View Article Online / Journal Homepage / Table of Contents for this issue

TUTORIAL REVIEW

www.rsc.org/csr | Chemical Society Reviews

Techniques for the study and development of microbial fuel cells:

an electrochemical perspective
Feng Zhao,* Robert C. T. Slade and John R. Varcoe

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

Received 6th November 2008

First published as an Advance Article on the web 7th April 2009
DOI: 10.1039/b819866g
Microbial fuel cells (MFCs) represent a clean and renewable energy resource. To date, power
generation in MFCs is severely limited. In order to improve performance, a wide range of
techniques have been utilised for a fundamental scientic understanding of the components and
processes and also to investigate MFC performance bottlenecks. In this tutorial review, we discuss
the electrochemical/electroanalytical techniques employed in recent MFC studies and discusses the
principles, experimental implementation, data processing requirements, capabilities, and
weaknesses of these techniques.

1. Introduction
Microbial fuel cells (MFCs) and enzymatic fuel cells (EFCs)
are the two principal types of biological fuel cells (also called
biofuel cells);1,2 these are devices that transform chemical
energy into electrical energy via electrochemical reactions
involving microbes or enzymes either located in both anode
and cathode compartments or in just one compartment;
enzyme (protein) containing electrodes can also be employed
in MFCs, thereby resulting in a hybrid of EFCs and MFCs.
MFCs generally consist of an anode, a cathode, an
electrolyte (which can be in membrane form) and an electrical
circuit; microbes located in one (or more) of the compartments
are used to degrade/convert organic and/or inorganic
substrates via their metabolisms. Electrons are produced at

Chemical Sciences, University of Surrey, Guildford UK GU2 7XH.

E-mail: [email protected]; Fax: +44 01483 686851;
Tel: +44 01483 682580

the anode and transferred through an external circuit to the

cathode, whilst a charge balancing number of cations and/or
anions are transferred between anode and cathode.
Utilising microbes or enzymes for electricity generation
extends the range of fuels (substrates) that can be oxidised
or oxidants that can be reduced; hence biological fuel cells can
carry out specic tasks, and/or have niche applications, in
addition to power generation. Two key issues that are required
for a sustainable society are environmental protection and the
generation of clean energy. MFCs are attracting worldwide
attention, driven by the promise of clean and renewable energy
from various wastes and wastewaters. Currently, MFCs can
only produce low power outputs (o6 W m2; r500 W m3)3
due to many factors related to the anode, the cathode, the
chemical species present in the electrolyte, the ion-exchange or
ltration membrane (if used), the microbial species present
and their metabolisms, fuel cell conguration, and operational
conditions. Scaling up of these devices can be problematic,
since what are considered as acceptable losses at lab scale,

Feng Zhao received his doctorate degree in chemistry under the

supervision of S. J. Dong at Changchun Institute of Applied
Chemistry, Chinese Academy of Sciences in 2004. He then spent
two years as a research assistant at the University of Greifswald,
with U. Schroder and F. Scholz, working on the development of
microbial fuel cells. In 2007, he came to the University of Surrey as
a research ocer. His current research focuses on the development
of enzymatic and microbial fuel cells.
Prof. Robert C. T. Slade trained at the University of Oxford. He
held academic posts at the Universities of Oxford and Exeter, and
moved to the University of Surrey in 1998. He leads an extended
team of researchers in the chemistry of energy materials and
functional devices and also leads the UKs SuperGen5 Biological
Fuel Cells Consortium (constituent teams in 6 Universities).
Dr John Varcoe trained at the University of Exeter, receiving a 1st
Left to right: Feng Zhao, Robert C. T. Slade and
class honours BSc Chemistry in 1995 and a PhD (studying
John R. Varcoe
inorganic polymer electrolytes) in 1999. He held post-doctoral
positions at the University of Surrey and was appointed Lecturer in Materials Chemistry in August 2006. His research interests are
polymer membranes for fuel cells and biological systems for energy generation. He is a specialist in the application of alkaline anionexchange membranes in fuel cells and is a key member of the Surrey team that is part of the UKs SuperGen5 Biological Fuel Cells
Consortium.
1926 | Chem. Soc. Rev., 2009, 38, 19261939

This journal is


c

The Royal Society of Chemistry 2009

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

View Article Online

become parasitic when scaled up to demonstration or

commercial products.
A better understanding of all of the components is required,
to allow the bottlenecks in MFC to be identied and to
improve power outputs; the selection of the most appropriate
techniques for MFC diagnosis and evaluation is vital in
achieving this. To fully understand MFCs, knowledge of
multiple technological disciplines is required, such as electrochemistry, microbiology, materials science and engineering,
molecular biology and environmental engineering. It is a
challenge for any research team to completely understand all
of the theories and techniques used in the study of MFCs.
An example is that some of the electrochemistry employed
and/or described in some research papers can be described as
ambiguous and/or inappropriate. In recent years a series of
detailed review papers have discussed the MFC history,
construction, reaction mechanisms, operational conditions
and limitations.29 However, review of the most appropriate
electrochemical techniques required for MFC diagnosis has
not been given the appropriate level of attention, with only a
few such techniques being discussed.10
New mechanisms (microbial, physiological, chemical
and electron transport) are being gradually elucidated.
Electrochemical/analytical techniques have an essential role
in this continuing process; they are vitally important in
analysing the limiting performances of each component, to
optimise MFC operation, and to allow continued innovation.
In this article, we review, from both electrochemistry and fuel
cell materials chemistry perspectives, the techniques that are
useful for the study and development of MFCs; the aim
is to complement previously published reviews, to highlight
some of the capabilities and to expose weaknesses of these
techniques.

2. Basic considerations
Generally speaking, the electrochemical techniques now being
more routinely used in MFC studies have been used in the
study of traditional electrochemical systems (e.g. chemical fuel
cells and electrocatalytic systems) for many decades. The
biological processes involved in electron transfer reactions in
MFCs are more complex when compared with classical
electrochemical systems that are, by comparison, relatively
simple and clean. The main dierence between chemical fuel
cells [such as a hydrogen/oxygen (air) polymer electrolyte fuel
cell (PEFC)] and MFCs is that the electricity generation in
MFCs arises from the metabolic activity of microbes; the
continuous and long term operation of MFCs mandates that
systems must be run under conditions that are generally
predened by the requirements for optimal microbial growth
and sustainability. Typical conditions include ambient
temperatures and pressures, electrolytes with a low concentration of ions, and near neutral pH.
Maximum current densities of MFCs are typically
o3 mA cm2; these levels are signicantly lower than those
of PEFCs, which are of the order of 41 A cm2
(both normalised to geometrical electrode area). The performances and requirements of the electrode materials and
catalysts in MFCs and chemical fuel cells dier substantially,
This journal is


c

The Royal Society of Chemistry 2009

even for the same reaction. One example is the oxygen

reduction reaction (ORR)1114 where anions such as Cl,
HSO4, HPO42, and HS are well known to be species that
poison Pt catalysts and seriously interfere with the electrode
reactions on Pt catalytic surfaces; however, the majority of
MFC and related studies involve the presence of phosphate
buers, ionic strength adjusters (e.g. NaCl) and other
nutrients. The low O2 dissolution concentrations in electrolyte
and the low H+ concentration (E107 mol dm3) have a
signicant impact on the cathode ORR performance in a MFC
operating at ambient temperature. It is essential that the
nature of all the species (including spectator ions and
microbes) are taken into account, especially when they can
come into contact with the catalysts. This could be one
explanation why non-Pt ORR catalysts show similar
performance levels with MFCs containing platinum. MFC
components (electrodes, catalysts, membranes etc.) must be
tolerant to these species for long term operation.
The following sections briey introduce MFCs from an
electrochemical point of view, before discussing the suitability
of the various techniques that can be used in studying these
complicated systems.
2.1 Choice of electrode congurations
Two- and three-electrode congurations are commonly
employed to determine MFC and electrode performances;
a four-electrode conguration is sometimes used when the
conductivity of an electrolyte (membrane) or electrode material
is tested by electrochemical impedance spectroscopy (EIS).
In practice, a MFC only needs two electrodes for electricity
generation, an anode and a cathode. A variety of materials
have been used as electrodes: graphite foil, rods, granules, and
bre brush; carbon paper, cloth, felt, and foam; activated
carbon cloth; reticulated vitreous carbon; electrodes modied
with conductive polymers; and metals such as aluminium,
nickel or stainless steel. The choice of materials used for the
anode or cathode generally depends upon the electrical
potential window required. It is sometimes overlooked that
the electrode materials and substances in the electrolyte are
not electrochemically inert and may themselves generate
currents in a potential range of the experiment in question.
For example: (1) metal electrodes will generate currents due to
chemical and/or microbial corrosion or passivation; (2) carbon
materials will most likely not have pristine surfaces composed
of 100% elemental carbon and will have heteroatom-containing
surface groups such as quinones, amines, carboxylic acids, or
other nitrogen containing functionalities, some of which can
be electrocatalytically active; (3) electroactive binders are
sometimes added for reasons of ease of electrode manufacture;
(4) unwanted electroactive species (contaminants) are easily
absorbed on the surface of electrodes prior to experiment
and can also generate currents; (5) Pt electrodes will cause
background currents by electrochemically oxidising a variety
of substrate materials that may be present in MFCs e.g.
sucrose or acetate; (6) electroactive chemicals exist naturally
in various wastewaters. The background currents from the
electrode materials and eletroactive chemicals present in the
MFC should be evaluated in the absence of microbial species if
Chem. Soc. Rev., 2009, 38, 19261939 | 1927

