Environ. Sci. Technol. 2009, 43, 4851–4857
Microbial Oxidation of Pyrite Coupled to Nitrate Reduction in Anoxic Groundwater Sediment
+ 2NO3 + 12H + 10e ⇒ N2 + 6H2O
(1)
Suggested electron donors capable of reducing nitrate are ferrous iron and organic carbon (13): + 5Fe2+ + NO3 + 12H2O ⇒ 5Fe(OH)3 + 1/2N2 + 9H
(2)
†,
CHRISTIAN JUNCHER JØRGENSEN, * OLE STIG JACOBSEN,‡ BO ELBERLING,† AND JENS AAMAND‡ Department of Geography and Geology, University of Copenhagen, Copenhagen, Denmark, Department of Geochemistry, The Geological Survey of Denmark and Greenland (GEUS), Copenhagen, Denmark
Received December 3, 2008. Revised manuscript received May 4, 2009. Accepted May 5, 2009.
+ 5CH2O + 4NO3 + 4H ⇒ 5CO2 + N2 + 7H2O
(3)
It is generally assumed that pyrite (FeS2) and other reduced sulfur compounds are important electron donors for in situ denitrification processes in anoxic groundwater environments (1, 2, 6-8, 10, 11, 14, 15): 2+ 4H+ ⇒ 7N2(g) + 10SO4(aq) + 5FeS2(s) + 14NO3(aq) 2+ 5Fe(aq) + 2H2O(aq) (4)
Although many areas in Denmark are intensively agricultured, the discharge of nitrate from groundwater aquifers to surface water is often lower than expected. In this study it is experimentally demonstrated that anoxic nitrate reduction in sandy sediment containing pyrite is a microbially mediated denitrification process with pyrite as the primary electron donor. The process demonstrates a temperature dependency (Q10) of 1.8 and could be completely inhibited by addition of a bactericide (NaN3). Experimentally determined denitrification rates show that more than 50% of the observed nitrate reduction can be ascribed to pyrite oxidation. The apparent zero-order denitrification rate in anoxic pyrite containing sediment at groundwater temperature has been determined to be 2-3 µmol NO3- kg-1 day-1. The in situ groundwater chemistry at the boundary between the redoxcline and the anoxic zone reveals that between 65 and 80% of nitrate reduction in the lower part of the redoxcline is due to anoxic oxidation of pyrite by nitrate with resulting release of sulfate. It is concluded that microbes can control groundwater nitrate concentrations by denitrification using primarily pyrite as electron donor at the oxic-anoxic boundary in sandy aquifers thus determining the position and downward progression of the redox boundary between nitratecontaining and nitrate-free groundwater.
Introduction Nitrate (NO3-) is a common inorganic pollutant in shallow groundwater aquifers and drinking water abstraction wells and is often seen as a pollutant linked to agricultural fertilization (1-4) and other nonagricultural sources (5). Marked variability in nitrate concentrations is present in the groundwater, which is usually linked to variations in groundwater flow and the reduction of nitrate (6-11). Various chemical and biological processes have been reported to be responsible for the natural reduction of nitrate in anoxic groundwater by a series of redox reactions summarized in the following generalized denitrification half reaction (12): * Corresponding author phone (+45) 3532 2500; e-mail: Cjj@ geo.ku.dk. † Department of Geography and Geology, Faculty of Science, University of Copenhagen. ‡ Department of Geochemistry, The Geological Survey of Denmark and Greenland (GEUS). 10.1021/es803417s CCC: $40.75
Published on Web 05/21/2009
2009 American Chemical Society
The oxidation of pyrite primarily in oxic conditions by abiotic as well as microbial mechanisms is generally recognized as a complicated process, and numerous reaction equations including oxygen (O2), manganese (Mn4+), and ferric iron (Fe3+) have been reported (16, 17): 22+ FeS2(s) + 3.5O2(g) + H2O ⇒ 2SO4(aq) + Fe(aq) + 2H+
(5) 2FeS2(s) + 7.5MnO2(s) + 11H+ ⇒ Fe(OH)3(s) + 2SO4(aq) + 2+ 7.5Mn(aq) + 4 H2O
(6)
3+ 22+ FeS2(s) + 8H2O + 14Fe(aq) ⇒ 2SO4(aq) + 15Fe(aq) + 16H+ (7)
It can be ruled out that oxygen is involved in anoxic pyrite oxidation, whereas the role of ferric iron under anoxic conditions is less clear. However, in order for Fe3+ to be present in a significant quantity in solution, pH must typically be below pH 3 (12), which is below typical groundwater pH levels. Since pure chemical reaction between nitrate and pyrite cannot interact kinetically in nature at significant rates (12), microbial catalysis has been proposed as an explanation of the electron transfer (6, 7). However, none of the few studies which have conducted controlled laboratory experiments investigating the pyrite oxidation capacity of nitrate (3, 18) have been able to prove the involvement of microorganisms in the process. Therefore, several