Sulfur transformations in pilot-scale constructed wetland treating high sulfate-containing contaminated groundwater: A stable isotope assessment

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 8 8 e6 6 9 8

Available online at www.sciencedirect.com

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

Sulfur transformations in pilot-scale constructed wetland treating high sulfate-containing contaminated groundwater: A stable isotope assessment Shubiao Wu a, Christina Jeschke b, Renjie Dong c,*, Heidrun Paschke d, Peter Kuschk e, Kay Kno¨ller b a

Key Laboratory of Agricultural Engineering in Structure and Environment of Ministry of Agricultural, College of Water Conservancy & Civil Engineering, China Agricultural University, 100083 Beijing, PR China b Department of Catchment Hydrology, Helmholtz Centre for Environmental Research - UFZ, Theodor-Lieser-Strasse 4, Halle D-06120, Germany c College of Engineering, China Agricultural University, 100083 Beijing, PR China d Department of Groundwater Remediation, Helmholtz Centre for Environmental Research - UFZ, Permoserstrasse 15, Leipzig D-04318, Germany e Department of Environmental Biotechnology, Helmholtz Centre for Environmental Research - UFZ, Permoserstrasse 15, Leipzig D-04318, Germany

article info

abstract

Article history:

Current understanding of the dynamics of sulfur compounds inside constructed wetlands is

Received 17 June 2011

still insufficient to allow a full description of processes involved in sulfur cycling. Experiments

Received in revised form

in a pilot-scale horizontal subsurface flow constructed wetland treating high sulfate-

30 September 2011

containing contaminated groundwater were carried out. Application of stable isotope

Accepted 7 October 2011

approach combined with hydro-chemical investigations was performed to evaluate the sulfur

Available online 19 October 2011

transformations. In general, under inflow concentration of about 283 mg/L sulfate sulfur, sulfate removal was found to be about 21% with a specific removal rate of 1.75 g/m2$d. The

Keywords:

presence of sulfide and elemental sulfur in pore water about 17.3 mg/L and 8.5 mg/L, respec-

Constructed wetland

tively, indicated simultaneously bacterial sulfate reduction and re-oxidation. 70% of the

Bacterial sulfate reduction

removed sulfate was calculated to be immobilized inside the wetland bed. The significant

Sulfide re-oxidation

enrichment of 34S and 18O in dissolved sulfate (d34S up to 16&, compared to average of 5.9& in

Stable isotopes

the inflow, and d18O up to 13&, compared to average of 6.9& in the inflow) was observed clearly correlated to the decrease of sulfate loads along the flow path through experimental wetland bed. This enrichment also demonstrated the occurrence of bacterial sulfate reduction as well as demonstrated by the presence of sulfide in the pore water. Moreover, the integral approach shows that bacterial sulfate reduction is not the sole process controlling the isotopic composition of dissolved sulfate in the pore water. The calculated apparent enrichment factor (3 ¼ 22&) for sulfur isotopes from the d34S vs. sulfate mass loss was significantly smaller than required to produce the observed difference in d34S between sulfate and sulfide. It indicated some potential processes superimposing bacterial sulfate reduction, such as direct reoxidation of sulfide to sulfate by oxygen released from plant roots and/or bacterial disproportionation of elemental sulfur. Furthermore, 41% of residual sulfate was calculated to be

* Corresponding author. Tel.: þ86 10 62737852; fax: þ86 10 62737885. E-mail address: [email protected] (S. Wu). 0043-1354/$ e see front matter ª 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.watres.2011.10.008

6689

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 8 8 e6 6 9 8

from sulfide re-oxidation, which demonstrated that the application of stable isotope approach combined with the common hydro-chemical investigations is not only necessary for a general qualitative evaluation of sulfur transformations in constructed wetlands, but also leads to a quantitative description of intermediate processes. ª 2011 Elsevier Ltd. All rights reserved.

1.

Introduction

Since the 1980s, constructed wetlands as an alternative ecological technology for wastewater treatment have been increasingly used for remediation of various contaminated waters (Bulc, 2006; Braeckevelt et al., 2008; Garcia et al., 2010). However, despite the successful application, the knowledge of the removal processes inside the systems is still insufficient, because of the variability of the redox states in the near-root zones and the complex interactions of different microbial transformations (Faulwetter et al., 2009). In the wetland bed, the spatial and temporal micro-scale gradients of oxygen concentrations and redox states established close to root surfaces enable the development of microbial biofilms of functionally different microorganisms. Those microorganisms can simultaneously mediate processes such as nitrification, denitrification, mineralization of organic carbon, methanogenesis, sulfate reduction, and sulfide oxidation on a small spatial scale (Lee et al., 1999; Holmer and Storkholm, 2001). As sulfate is a common constituent of wastewaters, different processes of sulfur cycling, depending on the availability of organic carbon and/or oxygen, are accordingly prevalent in constructed wetlands (Sturman et al., 2008; Wiessner et al., 2010). The wetland system can influence the sulfur cycling by e.g. releasing organic carbon compounds and/or oxygen from the plant roots to enhance the sulfate reduction or re-oxidation of the reduced sulfur compounds. Moreover, the processes of sulfur transformations, such as sulfate reduction can also influence the conditions for the biochemical processes, changing the pH and redox conditions (Leon et al., 2002; Geurts et al., 2009). In addition, the nitrogen removal and other removal processes under sulfur cycling can be influenced as well due to the toxicity of the reduced sulfur compounds like H2S, as well as competition for oxygen (Aesoy et al., 1998; Stein et al., 2007). Previous studies of sulfur transformations focus mainly on the removal of metals by precipitation with sulfide in constructed wetlands, particularly in treating acid mine drainage (Webb et al., 1998; Machemer et al., 1993; Woulds and Ngwenya, 2004), and some focus on negative effects caused by sulfide toxicity on wetland plants and microorganisms (Aesoy et al., 1998; Wiessner et al., 2010; Armstrong et al., 1996). However, the knowledge of evaluation on the dynamics of sulfur cycling in the constructed wetlands, particularly the quantitative description of intermediate turnover is still limited (Choi et al., 2006; Faulwetter et al., 2009). Moreover, the quantification of microbial sulfur transformations by concentration pattern of reactant (SO2 4 ) consumption or product formation (S2) in aquifers may be obscured by concurrent abiotic transformations, e.g. dilution due to the rainfall, concentration resulted from high

