Article

Recent progress and continuing challenges in bio-fuel cells. Part I: Enzymatic cells

School of Engineering Sciences, University of Southampton, Highfield, Southampton, Hants, UK.
Biosensors & Bioelectronics (Impact Factor: 6.41). 03/2011; 26(7):3087-102. DOI: 10.1016/j.bios.2011.01.004
Source: PubMed

ABSTRACT

Recent developments in bio-fuel cell technology are reviewed. A general introduction to bio-fuel cells, including their operating principles and applications, is provided. New materials and methods for the immobilisation of enzymes and mediators on electrodes, including the use of nanostructured electrodes are considered. Fuel, mediator and enzyme materials (anode and cathode), as well as cell configurations are discussed. A detailed summary of recently developed enzymatic fuel cell systems, including performance measurements, is conveniently provided in tabular form. The current scientific and engineering challenges involved in developing practical bio-fuel cell systems are described, with particular emphasis on a fundamental understanding of the reaction environment, the performance and stability requirements, modularity and scalability. In a companion review (Part II), new developments in microbial fuel cell technologies are reviewed in the context of fuel sources, electron transfer mechanisms, anode materials and enhanced O(2) reduction.

Full-text

Available from: Frank C. Walsh, Apr 12, 2016
Our reference: BIOS 4303 P-authorquery-v8
AUTHOR QUERY FORM
Journal: BIOS
Please e-mail or fax your responses and any corrections to:
E-mail: corrections.esch@elsevier.thomsondigital.com
Article Number: 4303 Fax: +353 6170 9272
Dear Author,
Please check your proof carefully and mark all corrections at the appropriate place in the proof (e.g., by using on-screen
annotation in the PDF file) or compile them in a separate list.
For correction or revision of any artwork, please consult http://www.elsevier.com/artworkinstructions.
Any queries or remarks that have arisen during the processing of your manuscript are listed below and highlighted by flags in
the proof. Click on the Q
link to go to the location in the proof.
Location in Query / Remark: click on the Q link to go
article
Please insert your reply or correction at the corresponding line in the proof
Q1 Please check the missing reference Liu (2004).
Thank you for your assistance.
Page 1
Please cite this article in press as: Osman, M.H., et al., Biosens. Bioelectron. (2011), doi:10.1016/j.bios.2011.01.004
ARTICLE IN PRESS
G Model
BIOS 4303 1–16
Biosensors and Bioelectronics xxx (2010) xxx–xxx
1
Contents lists available at ScienceDirect
Biosensors and Bioelectronics
journal homepage: www.elsevier.com/locate/bios
Review1
Recent progress and continuing challenges in bio-fuel cells. Part I: Enzymatic cells2
M.H. Osman, A.A. Shah
, F.C. Walsh3
Energy Technology Research Group, School of Engineering Sciences, University of Southampton, University Road, Highfield, Southampton, Hants SO17 1BJ, UK4
5
article info6
7
Article history:8
Received 26 October 20109
Received in revised form
30 November 2010
10
11
Accepted 4 January 201112
Available online xxx
13
Keywords:14
Bio-fuel cells
15
Enzymatic
16
Challenges17
Immobilization18
Materials19
Applications20
abstract
Recent developments in bio-fuel cell technology are reviewed. A general introduction to bio-fuel cells,
including their operating principles and applications, is provided. New materials and methods for the
immobilisation of enzymes and mediators on electrodes, including the use of nanostructured electrodes
are considered. Fuel, mediator and enzyme materials (anode and cathode), as well as cell configura-
tions are discussed. A detailed summary of recently developed enzymatic fuel cell systems, including
performance measurements, is conveniently provided in tabular form. The current scientific and engi-
neering challenges involved in developing practical bio-fuel cell systems are described, with particular
emphasis on a fundamental understanding of the reaction environment, the performance and stability
requirements, modularity and scalability. In a companion review (Part II), new developments in microbial
fuel cell technologies are reviewed in the context of fuel sources, electron transfer mechanisms, anode
materials and enhanced O
2
reduction.
© 2010 Published by Elsevier B.V.
Contents21
1. Introduction .......................................................................................................................................... 00
22
2. The operating principles of a bio-fuel cell ........................................................................................................... 0023
3. Secondary fuel production ........................................................................................................................... 00
24
4. Applications of bio-fuel cells ......................................................................................................................... 0025
5. Biofuel cell designs and configurations .............................................................................................................. 0026
6. Enzymatic fuel cells .................................................................................................................................. 0027
7. Enzyme and mediator immobilization .............................................................................................................. 00
28
7.1. Physical immobilisation of enzymes and mediators ......................................................................................... 0029
7.2. Enzyme immobilisation in polymers......................................................................................................... 0030
7.3. Reconstructed apoenzymes and sol–gels .................................................................................................... 0031
7.4. Nanostructured electrodes ................................................................................................................... 00
32
7.5. Fuel oxidation ................................................................................................................................ 0033
8. Summary and outlook ............................................................................................................................... 0034
References ........................................................................................................................................... 00
35
1. Introduction
36
Bio-fuel cells have been defined, in the broadest sense, as sys-37
tems capable of direct chemical to electrical energy conversion via38
biochemical pathways (Bullen et al., 2006; Katz et al., 2003; Shukla39
et al., 2004). Direct electrochemical conversion is a desirable fea-40
ture since it avoids the thermodynamic limitations associated with
41
combustion, in addition to being more environmentally friendly.
42
The conversion is achieved by coupling an oxidation reaction sup-43
Corresponding author. Tel.: +44 23 8059 8520; fax: +44 23 8059 3131.