questions remain unanswered regarding nitrate transformation rates, including the role of different electron donors and various rate-limiting environmental factors as well as the importance of soil microbes in nitrate reduction in deeper aquifers. Questions which all seem crucial to resolve in order to meet the threat that nitrate contamination poses in terms of securing future drinking water resources. This study determines experimentally whether or not in situ nitrate reduction in a sandy aquifer can be a microbially mediated denitrification process with pyrite as the electron donor. Furthermore, denitrification potentials inferred from laboratory experiments are compared to in situ groundwater chemistry profiles to investigate to which extent biotic denitrificationcan controlgroundwaternitrateconcentrations.
Materials and Methods Study Site and Sediments. The sediment used in the incubation experiments comes from the Fladerne Bæk VOL. 43, NO. 13, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4851
experimental site, on the Karup heath plain, Denmark (56°N, 9°E) approximately 9 km South West of Rabis Bæk (1, 19). Land use at the site is agriculture with potato production as the main crop. The site is situated in an unconfined sandy aquifer of approximately 20 m in depth where the lower boundary of the aquifer is delimited by the top of the prequarternary surface. The aquifer sediment consists of meltwater sand with an average grain size of 0.35 ( 0.15 mm. The groundwater table is situated 2-3 m below surface with seasonal variations of approximately 1.5 m (see Supporting Information (SI) Figure S1). Annual precipitation is approximately 800 mm with a net annual infiltration of approximately 400 mm. Horizontal and vertical groundwater transport rates of 40-60 m yr-1 and 0.5-1.0 m yr-1 have been reported in an adjacent aquifer (1), which correspond to values measured in the study area (unpublished data). The average pH of the rainwater in the area is between 5.0 and 5.5. Mean annual groundwater temperature is approximately 9 °C at the depth of zero degree annual temperature variation amplitude. Small amounts of carbon are identified as buried lignite in depths below 9 m. Sediment samples were collected using a standard sand bailer procedure in 0.5 m depth increments to a depth of 11 m below surface terrain. Sediment samples were collected as fully water-saturated sediments which were sealed off in an airtight glass jar to avoid contact with the atmosphere and kept at 5 °C. Sediment samples representing the oxic zone, the redoxcline and the anoxic zone (i.e., 4.5, 7.0, and 10.0 m below surface terrain, see Figure 3) were used in the incubation experiments. An 11 m long, 50 mm diameter PEHD casing equipped outside with 17 sintered stainless steel screens of each 5 cm in height in 0.5 m intervals was installed to allow for subsequent depth specific high resolution groundwater sampling. Sediment Incubation. Reactors were assembled using 100 ( 2.00 g of saturated sediment (corresponding to approximately 85 g dry weight), 450 mL Milli-Q water and 25 mL, 1000 mg L-1 KNO3 solution added to 585 mL glass bottles closed with chlorobutyl rubber septa and aluminum crimp. Reactors were incubated at temperatures corresponding to approximately 1, 1.5, 2, 3, and 4.5 times mean annual groundwater temperature: 9.1, 13.6, 21.5, 28.6, and 40.6 °C for a duration of 177 days. To overcome the natural inhomogeneous pyrite distribution in the natural sediment and to increase process rates, 1 ( 0.02 g of pyrite was added to the solution in a reactor batch parallel to the unmodified natural batch and mixed mechanically by shaking the reactors. Two stainless steel needles (TERUMO 18G, 1.2 × 40 mm and 17G, 1.2 × 150 mm) were fastened to each septum and fitted with sterile CODAN Steritex 3-way stopcocks for sample extraction and gas sparging. To achieve sterile conditions, all solutions, bottles, needles and septa where autoclaved at 115 °C at 15 psi. for 30 min prior to use. To eliminate any microbial activity, 10 mL 10% sodium azide (NaN3) was added as inhibitor (3, 20, 21) to a separate reactor batch. After assembly, the solution and headspace of the reactors were stripped of residual oxygen by sparging with reactive copper column purified nitrogen gas (5.0) for a minimum of 15 min (22). All reactors were prepared in duplicate (23, 24) and were stored in complete darkness during the experiment. For each sediment depth, parallel reactors with and without added pyrite were prepared. Additional reactors with sediment from the redoxcline (≈7.0 m below surface terrain) were incubated at various temperatures with and without the addition of NaN3. Sterile 50 mg NO3- L-1 blanks with and without pyrite and NaN3 were prepared as control units. Pyrite. Crushed pyrite from the Nanisivik mine in Northern Canada was used for the experiments (25). The elemental composition of the pyrite was analyzed by means 4852