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

View Article Online

meaningful results are to be collected. Those background/

control experiments are necessary for the complete understanding of a systems electrochemistry, especially for complex
devices such as MFCs.
The measurement of a cathodeanode two electrode
conguration represents the overall MFC performance, which
will be limited by many factors and is rarely completely
reproducible due to changes in the growth, evolution, and
metabolism of the consortium of microbes present. In order to
avoid unwanted perturbations to the system and to simplify
studies, a reference electrode (RE) and a counter electrode
(CE, sometimes called an auxiliary electrode) are often
introduced into a MFC in order to analyse the behaviour of
an individual working electrode (WE), which can either be the
anode or cathode: this is then a three-electrode conguration.
The potential of or the current generated at the WE can be
precisely controlled in a three-electrode mode (this is not
always possible with a two-electrode conguration).
The type of RE used must be clearly reported as the
potential dierence between practical REs of dierent types
and the standard hydrogen electrode diers; care is required
when a RE is employed over long periods of time as there will
inevitably be small potential changes e.g. from concentration
changes inside the RE. Due to the fact that silver/silver
chloride (Ag/AgCl) and saturated calomel electrodes (SCE)
are cheap, practical, and easily prepared, and in addition have
the ability to be revived after long term operation, they are
commonly the preferred option for REs.
Electrically conductive (e) and high surface area materials,
such as Pt web or graphite foil, are usually employed as the
CE, in order to prevent CE derived mass transport and
activity limitations interfering with the measurement of the
performance at the WE. The requirement of the CE is that it
does not produce byproducts or contamination that would
cause interferences to the anode or cathode reactions being
studied (the tiniest trace of Pt on a WE can increase its activity
towards a variety of species); it can, if required, be separated
from the chamber/compartment by a glass frit or sinter to
minimize CE-derived contamination.
2.2

Electrode reaction mechanisms and processes

Dierent catalysts for, and mechanisms of, MFC electrode

reactions have been proposed. For the cathode,9 catalysts such
as carbon-supported platinum, metal tetramethyphenylporphyrins and metal phthalocyanine have been used for the
ORR. As well as oxygen in air, electron accepting chemicals
such as ferricyanide or MnIV species can be used. Biocathodes
have also been reported.15 For the anode, direct electron
transfer (DET) from the microbes to the typically carbonbased anode via extracellular cytochromes16 or conductive
microbial pili (electrically conductive nanowires) have been
observed.17 Indirect (mediated) electron transfer (IET) to the
anode is also possible via electroactive metabolites, such as
hydrogen or sulde,18,19 or with the addition of non-natural
(chemical) redox mediators.
For abiotic cathodes immersed in electrolyte solutions, the
electrode reactions of MFCs are heterogeneous and take place
in the interfacial region between the solid electrode and the
1928 | Chem. Soc. Rev., 2009, 38, 19261939

Fig. 1 Schematic of the pathways involved in solution phase

electrochemical reactions composing of a series of simultaneous or
consecutive steps: mass transfer, chemical reactions or other surface
reactions, electron transfer. The dierent species (O, O0 , O00 , R, R0 and R00 )
represent the corresponding forms before and after the processes
such as adsorption, desorption, protonation and electron transfer.
(Adapted from ref. 20).

solution. An example of a pathway for a simple oxidation or

reduction process (O + ne 2 R, where O is the oxidised
species and R is the corresponding reduced form) on an
electrode is given in Fig. 1. The distribution of charged species,
which aects the electrode processes, diers in the electrode
surface region and in the bulk solution and can take place in
multiple stages as summarised below:20
1. Mass transport of the O00 species to the electrode surface
where the reaction (O + ne 2 R) occurs;
2. On the electrode surface, chemical reaction
[e.g. adsorption, desorption, protonation and decomposition
steps (O00 2 O 0 and R00 2 R 0 )] that might not involve the
oxidation or reduction processes may precede or follow the
electron transfer step (or steps);
3. Electron transfer (O + ne 2 R) at the electrode surface;
4. Transport of the R 0 from the electrode surface to the bulk
solution;
5. Transport of electrolyte including spectator ions and
chemical species.
Electrode processes involving electron transfer can take
place in several simultaneous or consecutive steps. The rate
of an overall reaction is aected not only by the reaction
processes at the electrode surface but also by the mass
transport of species at the interface between the bulk solution
and electrode surface region. Mass transfer can have contributions from: (i) diusion along concentration gradients;
(ii) migration when an electric eld is present; and (iii)
convection in the presence of any form of agitation or forced
motion. In traditional electrochemical systems, it is usual
practice to design experiments where the conditions present
facilitate the treatment of recorded data (e.g. leading to
simplication of the electrochemical equations required
for the calculation of kinetic parameters with the removal
of unwanted factors). For example, an inert supporting
This journal is


c

The Royal Society of Chemistry 2009

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

View Article Online

electrolyte of Z 0.1 mol dm3 concentration is often added to

reduce migration eects and ohmic losses, and convection
can either be avoided by preventing stirring and vibrations
in the system being studied or precisely controlled using
hydrodynamic techniques (see section 3.11). However,
the operational conditions typically found in MFCs
[e.g. r50 mmol dm3 (low concentrations) of stirred electrolyte]
are predened by the requirements of the microbes for
optimum growth and metabolism and are not necessarily
compatible with the objectives of an electrochemical study.
Careful experimental design is essential in MFC studies when
the aim of a study is to determine fundamental kinetic
parameters and the data must have some form of mass
transport as well as ohmic corrections.
A potential pathway for DET between biolm and electrode
is shown in Fig. 2a. The region of biolm formation on the
electrode surface is critical to the operation of a MFC and both
living and dead cells are present in the self-generating biolm.
The following steps can occur during the electrode reaction:
1. Transport of the substrate from the bulk solution to the
biolm region;

Fig. 2 Schematic of the pathways involved in solution during

electrochemical reactions with the presence of a solid biolm electrode:
(a) direct electron transfer between the biolm and the electrode;
(b) indirect electron transfer via the electroactive metabolites.

This journal is


c

The Royal Society of Chemistry 2009

2. Transport of the substrate in biolm region;

3. Metabolisms inside the dierent microorganisms that are
present;
4. DET between the biolm (i.e. the active sites of extracellular enzymes of the microorganisms) and the electrode;
5. Transport of the metabolites and the reaction products;
6. Transport of electrolyte including spectator ions and
chemical species.
For anode reactions, living cells in the biolm consume the
substrate (electron donor) that is transferred from the bulk
solution and the cellular metabolisms consist of a series of
enzyme reactions, some of which yield the electrons. The
intrinsic electron relay occurs in the microbes during the
biological metabolism and a proportion of these electrons
can be collected by the anode via extracellular cytochromes
(when present).
The biolm thickness is greater than the electrode surface
region and the positions of extracellular enzymes have an
important inuence on the DET reactions (Fig. 2a):
(a) DET can occur only when the active sites of extracellular
enzymes are suciently close to the electrode surface;
(b) When microbial cells in the biolm are a certain distance
from the electrode surface, it is impossible for DET generation
to occur via extracellular cytochromes;
(c) Some microbial cells contain electrically conductive pili,
which can extend for tens of micrometres and may serve as
electrical conduits for long range direct electron transfer.
Most redox proteins are known to undergo very slow
heterogeneous electron transfer at electrode surfaces and
exhibit irreversible electrochemical behaviour; one limitation
of DET is that the active sites are typically buried deep within
the enzymes, which results in poor electrode kinetics. Fig. 2b
illustrates the possible scenarios for indirect electron transfer
(IET). Electroactive metabolites, such as phenazine, H2 or
HS, are produced by the living cells in the bulk solution
and/or in the biolm and can react at the electrode surface.
There are two primary mechanisms by which this can
take place:
(a) When the cells are in contact with the electrode (i.e. the
electroactive metabolites released from the cells are already on
the electrode surface), the electron transfer reaction can occur
with preceding and/or subsequent chemical reactions and/or
surface reactions;
(b) When the electroactive metabolites are released away
from the electrode surface (e.g. by the suspended cells in the
bulk solution), mass transport must occur in addition to the
chemical reactions, electron transfer, and/or other surface
reactions.
In some earlier studies,2 articial redox mediators
(e.g. neutral red) were added to MFCs to enhance electron
transfer processes. There are, however, often undesirable
consequences when doing this (all of which make the use of
such articial mediators an unsuitable choice for sustainable
electricity generation in MFCs): these species tend to be toxic,
they require regular addition and they often cause environmental problems. This review will not further discuss the
functions of these articial mediators.
In summary, there is typically a large variety of microbial
species in commercially relevant MFCs, but only a part of
Chem. Soc. Rev., 2009, 38, 19261939 | 1929

