Chemical Geology 300-301 (2012) 123–132

Contents lists available at SciVerse ScienceDirect

Chemical Geology

journal homepage: www.elsevier.com/locate/chemgeo

Isotopic and microbiological signatures of pyrite-driven denitrification in a sandy aquifer

Yan-Chun Zhang a, Caroline P. Slomp a,⁎, Hans Peter Broers b,c,d, Benjamin Bostick e, Hilde F. Passier b, Michael E. Böttcher f,1, Enoma O. Omoregie a,g, Jonathan R. Lloyd g, David A. Polya g, Philippe Van Cappellen a,h,i a Department of Earth Sciences, Faculty of Geosciences, Utrecht University, Budapestlaan 4, 3584 CD Utrecht, The Netherlands b Deltares Research Institute, Subsurface and Groundwater, Utrecht, The Netherlands c TNO, Geological Survey of the Netherlands, Utrecht, The Netherlands d VU University Amsterdam, De Boelelaan 1085, 1081 HV Amsterdam, The Netherlands e Lamont-Doherty Earth Observatory of Columbia University, Palisades NY, USA f Max Planck Institute for Marine Microbiology, Bremen, Germany g School of Earth, Atmospheric and Environmental Sciences, University of Manchester, Manchester, UK h School of Earth and Atmospheric Sciences, Georgia Institute of Technology, Atlanta, USA i Department of Earth and Environmental Sciences, University of Waterloo, Ontario, Canada article info abstract

Article history: Denitrification driven by pyrite oxidation can play a major role in the removal of nitrate from groundwater Received 11 September 2011 systems. As yet, limited information is available on the interactions between the micro-organisms and aque- Received in revised form 23 December 2011 ous and mineral phases in aquifers where pyrite oxidation is occurring. In this study, we examine the ground- Accepted 14 January 2012 water and sediment composition along a well-characterized redox gradient in a heavily nitrate-polluted Available online 31 January 2012 pyritic sandy aquifer in Oostrum (the Netherlands) to identify the sequence of steps involved in denitrifica- δ15 − δ18 − δ34 2 − δ18 2 − Editor: B. Sherwood Lollar tion coupled to pyrite oxidation. Multi-isotope analyses ( N-NO3 , O-NO3 , S-SO4 , O-SO4 and 34 δ Spyrite) confirm that pyrite is the main electron donor for denitrification at this location. Enrichment Keywords: factors derived from the observed changes in nitrate isotopic composition range from −2.0 to −10.9‰ for Denitrification ε15N and from −2.0 to −9.1‰ for ε18O. The isotopic data indicate that pyrite oxidation accounts for approx- Pyrite imately 70% of the sulfate present in the zone of denitrification. Solid-phase analyses confirm the presence of Groundwater pyrite- and organic matter-rich clay lenses in the subsurface at Oostrum. In addition, sulfur XANES and iron Isotopes 2 − XAS results suggest the presence of a series of intermediate sulfur species (elemental sulfur and SO3 ) that Microbiology may be produced during denitrification. Consistent with geochemical analysis, 16S rRNA gene sequencing Sediment revealed the presence of bacteria capable of sulfide oxidation coupled to nitrate reduction and that are tolerant to high aqueous metal concentrations. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Pyrite oxidation leads to sulfate production and trace metal re- lease to the groundwater (Broers, 1998). This process can have a Nitrate is a common groundwater pollutant especially in regions major impact on local and regional water quality. Additional (inter- of intensive agriculture (Strebel et al., 1989). The major sources of ni- mediate) products may include nitrite and nitrous oxide (Appelo trate are manure, fertilizers, atmospheric deposition, soils and plants and Postma, 2005). Several studies have provided compelling evi- (Kendall et al., 2007). Denitrification is a natural microbial attenua- dence for the occurrence of nitrate reduction coupled to pyrite oxida- tion process (Korom, 1992) in which microorganisms use organic tion at field sites (Broers and Buijs, 1997; Pauwels et al., 1998; Postma matter or inorganic compounds as electron donors. When pyrite et al., 1991; Schwientek et al., 2008; Jørgensen et al., 2009; Tesoriero

(FeS2) acts as the electron donor, the process can be represented by et al., 2000; Zhang et al., 2009). These studies further suggest that de- the following idealized overall reaction: nitrification with pyrite can be the dominant pathway of nitrate re- moval from groundwater, even when organic matter is present. ð Þþ − þ þ→ 2þ þ ð Þþ 2− þ ð Þ 5FeS2 pyrite 14NO3 4H 5Fe 7N2 g 10SO4 2H2O 1 Recent laboratory experiments with sediment from an agricultural area have confirmed the microbial nature of the process (Jørgensen et al., 2009). However, sulfate production in pyritic aquifers is often ⁎ Corresponding author. Tel.: +31 302535514; fax: +31 302535302. E-mail address: [email protected] (C.P. Slomp). found to be lower than predicted by Eq. (1) (e.g. (Postma et al., 1 Present address: Leibniz Institute for Baltic Sea Research, Warnemünde, Germany. 1991; Miotliński1, 2008; Zhang et al., 2009). This may indicate that

0009-2541/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.chemgeo.2012.01.024 124 Y.-C. Zhang et al. / Chemical Geology 300-301 (2012) 123–132 other processes are active, such as incomplete pyrite oxidation to el- emental sulfur (Zhang et al., 2009). An alternative explanation for the mismatch between the observed changes in groundwater nitrate and sulfate with depth is that inputs of nitrogen from agriculture have de- clined with time. As a consequence, the present-day input of nitrate 14.0 to the zone of pyrite oxidation may not directly correspond to the

