Environ. Sci. Technol. 2009, 43, 7098–7104
Application of Electrolysis to Stimulate Microbial Reductive PCE Dechlorination and Oxidative VC Biodegradation SVENJA T. LOHNER AND ANDREAS TIEHM* Water Technology Center, Department of Environmental Biotechnology, Karlsruher Straβe 84, 76139 Karlsruhe, Germany
Received March 25, 2009. Revised manuscript received July 21, 2009. Accepted July 22, 2009.
A novel approach was applied to stimulate biodegradation of chloroethenes by a coupled bioelectro-process. In a flow-through column system, microbial dechlorination of tetrachloroethene to cis-dichloroethene, vinyl chloride, and ethene was stimulated by hydrogen produced by water electrolysis. Dechlorinating bacteria (Dehalococcoides spp. and Desulfitobacterium spp.) and also methanogens and homoacetogens were detected in the anaerobic column. Simultaneously, oxidative biodegradation of lower chlorinated metabolites (vinyl chloride) was stimulated by electrolytic oxygen formation in the corresponding aerobic column. The process was stable for more than 100 days and an average removal of approximately 23 µmol/d PCE and 72 µmol/d vinyl chloride was obtained with a current density of 0.05 mA/cm2. Abiotic electrochemical degradation of the contaminants was not observed. Microbial dechlorination correlated with the current densities in the applied range of 0.01-0.05 mA/cm2. The results are promising for environmental applications, since with electrolysis hydrogen and oxygen can be supplied continuously to chloroethene degrading microorganisms, and the supply rates can be easily controlled by adjusting the electric current.
Introduction Groundwater contamination is a major problem worldwide, in particular for countries deriving their drinking water from groundwater resources. Chloroethenes are among the most problematic contaminants because of their extensive use, toxicity, and their persistence in the environment (1). Chloroethenes form persistent plumes of contaminated groundwater, since microbial degradation in the field is often limited by a lack of electron donors for reductive dechlorination and electron acceptors for oxidative degradation (2, 3). During microbial reductive dechlorination, chloroethenes are sequentially dechlorinated from tetrachloroethene, via trichloroethene (TCE), cis-dichloroethene (cis-DCE), and vinyl chloride (VC), to ethene. Different complex organic substrates have been supplied to microbial communities to support reductive dechlorination (4-6). However, hydrogen was identified as the key electron donor for reductively dechlorinating bacteria such as Desulfitobacterium (7), Dehalo* Corresponding author phone: +49-721-9678-220; fax: +49-7219678-101; e-mail:
[email protected]. 7098
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 18, 2009
coccoides (8), Sulfurospirillum (9), and Dehalobacter (10), with Dehalococcoides being the only known microorganism capable of total dechlorination of PCE to ethene (11, 12). Transformation of cis-DCE and VC to ethene has been shown to be the rate limiting step (13) and in the field, long lasting plumes of VC or cis-DCE are observed (14). The lower chlorinated metabolites, such as VC and cis-DCE, can additionally undergo oxidative degradation with oxygen as electron acceptor (15-17). In particular, bacteria such as Polaromonas vacuolata, Mycobacterium sp., and Nocardioides sp. have been demonstrated to be capable of aerobic cisDCE and VC degradation (15, 16). In the field, bioremediation strategies often involve the addition of H2-releasing compounds and complex substrates such as molasses are applied in stoichiometric excess. Thus, other redox reactions and biomass growth are pronounced, leading to aquifer clogging due to excessive biomass growth or to the accumulation of fermentation products in the groundwater (18-20). A promising alternative to provide hydrogen and oxygen in bioreactors or in situ applications is water electrolysis. Recently there has been an increasing interest in the use of combined bioelectro-processes to improve bioremediation efficiencies (18, 21-24). So far, most studies deal with electrolytically enhanced biological nitrogen removal (23-30). A single study from Skadberg et al. (31) demonstrates the electrolytically enhanced degradation of 2,6-dichlorophenol. Also, the use of electrolytically formed hydrogen for methane production was reported that subsequently stimulated aerobic cometabolic chloroethene degradation (24). Another approach proposed by Aulenta et al. (18, 32) and Thrash et al. (22) involves the use of a redox mediator for electron shuttling from the electrode to the microorganisms for the stimulation of reductive dechlorination and perchlorate reduction, respectively. Also the direct transfer of electrons from the electrodes to bacteria to catalyze microbial reactions has been investigated (33). In this study, the feasibility of electrolytically stimulated biodegradation was examined for reductive dechlorination and oxidative degradation processes. According to the conceptual model (Figure 1), competing hydrogen consuming processes such as methanogenesis and homoacetogenesis were also taken into consideration. To our knowledge, this is the first report on reductive and oxidative bioelectroremediation of chloroethenes in a flow-through system. Degradation rates were investigated as a function of the applied electric current density, and the bioelectro-process efficiency was evaluated.
