduction in BTEX-contaminated aquifers.
Experimental Section Microcosm Batch Experiments. Altogether, eight different batch experiments were conducted in serum flasks with two different bacterial cultures. A marine reference culture, Desulfobacula toluolica, was provided by Deutsche Sammlung von Mikroorganismen, Germany (DSMZ 7467). For part of the experiments, we used the toluene-metabolizing freshwater enrichment culture Zz5-7 obtained from the BTEX-contaminated aquifer Zeitz, Eastern Germany. For both cultures, two different sulfate sources were applied (Na2SO4 and FeSO4). Furthermore, the FeSO4 experiments were carried out with three waters of significantly different oxygen isotopic composition [H2O (I), -35.5‰; H2O (II), -9.0‰; and H2O (III), +4.5‰ V-SMOW]. Detailed information on the experimental conditions is provided as Supporting Information. For each of the eight experimental setups, parallels of two (FeSO4 experiments) or three (Na2SO4 experiments) samples were sacrificed at a time in intervals of 5-8 days. All samples were analyzed for their hydrochemical and isotopic composition [concentration of dissolved sulfate/ sulfide, amount of precipitated sulfide, δ34S(SO42-), δ18O(SO42-), and δ34S(S(-II))]. The initial concentration and isotope data of the inoculated solution were determined by sacrificing one (Na2SO4) or two (FeSO4) parallel samples for each experimental setup immediately after the inoculation of the serum bottles. Isotope Analyses. For sulfur isotope analyses, any sulfide was precipitated as ZnS and subsequently converted to Ag2S. Dissolved sulfate was precipitated as BaSO4. Sulfur isotopic compositions were measured after conversion of BaSO4 or AgS2 to SO2 by use of an elemental analyzer (continuous flow flash combustion technique) coupled with an isotope ratio mass spectrometer (delta S, ThermoFinnigan, Bremen, Germany). Sulfur isotope measurements were performed with an analytical error of the measurement of better than (0.4‰ (2σ) and results are reported in δ notation (δ34S) as part per thousand (‰) deviation relative to the Can ˜ on Diablo Troilite (CDT) standard (according to eq 1):
δ (‰) ) {(Rsample - Rstandard)/Rstandard} × 1000
(1)
where R is the ratio of the heavy to light isotopes (e.g., 34S/32S or 18O/16O). Oxygen isotope analysis on barium sulfate samples was carried out by high-temperature pyrolysis at 1450 °C in a thermal combustion elemental analyzer (TC/EA) connected to a delta plus XL mass spectrometer (ThermoFinnigan, Bremen, Germany) with an analytical precision of better than (0.5‰ (2σ). According to eq 1, results of oxygen isotope measurements are expressed in δ notation (δ18O) as part per thousand (‰) deviation relative to Vienna standard mean ocean water (V-SMOW). Detailed information on isotope preparation and measurement is provided as Supporting Information. For our batch experiments, we computed enrichment factors (34) using a modified Rayleigh equation (eq 2) that allows an accurate calculation of 34, even for 34 values of e20‰ (40): 3880
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 12, 2006
103 ln 34
)
10-3 δ34S(SO24 )measured + 1 10-3 δ34S(SO24 )initial + 1
2ln [C(SO24 )measured/C(SO4 )initial]
(2)
Chemical Analyses. For all samples, the content of dissolved sulfate and dissolved/precipitated sulfide was determined gravimetrically by weighing the recovered dried substances of BaSO4 and Ag2S, respectively. For all experimental samples, the average recovery of sulfur [remaining SO42- plus S(-II)] was 96% of the initial amount of sulfur [SO42- from medium, SO42- and S(-II) from inoculum], with single values ranging from 87% to 101%.