evapotranspiration, matrix effects, and mineral precipitation such as precipitation of gypsum and sulfide with metals (Anderson and Lovley, 2000; Schroth et al., 2001; Kno¨ller et al., 2006; Rahman et al., 2008). Furthermore, the coexisting of biotic processes, such as bacterial dissimilatory sulfate reduction and sulfide re-oxidation as well as disproportionation of different reduced sulfur compounds under specific micro gradients, also makes the understanding of different turnover processes of sulfur species inside the constructed wetlands be insufficient (Finster et al., 1998; Liesack et al., 2000). As a tool to discern microbial activity from abiotic transformations and assess pathways and rates of sulfur transformations, stable isotope analyses have been found increasingly applied in recent years (Schroth et al., 2001; Kno¨ller et al., 2004, 2008). Microbial sulfate reduction usually results in significant isotope enrichment of 34S in residual sulfate coupled to a depletion of 34S in produced sulfide (Rees, 1973; Fry et al., 1988; Bottrell et al., 1995). Sulfur isotope fractionation appears to be a valuable indicator for microbial sulfate reduction in complex environments. Thus, sulfur isotope fractionation in groundwater was previously observed in forest hydrological studies (Robertson and Schiff, 1994; Alewell and Giesemann, 1996) as well as in contaminated aquifers (Kno¨ller et al., 2006, 2008). Presently, little is know about sulfur isotope fractionation in constructed wetlands with complex coexisting reductive and oxidative conditions. In this study, evaluation of application of stable isotope investigation combined with common hydro-chemical examination in pilot-scale constructed wetland treating high sulfate-containing contaminated groundwater was conducted for: 1) the identification of microbial sulfate reduction; 2) recognition of further bacterial sulfur transformations superimposing sulfate reduction, such as disproportionation of reduced sulfur compounds and re-oxidation of sulfide; 3) the impact of wetland plants on sulfur cycling which was facilitated by the oxygen and organic carbon compounds released from plant roots; 4) quantitative assessment of reoxidation of sulfide to sulfate.

2.

Materials and methods

2.1.

Site description and experimental setup

The experimental pilot-scale constructed wetland was built at the SAFIRA research site in Bitterfeld, Germany in 2003 (Braeckevelt et al., 2008). The local groundwater used for loading this system during this experiment contained monochlorobenzene (MCB) as the main organic compound with concentration of 6e12 mg/L and sulfate (710e920 mg/L) as the

6690

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 8 8 e6 6 9 8

main inorganic compound. The constructed wetland was designed to operate in a horizontal subsurface flow mode with dimension of 6 m length  1 m width and was filled to an average depth of 0.5 m with autochthonous quaternary aquifer material consisting predominantly of about 25% mica sand and about 67% gravel (porosity 0.35). The wetland bed was planted with common reed (Phragmites australis). The loading contaminated groundwater from a well installed in 22 m depth was pumped up and supplied continuously at a flow rate of 5.0 L/h to the wetland bed. The water level was maintained at approximate 0.1 m below the surface of the wetland. The samples took from 4 m (distance from inlet) were defined as effluent and the theoretical hydraulic retention time in the wetland bed was about 6 days. The outflow volumes were measured by a flow meter in order to determine water loss and allow the calculation of load removal. The system had been run with the same groundwater for 7 years before the current investigation was carried out from March to November in 2010.

2.2.

Sampling

Pore water samples for the evaluation of hydro-chemical parameters were taken from the wetlands at 0.5, 1, 2, 3 and 4 m distance from the inflow at 0.3, 0.4 and 0.5 m depth of the bed. Inflow samples were taken from the feeding pipe. The 18 2 isotope fractionation samples for d34SeSO2 4 , d OeSO4 were added 5% Zn-Acetate solution to removal the interference of sulfide. The samples for d34S in sulfide were only taken from 0.5 m depth of the bed and precipitated with 5% Zn-Acetate solution. Water sampling at each depth was carried out using stainless steel lancets (3.5 mm inner diameter) and peristaltic pumps at a rate of 78 ml/min. Water samples were stored without headspace at 4  C until analysis.

2.3.

Physical-chemical analysis

The redox potential was determined in the field using a SenTix ORP electrode (WTW, Weilheim, Germany) and the temperature was measured with a temperature sensor (PT 1000, PreSens, Regensburg, Germany). Sample filtration was carried out using a 5 mm-syringe filter (Ministart NML, Sartorius) for particles removal before the quantitative ions analysis. The pH value was measured by a SenTix41 electrode with pH 537 Microprocessor (WTW, Weilheim, Germany). For the conservation of Fe(II) samples, hydrochloric acid was added and after derivatisation with ferrocin, photometric measurement was carried out at 562 nm using a Cadas 100 photometer (Hach Lange, Dusseldorf, Germany) (Lovley and Phillips, 1986). Sulfate concentrations were determined by ion chromatography (DX 500) with an IonPacAG11 (4  250 mm) column (Dionex Corporation, Sunnyvale, USA) and conductivity detection (CD 20). Sulfide concentrations were measured by photometric method using Test kit LCW053 from HACH LANGE, Germany. Elemental sulfur in the pore water was estimated according to Rechmeier et al. (Rethmeier et al., 1997), by extracting samples with chloroform and subsequent detection by HPLC (Beckman, USA) using a Li-Chrospher 100, RP 18 column (5 mm, Merck, Germany) equipped with a UV-detector at 263 nm.

2.4.

Isotope analysis

Sulfur isotope analysis was conducted according to Kno¨ller and Schubert (2010). The precipitated ZnS was removed by filtration (0.45 mm). After adding concentrated hydrochloric acid in the laboratory, the hydrogen sulfide was stripped with N2 gas and then trapped as ZnS in Zn-acetate solution. The precipitated ZnS was subsequently converted to Ag2S by addition of 0.1 M AgNO3 solution. Dissolved sulfate was recovered by precipitation as BaSO4 at 70  C after the pH of the solution was adjusted to 3.0 and BaCl2 solution was added. Sulfur isotopic compositions were measured after conversion of BaSO4 (or Ag2S) to SO2 using an elemental analyzer (continuous flow flash combustion technique) coupled with an isotope ratio mass spectrometer (delta S, ThermoFinnigan, Bremen, Germany). Sulfur isotope measurements were performed with an analytical error of better than 0.3& and results are reported in delta notation (d34S) as part per thousand (&) deviations relative to the Vienna Can˜on Diablo Troilite (VCDT) standard (according to general Eq. (1)) dð&Þ ¼



Rsample  Rstandard


 Rstandard  1000

(1) 34

32

where R is the ratio of the heavy to light isotopes (e.g. S/ S or 18O/16O). Oxygen isotope analysis with barium sulfate samples was carried out by high temperature pyrolysis at 1450  C in a TC/EA connected to a delta plus XL mass spectrometer (ThermoFinnigan, Bremen, Germany) with an analytical error of better than 0.5&. According to Eq. (1), results of oxygen isotope measurements are expressed in delta notation (d18O) as part per thousand (&) deviations relative to Vienna Standard Mean Ocean Water (VSMOW). For normalizing the d34S data, the IAEA-distributed reference materials NBS 127 (BaSO4) and IAEA-S1 (Ag2S) were used. The assigned values were þ20.3& (VCDT) for NBS 127 and 0.3& (VCDT) for IAEA-S1. The normalization of oxygen isotope data of sulfate was carried out using the reference material NBS 127 with an assigned d18O value of þ8.7& (VSMOW).