E-mail address: A.Shah@soton.ac.uk (A.A. Shah).
plying electrons at the anode with a reduction reaction utilizing 44
electrons at the cathode. These two reaction are electronically sep- 45
arated inside the system to force electrons to flow through an 46
external circuit, while ion movement inside the system maintains 47
charge balance and completes the electrical circuit (see Fig. 1 for an 48
example). 49
Conventional inorganic fuel cells such as the
50
polymer–electrolyte, direct-methanol and solid-oxide systems
51
(Larminie and Dicks, 2003) rely on expensive rare metal catalysts 52
and/or operate on reformed fossil fuels. In bio-fuel cells (BFCs), the 53
chemical reactions are driven by diverse and abundant bio-fuels 54
and biological catalysts. The production/consumption cycle of 55
bio-fuels is considered to be carbon neutral and, in principle, more 56
0956-5663/$ see front matter © 2010 Published by Elsevier B.V.
doi:10.1016/j.bios.2011.01.004
Page 2
Please cite this article in press as: Osman, M.H., et al., Biosens. Bioelectron. (2011), doi:10.1016/j.bios.2011.01.004
ARTICLE IN PRESS
G Model
BIOS 4303 1–16
2 M.H. Osman et al. / Biosensors and Bioelectronics xxx (2010) xxx–xxx
Nomenclature
Abbreviations
ABTS 2,2
-azinobis (3-ethylbenzothiazoline-6-sulfonate)
diammonnium salt
AlcDH alcohol dehydrogenase
AldDH aldehyde dehydrogenase
BFC biofuel cell
BOD bilirubin oxidase
CDH cellobiose dehydrogenase
CF carbon fiber
CoTMPP cobalt tetramethylphenylporphyrin
DET direct electron transfer
EFC enzymatic fuel cell
FAD flavin adenine dinucleotide
FDH fructose dehydrogenase
GDH glucose dehydrogenase
GOx glucose oxidase
HQS 8-hydroxyquinoline-5-sulfonic acid
KB Ketjen black
CP carbon paper
LDH lactate dehydrogenase
MEA membrane-electrode assembly
MEC microbial electrolysis cell
MFC microbial fuel cell
MP microperoxidase
MWCNT multi-walled carbon nanotubes
NAD nicotinamide adenine dinucleotide
NR neutral red
OCV open circuit voltage
PANI polyaniline
PBS phosphate buffer solution
PLL poly-l-lysine
PPy polypyrrole
PQQ pyrroloquinoline quinone
PVP polyvinylpyrrolidone
SCC short circuit current
SWCNT single wall carbon nanotube
TBAB tetrabutylammoniumbromide
Symbols
E
0
formal potential (V)
j current density (mA cm
2
)
K
M
Michaelis constant (mol L
1
M)
P power density
R resistance ()
[S] substrate concentration (M)
E
cell
cell voltage (V)
v enzymatic reaction rate
sustainable than that of conventional fuel cells (Lovley, 2006).
57
Moreover, biocatalysts could offer significant cost advantages58
over traditional precious-metal catalysts through economies of59
scale. The neutral pH and low temperature of operation represents
60
further advantages (Bullen et al., 2006; Shukla et al., 2004).
61
This review considers major developments in enzymatic and62
microbial fuel cells over the past five years. Earlier developments63
will be reviewed briefly to provide context. For more detailed64
reviews of this work, the reader is referred to Bullen et al. (2006);65
Davis and Higson (2007); Kim et al. (2006). The review is divided66
into two parts, the present part focussing on enzymatic systems and67
Part II on microbial systems (Osman et al., 2010). An introduction
68
to biofuel cells is provided in the next section.69
2. The operating principles of a bio-fuel cell 70
Almost all biochemical processes are catalyzed by enzymes. A 71
group of these proteins, oxidoreductases, are responsible for reac- 72
tions involving electron transfer, and are the most commonly used 73
enzymes. Different subclasses of oxidoreductases are defined based 74
on the type of substrate they act on, as well as the reaction mech- 75
anism (e.g. dehydrogenases, oxidases, and peroxidases). 76
Bio-fuel cells can be classified according to the biocatalyst. Sys- 77
tems using specific isolated enzymes for at least part of their 78
operation are known as enzymatic fuel cells (EFCs), while those 79
utilizing whole organisms containing complete enzyme pathways 80
are know as microbial fuel cells (MFC). A third, intermediate group 81
based on organelles, namely mitochondria, has recently emerged 82
(Arechederra et al., 2009). Several operational differences between 83
these bio-fuel cell types can be identified immediately. Isolated 84
enzymes are substrate specific, while the diverse enzyme contents 85
of whole organisms can be used for a wide range of fuels. Moreover, 86
a complete breakdown of the organic fuel to carbon dioxide and 87
water is usually only possible with several reaction steps (several 88
enzyme catalysts). This is more easily achieved in MFCs, though it
89
also can be achieved in EFCs with an appropriate combination and 90
cascading of specific enzymes (Arechederra and Minteer, 2009). 91
A short lifetime is an inherent characteristic of enzymes, even in 92
their natural environment (Kim et al., 2006). This drawback is not 93
as severe in MFCs since the organisms are able to regenerate the 94
required enzymes as part of their natural functioning. These living 95
systems are also able to grow and adapt, a common and advanta- 96
geous phenomenon observed in MFCs (Kim et al., 2007). Fishilevich 97
et al. (2009) recently developed a microbial fuel cell in which GOx 98
was displayed on the surface of yeast in the anode compartment,
99
with glucose as the fuel and methylene blue as a mediator. The 100
use of micro-organism display in this manner opens up the pos- 101
sibility of self-regenerating enzyme systems from the continuous 102
growth and expression of enzymes on organism surfaces (Boder 103
and Wittrup, 1997). 104
The microorganism used is either a specific isolated species or 105
a mixed culture. It can be applied directly on the electrodes or 106
used in a suspension, or else the system may be inoculated with
107
a mixed culture in a nutrient solution under specific conditions
108
that will allow it to form a biofilm on the electrode. Due to the 109
living nature of organisms, MFC systems have an initial transient 110
operating period of bacterial growth and adaptation to the elec- 111
tron transport mechanism (to and from the electrodes). EFCs, on the 112
other hand, have a faster response time due to the simpler chemical 113
pathways involved.