Results and Discussion Sulfur Isotope Fractionation in Batch Experiments. Batch experiments were set up with toluene as electron donor and dissolved sulfate as the sole electron acceptor. Since earlier conducted control experiments (data not shown) without the addition of toluene did not show changes in sulfate concentration, any consumption of dissolved SO42- and simultaneous production of sulfide would give a clear indication for the occurrence of bacterial sulfate reduction (BSR) using toluene as an organic carbon source. Assuming a complete oxidation, the reaction equation of toluene degradation (eq 3) by BSR can be written as
C7H8 + 4.5SO24 + 3H2O f
+ 2.25H2S + 2.25HS- + 7HCO(3) 3 + 0.25H
As shown in Figure 1, a variable degree of sulfate consumption was observed in all batch experiments. One approximation to describe the time-dependent degree of sulfate consumption is the calculation of sulfate reduction rates (SRR). If zero-order kinetics are assumed, SRR represents the slope of the linear regression line for the relationship between time and SO42- concentrations [SRR ) dC(SO42-)/ dt]. Even though conclusions derived from SRR data alone have to be taken cautiously, this parameter gives qualitative information on the microbiological activity in the batch experiments. The highest SRR (356 µmol L-1 day-1) was observed for D. toluolica (Na2SO4 experiment). This value is very close to the range reported for similar experiments with the same strain (18). A comparably high SRR of 225 µmol L-1 day-1 was also obtained during the Na2SO4 experiment with the enrichment culture. For the same culture, significantly lower SRR (87-91 µmol L-1 day-1) were determined during the FeSO4 experiments. Surprisingly, SRR for the FeSO4 experiments with D. toluolica were more than 1 order of magnitude lower (19-37 µmol L-1 day-1) than the value observed for the Na2SO4 experiment with the same strain. Both marine reference strain and enrichment culture were cultured in a medium containing Na2SO4 as the sole sulfate source. Switching to a different sulfate source such as FeSO4 puts considerable stress on the microorganisms. According to the observed SRR, the freshwater enrichment culture seems to develop a higher potential to adjust to the changed environmental conditions than the marine strain. A wide range of enrichment factors (-18 to -48 ‰) was obtained for our different experiments by fitting eq 2 to the measured isotope and concentration data (Table 1). The value of -18‰ was observed for one set of the FeSO4 experiments with D. toluolica [H2O (I)] that showed the lowest sulfate consumption of all experiments (max 14%). The other two sets of FeSO4 experiments with D. toluolica, which were conducted under the same experimental conditions except for the use of isotopically different water, showed slightly higher absolute sulfate consumptions (18%) and significantly
Curves B in Figures 2a,b and 3a,b show the calculated isotopic development of the sulfide pool for different experiments. The curves were computed by use of the respective enrichment factors and initial δ34S and concentration data for sulfide. Very good agreement between predicted and measured δ34S(sulfide) data is observed for all Na2SO4 experiments and FeSO4 experiments with Zz5-7. A larger scattering of sulfide isotope data and hence a less good agreement between modeled and measured data is obvious for FeSO4 experiments with D. toluolica (Figure 3a). Again, this is most likely due to the low extent of sulfate consumption and the obvious differences of the growth condition between the single samples.