15.01 amount of sulfate produced at depth. 5 .0 Stable isotope analysis is a powerful tool to identify the sources and sinks of bioactive compounds in natural environments, because the 16.0 sources may have characteristic isotopic signatures and the isotopes are often strongly fractionated during biogeochemical transformations. 38 15 − 34 2− This holds for the commonly studied δ N-NO3 and δ S-SO4 ,aswell 17.017. as for the oxygen isotope signatures of nitrate and sulfate. The simulta- 0 neous determination of stable isotope ratios of multiple elements may 40 significantly strengthen the interpretation in terms of provenance and 41 18.01 transformation processes (Böttcher et al., 1990, 2001; Deutsch et al., 8.0 2006; Schwientek et al., 2008; Pauwels et al., 2010). For example, 1km 15 − Böttcher and coworkers showed that groundwater δ N-NO3 and 18 − 42 19.0 Forest δ O-NO3 values in a catchment area with high nitrate input were con- sistent with denitrification coupled to pyrite oxidation (Böttcher et al., 20.0 Farmland 1990). Similarly, isotope data of groundwater sulfate and sediment Town 7649 sulfur compounds for a carbonate-containing organic-poor aquifer suggest that pyrite oxidation is the dominant source of sulfate in the Fig. 1. Well locations at Oostrum. Wells 40, 41 and 42 are located under farmland, groundwater system (Schwientek et al., 2008). whereas well 38 is located in a forested area. The arrow indicates the general ground- water flow direction (Zhang et al., 2009). Two sediment cores were drilled at a distance The latter study also illustrates the need for combined studies of of ~3 m west (Core 1) and ~200 m northeast (Core 2) of well 41. the groundwater and solid phase composition for a correct under- standing of spatial and temporal trends in microbially-mediated pro- cesses in aquifers. Within this context, detailed sediment speciation farmland, well 38 is located in a nearby forested area (Broers and analyses with techniques such as X-ray spectroscopy are of particular Buijs, 1997). The investigated aquifer is unconfined and has an ap- value because they allow potential solid phase intermediates of pyrite proximate thickness of 45 m. Impermeable clay layers at its bottom oxidation to be identified (Ziegler et al., 2009). Detailed microbial clearly define the base of the aquifer. The main aquifer is usually sit- analysis of aquifer sediments can provide complimentary information uated between 15 and 45 m depth and consists of coarse fluvial on biogeochemical processes by providing insight into the organisms sands of Pliocene age with, in the upper part, unconnected clay that mediate these processes. Such studies have been performed for lenses, which often have high organic matter contents. Across the heterotrophic and autotrophic denitrifiers in marine sediments and area, there are several faults, which are revealed by horizontal discon- laboratory incubations. For example, Brettar et al. (2006) identified tinuities in hydraulic head. The depth, thickness and grain size distri- autotrophic denitrifiers at the oxic-anoxic interface in the Baltic Sea bution of the intermediate clay layer is spatially variable, indeed, the (Brettar et al., 2006). Various laboratory studies suggest Thiobacillus clay is sometimes absent in parts of the area. Thin centimeter thick denitrificans is capable of coupling denitrification to pyrite oxidation lenses of clay are present in the top of the Pliocene aquifer, just (Jørgensen et al., 2009; Torrentó et al., 2010). However, studies iden- below the clay layer which separates the Pliocene and Pleistocene se- tifying the bacteria responsible for denitrification coupled to pyrite quence. The Pliocene sequence is overlain by ~10 m of coarse fluvial oxidation in aquifer sediments from field sites are not yet available. sands and gravels of Pleistocene age which were deposited by the 34 2− 18 2− Here, we combine multi-isotope analyses (i.e., δ S-SO4 , δ O-SO4 , Meuse river and ~5 m of eolian sand deposits. The groundwater 15 − 18 − 34 δ N-NO3 , δ O-NO3 , and sediment δ Spyrite) with microbial analysis table is about 3–5 m below the land surface. Groundwater sampling of the sediment (16S rRNA gene amplification and sequencing) and was carried out in 1996 and 2006 and is described in detail elsewhere detailed field data of the aqueous and solid phase geochemistry for a (Zhang et al., 2009). Here, we focus on the results for 2006. Two sed- sandy aquifer. Our study site Oostrum (Netherlands) is characterized iment cores were drilled in 2007 at this site, at a distance of ~3 m by a nitrate-rich (up to 8 mM) upper layer overlying deeper, sulfate- west and ~200 m northeast of well 41. These cores are referred to rich (up to 4 mM) groundwater. Earlier work provides strong evi- as core 1 and 2, respectively. The cores were split and sub-sampled dence for denitrification coupled to pyrite oxidation at this location. under Ar. Samples from 28 depth intervals from both cores were This conclusion was based mainly on observed changes in groundwa- freeze-dried and analysed for total Fe with X-ray Fluorescence and ter nitrate and sulfate with depth in the aquifer (Zhang et al., 2009). for total C and S with a LECO SC 144DR. The goal of this current study is to better characterize the biogeochem- Solid phase samples from six depths in the aquifer for core 1 and 2 ical processes occurring along the redox gradient in the aquifer by were studied in detail (~16 m, ~18 m and ~23 m in core 1 and 11, 13 means of isotopic analyses and to gain insight into the microorgan- and 15 m in core 2) for their microbiological characteristics and Fe isms that are potentially involved in the processes affecting nitrogen and S chemistry. These samples are all from the part of the aquifer and sulfur in the aquifer. where major changes in groundwater nitrate and sulfate are ob- served. Subsampling was performed directly after drilling and open- 2. Materials and methods ing of the cores in a glovebox under argon. Samples for microbial analysis were stored at −20 °C. The other samples were stored at 2.1. Study site and sampling 4 °C under argon. For solid Fe and S speciation analysis, small aliquots of homogenized sediments were mixed with mineral oil in a Oostrum is located in an intensive agricultural area in the south of nitrogen-filled glovebox and refrigerated under nitrogen. To confirm the Netherlands (Limburg) (Fig. 1). The site covers approximately that these preservation methods were successful, moist mackinawite 1km2. Multi-screen sampling wells were installed roughly along the (nominally FeS, which reacts rapidly with oxygen) was treated in the direction of groundwater flow: wells 40, 41 and 42 are located in same manner, and its integrity following storage was confirmed by X- Y.-C. Zhang et al. / Chemical Geology 300-301 (2012) 123–132 125 ray absorption spectroscopy. Subsamples for S-isotope analysis were and internally calibrated using a metallic Fe reference foil freeze-dried and stored under argon. (7112.0 eV). The k3-weighted iron EXAFS spectra (spline from k=2.5–14) were fit over the k range 2.55–12.5 using the k3-weighted 2.2. Analytical methods chi functions of reference spectra. Pyrite, magnetite, mackinawite, hematite, goethite, ferrihydrite, biotite and hornblende model com- 2.2.1. Isotope analysis pounds were used in the final fits. Biotite and hornblende are represen- Groundwater major anion and cation analyses and 3H/3He tative of Fe silicates rather than explicit mineral components. Other groundwater age dating methods are described elsewhere (Zhang et common Fe minerals were also considered, but were not necessary to 15 − 18 − al., 2009). Samples for δ N-NO3 and δ O-NO3 analyses were pre- fit the data. Iron XANES spectra were fit over the range 7105 to pared using the AgNO3 method (Chang et al., 1999; Silva et al., 7150 eV, to confirm results obtained by EXAFS fitting or where 2000; Xue et al., 2009). The samples were prefiltered with 0.45 μm EXAFS spectra were not available, and were similarly accurate for membrane filters before passing through cation and anion exchange distinguishing ferric from ferrous phases. resin columns for nitrate purification and extraction. HCl and Ag2O were then added and AgCl was removed by filtration. The solution 2.2.4. Microbial community analysis was freeze-dried producing solid AgNO3. AgNO3 was converted to Sub-sections of mixed sediment (100 mg) within the sediment 15 18 N2 gas for δ N analysis. Finally, CO2 for δ O analysis was produced redox transition zone in both drill cores were used for DNA extrac- by combusting AgNO3 with graphite. Isotope analyses were carried tions. Total genomic DNA was extracted using the PowerSoil DNA ex- out by C-irmMS, and stable N and O isotope results are presented in traction kit (MOBIO, Carlsbad, USA), and the crude extracts were the usual δ-notation versus air, V-SMOW, and V-CDT, respectively. further purified using the Wizard DNA Clean-Up Kit (Promega, Mad- Dissolved sulfate was precipitated quantitatively from acidified ison, USA). The 16S rRNA gene was amplified from bacteria using uni- samples as BaSO4 by the drop-wise addition of 5% BaCl2 solution for versal bacterial primers. PCRs, cloning and sequencing were 34 2 − 18 2 − further determinations of δ S-SO4 and δ O-SO4 . The solid was performed according to the approach described by Niemann et al. filtered through a 0.45 μm membrane filter, washed and dried at (Niemann et al., 2006). Sequence analysis was carried out using the

