INTERNATIONAL JOURNAL OF ENERGY RESEARCH

Int. J. Energy Res. 2015; 39:543–556

Published online 11 December 2014 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/er.3270

Energy harvest with mangrove benthic microbial fuel cells

Paule Salvin1, Olivier Ondel2, Christophe Roos1 and Florent Robert1,*,†
1
Laboratoire des Matériaux et des Molécules en Milieu Amazonien, Université des Antilles et de la Guyane, UAG-UMR ECOFOG,
F-97337 Cayenne, French Guiana
2
Université de Lyon, F-69622, Lyon, France, Université Lyon 1, Villeurbanne; CNRS, UMR 5005, Laboratoire Ampère, 43, Boulevard du
11 Novembre 1918, Villeurbanne, 69622, France

SUMMARY
Benthic microbial fuel cells (BMFCs) are continuous electricity generators using electroactive microorganisms and organic
matter from aquatic environment, respectively, as catalysts and substrate. In this paper, first a low-cost PVC-made structure
is constructed to harvest electricity from mangrove environment located in French Guiana. An in situ BMFC has given power
density of 30 mW/m2 of the anodic surface area. This performance has been confirmed by experience in laboratory where
inter-electrode distance and electrode surface area appeared to be power increasing factors. However, the output power of
one BMFC is not used to supply real devices such as autonomous sensors. Second, to meet this expectation, in parallel
and in series associations were considered. These associations were made in order to increase the output voltage and conse-
quently the power, to reach levels that can supply small sensors (about 3 V). Finally, to improve the performance of the series
association and to avoid the voltage reversal phenomenon, a voltage balancing circuit was simulated and added to the series
connections. With balancing method, the cell voltage of BMFCs can be equalized, and the performances can be improved.
This allows an optimal energy harvesting and a better global efficiency of the set. Copyright © 2014 John Wiley & Sons, Ltd.

KEY WORDS
balancing circuit; benthic microbial fuel cell; efficiency; energy harvesting

Correspondence
*Florent Robert, Laboratoire des Matériaux et des Molécules en Milieu Amazonien, Université des Antilles et de la Guyane, UAG-UMR
ECOFOG, F-97337 Cayenne, French Guiana.
†
E-mail: florent.robert@guyane.univ-ag.fr

Received 9 July 2014; Revised 29 September 2014; Accepted 20 October 2014

1. INTRODUCTION the overlaying water. BMFC is also operated in laboratory.

For a few years now, microbial fuel cell (MFC) has be-
come of great interest as a future alternative in energy pro-
duction. It is based on the catalysis of the electrochemical
reactions realized by electroactive biofilms (EAB), adhered
on electrodes [1–3]. The anodic EAB catalyzes the oxida-
tion of the organic matter or the carbon source present in
the anodic chamber and transfers the generated electrons
to the conductive anode. Then, the difference of potentials
between the anode and the cathode induces that the electrons
pass through an external circuit composed of generally resis-
tive devices and go to the cathode where the reduction of the
dioxygen or another oxidizer takes place [3].
One MFC design seems to be closer to a marketable pro-
totype: benthic microbial fuel cell (BMFC) also called sedi-
ment microbial fuel cell (SMFC) [4]. Its name comes from
benthic area where this kind of MFC is usually installed,
which is at the interface between sediments and water in
aquatic environments (ocean, river, lake…) [5–8]. Anodes
are embedded in sediments whereas cathodes are placed in
There, compartments are designed to contain sediments and
water collected from natural sites [9,10]. The main advantage
of BMFC is its low cost of designing and functioning. Indeed:
i. Cheap materials, such as graphite, could be used to
build the electrodes, bacteria playing the role of
the catalyst [5];
ii. Organic matter and inoculum are naturally present
in the chosen environment;
iii. No membrane is needed because the sedimentary
bacteria greatly consume oxygen, limiting its pres-
ence in the anodic part [11].
The mechanisms of current generation are multiple and
complex, associating abiotic and microbial reactions, but
also direct and indirect electronic transfers [12]. The average
power obtained from BMFC is 10–20 mW per anodic sur-
face square meter, with maximal values of 109 mW/m2
[13] and current density of 104 mA/m2 [14]. The perfor-
mances are lesser than those announced for air-cathode