FIGURE 1. Temporal development of SO42- and S(-II) concentrations during laboratory batch experiments for strain D. toluolica and the enrichment culture Zz5-7. Experiments were conducted with FeSO4 or Na2SO4 as sulfate source and three isotopically different waters. larger enrichment factors of -40‰. 34 values for Na2SO4 experiments with D. toluolica and Zz5-7 were -34‰ and -30‰, respectively, and were in the range of enrichment factors reported by Bolliger et al. (18). Largest 34 values (-46‰ to -48‰) were obtained in FeSO4 experiments with the freshwater enrichment culture. Considering the statistically valid range of the enrichment factors reported in Table 1 (95% confidence interval), all determined 34 values are within the normal range for BSR (+3‰ to -47‰) proposed by Rees (32). If closed system conditions are assumed and the contribution of sulfide added to the system by the inoculum is considered, the isotopic development of the sulfide pool (PRmix) during our experiments can be described by the combined Rayleigh-mixing model given in eq 4, where f is the ratio C(SO42-)measured/C(SO42-)initial:
δ34Ssulfide-PR-mix ) Csulfide-measured - Csulfide-initial 34 δ S(SO24 )initial Csulfide-measured
[
)]
f ln f 1-f
(
+
Csulfide-initialδ34Ssulfide-initial (4) Csulfide-measured
Oxygen Isotope Fractionation in Batch Experiments. An increase of δ18O(SO42-) or a linear relationship of δ34S(SO42-) and δ18O(SO42-) during BSR could be interpreted as kinetic enrichment of the heavy oxygen isotopes in the remaining sulfate due to the preferential utilization of lighter isotopes by microorganism. Under certain circumstances, however, the same behavior would be expected for a process dominated by isotopic equilibration. During equilibration, the δ18O of residual sulfate will continuously approach a value defined by the δ18O of the ambient water plus an equilibrium isotope fractionation factor between sulfate and water [δ18OEQ ) δ18O(H2O) + ∆EQ]. Up to a certain progress of BSR, this approach will result in a positive linear relationship between δ18O and δ34S of residual SO42-. At higher progress of BSR, δ18O values of sulfate asymptotically move toward the equilibrium level. In case of an equilibration-dominated behavior, the assumption of a positive linear relationship between δ18O(SO42-) and δ34S(SO42-) is valid only if δ18OEQ is higher than δ18O(SO42-)initial. A lower δ18OEQ would result in a negative slope of the δ18O-δ34S relationship. Figures 2a,b and 3a,b show the relationship between the fraction of remaining sulfate and δ18O(SO42-). A variable enrichment of heavy oxygen isotopes in the remaining sulfate fraction was observed in experiments using isotopically heavy [H2O (III)] and intermediate water [H2O (II)]. A depletion of 18O was found in experiments with isotopically light water [H2O (I)]. Assuming a kinetic isotope fractionation of oxygen during BSR, eq 2 can be modified to calculate apparent enrichment factors for oxygen (18) analogously to the calculation of 34 (Table 1). While experiments with H2O (I) yield positive values for 18, negative values are obtained for all other experiments with larger 18 values for experiments with the isotopically heavier water. δ18O-δ34S plots for our experiments are shown in Figures 2c,d and 3c,d. The observed slopes of the correlation lines [equivalent to the ratio of the enrichment factors (34/18)] range from -2.7 to 7.2 (Table 1). A negative ratio and the relatively high variation in experiments with isotopically different waters under otherwise identical culturing conditions strongly indicate the effect of
TABLE 1. Sulfate Consumption and Isotope Fractionation Parameters for Laboratory Experiments and Field Investigations (20, 21) culture/field study Zz5-7
D. toluolica
field data (20) field data (21)
experiment/location
max SO42- consumption (%)
FeSO4 + H2O (I) FeSO4 + H2O (II) FeSO4 + H2O (III) Na2SO4 + H2O (II ) FeSO4 + H2O (I) FeSO4 + H2O (II) FeSO4 + H2O (III) Na2SO4 + H2O (II) MLS 15 MLS 17 TSL
53 56 66 65 14 17 18 47 68 65 60
34E
(‰)
-45.7 ( 3.2 -47.1 ( 1.9 -47.9 ( 3.4 -29.5 ( 2.5 -17.5 ( 5.2 -43.1 ( 5.2 -39.8 ( 4.4 -34.2 ( 1.1 -16.0 ( 3.0 -14.1 ( 3.0 -25.3 ( 1.4
apparent 18E (‰)
34E/18E
θa
17.2 ( 2.6 -10.9 ( 2.4 -27.1 ( 4.4 -4.1 ( 0.6 30.2 ( 7.8 -12.3 ( 6.2 -33.3 ( 9.2 -7.5 ( 1.1 -4.4 ( 2.0 -4.0 ( 2.0 -9.7 ( 0.8
-2.7 4.3 1.8 7.2 -0.6 3.5 1.2 4.6 2.6 3.6 3.5
1.7 ( 0.2 1.9 ( 0.3 1.9 ( 0.2 1.7 ( 0.2 1.7 ( 0.4 2.2 ( 0.6 2.0 ( 0.3 1.9 ( 0.2 0.5-0.6 0.3-0.4 1.5-1.9
a θ for batch experiments was calculated with a ∆ EQ value of 26.5 ‰; θ calculation from field data assumed a ∆EQ range of 28-29 ‰. Stated uncertainties refer to the 95% confidence interval.