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

View Article Online

these diverse consortia is directly involved in electrode

reactions. The pathways for the electrode reactions, DET or
IET, are interlinked and the processes involve a complex
sequence of coupled steps. Microbial kinetics are directly
linked to the reactions occurring at the electrode. Microbial
growth is dependent on the substrate, metabolites and
microbial concentrations, and an adaptation of the biolm
structure, and/or a change of the microbial community, will
inuence the electrochemical characteristics. The porosities
and pore dimensions of the biolm aect the transfer of
metabolites to the electrode surface. Thick biolms might
enhance DET, but yield increased mass transfer losses
(and associated impedancessee section 3). The transfer of
protons, substrate and metabolites between electrode surface
and bulk solution will also have an eect on the current
generation.21 For IET systems, higher power outputs are often
obtained with the presence of higher concentrations of metabolites. The rate of the overall electrode reaction is determined
by the rate of the slowest step in the sequence of coupled
reactions. The metabolism of the microbes is generally the
rate-limiting step and this limits the rate of current generation
to that dictated by the rate of substrate oxidation by the active
microbes; current cannot be generated at an electrode faster
than the rate that the microbes can oxidise a substrate and
transfer electrons to the electrode.
2.3

Microbes, conguration and operational conditions

Dierent MFC congurations have been designed for both

batch and continuous operation modes: examples include
two-chamber, single-chamber, upow tubular and miniMFC congurations. Many dierent separator types have
been used in MFCs and these include:22,23 cation or anion
exchange membranes, bipolar membranes, nano-porous lters
etc. A large variety of microbesubstrate combinations have
been studied and new types are being continuously discovered.
There are many factors that aect MFC performance and
even one small, seemingly insignicant change can lead to a
signicant substantial deterioration in performance. The
correct choice of experimental technique(s) for any specic
research objective is very important. Testing of identical
MFCs in parallel must be conducted to evaluate reproducibility and repeatability; data from only one test may lead to
misunderstandings due to the complex nature of MFC
systems.

3. Electrochemical/electroanalytical techniques
The power generated by a MFC is quantied in terms of
power output, P = Vcell (V)  I (A). The open circuit voltage
(OCV) is the voltage of a MFC under a no-load condition
(no current generation) and can be measured with a high
impedance voltmeter or potentiometer. The open circuit
potential (OCP) is the potential of an electrode (cathode or
anode) measured against a suitable RE, again using a high
impedance-measuring device. An overpotential, a measure of
the irreversibility of the processes occurring, is the dierence
between the equilibrium (reversible) potential of an electrode
with zero net current and its operating potential with a net
current ow; it broadly represents the extra energy (kick)
1930 | Chem. Soc. Rev., 2009, 38, 19261939

needed to force the electrode reactions to proceed at the

required rate.24 The cell voltage of a MFC can be expressed
in terms of the overpotentials associated with dierent fundamental phenomena as shown in the following equations:25
Ecell = Ec  Zact,c  Zconc,c  Ea  Zact,a  Zconc,a  Zohm (1)
Ecell = Ec  IRact,c  IRconc,c  Ea  IRact,a  IRconc,a
 IRohm
(2)
where Ec and Ea are the reversible OCPs for the cathodic and
anodic reactions respectively; Zact are the charge transfer
derived overpotentials of the anode or cathode (alternatively
called charge transfer losses or activation lossesthe
subscripts a = anode and c = cathode); Zconc are the
concentration overpotentials of the anode or cathode
(also called mass transport losses); Zohm are the ohmic overpotentials of MFCs (alternatively called ohmic lossesOhms
Law V = IR). Each of the overpotentials for various steps can
be represented with an associated resistance: Ract, Rconc, and
Rohm, (eqn (2)); Ract and Rconc are not real electrical elements
and are better described as impedances,20 but they can be
useful for characterizing MFCs and their constituent electrode
reactions e.g. Ract is mathematically related to the much more
important exchange current density, a parameter discussed
later in section 3.3. Caution is required in using the term of
internal resistance as the chemical fuel cell community
considers this to refer to Rohm only, whereas many MFC
studies use the term to mean the sum of all of the Ract, Rohm
and Rconc that are present in the system. A small impedance
(resistance) represents a low overpotential and a rapid process;
a key challenge for improving MFC power outputs is to
reduce these impedances.
3.1 MFC polarization techniques
The term polarization (which is old fashioned and sometimes misleading) is the change of electrode potential (or MFC
voltage) from its equilibrium state due to a ow of current.
Polarization curves are plots of electrode potential (or MFC
voltage) as a function of current or current density. Such plots
contain a wealth of information and can be obtained with a
suitable potentiostat (capable of measuring power inputsi.e.
power is supplied to the instrument from the power source
being studied) or a variable external resistance load; for highly
reproducible, precise measurements it is essential that an
electronic load with the appropriate current ranges is used.
There are four options for the measurement of MFC
polarization discharge curves: (a) constant resistance discharge
measured by connecting dierent resistors to the MFC and
measuring the resulting currents and voltages; (b) potentiodynamic polarization, i.e. linear sweep voltammetry (LSV)
where current is measured with slow voltage scan rate
(e.g. 1 mV s1); (c) galvanostatic discharge, where the current
is controlled and the resulting voltages measured; and (d)
potentiostatic discharge, where the voltage is controlled and
the resulting currents measured.
For the above options, constant potential (potentiostatic)
discharge measurements may be the most suitable for investigations that are probing the fundamental science that is
involved in biofuel cells, especially when the electrochemistry
This journal is


c

The Royal Society of Chemistry 2009

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

View Article Online

of enzymes is involved: (i) electroactive biological species

(enzymes) can rapidly switch their activities on or o
depending on the potentials present; hydrogenase fuel cell
catalysts are a classic example as they can suddenly switch
o (deactivate) when the potentials are raised beyond a certain
point, this being due to the oxidation of the metal centres to
non-active oxidation states.26 This is in contrast to chemical
fuel cell studies where constant current (galvanostatic)
discharges are preferred as constant reaction rates (electron
generation/consumption rates) are occurring at the electrodes.
(ii) for constant resistance discharge techniques, both the
current outputs and cell voltages will be varying with time
(two variables, however small their variation). In comparison,
potentiostatic experiments involve controlled and constant cell
voltages and require only measurement of the current (the only
major electronic variable). Constant potential and current
data are useful when MFCs are designed as power supplies
for practical systems; constant resistance values do not yield
much useful information for MFC studies, especially when
MFCs of dierent congurations and dimensions are being
compared. Even when chemical fuel cells have been tested for
real world scenarios, the variable load has been via a
controlled and time varying current prole (and not by a
external resistance prole). The use of electronic loads has
clearly beneted the development of chemical fuel cells.
Constant resistance measurements are not necessary when
accurate electronic loads or battery/fuel cell test equipment
is available. (iii) nally, it is often dicult to determine when
using LSV measurements (even with a slow scan rate) if a
system is achieving a steady state for each data point collected.
The factors aecting the steady state operation for MFCs are
much more complicated than for traditional electrochemical
systems.
The time that should be left for equilibration before
recording data points at each resistance/current/potential is
still being debated (seconds, minutes, hours). Unlike with a
chemical fuel cell (time frames in the range of several seconds
to several minutes are typically used when recording galvanostatic data, after which steady potentials are normally
achieved), it is dicult to record a credible steady state
polarization curve with a MFC for the following principal
reasons: (a) the nature of the microbial species/communities is
evolving with time and electrode potential, with concurrent
changes in the biolm (for chemical fuel cells, the composition
and structure of catalysts change only very slowly with time
via various degradation reactions, o100 mV h1 in the best
systems); (b) the concentrations of electroactive metabolites
and/or substrate may also be constantly changing with time
and potential (chemical fuel cells supplied with gases can be
tested or operated with very precise and constant fuel/oxidant
supplies); nally (c) the mechanisms of electrode reactions and
operational conditions may be dierent for each dierent type
of MFC, each requiring dierent sampling rates. MFCs can
exhibit stable behaviour in simple time-related voltage and
current experiments using xed resistance load, but long-term
polarization testing requires caution as there is a continual risk
of the presence of the above changes.10,27 In contemporary
MFC studies, pseudo-steady state conditions are generally
present during testing and a minimum of several minutes per
This journal is


