Available online at www.sciencedirect.

com

Journal of Power Sources 179 (2008) 274–279

Short communication

Scale-up of membrane-free single-chamber microbial fuel cells

Hong Liu a,∗ , Shaoan Cheng b , Liping Huang b,d , Bruce E. Logan b,c
a Department of Biological and Ecological Engineering, Oregon State University, 116 Gilmore Hall, Corvallis, OR 97331, USA
b Department of Civil and Environmental Engineering, The Pennsylvania State University, University Park, PA 16802, USA
c Penn State Hydrogen Energy (H E) Center, The Pennsylvania State University, University Park, PA 16802, USA
2
d School of Environmental and Biological Science & Technology, Dalian University of Technology, Dalian 116024, China

Received 14 December 2007; received in revised form 24 December 2007; accepted 31 December 2007
Available online 17 January 2008

Abstract
Scale-up of microbial fuel cells (MFCs) will require a better understanding of the effects of reactor architecture and operation mode on volumetric
power densities. We compared the performance of a smaller MFC (SMFC, 28 mL) with a larger MFC (LMFC, 520 mL) in fed-batch mode. The
SMFC produced 14 W m−3 , consistent with previous reports for this reactor with an electrode spacing of 4 cm. The LMFC produced 16 W m−3 ,
resulting from the lower average electrode spacing (2.6 cm) and the higher anode surface area per volume (150 m2 m−3 vs. 25 m2 m−3 for the SMFC).
The effect of the larger anode surface area on power was shown to be relatively insignificant by adding graphite granules or using graphite fiber
brushes in the LMFC anode chamber. Although the granules and graphite brushes increased the surface area by factors of 6 and 56, respectively,
the maximum power density in the LMFC was only increased by 8% and 4%. In contrast, increasing the ionic strength of the LMFC from 100
to 300 mM using NaCl increased the power density by 25% to 20 W m−3 . When the LMFC was operated in continuous flow mode, a maximum
power density of 22 W m−3 was generated at a hydraulic retention time of 11.3 h. Although a thick biofilm was developed on the cathode surface in
this reactor, the cathode potentials were not significantly affected at current densities <1.0 mA cm−2 . These results demonstrate that power output
can be maintained during reactor scale-up; increasing the anode surface area and biofilm formation on the cathode do not greatly affect reactor
performance, and that electrode spacing is a key design factor in maximizing power generation.
© 2008 Elsevier B.V. All rights reserved.

Keywords: Microbial fuel cell; Scale-up; Single chamber; Membrane free

1. Introduction There have been a few tests of power generation with rela-
tively large, two-chamber systems (∼1 L) and a ferricyanide
Microbial fuel cells (MFCs) use microorganisms as catalysts catholyte [16,17], but such systems are not sustainable due
to directly generate electricity from organic matter. MFCs have to the need to chemically regenerate ferricyanide. Air-cathode
great potential as a method of wastewater treatment and as power MFCs hold a greater promise for practical applications due to
sources for autonomous sensors [1–5]. Extensive recent stud- their simple configuration, sustainable operation, and relatively
ies have led to a better understanding of the electron transfer high power densities [1,14]. The liquid volumes of many air-
mechanisms between cells and surfaces [6–8], and have shown cathode MFCs, however, have been relatively small (tens of
several factors that can affect MFC performance, including: elec- mL). Although an air-cathode MFC of 1.5 L was developed [18],
trode materials [9–11], solution chemistry [12,13] and reactor the maximum volumetric power density of 2 W m−3 was up to
configuration [14–17]. three orders of magnitude lower than smaller air-cathode MFCs
Scale-up is an important issue for the application of MFCs, [15,19,20]. MFCs can be stacked together in series or in parallel
especially in the field of wastewater treatment, but there is little to achieve higher voltage or current [21–23]. However, stacking
information available on the effects of scaling on power output. multiple MFCs together in series can result in problems, such
as voltage reversal, contact voltage losses, and erratic opera-
tion [21]. Producing larger MFCs can alter electrode spacing,
∗ Corresponding author. Tel.: +1 541 737 6309; fax: +1 541 737 2082. and thus affect power density through changes in the area-
E-mail address: liuh@engr.orst.edu (H. Liu). specific internal resistance. In small air-cathode MFCs, electrode