VOL. 40, NO. 12, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3881
FIGURE 2. Relationship between the fraction of residual sulfate (C/C0) and δ34S of SO42- or S(-II) and δ18O(SO42-) (panels a and b); between δ18O(SO42-) and δ34S(SO42-) (panels c and d); and between the fraction of residual sulfate and normalized oxygen isotope data (eq 5) of sulfate (panels e and f) for batch experiments with the enrichment culture Zz5-7. SO42- (1), SO42- (2), and SO42- (3) represent samples from FeSO4 experiments with H2O (I) (-35.5‰), H2O (II) (-9.0‰), and H2O (III) (+4.5‰), respectively. SO42- (4) stands for Na2SO4 experiments with H2O (II). Curves A in panels a and b represent the logarithmic fit according to eq 2, while curves B represent the hypothetical isotopic composition of S(-II) computed from eq 4. a.p., error resulting from the respective analytical precision. the isotope composition of the applied water. The results lead to the assumption that oxygen isotope exchange dominated over kinetic isotope fractionation. δ18O(SO42-) during BSR depends on several parameters such as the initial δ18O(SO42-) value, δ18O(H2O), and ∆18OEQ. While ∆18OEQ is only temperature-dependent and has to be derived from suitable experiments, the first two parameters, which can be measured directly, are specific for different sites and experimental setups. Therefore, an appropriate normalization is needed in order to compare oxygen isotope data from different sites or experiments. Furthermore, a quantitative relationship between δ18O(SO42-) and C(SO42-) 3882
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 12, 2006
FIGURE 3. Relationship between the fraction of residual sulfate (C/C0) and δ34S of SO42- or S(-II), and δ18O(SO42-) (panels a and, b); between δ18O(SO42-) and δ34S(SO42-) (panels c and d); and between the fraction of residual sulfate and normalized oxygen isotope data (eq 5) of sulfate (panels e and f) for batch experiments with the marine reference strain D. toluolica. SO42- (1), SO42- (2), and SO42(3) represent samples from FeSO4 experiments with H2O (I) (-35.5‰), H2O (II) (-9.0‰), and H2O (III) (+4.5‰), respectively. SO42- (4) stands for Na2SO4 experiments with H2O (II). Curves A and A′ in panels a and b represent the logarithmic fit according to eq 2, while curves B and B′ represent the hypothetical isotopic composition of S(-II) computed from eq 4. a.p., error resulting from the respective analytical precision. needs to be defined. A method for normalizing δ18O(SO42-) data and relating them to sulfate concentrations during BSR is provided by the mathematical model suggested by Brunner et al. (31) and can be expressed as
ln (∆δ18OEQ/∆δ18OEQ-initial) ) 2θ ln [C(SO24 )/C(SO4 )initial] (5)
with 18 ∆δ18OEQ ) δ18O(SO24 )measured - δ O(H2O) - ∆EQ
and 18 ∆δ18OEQ-initial ) δ18O(SO24 )initial - δ O(H2O) - ∆EQ
The dimensionless coefficient θ in eq 5 is the slope of the respective regression line. A crucial parameter influencing the normalization and hence the slope of the regression line is the equilibrium oxygen isotope fractionation between SO42and H2O (∆EQ). However, the value of the temperature-dependent parameter ∆EQ has not yet been well defined. For a temperature of 28 °C, extrapolated ∆EQ values range from 25.3‰ (28) to 28.9‰ (31). To provide a valid ∆EQ for the temperature used in our experiments, we conducted a mathematical simulation with the oxygen isotope and sulfate concentration data from all experiments. By stepwise variation of the parameter ∆EQ, we determined that value of ∆EQ for which the average of the coefficients θ for all experimental setups showed the lowest standard deviation. Using this mathematical approach, we obtained a ∆EQ of 26.5‰. For that particular value, the mean value of θ was 1.9 with a standard deviation of 0.2. Assuming values for ∆EQ of 25.5‰ and 27.5‰, the corresponding mean values for θ were 2.2 ( 0.4 and 1.7 ( 0.3, respectively. Considering the analytical uncertainty related to the measurement of SO42- concentrations and δ18O(SO42-) values, the determination of ∆EQ remains relatively vague. Even though a value of 26.5‰ yielded the lowest standard deviation, stating a range of 25.5-27.5‰ seems more appropriate. The single θ values for all experiments calculated with a ∆EQ of 26.5‰ are shown in Table 1. In the strict sense, eq 5 is valid only for conditions of BSR where no cell-internal reoxidation of sulfite occurs (31). Since