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 13, 2009
of X-ray fluorescence. The chemical composition indicates the presence of a typical pyrite with minor contributions of silicates and oxides (SiO2, CaO and K2O) and zinc minerals (see SI Table S1). Massive pyrite crystals were crushed by hand in a stainless steel mortar providing fresh surface areas free of coatings and vibrated in brass sieves (45 and 200 µm) for 30 min. After crushing and sieving the pyrite fraction between 45 and 200 µm was washed in demineralized water and ultrasonicated for 1 h to remove fine pyrite particles adhering to the surface. The pyrite was then shaken in a 10% hydrochloric acid solution at 60 rpm for 1 h, washed with demineralized water, suction dried and sterilized with 95% ethanol (23) and left overnight to air-dry in a glovebox in a pure nitrogen atmosphere. During storage, the dry pyrite was kept in a pure nitrogen atmosphere (nitrogen grade 5.0). Sampling and Analysis. In each sampling, a 15 mL solution sample was extracted from the reactor with a sterile disposable syringe. Immediately after sampling, pH was measured (Mettler Toledo MP220 with Inlab 410 electrolyte 9823) and total alkalinity was determined by titration of 10 mL solution with 0.02 M hydrochloric acid (Schott Titroline easy with L300 electrolyte) to an end-point of pH 4.5 (26). The concentration of ferrous iron was measured by spectrophotometry at 520 nm (Jenway 6405 Spectrophotometer) upon reaction with bipyridine and acetate buffer. Concentrations of NO3-, NO2-, and SO42- were measured in an ion chromatograph in one determination within 24 h of sampling (Dionex Ionpac AS14 4 mm column, CD20 conductivity detector and GP50 gradient pump). Following sampling, each reactor was flushed with purified N2 gas for two minutes to eliminate any oxygen in the reactor. Sediment content of total carbon (TC) and total sulfur (TS) were determined by combustion in a LECO CS-200. Total pyrite was determined by stepwise extraction by first boiling the sediment sample in 20% hydrochloric acid for 4 h to remove all nonpyrite-associated iron and sulfide. Upon extraction the washed sediment was boiled in concentrated nitric acid for 16 h, which released pyrite-associated iron and sulfide to the solution (according to ref 27). Pyriteassociated iron was determined by atomic absorption spectroscopy (AAS) and recalculated to soil pyrite content assuming a typical FeS2 molar ratio of ≈1:2. The mean grain size, degree of sorting, and skewness were calculated, based on soil texture determined in half phi intervals (28, 29). Depth specific groundwater samples were extracted in 100 mL glass bottles sealed with butyl rubber septa and stored in the dark at 5 °C. Groundwater pH, dissolved O2 concentration and temperature were measured in the field, whereas alkalinity and concentrations of SO42-, NO3-, NO2-, and Fe2+ were measured in the laboratory as described above. Elemental Transformation Rates. The accumulated amount of moles n, of element i removed from or released to the reactor solution over time up to sampling occasion k was calculated from the concentration (Cmeas) measured, using the following equation: k
k + nik ) [Ci,k measVtotal
∑C
s s i, measVsample](moles)
s)1
k
where V total is the total volume of the solution in the reactor after removal of the kth sample, and Vssample is the volume of sample removed on sampling occasion k (23, 30). Final rate units are expressed as µmol kg-1 dry matter. Temperature Sensitivity. The temperature sensitivity for nitrate reduction (i.e., Q10 value, ) (R2/R1)10/(T2-T1)) were calculated directly from the rate constants (R1, R2) at different incubation temperatures. Statistical Analysis. Elemental transformation rates (dC dt-1) were calculated by means of Analysis Toolpak in EXCEL, using linear regression of the concentration changes over
FIGURE 1. Chemical development of NO3- (circles), SO42- (crosses), Fe2+ (diamonds), pH (triangles) and alkalinity (squares) of a single natural and pyrite amended reactor incubated at 21.5 °C over 177 days. Open symbols represent concentrations in the natural reactor and filled symbols represent concentrations in the pyrite-amended reactor.