Materials and Methods Dechlorinating Cultures and Media. The anaerobic culture was originally isolated from an anoxic aquifer in which dechlorination took place from PCE to ethene and enriched with PCE and yeast extract (1 g/L). The presence of Dehalococcoides species and Desulfitobacterium as well as methanogenic and homoacetogenic bacteria was demonstrated by polymerase chain reaction (PCR) (34). The anaerobic culture was incubated in 1 L borosilicate glass bottles sealed by gastight Teflon-coated caps. A chloride-free mineral medium was used to enable mass balances based on chloride analysis. The anaerobic basal medium (pH 7.2) contained the following constituents per liter of demineralized water: 80 mL phosphate buffer (1 mol/L KH2PO4, 2 mol/L Na2HPO4), 10 mL salt solution (330 mmol KH2PO4, 340 mmol (NH4)2HPO4), sodium carbonate (100 mg/L carbonate), 5.1 mL trace elements solution (35, 9), MgHPO4 × 3 H2O (0.8 mmol), and 1 mL 10.1021/es900835d CCC: $40.75
2009 American Chemical Society
Published on Web 08/13/2009
FIGURE 1. Conceptual model of the coupled bioelectro-process and generation of hydrogen and oxygen. MM: Mineral medium. vitamin solution (36). Transfer of 200-300 mL culture into 700-800 mL fresh mineral medium was done every 4 weeks after 2-3 PCE (10 µL) dechlorination cycles. The aerobic enrichment culture was originally isolated from an oxic aquifer in which aerobic degradation of VC was observed, and enriched with VC and atmospheric oxygen. Cultures were maintained in 2 L borosilicate glass bottles, sealed with Teflon-coated screw caps. A modified mineral medium according to Lochhead and Chase (37) was used for enrichment. A 1:1 transfer of aerobic culture into fresh mineral medium was done every 8 weeks. Vinyl chloride was regularly added as substrate and atmospheric oxygen was provided as electron acceptor. All cultures were incubated in the dark at room temperature on a rotary shaker. Bio-Electrochemical Column System and Operation. The column system consisted of two identical borosilicate columns (20 cm length, 3.5 cm inside diameter) connected by an s-shaped glass tube (Figure 1). A conducting bipolar membrane (Fumasep FBM, Fumatech GmbH, St. Ingbert) was inserted in the middle of the adjoining tube. The columns were filled with sterile DORSILIT 9S silica sand (porosity 0.38) on which the anaerobic and the aerobic cultures were immobilized by recirculating the corresponding enrichment culture through the column for several months prior to the experiment. Stainless steel (16% Cr, 10% Ni) meshes (10 cm2) were used as electrodes which were connected to a power supply (model 1134-150, Heiden power GmbH, Pu ¨ rgen) by a titanium wire (1 mm, Goodfellow GmbH, Friedberg). Constant current densities applied ranged between 0-0.1 mA/cm2, the voltage varied from 9-20 V. Oxygen removal in the feed solution was achieved by pumping the medium through a membrane tube in a bottle with NaSO3-solution (100 g/L). No external carbon source was added to the anaerobic medium. Since homoacetogenic bacteria were detected in the inoculum, acetate was considered the available carbon source. PCE and VC, respectively, were added to the reservoirs. The reservoirs had to be refilled one time during the experiment. The PCE concentration was 24 µmol/L during the first period (day 1-56) and approximately 45 µmol/L during the second period (day 56-115). VC concentrations were 352 µmol/L (first period) and 224 µmol/L (second period). The medium was pumped