0378-7753/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.jpowsour.2007.12.120
 H. Liu et al. / Journal of Power Sources 179 (2008) 274–279 275

spacing and anode surface area have been shown to affect power cific surface area A = 6θ/d [25] for d = 4 mm, and θ = 0.53, but
output [24]. More information is needed on how these key design decreased the liquid volume to 280 mL. Cloth (Pledge cloth;
factors affect reactor performance as reactor size is increased. S.C. Johnson & Son, Inc., USA) was placed between the gran-
In this study, we constructed and operated a larger MFC ules and the cathode to avoid short-circuiting of the electrodes.
(LMFC) with a relatively higher anode surface area and reduced In another set of test, the baffles were removed and four graphite
electrode spacing, and compared its performance to a similar but fiber brushes (5 cm in diameter and 7 cm long) with an average
smaller MFC (SMFC). To better understand the effects of anode fiber diameter of 7.2 ␮m (PANEX33 160K, Zoltek) were used as
spacing and surface area on performance, we altered the elec- the anode, increasing the total anode surface area to 42,200 cm2
trode orientation and examined the effect of additional surface (7680 m2 m−3 ) [26]. The brushes were placed horizontally in
area by adding graphite granules or using graphite fiber brushes the chamber, with the titanium core protruding through a hole
as the anode. on the reactor side. The four electrodes were connected with
titanium wire. The electrode spacing, defined as that from the
brush center to the side of the cathode, was 3.3 cm.
2. Materials and methods

2.1. Larger MFCs 2.2. Smaller MFCs

The LMFC was constructed from Plexiglas (15 cm × Single-chamber SMFCs were constructed as previously
15 cm × 7 cm). This MFC contained five equally spaced baffles reported with a 28 mL liquid volume and a 4 cm electrode spac-
(12.7 cm × 4.3 cm × 0.6 cm) glued to the bottom of the reactor, ing of the two electrodes oriented in parallel to each other [14].
forming six channels. Each baffle had a slot (0.25 cm × 14.5 cm, The anode surface area per volume, As , was 25 m2 m−3 . The
0.25 cm from the top edge) and a hole (d = 1 cm, 1 cm to the side effect of anode orientation was examined by placing the anode
edge) to allow the anode and water go through the reactor. The perpendicular to that of the cathode, at a distance of 1–4 cm from
anode was a single piece of carbon cloth (non wet-proofing; the cathode, producing an average electrode spacing of 2.5 cm.
Type A, E-Tek, USA) with a total surface area of 757 cm2 Since the carbon cloth was too soft to be fixed in a perpendicular
(146 m2 m−3 ). The cloth was placed onto one side of the chamber orientation to the cathode in an SMFC, we conducted our per-
wall, across the bottom of each channel and around each baffle pendicular anode tests using carbon paper anodes (plain toray
(through the slots to avoid the contact with cathode), and then carbon paper, 7 cm2 , no wet-proofing; E-Tek, USA). Cathodes
held to the surface using small plastic screws (Fig. 1). The carbon were made of carbon cloth (7 cm2 ; 0.5 mg cm−2 Pt; 10% of Pt/C
cloth cathode (161 cm2 , 0.35 mg cm−2 Pt; 30% wet-proofing, E- catalyst; 30% wet-proofing; E-Tek, USA).
Tek) was coated with four diffusion layers to reduce water loss
as previously described [9]. The cathode was covered with a 2.3. MFC operation
thick plate of Plexiglas containing holes (1.3 cm diameter) to
allow oxygen diffusion to its surface and to hold the cathode The MFCs were inoculated using domestic wastewater and
against the chamber. The average electrode spacing was 2.6 cm a nutrient medium amended with sodium acetate (1 g L−1 ) as
(range of 0.5–4.3 cm). Strips of anode and cathode carbon cloth previously reported [4]. After replacing this solution twice over
were left to extend outside the reactor to allow connection of the 2 days, the system was then operated using only the nutrient
electrode to the circuit using alligator clips. medium and acetate (ionic strength, IS = 100 mM). The system
To further investigate the effect of surface area, graphite gran- was considered to be operating under steady conditions when the
ules (300 g) were added to the anode chamber of the LMFC. The voltage output was reproducible after refilling the reactor with
addition of graphite granules with particle diameters of 2–6 mm medium at least two times. Stable power output was normally
(Graphite Sales, Inc., Chagrin Falls, OH) increased the total achieved in 2–3 batches (about 1–2 days per batch). Polarization
anode surface area to 4500 cm2 (870 m2 m−3 ), assuming spe- curves were obtained by measuring the stable power generated