c

The Royal Society of Chemistry 2009

data point is necessary. The optimum time for data point

recording needs further discussion by MFC researchers, as it
will inevitably be a function of the system being studied and
can aect numerous parameters.
An ideal polarization curve for a power generation device
(MFC or chemical fuel cell) includes three characteristic
regions located at dierent current ranges (either well segregated, as is typical with chemical fuel cells, or with varying and
signicant levels of overlap, as is more typical with MFCs) as
shown in Fig. 3a and summarised below:
(a) The region of charge transfer overpotentials (Zact,a +
Zact,c) is located at low currents and derives from the slowness
(irreversibility) of the reactions taking place on the surface of
the electrodes. The charge transfer overpotentials depend on
the nature of the electrode materials, catalysts, reactant
activities, electrolyte including any spectator species that
are present, electrochemical mediators, biolm, electrode
microstructure, microbial species and their metabolisms, and
operational conditions such as temperature;
(b) Ohmic overpotentials typically manifest themselves at
intermediate currents in the polarization curve and are caused
by ionic resistances in the electrolyte (especially for MFCs, due
to low ionic concentrations), membrane, and biolm, and by
electronic resistances in the electrodes, current collectors,

Fig. 3 (a) The ideal currentvoltage polarization curve of biological

fuel cell, and (b) the separate currentpotential polarization curves of
cathode and anode where a reference electrode has been introduced
into the system.

Chem. Soc. Rev., 2009, 38, 19261939 | 1931

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

View Article Online

interconnects, and the electronic components between

the measuring instrument and the power generation device
under study;
(c) Mass transport overpotentials result from the changes in
concentration of the reactants (or products) at the interface
between electrode surface region and bulk electrolyte; this
overpotential is prevalent at relatively high current densities
if the reactants cannot be supplied to the electrode reaction
zones at the rate required to sustain the generation of current.
The presence of high concentrations of product species can
also lead to the reduction of transport of reactants to where
they are needed. Mass transfer derived overpotentials are
aected by the geometry and the structure of electrodes and
biolm, the nature of the electrolyte, the metabolites, and the
products that are present.

carbide catalysts in a neutral pH electrolyte, whilst no activity

was observed for ethanol electro-oxidation. Potentiodynamic
methods were used by Manohar et al.31 to compare the
performances of anodes in buer, buersubstrate, and
buersubstratemicrobe systems; the analysis of their results
allowed the determination of the exchange current densities
and Tafel slopes.
3.3 ButlerVolmer equation and Tafel curves
Charge transfer overpotentials are controlled by the rate of
heterogeneous electron transfer, and the kinetics of this
process are described by the ButlerVolmer equation20 when
the reactants are abundant and the current is small enough
that the ohmic and concentration overpotentials are negligible:
I Aio fe

3.2

Electrode polarization techniques

A MFC polarization curve yields the overall fuel cell

performance under specic operating conditions, but fails to
give information on the performances of the individual
electrodes (anode or cathode), hence it is dicult to elucidate
the limiting factor of MFC performance. However, a RE can
be easily introduced in one or more of the MFC chambers so
that it is possible to record the individual potentials of anode
or cathode (as well as the overall cell voltage of the
MFCmodern electrochemical testers can measure the
impedance spectra of all of these simultaneously, see
section 3.5); single electrode polarization curves can, therefore,
be easily obtained as shown in Fig. 3b. The overpotentials of a
MFC are the sum of the (non-linear) overpotentials of the
anode and cathode and the (linear) internal ohmic overpotentials of the MFC. By analysing the potential changes
of the anode and the cathode at varying currents, the
performance limiting factor can often be revealed. For example, a previous study showed that a biocathode showed
inferior performance compared to the anode,28 and that
increasing the eciency or surface area of the cathode should
be targeted to enhance power generation. As it will become
apparent in section 3.5, to simplify analyses and to reveal more
information, polarization plots can be corrected for IRohm
losses i.e. IR-corrected voltages (=Vcell + IRint) can be
plotted against log10 I, where Rohm is the internal ohmic
resistances between the relevant electrodes; an example of this
is found in Fig. 4 of ref. 29.
In a three-electrode conguration, it is possible to obtain the
performance of an individual anode or cathode using a variety
of techniques including galvanodynamic or potentiodynamic
measurements (other relevant techniques are detailed in the
sections below). Galvanodynamic measurements12 have been
used to investigate the cathode performance of cobalt- and
iron-based ORR catalysts in dierent concentrations of
aqueous electrolyte. It was demonstrated that limited buering
capacities of the electrolyte will lead to undesirable decreases
in the cathode performance due to pH changes at the cathode.
Rosenbaum et al.30 used galvanodynamic electrochemical
methods to evaluate the anode performance and electrocatalytic activity with microbial fermentation products;
hydrogen, formate, and lactate can be oxidised by tungsten
1932 | Chem. Soc. Rev., 2009, 38, 19261939

anFZact;c
RT

 e

1anFZact;a

RT
g

where I is the current, A is the electrode active surface area, io

is the exchange current density, a is the charge transfer barrier
(symmetry coecient), n is number of electrons involved in the
electrode reaction, and Zact is the charge transfer overpotential. Typical electrode reactions occur in more than one
elementary step, and there is an overpotential associated with
each step. The io is a fundamental parameter in the rate of
electrooxidation or electroreduction of a chemical species at
an electrode at equilibrium; a large io indicates a fast reaction
(approaching reversible, as found with the hydrogen oxidation
reaction at a Pt electrode), whereas a small io indicates a slow
reaction (often an irreversible electrode reaction, such as the
reduction of oxygen, often irreversible even when employing
Pt electrodes).
The ButlerVolmer equation can be simplied to eqn (4) in
the high overpotential region (4118/n mV, where n is the
number of electrons exchanged), yielding the Tafel
equation:20,25
 
i
4
Zact b log10
io
where i is the current density and b is Tafel slope (mV dec1),
which is an important experimental parameter commonly used
to probe the mechanism of an electrode reaction. Plots of
overpotential against log10i are known as Tafel Plots
(see Fig. 4); io and b are obtained by extrapolation of the
linear region of the curve to Zact = 0.
Exchange currents or exchange current densities in MFCs
have previously been calculated from Tafel plots.12,31,32
Several methods have been proposed with the aim of reducing
charge transfer overpotentials and therefore increasing io:
1. Improved catalysts or biocatalysts, which can considerably decrease activation barriers, have been used to improve
the electrode performance;
2. Modication of the electrodes can signicantly alter the
biolm composition of electrode surface, which can facilitate
an electrode process such as the direct electron transfer
between a protein and the electrode, or can produce selectivity
towards a particular process (especially in EFC studies);32
3. Utilisation of electrode materials with higher specic
surface area (e.g. with a highly porous structure) has the
potential to provide more available reaction sites for electrode
This journal is


c

The Royal Society of Chemistry 2009

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

View Article Online

Fig. 4 A typical Tafel plot for an electrode reaction, from which

exchange current density and Tafel slope can be determined.

reactions and increase the geometric current density. Microbes

may not be able to access the surface areas inside smaller sized
pores, however, ignoring the specic surface area of the
electrode materials (especially with carbon based electrodes,
where these surface areas are easily obtained using gas
absorption/mercury porosimetry measurements) may lead to
a loss of vital information. An example is that a electrode with
higher specic surface area substantially improved performances for biologically produced sulde oxidation;19
4. The rates of chemical reactions involving electroactive
metabolites are generally faster than for reactions involving
biological metabolisms; employing optimised operational
conditions (e.g. raising the temperature to levels optimal for
microbial metabolism and for maximum enzyme reaction
activity and/or increasing the concentrations of reactants)33
may be eective;
5. The evolution of microbial species can be adapted to
allow enhanced electrode performance; analysis of gene
expression and identication of benecial mutations can allow
improvements via genetic engineering and the identication of
more of the electroactive species and more eective biolm
structures.
3.4

Current interruption (CI)

CI techniques have been used extensively to measure the

internal ohmic resistance of chemical fuel cells and more
recently of MFCs.27 As shown in Fig. 5, the basic principle
of the CI is to interrupt the current ow and to observe the
resulting voltage transients. On current interruption, the
ohmic overpotential is separated from other overpotentials
as the former is a near instantaneous process, whereas the
relaxation times for other voltage loss phenomena are
signicantly longer. CI measurements can be conducted using
low cost electronic equipment, and resulting data are easily
interpreted. The primary disadvantage is that very short
duration (o10 ms) measurements of the perturbation to the
system are required for precise and accurate determinations,
and it is possible to overestimate the voltage changes if data
collection is not rapid enough (sample rate is too low).
This journal is


c

The Royal Society of Chemistry 2009

Fig. 5 An ideal voltage transient in a MFC after current interruption.