2.5.

Calculation

Water loss generally occurs in constructed wetlands via evaporation from the filter surface and transpiration by plants, which is combined to be called evapotranspiration. The area specific water loss (ΔV) during a defined period is calculated by measuring the influent and effluent streams as well as rainfall, DV ¼ fðVin þ Prain  A  Vout Þ=Vin g  100

(2)

where ΔV is the water loss by evapotranspiration in %, Vin and Vout are the influent and effluent volumes in L/d, and Prain is the precipitation, which was measured by a weather station near the SAFIRA site and its amount was related to specific area in L/m2d. For the evaluation of treatment performance of constructed wetlands, the consideration of water loss and load calculation is necessary. The following assumptions were made: 1) Evapotranspiration along the wetland flow path increases in a linear way

6691

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 8 8 e6 6 9 8

2) An simplified ideal plug flow though the soil filter exists. Therefore, the load of the contaminants in the defined sampling points can be expressed as Eq. (3) and the residual fraction of the contaminants through the bed (e.g. 4 m sampling point) was defined as Eq. (4). Load4m ¼ ðVin þ Prain  A4m  DV4m  Vin Þ  C4m

(3)

Residual fraction ð%Þ ¼ fðVin þ Prain  A4m  DV4m  Vin Þ  C4m g=Vin  Cin

(4)

where Cin is the influent concentration in mg/L, C4m is the concentration at 4 m sampling point in mg/L, and A4m is the area from the inflow to the sampling point ¼ 4 m2.

3.

Results

Concentrations of sulfate, sulfide and elemental sulfur in the pore water of experimental constructed wetland along flow path from inlet to outlet are presented in Table 1. The sulfide and elemental sulfur depending on the distance from inlet to outlet gradually increased to 18.5 mg/L and 8.2 mg/L, respectively. However, the concentrations of sulfate-sulfur varied from 266 mg/L to 293 mg/L through the wetland and did not show any decreasing tendency. The clearly production of sulfide and elemental sulfur was not proved by the decrease of sulfate concentrations along the flow path, which was attributed to the concentrating effect of considerable water losses occurring via evapotranspiration. As shown in Fig. 1, the monthly average water loss calculated from inflow and outflow streams varied from 55% to 1%, correlating to the changes of temperature from 25  C to 8  C during sampling months. The loss of water from the wetland was mainly (about 92%) conducted by the plants transpiration.

Concerning the considerable water losses from wetland, the concentrations can not really reflect the microbial sulfur transformations here. Thus, according to Eqs. (2) and (3) the daily loads along the flow path from inlet to outlet were calculated and used to interpret the sulfur dynamics through the wetland. As shown in Fig. 2, with gradual increase of sulfide and elemental sulfur in the pore water through the wetland, sulfate-sulfur decreased correspondingly from 34 g/ d to 27 g/d, indicating the calculation of load with consideration of water loss fits the case study in the field. The reduction of sulfate was generally achieved 21% with a specific area removal rate of 1.75 g/m2$d. The compositions of isotopic sulfur and oxygen of sulfate and sulfide along the flow path from inlet to outlet were shown in Table 2 and Fig. 3. In general, d34S of sulfate and sulfide and d18O of sulfate increased from inlet with values of þ5.7&, 33.2& and þ6.7& to outlet with values of þ16&, 25& and þ13&, respectively. Significant enrichment of heavy isotopes in this experiment was observed. The pH measured in all sampling points was within the range of 6.5e6.9. Redox potential gradually decreased from about 50 mV in the inflow to 120 mV in the outflow (4 m sampling point). Ferrous iron from the pore water through the whole bed was constant around 0.5 mg/L.

4.

Discussion

Sulfide is a product of bacterial dissimilatory sulfate reduction (BSR) by using organic compounds as electron donors. Elemental sulfur is a product of sulfide oxidation, which may be performed by abiotic oxidation and/or biological oxidation by using different electron acceptors, such as oxygen, nitrite and nitrate (Buisman et al., 1990; Mahmood et al., 2009; Zheng, 2007; Zheng and Cai, 2007). In this study, the coexisting of sulfide and elemental sulfur in the pore water of experimental

Table 1 e Sulfur species concentrations of the soil pore water of experimental pilot-scale constructed wetland along flow path from inlet to outlet (in mg/L, n [ 15). Flow path (m)

0 (inflow) 0.5

1

2

3

4 (outflow)

SO2 4 S

Depth (m)

0.3 0.4 0.5 0.3 0.4 0.5 0.3 0.4 0.5 0.3 0.4 0.5 0.3 0.4 0.5

S2

S0

Mean

STDEV

Mean

STDEV

Mean

STDEV

283.1 281.5 282.0 282.7 280.5 284.6 292.7 261.2 270.5 269.2 288.4 285.2 263.9 266.0 257.9 288.1

12.1 12.1 10.4 12.4 11.7 11.1 15.3 19.8 16.9 18.5 15.8 18.3 18.5 17.2 24.1 14.8

B.D.L 2.2 1.1 1.1 3.2 2.9 1.8 14.2 10.3 8.4 15.5 15.8 18.5 14.1 14.6 17.9

B.D.L 0.9 0.6 0.8 2.8 2.8 1.7 5.1 2.5 3.6 7.1 3.6 4.9 4.6 3.1 6.2

B.D.L 1.3 1.3 0.4 1.6 0.8 0.4 7.2 4.3 4.1 6.7 6.9 8.2 6.1 6.5 6.8

B.D.L 0.8 1.0 0.3 1.4 0.6 0.3 1.7 1.4 1.7 2.6 3.8 3.7 2.7 5.1 2.4

B.D.L means the estimated concentration below detection limit. The detection limit for sulfide and elemental sulfur is 0.1 mg/L.

6692

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 8 8 e6 6 9 8

90

35

80

30

70 60 20

50

15

40 30

W ater loss (%)

Temperature (°C)

25

10 20 5

0 07/2010

10

08/2010

09/2010

10/2010

11/2010

0 12/2010

Date (month / year) Daily average temperature Monthly average temperature Monthly water loss

into the rhizosphere, the spatial and temporal micro-scale gradients of oxygen concentrations and redox states are consequently established close to root surfaces (Colmer, 2003; Bezbaruah and Zhang, 2004). The coexisting reductive and oxidative conditions in the root-near zones could enable microbial processes such as mineralization of organic carbon, sulfate reduction and sulfide re-oxidation realized on a small spatial scale simultaneously (Holmer and Storkholm, 2001). In concert with decreasing sulfate-sulfur load from inlet to outlet presented in Fig. 3, the significant enriched 34S of sulfate values along flow path (Fig. 4) are strong evidence for the occurrence of BSR. Generally, under closed system conditions, 34 the relationship between SO2 4 concentrations and d S values is expressed by a Rayleigh Eq. (5) where 3 is the respective enrichment factor for sulfur and f stands for the fraction of 2 residual sulfate expressed as CeSO2 4 /C0eSO4 . d34 S ¼ d34 SSO42initial þ εlnðfÞ

Fig. 1 e The dynamics of temperature and water loss in the experimental constructed wetland. The down arrows show the isotope sampling campaigns. Box plot shows median (central thick lines), 25% and 75% quartile ranges around the median, upper and lower edge (hinge) of the box. The ends of the vertical lines (whiskers) indicate the minimum and maximum data values.