FIGURE 2. Apparent nitrate reduction (A and C) and sulfate production (B and D) rates per kg redoxcline sediment incubated at 21.5 °C with and without added NaN3. Open symbols represent concentrations in the natural reactor and filled symbols represent concentrations in the pyrite-amended reactor. Error bars show standard error of the mean. Fit lines are applied when statistically significant (p < 0.05). A and B: nitrate and sulfate concentrations of the inhibited reactors. C and D: nitrate and sulfate concentrations of the noninhibited reactors. time with a significance level of p < 0.05. Average rates are listed with the standard error of the mean value ((SE).
Results and Discussion Elemental Transformation in Laboratory Reactors. An example of the chemical development of the measured parameters in eq 4 over 177 days in a single natural and pyrite-amended reactor with redoxcline sediment incubated
at room temperature is shown in Figure 1A and B. A clear tendency for nitrate concentrations to decrease over time is observed in both the natural and pyrite-amended reactor, with the most rapid nitrate removal occurring in the amended reactor. In the pyrite-amended reactor the nitrate removal was accompanied by an increase in sulfate concentrations, whereas the concentration of sulfate is stable in the natural reactor. Concentrations of dissolved ferrous iron are observed VOL. 43, NO. 13, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4853
TABLE 1. Zone Specific Nitrate and Sulfate Transformation Rates at Room Temperature ± Standard Error of the Mean, Expressed As Electron Equivalentsa Type
Zone
NO3- reduction µmol kg-1 day-1
SO42- production µmol kg-1 day-1
NO3- e-equivalents
SO42- e-equivalents
N P N P N P
oxic oxic redoxcline redoxcline anoxic anoxic
2.0 ( 0.0* 4.3 ( 0.4 2.3 ( 1.0 7.2 ( 0.1 4.9 ( 0.1 13.6 ( 0.2
0.0 ( 0.0 1.1 ( 0.0* 0.0 ( 0.0 2.4 ( 0.1 0.9 ( 0.0* 5.8 ( 0.3
-10 -22 -12 -36 -25 -68
0 8 0 17 6 41
a
N ) natural reactors, P ) pyrite-added reactors. * ) only 1 significant reactor.
