A Bench-Scale Constructed Wetland As a Model to Characterize

Oct 20, 2011 - Sherwood Lollar , B.; Slater , G. F.; Sleep , B.; Witt , M.; Klecka , G. M.; Harkness , M.; Spivack , J. Stable carbon isotope evidence...
0 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/est

A Bench-Scale Constructed Wetland As a Model to Characterize Benzene Biodegradation Processes in Freshwater Wetlands Jana Rakoczy,* Benjamin Remy,† Carsten Vogt, and Hans H. Richnow Department of Isotope Biogeochemistry, Helmholtz Centre for Environmental ResearchUFZ, Leipzig, Germany

bS Supporting Information ABSTRACT: In wetlands, a variety of biotic and abiotic processes can contribute to the removal of organic substances. Here, we used compound-specific isotope analysis (CSIA), hydrogeochemical parameters and detection of functional genes to characterize in situ biodegradation of benzene in a model constructed wetland over a period of 370 days. Despite low dissolved oxygen concentrations (98% removal), we applied CSIA to study in situ benzene degradation by indigenous microbes. Combining carbon and hydrogen isotope signatures by two-dimensional stable isotope analysis revealed that benzene was degraded aerobically, mainly via the monohydroxylation pathway. This was additionally supported by the detection of the BTEX monooxygenase gene tmoA in sediment and root samples. Calculating the extent of biodegradation from the isotope signatures demonstrated that at least 85% of benzene was degraded by this pathway and thus, only a small fraction was removed abiotically. This study shows that model wetlands can contribute to an understanding of biodegradation processes in floodplains or natural wetland systems.

’ INTRODUCTION During the past decades, constructed wetlands (CW) have become increasingly recognized as a low-cost treatment option for different kinds of wastewater. One large field of application is the treatment of wastewater from the petroleum industry,1 which often contains benzene, a constituent from the gasoline production. Benzene is a relatively water-soluble, highly abundant groundwater contaminant which can pose a risk to human health.2 A number of studies demonstrated that CWs are capable of improving water quality by efficiently removing benzene and other volatile organic compounds from groundwater or industrial waste waters.38 Ecosystems like natural wetlands or floodplains can as well provide physical and biochemical properties promoting biodegradation of contaminants, and may thus act as a reactive barrier that prevents contaminant fluxes from groundwater to adjacent surface waters. A variety of physicochemical and biological processes can facilitate the removal of benzene. Volatilization, sorption, and dilution merely relocate contaminants. Plants can contribute to benzene removal by uptake with the transpiration stream9,10 and subsequent transfer into the atmosphere.11 Phenols and polyaromatic hydrocarbons can also be metabolized by plants.12,13 In the present study, we focused on microbial degradation as it allows for complete mineralization of harmful substances and has been considered the most important removal process.6,7,14,15 In the last years, knowledge on microbial processes in wetlands proliferated and a recent review16 summarizes the techniques that have been employed to assess microbial activity in wetlands. Regarding the degradation of aromatic and aliphatic hydrocarbons, those techniques include r 2011 American Chemical Society

laboratory studies like batch cultivation or microbial counts of hydrocarbon degraders from contaminated sites. For example, microbial population studies on hydrocarbon-contaminated soils demonstrated that microbial densities were higher in the presence of hydrocarbons compared to unpolluted soils, which the authors ascribed to a higher density of hydrocarbon degraders.6,13,15,17 In a recent study,18 in situ microcosms were used to enrich benzene-degrading bacteria from methanogenic wetland sediments. Subsequently, benzene degradation rates were determined in laboratory tests and were found to be similar to those determined in the adjacent aquifer. Another widely used approach is the soil respiration test where the production of carbon dioxide is used to estimate the degradation capacity of a contaminated soil.6,17,19 Although these techniques can help to roughly estimate the biodegradation potential, most studies still lack an immediate proof for intrinsic microbial degradation activity16,20 and its detection among other removal processes remains challenging. As a complementary approach, hydrogeochemical parameters (e.g., consumption of electron acceptors, changes in redox potential) are often monitored to support the determination of aerobic and anaerobic degradation pathways as it has been shown for BTEX compounds6 and chlorinated ethenes.21 The availability of oxygen is of particular interest because oxygen acts as the most efficient electron acceptor in microbial respiration processes, and is also used as cosubstrate for the oxygenation of Received: July 29, 2011 Accepted: October 20, 2011 Revised: October 17, 2011 Published: October 20, 2011 10036