Since considerable water loss via plant transpiration occurred in this study, the concentrations of sulfate measured in the pore water can really not reflect the BSR. Thus, the 2 parameter of f in Eq. (5) expressed as CeSO2 4 /C0eSO4 can also not fit here. Accordingly, the load of sulfate calculated as a result of combined concentration and water loss was used in this study and the modified f in Eq. (5) was expressed as L2 SO2 4 /L0-SO4 . Besides, the application of Eq. (5) on field data often only yields an apparent enrichment factor 3 , because closed system conditions are rarely achieved under aquifer conditions (Kno¨ller et al., 2006). For the investigation in this

40

4

35

3

S (g/d)

30

2

2-

SO42--S (g/d)

wetland (Table 1 and Fig. 2) indicates the simultaneous reduction and re-oxidation of sulfur compounds. Regarding the oxygen and organic compounds released from the roots

(5)

25

1

20 2.0

0 0

1

2

3

4

Distance from inlet (m)

S0 (g/d)

1.5

1.0

0.5

0.0 0

1

2

3

4

Distance from inlet (m)

Fig. 2 e The load dynamics of sulfate-sulfur, sulfide and elemental sulfur along flow path in the experimental constructed wetland from inflow to outflow. The meaning of box plot was described in Fig. 2.

6693

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 8 8 e6 6 9 8

Table 2 e Isotope fractionation data for investigated sampling points of experimental pilot-scale constructed wetland along flow path from inlet to outlet (in &). Flow path (m) 0 (inflow) 0.5

1

2

3

4 (outflow)

Depth (m)

Sampling campaign in August SeSO2 4

34

d

d

5.8 6.7 6.0 5.9 6.5 6.2 6.3 12.0 9.8 9.3 13.0 13.6 13.2 16.6 16.3 14.3

0.3 0.4 0.5 0.3 0.4 0.5 0.3 0.4 0.5 0.3 0.4 0.5 0.3 0.4 0.5

OeSO2 4

18

7.0 7.9 7.7 7.9 8.0 7.6 7.3 11.2 9.6 8.3 11.4 10.2 11.4 12.8 12.4 11.2

33.2

28.7

28.4

24.9

13

‰ (VSMOW )

16

Delta- O-SO4

2-

2-

‰ (VCDT)

14

12

6.0 7.0 6.2 5.7 6.4 6.0 6.3 8.0 7.6 7.1 9.4 9.9 10.2 11.8 12.4 13.0

6.7 7.8 7.4 7.0 7.6 7.7 7.4 9.7 10.1 9.0 12.0 11.4 9.6 12.1 11.6 10.3

d34S-HS

31.0

32.0

28.3

12 11 10 9

18

34

10

d18OeSO2 4

Besides the enrichment of heavy sulfur isotope in the residual sulfate during BSR, the fractionation of oxygen isotopes in the residual sulfate molecule was also carried out. Even though the breaking of the SeO bondage will result in a kinetic isotopic effect as observed for sulfur, the kinetic isotopic effect of oxygen is masked by an isotope exchange between oxygen in sulfate and oxygen in ambient water in natural environments (Kno¨ller and Schubert, 2010). The enrichment of heavy oxygen

18

14

d SeSO2 4 34

d S-HS

study, the apparent isotopic enrichment factor (3 ) for 34S was 22&, which obtained by fitting the logarithmic Eq. (5) to the measured data (Fig. 4). The convincing logarithmic relationship shown in Fig. 4 as well as the high value of correlation coefficient (R2) suggest that BSR could be the main process determining the distribution and isotope composition of sulfate at the site, as compared to the processes like adsorption and/or desorption.

Delta- S-SO4

Sampling campaign in October 

34

8 6 4 -24

8 7 6 0

1

2

3

4

-28

34

Delta- S-S

2-

‰ (VCDT)

Distance from inlet (m) -26

-30

-32

-34 0

1

2

3

4

Distance from inlet (m)

Fig. 3 e The dynamics of isotopic compositions of d34S and d18O in sulfate and d34S in sulfide of pore water along flow path in the experimental constructed wetland from inflow to outflow.

6694

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 8 8 e6 6 9 8

6

y= -21.9 * ln(x) + 5.7 R2=0.86

10

4

2

5

0 1.0

0.9

0.8

0.7

0.6

0.5

Fraction residual sulfate (L/L0) 18

O-sulfate S-sulfate Logarithmic fitting of 34S-sulfate 34

Fig. 4 e The correlation of isotopic compositions of d34S and d18O in sulfate and fraction residual sulfate which expressed as fraction of the residual sulfate load to the initial inflow sulfate load.

in residual sulfate via isotope exchange usually was reached up to a certain equilibrium value. This equilibrium takes place within the sulfate reducing bacteria via intermediate intracellular sulfur compounds formed during BSR that are subject to re-oxidation (Brunner et al., 2005; Kno¨ller et al., 2006; Mangalo et al., 2007; Turchyn et al., 2010). As shown in Figs. 3 and 4, the significant enrichment of oxygen in the residual sulfate may again imply that the reduction of sulfate in constructed wetlands is exclusively due to BSR. As a product of BSR, sulfide is generally present in the pore water of the constructed wetlands. However, if reactive metals (e.g. iron) are also present in the aquifer, a considerable part of the sulfide may immediately be precipitated as metal sulfide and then immobilized in the wetland matrix. The isotopic evolution of the sulfide pool is depending on the isotopic composition of the precursor sulfate, enrichment factor (3 ) and the fraction of sulfide which is removed from the dissolved pool by precipitation with metals (Kno¨ller and Schubert, 2010). In closed systems, the estimation of isotopic composition of sulfide pool could be obtained by two different derivations according to Rayleigh Eq. (5). If the sulfide is precipitated under conditions with sufficient metals, the approximate d34S values of the immediate produced sulfide (d34Ssulfide-Instanous-t) can be calculated from Eq. (6). If in case of accumulation of sulfide in the pore water reservoir without any precipitation with metals, the isotopic composition of sulfide at a certain stage of BSR can be obtained according to Eq. (7). d34 SsulfideInstanoust ¼ d34 Ssulfatet þ ε

(6)

d34 Ssulfidereservoirt ¼ d34 Ssulfateinitial  f$ε$lnf=ð1  fÞ

(7)

12

y= x

10

0

8 15

8

2-

10

Delta-18O-SO42- ‰ (VSMOW )

20

2-

Delta-34S-SO4 ‰ (VCDT)