experimental setup. Biodegradation is stimulated by electrolytic into the columns with a flow rate of 0.5 L/d resulting in hydraulic retention times of approximately 3 h for each column. Samples were taken periodically from the influent and effluent of the columns for chemical analysis. Microbial Community PCR and MPN Analysis. For microbiological analysis, 5 g wet sand was taken out from the columns at ports 2 and 3 (SI Figure S1). The samples were rinsed in 5 mL 0.9% NaCl solution by moderate shaking by hand for 1 min followed by PCR and most probable number (MPN) analysis of the supernatant. Next, the samples were sonicated in a 2 L ultrasonic bath operated at 35 kHz with 2 × 200 W electrical power input. The sand samples were sonicated for 3 min in 100 mL of 0.9% NaCl solution. The extract was used for subsequent analysis. For PCR analysis, 4 mL of the washing solution was centrifuged at 6000g for 60 min at 4 °C and the supernatant was discarded. The genomic DNA was extracted from cell pellets and 1 g of the washed sand material, respectively, by using the UltraClean Soil DNA Kit (Mo Bio Laboratories, Carlsbad, CA). Specific PCR analysis of DNA extracts was performed with primer pairs for amplification of the homoacetogenic formyltetrahydrofolate synthetase (FHTS) gene, primers Ac-F and Ac-R (38), and the methanogenic methyl coenzyme-M reductase (mcrA) gene, primers LuF and LuR (39). For detection of dechlorinating bacteria, a twostep nested PCR was used. The initial amplification was performed using a pair of universal bacterial primers according to Hendrickson et al. (40) followed by a second PCR with Dehalococcoides, Desulfitobacterium, Dehalobacter, or Desulfuromonas targeted primers, which have been reported previously (40-42). PCR amplifications were performed with a T-personal thermocycler (Biometra, Goettingen, Germany). The MPN assay to determine the numbers of aerobic VC degrading bacteria was done as previously described (43). Chemicals. Tetrachloroethene was obtained from Fluka (99.9+%) and gaseous vinyl chloride (VC) was purchased from Linde (Chloroethene 3.7). All other chemicals were of analytical grade and used as received. Analytical Methods. Gas Chromatography. PCE, TCE, cis-DCE, VC, ethene, and methane were routinely determined after sampling of VOL. 43, NO. 18, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7099
a 5 mL aqueous sample. A drop of concentrated H3PO4 was amended for conservation. Samples were measured on the same day or stored in the fridge overnight. Analysis of chloroethenes was performed by a gas chromatograph from Hewlett-Packard (Series II 5890) equipped with a headspace autosampler (HP 19395 A, 70 °C, 3 h), flame ionization detector (FID) and electron capture detector (ECD) (detection limits in µmol/L: PCE: 0.0025, TCE: 0.0026, cDCE: 0.96, VC: 4.90, ethene: 4.64, methane: 2.88). Separation was accomplished in a 50 m-capillary column (HP PONA, 0.21 mm inside diameter, 0.5 µm methyl silicon film). Helium was used as carrier gas at a flow rate of 0.7 mL/min. The injector and detector temperatures were 180 °C and 250 °C (FID) and 300 °C (ECD), respectively. The temperature program was 15 min at 35 °C, increase to 60 °C (1.5 °C min-1), increase to 130 °C (15 °C min-1), increase to 200 °C (30 °C min-1), held for 5 min. At least duplicate measurements were conducted for every sample. In the abiotic control experiments, the formed hydrogen gas in the reductive column was collected in the effluent. The volume was measured regularly and 1 mL gas samples were taken with a gastight syringe (Hamilton, Reno, NV). Hydrogen measurement was performed by a gas chromatograph from Agilent Technologies (6890N) equipped with a thermal conductivity detector (detection limit H2: 1.5 µmol/ L). Separation was accomplished in a coupled column system (Poraplot, 27.5 m length, 0.53 mm inside diameter, 20 µm film thickness, and Molesieve, 25 m length, 0.53 mm inside diameter, 50 µm film thickness). Nitrogen was used as carrier gas at a flow rate of 29.5 mL/min. The injector and detector temperatures were 250 °C. Ion Chromatography. Chloride and acetate concentrations were determined with an ion chromatograph (Metrohm 761 Compact IC, Metrohm) with a MetrohmA-Supp-5 column (length 150 mm, 4 mm inside diameter) with suppression system and conductivity detector (detection limit: 1 mg/L). Measurements of pH, temperature and oxygen in the column effluents were done with a multimeter (model: MultiLine P4, WTW, Weilheim, Germany).