60 °C in a drying oven. For oxygen isotope analysis the BaSO4 precip- ARB (Ludwig et al., 2004) software package with the Silva 98 release itate was heated further in a porcelain crucible to 500 °C to remove database (Pruesse et al., 2007) and the Mothur software package potential contaminants. Reduced inorganic sulfur compounds (essen- (Schloss et al., 2009). Sequences generated in this study have been tially pyrite) in freeze-dried drill core sediments from cores 1 and 2 deposited in the Genbank database, and are accessible under the were extracted by hot acidic Cr(II)Cl2 solution (Canfield et al., 1986; following accession numbers HM641428-HM641687. Fossing and Jørgensen, 1989). The liberated hydrogen sulfide was 34 trapped quantitatively as Ag2S. Sulfur isotope measurements (δ S) 3. Results and discussion were carried out using a Thermo Finnigan Deltaplus gas isotope mass spectrometer coupled via a Thermo Conflo split interface to an Euro 3.1. Pore water profiles of aqueous components EA elemental analyzer. For oxygen isotope measurements barium sul- fate was combusted in a Thermo Quest TC-EA. Sulfur and O isotope The depth distributions of nitrate in the agricultural area at our ratios are reported in the conventional delta-notation versus V-CDT study site are characterized by a concentration maximum between and V-SMOW, respectively. Replicate measurements agreed within 10 and 15 m below the soil surface, as shown for well 41 in 2006 ±0.3‰ and ±0.5‰ for S and O isotope measurements, respectively. (Fig. 2). Most denitrification takes places below this maximum in ni- Several international intercomparison materials and in-house stan- trate. Here, a sharp decline in nitrate levels coincides with an increase dards were measured to calibrate the mass scales. in sulfate concentrations, down to a depth of ~23 m. The depth inter- val in which nitrate and sulfate exhibit opposing gradients is hereaf- 2.2.2. Solid sulfur XANES analysis ter referred to as the “reaction zone”. As shown previously, aqueous Solid S speciation was analyzed by S K-edge X-ray near-edge concentrations of Fe2+ and trace metals increase strongly within structure (XANES) spectroscopy at the National Synchrotron Light the reaction zone (Zhang et al., 2009). Below the reaction zone, ni- Source, Brookhaven National Laboratory, New York, on beamline trate concentrations are near or below detection, while sulfate con- X19A. Samples on filter paper were analyzed using a 1×7 mm centrations progressively decrease with depth. Qualitatively, the beam, in a He-purged atmosphere. The sulfate white line peak was compositional changes in the reaction zone are consistent with calibrated to 2483.1 eV. Sample fluorescence was measured using a those expected for denitrification coupled to the oxidative dissolution PIPS (passivated implanted planar silicon) detector. Data was of pyrite. background-corrected and normalized prior to least-squares fitting with known standards in WinXAS, over the range 2466 to 2487 eV. 3.1.1. Isotopic analysis of coupled denitrification-pyrite oxidation Data were best fit using reference spectra for sulfate, sulfite, elemen- Within the reaction zone under farmland where denitrification tal sulfur, pyrite, and iron sulfide (mackinawite). Organic sulfur com- driven by pyrite oxidation occurs in the aquifer, the average values 15 − 18 − pounds also were considered but were not present in sufficient of δ N-NO3 and δ O-NO3 increase with depth from 7.2‰ to quantities to affect fitting and thus are omitted in final fits. Speciation 18.2‰ and from −2.1‰ to 7.4‰, respectively (Fig. 3). These increases of S by K-edge XANES is possible for sediment with S contents down are qualitatively consistent with the preferential utilization of lighter to 0.01 wt.% (Prietzel et al., 2011). isotopes by denitrifying organisms. Isotope enrichment factors for ni- 15 − 18 − trate can be calculated from the changes in δ N-NO3 and δ O-NO3 , 2.2.3. Solid iron XAS analysis assuming that denitrification is the only fractionating process, that Iron speciation of solids was investigated by X-ray absorption water parcels can be treated as closed systems, and that the fraction- spectroscopy at the Stanford Synchrotron Radiation Laboratory, ation factor remained constant. The Rayleigh equation (Mariotti et al., beamline 11–2, which is configured with a Si(220) monochromator 1988; Böttcher et al., 1990) can then be used: and a phi angle of 0°. Sample fluorescence was measured with a 30- δ ¼ δ þ ε = ð Þ element Ge detector in combination with a 6 μmMnfilter. The r 0 ln C C0 2 beam was detuned as needed to reject higher-order harmonic fre- quencies and prevent detector saturation. Spectra were background- where δr and δ0 are initial and residual isotope values, ε is the isotopic corrected and normalized using linear pre- and post-edge functions, enrichment factor and C/C0 is the remaining fraction of the nitrate. 126 Y.-C. Zhang et al. / Chemical Geology 300-301 (2012) 123–132

34 2− 18 2 − 34 15 − 18 Fig. 2. Example of the depth distributions of a) dissolved nitrate and sulfate in groundwater; b) δ S-SO4 , δ O-SO4 and δ S in pyrite in solid phase and c) δ N-NO3 and δ O- − NO3 in the agricultural area (well 41). In the “reaction zone” (grey), nitrate concentrations decrease sharply whereas sulfate concentrations increase, suggesting denitrification coupled to pyrite oxidation. The depth intervals above and below the “reaction zone” are referred to as the “shallow part” and “deeper part”, respectively.

18 2 − As initial nitrate concentration (Co) we impose the maximal value at and δ O-SO4 from −0.7‰ to 15.3‰ (Table S1). The two major the top of the reaction zone. Then, using the pair-wise concentration sources of groundwater sulfate in the Oostrum area are sulfate leach- and δ values measured in the reaction zone, we obtain values of ing from the land surface (where it is introduced via atmospheric −10.9‰ (well 41) and −2.0‰ (well 42) for ε15N, and −9.1‰ (well input, manure and fertilizers) and sulfate from pyrite oxidation, 41) and −2.0‰ (well 42) for ε18O (Note: well 40 lacked sufficient ni- with no important sinks (Zhang et al., 2009). Sulfur isotopes undergo trate containing samples to perform the isotope analyses). essentially no isotope fractionation during pyrite oxidation to sulfate 34 2 − Culture experiments with heterotrophic denitrifiers have yielded (Stempvoort Van et al., 1994) Hence, the groundwater δ S-SO4 sig- wide ranges of enrichment factors: from −10‰ to −39‰ for ε15N, nature should only reflect the relative contributions of the sulfate and between −15‰ and 40‰ for ε18O(Lehmann et al., 2003; sources, according to the following mass balance (Moncaster et al., Toyoda et al., 2005). Field studies of denitrification suggest compara- 2000; Smith et al., 2009): ble ranges of ε15N and ε18O for autotrophic and heterotrophic denitri- fication. For example, Böttcher et al. (1990) reported values of −16‰ 34 34 C δ S ¼ C δ S for ε15N and −8‰ for ε18O for autotrophic denitrification coupled to sulfate total sulfate total sulfate input sulfate input þ C δ34S ð3Þ pyrite oxidation in a sandy aquifer in Germany, while Mariotti et al. sulfate pyrite pyrite (1988) estimated a ε15N value of ~−5‰ for heterotrophic denitrifica- tion in an aquifer of northern France. Although the enrichment factors where Csulfate_total =Csulfate_input +Csulfate_pyrite and Csulfate_total is exhibit large ranges, the δ15N:δ18O ratios reported in the literature for the measured (total) sulfate concentration in a groundwater sample, denitrification are mostly comprised between 0.9 and 2.1 (Böttcher Csulfate_input is the sulfate concentration derived from the land surface 34 et al., 1990). Our value for the wells located in the agricultural area and Csulfate_pyrite is that derived from pyrite oxidation. δ S sulfate_total, 34 34 (~1, Fig. S1) falls within this range. δ Ssulfate_input and δ Spyrite are the δ values of the groundwater, the The isotopic composition of dissolved sulfate in groundwater is historical input derived from the land surface and sedimentary pyrite, 34 2 − also quite variable, with δ S-SO4 ranging from −4.8 to 45.3‰ respectively.