12

The correlation between assumed bacterial sulfate consumption and the sum of existing reduced sulfur compounds such as sulfide and elemental sulfur are shown in Fig. 5. The sum of estimated sulfide and elemental sulfur in the pore water from bacterial sulfate consumption was lower than the expected product reservoir accumulation. In general, the measured reduced sulfur only takes 30% of the decreased sulfate in this experimental wetland bed and 70% was immobilized in the wetland matrix such as precipitation of sulpide with metals and elemental sulfur. In addition, as well as immobilization in the matrix, formation of dissolved organic sulfur compounds and plant uptake could also account for some of the lost sulfur as shown by the few studies that have examined these in peatlands (Steinmann and Shotyk, 1997; Bottrell et al., 2010; Bartlett et al., 2009). Unfortunately, the data on sulfur species in the substrate of wetland was not available in this study and should be further investigated. Considering the presence of sulfide measured in the pore water (Table 1 and Fig. 2) and considerable immobilization of sulfur compounds (Fig. 5), the d34S values of dissolved sulfide should be plotted within a theoretical range defined by Eqs. (6) and (7). As shown in Fig. 6, this range is enclosed by the two curves modeled from Eqs. (6) and (7). However, surprisingly, no sulfide samples were plotted within the expected isotopic range and all samples were quite below this range in this study. This finding indicated that the enrichment factor for sulfur isotopes calculated from the d34S vs. sulfate mass loss is significantly smaller than required to produce the observed large difference in d34S between sulfate and sulfide. Also, this

Sum of dissolved S and S (g/d)

14

25

6

70%

4

y= 0.3 x (R2= 0.53)

2

30% 0 0

2

4

6

8

10

12

Sulfate reduction (g/d)

Fig. 5 e Correlation between assumed bacterial sulfate reduction and sum of estimated sulfide and elemental sulfur in the pore water of constructed wetland. The straight dashed line illustrates the theoretical correlation for a complete accumulation of reduced sulfur reservoir. The straight line stands for the regression of measured sum of sulfide and elemental sulfur in the pore water. Regarding the slopes of the two lines, 30% of the reduced sulfate was expressed as dissolved sulfide and elemental sulfur, indicating 70% of the reduced sulfate was deposited in the wetland matrix such as precipitation of sulfide with metals and elemental sulfur.

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 8 8 e6 6 9 8

suggested one or more sulfur transformation processes superimposing on BSR, such as bacterial disproportionation of elemental sulfur (Bo¨ttcher et al., 2001; Holmkvist et al., 2011) and direct oxidation of sulfide to sulfate by elemental oxygen which was introduced by plant roots. In constructed wetlands, the release of oxygen from plant roots into rhizosphere was well reported (Armstrong et al., 1990). The re-oxidation of sulfide to elemental sulfur using oxygen as electron acceptor is accordingly quiet reasonable here. Besides, nitrite and nitrate as a product of ammonium oxidation can also easily stimulate sulfide oxidation to elemental sulfur in wetlands (Krishnakumar and Manilal, 1999; Londry and Suflita, 1999). The process associated with sulfide re-oxidation to elemental sulfur only provides a minor isotope effect and the isotopic composition of sulfide pool would not change significantly (Bo¨ttcher et al., 1990; Balci et al., 2007). However, if elemental sulfur undergoes microbial disproportionation into sulfate and sulfide (Eq. (8)), considerable sulfur isotope fractionation and significant isotopic difference between sulfate and sulfide can be easily reached. þ 4H2 O þ 4S0 /3H2 S þ SO2 4 þ 2H

(8)

The microbial disproportionation of elemental sulfur produces more enriched sulfate and more depleted sulfide. 20 15

y= -21.9 * ln(x) + 5.7 R2=0.86

34

2-

Delta- S-SO4 ‰ (VCDT)

10 5 0 -5 -10 -15 -20 -25 -30 -35 1.0

0.9

0.8

0.7

0.6

0.5

Fraction residual sulfate (L/L0) Sulfate Sulfide Calculated sulfide instantaneous Calculated sulfide product reservoir Fitting sulfate

Fig. 6 e Relationship between the fraction of the residual sulfate and the measured sulfur isotopic composition of d34S in dissolved sulfate and sulfide. The curves represent the calculated isotopic evolution of the sulfate pool (Eq. (5)), of the instantaneously produced sulfide (Eq. (6)), and of the sulfide pool in case of an accumulation of a product reservoir (Eq. (7)) during progressing bacterial sulfate reduction under closed system conditions. For the calculation, the field based apparent isotopic enrichment factor of L21.9& was used.

6695

When the new produced sulfate mixed with the precursor sulfate, the isotopic pool of 34S will be slightly enriched. In like manner, the new sulfide pool is isotopically lighter than the precursor pool (Bo¨ttcher et al., 2001). Consequently, the isotopic difference between sulfate and sulfide will increase, as compared to the difference as a result of BSR only. Moreover, if the newly formed sulfide is oxidized again to S0, its disproportionation will further increase the isotopic difference. In this case with periodical re-oxidation of sulfide and disproportionation of elemental sulfur, the isotopic difference between coexisting sulfate and sulfide would be significantly enlarged. So far, no direct evidence for the occurrence of bacterial sulfur disproportionation has been provided in constructed wetlands. A stable isotope fractionation during formation of sulfate via this disproportionation process was experimentally investigated in a pure culture. It was found that the dissolved sulfate was enriched in 18O by 17.4& (Bo¨ttcher et al., 2001). Compared to this, the smaller enrichment of 18O in dissolved sulfate (6&) in this study demonstrated the potential occurrence of disproportionation of reduced sulfur compounds in CWs. In addition to the microbial disproportionation, direct reoxidation of sulfide to sulfate by elemental oxygen released from wetland plant roots, may also be a possible explanation for the unexpected enlarged isotopic differences between the coexisting sulfate and sulfide. Considering the hypothesis of significant direct re-oxidation of sulfide to sulfate, the enrichment factor (3 ) obtained by fitting the logarithmic analysis of measured sulfate isotope data can be underestimated to reflect the real enrichment factor. The isotopic composition of sulfate pool can be depleted when new formed sulfate from reoxidation of sulfide with a more depleted sulfate isotopic composition mixed with the precursor sulfate. Compared to the sulfur isotope fractionation during BSR only, the sulfide pool does not significantly change its isotopic composition during re-oxidation of sulfide to sulfate. Consequently, if considerable re-oxidation of sulfide to sulfate occurred, the enrichment factor obtained by fitting the logarithmic analysis of measured sulfate isotope data should not be the real evaluation of BSR. According to Eq. (6) under the assumption that the produced sulfide from sulfate reduction was immediately precipitated in the matrix, the real enrichment factor obtained for pure BSR process can be presented as isotopic difference between coexisting sulfate and sulfide. Based on this case, the real isotope curves from process of BSR only in this study was modeled using the average isotopic difference (38.9&) between sulfate and sulfide as the enrichment factor (Fig. 7). In this way, the sulfide samples as shown in Fig. 7 were perfectly enclosed within the range which calculated from the modeled sulfide isotope fractionation curve. If one considers the hypothesis of re-oxidation as valid in this study, the fraction of newly produced sulfate from sulfide re-oxidation can be modeled according to Eq. (9). d34 Smixed ¼ d34 Sprecursor ð1  XÞ þ d34 Sproduced X