to decrease to concentrations close to the detection limit during the first few weeks of the incubation time in both the natural and the pyrite-amended reactors. The pH of the reactor solutions increase gradually over time in both reactor types with the highest increase in the pyrite-amended reactor. Values of total alkalinity of the reactor solutions are in the low range (0.07-0.14 meq L-1) of what is detectable by titration in Milli-Q bulk solution and show no significant (p > 0.05) development over time in neither the natural nor the pyrite-amended reactor. Nitrite was not observed above the detection limit. The chemical development of the reactors indicates a consumption of nitrate and protons in both reactor types which, in the presence of pyrite, results in accelerated nitrate and proton consumption and formation of sulfate. The low concentration of dissolved ferrous iron and visual observations of presumable ferric hydroxide accumulations on top of the incubated sediment in mainly the sulfate producing reactors over time may indicate simultaneous reaction of nitrate with ferrous iron as described in reaction eq 2 with a possible counteracting effect on the pH increase. Chemical data from pyrite-amended reactors with intended leaking stop cocks (data not shown) were used to evaluate the possibility of pyrite oxidation and iron precipitation due to oxygen contamination of the natural and pyrite-amended reactors. Accelerated pyrite oxidation and iron precipitation due to oxygen contamination of the anoxic natural and pyrite-amended reactors are not considered plausible pathways since the results from the leaking reactors showed pH levels in the range of 3.5 and sulfate concentrations of 3000 µmoles kg-1 at the end of the experiment. Oxidation of pyrite by Mn4+ as described in reaction eq 6 lead to production of sulfate and consumption of protons under stable nitrate concentrations. However, since concentrations of nitrate and sulfate as well as pH level were observed to be constant over time in the inhibited reactors, where Mn4+ would oxidize pyrite if present in significant amounts, it is concluded that Mn4+ is not an oxidant of pyrite in this system. Similarly, direct anoxic oxidation of pyrite by Fe3+ as described in reaction eq 7 is assumed to be negligible due to the extremely low solubility of Fe3+ at the current and the increasing pH levels. No indications of nitrate reduction by organic matter with resulting increases in solution alkalinity have been detected during the experiments. However, due to being close to the detection limit on the total alkalinity measurements, nitrate reduction by organic carbon cannot be completely ruled out as a possible simultaneous nitrate removal pathway. A comparison of both natural redoxcline and pyriteamended redoxcline reactors with and without the presence of a microbial inhibitor (NaN3) shows that no significant nitrate reduction or sulfate production occur in the inhibited reactors (Figure 2A-D). It is therefore concluded that no purely chemical reduction of nitrate takes place by either electron donors found in the natural sediment or by the added pyrite (Figure 2A and B). By contrast, the noninhibited reactors (Figure 2C and D) show significant nitrate transformation over time in both the natural and the pyrite4854
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 13, 2009
amended reactors in accordance to linear regression statistics indicating possible zero-order kinetics during the investigated time span. Nitrate reduction is observed in all investigated redox zones in both the unmodified and the pyrite-amended reactors (Table 1). Average nitrate reduction rates (dC dt-1) at room temperature in all three zones are found to be 2-3 times higher in the pyrite-amended reactors than in the parallel natural reactors. Increasing sulfate concentrations are not observed in the natural reactors containing sediment from the pyrite-depleted oxic and redoxcline zones. By contrast, significant production of sulfate is only observed in the pyrite-amended reactors from the oxic and redoxcline zones as well as in both the pyrite-amended reactors and one of the natural reactors of the anoxic zone. The nonsignificant sulfate production rate in the duplicate natural reactor of the anoxic zone may be explained by the natural heterogeneous distribution of pyrite in the anoxic sediments as also observed in ref 7. The highest denitrification rates and the largest natural nitrate reduction capacities are observed in the sediment from the pyrite containing anoxic zone where the addition of pyrite dramatically increases both nitrate reduction and sulfate production rates. In this way, sulfate production in the experimental reactors is only observed when pyrite is present in either the natural or added form. Consequently, a prerequisite for anoxic pyrite oxidation coupled to nitrate reduction is that both compounds are present at the same time. It is seen that microbial pyrite oxidation capacity under reduction of nitrate is available in both sediments from the oxic zone and at the redoxcline where the sediment is depleted of naturally occurring pyrite. It is furthermore observed that addition of pyrite to pyrite-depleted sediment from the oxic and redoxcline zones may stimulate denitrification rates to a similar level as the non-pyrite-amended anoxic sediment. It is therefore concluded that microbial pyrite oxidation by nitrate can be stimulated or initiated in sediment from all investigated sediment zones under anoxic conditions. Further, the addition of pyrite to pyrite-depleted sediment may increase denitrification rates by a factor 2-3 and may be responsible for at least 50% of the nitrate reduction measured. Incubation Temperature Response. Nitrate reduction rates for pyrite-amended redoxcline sediment reactors showed a temperature dependency (Q10) of 1.8 when incubated at temperatures corresponding to 1-3 times groundwater temperature (9.1-28.9 °C). Significantly lower reaction rates than would be expected from the exponential temperature increase of the significant fit line are observed at 40.6 °C (Figure 3). It appears that an apparent temperature maximum of the reaction have been exceeded at approximately 4.5 times groundwater temperature, which is a clear indication of biological process dependency. When correcting the observed denitrification rate from unmodified sediment from the anoxic zone (Table 1) to average groundwater temperature (∼9 °C) with the calculated Q10 value (Figure 3) it is found that the natural denitrification rate of
FIGURE 3. Normalized temperature dependency (Q10) relation (y ) 0.61 × 10(0.056x); r2 ) 0.96; p < 0.05) for measured nitrate reduction rate constants in pyrite-amended redoxcline sediment reactors at various incubation temperatures. X-axis shows incubation temperature and y-axis shows measured rate constants normalized to 10 °C. Dots show average rate constants of three replicates with standard errors as vertical error bars.