114
Ideally, the cell voltage for a BFC is independent of the current 115
drawn. In practice, the reversible cell voltage is not realized even 116
under open-circuit (zero current) conditions due to a number of 117
losses incurred when the cell is operated. The difference between 118
actual cell voltage (V
cell
) and the theoretical reversible cell voltage 119
for the overall cell reaction (E
cell
) at a generated current density j 120
(current I divided by the cross-sectional area of the electrodes, A), is
121
termed the overvoltage. As depicted in Fig. 2, there are three major 122
losses that contribute towards the overvoltage (or ‘overpotential’, , 123
for a single electrode): activation overpotentials, ohmic losses and 124
mass-transport (concentration) overpotentials (Clauwaert et al.,
125
2008). The cell voltage at zero current (open-circuit voltage, OCV) 126
E
OCV
can also deviate from E
cell
as a result of internal currents and 127
fuel crossover. At steady state, and assuming spatially distributed
128
reactants, the cell voltage can be approximated as follows:
129
V
cell
= E
cell
j
i
i
l
i
act
conc
(1) 130
Page 3
Please cite this article in press as: Osman, M.H., et al., Biosens. Bioelectron. (2011), doi:10.1016/j.bios.2011.01.004
ARTICLE IN PRESS
G Model
BIOS 4303 1–16
M.H. Osman et al. / Biosensors and Bioelectronics xxx (2010) xxx–xxx 3
Fig. 1. An example of a biofuel cell with oxygen reduction at the cathode. Oxidation of the substrate is catalysed by the enzymes/bacteria (preferably immobilised on the
electrodes), releasing protons and electrons. The electrons released are either transferred directly to the electrode or are transferred via redox mediators, M. Oxygen reduction
at the cathode can take place directly on the electrode or via enzymes/bacteria, possibly facilitated by mediators, N. The mediators can be freely suspended or immobilised
on the electrode to enhance electron transfer.
where the second term on the right-hand side represents ohmic
131
losses and the final two terms denote the activation and concen-
132
tration overpotentials, respectively (sums of the contributions from
133
the two electrodes). The reversible potential E
cell
can be calculated134
from the Gibbs free energy change for the anodic and cathodic135
reactions.136
At low currents, activation (charge transfer) losses dominate;
137
they arise from the energy barrier to charge transfer, from the
138
mediator or bacteria/enzyme to the electrodes. These overpo-139
tentials (separate for the two electrodes) can be approximated140
if expressions for the reaction rates are known, e.g. a Tafel’s or
141
Bultler–Volmer’s relation. Activation losses can be reduced by142
improving the electrode catalysis, increasing the electrode surface143
area, and by optimising the operating conditions (e.g. temperature144
and pH).145
Current density j / A m
2
Power density p / W m
2
Cell voltage V
cell
/ V
p
max
E
cell
E
OCV
Internal currents/fuel crossover
Activation losses
Ohmic
losses
Concentration
losses
Fig. 2. Typical variations of the cell voltage and power of an operating fuel cell
with current density. The major losses of cell voltage and the approximate ranges
of current density in which they occur are indicated.
Ohmic losses are due to the resistance to charge transport
146
through the various components in the cell, including contact resis- 147
tances. They include both ionic and electronic resistances through
148
the current collectors, electrolytes, membrane and electrodes, as 149
well as the interfaces between these components. Assigning a char- 150
acteristic resistivity
i
and thickness l
i
to each component i, the 151
ohmic losses may be approximated using Ohm’s law, as in Eq. (1).
152
To keep ohmic losses to a minimum, the membrane must possess a
153
low resistance, the gap between the electrodes should be optimal 154
and the components must be well contacted. The solution conduc- 155
tivity can also be increased by varying its composition, but this must
156
not affect the functioning of the bacteria/enzymes. 157
Concentration losses are caused by resistance to mass transport, 158
leading to large concentration gradients, notably in the vicinity of 159
the electrode surface. These losses tend to dominate at high cur- 160
rent densities. They can be lowered by ensuring that the solutions 161
are well-mixed (e.g. by stirring or recirculation) or, in the case of 162
an air-breathing cathode, that the ingress of O
2
is not severely 163
restricted. The electrical power density, P of a BFC is defined as
164
the product of the cell voltage and the generated current density: 165
P = IV
cell
= V
2
cell
/R
ext
, where R
ext
is a known, fixed external resis- 166
tance. The power density can be calculated by normalising the 167
power with respect to the electrode cross-sectional area or the elec- 168
trode volume. A typical profile for the power density, p = P/A = jV
cell
, 169
as a function of current density is shown in Fig. 2. 170
One of the most important measures of performance of a BFC
171
is the coulombic efficiency, which is defined as ratio of coulombs 172
transferred from the substrate to the anode, to the theoretical max- 173
imum coulombs produced if all of the substrate is oxidized (×100%) 174
(Liu et al., 2005a). The major causes of reduced coulombic efficiency 175
are (a) the occurrence of alternative reactions that do not result 176
in current production; (b) build-up of biomass; and (c) crossover 177
of the substrate or mixing of the anodic and cathodic reagents, a
178
particular problem in membrane-less systems (Clauwaert et al., 179
2008).
