This article was downloaded by: [University of New Hampshire]On: 13 March 2013, At: 06:18Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Environmental TechnologyPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/tent20

A cooperative microbial fuel cell system for wastetreatment and energy recoveryFei Zhang a & Zhen He aa Department of Civil Engineering and Mechanics, University of Wisconsin-Milwaukee,Milwaukee, USAAccepted author version posted online: 13 Feb 2013.Version of record first published: 18 Feb2013.

To cite this article: Fei Zhang & Zhen He (2013): A cooperative microbial fuel cell system for waste treatment and energyrecovery, Environmental Technology, DOI:10.1080/09593330.2013.770540

To link to this article: http://dx.doi.org/10.1080/09593330.2013.770540

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form toanyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses shouldbe independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims,proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly inconnection with or arising out of the use of this material.

Environmental Technology, 2013http://dx.doi.org/10.1080/09593330.2013.770540

A cooperative microbial fuel cell system for waste treatment and energy recovery

Fei Zhang and Zhen He∗

Department of Civil Engineering and Mechanics, University of Wisconsin-Milwaukee, Milwaukee, USA

(Received 15 November 2012; final version received 22 January 2013 )

A cooperative microbial fuel cell (MFC) system was developed and investigated for its performance in waste treatment andbioenergy production. The system consisted of a conventional MFC and a dual-cathode MFC that was specially designedfor nitrogen removal. Three feeding solutions, including a synthetic solution (acetate as an organic source), digested sludgeand landfill leachate, were used as substrates. The MFC system removed more than 99% of chemical oxygen demand (COD)from the synthetic solution, and the organic removal mainly occurred in the anodes of the MFCs. Nitrogen, on the other hand,was removed in the cathodes of the MFC system, which achieved more than 98% removal of ammonium and 96% removalof total nitrogen. The MFC system also effectively treated the actual wastes and removed more than 85% of the total CODand 50–70% of the total nitrogen from the digested sludge and landfill leachate. The organic removal rates with the actualwastes were significantly higher than in other MFC systems. In general, the MFC system consumed less than 0.06 kWh/m3 or0.1 kWh/kg COD, demonstrating that low energy consumption is a major advantage of MFC technology. The MFC systemproduced 0.1023 kWh/kg COD from the synthetic solution, much higher than the 0.0097 kWh/kg COD and 0.0019 kWh/kgCOD from the actual wastes, resulting in a theoretically positive energy balance with the synthetic solution; however, the netenergy with the actual wastes was negative.

Keywords: microbial fuel cells; bioenergy production; nitrogen removal; digested sludge; landfill leachate

1. IntroductionWastewater is generated from various sources and poses agreat threat to the ecosystem and to human health. The con-taminants in wastewater, including organic and inorganiccompounds, can seriously deteriorate water quality. Thus,wastewater must be effectively treated to reduce the con-centrations of these contaminants before being discharged.Wastewater treatment usually focuses on the removal oforganic carbon and nutrients (e.g. nitrogen and phospho-rus); waste streams from some industries also may requirethe treatment of heavy metals, colour and other special com-pounds. The treatment is an energy/resource input processassociated with high operating expense. To achieve sus-tainable wastewater treatment, it is necessary to reduce theconsumption of energy and resources (e.g. chemicals) andto recover value-added products (e.g. energy and usefulchemicals) from wastewater [1].

One of the promising technologies for sustainablewastewater treatment is microbial fuel cells (MFCs).MFCs are bioelectrochemical reactors that take advantageof the anaerobic microbial metabolism of organic com-pounds to directly produce electricity from wastewater[2]. The MFC concept was proposed about one hundredyears ago [3] but intensive research on various aspectsof MFCs [4] has increased greatly in the past decade.

∗Corresponding author. Email: zhenhe@uwm.edu

Consequently, researchers have gained much understandingof fundamental issues in microbiology, electrochemistryand materials in MFCs. The main challenges for futureapplications of MFC technology include:

(1) System scaling up – most MFC studies were con-ducted on very small scales; therefore, scaling upMFC reactors to larger sizes (eventually to full scalefor practical application) is a great challenge.

(2) Nutrient removal – the anode of an MFC is underan anaerobic condition and thus cannot effectivelyremove nutrients such as nitrogen.

(3) Material cost – the electrode and membrane materi-als used in MFCs result in a high capital investmentthat makes MFCs less competitive than the existingtechnologies.

(4) Energy efficiency – it is not clear how much energyMFC operation consumes and the energy produc-tion in MFCs, especially at a large scale, is still toolow to be practically used.

Those challenges need to be addressed through both funda-mental and applied research. The present study has focusedon two of the challenges – nitrogen removal and energy effi-ciency, and has expanded on our previous studies through

© 2013 Taylor & Francis

Dow

nloa

ded

by [

Uni

vers

ity o

f N

ew H

amps

hire

] at

06:

18 1

3 M

arch

201

3

2 F. Zhang and Z. He

coupling a conventional MFC (mainly for organic removal)with a dual-cathode MFC (for nitrogen removal). An impor-tant feature of this work is the use of two types of actualwastewaters (landfill leachate and digested sludge) thathave not been well studied before.