Fig. 1. (a) Anode chamber of the LMFC and (b) LMFC with cathode.
276 H. Liu et al. / Journal of Power Sources 179 (2008) 274–279

at various external resistances. For each resistance, the fuel cells

ran for at least two complete batch cycles to guarantee a stable
and sustainable voltage. In some tests using the LMFC, the IS
was increased to 300 mM using NaCl.
The LMFC was also operated in continuous flow mode at a
flow rate of 35–130 mL h−1 using a feed containing 800 mg L−1
acetate in nutrient medium. The theoretical retention time was
calculated from the volume of the medium and the flow rate
into the reactor. The actual hydraulic retention time (HRT) was
determined by spiking the feed line with a KCl solution, and
measuring the conductivity of the reactor effluent, as previously
described [27,28].

2.4. Analysis and calculations Fig. 2. Power generation as a function of current density in the SMFC, LMFC at
different ionic strengths (100 or 300 mM), and the LMFC with graphite granules
Voltage (V) was measured using a multimeter with a data or brushes.
acquisition system (2700, Keithley, USA), and used to calculate
the power (P) according to P = IV. Power was normalized by the
cross-sectional area (projected) of the cathode or reactor volume. the LMFC ranged from 38 to 52%, with an overall energy recov-
The internal resistance of the cell, Rint , was calculated from the ery of 8–12%. A low Coulombic efficiency is consistent with that
slope of plots of V and I using found in previous studies, and could be due to bacterial growth,
loss of substrate sustained by aerobic growth using oxygen that
V = Ecell − IR (1) diffused through the cathode, and other non-electricity generat-
ing processes such as methanogenesis, denitrification and sulfate
where Ecell is the electromotive force of the cell [1]. reduction [14].
Cathode potentials, either new or containing a biofilm (i.e.
from an MFC operated for 3 months), were measured at differ-
ent current densities by chronopotentiometry using a PC4/750 3.2. Effect of anode surface area on power generation
potentiostat (Gamry Instruments) and a three-chambered elec-
trochemical cell [9]. One chamber of the three-chambered To investigate that if increasing the anode surface area would
cell was used for the reference electrode, while the other two affect power generation, we added graphite granules or used
chambers housed the counter (anode) and working (cathode) graphite brushes as the anode which increased the overall sur-
electrodes. The catalyst-coated side of the cathode was placed face areas by factors of 6 and 56, respectively. A maximum
facing the solution with the other side exposed directly to air [9]. power density of 17 W m−3 (560 mW m−2 ) was obtained with
All reported potentials were therefore given as measured (based granules and 15 W m−3 (490 mW m−2 ) with brushes when the
on Ag/AgCl; +198 mV vs. standard hydrogen electrode). MFCs were operated in batch mode. Both are similar to the
The current–potential characteristics of cathode (with biofilm 16 W m−3 obtained with only the carbon cloth anode (Fig. 2).
and without biofilm) were fitted using This demonstrated that further increasing anode surface area had
a negligible effect on power generation, and that the cathode
E = A ln i + B (2) was the limiting electrode as a result of its significantly smaller
surface area in the LMFC.
where E (mV) is the cathode potential, i (mA cm−2 ) is the current
density of the cathode, A is the slope of the line, and B (mV) is
the cathode potential at 1 mA cm−2 . This equation is a variation
of the Tafel equation, which is commonly used in describ-
ing the overvoltage at the electrode surface in chemical fuel
cells [29].