15 − 18 − 34 2− 18 2 − Fig. 3. Box plots of δ N-NO3 , δ O-NO3 , δ S-SO4 and δ O-SO4 in the three depth zones in the agricultural area (wells 40, 41 and 42) in 2006. See Fig. 1 for the definition of the depth zones. Thick line: median; upper and lower boundary of the boxes: 25 and 75 percentile; whiskers: range; bullets: outliers; n: number of observations. Y.-C. Zhang et al. / Chemical Geology 300-301 (2012) 123–132 127

18 2− Assuming that the most shallow groundwater samples collected (Fig. 2 and Table S1; Zhang et al., 2009). The mean value of δ O-SO4 from wells 40, 41 and 42 are representative of surface input in the ag- in the deeper groundwater (4.9‰) also shows an increase compared 34 ricultural area, the average δ Ssulfate_input value is around 2.6‰ to the reaction zone. Taken together, the decrease in sulfate concentra- 34 (Fig. 3). Then, assigning a value to δ Spyrite of −4.2‰, which is the tion and the dual-isotope fractionation implies active sulfate-reducing mean value measured on the samples of drill core 1 which is located microorganisms below the zone of denitrification (Mizutani and adjacent to well 41 (Table S2), we can calculate the relative contribu- Rafter, 1973; Fritz et al., 1989; Böttcher et al., 1998). tions of pyrite oxidation to the measured groundwater sulfate con- centrations (i.e., Csulfate_pyrite/Csulfate_total). Input of sulfate leaching 3.1.3. Time scales from the land surface accounts for >95% of total sulfate in the Groundwater age data can provide important information on groundwater above the reaction zone. Across the reaction zone, how- groundwater flow paths and the time scales of biogeochemical pro- 34 ever, the average δ S sulfate_total drops to −2.1‰, implying a signifi- cesses in the subsurface, especially when coupled to groundwater cant contribution of lighter sulfate produced during pyrite concentrations of key elements and the historical input of concentra- oxidation, on the order of 70%. Below the reaction zone, Eq. (3) pre- tions in recharge (Visser et al., 2009; Zhang et al., 2009). For instance, 15 − 34 2 − dicts a decrease in the contribution of pyrite oxidation, which is due the simultaneous increase in δ N-NO3 and decrease in δ S-SO4 is to microbial sulfate reduction (see below). restricted to groundwater that entered the aquifer between 1985 and Sulfur isotope data for pyrite can vary over a wide range (Canfield 1995 (Fig. 4). Thus, it takes a maximum of 10 years for denitrification et al., 2005). For example, pyrite analyzed in core 2, which is located to remove the incoming nitrate from the groundwater. In contrast, 34 2 − ~200 m to the north west of well 41 (Fig. 1), has a much heavier iso- the general increase in δ S- SO4 with age beyond 1985 spans at tope composition at shallow depths (Table S2). The variation likely least three decades. Given the corresponding relatively small de- reflects heterogeneity of δ34S-pyrite in this Pliocene aquifer. However, crease in the sulfate concentration (Fig. 2), this points to much such high δ34S values in pyrite are not reflected in the groundwater lower rates of sulfate reduction compared to denitrification. composition of the wells at our study site. In addition to nitrate, molecular oxygen is a common electron ac- 3.2. Solid-phase analysis ceptor for pyrite oxidation in natural settings. The oxygen isotopic composition of sulfate can help distinguish between the potential Depth profiles of total sulfur, carbon and iron for drill cores 1 and 2 electron acceptors (e.g., (Taylor et al., 1984; Tuttle et al., 2009). confirm the strong heterogeneity of the sediment in the aquifer at According to the following reaction stoichiometry Oostrum observed visually (Fig. 5). In drill core 1, sediment layers enriched in total carbon, sulfur and iron and in fine-grained material ð Þþ þ → 2þ þ 2− þ þ ð Þ b μ – 2FeS2 pyrite 7O2 2H2O 2Fe 4SO4 4H 4 ( 63 m) are present at depths of ~14 15 and 21 m below the land surface. In drill core 2, similar layers are present at depths of 7/8 of the oxygen in sulfate is derived from molecular oxygen where- ~12–14, 16 and 21 m. Both drill cores show the ~2 m thick clay as the remaining oxygen comes from water. We can then use an layer at the top of the Pliocene Formation, which corresponds with δ18 2− δ18 2− isotope balance for O-SO4 to calculate the expected O-SO4 the samples with the highest proportion of the size fractionb63 μm. values when pyrite is oxidized by O2. Based on literature data, we assign The deeper samples with a larger fraction b63 μm coincide with inter- δ18 δ18 ‰ − ‰ OO2 and OH2O values of 23 (Tuttle et al., 2009)and 7 (Clark vals where thin clay lenses exist. The reaction zone identified from ε 18 and Fritz, 1997), respectively and enrichment factors of O(SO4-O2) the water chemistry data coincides with the top of the Pliocene aqui- ε 18 − ‰ ‰ and O(SO4-H2O) of 11 and 4 (Taylor et al., 1984; Toran and fer, in which the higher proportions of clay and silt material, organic δ18 2− Harris, 1989). The predicted groundwater O-SO4 in the reaction matter and sulfur are concentrated. Sampling sediments from such a zone should then be on the order of 10‰. This value is well outside heterogeneous system implies that the samples cannot be treated as the range of our measurements in the reaction zone in the agricultural being representative for a gradient in solid phase chemistry. Instead, area, which show a maximum value of 2.8‰ (Table S1). Therefore, we we consider the 6 samples that were used for XANES/XAS analysis can eliminate O2 as an important oxidant for pyrite at the Oostrum to be illustrative rather than to be exhaustive. site. The isotopic evidence is entirely consistent with the presence of a Insight in the mineralogical composition of the sediment from 6 10 m thick upper aquifer with low to negligible oxygen concentrations selected depth intervals in the reaction zone (at depths of ~16, ~18 above the reaction zone (Zhang et al., 2009). and ~23 m in core 1 and 11, 13 and 15 m in core 2) is obtained A similar calculation demonstrates that H2O molecules cannot be from sulfur XANES and iron XAS analysis (Table 1). In general, the the main source of oxygen in sulfate either. If all sulfate oxygen was fraction of Fe present in silicates is variable and reflects differences δ18 2 − − ‰ derived from H2O, O-SO4 should be on the order of 1 in in primary mineralogy and sediment sources. In these samples, sili- the reaction zone, compared to the observed mean value of 1.5‰ cates (e.g. glaucophane, biotite or glauconite) represent a minor frac- and a minimum value of 0.2‰. If we assume that nitrate is the sole tion (5–34%) of the total Fe. Here, we restrict our discussion to non- oxygen source for sulfate produced by pyrite oxidation (Eq. (1)), silicate Fe phases to effectively focus on the reactants and products ε18 the isotope balance yields a corresponding O(SO4-NO3) value of of redox reactions in the aquifer. 3‰. There are currently no other experimental data on oxygen iso- The 6 samples contain between 0.1 and 0.7% of Fe (Fig. 6, Table 1). tope fractionation during pyrite oxidation coupled to nitrate reduc- The amount of pyrite increases strongly at depth in both cores, which fi ε18 tion to which this eld-derived estimate of O(SO4-NO3) can be is in general agreement with the redox gradient. Note, however that compared. there is substantial heterogeneity in the fraction b63 μm, in S and or- ganic C between the individual samples (Fig. 5). The samples from the 3.1.2. Isotopic evidence for microbial sulfate reduction upper part of the reaction zone (16 and 11 m in cores 1 and 2, respec- Bacterial dissimilatory sulfate reduction generates a typical en- tively) are dominated by reduced ferrous and ferric iron oxide phases 34 2 − 32 2 − richment of δ S-SO4 in the residual sulfate (δ S-SO4 )(Kaplan and contain little pyrite Fe. The samples from intermediate depths and Rittenberg, 1964; Fritz et al., 1989; Böttcher et al., 2001; Habicht (13 and 18 m in cores 1 and 2, respectively) have lower iron contents and Canfield, 2001). Such a trend is observed in the isotope data and show a mixture of ferric and ferrous iron and pyrite Fe. The dee- 34 2 − from deeper parts of the aquifer (>20 m) where the δ S-SO4 pest samples analysed (23 and 15 m in cores 1 and 2, respectively) values (average of 8.9‰ with a maximum of 45.3‰) markedly exceed are dominated by pyrite Fe and contain about 30% of ferric and fer- 34 2 − those in the reaction zone (Fig. 3). The trend of increasing δ S-SO4 rous oxides. In well 40, close to core 2, groundwater at 15 m depth values with depth corresponds to decreasing sulfate concentrations still contained some nitrate in 1996 (Zhang et al., 2009) whereas 128 Y.-C. Zhang et al. / Chemical Geology 300-301 (2012) 123–132