The MFC is rstly operated at a xed current, and then current is
interrupted at to.

Another disadvantage of using CI is that it is dicult

to distinguish between charge transfer and mass transfer
impedances.
3.5 Electrochemical impedance spectroscopy (EIS)
EIS takes advantage of the large spectrum of time scales over
which dierent processes occur in the system being studied. It
is a powerful tool for examining chemical and physical
processes in solutions, at solidliquid interfaces and at
solidsolid interfaces, as it allows the separation of the
dierent voltage loss phenomena.20,29,34,35 There are two
common graphical representations used in EIS: Nyquist plots
and Bode plots (Fig. 6). The main shortcoming of Nyquist
plots is that they do not show the frequency represented by
each data point (each point being a representation of the
impedance vector in the complex plane at a particular
frequency); Bode plots show this frequency information as
they are plots of the magnitude and phase angle of the
impedance vector versus frequency.
Recently, MFC studies using EIS have been reported.31,36,37
A small amplitude (e.g. 10 mV rms where rms = root mean
squared) alternating current (a.c.sine wave) perturbation
signal is applied to the MFC being tested; small perturbations
are used to ensure non-linear harmonic eects are not interfering with the data collection and also to prevent damage to
the biolm attached to the electrode surface (systems are often
perturbed far from equilibrium in the cyclic voltammetry or
Chem. Soc. Rev., 2009, 38, 19261939 | 1933

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

View Article Online

Fig. 7 Schematic diagram illustrating the need for choosing the

amplitude of the stimulus signal correctly. The diagram shows the
measurement of impedance spectra at dierent MFC d.c. discharge
currents with potentiostatic control, a.c. potential perturbations and
the measurement of the resulting a.c. current responses: (a) nonlinear
response (distortion in impedance spectra); (b) ideal linear response;
and (c) linear response where the current amplitude is too small to
be accurately measured by the instrument (very noisy impedance
spectrum).

Fig. 6 (a) A Nyquist plot and (b) a Bode plot of the impedance
spectrum of a single chamber MFC measured with two-electrode
conguration under open circuit voltage. The frequency range
was 1060.1 Hz with a potential perturbation signal of 10 mV rms.
Rohm was the internal ohmic resistance, Ract, was the charge transfer
resistance.

potential step methods detailed in section 3.6). To avoid

non-linear responses when measuring EIS spectra at dierent
currents or potentials, it is vital that the d.c. polarization
curves are inspected rst so that the correct settings can be
deduced. The amplitude of either the voltage or current
perturbations (when conducting experiments with potential
or current control respectively) must be selected such that
responses will be linear at steady or pseudo-steady state
conditions (Fig. 7) and within the current/voltage ranges of
the electronic equipment used in the test e.g. do not use
a voltage perturbation signal that will result in a current
amplitude (e.g. 10 nA) that is smaller than the current
measuring resolution (e.g. 0.1 mA) of the instrument. In
general, it is best to use the smallest perturbation signal that
still produces noise free data; perturbations of large amplitude
will result in distortions of the EIS spectrum due to non-linear
responses.
Detailed information on the internal ohmic resistances, and
on the charge and mass transfer impedances, can be obtained
on the analysis of correctly collected EIS spectra. The Nyquist
plot and the Bode plots of a single chamber MFC with an air
cathode developed for wastewater treatment are shown in
1934 | Chem. Soc. Rev., 2009, 38, 19261939

Fig. 6. The internal ohmic resistance is often determined by

the high frequency intercept of the curve with the real
impedance axis (Zrethe x-axis).20 The internal ohmic resistances recorded using EIS data are much more accurate than
when using simple single frequency resistance measurements
(1 kHz measurements are commonly employed in cheap
ohmmeters). EIS can, like CI, overestimate internal ohmic
resistances, especially at high d.c. discharge current densities;
however, this overestimation tends to occur at current densities of 40.5 A cm2 (geometric) achievable in H2/O2 chemical
fuel cells, which are unlikely to be achieved with MFCs.
With EIS technique the conductivity of electrode materials
and membranes can be easily measured. The conductivity of
cation-exchange membranes (CEM), such as Naon, can be as
high as 0.1 S cm1, but only when the CEM is in the protonexchange form (H+ conduction). Conductivities of these
membranes are lower when in the Na+/K+/Ca2+ forms since
H+ ions have higher mobilities than metal cations, mainly
because proton conduction involves the Grotthuss
mechanism;38 CEMs are likely be present in MFCs in a form
where a variety of metal cations or other chemical species are
present. Careful selection of materials, MFC congurations,
electrode spacing, and electrolyte concentration will allow
reductions in ohmic losses. Salt-bridge-containing or high
internal resistance (ca. 3.9 MO) systems39 have limited power
generation because most of the electricity being generated is
dissipated as heat (not a desirable outcome: MFCs operating
at ambient temperatures will never be applicable as
co-generation systems). The conductivity of the electrode
material has a practical importance. It is rarely a serious issue
with lab scale MFCs, but low conductivity will lead to major
problems (uneconomic performances) at the pilot scale or with
full scale devices. The thickness and conductivity (an intrinsic
This journal is


c

The Royal Society of Chemistry 2009

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

View Article Online

property) of the electrolyte membrane also aect the internal

ohmic resistance of the energy generation device. Thin
membranes reduce ionic resistance but tend to have a higher
rate of oxygen crossover to anode compartment, which
undesirably reduces coulombic eciency and/or power
production; thin membrane electrolytes might also lead to
an increase in the crossover of other species to the cathode,
which leads to performance limitations i.e. additional mass
transport and/or activation derived losses (catalyst poisoning
or blocking of surface sites for ORR). A radical modication
is achieved through the removal of the electrolyte membrane
the membrane-less system. Cheng et al.40 reported the
maximum power production of a membrane-less MFC
achieved with a 2 cm dimensional spacing between the anode
and cathode; the power decreased with reduced space due to
oxygen crossover from the cathode side to the anode chamber;
the studies suggest that the anode and cathode spacing should
be less than 2 cm for high power generation when a separator
is present in MFCs.
Charge transfer processes occur with smaller time constants
(=RactCdl if modelled by parallel RC circuits in EIS, where
Ract is the charge transfer impedance and Cdl is the capacitance
of the electrodes double layer) when compared to mass
transport (diusion/migration) processes. Processes occurring
with dierent time constants can often be identied using EIS
measurements. For example, the dierence between the charge
transfer parallel RC impedance (kinetic control-semicircle at
medium-high frequencies) can be clearly distinguished from
the mass transfer derived impedances (sloped line at low
frequencies in Fig. 6a). Charge and mass transport related
impedances are common non-ohmic phenomena and require
overpotentials to enable a current ow; the diameter of the
semicircular charge transfer RC-related impedance response in
Nyquist plots will change exponentially with overpotential
(or electrode potential). The capacitance of the interface
between electrode and solution changes when a biolm
attaches to the surface of electrode. Manohar et al.31 found,
using EIS methods, that in the presence of the species MR-1
the OCP of an anode became more negative and the capacitance increased; both of these aected the MFC power output.
To reduce mass transfer impedances in MFCs, the electrode
structure and conguration should be designed to facilitate the
rapid transport of the reactants, products, supporting
electrolyte and substrate in order to avoid excessive concentration-derived polarizations. The accumulation of metabolites in the biolm can hinder microbial activity, whilst
mass transport between the anode and cathode chambers
tends to alter the pH in each and results in further performance decreases. Stirring and forced convection enhances
mass transport and have been shown to improve the power
output of EFCs;41 MFCs exhibit analogous behaviour e.g. the
anode reaction in a sulfur-pollutant-containing wastewater
treatment system was controlled by a diusion process.18
Appropriate stirring or forced convection may lead to denser
biolms in systems operating with a direct electron transfer
mechanism, leading to enhanced anode performances.
For systems operating with indirect electron transfer, the
concentration of electroactive metabolites has an important
eect on current outputs;19,42 higher concentrations of
This journal is


c

The Royal Society of Chemistry 2009

electroactive metabolites may reduce the mass transport

and charge transfer impedances, as well as internal ohmic
resistances by solution conductivity increase. A MFC
combined with a reservoir that is designed for the
accumulation of electroactive metabolites is a technology that
promises high current outputs.
In comparison with CI, EIS is a very sensitive technique
yielding a wealth of information such as kinetic parameters,
determination of the reaction mechanisms, electrolyte and
electrode conductivities, and biolm behaviours. However,
MFC studies, due to their complexity, need to be undertaken
carefully in terms of how they approach EIS data collection
and analysis, especially when equivalent circuit modelling is
used to interpret results and the reaction mechanism of
electrode is not known. Each EIS equivalent circuit element
introduced (resistor, capacitance or inductor) must represent
something in the real system: the tendency to unnecessarily
increase the number of circuit elements, or to replace
capacitors with constant phase elements without justication,
in order to gain better ts to the collected data must be
avoided.
3.6 Cyclic voltammetry (CV)
The determination of the mechanisms of electrode reactions
underlying oxidation or reduction reactions can be achieved
using a variety of electrochemical techniques. The most
common and straightforward technique is CV, which requires
a three-electrode conguration to obtain accurate results.
Fig. 8 shows the important parameters that can be obtained
from CV of a reversible redox couple (Fe(CN)63 + e Fe(CN)64). It is relatively easy to show using CV whether a
chemical system under study is reversible or irreversible: for a
fully electrokinetically reversible couple, the ratio of anodic to
cathodic peak currents = 1 and the potential separations
between the peak potentials = 59.2 mV/n at room temperature, where n is the number of electrons transferred in the