FIGURE 2. (A) Electric current density ((), (B) PCE transformation (b), (C) formation of metabolites ( × TCE, ∆ cis-DCE, 0 VC, 2 ethene, 9 sum metabolites), and (D) chloride formation (O) during electrolytically stimulated reductive dechlorination of PCE in the anaerobic column. Influent PCE concentrations were 24 µmol/L (day 1-56) and approximately 45 µmol/L (day 56-115). The flow was 0.5 L/d during the whole experiment.
Results Electrolytically Stimulated Reductive Dechlorination of PCE. Figure 2 shows the effect of the applied current (Figure 2A) on chloroethene transformation (Figure 2B, and C) and chloride formation (Figure 2D). In phase I of the experiment the feasibility of electrolytically stimulated reductive dechlorination was demonstrated. After applying 0.03 mA/cm2, followed by 0.05 mA/cm2, electrolytical production of hydrogen enabled microbial dechlorination. On day 19, PCE and TCE concentrations in the effluent were below the detection limit. Detected metabolites were cis-DCE, VC and ethene with 79% cis-DCE, 12% VC, and 9% ethene at 0.05 mA/cm2 (Figure 2C). When the current was switched off, dechlorination immediately stopped confirming that previously observed PCE transformation was triggered by electric current. In phase II, dechlorination was assessed as a function of the applied electric current density. The current density was incrementally increased from 0 to 0.05 mA/cm2 followed by a stepwise decrease (Figure 2A). After a short lag-phase a decrease of effluent PCE was observed and TCE and cis-DCE were increasingly detected in the effluent. Dechlorination to VC and ethene started when the current density exceeded 0.04 mA/cm2. At 0.05 mA/cm2 a similar metabolite composition was observed in the effluent as during phase I. On day 56 the reservoir was refilled and the influent concentration of PCE was doubled. Due to the less favorable ratio of PCE and produced hydrogen, TCE was detected again in the effluent. cis-DCE was the main metabolite of microbial reductive dechlorination. However, VC and ethene were also detected in significant amounts. After a period of 20 days, 7100
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 18, 2009
FIGURE 3. Correlation of reductive dechlorination activity with applied electric current density (9 chloroequivalent elimination, ( chloride formation). the current density was decreased again. The changing PCE/ mA ratio with decreasing current density resulted in increasing PCE concentrations in the effluent and decreasing metabolite formation. Again, microbial dechlorination stopped when no current was applied. For correlation of dechlorination activity with applied current, chloroequivalent concentrations were calculated accounting for the incorporated chlorine within the chloroethene molecules ([PCE] ) 4, [TCE] ) 3, [cis-DCE] ) 2, [VC] ) 1, [ethene] ) 0 chloroequivalents). The correlation was determined by regression analysis and the results show a linear correlation (R2 ) 0.96) (Figure 3). This indicates that at low current densities the availability of bioavailable hydrogen is limiting the biodegradation efficiency which can be overcome by increasing hydrogen production rates. The
highest chloroequivalent removal (49%) was achieved with an applied current density of 0.05 mA/cm2. Microbial community analysis revealed the presence of reductively dechlorinating bacteria as well as other hydrogenutilizing bacteria (SI Figure S1). Desulfitobacterium spp. which are known to transform PCE to cis-DCE, and Dehalococcoides spp. which can catalyze the whole dechlorination sequence from PCE to ethene, were detected in soil samples taken at ports 2 and 3. Other hydrogen-consuming microorganisms detected on the sand were homoacetogenic and methanogenic bacteria. Hydrogen consumption by dechlorinating bacteria