3. Results

3.1. LMFC performance

The maximum power density generated by the LMFC con-

taining a cloth electrode was 16 W m−3 (520 mW m−2 -cathode
area) at a current density of 0.18 mA cm−2 , which was slightly
higher than that (14 W m−3 ) of the SMFC (Fig. 2). Based on the Fig. 3. Voltage as a function of current in the LMFC at different ionic
strengths (100 or 300 mM). The internal resistance is calculated from the slopes,
slope of the polarization curve, the larger system had an internal with regression lines of: IS = 300 mM, V = −7.3068i + 0.5513; IS = 100 mM,
resistance of Rint = 9.4  (Fig. 3). The Coulombic efficiency of V = −9.3799i + 0.5568.
 H. Liu et al. / Journal of Power Sources 179 (2008) 274–279 277

3.3. Effect of solution ionic strength

The performance was improved to 20 W m−3 (630 mW m−2 )

(Fig. 2) at current density of 0.26 mA cm−2 when the solution IS
was increased from 100 to 300 mM as a result of the decreased
internal resistance (Rint = 7.3 ) (Fig. 3). Thus, ionic strength
had a greater impact on power generation than anode surface
area.

3.4. Effect of anode orientation on power generation of the

SMFC

The LMFC had part of the anode oriented perpendicular to Fig. 5. Power density and substrate (acetate) removal in the larger MFC as a
that of the cathode, and part oriented parallel. To investigate function of HRT at external resistance of 4 .
whether the relative anode orientation would affect the power
output, we examined power output using the SMFC with anodes increased from 52% at a HRT = 4.1 h to a maximum of 90% at
oriented in parallel and perpendicular to the cathode. The maxi- a HRT of 16 h (Fig. 5). The fouling of the electrodes caused
mum power density of the SMFC with electrodes placed parallel by cell decay or death at a longer HRT might also contribute
to each other was 14 W m−3 . When the electrode was placed to a reduction in MFC performance. These results show that
perpendicular to the cathode, the power increased by 36% to power generation can be increased by continuous flow, rather
19 W m−3 mainly due to the reduced electrode spacing (4 cm than fed-batch operation, by optimizing HRT.
vs. 2.5 cm) (Fig. 4). Future research is needed to investigate
how the electrode orientation affects the power output with the 3.6. Effect of biofilm on cathode performance
same electrode spacing.
During operation of the LMFC, a thick biofilm was devel-
3.5. Continuous flow operation (LMFC) oped on the cathode even after only 1 week of operation. To
determine if this biofilm affected cathode performance, we com-
The performance of the LMFC was further evaluated by pared the current–potential characteristics of a new cathode to
switching the reactor from batch feeding to continuous flow that of a biofilm-covered cathode using an electrochemical cell.
mode (IS = 100 mM). The power density and acetate removal The cathode potential as a function of current density was well
were found to be a function of HRT (fixed external resistance described using Eq. (2) with constants of Ac = −94.8 mV and
of 4 ) (Fig. 5). Power density increased from 17 W m−3 at Bc = −176 mV (R2 = 0.9932) for the cathode without a biofilm,
4.1 h, to a maximum 22 W m−3 (695 mW m−2 ) at 11.3 h, and and Ac = −99.6 mV and Bc = −201 mV (R2 = 0.9938) for the
then decreased to 20 W m−3 when HRT further increased to cathode with a biofilm (Fig. 6). The difference between the
16 h. The lower power density generated at 4.1 h than that at values of Ac for cathodes with/without biofilm is very small,
11.3 h was possibly due to the relatively higher oxygen con- indicating charge transfer kinetics were not significantly affected
centration in the cell because the influent contained dissolved by the formation of the biofilm. The cathode potentials were not
oxygen (∼8 mg L−1 at 30 ◦ C). The substrate concentration in significantly affected by the biofilm over the current density
this range should not affect the maximum power generation. range examined (0–1.0 mA cm−2 ), although the potentials for
The decrease in power generation at a longer retention time of cathode with a biofilm were slightly lowered than those without
16 h was possibly due to the decrease of substrate concentration a biofilm (Fig. 6). Possible reasons for the decreased potential
and a reduction in cell metabolism since the acetate removal

