Environ. Sci. Technol.

2004, 38, 2281-2285

but these technologies are generally only suitable for high-

Production of Electricity during strength wastewater streams typically produced by industry.
Wastewater Treatment Using a In addition, conventional anaerobic treatment produces
methane gas, which if released, can contribute to global
Single Chamber Microbial Fuel Cell warming. One method has been advocated to reduce
wastewater treatment costs, finding ways to produce useful
products from wastewater treatment (2). However, effective
HONG LIU,†
methods for generating products other than methane from
RAMANATHAN RAMNARAYANAN,‡ AND
wastewater have not yet been shown to be feasible.
B R U C E E . L O G A N * ,†,§
It has been known for several years that bacteria can be
Department of Civil and Environmental Engineering,
used to generate electricity that can be harvested in microbial
Department of Chemistry, and The Penn State Hydrogen
Energy (H2E) Center, The Pennsylvania State University,
fuel cells (MFCs) (3). In a MFC, bacteria that oxidize a
University Park, Pennsylvania substrate are kept physically separated from the electron
acceptor by a proton exchange membrane. Electrons pass
from the bacteria to the electrode (anode) in the same
chamber and then via a circuit to the cathode where they
Microbial fuel cells (MFCs) have been used to produce combine with protons and oxygen to form water. The
electricity from different compounds, including acetate, difference in the potential coupled to electron flow produces
electricity in this fuel cell. Electrons generated by bacteria
lactate, and glucose. We demonstrate here that it is also
through the anaerobic oxidation of organic matter (by
possible to produce electricity in a MFC from domestic nonfermentative reactions) are passed to respiratory enzymes
wastewater, while at the same time accomplishing biological normally bound to the inner cell membrane. Artificial electron
wastewater treatment (removal of chemical oxygen carriers (mediators), such as neutral red or anthraquinone-
demand; COD). Tests were conducted using a single 2,6-disulfonate (AQDS), or high concentrations of humic acids
chamber microbial fuel cell (SCMFC) containing eight can be used to carry electrons from inside the cell to an
graphite electrodes (anodes) and a single air cathode. The external electrode (3-6). However, the need for high
system was operated under continuous flow conditions concentrations of electron carriers, many of which are toxic
with primary clarifier effluent obtained from a local wastewater chemicals, is thought to make electricity generation on a
treatment plant. The prototype SCMFC reactor generated large scale in a mediator-type fuel cell impractical. Some
bacteria, such as Shewanella putrefaciens, however, can
electrical power (maximum of 26 mW m-2) while removing
produce their own mediators (soluble quinones) (7) that can
up to 80% of the COD of the wastewater. Power output eliminate the need to add mediators.
was proportional to the hydraulic retention time over a range It was only recently discovered that the respiratory
of 3-33 h and to the influent wastewater strength over enzymes of certain iron-reducing bacteria span their outer
a range of 50-220 mg/L of COD. Current generation was membrane allowing direct transfer of electrons to external
controlled primarily by the efficiency of the cathode. Optimal metals such as Fe (III) or Mn (IV) (8-10). The attachment of
cathode performance was obtained by allowing passive these iron-reducing bacteria to carbon electrodes results in
air flow rather than forced air flow (4.5-5.5 L/min). The electron transfer to the anode, with oxygen reduction at the
Coulombic efficiency of the system, based on COD removal cathode (11). The measured equilibrium potentials (-0.17
and current generation, was <12% indicating a substantial V) are close to those of purified preparations of c-type
fraction of the organic matter was lost without current cytochromes (-0.19 V) used in the respiratory chains of these
bacteria, suggesting these enzymes are responsible for
generation. Bioreactors based on power generation in MFCs
electron transfer to the electrode (11). Power generated in
may represent a completely new approach to wastewater various types of MFCs systems vary widely as a function of
treatment. If power generation in these systems can be the inoculum, substrate, and reactor, ranging from <1 mW/
increased, MFC technology may provide a new method to m2 (12) with lactate and pure cultures of S. putrefacians to
offset wastewater treatment plant operating costs, 3600 mW/m2 with glucose and an unidentified mixed culture
making advanced wastewater treatment more affordable (13). Most results with carbon electrodes generally report
for both developing and industrialized nations. power generation rates on the order of 10-100 mW/m2 (5,
11, 14-16).
It has not previously been examined whether electricity
Introduction can be generated using domestic wastewater or if MFCs could
be used for domestic wastewater treatment. In addition,
Over 126 billion liters of domestic wastewater is treated each current designs for mediator-less MFCs have not been
day in the U.S. at an annual cost of over $25 billion (1). Most developed specifically for continuous wastewater treatment.
of this wastewater is sent to centralized facilities that consume Two types of MFCs have so far been used: small batch-fed
large amounts of energy for treatment due to aeration. These laboratory systems primarily using defined substrates and
high energy costs for sanitation to levels required in the U.S. MFCs developed to harvest energy from marine sediments.
cannot be borne by a global population of six billion people, Laboratory systems typically use separate compartments for
particularly in developing countries. Anaerobic treatment the anode and cathode connected by a proton exchange
technologies provide potential for reducing treatment costs, membrane (PEM) (11, 14). This two-compartment design
would be difficult to apply to larger systems for continuously
* Corresponding author phone: (814)863-7908; fax: (814)863-7304;
treating wastewater. Marine sediment systems are based on
e-mail: blogan@psu.edu.
† Department of Civil and Environmental Engineering. the placement of a carbon pad into an anaerobic sediment
‡ Department of Chemistry. (anode) and a second pad placed in overlaying oxygenated
§The Penn State Hydrogen Energy (H E) Center.
2 waters (cathode) (5, 17). The latter systems cannot be adapted
10.1021/es034923g CCC: $27.50  2004 American Chemical Society VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 2281
Published on Web 02/21/2004
FIGURE 1. (A) Schematic and (B) laboratory-scale prototype of the SCMFC used to generate electricity from wastewater.