15 − 34 2 − 15 − Fig. 4. Values of δ N-NO3 and δ S-SO4 plotted versus groundwater age for wells 40, 41 and 42, which are all located in an agricultural area. The increase in δ N-NO3 and decrease in 34 2− 34 2− δ S-SO4 during the period 1985 to 1995 implies denitrification coupled to the oxidation of isotopically light pyrite. In older groundwater, the general increase in δ S-SO4 with age is 34 2− interpreted to reflect sulfate reduction. The anomalously low δ S-SO4 values in the groundwater older than 1970 observed in well 42 just below the reaction zone are probably due to local heterogeneity of the aquifer. nitrate was absent in 2006. Thus, in this part of the aquifer pyrite (Zhang et al., 2009). Together with the groundwater geochemistry oxidation with nitrate likely occurred in the past. This may explain data, the abundant presence of non-silicate Fe[III] mineral suggests the presence of ferrous iron at this depth. The sample from 23 m that, besides pyrite oxidation producing sulfate and Fe[II] (Eq. (1)), depth in core 1, in contrast, is located well below the reaction zone further oxidation of Fe(II) by nitrate (Eq. (5)) to Fe(III) could occur: where nitrate never has been present. Sulfur contents in the 6 samples roughly follow the pattern of the ½ þ − þ → þ : þ þ ð5Þ 5Fe II NO3 7H2O 5FeOOH 0 5N2 9H pyrite Fe contents, with almost no sulfur being present in samples 16 and 18 m depth in core 1 and from 11 m depth in core 2 (Fig. 6). The Previous field studies suggest that oxidation of Fe[II] to Fe[III] other 3 samples show significant amounts of sulfur, dominated by with nitrate (Eq. (5)) is less important for nitrate removal than pyrite-S and SO4-S. The presence of sulfate mineral throughout the sulfide (pyrite) oxidation (Postma et al., 1991; Zhang et al., 2009). reaction zone was unexpected (Table 1) and may be explained by In addition, the observed rise in pH in the reaction zone (Zhang et the partial oxidation of sulfide minerals in storage, but may also indi- al., 2009) is consistent with a more important role for pyrite oxidation cate the presence of solid phase sulfur (in adsorbed or mineral form) with nitrate following Eq. (1) at this depth. It is possible that the sul- in the aquifer. The sulfur speciation also reveals the presence of pre- fur intermediates which are found in many of the 6 samples are 2 − viously unidentified sulfur species (elemental sulfur and SO3 ). For formed as minor products through one or more separate reactions example, these intermediate oxidation-state species represent >90% coupling pyrite oxidation and denitrification. Further work would of the total S in the sample from 18 m depth in core 1 (Table 1) be needed to quantify their occurrence over the whole depth profile However, total S in this sample is low (Table 1). in this heterogeneous aquifer and to assess their potential role in con- Aqueous concentrations measured in the reaction zone suggest tributing to possible variations in the stoichiometry of the overall re- that the bulk of the denitrification coupled to pyrite oxidation is asso- actions as suggested for this site (Zhang et al., 2009). Alternatively, ciated with an increase of Fe[II] and sulfate and decrease of nitrate these intermediates also could be the major reaction products of

Fig. 5. Depth profiles of total sulfur, carbon and iron (in wt.%) and the grain size fraction b63 μm in sediment drill cores 1 and 2. The depths of samples used for sulfur XANES and iron XAS in core 1 and 2 are indicated with open circles. The lithology is indicated in the two borehole descriptions. A large fraction of material with a grain size b63 μm indicates clay and silt layers, some of which are 2 m thick, whereas others are only millimeters or a few cm thick. Y.-C. Zhang et al. / Chemical Geology 300-301 (2012) 123–132 129

Table 1 Weight percentages of iron and sulfur and their speciation as determined with XAS and XANES in samples from 3 depth intervals in Core 1 and Core 2.

Core Depth Total Fe Fe-silicate minerals Fe (II) oxides Fe (III) oxides Pyrite Pyrite (m) (wt.%) (% of total Fe) (% of total Fe) (% of total Fe (% of total Fe) (wt.% based on total Fe)

1 16 0.20 26 43 16 15 0.03 1 18 0.09 21 18 59 2 0.00 1 23 0.42 18 25 2 55 0.23 2 11 0.73 34 26 37 3 0.02 2 13 0.16 14 32 26 28 0.04 2 15 0.43 5 0 32 63 0.27

2− Core Depth Total S FeS Pyrite S(0) SO3 SO4 mineral Pyrite (m) (wt.%) (% of total S) (% of total S) (% of total S) (% of total S) (% of total S) (wt.% based on total S)

1 16 0.11 0 6 0 0 94 0.01 1 18 0.02 0 0 32 60 8 0.00 1 23 1.37 5 40 4 0 51 0.55 2 11 0.01 0 13 2 0 85 0.00 2 13 0.10 0 17 8 17 58 0.02 2 15 0.28 3 29 5 0 63 0.08 pyrite oxidation by nitrate, with the elemental sulfur being oxidized geochemical processes by relating both sediment mineralogy and or disproportionated by autotrophs leading to sulfate as one end- porewater geochemistry to genetic characterizations of the microbial product (Bak and Cypionka, 1987; Thamdrup et al., 1994; Finster, community. 2008). Between 40 and 45 16S rRNA gene sequences were obtained from The high contents of pyrite-Fe and pyrite-S in the samples at 15 m each of 6 samples in the reaction zone from drill cores 1 and 2 depth (Core 2) and 23 m depth (Core 1) coincide with the depth (Table 2). Organisms closely related to heterotrophic nitrate reducing range where nitrate concentrations are below the detection limit bacteria, such as Massilia brevitalea, (Zul et al., 2008), Herbaspirillum and dissolved sulfate concentrations are near peak values (~3 mM, autotrophicun (Ding and Yokota, 2004) and Cupriavidus metallidurans Fig.1 and Zhang et al., 2009). It is likely that no nitrate at this depth (Goris et al., 2001) were detected in the sediments from both cores. is consumed by pyrite oxidation, leaving all pyrite in the sediment The presence of these bacteria suggests heterotrophic denitrification matrix intact. The major redox reaction taking place below this cannot be excluded completely and these bacteria are either carrying depth range is likely sulfate reduction, as indicated by the depth out heterotrophic denitrification at a low rate or are involved in other 34 2 − trends in δ S-SO4 in the groundwater of wells 40 and 41 (Fig. 2 processes associated with the slow oxidation of recalcitrant sedimen- and Table S1). tary organic carbon. Note that heterotrophic denitrification would be expected to more favorable than autotrophic denitrification based on 3.3. Linking geochemistry to microbial community composition theoretical energy yields (Devlin et al., 2000). However, the pyrite in this system is likely more reactive than the organic matter allowing The solid-phase and aqueous geochemistry indicate that the aqui- autotrophic denitrification to dominate”. fer at this site shows a distinct reaction zone within an overall strong- The porewater and solid phase data indicate drastic changes in the ly heterogeneous sediment matrix. While in the samples from the chemical forms and concentrations of N, Fe and S with depth that are upper and middle part of the reaction zone Fe oxides are the most consistent with denitrification linked to pyrite oxidation (Zhang et al., abundant non-silicate Fe-phases, pyrite is quantitatively most impor- 2009). While microbially mediated nitrate-dependent sulfide (pyrite) tant in the lower samples. Here, micro-organisms are linked to oxidation is well-known, microbial evidence for a similar process (by

Fig. 6. Distribution of Fe (non-silicate) minerals and S minerals in solid phase in fresh sediment from the drill cores 1 and 2. The fractions of Fe and S minerals were determined with XAS (for Fe) and XANES (for S). For the solid phase total Fe (wt.%) and the total S (wt.%) see Table 1. 130 Y.-C. Zhang et al. / Chemical Geology 300-301 (2012) 123–132

Table 2 Breakdown of 16S rRNA gene sequences for three different depths [m] in the drill core 1 and core 2.