Nitrogen removal has been investigated in MFCs witha focus on bioelectrochemical denitrification in the cath-ode of an MFC, although ammonia recovery has become arecent interest [5,6]. Early studies demonstrated that nitratecould be reduced by autotrophic bacteria that acceptedelectrons from an electrode either in a three-electrode bio-electrochemical cell [7] or a cathode electrode of an MFC[8,9]. The concept of bioelectrochemical denitrificationwas further developed in an MFC system involving bothnitrification and denitrification [10–13]. We have previ-ously demonstrated the feasibility of nitrogen removal ina dual-cathode MFC system [14,15]. However, it is notclear how well the dual-cathode MFC can perform withactual wastewater, especially at a high organic loading rate.The organic loading rate could significantly affect nitrogenremoval: an insufficient supply of organics may limit theelectron supply to bioelectrochemical denitrification; oversupply of organics may stimulate growth of heterotrophicbacteria in the cathode and thus limit nitrification, which isa necessary step before denitrification.

In this study we have designed a cooperative MFCsystem consisting of an organic-based MFC (MFC-1) and anitrogen-based dual-cathode MFC (MFC-2) (Figure 1). Theorganic-based MFC is a conventional MFC that pre-treatsorganic compounds and acts as a major energy producerin the system; the nitrogen-based MFC is a tubular dual-cathode reactor, similar to that developed in our previousstudy, which mainly functions as a nitrogen remover [15].To operate this system, wastewater is fed into the anode ofMFC-1 (Figure 1(1)) and its effluent flows into the anodeof MFC-2 (Figure 1(2)); it is expected that the majorityof organic compounds are removed in those two anodes.Then, the anode effluent is supplied to the outer cathode(aerobic) of MFC-2 (Figure 1(3)) for nitrification by aero-bic bacteria. The effluent of the outer cathode moves intothe inner cathode (anoxic) for denitrification (Figure 1(4)).The treated effluent then flows into the cathode of MFC-1(Figure 1(5)) and is then discharged (Figure 1(6)). In thisway, a single stream flows through all of the compartmentsfor different purposes. We examined the system perfor-mance with a synthetic solution and actual wastes (digestedsludge and landfill leachate) and conducted a preliminaryenergy balance analysis.

2. Materials and methods2.1. MFC system setupThe MFC system consisted of two different tubular MFCs.MFC-1 was designed to treat organic carbons and MFC-2was designed to remove nitrogen (with additional organicremoval). MFC-1 contained two compartments, anode and

Figure 1. Diagram of the cooperative MFC system. (a) Flowpath of the feeding solution; (b) schematic of the MFC reactors.MFC-1: regular two-chamber MFC; MFC-2: dual-cathode MFC.The reactions in each compartment of the flow path (using sodiumacetate (NaOAc) as an example of organic source): MFC-1 anode,NaOAc to CO2; MFC-2 anode, NaOAc to CO2; MFC-2 outercathode (aerobic), O2 to H2O and NH+

4 to NO−3 ; MFC-2 inner

cathode (anoxic), NO−3 to N2; and MFC-1 cathode, O2 to H2O.

cathode, separated by a cation exchange membrane (CEM),(Membrane International Inc. Ringwood, NJ). The CEMtube was 29 cm long and 6 cm in diameter (Figure 1(b)),resulting in an anode liquid volume of 750 mL. The anodeelectrode was a carbon brush (Gordon Brush Mfg. Co.Inc. Commerce, CA) that was 20 cm in length and 5 cmin diameter; it was pre-treated before being used, preparedas previously [16]. The cathode electrode was a piece ofcarbon cloth (500 cm2, Zoltek Corporation, St. Louis, MO)coated with platinum (10% on carbon black) as an oxygenreduction catalyst (0.3 mg Pt/cm2) prepared as previously[17]. It should be noted that the use of Pt was to ensure a suf-ficient cathode reaction; this expensive catalyst will not bepractical for future large-scale MFC systems. The two elec-trodes were connected with copper wires to a resistance box.MFC-1 was placed in a tank that contained the catholyte(liquid volume of ∼2 L) aerated with air at 100 mL/min.

MFC-2 was built in a similar manner to that in ourprevious study [15]. As shown in Figure 1(b) it containedtwo membrane tubes. The CEM tube separated the anodeand the outer cathode, and the anion exchange membrane(AEM) tube formed an inner cathode. The anode liquid vol-ume was 250 mL and the inner cathode liquid volume was550 mL. The anode electrode was a piece of carbon cloth(200 cm2, Zoltek Corporation, St. Louis, MO). The outer

Dow

nloa

ded

by [

Uni

vers

ity o

f N

ew H

amps

hire

] at

06:

18 1

3 M

arch

201

3

Environmental Technology 3

cathode electrode was also a piece of carbon cloth (300 cm2,Zoltek Corporation, St. Louis, MO) but coated with a Ptcatalyst (0.3 mg Pt/cm2). The inner cathode electrode wasa carbon brush similar to the anode electrode of MFC-1.MFC-2 was placed in a different tank from that of MFC-1,but with the same liquid volume of ∼2 L (outer cathode).The tank was aerated with air at 100 mL/min. Two electriccircuits, one consisting of the anode electrode and the outercathode electrode (anode/outer cathode), and the other con-taining the anode electrode and the inner cathode electrode(anode/inner cathode), were formed by connecting eachelectrode couple to a resistance box using copper wires.

