Environ. Sci. Technol. 2009, 43, 5144–5149
Electron Fluxes in a Microbial Fuel Cell Performing Carbon and Nitrogen Removal B E R N A R D I N O V I R D I S , †,‡ KORNEEL RABAEY,† ZHIGUO YUAN,† ´ A. ROZENDAL,† AND RENE ¨ R G K E L L E R * ,† JU Advanced Water Management Centre (AWMC), The University of Queensland, St. Lucia, Queensland 4072, Australia, and Department of Geoengineering and Environmental Technologies (DIGITA), Universita` degli Studi di Cagliari, Piazza d’Armi, 09123 Cagliari, Italy
Received December 21, 2008. Revised manuscript received April 9, 2009. Accepted May 5, 2009.
The electron recovery in microbial fuel cells (MFCs) is decreased by processes like methanogenesis, bacterial growth, and the accumulation of intermediates. Using a suite of analytical techniques, including electrochemical monitoring, chemical analysis, microsensor analysis, and Titration and OffGas Analysis (TOGA), this study aimed to (a) identify and quantify the electron losses occurring at the anode and the cathode of a MFC removing acetate and nitrate (NO3-), respectively, and (b) to investigate the impact of the operational characteristics of the cathode on the denitrification process. Our results show that methane (CH4) production and estimated biomass formation at the anode and nitrous oxide (N2O) accumulation at the cathode were responsible for the reduction of Coulombic efficiency (ε) during continuous feeding conditions. At the anode, up to 40.1% of the acetate consumed was released as methane at closed circuit. At the cathode, N2O accumulation represented instead the main loss accounting for up to 10.0 ( 2.1% of the oxidation capacity of the electron acceptor provided as NO3-. Batch experiments at controlled potentials and currents revealed that for a given current the fraction of electron transferred and released as N2O is significantly reduced by low cathodic potentials.
Introduction Bio-Electrochemical Systems (BESs) use microorganisms to catalyze oxidation or reduction reactions (1). The best known BESs are Microbial Fuel Cells (MFCs), which potentially are an alternative to traditional wastewater treatment (2). Briefly, bacteria growing at the anode catalyze the electron transfer from the organic fraction of wastewater to the electrode. The electrons then migrate from the anode through an external circuitry to the cathode where they reduce an electron acceptor. Although oxygen is a prime cathodic oxidant due to its availability, wastewaters contain a variety of other compounds, such as nitrogen oxides, which are suitable as electron acceptors (1). It was recently shown that nitrate (NO3-) (3) and also nitrite (NO2-) (4) could effectively be * Corresponding author phone: +61 7 3365 4727; fax: +61 7 3365 4726; e-mail:
[email protected]. † The University of Queensland. ‡ Universita` degli Studi di Cagliari. 5144
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 13, 2009
used to drive current generation and simultaneous nitrogen and carbon removal. For both the anodic and cathodic reactions, the presence of competitive or incomplete processes can significantly reduce the efficiency of charge transfer, which is commonly referred to as the Coulombic efficiency (ε) (5). Processes like methanogenesis and bacterial growth have been shown to significantly decrease the Coulombic efficiency at the anode as they divert part of the electrons from the electricity generating process (6). At the cathode, intermediate compounds produced during denitrification can become electron sinks. Denitrification is in fact accomplished by a sequence of four reduction steps in which NO3- is reduced to NO2-, nitric oxide (NO), nitrous oxide (N2O), and dinitrogen (N2) gas (7). The accumulation of one of the intermediates would reduce the Coulombic efficiency at the cathode as it decreases the total oxidation capacity of the electron acceptor present. While NO accumulation is normally very limited (8), the accumulation of N2O has been frequently observed in nitrification and denitrification systems (9). We previously suggested that methane (CH4) and N2O production were responsible for the reduction of the Coulombic efficiency of a MFC performing carbon and nitrogen removal in a loop configuration. The process involved the sequential feeding of synthetic wastewater containing acetate and ammonia first to the anode for carbon oxidation, then to an external aerobic stage performing ammonia oxidation, and finally to the cathode for NO3- reduction (4). Since both CH4 and N2O are potent greenhouse gases, their emission needs to be prevented. Several studies recognized the role of the electrode potential in regulating the activity and the synergistic interactions of the microorganisms growing in the anode compartment (10-13). Hence, the aim of this study was, first, to identify and quantify the electron losses occurring in a MFC continuously fed with acetate at the anode and with NO3- at the cathode; and second, to specifically study the generation and emission of N2O during the denitrification process in relation to the applied current and cathodic polarization.