Copyright © 2014 John Wiley & Sons, Ltd. 543

P. Salvin et al.
MFC (6.98 W/m2 for example, [15]) but the low cost, the
durability of the performance and the unlimited availability
of the source of substrate are solid arguments to consider
those results. Several explanations have already been ex-
posed concerning the limitations of the BMFC such as slow
kinetics of the anodic and cathodic electrochemical reactions
and slow transfers of the reactants to both electrodes.
In the published studies on MFCs, increasing power was
already performed thanks to parameter modifications such as
decreasing electrode spacing [9,16–18], increasing electrode
surface [19–22] or reducing the membrane resistance [23].
Another improvement of the performance is brought by asso-
ciations of numerous MFC units (e.g. stack operation or in
parallel association) [24–27]. In the case of BMFC or SMFC,
these combinations were not explored yet because of possi-
ble limitations: first, connecting several cells in series could
not be possible in open systems such as aquatic environ-
ments [19,28] and second, overall potential could not be in-
creased in this configuration. To the best of our knowledge,
one research team has published about a self-stacked sub-
mersible microbial fuel cell installed in lake sediments
[29]. They conceived an air-space for the cathodes connected
by external tubes. Thus, cathodes were not in the same elec-
trolyte, and the voltages of the MFC could be cumulated.
In a recent published study we have lightened the pres-
ence of EAB in tropical mangrove sediments using electro-
chemical cell as an electro-activity detector [30]. In this last
case, acetate was provided as an electron source in the
Figure 1. A) Schematic representation of the in-situ BMFC placed in the mangrove. B) On the pontoon, the cell voltage is measured
thanks to a portative multimeter. C) Cubic glass-compartment containing the four laboratory-BMFCs. D) Laboratory BMFCs are con-
nected to electronic resistances placed in a plastic box. Pushing the red buttons opens or closes the electrical circuits.
544
Microbial fuel cell
medium to form bioanodes. In the present study, BMFCs
have been operated to harvest electricity from the same trop-
ical mangrove. With this design, no acetate or other electron
donors were needed to provide electricity as electron donors
were naturally present in the sediments. In-situ- and
laboratory-located BMFCs have been tested, with consider-
ations on electrode spacing and electrode surface. Associa-
tions (in parallel or in series) of BMFC units have been
considered as an adequate way to enhance power genera-
tion. In this case, an electrical model was defined for each
BMFC in order to simulate the behavior of the various
associations. An electrical circuit has been added in order
to avoid the different constraints (voltage reversal, non-
optimal efficiency) during the in series associations.
Finally, the results obtained with in parallel and in series
associations (with and without adding electrical circuit) are
discussed.
2. MATERIALS AND METHODS
2.1. Benthic microbial fuel cells setup
2.1.1. General descriptions of the BMFCs
BMFCs used in this study are based on a PVC-cylindrical-
tube-made structure in form of an inverse (head at the bottom)
letter F (Figure 1-A). Inside the structure, electrical junctions
between titanium current collectors (titanium wires, 1.5-mm
Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd.
DOI: 10.1002/er
Microbial fuel cell
diameter, Good Fellow) and copper cables are protected
from water. Outside of it, conductive electrodes are in con-
tact with water and sediments at the bottom part. Indeed, the
two parallel axes serve as inserts for the current collectors
of both anode and cathode in order to maintain them in par-
allel with anode in sediment and cathode in the overlaying
water. Plastic gained copper cables immerged from water
to air at the upper part (above the main vertical axis). The
latter link anode and cathode to a resistor (Rext) which
closes the electrical circuit. This output load is arbitrarily
set to 1 kΩ. Anode and cathode are both made with square
pieces of carbon cloth (carbon cloth, PaxiTech SA, France)
with a surface ratio of 1:1.
2.1.2. In-situ BMFC
In this case, the BMFC is placed in the benthic zone of
an estuary mangrove (Figure 1-A). It is attached to a dock
pontoon at 2 m below the water level at high tide. The re-
sistor is maintained on the dock, and it is never immersed
in water (Figure 1-B). Over time, the cathode is always
in water (at high and low tides). The projected surface of
both electrodes, anode and cathode, is 72.25 cm2, and the
distance between them is 12 cm.
2.1.3. Laboratory BMFCs
Sediments and water sampled the same place areand
placed in a 70-L cubic compartment made in glass
(Figure 1-C). Four BMFCs are fixed with metallic supports
and tongs which are outside and upper the compartment in
order to maintain the electrodes at the right level (anodes in
sediments and cathodes in water). The resistors are placed
in an electrical box where they are associated with electrical
switches, providing two operating modes: open circuit
(switch off equal to an infinite resistance) and closed circuit
(switch on—connected to a variable resistance) (Figure 1-D).
The BMFCs are named L1, L2, L3 and L4: L1 and L2 are rep-
licates with the same anode projected area (APA) 72.25 cm2
and inter-electrode distance (IED) 12 cm. L3 has a lesser
APA of 56.25 cm2 but the same IED as L1. L4 has the same
APA as L1 but a lesser IED (5 cm).
2.2. Mangrove water and sediments
The mangrove located in the Larivot port area (Matoury,
French Guiana) was a four-parameter sensor (HI9811-5,
Hanna Instruments) and the conductivity with inoculum
used in this study (water and sediments). The pH was mea-
sured in water samples with a conductivity meter (EC215,
Hanna Instruments) during both in-situ and laboratory ex-
periments. The pH varied between 7.0 and 7.6, and the
conductivity varied between 3.9 mS/cm and 33.4 mS/cm
following the tides and the moment of the day.
2.3. Electrochemical measurements
At the laboratory, the cell voltage (V) is continuously mea-
sured using a potentiostat as a voltmeter and recorded with
a data acquisition system (potentiostat VMP3—software
Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd.
DOI: 10.1002/er
P. Salvin et al.
Ec-Lab v.10.12, Biologic SA). In the in-situ mangrove, a
portable multimeter is used, and the voltage is measured
once a day. The current (I) is calculated according to
Ohm’s law (I = V / Rext) and then normalized by the
projected surface area of the anode to obtain the current
density (J). Each BMFC is characterized by a polarization
curve (V–I) obtained by the ‘Rext-variation’ procedure de-
scribed by Logan et al. [3]. A variable resistor box is used
to set external loads from 10 000 kΩ to 10 Ω (open-circuit
to short-circuit). The voltage for each load is taken after
pseudo-steady-state conditions have been established.
The time needed varies between 5 and 30 min according
to the difference between two consecutive values of Rext.
Power curve (P–I) is calculated from the polarization curve
and is a representation of power (P) versus current (I).
Evaluation of the internal resistance (Rint) of the cell is
done by calculating the slope of the linear region of the
polarization curve or calculating the resistance occurring
at maximal power point: Rint ~ Pmax / I2.
3. RESULTS AND DISCUSSION
3.1. Electrochemical results of the
individual BMFCs
3.1.1. BMFCs harvest electricity from natural
mangrove without substrate addition
3.1.1.1. In-situ BMFC. The first aim of this study is
to demonstrate that mangrove could be an adequate envi-
ronment to harvest electricity with a BMFC. Previous
experiments showed that this environment constitutes an
efficient inoculum to form bioanodes in three-electrode
cells thanks to its salinity and its microbial community
[30]. BMFC appears to be a good way to test the ability
to generate electricity from the mangrove microbial activ-
ity in real conditions. Our BMFC prototype stayed in place
only 10 days in the mangrove because of the difficulties
inherent to the chosen place (regular tides, presence of
branches in water and strong estuary currents). During this
period, the voltage versus time curve (V–t curve) shows the
establishment of a difference of potential between the
anode and the cathode when the BMFC is connected to a
1-kΩ resistor (Figure 2-A). Indeed, after 10 days of immer-
sion, the voltage reaches 453 mV. For logistical reasons,
the polarization curve and the power curve are realized af-
ter 7 days of immersion of the BMFC. The maximal power
point (MPP) at this date (noted Pmax7) is 30 mW/m2 in water
with pH 7.0 and conductivity of 17 mS/cm (Figure 2-B). The
voltage VMPP7 and current IMPP7 at MPP are, respectively,
300 mV and 0.58 mA (current density JMPP = 110 mA/m2).
The internal resistance of the BMFC estimated from the
polarization curve is 500 Ω. These first results indicate that
harvesting electricity from a natural environment such as
mangrove could be possible using a BMFC. The power
curve, shown in Figure 2-B, has not been obtained at the
maximal performance of the in-situ BMFC. Indeed, this
value is obtained after 7 days of immersion. However, the
545
P. Salvin et al. Microbial fuel cell