Fig. 4. Power generation in SMFC with anode parallel or perpendicular to Fig. 6. Cathode (with and without biofilm) potentials at various current densities
cathode. measured in electrochemical cell.
278 H. Liu et al. / Journal of Power Sources 179 (2008) 274–279

in the presence of the biofilm include decreased proton mass only resulted in a slight increase in power density from 14
transfer to the cathode surface, reduced availability of oxygen to 16 W m−3 . The power density increased by only an addi-
at the cathode surface due to oxygen utilization by aerobic bac- tional 6% when the specific anode surface area was increased
teria, and possible deactivation of the Pt catalyst by bacteria or to 870 m2 m−3 by adding graphite granules, indicating that the
chemicals produced by the bacteria. anode surface area did not limit the performance of the LMFC
at high anode/cathode surface area ratios. On the other hand,
4. Discussion increasing the specific cathode surface area can greatly improve
the MFC performance [31]. Future research is needed to reduce
The scale-up of MFCs can lead to changes in volumetric cathode limitations on power output through enlarging the cath-
power density depending on what factors are kept constant or ode surface area, such as using tubular systems [32] and/or
altered as reactor size is changed [30]. In experiments by others, developing new catalysts and cathode structures.
the power density (10 W m−3 ) of a small MFC (0.025 cm3 ) was
much higher than that (0.6 W m−3 ) of a very similar but larger 4.3. Ionic strength
MFC (5 cm3 ) [30]. The changes in power density during scale-up
result from changes in many important factors, such as elec- The power was increased by 25% when the IS was increased
trode spacing and electrode specific surface area (surface area from 100 to 300 mM. This increase is not as large as that found
per volume). Power density can be maintained or even increased in our previous tests (66%) possibly due to the different architec-
during scale-up if these factors, especially electrode spacing, are tures of the two MFCs and the fact that the LMFC had a smaller
considered and optimized during scale-up. As shown here, the average electrode spacing and lower area-specific internal resis-
power density increased by ∼15% in spite of a ∼20× increase tance. We previously found that there was little improvement
in volume. in power generation when the IS was increased from 100 to
400 mM if the internal resistance was reduced by decreasing the
4.1. Electrode spacing electrode spacing from 4 to 2 cm [24]. The reduced electrode
spacing from 4 to 2.6 cm in this study decreased the electrolyte
Electrode spacing and orientation was found here to be the internal resistance, lessening the importance of IS to the total
key factor that affected the area-specific internal resistance and internal resistance.
power density. Therefore, to maintain the power density during
scale-up, the larger reactor architecture must maintain or even 4.4. Other factors limiting the LMFC performance
reduce the electrode spacing. It was previously shown that power
density was increased when the anode and cathode spacing was Compared to power densities generated in other membrane-
decreased from 4 to 2 cm [24]. In our study here, the reduction free air-cathode MFCs, the maximum power generated in this
in average electrode spacing (2.6 cm in the LMFC vs. 4 cm in LMFC is 22 W m−3 , which is 69% higher than the 13 W m−3
the SMFC) accounted for most of the 15% increase in power previously obtained using an SMFC containing equally sized
density. The importance of electrode spacing on power density carbon paper anode and cathodes with an electrode spacing of
was further evidenced by altering the electrode orientations in 4 cm [12]. However, it is lower than the 30 W m−3 obtained
the SMFC. A 36% increase in power density was obtained when using an SMFC (14 mL) with an electrode spacing of 2 cm [12],
the electrode orientation was altered from parallel to perpendic- and 51 W m−3 produced by advective flow through the anode
ular, which was mainly due to the reduced electrode spacing of an SMFC (14 mL) with a 1 cm electrode spacing [24]. There
(4 cm vs. 2.5 cm). While further decreases in electrode spacing are several factors that contribute to the lower power genera-
can result in even smaller internal resistance, too close an elec- tion in the larger system tested here than in these other studies.
trode spacing can limit performance. Cheng et al. [24] reported First, the increased electrode surface area in the larger MFC
a 50% decrease in surface power density in air-cathode MFC increased the distance electrons travel from the point of gener-
when the electrode spacing was reduced from 2 to 1 cm due to ation on the anode to the location that they exit to the external
oxygen diffusion into the anode chamber. However, volumetric circuit (i.e. where the wire connects to the anode). This distance
power densities can be maintained or even increased by adding a is not a significant factor in the performance of an SMFC with
cloth separator and reducing the spacing between the electrodes a surface area of 7 cm2 since the electrode resistance is negli-
[15]. Reduced electrode spacing and the use of cloth separators gible in comparison to the external resistance (100–200 ) for
need to be explored further for their relative advantages in larger maximum power. However, in the LMFC, the maximum power
MFCs. output occurred at an external resistance smaller than 10 , and
therefore a small increase in electrode resistance could greatly
4.2. Specific electrode surface area increase power loss. Second, contact resistance between the cur-
rent collecting points (electrode strips) and clips (connected to
Specific electrode surface area can also be an important fac- the external circuit) may have become significant due to the
tor in reactor performance. It has been found in one study that increased current output in the LMFC. Considering the low inter-
increasing the anode surface area relative to the cathode can nal resistance (<10 ) of the LMFC, the voltage drop caused
increase power [26]. Increasing the anode surface area here by the contact resistance cannot be neglected. Third, while the
from 25 m2 m−3 (SMFC) to 150 m2 m−3 (LMFC), however, cover on the cathode had holes drilled to allow oxygen trans-
 H. Liu et al. / Journal of Power Sources 179 (2008) 274–279 279