dx.doi.org/10.1021/es2026196 | Environ. Sci. Technol. 2011, 45, 10036–10044

Environmental Science & Technology several compounds. The main pathways for oxygen entering the wetland are (i) through oxygen release from plant roots, (ii) by atmospheric diffusion or (iii) with the inflowing water.22 However, oxygen fluxes within wetlands are difficult to track as it is rapidly consumed during various processes like the microbial oxidation of plant-derived organic carbon or root exudates.2325 The evaluation of a wetland as a sustainable option for reducing contaminant fluxes requires a sound knowledge on the degradation performance of the residing microbial communities. An elegant way to detect in situ biodegradation in a complex environment is compound-specific isotope analysis (CSIA). It takes advantage of a shift in the compounds stable isotope composition (13C/12C, 2 H/1H), which occurs mainly during microbial degradation but remains largely unaffected from abiotic removal processes, for example, sorption or volatilization.2629 Beside the identification of degradation pathways using simultaneous isotope analysis of two elements (two-dimensional stable isotope analysis), CSIA can be employed to quantitatively assess biodegradation among other removal processes. These approaches have so far provided insights into natural attenuation of BTEX-contaminated groundwater systems,3034 but have not yet been used to evaluate biodegradation of benzene in wetlands. In the present study, we used CSIA to investigate in situ biodegradation of benzene in a constructed wetland which served as a model for a planted riparian zone connecting an aquifer with an open water body. Our model wetland was fed with groundwater from a refinery site containing benzene and ammonium as main contaminants. The main goal of the study was to characterize the dominant benzene degradation pathways and to quantify the microbial share within the total benzene removal using CSIA as an immediate detection method. Additionally, we monitored the hydrogeochemical development of the wetland as well as the distribution of a functional gene of the benzene degradation pathway.

’ MATERIALS AND METHODS Set-Up of a Bench-Scale Constructed Wetland System. The setup of the wetland system is described in detail in a previous study.21 Briefly, the system consisted of a stainless steel tank (201  60  5 cm), with a glass panel on the front side (Supporting Information SI-Figure 1). The tank was filled with sand (grain size 0.40.63 mm) and planted with Juncus effusus. In order to mimic groundwater conditions, the wetland was connected with a cooling system to sustain the temperature of the sand compartment at 1012 °C. Four inlet pipes for the injection of groundwater were located at a height of 4, 12, 20, and 36 cm above the bottom of the tank. At 150 cm from the inflow, an open water compartment was created, simulating water drainage to an open surface water body. Twelve sampling ports for extraction of pore water were located at 6, 49, 94, and 139 cm from the inflow and at 12, 24, and 36 cm from the bottom (Figure 1, Supporting Information SI-Figure 1b). The system was continuously supplied with groundwater from a benzenecontaminated aquifer at a former refinery site located in Leuna, Germany.35 The start of the groundwater supply to the yet unpolluted system was defined as day 0. The system was operated in a horizontal subsurface flow mode, with the water table at 46 ( 1 cm from the bottom. The groundwater was fed to the system from a 50 L tank which was kept at 0.5 bar nitrogen overpressure, and was integrated to the above-mentioned cooling system. The inflow rate was set to 2 L d1, using