Material and Methods Microbial Fuel Cell. A two-chamber MFC was built according to Freguia et al. (6) and filled with granular graphite as anode and cathode materials (2 to 6 mm diameter, El Carb 100, Graphite Sales, Inc., USA, available surface area 7.11 · 104 m2 · m-3). A cation exchange membrane separated the cell compartments (Ultrex CMI-7000, Membrane International, USA). Anode and cathode compartments (approximately 200 mL net liquid volume, 50 mL headspace) were inoculated with a microbial consortium previously operating in a MFC performing carbon and nitrogen removal (4). The anodic feed consisted of autoclaved modified M9 medium (6 g · L-1 Na2HPO4, 3 g · L-1 KH2PO4, 0.5 g · L-1 NaCl, 0.1 g · L-1 MgSO4 · 7H2O, 0.015 g · L-1 CaCl2 · 7H2O, 1 mL · L-1 trace nutrient solution as described elsewhere (14)) containing acetate as organic electron donor at a concentration of 0.393 g · L-1 as CH3COONa. The cathodic feed consisted of the same medium, to which 1 g · L-1 NaHCO3 was added as inorganic carbon source and NO3- was added as the sole electron acceptor at a concentration of 0.652 g · L-1 as NaNO3. Each feed (pH ∼ 7) was pumped into the reactor compartments at a flow rate of approximately 0.700 L · d-1, thus giving a loading rate of 1.398 mM as carbon (mM-C) per hour for acetate and 1.119 mM as nitrogen (mM-N) per hour for NO3-. Continuous recirculation was applied to both compartments 10.1021/es8036302 CCC: $40.75
2009 American Chemical Society
Published on Web 05/29/2009
at the rate of 200 mL · min-1. Ag/AgCl reference electrodes (MF-2052 Bioanalytical Systems, USA) were placed in both compartments. The electrodes were either connected over a 5 Ω resistor or connected to a potentiostat (VMP-3, Princeton Applied research, USA). The potentiostat was used to control and record the anodic potential, the cathodic potential, and the current. In case the external resistor was used, a data acquisition unit (34970A Data Acquisition Unit, Agilent Technologies, USA) was used to record voltage data. Current was then calculated from Ohm’s law. Off-Gas and Liquid Phase Analysis. CH4 production rate at the anode was measured by means of the Titration and Off-Gas Analysis (TOGA) sensor, developed by Pratt et al. (15). The sensor works by stripping the gases produced in the bioreactor with helium and sending the off-gas to a mass spectrometer (Omnistar Balzers AG, Liechtenstein), which gives an online measurement of its composition. For liquid phase composition analysis, samples obtained from the anode and cathode liquid phases were immediately filtered with a 0.22 µm sterile filter. The acetate content was determined with the use of high-performance liquid chromatography (HPLC), whereas NH4+, NO2-, and NO3- were measured using a Lachat QuikChem 8000 Flow Injection Analyzer (FIA). Unless stated otherwise, N2O was measured with a N2O microsensor (Unisense A/S, Denmark) placed in a separate glass chamber located in the cathode recirculation system. Tests under Continuous Feeding Conditions. Experiments under continuous feeding conditions were performed on both anode and cathode compartments. These experiments are detailed below, while a full description of the equations representing the balances as well as the method used to estimate the bacterial growth is provided in S1 and S2 in the Supporting Information (SI). Effect of Electrode Potential on the Anodic Processes. These experiments were performed under continuous anodic polarization of -100 and -200 mV vs Standard Hydrogen Electrode (SHE) and at open circuit. Each potential was maintained for 2 days before performing the test in order to allow the establishment of steady state conditions, evidenced by constant current production and stable anodic potential. The anode was then flushed with helium (100 mL · min-1) for 3 h before starting the experiment. This was done to strip out the CH4 that might have accumulated in the anodic chamber. Each test was then run over 3 h, during which electric current and CH4 production rate were continuously monitored with the methods previously described. Liquid phase samples were collected hourly and analyzed for acetate content. The data were averaged and statistical variability was determined. Effect of Electrode Potential on the Cathodic Processes. These tests were performed at applied cathodic polarizations of +100, 0, -100, and -200 mV SHE. Each condition was maintained for 2 days before performing the tests. Steady state conditions were evidenced by constant current production and stable cathodic potential. Each test was then run over 3 h, during which N2O concentration was monitored online and samples of the liquid phase were taken hourly for the analysis of NH4+, NO2-, and NO3-. The results were averaged to obtain statistical variability. Each test was done in triplicate. Tests under Batch Feeding Conditions. Batch tests were designed to specifically investigate the effects of the cathodic potential and of the applied current on the dynamics of the denitrification process. Current, electrodes potential, voltages, and N2O concentrations were continuously measured throughout the tests. Samples of the liquid phase were taken every 15 or 30 min during the experiments with nitrate for analysis of NO2-, NO3-, and NH4+. Production or consumption rates were determined from the respective concentration profiles.