(9)

in which the X stands for the fraction of newly produced sulfate from oxidation of sulfide. Using liner regression analysis of the modeled real sulfate isotope (as precursor pool) and fitting sulfate by measured values (as mixed pool), the fraction

6696

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 8 8 e6 6 9 8

of newly produced sulfate from oxidation of sulfide (X) was calculated as 41%. This fraction indicates that there was 41% of the measured sulfate resulted from sulfide re-oxidation. The mixture of 41% new produced sulfate with precursor sulfate made the enrichment factor of the whole sulfate pool decreased from 38.9& to 21.9&. Moreover, the reaction of sulfide oxidation to sulfate drived by oxidizing reactant like oxygen can be expressed as Eq. (10). Considering the fraction of 41% new produced sulfate, the oxygen equivalent consumption was calculated up to be 4.1 g/ m2d by using the minimum mean sulfate load of 27 g/d in 4 m from the inlet of the wetland with corresponding area of 4 m2 in this study. S2 þ 2O2 /SO2 4

(10)

Regarding the fine sand as the matrix used in the wetland and the horizontal saturated flow model, the oxygen diffusion from atmosphere into the wetland could be neglected in this

30

y= -38.9 * ln(x) + 5.7

25

Delta-34S-SO42- ‰ (VCDT)

20 15

y= -21.9 * ln(x) + 5.7 R2=0.86

10 5 0 -5 -10 -15

study. The oxygen supply here was mainly attributed to the plants. The oxygen flux from the roots of P. australis into their surroundings reported by Lawson (1985) and Armstrong et al. (1990) was up to be 4.3 g/m2$d and 5e12 g/m2$d, respectively. The calculated plant oxygen transfer capacity of 4.1 g/m2d using isotopic technology here was well agreed. This agreement strongly indicates the important role of plants in constructed wetlands and also indicates the potential application of isotope fractionation in constructed wetlands for deep understanding of the internal complex sulfur transformation processes. Besides the role of plants in constructed wetlands for the release of oxygen, the release of organic carbon as providing extra electron donors is as well important to influence the various microbial processes. The inflow sulfate load of the experimental wetland is about 34 g/d and outflow (4 m) sulfate is reduced to be approximate 27 g/d, yielding a net difference of 7 g/d. Regarding the 41% (11.6 g/d) of the outflow sulfate coming from the re-oxidation of sulfide, the net flux of bacterial sulfate reduction could be calculated up to be 18.6 g/ d with a specific area reduction rate of 4.65 g/m2d. The concentration of MCB as the main organic compound in the inflow water is around 0.3 g/m2d in this study. According to the reaction of MCB as electron donor coupled to the reduction of sulfate as electron acceptor expressed as Eq. (11) (Colberg, 1990), the reduction of 4.65 g/m2d sulfate needs 1.56 g/m2d consumption of MCB. The net difference of 1.26 gMCB/m2d from inflow supply and theoretical consumption strongly gives the indication of extra electron donors involved in the sulfate reduction process, underlining the role of organic compounds release from plant roots deriving sulfur cycling in constructed wetlands.

-20 

-25

 þ 2C6 H5 Cl þ 7SO2 4 þ 5H /12CO2 þ 7HS þ 2Cl þ 4H2 O

(11)

-30 -35 -40 1.0

0.9

0.8

0.7

0.6

0.5

Fraction residual sulfate (L/L0) Sulfide Sulfate Initial sulfate Fitting sulfate Calculated sulfide instantaneous Calculated sulfide product reservoir

Fig. 7 e Relationship between the fraction of the residual sulfate and the measured sulfur isotopic composition of d34S in dissolved sulfate and sulfide. The curves represent the calculated isotopic evolution of the initial sulfate pool which processing by BSR only, of the measured sulfate pool (Eq. (5)), of the instantaneously produced sulfide (Eq. (6)), and of the sulfide pool in case of an accumulation of a product reservoir (Eq. (7)) during progressing bacterial sulfate reduction under closed system conditions. For the calculation, the isotopic enrichment factor for the initial sulfate pool was modified to L38.9& according to the average isotopic difference between coexisting dissolved sulfate and sulfide.

In general, the effect of vegetation in constructed wetland was well reported (Brix, 1997; Chazarenc et al., 2009) and the role of plants improving the performance by releasing organic carbon compounds and/or oxygen from roots is also proved (Merbach et al., 1999; Picek et al., 2007). However, due to quite a biodegradation availability of the released organic carbons, the fast consumption by various microbes sitting around the roots surface makes the quantification of the released organic compounds into the wetland bed become extreme difficult. The consumption of some pollutants only under irreversible reaction using organic compounds as electron donor can be used to calculate the theoretical amount of organic carbon released from roots. However, as sulfate is undergoing microbial reversible reactions including sulfate reduction and re-oxidation of reduced sulfur compounds, the estimation of organic carbon consumption here is extremely difficult. In this study, a progressive step was made by using the approach of stable isotope combined with the common hydro-chemical parameters, and the capacity of organic carbon release from roots was calculated to 1.26 g/m2d MCB equivalent. But this value was derived by only considering the process of bacterial sulfate reduction. If considering the consumption of organic compounds by denitrification, methanogenesis and microbial respiration, the capacity of organic matter release from roots should be larger.

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 8 8 e6 6 9 8

5.

Conclusion

The significant enrichment of heavy isotopes of sulfur and oxygen in dissolved sulfate was observed to be clearly correlated to the decrease of sulfate loads along the flow path through an experimental horizontal subsurface flow wetland bed. This strongly indicates the occurrence of bacterial dissimilatory sulfate reduction. - Lack of sulfur isotope mass balance between sulfate removed and sulfide produced implies that other processes are superimposed on bacterial sulfate reduction. These include re-oxidation of sulfide to sulfate by oxygen and bacterial disproportionation of elemental sulfur in constructed wetlands. - The application of the stable isotope approach combined with common hydro-chemical investigations enables a general qualitative evaluation of sulfur transformations in constructed wetlands, but also leads to a quantitative description of intermediate processes. -

Acknowledgement This work was supported by a grant of China Scholarship Council (CSC) and by the Helmholtz Centre for Environmental Research e UFZ within the scope of the SAFIRA II Research Program (Revitalization of Contaminated Land and Groundwater at Megasites, subproject ‘‘Compartment Transfer e CoTra”). We are grateful to Martina Neuber and Sandra Zuecker-Gerstner of the stable isotope laboratory Halle/Salle for conducting isotope analyses of the samples. Thanks are also addressed to the A. Al-Dahoodi and M. Schro¨te for their valuable assistance in the laboratory and field. Furthermore, we would like to thank Simon Botrell for his valuable comments that considerably helped improve the paper.