ARTICLE

a peristaltic pump (IPC 8, Ismatec, Wertheim, Germany), leading to a retention time of approximately 14 days. The composition of the groundwater is reported in the Supporting Information (SI-Table 1). Analysis of Hydrogeochemical Parameters and Benzene Concentration in Pore Water. Water samples were extracted at the inflow, at the 12 sampling ports across the sediment compartment and at the outflow. Four sampling campaigns were conducted within one year (day 41, 231, 293, 345). Samples were analyzed for concentrations of dissolved oxygen, nitrite, nitrate, ammonium, ferrous iron, total manganese, sulfate, sulfide, chloride, phosphate, pH, and redox potential. Benzene concentrations were analyzed at day 41, 231, 293, 345, 370 by gas chromatography as described previously.36 Details on the analytical methods and detection limits are given in the Supporting Information. Solid-Phase Microextraction and Compound-Specific Isotope Analysis. Due to limited volume of extractable pore water, solid-phase microextraction (SPME) was used to extract benzene from pore water samples. For preparation of each sample, 1 mL pore water was filled into a 10 mL vial containing 1 mL saturated sodium chloride solution. Vials were sealed with Teflon-coated caps and stored at 20 °C until analysis, whereby no measurable isotope effect was observed.37 Benzene was extracted by exposing a 75 μm Carboxen-PDMS fiber (Supelco, Bellefonte, PA) to the headspace of the vial for 20 min. During extraction, samples were kept at 60 °C to enhance the partitioning of benzene into the gas phase which, in turn, facilitates adsorption to the fiber. Benzene desorption from the fiber was achieved in a split/splitless injector of a gas chromatograph (GC) held at 300 °C. The inlet of the GC was purged for 5 min and the sample was transferred to the GC column (column properties and temperature program are specified in the Supporting Information). For determination of carbon and hydrogen isotope signatures of benzene, samples were injected at two different gas chromatography isotope ratio mass spectrometry (GC IRMS) systems (detailed information on the analytical procedure is provided in the Supporting Information). Carbon and hydrogen isotope signatures are given in δ-notation (%) relative to the Vienna Pee Dee Belemnite standard (VPDB) and Vienna Standard Mean Ocean Water (VSMOW), respectively (eq 1).  δ13 Csample or δ2 Hsample ð%Þ ¼

Rsample  Rstandard Rstandard

  1000

ð1Þ

Rsample and Rstandard are the ratios 13C/12C or 2H/1H of sample and standard. Quantification of Stable Isotope Fractionation. Carbon and hydrogen enrichment factors (εC, εH) were calculated according the Rayleigh equation (eq 2).38 The Rayleigh equation relates the isotope ratio at time point zero (R0) and a given time point t (Rt) to the corresponding contaminant concentration at time point zero (c0) and t (ct). In our system, in each individual sampling campaign, zero is the concentration or isotope ratio determined at the inflow (c0, R0) to which all concentrations (ct) or isotope ratios (Rt) from this sampling campaign refer. ε is obtained from the slope (m) of the linear regression of ln(Rt/R0) on ln(ct/c0), m = ε/1000.     Rt ε ct ln ln ¼ ð2Þ 1000 c0 R0 Extent of Biodegradation. To calculate the extent of biodegradation [B(%)], the Rayleigh equation is solved for the 10037

dx.doi.org/10.1021/es2026196 |Environ. Sci. Technol. 2011, 45, 10036–10044

Environmental Science & Technology

ARTICLE

Figure 1. Distribution map of hydrogeochemical parameters measured along the water flow path: from the inflow (triangles), across the sediment compartment (circles), to the outflow (big square) (exemplified by day 293). Values 13 50,52,56). In contrast, the highest slope for benzene degradation under oxic conditions was Λ = 11 ( 6, detected in a culture of Cupriavidus necator, which can be considered as the current