2.2. Operating conditionsThe MFC system was operated continuously at a room tem-perature of around 20◦C. The anodes of both MFC-1 andMFC-2 and the inner cathode of MFC-2 were inoculatedwith mixed activated sludge and digested sludge (1:1) col-lected from the Southshore Wastewater Treatment Plant,Milwaukee, WI. Either a synthetic solution or actual wastewas fed into the anode of MFC-1 at 2 mL/min, resultingin an anode hydraulic retention time (HRT) of 6.3 h inMFC-1 and 2.1 h in MFC-2. The synthetic solution withan initial pH of 6.84 was prepared as: 1.5 g CH3COONa,0.5 g NaCl, 0.26 g NH4Cl, 0.015 g MgSO4, 0.02 g CaCl2,0.53 g KH2PO4, 1.07 g K2HPO4, 1 g NaHCO3 and 1 mLtrace elements per L of tap water [18]. The digestion sludgewas obtained from an anaerobic digester at the SouthshoreWastewater Treatment Plant (Milwaukee, WI) and was usedeither in original form (X0, without any pre-treatment) ordiluted with tap water for 5 (X5), 10 (X10) and 20 (X20)times. Landfill leachate was collected from a landfill (Emer-ald Park Landfill, Muskego, WI) and used without anypre-treatment. The anolytes of both MFCs and the innercatholyte of MFC-2 were recirculated at 60 mL/min. Theinitial external resistance for both MFCs was 1000 �, andwas gradually reduced to the values introduced in Section 3.

2.3. Measurement and analysisA digital multimeter (2700, Keithley Instruments, Inc.Cleveland, OH) was used to monitor the voltage and recorda point every 5 min. The pH was measured using a bench-top pH meter (Oakton Instruments, Vernon Hills, IL). Theconcentrations of NH+

4 -N, NO−2 -N, NO−

3 -N and COD weremeasured using a colorimeter (DR/890, Hach Company,Loveland, CO) according to the manufacturer’s procedures.The densities of power and current were calculated basedon the liquid volume of the anode compartment. A polar-ization test was performed in a two-electrode mode using apotentiostat (Reference 600, Gamry Instruments, Warmin-ster, PA) to scan from the open-circuit voltage to zero ata scan rate of 0.2 mV/s. The internal resistance was esti-mated by the voltage and the current at the maximum poweroutput. The COD loading rates and removal rates were cal-culated based on either the total liquid volume of the MFC

system (5.55 L), or the total anode liquid volume (1.00 L).Nitrogen loading rates and removal rates were calculatedbased on the liquid volume of the two cathodes of MFC-2 (2.55 L). Columbic efficiency (CE) shows the efficiencyof substrate-to-electricity by the ratio between the coulomboutput (integrating current and time) and the total coulombinput [15]:

CEc = Qoutput

Qinput=

∑I × t

96485 × CODtotal × 4

CEN = Qoutput

Qinput=

∑Iinnert

96485 × N × 5

where CEC is the coulombic efficiency based on the organicsubstrate, CEN is the coulombic efficiency based on nitrate,Qoutput is the produced charge, Qinput is the total chargeavailable in the substrate that has been removed, Iinner isthe electric current (A) generated from the inner cathode, tis time (s), CODtotal is the total COD removed (mol) by theMFC system in the period of time t and N is the amount ofnitrate removed (mol) in the inner cathode within time t.

The theoretical power requirement for the pumpingsystem was estimated as [19]:

Ppumping = Qγ E1000

where P is power requirement (kW), Q is flow rate (m3/s), γis 9800 N/m3 and E is the hydraulic pressure head (m). Theenergy consumption by aeration was estimated as describedin a previous publication [20].

3. Results and discussion3.1. Synthetic solutionTo examine the feasibility of this cooperative MFC system,the operation began with a synthetic solution containingacetate as an electron donor. Polarization tests were con-ducted to examine the overall electricity generation afterboth MFCs exhibited stable performance (after about 60 d ofoperation). The open-circuit potentials (OCP) were 0.79 Vin MFC-1 and 0.64 V in MFC-2. The maximum poweroutput was 9.5 and 4.3 mW, and the maximum current gen-eration was 59.4 and 20.0 mA for MFC-1 and MFC-2,respectively (Figure 2(a)). As expected, MFC-1 producedmore electricity than MFC-2 (the electric couple of theanode/outer cathodes), most likely due to a greater organicsupply and a larger cathode electrode in MFC-1. The com-petition for electrons between the two cathodes in MFC-2could also lower its power production from the anode/outercathode circuit. According to the polarization curves, theinternal resistances of MFC-1 and MFC-2 (outer cathode)were estimated to be 19 and 27 �. Therefore, to maximizethe power output, the external resistance of MFC-1 was setto 20 �, and the resistance between the anode and the outercathode of MFC-2 was set to 30 �. Because the inner cath-ode of MFC-2 was designed for nitrate removal through