fer to the electrode, some of the surface area was obscured by [4] D.R. Bond, D.R. Lovley, Appl. Environ. Microbiol. 69 (2003) 1548–
the cover plate, perhaps limiting cathode performance. It will be 1555.
necessary to continue to address these design factors in future [5] C.E. Reimers, L.M. Tender, S. Fertig, W. Wang, Environ. Sci. Technol. 35
(2001) 192–195.
scale-up designs. For example, a collector plate could be used [6] K. Rabaey, N. Boon, S.D. Siciliano, M. Verhaege, W. Verstraete, Appl.
to reduce the electron travel distance and contact resistance. Environ. Microbiol. 70 (2004) 5373–5382.
[7] G. Reguera, K.D. McCarthy, T. Mehta, J.S. Nicoll, M.T. Tuominen, D.R.
5. Conclusions Lovley, Nature 435 (2005) 1089–1101.
[8] N. Ryckelynck, H.A. Stecher III, C.E. Reimers, Biogeochemistry 76 (2005)
113–139.
The performance of a SMFC (28 mL) was compared with a [9] S. Cheng, H. Liu, B.E. Logan, Electrochem. Commun. 8 (2006) 489–494.
newly designed LMFC (520 mL) in fed-batch mode. The power [10] D.H. Park, J.G. Zeikus, Appl. Microbiol. Biotechnol. 59 (2002) 58–61.
density increased by about 15% from 14 W m−3 (SMFC) to [11] J. Niessen, U. Schroder, M. Rosenbaum, F. Scholz, Electrochem. Commun.
16 W m−3 (LMFC) in spite of the increase in volume by a 6 (2004) 571–575.
factor of nearly 20, mainly as a result of the lower average elec- [12] H. Liu, S. Cheng, B.E. Logan, Environ. Sci. Technol. 39 (2005) 5488–5493.
[13] G.C. Gil, I.S. Chang, B.H. Kim, M. Kim, J.K. Jang, H.S. Park, H.J. Kim,
trode spacing and higher anode surface area per volume of the Biosens. Bioelectron. 18 (2003) 327–334.
LMFC. The effect of the larger anode surface area on power was [14] H. Liu, B.E. Logan, Environ. Sci. Technol. 38 (2004) 4040–4046.
shown to be relatively insignificant by adding graphite granules [15] Y. Fan, H. Hu, H. Liu, J. Power Sources 171 (2007) 348–354.
or graphite fiber brushes to the LMFC anode chamber. Thus, it [16] Z. He, N. Wagner, S.D. Minteer, L.T. Angenent, Environ. Sci. Technol. 40
was concluded that the cathode was the electrode limiting power (2006) 5212–5217.
[17] K. Rabaey, P. Clauwaert, P. Aelterman, W. Verstraete, Environ. Sci. Tech-
generation as a result of its significantly smaller surface area. nol. 39 (2005) 8077–8082.