ARTICLE

upper limit of oxic degradation. Below this ratio, oxic degradation can additionally be divided into the monohydroxylation and the dihydroxylation pathway. Since dihydroxylation involves cleavage of the CdC double bond but no cleavage or formation of a C—H bond in the rate-limiting step of the biochemical reaction, it will produce a carbon isotope effect but no significant primary hydrogen isotope effect57,58 and thus, data distribute along the δ13C axis. The initial activation step of the monohydroxylation reaction is less well understood and several mechanisms are conceivable.50 As this can include either a C—C or a C—H bond cleavage, depending on the reaction mechanism of the enzyme, it results in a low but significant hydrogen isotope fractionation, with Λ values ranging from 3 to 11.50 This variability in the initial activation step might as well explain the scattering of the isotope signatures in the Rayleigh plots (Supporting Information SIFigure 5). Using these fractionation patterns as a template, we plotted the isotope signatures collected in four sampling campaigns (day 231, 293, 345, 370) in order to determine which of the two pathways was dominant in our wetland system (Figure 2). In detail, data from day 231 showed significant carbon fractionation but no hydrogen fractionation, resulting in a slope of Λ = 1 ( 2 (R2 = 0.39), signifying the dihydroxylation pathway. Some data points overlapped with the monohydroxylation pathway and should therefore be considered with caution. Still, the general absence of hydrogen fractionation combined with significant carbon fractionation suggested that benzene was degraded mainly by dihydroxylation (Λ = 0 ( 5, according to the uncertainty of hydrogen analysis of (5%). At day 293, a significant shift in the hydrogen isotope signature was observed, resulting in a higher Λ value (Λ = 4 ( 5, R2 = 0.51). Taking the relatively high uncertainty of the Λ value into account, degradation could be attributed either to dihydroxylation or monohydroxylation pathway, suggesting that this period is a transition phase in which both pathways coexisted. Day 345 and day 370 could be clearly attributed to monohydroxylation with slopes of 9 ( 2 and 9 ( 3, respectively, showing a good correlation (R2 = 0.98 and 0.86, respectively). Additionally, Figure 2 shows that the isotope pattern for oxic degradation was found across the entire sediment compartment, regardless of the sampling depth. Although the density of the plant roots was observed to be lower in the deeper sediment layers, probably resulting in less oxygen release, the isotope data confirmed that oxic benzene degradation was still the dominant process here. The lower root density might have led to a formation of larger anoxic zones in the bulk sediment which might explain that the remaining benzene concentrations increased with sampling depth since benzene was not as efficiently degraded as in the upper sediment layers (Figure 1). Although, based on the hydrogeochemical data, anoxic benzene degradation did not substantially contribute to the overall benzene removal, it should be noted that a mixing of isotope signatures from adjacent oxic and anoxic degradation zones may have added to the observed increasing Λ values. Such adaptations in the degradation pathway together with hydrogeochemical parameters reflect the systems’ development over time, at a level of complexity between well-defined laboratory microcosms and highly complex field studies. Quantification of Benzene Degradation. Both carbon and hydrogen isotope signatures at each sampling point were used to assess the extent of benzene biodegradation [B(%)] in the wetland system (eq 3). To avoid an overestimation, the calculation was done conservatively by choosing the highest enrichment factors so far known for the particular degradation pathways,59 10040

dx.doi.org/10.1021/es2026196 |Environ. Sci. Technol. 2011, 45, 10036–10044

Environmental Science & Technology

Figure 3. Decreasing benzene concentration measured across the sediment compartment (b) compared to the expected benzene concentration calculated from carbon (0) or hydrogen (Δ) isotope signatures according eqs 3 and 4 (exemplified by day 345). Calculations were performed with εC = 4.3% and εH = 17% as the currently highest enrichment factors detected for the monohydroxylation pathway.

which have been determined beforehand by means of twodimensional stable isotope analysis. For monohydroxylation, B(%) was calculated with εC = 4.3% and εH = 17%,50 and for dihydroxylation, B(%) was only calculated based on carbon signatures with εC = 1.3%,50 since hydrogen enrichment is not observed here. In a second step, B(%) values were used to quantify the amount of theoretically biodegraded benzene taking the inflow concentration as initial concentration (eq 4). These benzene concentrations calculated from each carbon or hydrogen isotope signature were then compared to the benzene concentrations measured across the wetland. All sampling campaigns showed a good accordance of calculated and measured values: For the majority of the samples, measured benzene losses differed from the calculated benzene losses less than 17% and 10% for carbon and hydrogen signatures, respectively (Figure 3 shows day 345 as an example). That would mean, at 100% removal efficiency, at least 85% of the benzene was degraded by bacteria and only a small portion was removed by abiotic processes. Similar calculations have been conducted to quantify microbial degradation in other complex systems like contaminated aquifers. Here, degradation efficiencies ranged from 40% in a TCE-contaminated groundwater60 to up to 90% for BTEX compounds.31,39,6163 Another way to assess biodegradation, is to monitor concentration changes of terminal electron acceptors along the groundwater flow.64 Unlike an aquifer, this approach cannot be applied to a wetland since beside microbial consumption, the plant roots as well can serve as sink or source for electron acceptors (e.g., oxygen release, sulfate uptake), precluding a reliable calculation. For example, to achieve 85% benzene mineralization in the wetland system (which equates approximately 130 μM benzene), 975 μM oxygen is needed. Additionally, oxidation of 110 μM ferrous iron and formation of 220 μM nitrate from ammonium requires 28 μM and 440 μM oxygen, respectively. Within the retention time of about two weeks, at least 1.4 mM oxygen must be released into the sediment in order to render these reactions possible. This in turn shows that by a measurement of oxygen concentrations, which were mostly