references

Aesoy, A., Odegaard, H., Bentzen, G., 1998. The effect of sulphide and organic matter on the nitrification activity in a biofilm process. Water Science and Technology 37 (1), 115e122. Alewell, C., Giesemann, A., 1996. Sulfate reduction in a forested catchment as indicated by d34S values of sulfate in soil solutions and runoff. Isotopes in Environmental and Health Studies 32, 203e210. Anderson, R.T., Lovley, D.R., 2000. Anaerobic bioremediation of benzene under sulfate-reducing conditions in a petroleumcontaminated aquifer. Environmental Science & Technology 34, 2261e2266. Armstrong, J., Armstrong, W., Beckett, P.M., 1990. Measurement and modeling of oxygen release from roots of Phragmites australis. In: Cooper, P.F., Findlater, B.C. (Eds.), Constructed Wetlands in Water Pollution Control. Pergamon Press, Oxford, United Kingdom, pp. 41e52. Armstrong, J., Afreen-Zobayed, F., Armstrong, W., 1996. Phragmites die-back: sulphide- and acetic acid-induced bud and root death, lignifications, and blockages within aeration and vascular systems. New Phytologist 134, 601e614.

6697

Balci, N., Shanks III, W.C., Mayer, B., Mandernack, K.W., 2007. Oxygen and sulfur isotope systematics of sulfate produced by bacterial and abiotic oxidation of pyrite. Geochimica et Cosmochimica Acta 71, 3796e3811. Bartlett, R., Bottrell, S.H., Coulson, J.P., Lee, J., Forbes, L., 2009. 34S tracer study of pollutant sulfate behaviour in a lowland peatland. Biogeochemistry 95, 261e275. Bezbaruah, A.N., Zhang, T.C., 2004. pH, redox, and oxygen microprofiles in rhizosphere of bulrush (Scirpus validus) in a constructed wetland treating municipal wastewater. Biotechnology and Bioengineering 88, 60e70. Bo¨ttcher, J., Strebel, O., Voerkelius, S., Schmidt, H.L., 1990. Using isotope fractionation of nitrate nitrogen and nitrate oxygen for evaluation of microbial denitrification in a sandy aquifer. Journal of Hydrology 114, 413e424. Bo¨ttcher, M.E., Thamdrup, B., Vennemann, T.W., 2001. Oxygen and sulfur isotope fractionation during anaerobic bacterial disproportionation of elemental sulfur. Geochimica et Cosmochimica Acta 65, 1601e1609. Bottrell, S.H., Hayes, P.J., Bannon, M., Williams, G.M., 1995. Bacterial sulfate reduction and pyrite formation in a polluted sand aquifer. Geomicrobiology Journal 13, 75e90. Bottrell, S.H., Hatfield, D., Bartlett, R., Spence, M.J., Bartle, K.D., Mortimer, R.J.G., 2010. Concentrations, sulfur isotope compositions and origin of organosulfur compounds in pore waters of a highly polluted raised peatland. Organic Geochemistry 41, 55e62. Braeckevelt, M., Mirschel, G., Wiessner, A., Rueckert, M., Reiche, N., Vogt, C., Schultz, A., Paschke, H., Kuschk, P., Kaestner, M., 2008. Treatment of chlorobenzenecontaminated groundwater in a pilot-scale constructed wetland. Ecological Engineering 33, 45e53. Brix, H., 1997. Do macrophytes play a role in constructed treatment wetlands? Water Science and Technology 35 (5), 11e17. Brunner, B., Bernasconi, S.M., Kleikemper, J., Schroth, M.H., 2005. A model for oxygen and sulfur isotope fractionation in sulfate during bacterial sulfate reduction processes. Geochimica et Cosmochimica Acta 69, 4773e4785. Bulc, T.G., 2006. Long term performance of a constructed wetland for landfill leachate treatment. Ecological Engineering 26, 365e374. Buisman, C., Uspeert, P., Janssen, A., Lettinga, G., 1990. Kinetics of chemical and biological sulphide oxidation in aqueous solutions. Water Research 24, 667e671. Chazarenc, F., Gagnon, V., Comeau, Y., Brisson, J., 2009. Effect of plant and artificial aeration on solids accumulation and biological activities in constructed wetlands. Ecological Engineering 35, 1005e1010. Choi, J.H., Park, S.S., Jaffe´, P.R., 2006. The effect of emergent macrophytes on the dynamics of sulfur species and trace metals in wetland sediments. Environmental Pollution 140, 286e293. Colberg, P.J.S., 1990. Role of sulfate in microbial transformations of environmental contaminants: chlorinated aromatic compounds. Geomicrobiology Journal 8, 147e165. Colmer, T., 2003. Long distance transport of gases in plants: a perspective on internal aeration and radial oxygen loss from roots. Plant, Cell & Environment 26, 17e36. Faulwetter, J.L., Gagnon, V., Sundberg, C., Chazarenc, F., Burr, M. D., Brisson, J., Camper, A.K., Stein, O.R., 2009. Microbial processes influencing performance of treatment wetlands: a review. Ecological Engineering 35, 987e1004. Finster, K., Liesack, W., Thamdrup, B., 1998. Elemental sulfur and thiosulfate disproportionation by Desulfocapsa sulfoexigens sp. nov., a new anaerobic bacterium isolated from marine surface sediment. Applied and Environmental Microbiology 64, 119. Fry, B., Gest, H., Hayes, J., 1988. 34S/32S fractionation in sulfur cycles catalyzed by anaerobic bacteria. Applied and Environmental Microbiology 54, 250.