for continuous treatment of a wastewater due to the need MA) placed in a concentric arrangement about a single
for large volumes of oxygenated water and the limited cathode (Figure 1). The graphite rods were abraded by sand
solubility of dissolved oxygen in water. Thus, changes in paper to enhance bacterial attachment. The air-porous
reactor design are needed to produce MFCs suitable for cathode consisted of a carbon/platinum catalyst/proton
treating continuous streams of dissolved organic matter. exchange membrane (PEM) layer fused to a plastic support
To continuously generate electricity from organic matter tube. The PEM (Nafion 117, Dupont, Wilmington, DE) was
in water, a MFC reactor must have both a large surface area sequentially boiled in H2O2 (30%), deionized water, 0.5 M
for biofilm formation and a high void volume. Lessons learned H2SO4, and then deionized water (each time for 1 h). The
from the design of fixed-film bioreactors used for wastewater PEM was then hot pressed directly onto carbon cloth loaded
treatment can be applied to MFCs. Wastewater treatment with 0.5 mg cm-2 of Pt (E-Tek, Miami, FL) by heating it to
reactors are designed to be highly porous to avoid clogging 140 °C at 1780 kPa for 3 min. The cathode/PEM was placed
the reactor with high concentrations of particles in the onto a 25 mm diameter plastic (Plexiglas) tube containing
wastewater. This requirement for high porosity is at odds 2 mm diameter pores at 2 mm intervals (cathode tube). Air
with the need for a large surface area for biofilm formation. flow through the tube was initially 4.5-5.5 mL/min but was
Thus, fixed-film wastewater reactors have been designed to changed to passive oxygen transfer (no forced air flow) except
produce thin fluid films and convective flow rates sufficiently as noted. Copper wire was used to connect the circuit
high to shear off excess biofilm growth from surfaces. These containing a 465 Ω load unless stated otherwise.
reactors cannot be used without modification for energy Local domestic wastewater (primary clarifier effluent) was
production because MFCs have an additional requirement obtained from the primary clarifier of the Pennsylvania State
that the biomass be kept separated from any dissolved University Wastewater Treatment plant. Wastewater was
oxygen. These design challenges are similar to requirements pumped into the reactor and used as the fuel without any
for membrane bioreactors used to treat water and wastewater. modifications except for dilution of the wastewater with
In these reactors, a series of fibrous nanopore-sized mem- deionized water in some experiments. Wastewater fed to the
branes are inserted into a single, sheared compartment reactor had a pH ranging from 7.3 to 7.6 and a COD of 210
providing a high porosity tank containing the membrane or 220 mg/L. The hydraulic retention time in the reactor
(18). The high shear generated in the reactor by bubbles is ranged from 3 to 33 h as noted. All measurements were taken
used to help shear biofilms from the fibers. after the reactor had been operated for at least six hydraulic
The methods used to achieve a high porosity in membrane retention times, when power output was consistent. Mea-
reactors suggested a method to improve the design of a MFC surements reported here are averages of triplicate measure-
by providing a chamber that contained both the anode and ments taken over three consecutive retention times. Two
the cathode in a single compartment MFC (SCMFC). A different controls were used to account for COD losses. The
prototype reactor was built and tested using wastewater first control consisted of a second tank of similar volume
obtained from a local wastewater treatment plant. We report (400 mL) to account for COD or BOD losses unrelated to fuel
here our preliminary findings using this reactor that dem- cell operation (volume control). The second control consisted
onstrate for the first time electricity generation accompanied of a separate set of experiments using the SCMFC with an
by wastewater treatment demonstrated by the removal of open circuit (disconnected electrodes; electrode control). All
organic matter in the form of chemical oxygen demand (COD) experiments were conducted in a temperature-controlled
or biochemical oxygen demand (BOD). room set at 30 °C.
In the first set of experiments, the SCMFC was inoculated
Materials and Methods with a pure culture (Geobacter metallireducens, 10 mL at 0.5
The SCMFC consisted of a single cylindrical plexiglass g-DW/L), filled with wastewater, and operated in batch mode
chamber (15 cm long by 6.5 cm diameter; empty bed volume for one week. However, other experiments in our laboratory
of 388 mL) containing eight graphite rods (anode) each 6.15 and by others (19) demonstrated that this inoculation step
mm in diameter and 150 mm long (Alfa Aesar, Ward Hill, was not needed to produce electricity in a MFC. Therefore,