180
The operating voltage of a fuel cell has an upper limit dictated
181
by the difference in potential between the oxidant and reductant 182
Page 4
Please cite this article in press as: Osman, M.H., et al., Biosens. Bioelectron. (2011), doi:10.1016/j.bios.2011.01.004
ARTICLE IN PRESS
G Model
BIOS 4303 1–16
4 M.H. Osman et al. / Biosensors and Bioelectronics xxx (2010) xxx–xxx
Enzyme
Cofactor
Apo-
enzyme
Redox
relay
Electrode
e
e
R
R
R
Fuel Product
ab
N
N
N
NH
O
Os
N
N
N
N
N
N
CH
3
N
N
CH
3
N
N
H
3
C
N
N
H
3
C
H
3
C
2+/3+
4Cl
++
11
85
4
15
O
2
C
Fig. 3. The structure of polyvinylpyridine-[Os(N,N
-dialkylated-2, 2
-biimidazole)
3
]
2+/3+
(Mao et al., 2003). A tris-dialkylated N, N
-biimidazole Os
2+/3+
complex was tethered
to the backbone of a PVP polymer via 13-atom spacers. (b) Illustration of the concept of enzyme reconstitution. An electron relay unit (molecule, redox polymer or nanoparticle)
is linked to an electrode. The cofactor of the enzyme is eliminated and tethered to the relay unit. The apoenzyme is then reconstituted on the relay-cofactor monolayer.
Adapted from Zayats et al. (2008).
and the potential difference between the final electron donor and183
initial acceptor at the electrodes. In bio-fuel cells, this upper limit184
is determined largely by the redox potential of the active sites act-
185
ing on the substrate. If mediators (redox active species) are used186
to shuttle the electrons to/from the electrode, inevitable thermo-187
dynamic losses will occur; the mediators require a potential that
188
is shifted from that of the active site to promote electron trans-189
fer. Mediated electron transfer can, however, yield higher currents190
when the mediator concentration is sufficiently large (Kamitaka191
et al., 2007).
192
Many reports have categorized BFCs into direct and mediated193
electron transfer (DET and MET respectively), with often differ-194
ing definitions. Systems utilizing non-diffusive mediators that are195
attached along with enzymes on electrodes, or those utilizing mix-196
tures of carbon nanotubes and redox polymers, for instance, can be197
considered both direct and mediated. Other reports have classified198
systems based on the materials and methods used for electrode
199
preparation, such as apoenzyme reconstruction, immobilization200
in redox polymers and the use of nanostructured elements (Kim201
et al., 2006; Willner et al., 2009; Sarma et al., 2009). In this review,202
discussion will be divided according to the methods and purposes.
203
MET usually refers to cases where a mediator is used to enhance204
electron transfer between the electrochemically active part of the205
enzyme and the electrode. Conversely, the transfer of electrons
206
directly from the enzyme to the electrode is termed as DET. The vast207
majority of enzymes are not capable of DET (Sucheta et al., 1993;208
Ghindilis et al., 1997; Ferapontova et al., 2003; Tasca et al., 2008a)209
so that most systems employ a mediator, which is usually enzyme210
specific. For glucose oxidation on glucose oxidase (GOx), examples
211
of mediators are ferrocene monocarboxylic acid, pyrroloquino-212
line quinone (PQQ), methylene blue, ferrocenecarboxaldehyde and213
ferrocenemethanol (Harper and Anderson, 2010). The potential 214
applications of bio-fuel cells are diverse. Non-electrochemical 215
applications of bacterial reactions for the production of hydrogen
216
through fermentation, or methane via methanogens are known 217
technologies. Although these bioreactors may be connected to con- 218
ventional fuel cells for electricity production, either as an external
219
unit supplying the fuel (Ishikawa et al., 2006), or by incorporat- 220
ing the fuel production process with the oxidation reaction on the 221
same anode (Niessen et al., 2005), the biological pathway remains 222
separate from the process of electricity production.
223
3. Secondary fuel production 224
Microbial electrolysis cells (MECs) are similar in configuration to 225
microbial fuel cells but require an electrical energy input to initiate 226
a normally unfavorable reaction producing a secondary fuel. For 227
example, hydrogen can be produced on an anaerobic cathode by 228
the reduction of protons (the product of acetic acid oxidation at 229
the anode) (Liu et al., 2005b; Rozendal et al., 2006; Call and Logan, 230
2008). The electrode reactions can be written as (Liu et al., 2005b): 231
Anode : CH
3
COOH + 2H
2
O 2CO
2
+ 8e
+ 8H
+
(2) 232
Cathode : 8H
+
+ 8e
4H
2
(3) 233
which can be combined with the fermentation of glucose into 234
acetate to produce hydrogen. 235
In another example Cheng et al. (2009) produced methane at 236
a cathode by ‘electromethanogenesis’ combined with the oxida- 237
tion of an organic fuel at the anode. Several reports have suggested 238
a mechanism of methane production in microbial electrochem- 239
ical cells from acetate through acetoclastic methanogenesis, or 240
Page 5
Please cite this article in press as: Osman, M.H., et al., Biosens. Bioelectron. (2011), doi:10.1016/j.bios.2011.01.004
ARTICLE IN PRESS
G Model
BIOS 4303 1–16
M.H. Osman et al. / Biosensors and Bioelectronics xxx (2010) xxx–xxx 5
from the intermediate hydrogen product. The work of Cheng et al.241
(2009) presents preliminary evidence that methane can be pro-242
duced from microorganisms (combined with CO
2
capture). MECs243
that use electricity for the production of a secondary fuel may244
be used with renewable energy systems to generate usable fuels245
that are easily transported and stored. Call and Logan (2008) have246
projected that such systems can provide hydrogen gas at $0.62247
per kg compared to $3.8 per kg by water electrolysis. Another248
class of microbial electrochemical cells not strictly adhering to249
the definition of a fuel cell is that based on phototroph organ-250
isms that use light energy to produce electricity. A two-step251