FIGURE 4. Depth-specific in situ groundwater chemistry and sediment content of total carbon and pyrite at well P1. The approximate positions of the oxic zone, the redoxcline and the anoxic zone are shown by dashed lines.
FIGURE 5. In situ concentrations of nitrate, sulfate, total alkalinity, and ferrous iron expressed as electron equivalents across the nitrate disappearance front across the lower boundary of the redoxcline, as indicated by the dotted line. one kilogram of pyrite containing anoxic zone sediment will groundwater will inevitably show temporal variation and the be approximately 850 µmol NO3- kg-1 year-1. shown geochemical data only represent a single point in In Situ Groundwater Chemistry and Aquifer Zonation. time. Consequently, the position of the nitrate front is Because of natural variation in precipitation and fertilizer nonstationary and the depth-specific concentration of NO3will change due to source water concentration, groundwater input to the aquifer system the chemical composition of the VOL. 43, NO. 13, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4855
flow, and in situ nitrate reduction. Consequently, a prerequisite for microbial pyrite oxidation coupled with nitrate reduction is that nitrate containing groundwater is present in close proximity to sediment containing pyrite. The aquifer sediment is virtually free of inorganic carbon and total carbon is equal to total organic carbon (TOC). FeS2 is absent in the oxic parts of the aquifer gradually rising to around 8 mmol kg-1 from 9 m below surface terrain (Figure 4). Alkalinity and pH profiles show close correlation with higher values in close proximity to the groundwater table and at the depth where pyrite and lignite appear. The pH profile of the aquifer indicates that pyrite oxidation by atmospheric oxygen, as shown in reaction eq 5, is likely to be of negligible proportions since the absolute pH values matches those of the infiltrating rainwater rather than an acidified percolate due to oxic pyrite oxidation. Dissolved oxygen concentrations decrease to below 1 mg L-1 at depth 5.3, indicating the transition from the oxic zone to the redoxcline. Nitrate concentration varies from 2 to 5 mmol L-1 across the oxic zone into the redoxcline with a steep decrease below the detection limit at the transition from the redoxcline to the anoxic zone corresponding to the depth with increasing concentrations of sulfate and ferrous iron (Figure 3). Large seasonal variations in nitrate and sulfate concentrations have been observed (see SI Figure S2), probably controlled by nitrogen input and infiltration rates. Despite these marked variations, all geochemical profiles from the aquifer show that NO3- is completely reduced over a very narrow depth interval at the lower boundary of the redoxcline (see also SI Figure S2) with a resulting production of SO42-, Fe2+ and alkalinity as previously described for other sediments (1, 6-8, 10, 11). It is observed that significant amounts of primarily sulfate electron equivalents appear immediately below the nitrate disappearance depth along with lesser increases of both alkalinity and ferrous iron electron equivalents (Figure 5). Using the SO42-:NO3- electron equivalent ratio approach (1, 11), it is suggested that between 65 and 80% of the reduced nitrate electron equivalents at the lower boundary of the redoxcline may be explained by reaction with reduced pyritic sulfide producing sulfate, ferrous iron and alkalinity. The residual fraction may be explained by a combination of various electron donors such as organic compounds and dissolved metal ions as described in other studies (15, 31). Environmental Implications and Future Perspectives. Improved understanding of both processes and reaction rates associated with nitrate removal from groundwater will assist in future targeting of zones for drinking water wells that are less likely to become affected by nitrate. Various in situ denitrification studies have reported both zero-order (15) and first-order denitrification kinetics (8, 9), and it is therefore still unknown if denitrification with pyrite as an electron donor is independent of nitrate concentrations and electron donor availability. Even though the process of nitrate reduction by pyrite oxidation may be important for in situ nitrate removal when agricultural nitrate leaches into the groundwater in organic poor sandy aquifers, the risk of potential trace metal mobilization following pyrite oxidation (32) will continue to threaten the drinking water quality and will remain a negative consequence of nitrate reduction by pyrite oxidation in aquifers. Further information on potential rate-limiting factors and environmental controls on the actual microbes involved in the process would be valuable in future assessments of in situ denitrification potentials in a broader range of sediment and aquifer types. Preliminary results using Thiobacillus denitrificans indicate that specific bacterial strains are able to oxidize pyrite by nitrate reduction under anoxic conditions (see SI Figure 4856
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 13, 2009
S3). This supports our main conclusion that microbes can control groundwater nitrate concentrations by denitrification using pyrite as electron donor at the oxic-anoxic boundary in sandy aquifers.