2282 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004

 FIGURE 3. Removal of organic matter in terms of COD (b) and BOD
(O) by the SCMFC as a function of wastewater strength (465 Ω, HRT
FIGURE 2. Power generated in a SFMFC as a function of influent ) 12 h, and air flow rate of 5.5 L/min; reactor initially inoculated
wastewater concentration (465 Ω, HRT ) 12 h, and air flow rate with G. metallireducens). (Error bars (SD based on triplicate
of 5.5 L/min; reactor initially inoculated with G. metallireducens). measurements of the reactor effluent.)

except as stated otherwise, the SCMFC was inoculated using

only the bacteria present in the wastewater.
Oxygen transfer through the membrane was estimated
based on measurements of oxygen flux through Nafion
membranes in hydrogen fuel cells (20). The rate (W) of oxygen
transfer through a membrane is a product of the cross section
A and flux (J), or

dC ∆C
W ) JA ) -DA = -DA (1)
dx ∆x

For a cross-sectional area of 102 cm2, and assuming saturation

of oxygen in the membrane (3.7-6.2 × 10-7 mol/cm3), no
oxygen in the anode section, a membrane thickness of 190
µm, and diffusion constants of oxygen in the membrane that
range from 1 to 6 × 10-6 cm2/s (20), oxygen transport into
the anode chamber would range from 0.23 to 2.3 mg of O2/h. FIGURE 4. Effect of oxygen flow on electricity generation (HRT of
Analyses. Voltage was continuously measured by a 3 h). Values indicate air flow rate (L/min) (COD ) 210 mg/L, 465 Ω).
multimeter with a data acquisition system (2700, Keithly,
Cleveland, OH) and converted to power according to P ) iV, decreased the voltage in the reactor by 25% over a 7.5 h
where P ) power, i ) current, and V ) voltage. Power was period. Therefore, in subsequent experiments only passive
normalized by the total surface area of the anodes. COD and air transfer was used at the cathode.
BOD were measured according to standard methods (21). Power output was a function of circuit load (16-5000 Ω)
Total electron flow was calculated by summing the product in the system, consistent with trends observed by others in
of the time interval and the current passed through the circuit. other MFC experiments (11, 23). At a hydraulic detention
Coulombic efficiency was calculated as the total coulombs time of 6 h, power reached a maximum of 26 mW/m2, with
measured divided by the moles of COD removed assuming a current density of 125 mA/m2 (69 Ω resistor) in the absence
4 mol of electrons/mol of COD. of forced air flow through the cathode tube (Figure 5). In
contrast, a maximum of only 12.2 mW/m2 was achieved in
Results and Discussion the system with air flow at 5.5 L/min at a longer detention
Electricity in the SCMFC was generated using domestic time (HRT ) 12 h).
wastewater in proportion to wastewater strength (Figure 2). COD removal and power were a function of hydraulic
The voltage across the circuit containing a 465 Ω resistor detention times of the wastewater in the reactor. COD
increased to 0.32 V for wastewater containing up to 220 mg/L removal increased from 40% at a HRT ) 3 h to a maximum
of COD, producing up to 9 mW/m2 of power based on anode of 80% at a HRT of 33 h (passive air cathode; 465 Ω) (Figure
surface area (HRT ) 12 h; air flow rate ) 5.5 L/min). Organic 6). A large percent of the COD removal was not associated
matter removal in terms of BOD was nearly constant at with power generation. When the circuit was disconnected,
78(2%, while the reduction in COD ranged from 50 to 70% COD removal ranged from 20 to 50% over the same range
(Figure 3). Changes in the amount of BOD removed relative of HRT (3-33 h), with an average of 36(13% of the COD
to COD removed reflect differences in the biodegradability removal due to processes not associated with power genera-
of the wastewater samples fed to the reactor, large statistical tion (Figure 6). Power density decreased from 17 to 12 mW/
variations in BOD test results ((20% even at a single m2 over this range of HRTs. The Coulombic efficiency of the
laboratory), and some losses of BOD that may have occurred system ranged from 3 to 12%, indicating there was substantial
in the volume-control tank (no electrodes) (5-20% variation COD that was not associated with power generation. While
in BOD). Not all of the organic matter in a wastewater can these Coulombic efficiencies are low, those reported by others
be biologically degraded, so that COD removals are typically have varied over a large range of 0.04-97% (5, 11, 16).
lower than those for BOD removal (21). COD removal in the absence of power generation cannot
Forced air flow through the cathode tube was found to be explained solely on the basis of oxygen transfer across the
decrease power generation (Figure 4). Compared to passive air-cathode. On the basis of values reported in Basura et al.
air flow, forcing air through the cathode tube at 4.5 L/min (19), the oxygen flux through the membrane is estimated as

VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 2283

 FIGURE 6. (A) COD removal rate in MFC (4) as compared to control
reactor (2; circuit disconnected). (B) Power density (b) and current
recovery (O) in MFC as a function of HRT (COD ) 210 mg/L, 465 Ω).

FIGURE 5. Voltage (O) and power (b) generated in the SCMFC expensive electrodes and membranes and that minimize
when resistance (electrical load) on the system is changed from catalyst content. Several reactor systems have used graphite
16 to 5000 Ω. (A) HRT ) 12 h, COD ) 220 mg/L (cathode air flow rods as electrodes (5, 14, 18), but graphite felt or carbon
rate ) 5.5 L/min; reactor initially inoculated with G. metallireducens). cloth may provide less expensive alternatives (15). Other
(B) HRT ) 6 h, COD ) 210 mg/L (passive cathode air flow; wastewater alternatives include conductive materials with coatings such
inoculum only).
as polyaniline that were recently shown to boost current
much as 2.3 mg/h. While this could achieve as much as 93% densities in MFCs by more than an order of magnitude (24).
of the COD removal at the longest detention time (33 h), it Advances are also being made in membranes that may result
could only produce a total of 9% of the COD removal at the in the replacement of the Nafion membrane used here with
3 h HRT in Figure 6. There could have been a loss of COD less expensive membranes (22).
using other electron acceptors (such as nitrate and sulfate) Providing adequate sanitation and clean water to a global
present in the wastewater. Biomass production could account population is an economic challenge. Advanced wastewater
for additional COD removal, but it was not possible to treatment systems that are mandated by law in the U.S. to
establish a complete mass balance of COD in this system. protect human health and the environment are too costly to
Implications for Using MFCs for Wastewater Treatment. be installed or operated and maintained in developing
Electricity was produced in a single-chamber MFC, but it countries where the need for sanitation is greatest to protect
was found that a large percentage of the organic matter in human health and the global spread of disease. Even in the
the wastewater was removed by processes that did not U.S., there are substantial economic challenges for main-
generate electricity. To increase the power density of the taining the current levels of water and wastewater treatment.
system, we will need to find ways to increase the fraction of It is estimated that $2 trillion is needed in the U.S. over the
the organic matter that is converted into electricity. However, next 20 years for building, operating, and maintaining
the organic matter in wastewater is essentially free. As long wastewater and drinking water facilities (1). About $45 billion
as COD removal is accomplished, by electricity generation is needed for wastewater infrastructure improvements in
or other methods, the goal of wastewater treatment is addition to current annual costs of $25 billion. Operating
achieved. It appears that substantial losses of COD here costs are a substantial factor in wastewater treatment, with
resulted from passive oxygen transport into the reactor by over half of the operating expenditures typically needed for
diffusion across the proton exchange membrane. COD wastewater aeration. Power production measured here could
reduction due to passive oxygen transfer may actually be provide much needed energy in the U.S. from domestic
beneficial when compared to the high cost of forced-air wastewater alone. Processes that can generate electricity
oxygen transfer using blowers in wastewater treatment during domestic and industrial treatment will help to reduce
systems such as activated sludge. Clearly, additional work the economic burden of treatment and provide access to
will need to be done to better understand how electricity sanitation technologies throughout the world.
generation affects COD removal and how operational factors
such as hydraulic retention time and air flow rate through
the cathode can be optimized.
Acknowledgments
The capital cost of the SCMFC used here could be We thank Tom Mallouk for advice on electrode construction
substantially reduced through system designs that use less and fuel cells and Booki Min for help with wastewater test