approach where Rhodobacter sphaeroides converts sunlight and an252
organic substrate into hydrogen gas, which is then oxidized at253
a Pt anode, has achieved power densities up to 0.079 mW cm
2
254
(Cho et al., 2008; Rosenbaum et al., 2005) in a single compartment255
cell. In an approach relying solely on light energy, Furukawa et al.256
(2006) designed a miniature fuel cell that replicated the photo-257
synthetic/metabolic processes to provide direct electrical energy258
in light/dark conditions, via alternate conversion between CO
2
,259
H
2
O and electricity. Cao et al. (2008) used an enriched consortium260
of phototrophic bacteria from a wastewater treatment plant in a
261
two-chamber MFC. A maximum power density of 0.265 mW cm
2
262
was obtained. A sediment type MFC using mixed communities263
of photosynthetic and heterotrophic microorganisms, capable of264
power production in both light and dark without the need of265
organic substrate additions, was recently reported by He et al.266
(2009).267
4. Applications of bio-fuel cells268
An immediately obvious area of application is static power gen-269
eration, with microbial fuel cells being the more likely candidate.270
These systems can be fuelled by widely available, carbon-neutral271
complex fuels such as cellulose (Ishii et al., 2008; Niessen et al.,272
2005). Alternatively, they can form the basis for waste treatment
273
systems, combined with energy generation from the organic mat-274
ter found in sewage. In this application area, MFCs will have to275
compete with traditional anaerobic digesters producing methane
276
or hydrogen. At present, this is not possible considering target277
power densities of around 1 kW m
3
for economic competitive-278
ness (Watanabe, 2008; Rabaey and Verstraete, 2005). Despite279
these drawbacks, direct electrical output with high efficiency, low
280
operating temperatures, and good organic treatment efficiency,281
with the possibility of operating on low strength wastewater, are
282
some of the advantages of MFCs (Watanabe, 2008; Rabaey and
283
Verstraete, 2005). Moreover, bio-electrodes can be used for the oxi-284
dation/reduction of specific target substrates (such as nitrate, iron285
and sulfate) in waste removal or metal extraction from minerals286
(He and Angenent, 2006). The operation of MFCs on marine sedi-287
ment to power remote marine instruments has also been explored288
in several reports.
289
For applications on a smaller scale, BFCs operating on high
290
energy density fuels have the potential to power portable electronic
291
devices, though current power densities are still far from the tar-292
get figures of 100 mW (Sakai et al., 2009). In vivo application of293
bio-fuel cells, either for powering small implantable devices or as294
biosensors, are more promising for the short term due to their low295
power requirements. Short life times are, however, a major issue.296
5. Biofuel cell designs and configurations297
The classical design of an BFC is based on two chambers contain-298
ing the anode and cathode, separated by a ion-selective membrane299
(Bullen et al., 2006). Such cells can be operated in either batch or
300
continuous mode. For wastewater treatment (MFCs), an up-flow,301
two-chamber design was developed by He et al. (2005), operating in 302
continuous mode. The system exhibited a high internal resistance 303
of 84 . A membrane-less version was constructed by Jang et al. 304
(2004), with, however, a considerably higher internal resistance of 305
3.9 M. Removing the membrane can lead to higher power outputs 306
but the cells must be carefully designed for high reaction selectiv- 307
ity in order to avoid low coulombic efficiencies (due to transport of 308
oxygen to the anode). For scale up and reduced cost, on the other 309
hand, the concept of a membrane-less, single chamber design is 310
highly attractive. Moreover, the use of a ferricyanide solution and 311
aeration in the cathode compartment are not desirable. Park and 312
Zeikus (2003) developed an MFC with a Mn
4+
graphite anode and air 313
cathode containing an internal, proton-permeable porcelain layer. 314
Liu and Logan (2004) simplified the design by using a carbon-paper 315
air-breathing-cathode (direct O
2
reduction, catalysed by platinum) 316
without a membrane in a tubular arrangement. The cell exhibited 317
a higher power density than an equivalent membrane-containing 318
cell but with a much reduced coulombic efficiency. An alternative 319
arrangement was developed by Rabaey and Verstraete (2005),in 320
which a granular graphite matrix anode was housed in a tubular, 321
sealed membrane covered by a woven-graphite cathode (soaked in
322
a ferricyanide solution). 323
Immobilisation of the enzymes/mediators opens up the possi- 324
bility of single compartment EFCs. There are, however, very few 325
examples of membrane-less or separator-free EFCs. The first single- 326
chamber EFC was developed by Katz et al. (1999b), consisting of 327
two immiscible electrolytes separated by a liquid–liquid interface, 328
allowing DET to take place. GOx apo-enzyme was reconstituted a 329
PQQ-flavin adenine dinucleotide phosphate (FAD) monolayer asso- 330
ciated with an Au electrode (see Section 7). The cathode consisted 331
of an Au electrode onto which a microperoxidase-11 monolayer
332
was assembled and for which cumene peroxide was used as the 333
oxidiser. Ramanavicius et al. (2005) constructed a single-chamber 334
EFC operating with immobilised alcohol dehydrogenase (AlcDH) 335
on a carbon-rod anode and co-immobilized GOx/microperoxidase 336
on a carbon-rod cathode. The power density, around 10 nW cm
2
, 337
was low and the operational half-life was only 2.5 days. A DET, 338
single-chamber H
2
/O
2
cell with hydrogenase at the anode and 339
fungal laccase at the cathode was constructed by Vincent et al.