Dow

nloa

ded

by [

Uni

vers

ity o

f N

ew H

amps

hire

] at

06:

18 1

3 M

arch

201

3

4 F. Zhang and Z. He

Figure 2. Electricity generation in the cooperative MFC systemwhen treating the synthetic solution. (a) Polarization curves ofMFC-1 and MFC-2 (outer cathode); (b) current generation in theMFC-1 and two electric circuits of the MFC-2. MFC-2-OC repre-sents the circuit of the anode/outer cathode, and MFC-2-IC is theanode/inner cathode circuit.

bioelectrochemical denitrification and more electrons (highcurrent) would benefit nitrate reduction, the anode/innercathode circuit was connected through a low resistance of1 � to achieve a high current flow. Figure 2(b) shows currentgeneration from three electric circuits: the average currentof MFC-1 was 25 mA, and the average currents of MFC-2anode/outer cathode and anode/inner cathode were 7 and9 mA, respectively.

The MFC system effectively removed both COD andnitrogen compounds from the synthetic solution. The CODremoval mainly occurred in the two anodes: the COD con-centration decreased from 1170 ± 121 to 568 ± 32 mg/Lafter the solution flowed through the anode of MFC-1;the MFC-2 anode further decreased its concentration to17 ± 29 mg/L (Figure 3(a)). The following cathode com-partments maintained a lower COD concentration. Thetotal COD removal rate for the whole system was 3.37 kgCOD/m3/d based on the anode liquid volume, or 0.61 kgCOD/m3/d based on the total liquid volume of the wholeMFC system.

Figure 3. The concentrations of organics and nitrogen in differ-ent compartments of the cooperative MFC system when treatingthe synthetic solution. (a) COD concentration; (b) nitrogen con-centrations. MFC1-A: MFC-1 anode; MFC2-A: MFC-2 anode;MFC2-OC: MFC-2 outer cathode; MFC2-I: MFC-2 inner cathode;MFC1-C: MFC-1 cathode.

In contrast, nitrogen removal occurred mainly in thecathodes of the MFCs, especially MFC-2. Figure 3(b) showsthat the concentration of ammonium nitrogen did not obvi-ously change in the two anodes, confirming that anaerobicammonium oxidation cannot easily be accomplished in anMFC anode. Almost all ammonium nitrogen was removedin the outer cathode of MFC-2 through nitrification, witha significant production of nitrate nitrogen suggesting thepresence of nitrification. The total nitrogen in the outer cath-ode was lower than that in the feed solution, most likelybecause some ammonium nitrogen was removed throughair stripping because of aeration and the elevated pH ofthe catholyte due to oxygen reduction [21]. Nitrate wasgreatly reduced in the inner cathode via bioelectrochemicaldenitrification, resulting in an effluent nitrate concentrationof 1.4 ± 0.2 mg NO−

3 -N/L. Incomplete denitrification alsooccurred, as indicated by a nitrite concentration of 4.5 ±0.1 mg NO−

2 -N/L in the effluent of the MFC-2 inner cath-ode. Both nitrate and nitrite were slightly further reduced

Dow

nloa

ded

by [

Uni

vers

ity o

f N

ew H

amps

hire

] at

06:

18 1

3 M

arch

201

3

Environmental Technology 5

in the cathode of MFC-1 where a micro-anaerobic condi-tion might exist for denitrification. The concentrations ofnitrate nitrogen and nitrite nitrogen in the final effluent (fromthe cathode of MFC-1) were 0.8 ± 0.4 and 0.4 ± 0.1 mg/L,respectively. The final concentration of ammonium nitrogenwas 1.0 ± 1.0 mg/L. Therefore, the removal efficiencies ofammonia and nitrate were higher than 98%. The total nitro-gen removal efficiency was more than 96% with a removalrate 0.08 ± 0.00 kg N/m3/d.

Nitrate was mostly removed via bioelectrochemical den-itrification in the inner cathode of MFC-2, and the electronsupply to the inner cathode was expected to play a criti-cal role in denitrification. Because two cathodes shared thesame anode (to compete for electrons) in MFC-2, currentgeneration from the outer cathode would affect the elec-tron flow to the inner cathode. To investigate this effect, theexternal resistance of the anode/outer cathode was adjustedto 1, 50 and 1000 � for different levels of current gen-eration, and nitrogen removal was monitored under thoseconditions. Clearly, both the current generation from theinner cathode and the nitrogen removal were affected by theexternal resistance between the anode and the outer cath-ode (Figure 4). At 1 �, the average current of the innercathode was nearly zero, because the current generationof the outer cathode was greatly improved at a low exter-nal resistance, and oxygen (in the outer cathode) was amore preferable electron accepter than nitrate (in the innercathode) due to its higher reducing potential and strongerreduction reaction assisted by the Pt catalyst, thereby morecompetitive for electrons from the anode. The total nitro-gen removal efficiency was 42% and the nitrate removalefficiency was 24% at 1 �. At a high external resistanceof 1000 �, the current of the inner cathode increased to9.0 ± 0.5 mA. This increased current improved the removalof both total nitrogen and nitrate to 65% and 59%, respec-tively, higher than those at 1 �. Those results demonstrate