A Knowing the evolution law and knowing V1kΩ to 10 days

500
(parameter b), V500Ω can be determined at 10 days (noted
Cell voltage in mV
400 VMPP10) and Pmax10 inferred. Calculations are summarized
in Table I and give an estimated maximum power
300
Pmax10 = 54 mW/m2 after 10 days. Depending on the evolu-
200 tion of the voltage after 10 days, this power can be further
increased.
100 This performance could be very interesting if it was du-
rable in time because there are few studies which present a
0
0 2 4 6 8 10 performance equivalent in our conditions of work. Indeed,
Time in days in general, higher power densities (>30 mW/m2) are ob-
tained when improvements are made to get better electro-
B chemical reactions kinetics in the BMFC. For example,
500 40 the electrode material could be optimized adding catalysts
on cathode or increasing the anode surface [31–33]. Other
400
Cell voltage in mV

30 way, the concentration of reactants is increased near the

300
200
100
0 0
0 25 50 75 100
Current density in mA/m2
Figure 2. In-situ BMFC characterization—A) Cell voltage in func-
tion of time. B) Polarization (white diamonds) and power density
(grey diamonds) curves after 7 days of immersion.
establishment of cell voltage V1kΩ (Figure 2-A) shows that it
continues to increase almost linearly until the 10th day. So,
the maximal power density data of that BMFC would cer-
tainly be higher than 30 mW/m2. To have an approximate
idea of the sustainable power after 10 days, from the polari-
zation curve obtained, we assume a linear evolution of the
voltage versus time starting at day 7. The maximal power
is calculated with Pmax = VMPP2/Rint (VMPP is obtained when
Rext = Rint). Rint (equal to 500 Ω) was calculated from the po-
larization curve. It remains to determine the cell voltage after
10 days with Rext = 500 Ω (noted VMPP10). For Rext = 500 Ω
and 1 kΩ, it is within the linear region of the polarization
curve (Figure 2-B). We can define the evolution law in this
area from this equation: V = a.ΔRext + b, with a the slope
and b the value of V for Rext = 1 kΩ (noted V1kΩ).
As the voltage evolution between 7th and 10th days
(Figure 2-A) is linear, we hypothesize that the polarization
curve after 10 days will have the same evolution in its
linear region.
P in mW/m2
electrode as in Rezaei et al. experiments, with the addition
20 of chitin at the anode [9]. Reimers et al. recommended the
installation of their BMFC in an environment naturally
richer in organic substrate [34]. Nielsen et al. proposed a
10
new architecture for the BMFC with a benthic chamber
which concentrates the organic substrate around the anode
125 150
[35]. Recently, Song and Jiang showed that a pre-treatment
of the sediments by heat could increase the generation of
current [36].
3.1.1.2. Laboratory BMFCs. After the test in real
conditions, the same experience is reproduced at the labo-
ratory, so that the problems linked to the in-situ BMFC are
avoided (tides, high currents, limited access…). Four
BMFCs are placed in a compartment of 70 L and are
named L1, L2, L3 and L4. L1 and L2 are similar with same
dimensions as the in-situ one. L3 has a smaller electrode
surface and L4 a shorter inter-electrode distance.
During 30 days, the voltage of each BMFC (V1kΩ) in-
creases until it reaches a stable value (data not shown).
This evolution could be attributed to the formation of the
biofilm at the surface of the electrode. Indeed, this link
between the growth of an electroactive biofilm and the
increase of the performance has already been shown [30].
No supplementary organic substrate has been added in
the compartment during the experiment. After 40 days of
operation, the cell voltages continue to be stable while no
supplementary substrate has been ever provided to the
system. So, these results demonstrate that the duration of
an in-situ BMFC could be more important if the system
could be resistant enough to the environmental and techni-
cal difficulties.

Table I. Maximal power determination after 10 days (Pmax10) by linear extrapolation.

Evolution law V1kΩ V500Ω(VMPP) Pmax(500Ω)

th 2
7 days a =
0.13 V = a. Rext + b 355 mV 290 mV 30 mW/m
b = 355
th 2
10 days a =
0.13 V = a. Rext + b 455 mV 390 mV 54 mW/m
b = 455

546 Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd.
DOI: 10.1002/er
Microbial fuel cell P. Salvin et al.

800
3.1.2. Electrode surface area and inter-electrode
distance enhance the power output

Cell voltage in mV
Laboratory BMFCs L1 and L2 are duplicates of the 600
in-situ BMFC (same anode surface area and same inter-
electrode distance). The analysis of their V–t curve 400
shows near close cell voltages’ values once the plate
value was achieved. At day 40, the cell voltages are
373 mV for L2 and 404 mV for L1. These voltages are 200
lower than those obtained with the in-situ BMFC despite
the same parameters of electrode surface and inter- 0
electrode distance. This lower performance could be 0.00 0.25 0.50 0.75 1.00 1.25
explained by the constant renewal of the media and the Current intensity in mA
reactants around the in-situ BMFC.
0.25
L3, with smaller electrodes’ surface area, presents the
lowest performance, i.e. 221 mV. This result indicates that
0.20
increasing the surface available for the microorganisms