Acknowledgments This study was funded with a grant from Geocenter Copenhagen (“Geocenterbevilling”) (www.geocenter.dk). We would like to thank P. Venslev and J. C. Bailey at the Department of Geography and Geology at the University of Copenhagen for preparing and analyzing the pyrite and the staff at the Department of Geochemistry (GEUS) for valuable advice and help in the laboratories. We would also like to thank our three anonymous reviewers for their generous comments and suggestions.
Supporting Information Available Temporal variations in groundwater level and temperature; Temporal variations in groundwater nitrate and sulfate concentrations; Chemical composition of the used bulk pyrite; Preliminary results on anoxic pyrite oxidation coupled to nitrate reduction by pure culture strain Thiobacillus denitrificans. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Postma, D.; Boesen, C.; Kristiansen, H.; Larsen, F. Nitrate reduction in an unconfined sandy aquifer - water chemistry, reduction processes, and geochemical modeling. Water Resour. Res. 1991, 27 (8), 2027–2045. (2) Korom, S. F. Natural denitrification in the saturated zonesA review. Water Resour. Res. 1992, 28 (6), 1657–1668. (3) Devlin, J. F.; Eedy, R.; Butler, B. J. The effects of electron donor and granular iron on nitrate transformation rates in sediments from a municipal water supply aquifer. J. Contam. Hydrol. 2000, 46 (1-2), 81–97. (4) Darbi, A.; Viraraghavan, T.; Butler, R.; Corkal, D. Column studies on nitrate removal from potable water. Water Air Soil Pollut. 2003, 150 (1-4), 235–254. (5) Wakida, F. T.; Lerner, D. N. Non-agricultural sources of groundwater nitrate: A review and case study. Water Res. 2005, 39 (1), 3–16. (6) Ko¨lle, W.; Werner, P.; Strebel, O.; Bo¨ttcher, J. Denitrifikation in einem reduzierenden Grundwasserleiter. Vom Wasser. 1983, 61, 125–147. (7) Ko¨lle, W.; Strebel, O.; Bo¨ttcher, J. Formation of sulphate by microbial denirification in a reducing aquifer. Water Supply. 1985, 3, 35–40. (8) Bo¨ttcher, J.; Strebel, O.; Duynisveld, W. H. M. Vertikale Stoffkonzentrationsprofile im Grundwasser eines Lockergesteins-Aquifers und deren Interpretation (Beispiel Fuhrberger Feld). Z. Dt. Geol. Ges. 1985, 136, 543–552. (9) Frind, E. O.; Duynisveld, W. H. M.; Strebel, O.; Boettcher, J. Modeling of multicomponent transport with microbial transformation in groundwatersThe Fuhrberg case. Water Resour. Res. 1990, 26 (8), 1707–1719. (10) Robertson, W. D.; Russell, B. M.; Cherry, J. A. Attenuation of nitrate in aquitard sediments of southern Ontario. J. Hydrol. 1996, 180 (1-4), 267–281. (11) Tesoriero, A. J.; Liebscher, H.; Cox, S. E. Mechanism and rate of denitrification in an agricultural watershed: Electron and mass balance along groundwater flow paths. Water Resour. Res. 2000, 36 (6), 1545–1559. (12) Stumm, W.; Morgan, J. J. Aquatic ChemistrysChemical Equilibria and Rates in Natural Waters; John Wiley & Sons: New York. 1996. (13) Pedersen, J. K.; Bjerg, P. L.; Christensen, T. H. Correlation of nitrate profiles with groundwater and sediment characteristics in a shallow sandy aquifer. J. Hydrol. 1991, 124 (3-4), 263–277. (14) Garcia-Gil, L. J.; Golterman, H. L. Kinetics of FeS-mediated denitrification in sediments from the Camargue (Rhone Delta, Southern France). FEMS Microbiol. Ecol. 1993, 13 (2), 85–91. (15) Korom, S. F.; Schlag, A. J.; Schuh, W. M.; Schlag, A. K. In situ mesocosms: Denitrification in the Elk Valley aquifer. Ground Water Monit. Rem. 2005, 25 (1), 79–89.