340
(2005), again, however, with a low power density. A more system-
341
atic selection of the enzymes and electrode materials by Kamitaka 342
et al. (2007) led to a single-chamber, membrane-less fructose/O
2
343
cell capable of power densities on the order of 1 mW cm
2
; fruc- 344
tose dehydrogenase (FDH) was immobilised on a Ketjen–Black (KB) 345
modified carbon paper and multi-copper oxidases were immo- 346
bilised on a carbon paper cathode modified with KB and a carbon
347
aerogel. Coman et al. (2008) instead used cellobiose dehydro- 348
genase (CDH) and lacasse for a glucose/O
2
system, which was 349
capable of only 5 Wcm
2
. More recently, Wang et al. (2009) 350
immobilised GOx (anode) and laccase (cathode) on porous sili- 351
con substrates with pre-deposited carbon nanotubes to form a 352
membrane-less, mediator-free glucose EFC. Again, the power den- 353
sity (1.38 Wcm
2
) was low and decreased by a factor of almost 5 354
after 24 h. 355
6. Enzymatic fuel cells
356
The two major problems in enzyme-based systems are the short 357
lifetime of the enzyme caused by a reduction in its stability when
358
functioning in a foreign environment, and the low power densities
359
resulting from a low electron transfer rate from the enzyme active
360
site to the electrode (Kim et al., 2006). The bulk of the research in 361
enzymatic fuel cells has been directed at enzyme/electrode inte- 362
gration methods that alleviate these problems. The short lifetime
363
(a few hours) is an inherent characteristic of enzymes even in their 364
Page 6
Please cite this article in press as: Osman, M.H., et al., Biosens. Bioelectron. (2011), doi:10.1016/j.bios.2011.01.004
ARTICLE IN PRESS
G Model
BIOS 4303 1–16
6 M.H. Osman et al. / Biosensors and Bioelectronics xxx (2010) xxx–xxx
natural environment, but the lifetime may be increased to a few365
days by immobilization (Kim et al., 2006).366
Enzymes used in fuel cells are of the oxidoreductase family367
(capable of catalysing oxidation and reduction reactions). They can368
be categorised into three groups, according to the type of electrical369
communication (Heller, 1992) or to their associated redox cofac-370
tors. The first group consists of PQQ-dependent dehydrogenases,371
e.g. AlcDH, glucose dehydrogenase and glycerol dehydrogenase.372
Each is structurally different, but most have multiple metal cen-373
ters and the coenzyme PQQ is bound to the enzyme. The second374
group includes those with a nicotinamide adenine dinucleotide375
(NADH/NAD
+
) or nicotinamide adenine dinucleotide phosphate376
(NADPH/NADP
+
) cofactor, e.g. glucose dehydrogenase and AlcDH.377
In this case, the redox center is usually loosely bound and may378
diffuse away. This allows the enzyme to transfer electrons to the379
electrode by the diffusing center, although the diffusing enzyme380
site may be lost, especially in continuous flow systems. Covalent381
linking of such enzymes needs to maintain a flexible link, allow-382
ing reversible movement between the protein structure and the383
electron acceptor. Enzymes in the third category have a tightly384
bound FAD redox cofactor that is buried deep inside the pro-
385
tein structure, which makes extraction of the electrons difficult.386
The most commonly used enzyme, GOx, belongs to this group. In387
an aqueous solution, the redox potential of FAD at the enzyme388
active site is negative, making it ideal for the anode side of a389
biofuel cell if DET can be achieved. However, systems employ-390
ing GOx are typically mediated, although it is possible to achieve391
DET using nanostructured electrodes as discussed later (Xiao et al.,392
2003; Zayats et al., 2005; Cai and Chen, 2004; Patolsky et al.,393
2004).394
For biological cathodes, the enzymes are typically multi-copper
395
oxidases, which are capable of a four-electron reduction of O
2
to396
water and have a high specificity towards this reaction (Solomon397
et al., 1996). Examples include plant and fungal laccases (Chen398
et al., 2001) and BOD (Mano et al., 2003). Laccases are generally399
employed under slightly acidic conditions, while BOD has activity400
in more alkaline media, which allows it to be used at neutral pH.401
Cytochrome oxidase and cytochrome c have also been employed. In402
the case of H
2
O
2
reduction, microperoxidase (Willner et al., 1998a;403
Katz et al., 1999b) and horseradish peroxidase (Pizzariello et al.,
404
2002) are commonly used as enzymes.405
7. Enzyme and mediator immobilization406
Immobilisation of the enzyme can has several advantages,407
including isolation of the enzyme for reaction, increased selectiv-408
ity, improved mass transfer and long-term stability (Cao, 2005). It409
also has the advantage of separating the enzyme from the mixture410
containing the substrate, allowing for more modular cell designs.411
On the other hand, it can affect the stability and/or activity of the412
enzyme, it can introduce additional mass-transfer limitations on413
the substrate and it involves additional costs. Stability is clearly
414
a key consideration. The stability of the immobilised enzyme will415
depend on the nature and strength of the bonds to the support416
material, the conditions required for immobilisation, the degree of417
confinement and the conditions under which the enzyme reactions
418
occur in a functioning electrode. The method of immobilization419
must be selected carefully to avoid denaturing of the enzymes and420
loss of structural freedom required for their activity (Moehlenbrock
421
and Minteer, 2008; Cooney et al., 2008).