could enhance the efficiency of the BMFC. L4, with a
reduced inter-electrode distance, reaches the highest maxi-
mum cell voltage, i.e. 471 mV. Reducing the gap between
the anode and the cathode seems to increase the cell volt-
ages, as detailed in introduction. This could be explained
by the decrease in the internal resistance of the BMFCs
through ion exchange facilities when the distance is shorter
between the electrodes. In the case of the in-situ BMFC,
ion exchange and redox reactions are certainly promoted
by the constant renewal of the environment brought by
tides phenomena (among others). In the case of laboratory
BMFCs, the lack of stirring of the medium, in particularly
in the anodic part, could slow the reagent diffusion. More-
over, L4 has the closest performance to the 10 days in situ
BMFC’s one. However, this approach is not suitable for
the in-situ device due to the difficulty to control the sedi-
ment level at the interface sediment-water over time (peri-
odical movement of the water).
The polarization curves and power curves of the labora-
tory BMFCs, made in their stable phase, are shown in
Figure 3. All BMFCs have an open circuit voltage VOC
which tends to range from 650 to 750 mV whereas the
short-circuit current varies between 0.2 mA (L3) and
1.2 mA (L4). This amplitude corresponds to that usually
found in the literature on BMFC [31]. Calculations of their
mean V, I and P values associated with standard deviation
have been realized for L1 and L2 BMFCs. They show that
their comportment is quite similar in comparison with
BMFCs L3 and L4 (Figure 3) as remarked with the V–t
curve. The similarity is maximal in the linear region
between 485 and 300 mV of the polarization curve. The
maximum power point MPP is the highest with 0.23 mW
at 0.61 mA for L4 and the lowest for L3 with 0.08 mW at
0.20 mA (Figure 3 and Table II). Duplicates L1 and L2
have a mean MPP of 0.154 ± 5.10
3 mW at 0.438
± 4.10
3 mA. The voltage corresponding to MPP (noted
VMPP) for the duplicates L1 and L2 is 350 mV ± 6 mV.
VMPP for L3 and L4 are 400 mV and 370 mV (Figure 3
and Table II). Internal resistance Rint calculated from
the slope in the linear region of the polarization curve,
respectively, for duplicates, L3 and L4 are 686 Ω, 1730 Ω
and 510 Ω.
Power in mW
0.15
0.10
0.05
0.00
0.00 0.25 0.50 0.75 1.00 1.25
Current intensity in mA
Figure 3. Polarization curves (upper) and power curves (bottom) of
laboratory BMFCs. Mean values have been used for the duplicates
L1 and L2 and red error bars are indicated on their curve. Circle
symbols are for L3 BMFC and cross symbols are for L4 BMFC.
3.2. Increasing power via in parallel or/and
in series combinations of BMFCs
In regards of the application of BMFCs to supply real ap-
plication as wireless sensor for example, the power levels
described for the individuals laboratory BMFCs (between
300 and 400 mV for 1 kΩ) are not sufficient to supply an
electronic device. In order to improve the voltage or cur-
rent characteristics, the laboratory BMFCs are connected
either in parallel or in series. The differences between the
tests realized in situ and at the laboratory lead to test
BMFCs under different conditions. From now on, we only
consider the results of the identically built BMFCs (i.e.
BMFC in situ, L1, L2). It was found that the internal
equivalent resistance (Rint) increases when fuel concentra-
tion decreases. In this case, the voltage at maximum power
point VMPP is identical for reactors, while the MPP current
(IMPP) is different (Figure 4). For this reason, parallel asso-
ciation was preferred as a first step, in order to increase and
optimize the harvesting energy.
In fact, associating BMFCs in parallel allows to impose a
common voltage but with a different current through them.
As we have seen (Figure 4) the cell voltage at the MPP
(VMPP) being almost identical for each BMFC. If this voltage
is imposed on each cell (in our case VMPP = 300 mV), then
they provide their maximum power. Thus the parallel associ-
ation will present the best energy efficiency.

Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd. 547
DOI: 10.1002/er
P. Salvin et al. Microbial fuel cell

Table II. Electrochemical characteristics corresponding to the maximal power Pmax of individual and in parallel associated laboratory
BMFCs.

BMFC names Theoretical

(individual or association L1–L2 in parallel Theoretical association L1–L2–L3–L4 in
associated) L1 L2 L3 L4 L1//L2 (experiment) L1//L2//L3//L4 parallel (experiment)

Voc in mV 659 710 663 647 – 664 – 653

Pmax in mW 0.15 0.16 0.08 0.23 0.31 0.28 0.62 0.56
VMPP in mV 315 310 400 370 310 290 310 330
IMPP in mA 0.46 0.52 0.2 0.61 0.98 0.96 1.79 1.67
Rint in Ω 700 680 1730 510 340 300 182 200

Thus all BMFCs are connected to a variable resistance Rext

(Figure 5-A). When the voltage is stable, Rext is varied to
characterize the BMFCs’ association.
The parallel combination of L1 and L2 involves two
BMFCs that have similar characteristics. The association
has quasi doubled Pmax (0.28 mW) in comparison with a
single BMFC (0.15 mW) (Figure 5-B & Table II). As
attempted, at the MPP, the cell voltages of the association
(VMPP) stay equivalent to the single BMFC’s one, near
300 mV (Table II). The internal resistance (Rint) of the

Figure 4. A) Theoretical dispersion of polarization curves in par-

allel associations. B) Polarization and power curves for identical
BMFCs in situ, L1 and L2 of the study. C) Power curves for iden-
tical BMFCs in situ, L1 and L2 of the study.

3.2.1. In parallel association

When electrical generators are connected in parallel, the
currents supplied by them are cumulative. Thus, current in-
tensity I recorded in the circuit is the sum of all currents
flowing through each cell. The power is then greater for a
given voltage. In addition, when a generator is down, the
others could still work (unlike the in series connection).
However, the disadvantage of this arrangement is that the Figure 5. A) Electrical chart of the in parallel association of
lowest generator voltage is imposed to all others. In our BMFCs L1, L2, L3 and L4. B) Power curves of in parallel BMFC
study, L1 and L2 are first connected in parallel, and their associations: □ to the L1–L2 association and ◊ to the L1–L2–L3–
association is characterized. Then, L3 and L4 are also con- L4 association, compared with the individuals BMFCs L1 (Δ), L3
nected in the circuit, in derivation with the first association. (x) and L4 (o).

548 Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd.
DOI: 10.1002/er
Microbial fuel cell P. Salvin et al.