Figure 4. Nitrogen removal in the cooperative MFC system,and current generation of the inner cathode of MFC-2 at differentexternal resistances of the outer cathode of the MFC-2.

that current generation in the outer cathode significantlyaffected the denitrifying performance of the inner cathode,possibly through distribution of organic compounds. Whenthe outer cathode produced a low current, it also had a lowdemand of organic compounds; as a result, more electronscould flow to the inner cathode to improve bioelectrochem-ical denitrification. Therefore, the outer cathode may beused to control the denitrification in the inner cathode viaadjusting the current generation (from the outer cathode).

3.2. Digested sludge and landfill leachateTo further demonstrate the cooperative MFC system forwaste treatment, two actual wastes, digested sludge andlandfill leachate, were examined for electricity genera-tion and contaminant removal. Those wastes were cho-sen because of their high organic and nitrogen con-centrations and potentially serious environmental effects.The raw digested sludge contained a total COD con-centration of ∼10,500 mg/L, a soluble COD concentra-tion of ∼660 mg/L and a total nitrogen concentration of∼550 mg/L. The digested sludge also contained a highlevel of solids: the concentrations of total and volatilesuspended solids were about 6000 mg/L and 3000 mg/L,respectively. The landfill leachate had a total COD of∼11400 mg/L, a soluble COD of ∼1000 mg/L and totalnitrogen of ∼550 mg/L. The significant difference betweenthe total COD and the soluble COD indicated that a largefraction of the organic content in those wastes is partic-ulate and/or recalcitrant compounds, which makes theirbiodegradation more difficult.

The experiment with the digested sludge was conductedat four dilutions: 0 (X0), 5 (X5), 10 (X10) and 20 (X20)times. At the highest dilution X20 (the lowest influent con-centrations), it was observed that the MFC system removedmore than 99.0% of the total COD, 97.3% of the solubleCOD and 93.2% of the total nitrogen (Figure 5). Whenthe original digested sludge (X0, no dilution) was fed, the

Figure 5. Removal of COD and nitrogen in the cooperative MFCsystem at different dilution ratios of the digested sludge.

Dow

nloa

ded

by [

Uni

vers

ity o

f N

ew H

amps

hire

] at

06:

18 1

3 M

arch

201

3

6 F. Zhang and Z. He

MFC system still removed 91.0% of the total COD, but theremoval of the soluble COD and the total nitrogen decreasedto 78.5% and 53.7%, respectively. The larger loading ratesat the original concentrations contributed to a lower removalefficiency; in addition, the total COD could be decomposedinto the soluble COD, thereby increasing the concentrationof the soluble COD in the final effluent. The total CODremoval rate significantly increased from 0.26 ± 0.00 kgCOD/m3/d at X20 to 4.94 ± 0.15 kg COD/m3/d at X0based on the total liquid volume of the MFC system; like-wise, the total nitrogen removal rates also increased from0.04 ± 0.00 kg N/m3/d at X20 to 0.34 ± 0.19 kg N/m3/dat X0 based on the cathode liquid volume. When the feedingsolution was switched to landfill leachate, the MFC systemremoved 87.0% of the total COD, 77.6% of the solubleCOD and 69.1% of the total nitrogen. The treated effluentcontained a total COD concentration of 1481 ± 387 mg/Land a soluble COD concentration of 224 ± 56 mg/L. Thetotal nitrogen in the final effluent included 56.7 ± 5.8 mg/Lof ammonium nitrogen, 83.3 ± 24.0 mg/L of nitrate nitro-gen and 29.2 ± 16.0 mg/L of nitrite nitrogen, suggest-ing that denitrification was a limiting factor to nitrogenremoval.

Those results demonstrate that the cooperative MFCsystem can effectively treat wastes at high loading ratesof both organics and nitrogen. The total COD loadingrates with raw digested sludge (X0) and landfill leachatewere 5.4 ± 0.3 and 5.9 ± 0.3 kg COD/m3/d, and the totalnitrogen loading rates of the two wastes were similar at0.61 kg N/m3/d (Figure S1, see supplementary informa-tion). The cooperative MFC system achieved much higherremoval rates of both organics and nitrogen than thoseobtained in other MFCs designed for nitrogen removal, andalso higher than when the system was treating the syntheticsolution (Figure S1 and Table 1). The high removal rates inthe present study were due to the high input concentrationsand additional treatment in the cathode compartments viaaerobic reactions. For example, at an input COD loading

rate of 5.4 ± 0.3 kg COD/m3/d with the digested sludge,about 40% of the total COD was removed by the twoanodes and 39% of the total COD was removed by the aer-ated cathode tanks. The remaining COD flowed into theinner cathode where it provided electrons for denitrifica-tion, although that would decrease the nitrogen coulombicefficiency because of the competition of electron donationbetween the COD and the inner cathode electrode.