Increasing the ionic strength of the LMFC from 100 to 300 mM [18] Z. Li, L. Yao, L. Kong, H. Liu, Bioresour. Technol. 99 (2008) 1650–1655.
increased the power density to 20 W m−3 . When the LMFC [19] S. Cheng, B.E. Logan, Electrochem. Commun. 9 (2007) 492–496.
was operated at continuous mode, a maximum power density of [20] Y. Fan, H. Hu, H. Liu, Environ. Sci. Technol. 41 (2007) 8154–8158.
22 W m−3 was generated at an HRT of 11.3 h. Although a thick [21] S.-E. Oh, B.E. Logan, J. Power Sources 167 (2007) 11–17.
[22] P. Aelterman, K. Rabaey, T.H. Pham, N. Boon, W. Verstraete, Environ. Sci.
biofilm was developed on the cathode surface in this reactor, Technol. 40 (2006) 3388–3394.
the cathode potentials were not significantly affected at current [23] S.H. Shin, Y.J. Choi, S.H. Na, et al., Bull. Kor. Chem. Soc. 27 (2006)
densities <1.0 mA cm−2 . These results demonstrate that power 281–285.
output can be maintained or even increased during reactor scale- [24] S. Cheng, H. Liu, B.E. Logan, Environ. Sci. Technol. 40 (2006) 2426–2432.
up if these factors, especially electrode spacing, are considered [25] B.E. Logan, Environmental Transport Processes, John Wiley & Sons, Inc.,
New York, 1999.
and optimized during scale-up. [26] B.E. Logan, S. Cheng, V. Watson, G. Estadt, Environ. Sci. Technol. 41
(2007) 3341–3346.
References [27] H. Zhang, M.A. Bruns, B.E. Logan, Water Res. 40 (2006) 728–734.
[28] B. Min, P.J. Evans, A.K. Chu, B.E. Logan, Water Res. 38 (2004) 47–60.
[1] B.E. Logan, B. Hamelers, R. Rozendal, U. Schroder, J. Keller, S. Freguia, [29] J. Larminie, A. Dicks, Fuel Cell Systems Explained, 2nd edition, John
P. Aelterman, W. Verstraete, K. Rabaey, Environ. Sci. Technol. 40 (2006) Wiley & Sons, 2003.
5181–5192. [30] J.C. Biffinger, R. Ray, B. Little, B.R. Ringeisen, Environ. Sci. Technol. 41
[2] H. Liu, R. Ramnarayanan, B.E. Logan, Environ. Sci. Technol. 38 (2004) (2007) 1444–1449.
2281–2285. [31] S.-E. Oh, B.E. Logan, Appl. Microbiol. Biotechnol. 70 (2006) 162–169.
[3] K. Rabaey, G. Lissens, S.D. Siciliano, W. Verstraete, Biotechnol. Lett. 25 [32] Y. Zuo, S. Cheng, D. Call, B.E. Logan, Environ. Sci. Technol. 41 (2007)
(2003) 1531–1535. 3347–3353.