associated BMFCs, got from the linear region slope of the
polarization curve, is 300 Ω. It almost corresponds to the
value of the equivalent resistance of the two BMFCs in
parallel:
RL1;int RL2;int 700680
RL1-L2;int ¼ ¼ ¼ 340Ω
RL1;int þ RL2;int 700 þ 680
By adding L3 and L4 in parallel to the previous associ-
ation (finally L1//L2//L3//L4), the maximal power point
(MPP) obtained is 0.55 mW for IMPP equal to 1.67 mA.
The theoretical association of these 4 BMFCs in parallel
gives a maximal power of 0.62 mW for a current of
1.79 mA (Table II).
The maximum efficiency for this configuration at MPP
is defined in Equation 1 where POUT is the output power
for the associated BMFCs and ΣPMPP is the sum of MPP
for each BMFC. It is about 88.7% at MPP. In this case,
the BMFCs work at 99%, 98%, 87% and 91% of MPP,
respectively.
 the maximum potential power available.
η% ¼ POUT ∑ PMPP
100
We note that reactors L1 and L2 operate at maximum
efficiency which is not the case for both others reactors
L3 and L4. This is normal since the voltage imposed on
all BMFCs is around 300 mV. That corresponds to VMPP
(and thus Pmax) for L1 and L2 while this is not the case
for L3 and L4 (VMPP being around 385 mV). The last two
do not operate at MPP. The parallel association has
allowed increasing the final power output (0.55 mW when
the best individual performance is 0.23 mW).
With four equivalent BMFCs with the best performance
(L4), the maximal power output could be 0.92 mW, a value
close to 1 mW. So, several mW could be obtained if
the surface area of the electrode was increased. In addition
to this, this experiment is technically realizable in situ.
However, the inconvenient is still the low cell voltage
which is not sufficient to supply the high majority of
electronic materials. Researchers have conceived ‘power
management system’ and ‘voltage conditioner’ which are
devices capable of amplifying the output cell voltage of
BMFC [11,28,37].
3.2.2. In series association
Another way to increase the cell voltage could be to as-
sociate BMFCs in series. In series-connected MFCs, the
individual voltages of each cell add up at the output while
a common current flows through fuel cells. The serial asso-
ciation is considered not only in terms of increased power
production but also to increase voltage to an acceptable
level for real applications.
Nevertheless, the drawbacks of this type of association
are:
i. Internal resistance generators are added to each
other and can reduce the current gradually as the
number of cells increases;
ii. If one generator is out of service, so is the entire
circuit;
iii. The generator with the lowest intensity imposes it to
all others. The power loss could be important
because some of the generators will not be used to
their maximum power point.
The latter indicate that this association cannot provide
(1)
In particular, in the case of BMFC, an application in situ
is not possible because the cathodes are in the same elec-
trolyte, without separation [28]. So, in this study, only
the laboratory BMFCs are considered for the in series asso-
ciation. The individual BMFCs L1 and L2 were stacked
after separating the two cells by transferring to another
compartment the L1 BMFC system. Thus, the possible
ion exchange inter-cell was avoided. L1 was then placed
alone in a 5-L bucket. There was a significant decrease in
its voltage value. Indeed, the cell voltage V1kΩ was
260 mV instead of 400 mV before the transfer. The transfer
and the volume change have resulted in the loss of perfor-
mance of L1. So for a better comprehension, L1 was then
called L1*, and a new characterization was made by a po-
larization curve and power curve after the transfer in its
new container.
L1* and L2 cells have no similar characteristics (Table III).
The performance of the in series combination will be limited
by the less efficient BMFC, L1*. VOC, the open circuit
voltage, has increased significantly since it corresponds to
the sum of the two elementary open circuit cell voltages,
approximately 1470 mV. The voltage increase is significant

Table III. Electrochemical characteristics corresponding to the maximal power Pmax of individual and series associated laboratory
BMFCs.

BMFC names L1*–L2 in series L1*–L2 in series L1*–L2 in series without L1*–L2 in series with
(individual or associated) L1* L2 (theoretical) (experiment) SC (simulation) SC (simulation)

Pmax in mW 0.07 0.16 0.23 0.20 0.196 0.226

Voc in mV 767 710 1477 1420 1445 1437
VMPP in mV 370 310 680 625 683 482
IMPP in mA 0.19 0.52 – 0.31 0.34 0.48
Rint in Ω 1620 680 2300 2120 2000 2000
Efficiency (%) – – 100 86 86 98.6
L1* is the L1 BMFC system after its replacement in an independent container.

Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd. 549
DOI: 10.1002/er
P. Salvin et al. Microbial fuel cell

for power electrical applications. A combination of several 3.3.1. Modeling of BMFCs

equivalent BMFCs could achieve a voltage interesting to
recharge a micro-battery for example.
The maximum efficiency for this configuration at MPP
is defined by (1) and equal to 86.5%.
As it was described in Figure 4, the non-uniform
manufacturing processes or microbial heterogeneity causes
dispersions between the BMFCs. Generally, during the ex-
periments, dispersions in the internal resistance were noted
(Tables II and III). The result is a non-optimized operation
of in series association since traversed by the same current,
BMFCs do not function at their MPP.
The theoretical maximum cell voltage of a BMFC is
never achieved, particularly with BMFCs because of volt-
age drops and current drops. The output voltage is there-
fore expressed as [49]:
V ¼ V OC
ΔV ohm
ΔV act
ΔV trans (2)
where VOC is the open circuit voltage.ΔVohm is the ohmic
drop and it is the simplest to understand and to model.
The voltage drop value is simply proportional to the cur-
rent, which is given by Ohm Equation 2

3.3. Simulation of operation with voltage ΔV ohm ¼ Rohm :I (3)

balancing circuit

The absence of bacterial activity in cells is one of the I is the current density given in A or mA.cm
2, and Rohm is
reasons that cause the voltage reversal. The dispersions the ohmic resistance and should be given in Ω or kΩ.cm2.
between the associated BMFCs may also cause voltage re- ΔVact is the activation voltage drop and given by the so-
versal. Voltage reversal is detrimental to power generation called Tafel Equation 4:
because the concerned BMFC (the weakest) will absorb
the electrical energy of the others BMFCs. However, to ΔV act ¼ B:lnðI Þ (4)
optimize the efficiency, it is possible to use a voltage
balancing circuit with in series-connected BMFCs. The expression of the activation losses is valid when
Voltage balancing circuits are usually used for batteries I > 1A which is not a valid assumption in the case of
to expand their life [38,39]. Applications to supercapacitors MFCs. When currents below 1 A are considered, it can
and photovoltaic cells are made to improve the power gen-
eration [40,41]. Several methods are used to achieve the
voltage balancing for the batteries, and other methods are
used only to prevent the voltage reversal as performed with
solar cell.
The voltage balancing circuits designed for batteries
and photovoltaic cells are not adapted for BMFCs. The
main drawback of passive balancing circuits is that they
have a poor efficiency in the case of low-power generators
like BMFCs [42,43]. Methods that use inductors or trans-
formers to transfer energy have low efficiency due to
switching losses and magnetic losses [43,44]. Converter
architecture such as cûk [45], buck or/and boost [46],
full-bridge, flyback and quasi-resonant can be used for cell
balancing. Although these converters have an acceptable
efficiency, they are expensive and complex. The switched
capacitor and inductive shuttling methods are options
presenting theoretically higher efficiency. The switched
capacitor circuit (SCC) only requires switches (SW), ca-
pacitors (C) and a single oscillator [47,48]. In this method,
a capacitor is alternatively connected in parallel with the
BMFCs at FS frequency. When the BMFCs are unbal-
anced, the capacitor charges when it is connected to the
strongest BMFCs and discharges when it is connected to
the weakest BMFCs. The transfer of charges is limited by
the in series resistances of the voltage sources inside
BMFCs and the capacitor. These resistances include the
on-resistances RSW of the switches, the ohmic resistance
of the BMFCs and the resistance of the capacitors. These
are less area-consuming than inductive-based circuits and Figure 6. A) Features of analytical model and comparison with
countless in components. Therefore, the switched capacitor experimental polarization curve of L1*; B) features of analytical
circuits offer good perspectives for BMFCs. model and comparison with experimental polarization curve of L2.