It was observed that with raw digest effluent or land-fill leachate, there was a high concentration of ammoniumnitrogen in the final effluent (the cathode effluent of MFC-1); however, this was not found with the synthetic solutionor the diluted digest effluent. The high ammonium concen-tration was most likely caused by ammonium migrationthrough the CEM in MFC-1, which has been reported inprevious studies [5,6]. To prevent ammonium loss to thefinal effluent in MFC-1, we replaced the CEM with anAEM and found a significant reduction of ammonium nitro-gen concentration from 155 ± 5 mg/L to 52 ± 6 mg/L.As a result, the removal rate of total nitrogen increasedfrom 0.34 ± 0.20 kg N/m3/d to 0.43 ± 0.11 kg N/m3/d,and the nitrogen removal efficiency increased from 54%to 70%. The use of an AEM may also reduce the con-centrations of nitrate/nitrite in the final effluent, becausethose anions could migrate into the anode of MFC-1 through the AEM and be removed by conventionaldenitrification.

The efficiency of substrate-to-electricity was repre-sented by coulombic efficiency (CE). There were twodifferent CEs in this study, CEC based on organics (COD)and CEN based on nitrogen (nitrate). The cooperative MFCsystem achieved a CEC of 8.9% when treating the syn-thetic solution (acetate); this efficiency decreased to 0.3%(or 4.0% based on soluble COD) with digested sludge and0.6% (or 6.2% based on soluble COD) with landfill leachate,indicating that the majority of the removed organics wasnot associated with electricity generation. The low CEswere most likely due to the high loading rates, especially

Table 1. Comparison of COD removal rate and nitrogen removal rates (kg/m3/d) between this study and the literature (the studies usingnitrate instead of ammonium as the initial nitrogen compounds are excluded from this comparison).

COD COD Nitrogen NitrogenMFC configuration Organic substrate (kg/m3/d)∗ (kg/m3/d)† (kg/m3/d)‡ (kg/m3/d)¶ References

Two chambers Acetate 0.84 0.42 0.20 0.100 [11]Two chambers Glucose 1.44 0.26 0.06 0.052 [23]Two chambers Acetate 2.00 1.00 0.06 0.051 [10]Three chambers Glucose 4.40 0.88 0.03 0.030 [12]Three chambers Acetate 0.01 0.003 0.005 0.003 [14]Three chambers Acetate 2.09 0.19 0.01 0.013 [15]Cooperative MFC system Acetate 3.37 0.61 0.08 0.037 This studyCooperative MFC system Digested sludge 27.44 4.94 0.34 0.158 This studyCooperative MFC system Leachate 28.53 5.14 0.43 0.199 This study

Notes: ∗Based on the total anode liquid volume; †based on the total liquid volume of the MFC system; ‡based on the total cathode liquidvolume; ¶based on the total liquid volume of the MFC system.

Dow

nloa

ded

by [

Uni

vers

ity o

f N

ew H

amps

hire

] at

06:

18 1

3 M

arch

201

3

Environmental Technology 7

with actual wastes, which were beyond the conversioncapacity of the electrochemically-active microorganisms onthe anode electrodes. Another important reason for a lowCEC was that both the MFC-1 and MFC-2 outer cathodeswere operated at high power output, which produced lowercurrent compared with the high-current condition. Powerproduction (or energy production) at the high-current con-dition would be very low. Therefore, a high CEC is notnecessarily the goal of MFC operation. On the other hand,a high CEN is desired because bioelectrochemical denitrifi-cation is a critical factor in nitrate removal as well as in totalnitrogen removal. With the synthetic solution, the MFC sys-tem achieved a CEN of 17%. When the feeding solution waschanged to high organic-loading solutions such as digestedsludge or landfill leachate, the CEN decreased to 1.3% (atfull strength X0), 3.2% (at X20) or 8.4% (leachate). Theoversupply of organics at a high loading rate resulted ina high concentration of remaining organics in the effluentof the MFC-2 outer cathode, which provided electrons tonitrate in the inner cathode for conventional denitrification.Therefore, current generation from the inner cathode greatlydecreased because of the competition of the electron sup-ply between the organic compounds and the inner cathodeelectrode.

The study with the actual wastes revealed a great chal-lenge in the future application of the cooperative MFCsystem for treating high-strength/high-solid wastes: theoverload of organic compounds makes it difficult to pre-cisely control the distribution of organic content in differentcompartments. The excessive organic flux into the cath-odes (both inner and outer) will reduce energy productionand thus lower the significance of using MFC technology.Therefore, high-strength/high-solid wastes may not be suit-able for MFC treatment unless the treatment performanceof the anode can be further improved. System scale-up isanother challenge for all MFC processes.