Phylogentic groupa Closest relativeb % Identityc Core 1 Core 1 Core 1 Core 2 Core 2 Core 2

16 m 18 m 23 m 11 m 13 m 16 m

Janthinobacterium/ Massilia brevitalea, EF54677 99–95 15 2 13 11 13 Oxalobacter Herbaspirillum Herbaspirillum autotrophicum, AB074524 97–962 678 Ralstonia/Cupriavidus Cupriavidus metallidurans CH34, Y10824 99–97 16 1 5 10 13 Acidovorax BrG1, U51101 99–94442 Thiobacillus Thiobacillus denitrificans, AJ243144 98 1 Thiothrix Thiothrix nivea, L40993, Thiothrix unzii, L79961 99,98 2 Pseudomonas Pseudomonas sp. QZ1, EF542804 99–97 3133 Acinetobacter Acinetobacter baumannii, EF672505 99–97 23 Sphingomonas ––2 Gallionella Gallionella ferruginea, L078797 94–96 2 1 0 Other Gammaproteobacteria ––12 2 1 Methylobacterium Methylobacterium thiocyanatum, U58018 99–971 215 Other Alphaproteobacteria ––1 411 Deltaproteobacteria ––18 1 Acidobacteria Acidobacteriaceae bacterium TAA48, AY587229 87–72 1 8 11 Nitrospirae Thermodesulfovibrio thiophilus, AB231857 86–85 3 1 Actinobacter – 24 21 Firmicutes ––2 Bacteriodetes ––21 Total number of sequences 45 40 43 46 42 43

(a) Phylogenetic affiliation; (b) Closest type strain or cultured representative; (c) Sequence identity to closest type strain or cultured representative.

Thiobacillus denitrificans) with pyrite has only recently been pub- 4. Conclusions lished (Jørgensen et al., 2009). Bacteria closely related (99–98%) to sulfide-oxidizing, denitrifying species, such as Thiobacillus denitrificans Denitrification coupled to pyrite oxidation is identified as the prin- (Kelly and Wood, 2000), Thiotrix nivea (Larkin and Shinabarger, 1983) cipal biogeochemical process causing large changes in nitrate and sul- and Thiotrix unzii (Polz et al., 1996) were present at a depth of 23 m in fate concentrations in groundwater below agricultural fields at our core 1. Furthermore, bacteria closely related to the sulfide-oxidizing, field site, at Oostrum (the Netherlands). Isotopic data for sulfate 34 2 − 18 2 − 15 − 18 denitrifying bacterium Pseudomonas stutzeri (Mahmood et al., 2009) (δ S-SO4 , δ O-SO4 ), nitrate (δ N-NO3 , δ O-NO3) and pyrite 34 were detected within the same interval of core 1 and in all 3 intervals (δ Spyrite) confirm our earlier work (Zhang et al., 2009) that pyrite analysed for core 2. The presence of other solid-phase products in the is the main electron donor for denitrification at this location and reaction zone, i.e. elemental sulfur and sulfite suggests that some of that sulfate reduction occurs at depth. Solid phase analyses confirm the S is incompletely oxidized and that the complete oxidation of pyrite the highly heterogeneous distribution of organic matter and pyrite- to sulfate involves an additional step (and possibly an additional rich clay lenses in the aquifer. Iron XAS and sulfur XANES confirm microbe) that can oxidize these species. Organisms closely related to the presence of pyrite and suggest that various intermediate sulfur 2 − Methylobacterium thiocyanatum (Anandham et al., 2009)(99–97%) are species (elemental sulfur and SO3 ) are formed in the zone of deni- present in many of the samples. This last organism and its relatives trification. Observed differences in microbial populations with depth are known to oxidize intermediate nitrogen and sulfur compounds are suggestive of a series of reactions along the redox gradient in such as thiocyanate and thiosulfate (Anandham et al., 2009). These or- this aquifer. Besides sulfide (pyrite) oxidation with nitrate to sulfate, ganisms may play a role in pyrite oxidation coupled to denitrification, these processes include, heterotrophic denitrification (at a very slow by recycling reduced or intermediate sulfur compounds that are rate), incomplete oxidation of sulfide to elemental sulfur and Fe(II) produced upon the interaction between aqueous solutions and sulfide oxidation with nitrate. The microbial populations are also related to minerals (Schippers and Jørgensen, 2002). Bacteria closely related to metal-tolerant bacteria which is consistent with the high concentra- Gallionella ferruginea, which is known to oxidize Fe[II] (Hallbeck et al., tions of metals in the groundwater at this location. 1993), were detected along with bacteria closely related to the iron oxidizer BrG1(Straub et al., 1996) in the samples from core 1 Acknowledgments (Table 2). These organisms are likely responsible for Fe[II] oxidation, possibly with nitrate. Financial support was provided by the Netherlands Organisation Over 50% of the sequences recovered from these sediments were ei- for Scientific Research (NWO-water and Vidi grant 86405.004 to ther closely related to those of bacterial genera that are often detected in CPS). The authors thank A. Visser, J. Rozemeijer, A. Dale, D. van de environments with elevated metal/metalloid concentrations (such as Meent, H. de Waard, G. Nobbe, G. Laruelle, M. Verheul and G. Klaver observed in natural waters impacted by mining or waste water), or to for help with the field work. We thank TNO for making available those of genera with members that have shown a specific tolerance to data from an earlier project. We also thank the UU-TNO lab for elevated metal/metalloid concentrations, e.g. Acinetobacter, Cupriavidus, laboratory analysis. EO acknowledges support from the European Herbaspirillum, Janthinobacterium, Methylobacterium, Oxalobacter, Commission through the FP6 Mobility Actions AquaTRAIN (Contract Pseudomonas,andRahnella (Hery et al., 2003; Mergeay et al., 2003; no. MRTN-CT-2006-035420). MEB wishes to thank R. Rosenberg Reardon et al., 2004; Idris et al., 2006; Palmroth et al., 2007; Zakaria et (IOW) for lab assistance and T. Max (MPI-MM) for mass spectrometric al., 2007). The presence of these organisms coincides with the high support. metal concentrations observed in groundwater from the reaction zone in this aquifer (e.g. Ni concentrations of ~200 μg/L; Zhang et al., 2009). Appendix A. Supplementary data These geochemical conditions are the result of the recent input of anthropogenic nitrate into the pyrite containing aquifer and the corre- Supplementary data to this article can be found online at doi:10. sponding mobilization of trace metals. 1016/j.chemgeo.2012.01.024. Y.-C. Zhang et al. / Chemical Geology 300-301 (2012) 123–132 131