550 Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd.
DOI: 10.1002/er
Microbial fuel cell
be modified by an almost equivalent expression. The acti-
vation voltage drop is modified as in (5):
 
I
ΔV act ¼ B:ln þ1
C
where B and C are constants depending on the electrode
and cell conditions.
ΔVtrans is the mass transport or concentration voltage
drop and given by Nernst equation ΔVtrans = D. exp(E. I),
D and E are constants to be determined.
The equation for the operating voltage of a fuel cell at a
current density can be written as (6):
 
I the other losses are non-linear and are represented in
V ¼ V oc
Rohm :I
B:ln þ 1
D:expðE:I Þ
C Figure 7 (model a) with variable resistances. This model
This equation defines the analytical expression of a po-
larization curve with five parameters to be defined. A direct
fitting of a polarization curve does not necessarily enable
to identify all parameters precisely because different
parameters can govern similar effects on the polarization
curve. The ohmic resistance is therefore primarily deter-
mined using electrical impedance spectroscopy (EIS). The
other parameters are determined through fitting of the exper-
imental polarization curve with the four-parameter equation
(Prism software). It is simple to model this equation using
a spreadsheet (such as EXCEL), or programs such as
MATLAB®. Voc was measured and it was determined for
Figure 7. Electrical equivalent circuits for a BMFC.
Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd.
DOI: 10.1002/er
P. Salvin et al.
L1* and L2 (0.767 V and 0.710 V, respectively, Table III).
The slope of the polarization curve in the ohmic region allowed
determining the ohmic resistance (1620 Ω and 680 Ω, respec-
tively, for L1* and L2, Table III). This was validated using
(5) Prism software. The other parameters B, C, D, and E can be
determined using Prism software. The fitting method used is
nonlinear regression (least squares fit). The results obtained
are presented in Figure 6. These figures show that the compar-
ison between analytical model (defined by Prism software) and
the experimental polarization curve is very close to each other.
In terms of electrical equivalent circuit, the above ana-
lytical equations can be represented by an ideal voltage
source and a number of electrical components connected
in series and responsible for the different voltage losses.
A constant resistance represents the ohmic losses while
(6)
is valid in steady state because the experiences show
non-immediate reaction of BMFC caused by the internal
mechanisms (i.e. activation and concentration). The inter-
face between each electrode and its surrounding electrolyte
forms a capacitor. The combination of static and dynamic
mechanisms can be represented in an electrical model:
the effects of both concentration and activation overvolt-
ages are modeled by a variable resistance connected in
parallel with a variable capacitor as in Figure 7 (model b).
It is simple to implement this model in MATLAB®: the
simulation results show that the capacitive effect of concen-
tration is almost negligible compared to the activation.
Therefore, the electrical model can be simplified as shown
551
P. Salvin et al. Microbial fuel cell

Figure 8. Results of simulation. A) Serial association of two MFCs represented by their electrical models. B) Polarization (solid line) and
power (dotted) curves Simulation (+, red) and experimental (o, blue). C) Cell voltage for each BMFC.