Batch Tests under Controlled Cathodic Potential with NO3-. The tests were performed at applied cathodic polarizations +100, 0, -100, and -200 mV SHE. Prior to the experiment, the cathodic compartment was fed overnight with only medium solution containing no nitrate, and the system was run over a 5 Ω resistor in order to deplete NO3in the catholyte. During the experiments, the electrodes were connected through the potentiostat, which first operated the system at open circuit for one hour before applying the required conditions. 0.100 mmol-N as nitrate were injected to the cathode chamber at the beginning of each test. Each test was done in triplicate. Batch Tests under Controlled Current with NO3-. The electric current across the MFC compartments was controlled by the potentiostat. Three currents were used: 5, 10, and 15 mA. Previously to each test, the cathode chamber was flushed overnight with only medium solution, and the MFC was operated over a 5 Ω resistor. During the tests, 0.100 mmol-N as nitrate was injected to the cathode chamber, and the potentiostat was set to operate the system at open circuit for one hour and then under controlled current until the cell voltage dropped to 0 mV. Each test was done in triplicate.
Results Balances at the Anode during Continuous Feeding. The results of the acetate-fed anode are detailed in Table 1. Up to 10.8 ( 0.1 mA of current were produced at -100 mV SHE, which corresponded to an acetate removal rate of 1.100 ( 0.087 mM-C · h-1. A slightly lower current of 9.9 ( 0.1 mA (p < 0.001, see S4 in the SI) was instead obtained at -200 mV SHE, which resulted in an acetate consumption rate of 0.985 ( 0.002 mM-C · h-1. CH4 production occurred at any of the conditions applied, indicating the presence of a methanogenic community alongside the anodophiles. CH4 production rates ranged from 0.158 ( 0.0004 mM-C · h-1 at -100 mV SHE to 0.198 ( 0.0004 mM-C · h-1 at -200 mV. Much higher CH4 production (0.462 ( 0.0010 mM-C · h-1) was detected during the measurements at open circuit. Analysis of variance performed on the results shows that while acetate consumption was not significantly affected by the anodic potential, methane production rates and current production were instead significantly affected by this parameter (see S4 in the SI). Higher total biomass growth rates were estimated at higher anodic potentials, as up to 0.271 mM-C · h-1 of biomass was produced at -100 mV whereas 0.124 mM-C · h-1 and only 0.054 mM-C · h-1 were estimated at -200 mV and at open circuit, respectively. The total growth was split into the contribution of the two communities considered: electrochemically active bacteria (EAB), able to electrically interact with an electrode, and methanogenic Archaea. Based on the method detailed in S2 in the SI, while the growth of EAB was negligible at open circuit, methanogens were estimated to produce 0.054 mM-C · h-1 as biomass under the same operating conditions. At higher applied anodic potentials, lower growth rates for methanogens resulted from our calculations (0.023 mM-C · h-1 and 0.018 mM-C · h-1 at -200 mV and -100 mV SHE, respectively), whereas growth rates for EAB were estimated to range from 0.101 mM-C · h-1 to 0.253 mM-C · h-1 at -200 mV and -100 mV SHE, respectively. CH4 production and bacterial growth, considered as the only alternative anodic electron sinks in this study, had the effect of reducing the fraction of electrons provided as acetate that was used for current generation. As a consequence, Coulombic efficiencies of only 45.6% and 46.7% were obtained at -100 mV and -200 mV. The distribution of electrons across the different electron sinks is depicted in Figure 1a. When a potential of -100 mV SHE was applied, up to 25.6% of the total charge obtained from acetate oxidation was estimated to accumulate in the VOL. 43, NO. 13, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5145
TABLE 1. Summary of Results of Experiments under Continuous Feeding Conditions As Averages ± Standard Errorsf evaluated variablesa
measured variables applied Eanode, mV SHE
current, mA