References NCIMB 9548 as the type strain. International Journal of Systematic and Evolution- ary Microbiology 50, 547–550. Anandham, R., Indiragandhi, P., Madhaiyan, M., Chung, J., Ryu, K.Y., Jee, H.J., Sa, T., 2009. Kendall, C., Elliott, E.M., Wankel, S.D., 2007. Tracing anthropogenic inputs of nitrogen to Thiosulfate oxidation and mixotrophic growth of Methylobacterium goesingense and ecosystems. In: Michener, R., Lajtha, K. (Eds.), Stable isotopes in ecology and envi- Methylobacterium fujisawaense. Journal of Microbiology and Biotechnology 19 (1), ronmental science. Blackwell, pp. 375–449. 17–22. Korom, S.F., 1992. Natural denitrification in the saturated zone: a review. Water Appelo, C.A.J., Postma, D., 2005. Geochemistry, groundwater and pollution, 2nd edition. Resources Research 28 (6), 1657–1668. CRC Press. 649 pp. Larkin, J.M., Shinabarger, D.L., 1983. Characterization of Thiotrix-Nivea. International Bak, F., Cypionka, H., 1987. A novel type of energy metabolism involving fermentation Journal of Systematic and Evolutionary Bacteriology 33, 841–846. of inorganic sulphur compounds. Nature 326 (6116), 891–892. Lehmann, M.F., Reichert, P., Bernasconi, S.M., Barbieri, A., McKenzie, J.A., 2003. Böttcher, J., Strebel, O., Voerkelius, S., Schmidt, H.L., 1990. Using isotope fractionation of Modelling nitrogen and oxygen isotope fractionation during denitrification in a nitrate-nitrogen and nitrate-oxygen for evaluation of microbial denitrification in a lacustrine redox-transition zone. Geochimica et Cosmochimica Acta 67 (14), sandy aquifer. Journal of Hydrology 114 (3–4), 413–424. 2529–2542. Böttcher, M.E., Oelschläger, B., Höpner, T., Brumsack, H., Rullkötter, J., 1998. Sulfate re- Ludwig, W., Strunk, O., Westram, R., Richter, L., Meier, H., al., Y.e., 2004. ARB: a software duction related to early diagenetic degradation of organic matter and ‘black spot’ environment for sequence data. Nucleic Acids Research 32, 1363–1371. formation in tidal sandflats of the German Wadden Sea: Stable isotope (13C, 34S, Mahmood, Q., Zheng, P., Hu, B.L., Jilani, G., Azim, M.R., Wu, D.L., Liu, D., 2009. Isolation 18O) and other geochemical results. Organic Geochemistry 29, 1517–1530. and characterization of Pseudomonas stutzeri QZ1 from an anoxic sulfide-oxidiz- Böttcher, M.E., Thamdrup, B., Vennemann, T.W., 2001. Oxygen and sulfur isotope frac- ing bioreactor. Anaerobe 15, 108–115. tionation during anaerobic bacterial disproportionation of elemental sulfur. Geo- Mariotti, A., Landreau, A., Simon, B., 1988. 15N isotope biogeochemistry and natural de- chimica et Cosmochimica Acta 65 (10), 1601–1609. nitrification process in groundwater: application to the chalk aquifer of northern Brettar, I., Labrenz, M., Flavier, S., Botel, J., Kuosa, H., Christen, R., Hofle, M.G., 2006. France. Geochimica et Cosmochimica Acta 52 (7), 1869–1878. Identification of a Thiomicrospira denitrificans-like epsilonproteobacterium as Mergeay, M., Monchy, S., Vallaeys, T., Auquier, V., Benotmane, A., Bertin, P., Taghavi, S., a catalyst for autotrophic denitrification in the Central Baltic Sea. Applied and Dunn, J., van der Lelie, D., Wattiez, R., 2003. Ralstonia metallidurans, a bacterium Environmental Microbiology 72 (2), 1364–1372. specifically adapted to toxic metals: towards a catalogue of metal-responsive Broers, H.P., 1998. Release of Ni, Co and As from aquifer sediments at the Oostrum genes. FEMS Microbiology Reviews 27 (2–3), 385–410. well field. In: Herbert, M., Kovar, K. (Eds.), Groundwater Quality, Remediation Miotliński1, K., 2008. Coupled reactive transport modeling of redox processes in a and Protection. Tuebingener Geowissenschaftiche Arbeiten, Tuebingen, Germany, nitrate-polluted sandy aquifer. Aquatic Geochemistry 14 (2), 117–131. pp. 106–108. Mizutani, Y., Rafter, T.A., 1973. Isotopic behavior of sulfate oxygen in the bacterial re- Broers, H.P., Buijs, E.A., 1997. The origin of trace metals and arsenic at the drinking duction of sulfate. Geochemical Journal 6, 183–191. water well field. Report TNO-NITG97-198-A (in Dutch). Moncaster, S.J., Bottrell, S.H., Tellam, J.H., Lloyd, J.W., Konhauser, K.O., 2000. Migration Canfield, D.E., Raiswell, R., Westrich, J.T., Reaves, C.M., Berner, R.A., 1986. The use of and attenuation of agrochemical pollutants: insights from isotopic analysis of chromium reduction in the analysis of reduced inorganic sulfur in sediments and groundwater sulphate. Journal of Contaminant Hydrology 43 (2), 147–163. shales. Chemical Geology 54 (1–2), 149–155. Niemann, H., Duarte, J., Hensen, C., Omoregie, E., Magalhães, V.H., Elvert, M., Pinheiro, Canfield, D.E., Kristensen, E., Thamdrup, B., 2005. Advances in Marine Biology: Chapter L.M., Kopf, A., Boetius, A., 2006. Microbial methane turnover at mud volcanoes of 9 The sulfur cycle. Advances in Marine Biology 48, 314–381. the Gulf of Cadiz. Geochimica et Cosmochimica Acta 70 (21), 5336–5355. Chang, C.C.Y., Langston, J., Riggs, M., Campbell, D.H., Silva, S.R., Kendall, C., 1999. A method Palmroth, M.R.T., Koskinen, P.E.P., Kaksonen, A.H., Munster, U., Pichtel, J., Puhakka, J.A., for nitrate collection for d15N and d18O analysis from waters with low nitrate con- 2007. Metabolic and phylogenetic analysis of microbial communities during phy- centrations. Canadian Journal of Fisheries and Aquatic Sciences 56, 1856–1864. toremediation of soil contaminated with weathered hydrocarbons and heavy Clark, I.D., Fritz, P., 1997. Environmental isotopes in hydrogeology. CRC Press/Lewis metals. Biodegradation 18 (6), 769–782. Publishers, Boca Raton, FL. 328 pp. Pauwels, H., Ayraud-Vergnaud, V., Aquilina, L., Molénat, J., 2010. The fate of nitrogen Deutsch, B., Mewes, M., Liskow, I., Voss, M., 2006. Quantification of diffuse nitrate and sulfur in hard-rock aquifers as shown by sulfate-isotope tracing. Applied Geo- inputs into a small river system using stable isotopes of oxygen and nitrogen in chemistry 25 (1), 105–115. nitrate. Organic Geochemistry 37 (10), 1333–1342. Pauwels, H., Kloppmann, W., Foucher, J.-C., Martelat, A., Fritsche, V., 1998. Field tracer Devlin, J.F., Eedy, R., Butler, B.J., 2000. The effects of electron donor and granular iron on test for denitrification in a pyrite-bearing schist aquifer. Applied Geochemistry nitrate transformation rates in sediments from a municipal water supply aquifer. 13 (6), 767–778. Journal of Contaminant Hydrology 46, 81–97. Polz, M.F., Odintsova, E.V., Cavanaugh, C.M., 1996. Phylogenetic relationships of the fil- Ding, L., Yokota, A., 2004. Proposals of Curvibacter gracilis gen. nov., sp. nov. and amentous sulfur bacterium Thiothrix ramosa based on 16S rRNA sequence analy- Herbaspirillum putei sp. nov. for bacterial strains isolated from well water and reclas- sis. International Journal of Systematic and Evolutionary Bacteriology 46, 94–97. sification of [Pseudomonas] huttiensis,[Pseudomonas] lanceolata,[Aquaspirillum] Postma, D., Boesen, C., Kristiansen, H., Larsen, F., 1991. Nitrate reduction in an uncon- delicatum and [Aquaspirillum] autotrophicum as Herbaspirillum huttiense comb. nov., fined sandy aquifer – water chemistry, reduction processes, and geochemical Curvibacter lanceolatus comb. nov., Curvibacter delicatus comb.nov.andHerbaspirillum modeling. Water Resources Research 27 (8), 2027–2045. autotrophicum comb. nov. International Journal of Systematic and Evolutionary Prietzel, J., Botzaki, A., Tyufekchieva, N., Brettholle, M., Thieme, J., Klysubun, W., 2011. Microbiology 54 (6), 2223–2230. Sulfur speciation in soil by S K-Edge XANES spectroscopy: comparison of spectral Finster, K., 2008. Microbiological disproportionation of inorganic sulfur compounds. deconvolution and linear combination fitting. Environmental Science and Technol- Journal of Sulfur Chemistry 29 (3), 281–292. ogy 45, 2878–2886. Fossing, H., Jørgensen, B.B., 1989. Measurement of bacterial sulfate reduction in sedi- Pruesse, E., Quast, C., Knittel, K., Fuchs, B.M., Ludwig, W.G., Peplies, J., Glockner, F.O., ments: evaluation of a single-step chromium reduction method. Biogeochemistry 2007. SILVA: a comprehensive online resource for quality checked and aligned 8 (3), 205–222. ribosomal RNA sequence data compatible with ARB. Nucleic Acids Research 35 Fritz, P., Basharmal, G.M., Drimmie, R.J., Ibsen, J., Qureshi, R.M., 1989. Oxygen isotope (21), 7188–7196. exchange between sulphate and water during bacterial reduction of sulphate. Reardon, C.L., Cummings, D.E., Petzke, L.M., Kinsall, B.L., Watson, D.B., Peyton, B.M., Chemical Geology: Isotope Geoscience Section 79 (2), 99–105. Geesey, G.G., 2004. Composition and diversity of microbial communities recovered Goris, J., De Vos, P., Coenye, T., Hoste, B., Janssens, D., Brim, H., Diels, L., Mergeay, M., from surrogate minerals incubated in an acidic uranium-contaminated aquifer. Kersters, K., Vandamme, P., 2001. Classification of metal-resistant bacteria from in- Applied and Environmental Microbiology 70 (10), 6037–6046. dustrial biotopes as Ralstonia campinensis sp. nov., Ralstonia metallidurans sp. nov. Schippers, A., Jørgensen, B.B., 2002. Biogeochemistry of pyrite and iron sulfide oxida- and Ralstonia basilensis Steinle et al. 1998 emend. International Journal of Systematic tion in marine sediments. Geochimica et Cosmochimica Acta 66 (1), 85–92. and Evolutionary Microbiology 51 (5), 1773–1782. Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., 2009. In- Habicht, K.S., Canfield, D.E., 2001. Isotope fractionation by sulfate-reducing natural troducing mothur: open-source, platform-independent, community-supported populations and the isotopic composition of sulfide in marine sediments. Geology software for describing and comparing microbial communities. Appl Environmen- of Ore Deposits 29, 555–558. tal Microbiology 75, 7537–7541. Hallbeck, L., Stahl, F., Pedersen, K., 1993. Phylogeny and phenotypic characterization of Schwientek, M., Einsiedl, F., Stichler, W., Stögbauer, A., Strauss, H., Maloszewski, P., the stalk-forming and iron-oxidizing bacterium Gallionella–Ferruginea. Journal of 2008. Evidence for denitrification regulated by pyrite oxidation in a heterogeneous General Microbiology 139, 1531–1535. porous groundwater system. Chemical Geology 255 (1–2), 60–67. Hery, M., Nazaret, S., Jaffre, T., Normand, P., Navarro, E., 2003. Adaptation to nickel Silva, S.R., Kendall, C., Wilkison, D.H., Ziegler, A.C., Chang, C.C.Y., Avanzino, R.J., 2000. A spiking of bacterial communities in neocaledonian soils. Environmental Microbiol- new method for collection of nitrate from fresh water and the analysis of nitrogen ogy 5 (1), 3–12. and oxygen isotope ratios. Journal of Hydrology 228 (1–2), 22–36. Idris, R., Kuffner, M., Bodrossy, L., Puschenreiter, M., Monchy, S., Wenzel, W.W., Sessitsch, Smith, R., Böhlke, J., Repert, D., Hart, C., 2009. Nitrification and denitrification in a mid- A., 2006. Characterization of Ni-tolerant methylobacteria associated with the hyperac- western stream containing high nitrate: in situ assessment using tracers in dome- cumulating plant Thlaspi goesingense and description of Methylobacterium goesingense shaped incubation chambers. Biogeochemistry 96 (1), 189–208. sp nov. Systematic and Applied Microbiology 29 (8), 634–644. Stempvoort Van, D.R., Hendry, M.J., Schoenau, J.J., Krouse, H.R., 1994. Sources and dy- Jørgensen, C.J., Jacobsen, O.S., Elberling, B., Aamand, J., 2009. Microbial oxidation of py- namics of sulfur in weathered till, Western Glaciated Plains of North America. rite coupled to nitrate reduction in anoxic groundwater sediment. Environmental Chemical Geology 111 (1–4), 35–56. Science and Technology 43, 4851–4857. Straub, K.L., Benz, M., Schink, B., Widdel, F., 1996. Anaerobic, nitrate-dependent micro- Kaplan, I.R., Rittenberg, S.C., 1964. Microbiological fractionation of sulfur isotopes. Journal bial oxidation of ferrous iron. Applied and Environmental Microbiology 62 (4), of General Microbiology 34, 195–212. 1458–1460. Kelly, D.P., Wood, A.P., 2000. Confirmation of Thiobacillus denitrificans as a species of Strebel, O., Duynisveld, W.H.M., Böttcher, J., 1989. Nitrate pollution of groundwater in the genus Thiobacillus, in the beta-subclass of the Proteobacteria, with strain western Europe. Agriculture, Ecosystems & Environment 26 (3–4), 189–214. 132 Y.-C. Zhang et al. / Chemical Geology 300-301 (2012) 123–132