during sulfite oxidation one oxygen atom from water has to be incorporated in to the SO42- molecule, the resulting sulfate will be isotopically lighter than the precursor sulfite (33). Rees (32) proposed that sulfur enrichment factors of larger than -25‰ require a cell-internal reoxidation of sulfite. Except for one FeSO4 experiment with D. toluolica, all our experiments yielded 34 values of larger than -25‰ and may be affected by sulfite reoxidation. The contribution of sulfitederived sulfate can vary considerably between single batches (31). This would result in a missing continuous trend of δ18O(SO42-) during BSR progress or at least in a relatively poor correlation between normalized δ18O(SO42-) data (eq 5) and the fraction of the residual sulfate (large data scattering). Furthermore, sulfite reoxidation could lead to an underestimation of ∆EQ by the mathematical approach described above. In our experiments, we observe a relatively good correlation between normalized δ18O(SO42-) data and the fraction of the residual sulfate. Hence, no clear indication is given for an influence of sulfite reoxidation on oxygen isotopes in the residual SO42-. Nevertheless, such an influence cannot be excluded. One possibility could be that the effect is constant throughout all our experiments. However, that appears to be relatively unlikely regarding the observed large variation range of 34. On the other hand, the effect of sulfite reoxidation on δ18O(SO42-) pattern may be repressed by a cell-internal recycling of sulfite-derived sulfate and a subsequent reequilibration with the ambient water, especially for lower BSR rates as observed in our experiments. Field-Scale Assessment of BSR by Use of the Isotopic Composition of Sulfate. To test the applicability of oxygen and sulfur isotope analysis for assessing BSR under realistic field conditions, we evaluated isotope data from BTEXcontaminated aquifers presented in recent publications. The first example refers to a sulfate-rich BTEX-contaminated urban aquifer in Germany [C(SO42-)max 460 mg L-1] where BSR seems to be the main electron-accepting process for anaerobic BTEX degradation (21). Sulfate concentration decrease and the enrichment of heavy oxygen and sulfur
FIGURE 4. Relationship between the fraction of residual sulfate (C/C0) and δ34S and δ18O of SO42- for field samples from BTEXcontaminated aquifers (TSL data after ref 21; MLS data after ref 20). isotopes in the remaining sulfate downstream of the source area prove the occurrence of BSR (Figure 4). By fitting the Rayleigh equation (eq 2) to the presented field data (21), we obtained a 34 value of -25‰ (Table 1). The defined environmental conditions of batch experiments make it possible to attribute the decrease of sulfate concentration to BSR and thus to calculate 34. Under aquifer conditions, however, further processes such as dispersion and adsorption may contribute to the reduction of sulfate concentrations. This would result in an overestimation of BSR and consequently in an underestimation of 34. Considering the large variation range for 34 found in our experiments and described in the recent literature, the decision whether BSR is the sole process causing C(SO42-) decrease is not possible by investigating δ34S(SO42-) alone. Unless a reliably defined 34 is provided, any δ34S-based assumption beyond the qualitative recognition of BSR remains quite speculative. The presence of concentration-relevant processes other than BSR may also lead to an underestimation of the oxygen fractionation parameter θ if computed from field data. In contrast to the significant variability of 34, our experiments yielded very consistent θ values (Table 1) despite the varying experimental settings and the utilization of two very different cultures. This suggests that oxygen isotope fractionation in sulfate during BSR is less sensitive to changes of environmental conditions. Consequently, oxygen isotope ratios may a priori be more appropriate to assess BSR under field conditions since experimentally defined values for θ, that can be applied on a field scale, seem more reliable than respective values for 34. The oxygen isotope data reported by Kno¨ller and Schubert (21) for the field investigation in the BTEX-contaminated urban aquifer yielded a value for θ between 1.5 and 1.9. This variation range is attributed to the uncertainty of the parameter ∆EQ, which has not been welldefined