422
The main immobilisation techniques for biosensors and EFCs are
423
(Tischer and Wedekind, 2000): physical surface adsorption with424
diffusional mediators, or mediators co-adsorbed with an enzyme;425
entrapment in conducting polymer matrices or gels; wiring or
426
covalent attachment to functionalised polymers; and apoenzyme427
reconstruction. Nano-structured elements can also be used as sub- 428
strates for binding, or incorporated with one of the aforementioned 429
techniques for enhanced electrical conductivity and stability. 430
The simplest method of enzyme immobilization is physical 431
adsorption or entrapment. Enzymes can be adsorbed, for example, 432
onto conductive particles such as carbon black or graphite powder 433
(Pizzariello et al., 2002). The methods are straightforward and cost- 434
effective. If the binding forces (primarily electrostatic) between the 435
enzyme and the support are too weak, however, the enzymes can 436
desorb and contaminate the solution; if they are too strong, denat- 437
uration can occur during the immobilisation process. Entrapment 438
involves the confinement of the enzyme within a polymer matrix, 439
a sol–gel (Kandimalla et al., 2006), a redox hydrogel (Gregg and 440
Heller, 1991) or behind a semi-permeable membrane. The structure 441
must permit a sufficient degree of enzyme movement, while simul- 442
taneously preventing any leaching of the enzyme and/or mediator. 443
Isolated enzymes can be covalently bonded to supports (e.g. 444
porous glass, cellulose, ceramics, and metallic oxides) via different 445
functional groups on the support and enzyme, often in the presence 446
of enzyme inhibitors. Reagents are used to activate the functional 447
groups on the support. The functional groups on the enzyme, which
448
include amino, carboxylic acid and hydroxyl groups, should not be 449
essential for catalytic activity. The conditions for this type of immo- 450
bilization are important since they determine the level of enzyme 451
activity retention. Cross-linking consists of joining enzymes to form 452
three dimensional aggregates via covalent bonding between active 453
groups within the enzymes. The aggregates exhibit low mechanical 454
stability and the retained enzymatic activity can be low using this 455
method (Sheldon, 2007). 456
Covalent bonding and cross-linking are commonly used 457
to immobilise enzymes on self-assembled monolayers (SAMs)
458
(Gooding and Hibbert, 1999). In the context of biosensors, the 459
most studied SAMs are those formed by alkanethiols chemisorbed 460
from solution onto gold surfaces (Dubois and Nuzzo, 1992). Despite 461
the many advantages of these SAMs (simplicity of preparation, 462
densely packed structures and control over functional groups at the 463
monolayer surface), they are prone to instability (Schoenfisch and 464
Pemberton, 1998; Delamarche et al., 1994; Gooding et al., 2003). 465
To improve stability, several research groups have used covalent
466
modification of carbon surfaces via electrochemically reductive
467
adsorption of aryldiazonium salts (Allongue et al., 1997; Saby et al., 468
1997; Kariuki and McDermott, 2001; Brooksby and Downward, 469
2004). The resulting monolayers are highly stable over a wide 470
potential window (Allongue et al., 1997). In recent studies, gold, 471
graphite and glassy-carbon (GC) electrodes were functionalised 472
using aryldiazonium salts bearing carboxylic acid groups (Pellissier
473
et al., 2008; Boland et al., 2009a). Pellissier et al. (2008) grafted a 474
GOx layer on a GC electrode modified using this method, through 475
coupling with peripheral amine groups of the GOx. This enzyme 476
layer was used as an anchoring base onto which a cross-linked 477
enzyme layer was subsequently deposited, before testing the elec- 478
trode using a GC rod counter electrode. The authors demonstrated 479
that these modified electrodes retained much of their activity after
480
6 weeks, while control electrodes prepared by depositing the cross- 481
linker and GOx directly onto the GC had lost all activity within only 482
1 week. 483
Sol–gel glass is produced by the hydrolysis and polycondensa-
484
tion of organometallic compounds (typically silicon alkoxides) at 485
low temperature (Lin and Brown, 1997). Enzymes can be intro- 486
duced during the formation of the sol–gel (the ‘sol–gel process’),
487
leaving them entrapped around siloxane polymer chains within
488
an inorganic oxide network. The final matrix structure can be
489
controlled by the pH, the temperature, the choice of solvent and 490
the choice of catalyst, amongst other considerations. The main 491
advantages of this method in the context of biosensor and biofuel
492
applications are: simplicity of preparation; the ability to control 493
Page 7
Please cite this article in press as: Osman, M.H., et al., Biosens. Bioelectron. (2011), doi:10.1016/j.bios.2011.01.004
ARTICLE IN PRESS
G Model
BIOS 4303 1–16
M.H. Osman et al. / Biosensors and Bioelectronics xxx (2010) xxx–xxx 7
Table 1
Summary of key enzymatic fuel cell developments
Anode Cathode Electrolytes/membrane P
max
(cw cm
2
) V or j at P
max
OCV (V) Remarks Reference
Mitochondria immob.
in modified Nafion on
C-electrode
Air-breathing
Pt-C/membrane
assembly
10 mM, 7.45 PBS, 6 M
NaNO
3
, 100 mM
pyruvate,1gL
1
ADP
0.0315
, 0.024 (average) 0.1
mA cm
2
Air cathode/
membrane assembly
Arechederra et al.
(2009)
AlcDH/AldDH/oxalate
oxidase in modified
Nafion on C-paper
Air-breathing Pt
–C/membrane
assembly
7.15 pH PBS, 6 M
NaNO
3
100 mM
glycerol
1.32 2
mA cm
2
(0.66 V
) Air cathode/membrane
assemble
Arechederra and
Minteer (2009)
GOx /HQS (mediator)
in PPy on Carbon rod
Laccase /ABTS
(mediator) in PPy on
porous carbon tube
5 pH PBS, 10 mM
glucose,N
2
purged,
37
C, separate O
2
solution circulated
inside cathode
0.027 0.25 V 0.41
Brunel et al. (2007)
CDH adsorbed on
graphite
Laccase adsorbed on
graphite
0.1 M citrate buffer, 4.5
pH, 5 mM glucose, air
saturated
5 × 10
3
0.5 V 0.73 Membrane/mediator-
less. Enzyme
desorption causes
current/power loss.
Coman et al. (2008)
5 bilayers of
AuNPs/GDH on three
dimensional ordered
macroporous,
cystamine treated Au
electrode
Similar to anode, using
laccase as catalyst
0.1 M, 6 pH PBS, 5 mM
NADH, 30 mM glucose
0.178 0.226 V 0.32 SCC = 0.752 mA cm
2
Deng et al. (2008)
Cross-linked clusters of
GOx and CNTs on CF
electrode (0.332 cm
2
)
Air breathing Pt –C
cathode (0.332 cm
2
)
Un-buffered, 200 mM
glucose,10mM
benzoquinone
0.12 0.1 V 0.33 MEA assembly. Better
initial performance but
degrades quickly in
buffered solutions due
to cation interference
with proton transport.