Fig. 8 A classic cyclic voltammogram of a fully reversible redox

couple showing the important parameters: (i) peak potentials; (ii) peak
currents and (iii) the potential dierence DE between the reduction and
oxidation peaks.

Chem. Soc. Rev., 2009, 38, 19261939 | 1935

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

View Article Online

electrode reaction. Compared to reversible systems, quasi- or

irreversible phenomena will exhibit larger separations between
peak potentials and/or one or more peaks that are reduced in
size (the electron transfer process takes more time to respond
to the applied potential in these non-reversible systems). MFC
studies employing CV generally use forward and backward
voltage sweeps with rates in the range of 1100 mV s1.
Multiple peaks in the cyclic voltammograms of bioelectrochemical system may be observed due to multi-step
parallel or consecutive (series) mechanisms, or to the presence
of several dierent redox species.
In protein lm voltammetry, the analysis of the rst
derivative of CV allows the estimation of the potential at
which the rate of increase of the catalytic wave reaches a
maximum; this has been used to reveal that the oxidative and
reductive potential sweep possess two (or more) inection
points in some systems.26,43
In MFC studies, CV experiments have been used extensively
to: (i) investigate the mechanisms of electrode reactions
involving both direct and indirect electron transfer between
the biolm and the electrode; (ii) determine the redox
potentials of the chemical or biological species involved at
the anode or cathode (for a reversible redox couple, the
average of the cathodic and anodic peak gives the reversible
potential for that couple referenced against the RE being
employed); and (iii) to evaluate the performance of the
catalysts being studied.
For anode reactions, only microbial extracellular cytochromes have been proven, to date, to be active for direct
electron transfer mechanisms. However, some CV-based MFC
studies shows characteristics of inter-conversion between
activeinactive states, i.e. inactivation is observed during the
positive potential sweep and reactivation is observed on the
return sweep; this is not an expected behaviour of cytochromes
but is similar with that of enzymes (e.g. hydrogenases) that
have been studied in EFCs.26 The shape of these voltammograms is an initial indication that turnover enzymes or multienzyme systems might be contributing to the anodic current.
CV is a simple technique and results are obtained in a
relatively short time. However, background experiments
(with blank electrolyte) are mandatory for high quality mechanistic studies. When microbial communities are present in
MFC chambers, the peak current and the peak potential
coming from the electrode interface reactions might involve
direct electron transfer and/or indirect electron transfer. For
complex systems, where there are many unknowns (especially
relevant to wastewaters), the following CV testing protocols
are recommended, as a minimum, to aid understanding:
(a) Eects of electrode and electrolyte: Employ a biolmless electrode in un-inoculated or abiotic electrolyte;
(b) Eect of the biolm: Employ a biolm-coated electrode
in the un-inoculated or abiotic electrolyte;
(c) Eect of the substrate: Employ a biolm-less and a
biolm-coated electrode in the un-inoculated or abiotic
electrolyte with the addition of fresh substrate;
(d) Eects of the suspended cells: Employ a biolm-less
electrode in an inoculated electrolyte;
(e) Eect of the metabolites: Employ a biolm-less electrode
in the inoculated electrolyte with consumed substrate(s).
1936 | Chem. Soc. Rev., 2009, 38, 19261939

By comparing the results of these CV experiments (i.e. the

peak currents and the peak potentials at same scan rate using a
same size electrode) the electron transfer mechanisms of the
electrode should be claried to an extent. The collection of
reproducible CV scans is required for rigorous understanding,
especially in systems with low concentrations of electroactive
metabolites.
The characteristics of CV depend on several factors, such as
the electrode surface pretreatment, the rate of the electron
transfer reactions, the chemical and biological species present
and their thermodynamic properties, the concentration of
electroactive species and their rates of diusion and the
sweep rate. It should be noted that many electrode materials
used in MFCs cannot produce reversible electrochemical
reactions even for the classic reversible redox couple
Fe(CN)63/Fe(CN)64. The main reason for this is that the
heterogeneous processes of electrode reactions can be signicantly aected by the microstructure, roughness and function
groups present on the electrode surface. One example is that a
simple electrode will often produce dierent electrochemical
reaction rates before and after polishing, and will show
dierent CV proles. When the accurate determination of
kinetic parameters is investigated, it is essential that the
electrodes and the operational conditions are appropriate.
3.7 Dierential pulse voltammetry (DPV)
DPV is a voltammetric technique with an improved sensitivity
compared to CV and LSV methods. An important limitation
of both CV and LSV is the substantial background levels from
capacitive, non-Faradaic currents. Enhanced discrimination of
Faradaic currents (electron transfer to and from an electrode)
can be obtained using DPV, where the potential perturbation
consisting of small pulses is superimposed upon a staircase
waveform. DPV studies can also provide improved selectivity
for observing dierent redox processes compared with CV
and LSV.
In DPV measurements using sterile carbon electrodes in a
growth medium and with carbon electrodes colonised by
Shewanella, Marsili et al.44 found that a layer of avins was
adsorbed on the electrodes and these acted as external electron
transfer acceptors (especially in older biolms), lower
concentrations of secondary compounds were present in the
electrolyte.
3.8 Chronoamperometry (CA)
CA is an electrochemical technique where the potential of
anode or cathode is controlled (or stepped) and then held
constant with the resulting currents being monitored as a
function of time. Bond et al.45 inoculated a MFC chamber
containing Geobacter sulfurreducens, acetate as the electron
donor, and a graphite electrode held at an oxidising potential
(+200 mV versus Ag/AgCl). A DET process was indicated for
the MFC as the biolm electrode continued, unaected, to
generate an acetate-dependent current when the medium was
replaced with an anaerobic buer lacking nutrients. It is
important to appreciate that the anode potentials inuence
and regulate biolm activity as well as the microbial cell
growth rate. The study of three comparable reactors,
This journal is


c

The Royal Society of Chemistry 2009

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

View Article Online

containing microbial communities that were continuously fed

with acetate, with anode potentials poised at 0, 200 and
400 mV versus Ag/AgCl respectively, led to the conclusion
that the optimal anode was at 200 mV under the conditions
tested.27 In another experiment, a tungsten carbide modied
anode was immersed in a glucose substrate solution that had
been freshly inoculated with heat-treated soil, and data on
microbial growth and oxidation of metabolites as a function of
time were recorded.30 A nal example is where both carbon
bre veil and activated carbon cloth anodes were compared
during a study of the oxidation current for microbially
produced sulde at anode potentials of +200 mV versus
Ag/AgCl; the highly porous activated carbon cloth electrode
gave substantially better performances.18
3.9

Chronopotentiometry (CP)