yet. For our calculation, we used a range of ∆EQ between 28‰ and 29‰ as suggested by Fritz et al. (28) for groundwater temperatures. Despite this uncertainty, the value of θ derived from field data is strikingly close to experimentally determined values. This agreement is promising and may imply that the reduction of sulfate concentration at the investigated field site is exclusively due to BSR. Nevertheless, further testing seems indicated to confirm the reliability of the obtained oxygen fractionation parameter θ. The second example refers to a field investigation conducted by Spence et al. (20) in a BTEX-contaminated chalk aquifer in southern England. Maximum SO42- concentrations in the investigated aquifer are ca. 140 mg L-1. VOL. 40, NO. 12, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3883
The sulfur and oxygen isotope data for two locations, showing enrichment of the heavy isotopes in the remaining sulfate downstream of the source area, are presented in Figure 4. Despite comparable relative sulfate consumption, computed enrichment factors for sulfur are between -14‰ and -16‰ (Table 1) and hence significantly lower than in the first example described above. Again, sulfur isotope ratios provide only qualitative evidence for the occurrence of BSR and do not allow for the recognition of other concentration-relevant processes superimposing BSR. θ was calculated from a range of ∆EQ between 28‰ and 29‰ and a δ18O(H2O) of -7.5‰ (V-SMOW) (34). The obtained values for θ between 0.3 and 0.6 are much lower than in our experiments and in the field study described above. One possible reason is given by Brunner et al. (31), who theoretically related low θ values to very high sulfate reduction rates (SRR). However, during our experiments we observed very consistent θ values despite differences in the SRR by 1 order of magnitude (Figure 1). Therefore, we believe that the low θ values in the chalk aquifer result from sulfur transformation processes superimposing BSR. For example, if we assume that only 50% of the SO42concentration reduction is contributed by BSR, the recalculation yielded values for 34 and θ of -45‰ and 1.6, respectively. When one looks at the hydrochemical data presented by Spence et al. (20), a further explanation has to be taken into consideration. At both described locations, dissolved nitrate is present at certain depth ranges. If the redox zonation changes due to variations in the hydrodynamics, a reoxidation of dissolved or precipitated sulfide by nitrate or even dissolved oxygen may occur. As a result, δ34S(SO42-) values decrease due to the contribution from isotopically light sulfide and δ18O(SO42-) may be affected by the incorporation of isotopically light oxygen from ambient H2O (e.g., refs 31 and 35). In general, reoxidation of reduced sulfur compounds will result in an underestimation of both 34 and θ. If reoxidation is incomplete, intermediate sulfur compounds such as elemental sulfur, thiosulfate, or sulfite are formed. Those compounds are subject to bacterial disproportionation, as observed in batch experiments and in marine environments. Bacterial disproportionation is associated with isotope effects (36-38) that can interfere with the isotope fractionation pattern originally determined by BSR, resulting in an underestimation of both 34 and θ. Besides chemical transformation, the isotopic composition of sulfate may also be affected by mixing of sulfate from isotopically different sources. The dual isotope approach considering δ34S(SO42-) as well as δ18O(SO42-) allows for the recognition of such mixing processes in contaminated (8) and noncontaminated aquifers (39). Even though an assessment of BSR beyond the qualitative recognition is not always possible under aquifer conditions, δ34S(SO42-) and δ18O(SO42-) data enable the identification of sulfur transformation processes superimposing BSR. This identification is especially vital for a general qualitative evaluation of the natural attenuation potential of the contaminated aquifer. Nevertheless, the applicability, especially of oxygen isotope ratios, may in the future be improved by an exact definition of the parameters controlling the oxygen isotope fractionation during BSR.
Acknowledgments This work is integrated in the research and development program of the UFZ. We thank the technical staff of the Stable Isotope Laboratory Halle/Saale for conducting the isotope analyses. Special thanks is addressed to Ms. Katja Tro¨ger, who did the complete hydrochemical and isotope preparation of the incubation experiments. 3884