2284 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004

measurements. This research was supported by National (12) Kim, H. J.; Park, H. S.; Hyun, M. S.; Chang, I. S.; Kim, M.; Kim,
Science Foundation Grant BES-0331824. B. H. Enzyme Microbiol. Technol. 2002, 30, 145-152.
(13) Rabaey, K.; Lissens, G.; Siciliano, S. D.; Verstraete, W. Biotechnol.
Note Added after ASAP Posting Lett. 2003, 25, 1531-1535.
(14) Kim, H. J.; Hyun, M. S.; Chang, I. S.; Kim, B. H. J. Microbiol.
This paper was released ASAP on 02/21/2004 with an error Biotechnol. 1999, 9, 365-367.
in the second line of the Introduction. The correct version (15) Chaudhuri, S. K.; Lovley, D. R. Nat. Biotechnol. 2003, 21(10),
was posted on 03/03/2004. 1229-1232.
(16) Park, D. H.; Zeikus, J. G. Appl. Microbiol. Biotechnol. 2002, 59,
Literature Cited 58-61.
(1) Water infrastructure network (WIN). Clean safe water for the (17) Reimers, C. E.; Tender, L. M.; Ferig, S.; Wang, W. Environ. Sci.
21st century; 2001; http://www.amsa-cleanwater.org/advocacy/ Technol. 2001, 35, 192-195.
winreport/ winreport2000.pdf. (18) Rosenberger, S.; Krüger, U.; Witzig, R.; Manz, W.; Szewzyk, U.;
(2) M. E. Watanabe Consulting Inc. In Research Needs to Optimize Kraume, M. Water Res. 2002, 36, 413-420.
Wastewater Resource Utilization; Watanabe, M. E., Ed.; Water
(19) Basura V. I.; Beattie, P. D.; Holdcroft, S. J. Electroanal. Chem.
Environment Research Foundation (WERF): New York, 1999.
1998, 458, 1-5.
(3) Allen, R. M.; Benneto, H. P. Appl. Biochem. Biotechnol. 1993,
39/40, 27-40. (20) APHA, AWA, WPCF. In Standard Methods for the Examination
(4) Park, D. H.; Zeikus J. G. Appl. Environ. Microbial. 2000, 66, of Water and Wastewater, 18th ed.; American Public Health
1292-1297. Association: Washington, D.C., 1992.
(5) Bond, D. R.; Holmes, D. E.; Tender, L. M.; Lovley, D. R. Science (21) Grady, C. P. L., Jr.; Daigger, G. T.; Lim, H. C. In Biological
2002, 295, 483-485. Wastewater Treatment, 2nd ed.; Marcel Dekker: New York, 1999.
(6) Lovley, D. R.; Fraga, J. L.; Coates, J. D.; Blung-Harris, E. L. Environ. (22) Jang, J. K.; Pham, T. H.; Chang, I. S.; Kang, K. H.; Moon, H.; Cho,
Microbiol. 1999, 1(1), 89-98. K. S.; Kim, B. H. Proc. Biochem., 2004, in press.
(7) Newman, D. K.; Koler, R. Nature 2000, 405, 94-97. (23) Park, D. H.; Zeikus J. G. Biotechnol. Bioeng. 2003, 81, 348-355.
(8) Nevin, K. P.; Lovley, D. R. Appl. Environ. Microbiol. 2000, 66,
2248-2251. (24) Schröder, U.; Nieβen, J.; Scholz, F. Angew. Chem., Int. Ed. 2003,
(9) Lovley, D. R. Environmental microbe-metal interactions; ASM 42, 2880-2883.
Press: Washington, D.C.
(10) Leang, C.; Coppi, M. V.; Lovley, D. R. J. Bacteriol. 2003, 185(7), Received for review August 21, 2003. Revised manuscript
2096-2103. received January 9, 2004. Accepted January 15, 2004.
(11) Bond, D. R.; Lovely, D. R. Appl. Environ. Microbiol. 2003, 69,
1548-1555. ES034923G

VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 2285