∆Acetate, mM-C · h-1
∆X, ∆XCH4, ∆XEAB, YCH4, YEAB, ∆CH4, mM-C · h-1 mM-C · h-1 mM-C · h-1 mM-C · h-1 mol-C/mol-C mol-C/mol-C εanode, %
-100 10.8 ( 0.1 -1.100 ( 0.087 0.158 ( 0.0004 -200 9.9 ( 0.1 -0.985 ( 0.002 0.198 ( 0.0004 b -301.2 ( 0.4 (OC ) -0.980 ( 0.002 0.462 ( 0.0010
0.271 0.124 0.054
0.018 0.023 0.054
0.253 0.101 0
measured variablesc
0.058 0.058 0.058
0.322 0.172 0
45.6 46.7 -
evaluated variables
applied Ecathode, mV SHE
current, mA
∆NO3-, mM-N · h-1
∆N2O, mM-N · h-1
N2O/i, e-/e-
Error on ebalance,d %
εcathode,e %
-200 -100 0 +100
11.3 ( 0.7 12.4 ( 1.3 10.4 ( 0.6 7.8 ( 0.9
-0.424 ( 0.020 -0.484 ( 0.050 -0.432 ( 0.025 -0.327 ( 0.046
0.072 ( 0.029 0.141 ( 0.069 0.209 ( 0.040 0.173 ( 0.060
0.033 ( 0.013 0.024 ( 0.015 0.118 ( 0.020 0.114 ( 0.043
2.5 ( 2.6 1.0 ( 3.0 1.2 ( 3.6 1.1 ( 4.1
99.2 ( 1.3 95.6 ( 1.0 89.7 ( 4.6 89.4 ( 4.3
a The method used to estimate growth rates and yields is detailed in S1 and S2 in the Supporting Information. b Open Circuit. c NO2- and NH4+ were below detection limit at all conditions. d Balance equation detailed in S1 in the Supporting Information. e Coulombic efficiency for cathodic reaction as described in section S3 in the Supporting Information. f Each electrode potential was applied for 2 days prior to each experiment. Acetate, methane, current, nitrate, nitrite, and ammonium were recorded for 3 h. The experiment at the cathode was done in triplicate.
FIGURE 1. Breakdown of electron consumption among the different electron sinks as percentages of the total ∆Acetate at the anode (a) and of the total ∆NO3- at the cathode (b), based on data reported in Table 1. CH4 ) methane, I ) current, X ) estimated total bacterial growth, XEAB ) estimated growth of electrochemically active bacteria, XCH4 ) estimated growth of methanogens, N2 ) dinitrogen gas, N2O ) nitrous oxide. system as biomass, whereas 28.8% was released as CH4 and the remaining 45.6% was converted into current. When -200 mV SHE were applied, fewer electrons were estimated to shift to growth processes (13.2%), while 40.1% of the total charge was released as CH4. At open circuit, no current was produced, and up to 94.3% of the substrate removed was converted to CH4. Balances at the Cathode during Continuous Feeding. Table 1 also summarizes the results for the cathodic denitrification at different poised cathodic potentials during continuous feeding. The method established to describe the electron fluxes at the cathode could reasonably close the electron balance with errors ranging from 1.0 ( 3.0% to 2.5 ( 2.6%. In spite of the variability of results, higher applied potentials resulted in higher NO3- conversion rates (see analysis of variance in S4 in the SI). The maximum current measured was 12.4 ( 1.3 mA at -100 mV, which corresponded to a NO3- turnover of 0.484 ( 0.050 mM-N · h-1. The minimal current was instead obtained at +100 mV SHE, when 7.8 ( 5146
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 13, 2009
0.9 mA were produced, which corresponded also to the lowest NO3- removal rate of 0.327 ( 0.046 mM-N · h-1. NH4+ and NO2- were below detection limit (0.03 and 0.02 ppm, respectively) at all the applied conditions. In contrast, N2O measurements revealed a significant accumulation of this intermediate during the denitrification process. N2O production rates ranged from 0.072 ( 0.029 mM-N · h-1 at -200 mV SHE to 0.209 ( 0.040 mM-N · h-1 at 0 mV SHE. In terms of electron breakdown, the discharge of N2O corresponded to a loss of electrons ranging from 3.3 ( 1.3% at -200 mV SHE to 10.0 ( 2.1% and at +100 mV SHE (Figure 1b). Batch Tests: Effect of Applied Cathodic Potential and Current. Faster NO3- reduction was observed when a lower electrode potential was applied (Table 2), 0.256 ( 0.031 versus 0.310 ( 0.024 mM-N · h-1 at +100 and -200 mV SHE, respectively. Concurrently with NO3- reduction, N2O accumulated in the system at rates that averaged from 0.266 ( 0.083 mM-N · h-1 at +100 mV SHE to 0.352 ( 0.023 mMN · h-1 -200 mV SHE. After depletion of NO3-, N2O reduction