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 12, 2006
Supporting Information Available Detailed information on experimental conditions of batch experiments and on preparation of sulfate and sulfide for isotope analysis. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Hunkeler, D.; Anderson, N.; Aravena, R.; Bernasconi, S. M.; Butler, B. J. Hydrogen and carbon isotope fractionation during aerobic biodegradation of benzene. Environ. Sci. Technol. 2001, 35, 3462-3467. (2) Morasch, B.; Richnow, H. H.; Schink, B.; Meckenstock, R. U. Stable carbon and hydrogen isotope fractionation during microbial toluene degradation: mechanistic and environmental aspects. Appl. Environ. Microbiol. 2002, 68, 5191-5194. (3) Chapelle, F. H.; Bradley, P. M.; Lovley, D. R.; Vroblesky, D. A. Measuring rates of biodegradation in a contaminated aquifer using field and laboratory methods. Ground Water 1996, 34, 691-698. (4) Reinhard, M.; Shang, S.; Kitanidis, P. K.; Orwin, E.; Hopkins, G. D.; Lebron, C. A. In situ BTEX biotransformation under enhanced nitrate- and sulfate-reducing conditions. Environ. Sci. Technol. 1997, 31, 28-36. (5) Vroblesky, D. A.; Bradley, P. M.; Chapelle, F. H. Influence of electron donor on the minimum sulfate concentration required for sulfate reduction in a petroleum hydrocarbon-contaminated aquifer. Environ. Sci. Technol. 1996, 30, 1377-1381. (6) Davis, G. B.; Barber, C.; Power, T. R.; Thierrin, J.; Patterson, B. M.; Rayner, J. L.; Wu, Q. L. The variability and intrinsic remediation of a BTEX plume in anaerobic sulphate-rich groundwater. J. Contam. Hydrol. 1999, 36, 265-290. (7) Wiedemeier, T. H.; Rifai, H. S.; Newell, C. J.; Wilson, J. T. Natural Attenuation of Fuels and Chlorinated Solvents in the Subsurface; John Wiley & Sons: New York, 1999. (8) Spence, M. J.; Bottrell, S. H.; Thornton, S. F.; Lerner, D. N. Isotopic modeling of the significance of bacterial sulphate reduction for phenol attenuation in a contaminated aquifer. J. Contam. Hydrol. 2001, 53, 285-304. (9) Kaplan, I. R.; Rittenberg, S. C. Microbiological fractionation of sulphur isotopes. J. Gen. Microbiol. 1964, 34, 195-212. (10) Canfield, D. E. Isotope fractionation by natural populations of sulfate-reducing bacteria. Geochim. Cosmochim. Acta 2001, 65, 1117-1124. (11) Harrison, A. G.; Thode, H. G. Mechanism of the bacterial reduction of sulphate from isotope fractionation studies. Trans. Faraday Soc. 1958, 54, 84-92. (12) Habicht, K. S.; Canfield, D. E. Sulfur isotope fractionation during bacterial sulfate reduction in organic-rich sediments. Geochim. Cosmochim. Acta 1997, 61, 5351-5361. (13) Detmers, J.; Bru ¨ chert, V.; Habicht, K. S.; Kuever, J. Diversity of sulfur isotope fractionations by sulfate-reducing prokaryotes. Appl. Environ. Microbiol. 2001, 67, 888-894. (14) Chambers, L. A.; Trudinger, P. A.; Smith, J. W.; Burns, M. S. Fractionation of sulfur isotopes by continuous cultures of Desulfivibrio desulfuricans. Can. J. Microbiol. 1975, 21, 16021607. (15) Aharon, P.; Fu, B. Microbial sulfate reduction rates and sulfur and oxygen fractionations at oil and gas seeps in deepwater Gulf of Mexico. Geochim. Cosmochim. Acta 2000, 64, 233-246. (16) Kleikemper, J.; Schroth, M. H.; Bernasconi. S. M.; Brunner, B.; Zeyer, J. Sulfur isotope fractionation during growth of sulfatereducing bacteria on various carbon sources. Geochim. Cosmochim. Acta 2004, 68, 4891-4904. (17) Boudreau, B. P.; Westrich, J. T. The dependence of bacterial sulfate reduction on sulfate concentration in marine sediments. Geochim. Cosmochim. Acta 1984, 48, 2503-2516. (18) Bolliger, C.; Schroth, M. H.; Bernasconi, S. M.; Kleikemper, J.; Zeyer, J. Sulfur isotope fractionation during microbial sulfate reduction by toluene-degrading bacteria. Geochim. Cosmochim. Acta 2001, 65, 3289-3298. (19) Schroth, M. H.; Kleikemper, J.; Bolliger, C.; Bernasconi, S. M.; Zeyer, J. In-situ assessment of microbial sulfate reduction in a petroleum-contaminated aquifer using push-pull tests and stable sulfur isotope analyses. J. Contam. Hydrol. 2001, 51, 179195. (20) Spence, M. J.; Bottrell, S. H.; Thornton, S. F.; Richnow, H. H.; Spence, K. H. Hydrochemical and isotope effects associated with petroleum fuel biodegradation pathways in a chalk aquifer. J. Contam. Hydrol. 2005, 79, 67-88.