was compared to electrolytically produced hydrogen for assessment of process efficiency. Based on Faraday’s Laws, which can be summarized as I · t ) z · F · n (with I ) current intensity [A], t ) time [s], z ) electrons transferred, F ) Faradayconstant [96485.3 C/mol], n ) amount of substance produced at the electrode [mol]), the theoretical formation of hydrogen and oxygen by water electrolysis was calculated. Theoretically, at current densities exceeding 0.02 mA/cm2, hydrogen was produced in amounts higher than stoichiometrically necessary for complete reductive dechlorination of influent PCE. In abiotic control experiments, the production of hydrogen was proven with amounts increasing with increasing current density (SI Table S1). Based on PCE transformation and metabolite formation, about 10-13% of electrolytically generated hydrogen was actually used for dechlorination. Additionally, in the cathodic column a significant amount of methanogenic activity was observed, which was also dependent on the applied current density (data not shown). The calculated hydrogen consumption for methanogenesis was about 25%. As indicated by PCR (SI Figure S1), homoacetogenesis also has to be considered in the anaerobic column although acetate never exceeded the detection limit. Electrolytically Stimulated Oxidative Degradation of VC. Oxidative VC degradation was stimulated simultaneously with reductive dechlorination in the aerobic column. Figure 4 shows the applied current density (Figure 4A), VC degradation (Figure 4B) and chloride formation (Figure 4C) throughout the experiment. VC was added in excess to ensure that it was not the limiting factor for biodegradation. When the electrolysis was started, degradation of VC began without a lag-phase. With 0.05 mA/cm2, an average VC degradation of 72 µmol/d was achieved. After the current was switched off, degradation stopped, indicating that electrogenerated oxygen was responsible for the observed VC degradation. Again, no pH changes were detected during electrolytic treatment. In phase II of the experiment, the effect of current on oxidative degradation was investigated as compared to reductive dechlorination (Figure 4). Based on VC degradation and corresponding chloride formation, the oxidative degradation activity showed a significant linear correlation with the applied current density (R2 ) 0.92) (Figure 5), similar to reductive dechlorination. These results demonstrate that bioavailable oxygen was the limiting factor for biodegradation and that the mineralization of VC could be stimulated with increasing oxygen production. Microbial analysis of soil samples taken at port 2 and 3 of the aerobic column demonstrated the presence of VC degrading bacteria. At port 2, 1.7 × 104 VC-degraders/g sand were detected by the three moderate washing steps and 1.7 × 103 VC-degraders/g soil by the rigorous elution done to transfer immobilized bacteria into the aqueous phase. In the soil from port 3 the VC-degraders were detected in slightly lower numbers. Thus, VC degrading bacteria were present in the whole column, and a significant fraction was immobilized in the biofilm covering the soil particles. A stoichiometric ratio for VC and oxygen of 1:2.5 requires 375 µmol/L oxygen for complete mineralization of 150
FIGURE 4. (A) Electric current density (s2s), (B) VC degradation (9), and (C) chloride formation (b) during electrolytically stimulated oxidative degradation of VC in the aerobic column. Influent VC concentrations were 352 µmol/L (day 1-56) and approximately 224 µmol/L (day 56-115). The flow was 0.5 L/d during the whole experiment.
FIGURE 5. Correlation of oxidative VC degradation activity with applied electric current density (9 VC degradation, ( Chloride formation). µmol/L VC. Effluent oxygen concentrations were generally