TABLE 2. Summary of Results of Batch Tests at Applied Cathodic Potential As Averages ± Standard Errorsa results in the presence of NO3- in the bulk -
results after NO3- depletion
applied Ecathode, mV SHE
maximum NO3 reduction rate, mM-N · h-1
maximum N2O accumulation rate, mM-N · h-1
current, mA
N2O/i, e-/e-
N2O reduction rate, mM-N · h-1
residual current on N2O reduction, mA
-200 -100 0 +100
-0.310 ( 0.024 -0.316 ( 0.028 -0.307 ( 0.006 -0.256 ( 0.031
0.352 ( 0.023 0.331 ( 0.037 0.314 ( 0.009 0.266 ( 0.083
20.4 ( 1.0 19.6 ( 0.7 14.6 ( 0. 8 9.4 ( 0.4
0.093 ( 0.002 0.090 ( 0.007 0.116 ( 0.007 0.151 ( 0.043
-0.136 ( 0.012 -0.115 ( 0.015 -0.060 ( 0.005 -0.049 ( 0.011
2.3 ( 0.1 1.9 ( 0.1 1.0 ( 0.1 0.7 ( 0.3
a
Each batch test was done in triplicate.
that increased from 0.082 ( 0.0001 mM-N · h-1 at 5 mA to 0.211 ( 0.0044 mM-N · h-1 at 15 mA. Very large differences in the cycle times were observed during the batches at fixed current. While at 15 mA the system was not able to keep the voltage within positive values for more than 37 ( 1 min, more than four times longer cycle time was instead measured at 5 mA. Figure 2b shows that during one of the replicates at 5 mA, as soon as the NO3- reached lower levels in the catholyte, the voltage dramatically dropped even if N2O was still present in the liquid bulk.
Discussion
FIGURE 2. (a) Example of batch test at fixed potential (0 mV SHE). (b) Example of batch at fixed current (5 mA applied). 0.100 mmol-N as NO3- were injected at time 0. NH4+ and NO2- were below detection limit at all conditions. An explanation for the N2O release during the open circuit phase is provided in S7 in the Supporting Information. accounted for the residual current during the last part of the batch experiment, as also shown by Figure 2a. Higher residual currents as well as higher N2O reduction rates were observed at lower cathodic potentials (Table 2). During the experiments at fixed current (results detailed in Table 3), higher nitrate removal rates were measured at higher applied currents. In fact, while 0.104 ( 0.005 mMN · h-1 were removed during the tests at 5 mA, 0.166 ( 0.014 mM-N · h-1 and 0.225 ( 0.004 mM-N · h-1 were removed at 10 mA and 15 mA, respectively. N2O accumulated at rates
Electron Sinks at the Anode. While effective in assessing the efficiency of electron utilization, the Coulombic efficiency provides only a partial explanation to the possible electron fluxes in MFC compartments. Processes like bacterial growth, methanogenesis, or fermentation, for instance, can occur simultaneously with current generation. Our results suggest that the Coulombic efficiency in this study was reduced by CH4 production and by the estimated bacterial growth during continuous operation of the acetate-fed anode. In addition, both measured and estimated processes were dependent on the electrode potential (Table 1 and Figure 1a). With respect to bacterial growth yields, we estimated that 0.058 mol-C of methanogenic biomass was produced per mol-C acetate consumed for CH4 production (Table 1). Previous studies reported yields ranging from 0.024 to 0.055 mol-C biomass per mol-C acetate for these organisms (16, 17). For electrochemically active bacteria (EAB), however, higher anodic potentials corresponded to higher growth yields. While 0.322 mol-C of EAB biomass was produced per mol-C of acetate at -100 mV SHE, the yield decreased to 0.172 mol-C biomass per mol-C acetate at -200 mV. These results are in agreement with the work of Freguia et al. (6), who obtained yields ranging from 0.238 to 0.292 mol-C biomass per mol-C in their acetate-fed anode operating at 20 Ω and 5 Ω, respectively. Acetate consumptions rates were comparable during the experiments at different applied conditions. Nevertheless, CH4 production rates significantly increased when the anodic potential was lowered from -100 mV to -200 mV SHE, and, furthermore, it more than doubled during the measurements at open circuit (see Table 1), thus demonstrating that acetoclastic methanogenesis was competing with electricity generation in our system. CH4 production from acetate yields a very low energy for the methanogens (∆G°’ ) -18 kJ per mol-C of acetate). This value is of the same order of magnitude as the theoretical energy that can be obtained at an anode fed with acetate (E°’CO2/CH3COO- ) -290 mV SHE), estimated through the Gibbs free energy relation (∆G°’ ) -n · F · (E°’donor-E°’acceptor)), and equal to -34.7 kJ per mol-C for an anode poised at -200 mV SHE and -73.3 kJ per mol-C for an anode poised at -100 mV SHE. It is important to note that the MFC was operated through a manual resistor (5 to 100 Ω) for several months prior to the experiments reported here. Rather low anodic potentials resulted from the relatively low currents that were VOL. 43, NO. 13, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5147
TABLE 3. Summary of Results of Batch Tests at Applied Current As Averages ± Standard Errorsa applied current, mA
maximum NO3reduction rate, mM-N · h-1
maximum N2O accumulation rate, mM-N · h-1
cathode %
N2O/i, e-/e-
cycle time, min
N2O still present at the end of the batch, mM-N
5 10 15
-0.104 ( 0.005 -0.166 ( 0.014 -0.225 ( 0.004
0.082 ( 0.0001 0.193 ( 0.0067 0.211 ( 0.0044
92.8 ( 2.4 75.3 ( 2.0 68.9 ( 1.2
0.087 ( 0.0002 0.207 ( 0.0071 0.226 ( 0.0047
149 ( 4 60 ( 2 37 ( 1
0.116 ( 0.021 0.174 ( 0.004 0.164 ( 0.004
a
Each batch test was done in triplicate.
generated. As a result, the low energy yields available for the EAB may have favored the establishment of a methanogenic community alongside the electrochemically active bacteria. Moreover, while EAB need an electrical contact with the electrode (1), methanogens may be advantaged by the fact that they can grow in any location in the reactor, regardless of the distance from the graphite granules. Electron Fluxes at the Cathode. Denitrification of NO3to N2 gas is obtained through four reduction steps involving NO2-, NO, and N2O as intermediates (7). When evaluated for the cathodic reaction, a low Coulombic efficiency (eq S5 in the SI) indicates possible intermediate accumulation and thus a loss of oxidation capacity of the electron acceptor used. The Coulombic efficiency in this study ranged from 89.4 ( 4.3% to 99.2 ( 1.3% (Table 1). NH4+ was always below detection limit during the tests, thus excluding possible alternative pathways for nitrate consumption such as nitrate ammonification. NO2- was also below detection limit throughout the experiments. NO was normally not measured during the tests at the cathode. However, very low concentrations can be expected for this intermediate (8), as it was also confirmed during a separate measurement (see S8 in the SI). Denitrification was confirmed by measurements of N2 and N2O in the gas phase during an additional batch test using the TOGA sensor. The nitrogen balance could be closed within 4.6% (see S6 in the SI). It can be concluded that apart from N2 production, N2O represented the major electron sink during nitrate removal in the MFC. Effect of Current and Cathodic Potential on N2O Accumulation. N2O was detected under all of the conditions applied at the cathode. Higher currents were achieved during the batches at poised potential. As a result, the conversion rates obtained for NO3- and N2O were higher during these tests. NO3- reduction occurred at rates that were comparable with that of N2O accumulation (Tables 2 and 3). This may suggest that very little N2O was used during the periods when NO3- was still present. Nevertheless, Table 3 shows that during the tests at a fixed current of 5 mA, NO3- was reduced at a rate of 0.104 ( 0.005 mM-N · h-1, whereas only 0.082 ( 0.0001 mM-N · h-1 accumulated as N2O. This implies that in this instance, some N2O was simultaneously reduced together with NO3-. The MFC was able to reduce N2O, although at a slower rate compared with NO3- reduction. Further experiments at controlled current with only N2O as electron acceptor (see S5 in the SI) indicated that while the biocathode was capable of sustaining low currents (about 3-4 mA) without substantial decrease of the MFC voltage over time, any higher applied current resulted instead in a rather fast drop of the voltage, in spite of the availability of N2O in the bulk. This may suggest a possible limitation through which the cathodic biofilm was not able to reduce N2O at the same rate as it was produced by the previous step of denitrification. The hypothesis of an inhibition of the enzyme responsible for N2O reduction, N2O reductase, is often mentioned in literature. Elevated NO2concentrations and/or low pH (through free nitrous acid formation) can result in N2O accumulation during denitrification in activated sludge (18). However, none of these conditions apply to the present study as the highly buffered 5148