3.3. Energy balanceAn important feature of MFC technology is its potentialenergy benefits, which can be understood from the aspectsof both energy saving and energy production [22]. Asan anaerobic treatment technology, MFCs are expected to

consume less energy than the activated sludge processes.In the present MFC system, the energy consumption wasmainly due to the pumping system (feeding and recircu-lation) and the aeration in the cathodes, and the data canbe expressed in either kWh/m3 treated water, or kWh/kgCOD (removed). Because the MFC system was operatedunder the same condition (feeding rate, recirculation andaeration) when treating different solutions, the total energyconsumption was the same at 0.0535 kWh/m3 (Table S1).In comparison, the energy consumption of an activatedsludge process is about 0.30 kWh/m3, which was estimatedaccording to the information in a previous publication [1].However, when the energy consumption was expressedin kWh/kg COD, which is expected to be more precisebecause of different organic concentrations in the feedingsolutions, the total energy consumption of the present MFCsystem varied between 0.0054 and 0.1071 kWh/kg CODbecause of different COD loading rates (Table 2). In the caseof the digested sludge, the energy consumption per kg CODdecreased with the increasing organic loading rates, indicat-ing the potential advantages of lower energy consumptionof MFC technology for treating high-strength wastes. How-ever, this advantage needs to be carefully evaluated bycomparing it with anaerobic digestion that is commonlyapplied to treat high-strength wastes, especially from theaspect of energy production. The activated sludge process,on the other hand, consumes about 0.67 kWh/kg COD whentreating domestic wastewater [1]. Clearly, the MFC sys-tem consumes much less energy than an activated sludgeprocess.

The energy production in the MFC system was thesum of the electric energy produced in the MFC-1 andMFC-2 outer cathodes. In general, the MFC system pro-duced 0.0049 to 0.1196 kWh/m3 (Table S1) or 0.011 to0.1023 kWh/kg COD (Table 2). An energy balance wasestablished and the difference (net energy) between theenergy production and the energy consumption is shown inTable S1 and Table 2. The MFC system could theoreticallyachieve a positive net energy when treating the syntheticsolution while the energy balance was generally negativewith either digested sludge or landfill leachate. Becauseof the same energy consumption (kWh/m3) when treat-ing different solutions, the negative energy balance with

Table 2. Energy production and consumption in the cooperative MFC system.

Produced energy Feeding energy Recirculation energy Aeration energy Total consumed Net energySubstrate (kWh/kg COD) (kWh/kg COD) (kWh/kg COD) (kWh/kg COD) (kWh/kg COD) (kWh/kg COD)

Acetate 0.1023 0.0002 0.0105 0.0351 0.0458 0.0565DS-X20 0.0097 0.0005 0.0245 0.0821 0.1071 −0.0974DS-X10 0.0038 0.0002 0.0091 0.0305 0.0398 −0.0361DS-X5 0.0028 0.0001 0.0055 0.0185 0.0241 −0.0213DS-X0 0.0011 0.0000 0.0013 0.0043 0.0056 −0.0045Leachate 0.0019 0.0000 0.0012 0.0041 0.0054 −0.0035

Note: DS = digested sludge.

Dow

nloa

ded

by [

Uni

vers

ity o

f N

ew H

amps

hire

] at

06:

18 1

3 M

arch

201

3

8 F. Zhang and Z. He

the actual wastes was due to the low energy production.The high organic loading rates in the actual wastes couldpotentially provide more electrons for electricity genera-tion than the synthetic solution; however, the large amountof organic compounds remaining from the anode oxidationwent into the (aerobic) cathode compartments and stimu-lated the growth of heterotrophic bacteria. Consequently,those organic compounds competed with the cathode elec-trodes for oxygen. The biofilm formed on the cathodeelectrode would also reduce oxygen transfer to the cathodeelectrode for electrochemical reduction, thereby decreas-ing cathode performance as well as electricity generation.Those results suggest that, although anaerobic treatment inthe anodes resulted in low energy consumption in the MFCsystem when treating the high-strength wastes, the anodeprocesses must be improved to ensure a sufficient removal oforganic compounds that is important to the cathode reactionand electricity generation.

4. ConclusionsThis study has developed and demonstrated a cooperativeMFC system for simultaneous waste treatment and bioen-ergy production. This system, consisting of a regular MFC(for organic removal) and a dual-cathode MFC (for nitro-gen removal), could effectively reduce the concentrationsof both organic and nitrogen compounds in either syn-thetic solutions or actual wastes (digested sludge or landfillleachate). The energy analysis showed that the MFC sys-tem could theoretically achieve positive net energy withthe synthetic solution; however, the energy consumptionwhen treating the actual wastes was higher than the energyproduction. These results indicate that the primary benefitof MFC technology should be energy saving (low energyconsumption), especially compared with activated sludgeprocesses when treating medium/low-strength wastewaters.For high-strength wastes, the MFC system would require alonger HRT or dilution of the feeding solution (e.g. withthe treated effluent) to prevent organics from entering thecathode. From a perspective of energy production (and anenergy-neutral system), MFC technology may be more suit-able for treating low-strength wastewater containing highsoluble COD contents.

AcknowledgementsThis project is financially supported by the National ScienceFoundation Industry/University Cooperative Research Center(I/UCRC) – the Center for Water Equipment and Policy. Theauthors would like to thank Dr Marjorie Piechowski (UW-Milwaukee) for her help with manuscript proofreading.

Supplementary informationThe removal rates of COD and nitrogen (Figure S1) and energyanalysis in kWh/m3 (Table S1) appear in the supplementaryinformation.