Taylor, B.E., Wheeler, M.C., Nordstrom, D.K., 1984. Stable isotope geochemistry of acid Visser, A., Broers, H.P., Heerdink, R., Bierkens, M.F.P., 2009. Trends in pollutant concen- mine drainage: experimental oxidation of pyrite. Geochimica et Cosmochimica trations in relation to time of recharge and reactive transport at the groundwater Acta 48 (12), 2669–2678. body scale. Journal of Hydrology 369 (3–4), 427–439. Tesoriero, A.J., Liebscher, H., Cox, S.E., 2000. Mechanism and rate of denitrification in an Xue, D., Botte, J., De Baets, B., Accoe, F., Nestler, A., Taylor, P., Van Cleemput, O., agricultural watershed: electron and mass balance along groundwater flow paths. Berglund, M., Boeckx, P., 2009. Present limitations and future prospects of stable Water Resources Research 36 (6), 1545–1559. isotope methods for nitrate source identification in surface- and groundwater. Thamdrup, B., Finster, K., Fossing, H., Hansen, J.W., Jørgensen, B.B., 1994. Thiosulfate Water Research 43 (5), 1159–1170. and sulfite distributions in porewater of marine sediments related to manganese, Zakaria, Z.A., Zakaria, Z., Surif, S., Ahmad, W.A., 2007. Hexavalent chromium reduction iron, and sulfur geochemistry. Geochimica Et Cosmochimica Acta 58 (1), 67–73. by Acinetobacter haemolyticus isolated from heavy-metal contaminated wastewater. Toran, L., Harris, R.F., 1989. Interpretation of sulfur and oxygen isotopes in biological Journal of Hazardous Materials 146 (1–2), 30–38. and abiological sulfide oxidation. Geochimica et Cosmochimica Acta 53 (9), Zhang, Y.-C., Slomp, C.P., Broers,H.P.,Passier,H.F.,VanCappellen,P.,2009. 2341–2348. Denitrification coupled to pyrite oxidation and changes in groundwater Torrentó, C., Cama, J., Urmeneta, J., Otero, N., Soler, A., 2010. Denitrification of groundwater quality in a shallow sandy aquifer. Geochimica et Cosmochimica Acta 73 (22), with pyrite and Thiobacillus denitrificans. Chemical Geology 278 (1–2), 80–91. 6716–6726.

Toyoda, S., Mutobe, H., Yamagishi, H., Yoshida, N., Tanji, Y., 2005. Fractionation of N2O Ziegler, S., Ackermann, S., Majzlan, J., Gescher, J., 2009. Matrix composition and com- isotopomers during production by denitrifier. Soil Biology and Biochemistry 37 munity structure analysis of a novel bacterial pyrite leaching community. Environ- (8), 1535–1545. mental Microbiology 11 (9), 2329–2338. Tuttle, M.L.W., Breit, G.N., Cozzarelli, I.M., 2009. Processes affecting δ34S and δ18O Zul, D., Wanner, G., Overmann, J., 2008. Massilia brevitalea sp. nov., a novel betaproteo- values of dissolved sulfate in alluvium along the Canadian River, central Oklahoma, bacterium isolated from lysimeter soil. International Journal of Systematic and USA. Chemical Geology 265 (3–4), 455–467. Evolutionary Microbiology 58 (5), 1245–1251.