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 13, 2009
media, the tendency of the cathode pH to increase (19), and the very low NO2- levels measured during the study exclude the presence of significant levels of free nitrous acid that might inhibit the N2O reductase (20). Schalk-Otte and co-workers (21) postulated that the competition between the different reductases for the electrons from the cytochrome C pool was responsible for N2O production by Alcaligenes faecalis. They suggested that under conditions of low electron availability, a lower affinity of the N2O reductase toward the electron donor drives the accumulation of this intermediate. Our study shows that very little N2O was reduced as long as NO3- was present in the bulk liquid during batch measurements, thus suggesting a preference toward nitrate as electron acceptor. The cathodic potential has emerged as an important factor in this study. During the experiments at controlled cathodic potential under continuous feeding, lower amounts of N2O accumulated in the system at lower applied potentials, resulting in higher Coulombic efficiencies (Table 1). During batch measurements at controlled potentials, the residual N2O present in the bulk after the depletion of NO3- was still responsible for current generation in the last part of the experiment (Figure 2a and Table 2). Interestingly, also in this case higher N2O reduction rates were observed under lower cathodic poised potentials. The N2O to current ratio represents the fraction of electrons that accumulates as N2O in electron equivalents with respect to the electrons (charge) that flow through the system. Higher currents resulted in higher substrate turnover and N2O production (Tables 2 and 3). Nevertheless, the fraction of charge that was lost as N2O was considerably decreased at lower applied potentials (Tables 1 and 2). Similarly to the anodic processes, bacteria growing at the cathode also gain energy proportionally to the potential difference existing between the electron donor (the cathode) and the electron acceptor (nitrogen oxides). When the cathode is poised at low potentials, more energy is theoretically available for N2O reduction as a result of the higher electromotive force that can be generated. From a purely thermodynamic point of view, N2O should be a more favorable electron acceptor than NO3- (E°’N2O/N2 ) +1.355 V SHE, E°’NO3/NO2 ) +0.433 V SHE). However, the energy generated from N2O reduction is in general dissipated as heat, whereas NO3- reduction by membrane bound NO3- reductase (NAR) results in net translocation of protons and therefore proton motive force is produced (22). It might be possible that bacteria preferentially utilize NO3- instead of N2O because of the higher “relevant” energy gain resulting from the reduction of the former. Further investigation in this regard is certainly encouraged.
Acknowledgments This research was supported by the Australian Research Council (Grant DP0666927). Bernardino Virdis was supported by the Ph.D. program in Engineering and Environmental Sciences granted by the University of Cagliari, Italy.
Supporting Information Available Equations representing the electron balances, the method used to estimate the bacterial growth, the cathodic Coulombic efficiency, the analysis of variance of results obtained during continuous feeding, additional N2, N2O, and NO measurements, and measurements of the observed capacitance. This material is available free of charge via the Internet at http:// pubs.acs.org.