(16) Elberling, B.; Schippers, A.; Sand, W. Bacterial and chemical oxidation of pyritic mine tailings at low temperatures. J. Contam. Hydrol. 2000, 41 (3-4), 225–238. (17) Schippers, A.; Jorgensen, B. B. Oxidation of pyrite and iron sulfide by manganese dioxide in marine sediments. Geochim. Cosmochim. Acta 2001, 65 (6), 915–922. (18) Schippers, A.; Jorgensen, B. B. Biogeochemistry of pyrite and iron sulfide oxidation in marine sediments. Geochim. Cosmochim. Acta 2002, 66 (1), 85–92. (19) Engesgaard, P.; Kipp, K. L. A geochemical transport model for redox-controlled movement of mineral fronts in groundwaterflow systemssA case of nitrate removal by oxidation of pyrite. Water Resour. Res. 1992, 28 (10), 2829–2843. (20) Jacobsen, O. S.; Bossi, R. Degradation of ethylenethiourea (ETU) in oxic and anoxic sandy aquifers. FEMS Microbiol. Rev. 1997, 20 (3-4), 539–544. (21) Madigan, M. T.; Martinko, J. M. Brock Biology of Microorganisms; Prentice Hall: New York, 2005. (22) Chander, S.; Briceno, A. Kinetics of pyrite oxidation. Miner. Metall. Process. 1987, 4, 171–176. (23) Gleisner, M.; Herbert, R. B.; Kockum, P. C. F. Pyrite oxidation by Acidithiobacillus ferrooxidans at various concentrations of dissolved oxygen. Chem. Geol. 2006, 225 (1-2), 16–29. (24) Ghoshal, S.; Larsson, B. Studies on nitrate reduction. Effect of chemical reduction potential on nitrate reduction in subsoil. Acta Agric. Scand. 1977, 27 (3), 233–241.
(25) Elberling, B.; Balic-Zunic, T.; Edsberg, A. Spatial variations and controls of acid mine drainage generation. Environ. Geol. 2003, 43 (7), 806–813. (26) Appelo, C. A. J.; Postma, D. Geochemistry, Groundwater and Pollution; A. A. Balkema Publishers: Leiden, 2005. (27) Huertadiaz, M. A.; Morse, J. W. Pyritization of trace-metals in anoxic marine-sediments. Geochim. Cosmochim. Acta 1992, 56 (7), 2681–2702. (28) Krumbein, W. C. Size frequency distribution of sediments. J. Sed. Petrol. 1934, 4, 65–77. (29) Folk, R. L.; Ward, W. Brazos River Bar: A study in the significance of grain size parameters. J. Sed. Petrol. 1957, 27, 3–26. (30) Salmon, S. U.; Malmstrom, M. E. Quantification of mineral dissolution rates and applicability of rate laws: Laboratory studies of mill tailings. Appl. Geochem. 2006, 21 (2), 269–288. (31) Bohlke, J. K.; Wanty, R.; Tuttle, M.; Delin, G.; Landon, M. Denitrification in the recharge area and discharge area of a transient agricultural nitrate plume in a glacial outwash sand aquifer, Minnesota. Water Resour. Res. 2002, 38 (7). (32) Larsen, F.; Postma, D. Nickel mobilization in a groundwater well field: Release by pyrite oxidation and desorption from manganese oxides. Environ. Sci. Technol. 1997, 31 (9), 2589–2595.
ES803417S
VOL. 43, NO. 13, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4857