CP involves the study of potential as a function of time at an

electrode operating with a constant current (the opposite of
CA above). A recent study using this technique investigated air
cathodes coated with polytetrauoroethylene containing
layers.46 The electrode potentials of the layered carbon cloth
electrodes were measured whilst applying a constant current.
The resulting potentials were then plotted as a function of
current densities to evaluate the electrode performances.
A maximum performance was achieved with coulombic
eciencies of up to 42% when a four-layer-coated electrode
was employed.
3.10 Other simple time-related voltage, potential and current
measurements
There are other basic techniques that can be employed to
investigate both MFCs and the individual performances of
electrodes as a function of time without precise control of
potential or current. Information such as voltage versus time
proles, acclimation times, the eects of substrate addition
and/or depletion, electrolyte eects, and the changes in current
on bacterial evolution and enrichment can all be obtained.47,48
The potential of various carbon based anodes as a function
of time in wastewater containing sulfur pollutants was recently
investigated.19 The results indicated that the anodic potential
was primarily controlled by the concentration of microbially
produced sulde in the solution. Another study, that analysed
potential/time curves, showed that biologically produced H2
also had an eect on the behavior of modied Pt anodes.49
These studies demonstrate that the electroactive species
(not only the biolm) should be considered when probing
electrode mechanisms, such species are commonly found in
wastewaters and signicantly aect the electrode performance.
3.11 Rotating-disk electrode (RDE) and rotating ring disk
electrode (RRDE)
Hydrodynamic techniques, such as RDE and RRDE, are
electrochemical experiments where mass transfer is carefully
controlled and are essential for studying the precise kinetic
parameters of electron transfer and for detailed probing
of electrochemical reaction mechanisms including enzyme
electrodes. These techniques have been used in the evaluation
of catalyst or modied electrode performances and for
This journal is


c

The Royal Society of Chemistry 2009

quantifying the number of electrons involved in ORR.20 The

desired cathode process is the full (4e) reduction of oxygen to
water, but partial (2e) ORR will occur on carbon-based
electrodes and results in the production of signicant
quantities of highly reactive hydrogen peroxide (H2O2), which
may aect microbial metabolism. RRDE studies involve a
RDE with an additional ring electrode (poised at a separately
controlled potentialhence the need for a bipotentiostat)
to identify the products (including peroxides) of the electrochemical reactions occurring on the central disk electrode.
These techniques cannot, however, always be used to probe
the electrochemical behaviour of biolms on electrode surfaces
since biolm can be fragile and likely to be destroyed under
conditions of high speed rotation.
Another possibility that will allow experiments with rapid
mass transfer is to design experiments, which can study the
electrochemistry of a single microbe on some form of electrochemically inactive but electronically conductive electrode
surface. The hydrogen oxidation reaction has been studied
on electrode tips containing single Pt nanoparticles of
dened size.50
3.12 Denaturing gradient gel electrophoresis (DGGE)
DGGE is one of the most sensitive electrophoretic techniques
available and has become a routine and essential technique in
the eld of environmental microbiology for characterisation of
microbial population structures and ecology. In many MFC
studies, the development (evolution) of microbial communities
over time have been monitored using DGGE involving
polymerase chain reaction (PCR)amplied 16S rRNA gene
fragments.51 Shewanella and Geobacter species have been
proven to be electroactive microbes for direct electron
transfer, but analysis of the communities in the anode biolm
revealed a diversity of microbes much greater than these
simple iron reducing bacteria alone; many novel microbes
appear to be enriched in the MFCs. The types of microbes
that are enriched in a MFC depend strongly on the substrate
and environmental conditions; no typical consortium has been
detected from any clone library to date.52,53
PCR-DGGE is dependent on intact nucleic acid rather than
viable or non-viable cells, and does not distinguish between
living and dead cells in MFCs. As a consequence, it may result
in misleading results with the complex communities in wastewater treatment systems. Much remains unknown about the
role and the interactions of microbes within both the electrode
biolm and the suspended communities, and the contribution
towards electricity generation of microbial cells at dierent
electrode surface regions. Understanding how microbial
populations and biolm structure evolve with time is essential
if superior MFC power levels are to be achieved.
3.13 Combined electrochemistry-spectroscopy techniques
Spectroelectrochemical techniques facilitate an in vivo
understanding of molecular structures, electron transfer
mechanisms, and the interactions between microbes and
electrodes. Using direct current surface enhanced infrared
(IR) absorption spectroscopy and subtractive interfacial
Fourier transform IR spectroscopy, a recent study investigated
Chem. Soc. Rev., 2009, 38, 19261939 | 1937

View Article Online

the interface between Geobacter sulfurreducens and a gold

electrode.54 The analysis of the IR spectral bands linking the
increasing presence of protein at the interface with time to the
increase in current, thereby demonstrating that the extracellular cytochromes were responsible for the electron transfer
to the gold electrode.

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

3.14 Scanning tunnelling microscopy (STM) and atomic force

microscopy (AFM)
STM and AFM are both high-resolution and powerful types
of microscopy. These are not direct electrochemical techniques
but merit inclusion as they can provide complementary
information on electrode materials. STM or AFM can be
used to image conducting substrates and microbial pili in
MFC studies. For example, the pili on Geobacter species were
recently recognised using these techniques;17 the presence of
such pili raises the possibility of long range DET from one
microbe to the electrode.
The understanding of the function of these nanowires is in
its infancy and the mechanism of how the proteins of these
microbial pili interact with extracellular electron acceptors is
poorly understood. This is an important area for future study,
in which in situ electrochemical (and associated) techniques
will play a vital part. Techniques combining AFM with
electrochemical measurements have recently been developed
for the study of the ionically conductive channels in
proton-exchange membrane,55 and these may have uses in
MFC studies.
3.15

Other techniques worth consideration

Alternative electroanalytical techniques that aid in the

investigation of MFCs in the future and to further improve
the fundamental understanding of the processes occurring are:
(i) square wave voltammetry, which has detection limits as low
as 108 mol dm3 when used as an electroanalytical technique
to determine the nature and concentrations of electroactive
metabolites at trace levels;20 (ii) electrochemical quartz crystal
microbalance measurements can be used as a non-destructive
on-line monitoring tool with regards to the formation of
biolms, and has been used to obtain information on the
biolmelectrode interface;56 (iii) a.c. voltammetry is receiving
increasing interest in the electrochemical community;
(iv) microelectrode techniques may be useful for in situ
investigations of the conductive pili; and (v) electrochemistry
combined with other techniques, such as microarrays,
microscopy, or uorescent in situ hybridization experiments,
may provide a great deal of insight about the behaviour of the
DNA, microbes, and other electroactive species that are
present in the systems being studied.

4. A summary of the key points

To date, the commercial viability of MFCs will rely on side
benets and not purely on the generation of electrical power.
An ever expanding range of cross-disciplinary techniques
are now being employed to study MFCs with the aim of
enhancing both fundamental understanding and innovation
in the eld. Each of the electrochemical/electoanalytical
techniques described in this review provides a range of
1938 | Chem. Soc. Rev., 2009, 38, 19261939

valuable information that is useful in the study of MFCs.

More details about these techniques can be obtained from the
classical electrochemistry literature e.g. ref. 20. The following
summarises some of the key points from the main text.
1. There are various biological and chemical reactions
occurring in MFCs, but only a few steps are directly involved
in electron transfer processes of the electrode reaction.
Background experiments should be conducted to identify
and clarify the electrochemical reaction mechanisms, and the
eects of the electrode materials, biolm, substrate and metabolites. The spectator ions/chemicals present in MFC systems
cannot be ignored and will fundamentally aect any electrocatalysts that are present. This is especially true of species such
as HS, Cl, HPO42, and HCO3/CO32.
2. Experiments that measure reproducibility and repeatability
(such as running identical MFCs in parallel and at dierent
times) should be conducted.
3. The inclusion of a reference electrode will allow the
routine (and often simultaneous) collections of anode, cathode
and whole cell potentials at dierent MFC current outputs.
Modern equipment will also allow the simultaneous collection
of the impedance spectra at steady or pseudo-steady state
conditions for each of the above.
4. With the availability of low cost electronic loads or
battery test systems, a constant potential (potentiostatic) or
a constant current (galvanostatic) measurement may be
preferable when studying MFC discharge performances.
5. Each study must explicitly dene the use of the term
internal resistance. This should refer only to internal ohmic
resistances. Ideally, the collection of IRohm corrected voltage
data should be routine.
6. Keep the analysis of impedance spectra as simple as
possible, and ensure that each element (e.g. capacitor) that is
introduced in equivalent circuit modelling is fully justied by
the system under study, i.e. each circuit element should
represent an identied parameter.
7. Carbon materials from dierent suppliers/sources will not
be the same. The surface chemistry (specic surface area,
surface functionalities and morphology) of the electrode
materials should be evaluated because they have an eect on
biolm and/or on other species.
8. Oxygen reduction on carbon-based electrodes will
produce peroxide species under neutral pH conditions, with
2e oxygen reduction being predominant on carbon surfaces.
The eect of peroxide on the system performance (and durability)
should be taken into account or even studied.

Acknowledgements
This work was supported by the Engineering and
Physical Sciences Research Council as part of the UKs
Supergen5 Biological Fuel Cells Consortium programme
(contract EP/D047943/1).

References
1 M. C. Potter, Proc. R. Soc. Lond. B. Biol. Sci., 1911, 260276.
2 R. A. Bullen, T. C. Arnot, J. B. Lakeman and F. C. Walsh,
Biosens. Bioelectron., 2006, 21, 20152045.

This journal is


c

The Royal Society of Chemistry 2009

Published on 07 April 2009. Downloaded by Mahidol University on 25/09/2015 07:21:12.