Literature Cited (1) Rabaey, K.; Rodriguez, J.; Blackall, L. L.; Keller, J.; Gross, P.; Batstone, D.; Verstraete, W.; Nealson, K. H. Microbial ecology meets electrochemistry: electricity-driven and driving communities. ISME J. 2007, 1 (1), 9–18. (2) Aelterman, P.; Rabaey, K.; Clauwaert, P.; Verstraete, W. Microbial fuel cells for wastewater treatment. Water Sci. Technol. 2006, 54 (8), 9–15. (3) Clauwaert, P.; Rabaey, K.; Aelterman, P.; DeSchamphelaire, L.; Pham, T. H.; Boeckx, P.; Boon, N.; Verstraete, W. Biological Denitrification in Microbial Fuel Cells. Environ. Sci. Technol. 2007, 41 (9), 3354–3360. (4) Virdis, B.; Rabaey, K.; Yuan, Z.; Keller, J. Microbial fuel cells for simultaneous carbon and nitrogen removal. Water Res. 2008, 42 (12), 3013–3024. (5) Logan, B. E.; Hamelers, B.; Rozendal, R.; Schroder, U.; Keller, J.; Freguia, S.; Aelterman, P.; Verstraete, W.; Rabaey, K. Microbial fuel cells: methodology and technology. Environ. Sci. Technol. 2006, 40 (17), 5181–5192. (6) Freguia, S.; Rabaey, K.; Yuan, Z.; Keller, J. Electron and Carbon Balances in Microbial Fuel Cells Reveal Temporary Bacterial Storage Behavior During Electricity Generation. Environ. Sci. Technol. 2007, 41 (8), 2915–2921. (7) Zumft, W. G. Cell biology and molecular basis of denitrification. Microbiol. Mol. Biol. Rev. 1997, 61 (4), 533–616. (8) Cervantes, F.; Monroy, O.; Gomez, J. Accumulation of intermediates in a denitrifying process at different copper and high nitrate concentrations. Biotechnol. Lett. 1998, 20 (10), 959–961. (9) Wrage, N.; Velthof, G. L.; van Beusichem, M. L.; Oenema, O. Role of nitrifier denitrification in the production of nitrous oxide. Soil Biol. Biochem. 2001, 33 (12-13), 1723–1732. (10) Aelterman, P.; Freguia, S.; Keller, J.; Verstraete, W.; Rabaey, K. The anode potential regulates bacterial activity in microbial fuel cells. Appl. Microbiol. Biotechnol. 2008, 78 (3), 409–418.
(11) Finkelstein, D. A.; Tender, L. M.; Zeikus, J. G. Effect of electrode potential on electrode-reducing microbiota. Environ. Sci. Technol. 2006, 40 (22), 6990–6995. (12) Freguia, S.; Rabaey, K.; Yuan, Z.; Keller, J. Syntrophic Processes Drive the Conversion of Glucose in Microbial Fuel Cell Anodes. Environ. Sci. Technol. 2008, 42 (21), 7937–7943. (13) Schroder, U. Anodic electron transfer mechanisms in microbial fuel cells and their energy efficiency. Phys. Chem. Chem. Phys. 2007, 9 (21), 2619–2629. (14) Lu, H.; Oehmen, A.; Virdis, B.; Keller, J.; Yuan, Z. Obtaining highly enriched cultures of Candidatus Accumulibacter phosphates through alternating carbon sources. Water Res. 2006, 40 (20), 3838–3848. (15) Pratt, S.; Yuan, Z. G.; Gapes, D.; Dorigo, M.; Zeng, R. J.; Keller, J. Development of a novel titration and off-gas analysis (TOGA) sensor for study of biological processes in wastewater treatment systems. Biotechnol. Bioeng. 2003, 81 (4), 482–495. (16) Liu, J. S.; Marison, I. W.; von Stockar, U. Microbial growth by a net heat up-take: A calorimetric and thermodynamic study on acetotrophic methanogenesis by Methanosarcina barkeri. Biotechnol. Bioeng. 2001, 75 (2), 170–180. (17) Sowers, K. R.; Nelson, M. J.; Ferry, J. G. Growth of acetotrophic, methane-producing bacterial in a pH auxostat. Curr. Microbiol. 1984, 11 (4), 227–229. (18) Zhou, Y.; Pijuan, M.; Yuan, Z. G. Free nitrous acid inhibition on anoxic phosphorus uptake and denitrification by poly-phosphate accumulating organisms. Biotechnol. Bioeng. 2007, 98 (4), 903–912. (19) Rozendal, R. A.; Hamelers, H. V. M.; Buisman, C. J. N. Effects of membrane cation transport on pH and microbial fuel cell performance. Environ. Sci. Technol. 2006, 40 (17), 5206–5211. (20) Zhou, Y.; Pijuan, M.; Zeng, R. J.; Yuan, Z. Free Nitrous Acid Inhibition on Nitrous Oxide Reduction by a DenitrifyingEnhanced Biological Phosphorus Removal Sludge. Environ. Sci. Technol. 2008, 42 (22), 8260–8265. (21) Schalk-Otte, S.; Seviour, R. J.; Kuenen, J. G.; Jetten, M. S. M. Nitrous oxide (N2O) production by Alcaligenes faecalis during feast and famine regimes. Water Res. 2000, 34 (7), 2080–2088. (22) Wasser, I. M.; de Vries, S.; Moenne-Loccoz, P.; Schroder, I.; Karlin, K. D. Nitric oxide in biological denitrification: Fe/Cu metalloenzyme and metal complex NOx redox chemistry. Chem. Rev. 2002, 102 (4), 1201–1234.
ES8036302
VOL. 43, NO. 13, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5149