6698

w a t e r r e s e a r c h 4 5 ( 2 0 1 1 ) 6 6 8 8 e6 6 9 8

Garcia, J., Rousseau, D.P.L., Morato, J., Lesage, E., Matamoros, V., Bayona, J.M., 2010. Contaminant removal processes in subsurface-flow constructed wetlands: a review. Critical Reviews in Environmental Science and Technology 40, 561e661. Geurts, J.J.M., Sarneel, J.M., Willers, B.J.C., Roelofs, J.G.M., Verhoeven, J.T.A., Lamers, L.P.M., 2009. Interacting effects of sulphate pollution, sulphide toxicity and eutrophication on vegetation development in fens: a mesocosm experiment. Environmental Pollution 157, 2072e2081. Holmer, M., Storkholm, P., 2001. Sulphate reduction and sulphur cycling in lake sediments: a review. Freshwater Biology 46, 431e451. Holmkvist, L., Ferdelman, T.G., Jørgensen, B.B., 2011. A cryptic sulfur cycle driven by iron in the methane zone of marine sediment (Aarhus Bay, Denmark). Geochimica et Cosmochimica Acta. Kno¨ller, K., Fauville, A., Mayer, B., Strauch, G., Friese, K., Veizer, J., 2004. Sulfur cycling in an acid mining lake and its vicinity in Lusatia, Germany. Chemical Geology 204, 303e323. Kno¨ller, K., Schubert, M., 2010. Interaction of dissolved and sedimentary sulfur compounds in contaminated aquifers. Chemical Geology, 284e293. Kno¨ller, K., Vogt, C., Feisthauer, S., Weise, S., Weiss, H., Richnow, H., 2008. Sulfur cycling and biodegradation in contaminated aquifers: insights from stable isotope investigations. Environmental Science & Technology 42, 7807e7812. Kno¨ller, K., Vogt, C., Richnow, H.-H., Weise, S.M., 2006. Sulfur and oxygen isotope fractionation during benzene, toluene, ethyl benzene, and xylene degradation by sulfate-reducing bacteria. Environmental Science & Technology 40, 3879e3885. Krishnakumar, B., Manilal, V., 1999. Bacterial oxidation of sulphide under denitrifying conditions. Biotechnology Letters 21, 437e440. Lawson, G., 1985. Cultivating Reeds (Phragmites Australis) for Root Zone Treatment of Sewage. Contract Report ITE Project 965. Water Research Centre, Cumbria, UK. Lee, R., Kraus, D., Doeller, J., 1999. Oxidation of sulfide by Spartina alterniflora roots. Limnology and Oceanography 44, 1155e1159. Leon, P.M.L., Falla, S.-J., Samborska, E.M., Dulken, I.A.R.v., Hengstum, G.v., Roelofs, J.G.M., 2002. Factors controlling the extent of eutrophication and toxicity in sulfate-polluted freshwater wetlands. Limnology and Oceanography 47, 585e593. Liesack, W., Schnell, S., Revsbech, N.P., 2000. Microbiology of flooded rice paddies. FEMS Microbiology Reviews 24, 625e645. Londry, K., Suflita, J., 1999. Use of nitrate to control sulfide generation by sulfate-reducing bacteria associated with oily waste. Journal of Industrial Microbiology and Biotechnology 22, 582e589. Lovley, D., Phillips, E., 1986. Organic matter mineralization with reduction of ferric iron in anaerobic sediments. Applied and Environmental Microbiology 51, 683. Machemer, S.D., Reynolds, J.S., Laudon, L.S., Wildeman, T.R., 1993. Balance of S in a constructed wetland built to treat acid mine drainage, Idaho Springs, Colorado, USA. Applied Geochemistry 8, 587e603. Mahmood, Q., Zheng, P., Hayat, Y., Ren-cun, J., Azim, M.R., Jilani, G., Islam, E., Ahmed, M., 2009. Effect of nitrite to sulphide ratios on the performance of anoxic sulphide oxidizing reactor. Arabian Journal for Science and Engineering 34, 45e54. Mangalo, M., Meckenstock, R.U., Stichler, W., Einsiedl, F., 2007. Stable isotope fractionation during bacterial sulfate reduction

is controlled by reoxidation of intermediates. Geochimica et Cosmochimica Acta 71, 4161e4171. Merbach, W., Mirus, E., Knof, G., Remus, R., Ruppel, S., Russow, R., Gransee, A., Schulze, J., 1999. Release of carbon and nitrogen compounds by plant roots and their possible ecological importance. Journal of Plant Nutrition and Soil Science 162, 373e383. Picek, T., Cizkova, H., Dusek, J., 2007. Greenhouse gas emissions from a constructed wetland-Plants as important sources of carbon. Ecological Engineering 31, 98e106. Rahman, K., Wiessner, A., Kuschk, P., Mattusch, J., Ka¨stner, M., Mu¨ller, R., 2008. Dynamics of arsenic species in laboratory-scale horizontal subsurface-flow constructed wetlands treating an artificial wastewater. Engineering in Life Sciences 8, 603e611. Rees, C., 1973. A steady-state model for sulphur isotope fractionation in bacterial reduction processes. Geochimica et Cosmochimica Acta 37, 1141e1162. Rethmeier, J., Rabenstein, A., Langer, M., Fischer, U., 1997. Detection of traces of oxidized and reduced sulfur compounds in small samples by combination of different highperformance liquid chromatography methods. Journal of Chromatography A 760, 295e302. Robertson, W., Schiff, S., 1994. Fractionation of sulphur isotopes during biogenic sulphate reduction below a sandy forested recharge area in south-central Canada. Journal of Hydrology 158, 123e134. Schroth, M.H., Kleikemper, J., Bolliger, C., Bernasconi, S.M., Zeyer, J., 2001. In situ assessment of microbial sulfate reduction in a petroleum-contaminated aquifer using pushpull tests and stable sulfur isotope analyses. Journal of Contaminant Hydrology 51, 179e195. Stein, O.R., Borden-Stewart, D.J., Hook, P.B., Jones, W.L., 2007. Seasonal influence on sulfate reduction and zinc sequestration in subsurface treatment wetlands. Water Research 41, 3440e3448. Steinmann, P., Shotyk, W., 1997. Chemical composition, pH and redox state of sulphur and iron in complete vertical profiles from two Sphagnum peat bogs, Jura mountains, Switzerland. Geochimica et Cosmochimica Acta 61, 1143e1163. Sturman, P., Stein, O., Vymazal, J., Kro¨pfelova´, L., 2008. Sulfur cycling in constructed wetlands. In: Vymazal, J. (Ed.), Wastewater Treatment, Plant Dynamics and Management in Constructed and Natural Wetlands. Springer, pp. 329e344. Turchyn, A.V., Bruchert, V., Lyons, T.W., Engel, G.S., Balci, N., Schrag, D.P., Brunner, B., 2010. Kinetic oxygen isotope effects during dissimilatory sulfate reduction: a combined theoretical and experimental approach. Geochimica et Cosmochimica Acta 74, 2011e2024. Webb, J., McGinness, S., Lappin-Scott, H., 1998. Metal removal by sulphate-reducing bacteria from natural and constructed wetlands. Journal of Applied Microbiology 84, 240e248. Wiessner, A., Rahman, K.Z., Kuschk, P., Ka¨stner, M., Jechorek, M., 2010. Dynamics of sulphur compounds in horizontal subsurface flow laboratory-scale constructed wetlands treating artificial sewage. Water Research 44, 6175e6185. Woulds, C., Ngwenya, B.T., 2004. Geochemical processes governing the performance of a constructed wetland treating acid mine drainage, Central Scotland. Applied Geochemistry 19, 1773e1783. Zheng, P., 2007. Anoxic sulfide biooxidation using nitrite as electron acceptor. Journal of Hazardous Materials 147, 249e256. Zheng, P., Cai, J., 2007. Comparison of anoxic sulfide biooxidation using nitrate/nitrite as electron acceptor. Environmental Progress 26, 169e177.