552 Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd.
DOI: 10.1002/er
Microbial fuel cell
in the Figure 7 (model c). This model will be used to realize
simulation of BMFCs in series without balancing (and
compared to experimental results) and with balancing
circuit (SC).
3.3.2. Simulation of two BMFCs in series without
switched capacitor circuit (SCC)
Figure 8-A presents the in series association of two
BMFCs represented by their electrical models as defined
above. The V–I and P–I simulation curves for associated
BMFCs are given in Figure 8-B. The polarization and
power experimental curves are also shown in these figures.
We see that the simulated curves are very close to experi-
mental curves, which validates the relevance of our electric
model. At the maximum power point, the voltage of each
BMFC is different, leading to a non optimal operating
point. For high values of current, the voltages V1 and V2
drift apart, eventually leading to a negative value of V1 as
represented Figure 8-C. This phenomenon is discussed as
‘voltage reversal’. Voltage reversal is detrimental to power
generation because the concerned BMFC (the weakest)
will absorb the energy instead of producing some. The
maximum efficiency for this configuration is of about
86% at MPP for Rext = 2000 Ω (Table III). To increase
the efficiency and avoid the voltage reversal, a voltage
balancing circuit was added to the in series association.
Figure 9. A) Schematic switched capacitor circuit. B) Efficiency of the association vs the switching frequency Fs for various values of
capacitors. C) Polarization (solid line, +) and power (dotted, o) curves Simulation with SC. D) Cell voltage for each BMFC with SC.
Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd.
DOI: 10.1002/er
P. Salvin et al.
3.3.3. Simulation of two BMFCs in series with SCC
The application of SCC is represented in Figure 9-A. The
global efficiency of the association is defined as Equation 1.
The simulated efficiency of the association with changes in
switching frequency is plotted in Figure 9-B for RSW = 0.5 Ω
(low on resistance) and for various capacitors values C12. At
low switching frequency, the circuit is not useful and volt-
ages are not balanced. Efficiency for FS = 1 Hz is around
80% (66% for C12 = 500 μF). The efficiency increases with
switching frequency until 100 Hz, reaches up to 98% and
finally decreases for frequency beyond 100 Hz. The shape
of the curve seems logical: indeed, efficiency decreases
with increased frequency because it causes the increase of
the switching losses. The maximal efficiency (98.6%) is
reached for Fs = 100 Hz and a capacitor C12 = 100 μF.
Figure 9-C shows V–I and P–I curves for associated
BMFCs with voltage balancing circuit for 10 KHz. When
the BMFCs’ performances are widespread, the SCC
achieves good overall performances. Figure 9-D shows
the cell voltage V1 and V2 for each BMFC. We note that
the voltage reversal phenomenon disappeared (compared
to Figure 8-C for V1) thanks to the balancing circuit which
allows increasing the global efficiency.
The switched capacitor circuit is the most efficient volt-
age balancing circuit. It benefits from the fact that there is
no switch in series with the BMFCs. The switches are only
crossed by the balancing current. A drawback is that it
553
P. Salvin et al.
requires a minimum of (n
1) capacitors with quite high
values (several μF).
With adding voltage balancing circuit, the global effi-
ciency is increased of 12% (Table III). Moreover, this
circuit allows avoiding voltage reversal. It is simple to
implement and offers good perspectives for the serial asso-
ciation of several BMFCs.
Systems in series or in parallel have achieved energy
sources greater than that provided by a single BMFC.
The parallel combination has increased current and power.
However, the performance obtained is difficult to exploit in
order to supply a conventional electric system, without
voltage amplification. The in series association allows
increasing output voltage to achieve acceptable levels for
supply wireless sensors. All these results point out the pos-
sibility of the construction of prototypes combining both in
parallel and in series BMFCs to access the power settings
features of small electrical devices.
4. CONCLUSION
The study presented in this article has put forward several
points:
1. A benthic microbial fuel cell, located in a tropical
mangrove, was operated during ten days and the
operational characteristic at this date were 453 V.
2. Similar laboratory BMFCs were operated over 40 days
with stables cell voltages and without any addition of
organic substrate.
3. In this study, BMFCs in situ and in laboratory appeared
to be autonomous electricity harvesting systems, only
using the nutrients in the mangrove environment.
4. Associations in parallel and in series of the labora-
tory BMFCs have shown maximum efficiencies of
88.7% and 86.5%, respectively.
5. The voltage balancing circuits enabled efficient serial
association of non-uniform BMFCs and avoided
voltage reversal phenomena. The switched-capacitor
circuit achieved very high performances.
ACKNOWLEDGEMENTS
This work was supported by the Bio-Energies programme
of the French ANR through the Agri-Elec project (ANR-
08-BioE-001), European Union through ‘AMABIO’
framework (European Regional Development Funds, PRE-
SAGE 30927), CNRS PEPS Energie INSIS ‘GEPAM’ and
by the French Ministry in charge of 2013 MOM projects.
The authors would like to thank the city of Matoury for
allowing us accessing to the port of Larivot.
REFERENCES
1. Kim BH, Kim HJ, Hyun MS, Park DH. Direct elec-
trode reaction of Fe (III)-reducing bacterium,
554
Microbial fuel cell
Shewanella putrefaciens. Journal of Microbiology
and Biotechnology 1999; 9:127–131.
2. Rabaey K, Verstraete W. Microbial fuel cells: novel
biotechnology for energy generation. Trends in Bio-
technology 2005; 23:291–298.
3. Logan BE, Hamelers B, Rozendal R, Schröder U,
Keller J, Freguia S, et al. Microbial fuel cells: method-
ology and technology. Environmental Science and
Technology 2006; 40:5181–5192.
4. Lovley DR. Bug juice: harvesting electricity with mi-
croorganisms. Nature Reviews Microbiology 2006;
4:497–508.
5. Reimers CE, Tender LM, Fertig S, Wang W. Harvest-
ing energy from the marine sediment - water interface.
Environmental Science and Technology 2001; 35:192–195.
6. Dumas C, Mollica A, Féron D, Basseguy R,
Etcheverry L, Bergel A. Checking graphite and stain-
less anodes with an experimental model of marine mi-
crobial fuel cell. Bioresource Technology 2008;
99:8887–8894.
7. Martins G, Peixoto L, Ribeiro DC, Parpot P, Brito AG,
Nogueira R. Towards implementation of a benthic mi-
crobial fuel cell in lake Furnas (Azores): phylogenetic
affiliation and electrochemical activity of sediment
bacteria. Bioelectrochemistry 2010; 78:67–71.
8. Venkata Mohan S, Srikanth S, Veer Raghuvulu S,
Mohanakrishna G, Kiran Kumar A, Sarma PN. Evalu-
ation of the potential of various aquatic eco-systems in
harnessing bioelectricity through benthic fuel cell: ef-
fect of electrode assembly and water characteristics.
Bioresource Technology 2009; 100:2240–2246.
9. Rezaei F, Richard TL, Brennan RA, Logan BE.
Substrate-enhanced microbial fuel cells for improved
remote power generation from sediment-based sys-
tems. Environmental Science and Technology 2007;
41:4053–4058.
10. Song TS, Yan ZS, Zhao ZW, Jiang HL. Removal of
organic matter in freshwater sediment by microbial fuel
cells at various external resistances. Journal of Chemical
Technology and Biotechnology 2010; 85:1489–1493.
11. Tender LM, Gray SA, Groveman E, Lowy DA,
Kauffman P, Melhado J, et al. The first demonstration
of a microbial fuel cell as a viable power supply:
powering a meteorological buoy. Journal of Power
Sources 2008; 179:571–575.
12. De Schamphelaire L, Rabaey K, Boeckx P, Boon N,
Verstraete W. Outlook for benefits of sediment micro-
bial fuel cells with two bio-electrodes. Microbial Bio-
technology 2008; 1:446–462.
13. Fu Y, Yu J, Zhang Y, Meng Y. Graphite coated with
manganese oxide/multiwall carbon nanotubes compos-
ites as anodes in marine benthic microbial fuel cells.
Applied Surface Science 2014; 317:84–89.
Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd.
DOI: 10.1002/er
Microbial fuel cell
14. Ryckelynck N, Stecher Iii HA, Reimers CE. Under-
standing the anodic mechanism of a seafloor fuel cell:
interactions between geochemistry and microbial
activity. Biogeochemistry 2005; 76:113–139.
15. Fan Y, Sharbrough E, Liu H. Quantification of the
internal resistance distribution of microbial fuel
cells. Environmental Science and Technology 2008;
42:8101–8107.
16. Cheng S, Liu H, Logan BE. Increased power generation
in a continuous flow MFC with advective flow through
the porous anode and reduced electrode spacing. Environ-
mental Science and Technology 2006; 40:2426–2432.
17. Liu H, Cheng S, Logan BE. Power Generation in Fed-
Batch Microbial Fuel Cells as a Function of Ionic
Strength, Temperature, and Reactor Configuration. Envi-
ronmental Science and Technology 2005; 39:5488–5493.
18. Zhang Y, Angelidaki I. Submersible microbial fuel cell
sensor for monitoring microbial activity and BOD in
groundwater: focusing on impact of anodic biofilm
on sensor applicability. Biotechnology and Bioengi-
neering 2011; 108:2339–2347.
19. Dewan A, Beyenal H, Lewandowski Z. Scaling up mi-
crobial fuel cells. Environmental Science and Technol-
ogy 2008; 42:7643–7648.
20. Logan B, Cheng S, Watson V, Estadt G. Graphite fiber
brush anodes for increased power production in air-
cathode microbial fuel cells. Environmental Science
and Technology 2007; 41:3341–3346.
21. Kim HJ, Park HS, Hyun MS, Chang IS, Kim M, Kim
BH. A mediator-less microbial fuel cell using a metal
reducing bacterium, Shewanella putrefaciens. Enzyme
and Microbial Technology 2002; 30:145–152.
22. Jiang Y, Xu Y, Yang Q, Chen Y, Zhu S, Shen S.
Power generation using polyaniline/multi-walled
carbon nanotubes as an alternative cathode catalyst in
microbial fuel cells. International Journal of Energy
Research 2014; 38:1416–1423.
23. Timmers RA, Strik DPBTB, Hamelers HVM, Buisman
CJN. Increase of power output by change of ion trans-
port direction in a plant microbial fuel cell. International
Journal of Energy Research 2013; 37:1103–1111.
24. Liu Z, Liu J, Zhang S, Su Z. A novel configuration of
microbial fuel cell stack bridged internally through an
extra cation exchange membrane. Biotechnology Letters
2008; 30:1017–1023.
25. Aelterman P, Rabaey K, Pham HT, Boon N, Verstraete
W. Continuous electricity generation at high voltages
and currents using stacked microbial fuel cells. Environ-
mental Science and Technology 2006; 40:3388–3394.
26. Zhuang L, Zheng Y, Zhou S, Yuan Y, Yuan H, Chen Y.
Scalable microbial fuel cell (MFC) stack for continuous
real wastewater treatment. Bioresource Technology
2012; 106:82–88.
Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd.
DOI: 10.1002/er
P. Salvin et al.
27. Ieropoulos I, Greenman J, Melhuish C. Microbial fuel
cells based on carbon veil electrodes: stack configura-
tion and scalability. International Journal of Energy
Research 2008; 32:1228–1240.
28. Donovan C, Dewan A, Heo D, Beyenal H. Batteryless,
wireless sensor powered by a sediment microbial fuel
cell. Environmental Science and Technology 2008;
42:8591–8596.
29. Zhang Y, Angelidaki I. Self-stacked submersible mi-
crobial fuel cell (SSMFC) for improved remote power
generation from lake sediments. Biosensors and Bio-
electronics 2012; 35:265–270.
30. Salvin P, Roos C, Robert F. Tropical mangrove sedi-
ments as a natural inoculum for efficient electroactive
biofilms. Bioresource Technology 2012; 120:45–51.
31. Lowy DA, Tender LM, Zeikus JG, Park DH, Lovley
DR. Harvesting energy from the marine sediment–
water interface II: kinetic activity of anode materials.
Biosensors and Bioelectronics 2006; 21:2058–2063.
32. Scott K, Cotlarciuc I, Head I, Katuri KP, Hall D,
Lakeman JB, et al. Fuel cell power generation from
marine sediments: investigation of cathode materials.
Journal of Chemical Technology and Biotechnology
2008; 83:1244–1254.
33. Scott K, Cotlarciurc I, Hall D, Lakeman JB, Browning
D. Power from marine sediment fuel cells: the influ-
ence of anode material. Journal of Applied Electro-
chemistry 2008; 38:1313–1319.
34. Reimers CE, Girguis P, Stecher Iii HA, Tender LM,
Ryckelynck N, Whaling P. Microbial fuel cell
energy from an ocean cold seep. Geobiology 2006;
4:123–136.
35. Nielsen ME, Reimers CE, Stecher HA. Enhanced
power from chambered benthic microbial fuel cells.
Environmental Science and Technology 2007;
41:7895–7900.
36. Song T, Jiang H. Effects of sediment pretreatment on
the performance of sediment microbial fuel cells.
Bioresource Technology 2011; 102:10465–10470.
37. Donovan C, Dewan A, Peng H, Heo D, Beyenal H.
Power management system for a 2.5 W remote sensor
powered by a sediment microbial fuel cell. Journal of
Power Sources 2011; 196:1171–1177.
38. Altemose G. A battery Electronics Unit (BEU) for
Balancing Lithium-Ion Batteries. SAE Technical Pa-
per 2008-01-2097 2008. doi:10.4271/2008-01-2907
39. Baronti F, Fantechi G. Hierarchical platform for mon-
itoring, managing and charge balancing of LiPo batte-
ries. IEEE Vehicle Power and Propulsion Conference
Sep. 2011; pp. 1–6, Chicago, USA.
40. Diab Y et al. Comparison of the different circuits used
for balancing the voltage of supercapacitors: studying
performance and lifetime of supercapacitors. Proceeding
555
P. Salvin et al.
of the 2nd European Symposium on Super Capacitors &
Applications, 2006, Laussane, Switzerland.
41. Vighetti S et al. Photovoltaic module and shadow:
study and integration of a current balancing system.
Proceedings of EUPVSEC, 2010; 4139–4143,
Valencia, Spain.
42. Lindemark B. Individual Cell voltage Equalizers (ICE)
For Reliable Battery Performance. IEEE 13th Interna-
tional Telecommunications Energy Conference, INTELEC
’91, 1991; 196–201, Kyoto, Japan.
43. Dorin V et al. An Energy Converter Method for Bat-
tery Cell Balancing. IEEE 33rd International Spring
Seminar on Electronics Technology (ISSE), 2010;
290–293, Warsaw, Poland.
44. Yan J et al. Fuzzy Control for Battery Equalization
Based on State of Charge. Vehicular Technology Con-
ference Fall (VTC 2010-Fall), 72nd, IEEE, Sept. 2010;
pp. 1–7, Ottawa, Canada.
45. Hong W et al. Charge Equalization of Battery Power
Modules in Series. The 2010 International Power
556
Microbial fuel cell
Electronics Conference, 2010; pp. 1568–1572,
Sapporo, Japan.
46. Kim CH et al. Individual Charge Equalization Con-
verter with Parallel Primary Winding of Transformer
for Series Connected Lithium-Ion Battery Strings in
an HEV. Journal of Power Electronics 2009; 9(3):
472–480.
47. Pascual C et al. Switched capacitor system for auto-
matic series battery equalization. Applied Power Elec-
tronics Conference and Exposition, 1997. APEC ’97
Conference Proceedings 1997, Twelfth Annual, Feb.
1997; Vol. 2: pp. 848–854, Atlanta, USA.
48. West S et al. Switched-Capacitor Systems for
Battery Equalization. IEEE Modern Techniques and
Technology (MTT 2000). Proceedings of the VI
International Scientific and Practical Conference
of Students, Post-graduates and Young Scientists,
57–59, Russia.
49. Larmini J, Dicks A. Fuel cell systems explained (2nd).
John Wiley & Sons, 2003.
Int. J. Energy Res. 2015; 39:543–556 © 2014 John Wiley & Sons, Ltd.
DOI: 10.1002/er