References[1] McCarty PL, Bae J, Kim J, Domestic wastewater treatment

as a net energy producer – Can this be achieved? EnvironSci Technol. 2011;45:7100–7106.

[2] Logan BE, Hamelers B, Rozendal RA, Schroder U, KellerJ, Freguia S, Aelterman P, Verstraete W, Rabaey K. Micro-bial fuel cells: methodology and technology. Environ SciTechnol. 2006;40:5181–5192.

[3] Potter MC. Electrical effects accompanying the decompo-sition of organic compounds. P Roy Soc Lond. 1912;84:260–276.

[4] Logan B. Essential data and techniques for conductingmicrobial fuel cell and other types of bioelectrochemicalsystem experiments. ChemSusChem. 2012;5:988–994.

[5] Cord-Ruwisch R, Law Y, Cheng KY. Ammonium as asustainable proton shuttle in bioelectrochemical systems.Bioresource Technol. 2011;102:9691–9696.

[6] Kuntke P, Smiech KM, Bruning H, Zeeman G, Saakes M,Sleutels TH, Hamelers HV, Buisman CJ. Ammonium recov-ery and energy production from urine by a microbial fuelcell. Water Res. 2012;46:2627–2636.

[7] Gregory KB, Bond DR, Lovley DR. Graphite electrodes aselectron donors for anaerobic respiration. Environ Microbiol.2004;6:596–604.

[8] Clauwaert P, Rabaey K, Aelterman P, Schamphelaire LD,Pham TH, Boeckx P, Boon N, Verstraete W. Biologicaldenitrification in microbial fuel cells. Environ Sci Technol.2007;41:3354–3360.

[9] Clauwaert P, Desloover J, Shea C, Nerenberg R, BoonN, Verstraete W. Enhanced nitrogen removal in bio-electrochemical systems by pH control. Biotechnol Lett.2009;31:1537–1543.

[10] Virdis B, Rabaey K, Yuan Z, Keller J. Microbial fuel cellsfor simultaneous carbon and nitrogen removal. Water Res.2008;42:3013–3024.

[11] Virdis B, Rabaey K, Rozendal RA, Yuan Z, Keller J. Simul-taneous nitrification, denitrification and carbon removal inmicrobial fuel cells. Water Res. 2010;44:2970–2980.

[12] Xie S, Liang P, Chen Y, Xia X, Huang X. Simultaneous car-bon and nitrogen removal using an oxic/anoxic-biocathodemicrobial fuel cells coupled system. Bioresource Technol.2011;102:348–354.

[13] Virdis B, Rabaey K, Yuan Z, Rozendal RA, Keller J. Electronfluxes in a microbial fuel cell performing carbon and nitrogenremoval. Environ Sci Technol. 2009;43:5144–5149.

[14] Zhang F, He Z. Simultaneous nitrification and denitrifica-tion with electricity generation in dual-cathode microbialfuel cells. J Chem Technol Biot. 2012;87:153–159.

[15] Zhang F, He Z. Integrated organic and nitrogen removal withelectricity generation in a tubular dual-cathode microbial fuelcell. Process Biochem. 2012;47:2146–2151.

[16] Wang X, Cheng S, Feng Y, Merrill MD, Saito T, Logan BE.Use of carbon mesh anodes and the effect of different pre-treatment methods on power production in microbial fuelcells. Environ Sci Technol. 2009;43:6870–6874.

[17] Zhang F, Jacobson KS, Torres P, He Z. Effects of anolyterecirculation rates and catholytes on electricity generation ina liter-scale upflow microbial fuel cell. Energy Environ Sci.2010;3:1347–1352.

[18] Angenent LT, Sung S. Development of anaerobic migrat-ing blanket reactor (AMBR), a novel anaerobic treatmentsystem. Water Res. 2001;35:1739–1747.

[19] Kim J, Kim K, Ye H, Lee E, Shin C, McCarty PL,Bae J. Anaerobic fluidized bed membrane bioreactorfor wastewater treatment. Environ Sci Technol. 2011;45:576–581.

Dow

nloa

ded

by [

Uni

vers

ity o

f N

ew H

amps

hire

] at

06:

18 1

3 M

arch

201

3

Environmental Technology 9

[20] Larson L, Rosso D, Leu S-YB, Stenstrom MK. Energy-conservation in fine pore diffuser installations in activatedsludge processes. University of California – Los Angeles:Los Angeles (US); 2007.

[21] Zhao F, Hamisch F, Schroder U, Scholz F, Bogdanoff P,Hermann I. Challenges and constraints of using oxygen cath-odes in microbial fuel cells. Environ Sci Technol. 2006;40:5193–5199.

[22] He Z. Microbial fuel cells: Now let us talk about energy.Environ Sci Technol. 2013;47:332–333.

[23] Cha J, Kim C, Choi S, Lee G, Chen G, Lee T. Evalua-tion of microbial fuel cell coupled with aeration chamberand bio-cathode for organic matter and nitrogen removalfrom synthetic domestic wastewater. Water Sci Technol.2009;60:1409–1418.

Dow

nloa

ded

by [

Uni

vers

ity o

f N

ew H

amps

hire

] at

06:

18 1

3 M

arch

201

3