View Article Online

3 B. H. Kim, I. S. Chang and G. M. Gadd, Appl. Microbiol.

Biotechnol., 2007, 76, 485494.
4 L. T. Angenent, K. Karim, M. H. Al-Dahhan, B. A. Wrenn and
R. Domiguez-Espinosa, Trends Biotechnol., 2004, 22, 477485.
5 K. Rabaey and W. Verstraete, Trends Biotechnol., 2005, 23, 291298.
6 U. Schroder, Phys. Chem. Chem. Phys., 2007, 9, 26192629.
7 R. A. Rozendal, H. V. M. Hamelers, K. Rabaey, J. Keller and
C. J. N. Buisman, Trends Biotechnol., 2008, 26, 450459.
8 D. R. Lovley, Nat. Rev. Microbiol., 2006, 4, 497508.
9 H. Rismani-Yazdi, S. M. Carver, A. D. Christy and
O. H. Tuovinen, J. Power Sources, 2008, 180, 683694.
10 B. E. Logan, B. Hamelers, R. Rozendal, U. Schroder, J. Keller,
S. Freguia, P. Aelterman, W. Verstraete and K. Rabaey, Environ.
Sci. Technol., 2006, 40, 51815192.
11 R. OHayre, S.-K. Cha, W. Colella and F. B. Prinz, Fuel Cell
Fundamentals, Wiley, New York, 2006.
12 F. Zhao, F. Harnisch, U. Schroder, F. Scholz, P. Bogdano and
I. Herrmann, Environ. Sci. Technol., 2006, 40, 51935199.
13 F. Zhao, F. Harnisch, U. Schroder, F. Scholz, P. Bogdano and
I. Herrmann, Electrochem. Commun., 2005, 7, 14051410.
14 A. E. S. Sleightholm, J. R. Varcoe and A. R. Kucernak,
Electrochem. Commun., 2008, 10, 151155.
15 P. Clauwaert, D. van der Ha, N. Boon, K. Verbeken, M. Verhaege,
K. Rabaey and W. Verstraete, Environ. Sci. Technol., 2007, 41,
75647569.
16 H. J. Kim, H. S. Park, M. S. Hyun, I. S. Chang, M. K. and
B. H. Kim, Enzyme Microbiol. Technol., 2002, 30, 145152.
17 G. Reguera, K. D. McCarthy, T. Mehta, J. S. Nicoll,
M. T. Tuominen and D. R. Lovley, Nature, 2005, 435, 10981101.
18 F. Zhao, N. Rahunen, J. R. Varcoe, A. Chandra, C. AvignoneRossa, A. E. Thumser and R. C. T. Slade, Environ. Sci. Technol.,
2008, 42, 49714976.
19 F. Zhao, N. Rahunen, J. R. Varcoe, A. J. Roberts, C. AvignoneRossa, A. E. Thumser and R. C. T. Slade, Biosens. Bioelectron.,
2009, 24, 19311936.
20 A. J. Bard and L. R. Faulkner, Electrochemical Methods:
Fundamentals and Applications, John Wiley & Sons, New York,
2nd edn, 2000.
21 C. I. Torres, A. K. Marcus and B. E. Rittmann, Biotechnol.
Bioeng., 2008, 100, 872881.
22 A. Heijne, H. V. M. Hamelers, V. Wilde, R. A. Rozendal and
C. J. N. Buisman, Environ. Sci. Technol., 2006, 40, 52005205.
23 J. C. Binger, R. Ray, B. Little and B. R. Ringeisen, Environ. Sci.
Technol., 2007, 41, 14441449.
24 H. H. Lou and Y. Huang, Encyclopedia of Chemical Processing,
ed. S. Lee, Taylor & Francis, 2005, vol. 2.
25 D. Linden and T. B. Reddy, Handbook of batteries, McGraw-Hill,
2002.
26 K. Vincent, A. Parkin and F. A. Armstrong, Chem. Rev., 2007,
107, 43664413.
27 P. Aelterman, S. Freguia, J. Keller, W. Verstraete and K. Rabaey,
Appl. Microbiol. Biotechnol., 2008, 78, 409418.
28 K. Rabaey, S. Read, P. Clauwaert, S. Freguia, P. L. Bond,
L. L. Blackall and J. Keller, The ISME Journal, 2008, 2, 519527.
29 J. R. Varcoe, R. C. T. Slade, G. L. Wright and Y. Chen, J. Phys.
Chem. B, 2006, 110, 2104121049.

This journal is


c

The Royal Society of Chemistry 2009

30 M. Rosenbaum, F. Zhao, M. Quaas, H. Wul, U. Schroder and

F. Scholz, Appl. Catal., B, 2007, 74, 262270.
31 A. K. Manohar, O. Bretschger, K. H. Nealson and F. Mansfeld,
Electrochim. Acta, 2008, 53, 35083513.
32 J. L. Liu, D. A. Lowy, R. G. Baumann and L. M. Tender, J. Appl.
Microbiol., 2007, 102, 177183.
33 T. Chen, S. C. Barton, G. Binyamin, Z. Gao, Y. Zhang, H. Kim
and A. Heller, J. Am. Chem. Soc., 2001, 123, 86308631.
34 E. Barsoukov and J. R. Macdonald, Impedance Spectroscopy:
Theory, Experiment, and Applications, John Wiley & Sons, Inc.,
Hoboken, New Jersey, 2005.
35 F. Zhao, X. E. Wu, M. K. Wang, Y. Liu, L. X. Gao and
S. J. Dong, Anal. Chem., 2004, 76, 49604967.
36 Z. He, N. Wagner, S. D. Minteer and L. T. Angenent, Environ. Sci.
Technol., 2006, 40, 52125217.
37 B. E. Logan, S. Cheng, V. Watson and G. Estadt, Environ. Sci.
Technol., 2007, 41, 33413346.
38 M. Saito, K. Hayamizu and T. Okada, J. Phys. Chem. B, 2005,
109, 31123119.
39 J. K. Jang, T. H. Pham, I. S. Chang, K. H. Kang, H. Moon,
K. S. Cho and B. H. Kim, Process Biochem., 2004, 39, 10071012.
40 S. Cheng, H. Liu and B. E. Logan, Environ. Sci. Technol., 2006, 40,
24262432.
41 Y. Kamitaka, S. Tsujimura, N. Setoyama, T. Kajino and K. Kano,
Phys. Chem. Chem. Phys., 2007, 9, 17931801.
42 I. Ieropoulos, J. Greenman, C. Melhuish and J. Hart, J. Power
Sources, 2005, 145, 253256.
43 S. Srikanth, E. Marsili, M. C. Flickinger and D. R. Bond,
Biotechnol. Bioeng., 2007, 99, 10651073.
44 E. Marsili, D. B. Baron, I. D. Shikhare, D. Coursolle,
J. A. Gralnick and D. R. Bond, Proc. Natl. Acad. Sci. U. S. A.,
2008, 105, 39683973.
45 D. R. Bond and D. R. Lovley, Appl. Microbiol. Biotechnol., 2003,
69, 15481555.
46 S. Cheng, H. Liu and B. E. Logan, Electrochem. Commun., 2006, 8,
489494.
47 G. C. Gil, I. S. Chang, B. H. Kim, M. Kim, J. K. Jang, H. S. Park
and H. J. Kim, Biosens. Bioelectron., 2003, 18, 327334.
48 J. R. Kim, B. Min and B. E. Logan, Appl. Microbiol. Biotechnol.,
2005, 68, 2330.
49 J. Niessen, F. Harnisch, M. Rosenbaum, U. Schroder and
F. Scholz, Electrochem. Commun., 2006, 8, 869873.
50 S. Chen and A. R. Kucernak, J. Phys. Chem. B, 2005, 108,
1398413994.
51 J. Lee, N. T. Phung, I. S. Chang, B. H. Kim and H. C. Sung,
FEMS Microbiol. Lett., 2003, 223, 185191.
52 P. Clauwaert, P. Aelterman, T. H. Pham, L. Schamphelaire,
M. Carballa, K. Rabaey and W. Verstraete, Appl. Microbiol.
Biotechnol., 2008, 79, 901913.
53 B. E. Logan and J. Regan, Trends Microbiol., 2006, 14, 512518.
54 J. P. Busalmen, A. Esteve-Nunez, A. Berna and J. M. Feliu,
Angew. Chem., Int. Ed., 2008, 47, 48744877.
55 X. Xie, O. Kwon, D. M. Zhu, T. Nguyen and G. Lin, J. Phys.
Chem. B, 2007, 111, 61346140.
56 D. E. Nivens, J. Q. Chambers, T. R. Anderson and D. C. White,
Anal. Chem., 1993, 65, 6569.

Chem. Soc. Rev., 2009, 38, 19261939 | 1939