(21) Kno¨ller, K.; Schubert, M. Sulfur and carbon isotope investigation for assessing natural attenuation in a hydrocarbon contaminated urban aquifer. V-SSAGI short papers, Punta del Este: Uruguay, 2006, ISBN 9974-0-0327-X, 448-450. (22) Aharon, P.; Fu, B. Microbial sulfate reduction rates and sulfur and oxygen isotope fractionations at oil and gas seeps in deepwater Gulf of Mexico. Geochim. Cosmochim. Acta 2000, 64, 233-246. (23) Mizutani, Y.; Rafter, T. A. Oxygen isotopic composition of sulphatessPart 4; Bacterial Fractionation of oxygen isotopes in the reduction of sulphate and in the oxidation of sulphur. N. Z. J. Sci. 1969, 12, 60-68. (24) Aharon, P.; Fu, B. Sulfur and oxygen isotopes of coeval sulfates sulfide in pore fluids of cold seep sediments with sharp redox gradients. Chem. Geol. 2003, 195, 201-218. (25) Mandernack, K. W.; Krouse, H. R.; Skei, J. M. A stable sulfur and oxygen isotopic investigation of sulfur cycling in an anoxic marine basin, Framvaren Fjord, Norway. Chem. Geol. 2003, 195, 181-200. (26) Kno¨ller, K.; Fauville, A.; Mayer, B.; Strauch, G.; Friese, K.; Veizer, J. Sulfur cycling in an acid mining lake and its vicinity in Lusatia, Germany. Chem. Geol. 2004, 204, 303-323. (27) Mizutani, Y.; Rafter, T. A. Isotopic behaviour of sulphate oxygen in the bacterial reduction of sulphate. Geochem. J. 1973, 6, 183191. (28) Fritz, P.; Basharmal, G. M.; Drimmie, R. J.; Ibsen, J.; Qureshi, R. M. Oxygen isotope exchange between sulphate and water during bacterial reduction of sulphate. Chem. Geol. 1989, 79, 99-105. (29) Bo¨ttcher, M. E.; Bernasconi, S. M.; Brumsack, H.-J. Carbon, sulfur, and oxygen isotope geochemistry of interstitial waters from the western mediterranean. Proc. Ocean Drill. Program 1999, 161, 413-421. (30) Lloyd, R. M. Oxygen isotope behavior in the sulfate-watersystem. J. Geophys. Res. 1968, 73, 6099-6110. (31) Brunner, B.; Bernasconi, S. M.; Kleikemper, J.; Schroth, M. H. A model for oxygen and sulphur isotope fractionation in sulphate during bacterial sulphate reduction processes. Geochim. Cosmochim. Acta 2005, 69, 4773-4785. (32) Rees, C. E. A steady-state model for sulphur isotope fractionation in bacterial reduction processes. Geochim. Cosmochim. Acta 1973, 37, 1141-1162.
(33) Taylor, B. E.; Wheeler, M. C.; Nordstrom, D. K. Isotope composition of sulfate in acid mine drainage as measure of bacterial oxidation. Nature 1984, 308, 538-541. (34) Elliot, T.; Andrews, J. N.; Edmunds, W. M. Hydrochemical trends, palaeorecharge and groundwater ages in the fissured Chalk aquifer of the London and Berkshire Basins, UK. Appl. Geochem. 1999, 14, 333-363. (35) Toran, L.; Harris, R. F. Interpretation of sulfur and oxygen isotopes in biological and abiological sulfide oxidation. Geochim. Cosmochim. Acta 1989, 53, 2341-2348. (36) Habicht, K. S.; Canfield, D. E.; Rethmeier, J. Sulfur isotope fractionation during bacterial sulfate reduction and disproportionation of thiosulfate and sulfite. Geochim. Cosmochim. Acta 1998, 62, 2585-2595. (37) Bo¨ttcher, M. E.; Thamdrup, B.; Vennemann, T. W. Oxygen and sulfur isotope fractionation during anaerobic bacterial disproportionation of elemental sulfur. Geochim. Cosmochim. Acta 2001, 65, 1601-1609. (38) Bo¨ttcher, M. E.; Thamdrup, B.; Gehre, M.; Theune, A. 34S/32S and 18O/16O fractionation during sulphur disproportionation by Desulfobulbus propionicus. Geomicrobiol. J. 2005, 22, 219226 (39) Kno¨ller, K.; Trettin, R.; Strauch, G. Sulfur cycling in the drinking water catchment area of Torgau-Mockritz (Germany): insights from hydrochemical and stable isotope investigations Hydrol. Proc. 2005, 19, 3445-3465. (40) Mariotti, A.; Germon, J. C.; Hubert, P.; Kaiser, P.; Letolle, R.; Tardieux, A.; Tardieux, P. Experimental determination of nitrogen kinetic isotope fractionation: Some principles. Plant Soil 1981, 62, 413-430.
Received for review November 18, 2005. Revised manuscript received March 19, 2006. Accepted April 7, 2006. ES052325R
VOL. 40, NO. 12, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3885