Fischback et al. (2006)
GDH on poly(brilliant
cresyl blue)/SWCNT/GC
rod (3 mm diameter)
Cross-linked BOD on
SWCNT on same
carbon electrode
0.1 M PBS, 7 pH, 10 mM
NAD
+
,40mMglucose,
ambient air
0.054 0.5 V 0.73 Membrane-Less. 5% P
loss in first day. 46%
loss in one week
Gao et al. (2007)
GOx ‘wired’ through
PVP-Os complex with
cross-linking on 2 cm
long, 7 m diam. CF
BOD ‘wired’ through
PAA-PVI-Os complex
with cross-linking on
similar electrode
20 mM PBS, 7.24 pH,
0.14 M NaCl, 15 mM
glucose,37
C
0.315 0.46
V Membrane-less.
Commercial enzyme
stock purified before
usage. Operating cell
for 1 week at 0.52 V
lost 6% of power output
per day
Gao et al. (2009)
Au
70
Pt
30
bi-metallic
nanoparticles on inner
surface of carbon tube
(4.4 cm
2
, 1.4 cm diam.)
BOD /ABTS in modified
Nafion on inner surface
of porous carbon tube
(6 mm diam.)
7.4 pH, PBS, 0.7 M
glucose,37
C
0.19, 0.09 (10 mM glucose) 0.52 V, 0.4 V (10 mM glucose) 0.89
Abiotic anode.
Concentric design.
Membrane-less
Habrioux et al. (2009)
FDH adsorbed on
Ketjen’s black
(0.282 cm
2
)
Laccase adsorbed on
carbon aerogel
particles (0.282 cm
2
)
McIlvaine buffer, 5 pH,
200 mM fructose,O
2
saturated, 25
C
0.85 (stirred), 0.39 (unstirred) 0.41 V 0.79 SCC = 2.8 mA cm
2
(stirred), 1.1 mA cm
2
(unstirred). Power
drops to 63% after
12 hours. 4 cells in
series operate 1.8 V
LED for 60 days.
Kamitaka et al. (2007)
GOx /HQS immobilized
in polypyrrole
nanowires (0.15 cm
2
)
BOD /ABTS in
polypyrrole film
(0.35 cm
2
)
7.4 pH, PBS, 15 mM
glucose
0.28 0.15 V 0.35 SCC = 2.9 mA cm
2
and
maximum power
density with 200 nm
diam. 16 m length
nanowires.
Membrane-Less
Kim et al. (2009)
Page 8
Please cite this article in press as: Osman, M.H., et al., Biosens. Bioelectron. (2011), doi:10.1016/j.bios.2011.01.004
ARTICLE IN PRESS
G Model
BIOS 4303 1–16
8 M.H. Osman et al. / Biosensors and Bioelectronics xxx (2010) xxx–xxx
Table 1 (Continued)
Anode Cathode Electrolytes/membrane P
max
(cw cm
2
) V or j at P
max
OCV (V) Remarks Reference
GOx covalently
attached to
3-methylthiophene
(3MT) and
thiophene-3-acetic
acid (T3A) copolymer
BOD covalently
attached to same
copolymer
0.1 M PBS, 7 pH and
either 0.1 M glucose,
1mMN, N, N
,
N
-tetramethyl-p-
phenylenediamine, N
2
saturated or 1 mM
ABTS, O
2
saturated in
either compartment
separated by Nafion
membrane
0.15 0.35 V 0.61 Anodic current
decreased to 50% while
cathodic current
decreased to 75% the
initial values after 1
month
Kuwahara et al. (2009)
Au electrode-
cystamine-PQQ-LDH
monolayer
Au
electrode-cystamine-
microperoxidase
11
Anolyte: 0.1 M tris
buffer, 7 pH, 20 mM
CaCl
2
, 20 mM NAD
+
,
20 mM lactate.
Catholyte 0.1 M PBS, 7
pH, 1 mM H
2
O
2
, ABTS.
Nafion separator.
0.142 0.1
V 0.34
LDH immobilization
carried in presence of
CaCl
2
promoter, NAD
+
,
and lactate found to
increase power by 26%
compared to
immobilization
without.
Lee et al. (2009)
Latex draw chemical
structure GOx
/single-stranded
DNA-wrapped SWCNT
on cystamine
dihydrochloride
treated Au electrode
(0.0314 cm
2
)
Similar immobilization
for laccase
0.1 M PBS, 7 pH,
glucose,O
2
,25
C
0.442 0.46 V 1.5
Cell operated for more
5 days with power in
excess of
0.43 mW cm
2
.
DNA-wrapped SWCNT
found to increase
enzyme loading.
Lee et al. (2010)
Polyethylene glycol
diglycidyl ether) GOx
/SWCNTs in silica gel
BOD /SWCNT in silica
gel
Anolyte: 4 mM
ferrocene methanol,
100 mM glucose.
Catholyte: 8 mM ABTS,
O
2
saturated. Room
temperature. Nafion
separator.
0.12, 0.086 (ambient air) 0.24 V, 0.21 V (ambient air) 0.48 Lim et al. (2007)
Penicillium pinophilum
sourced GOx /PVP-Os
complex, cross-linked,
on 2 cm long, 7 m
diam. CF
Laccase /PVP-Os
complex, cross-linked
on 2 cm long, 7 m
diam. CF
20 mM citrate buffer, 5
pH, 37
C,5mM
glucose
0.28 0.88 V GOx sourced from P.
pinophilum allows
higher power density
at lower fuel
concentration than
tradition A. niger but
unstable at neutral pH.
3% power loss per day
for first 2 weeks.
Mano (2008)
GDH /NAD
+
in Ketjen’s
black on GC (0.07 cm
2
)
BOD in Ketjen’s black PBS, 50 mM glucose,
O
2
saturated
0.052 0.3 V 0.64 SCC = 0.223 mA cm
2
.
Membrane/mediator-
less
Miyake et al. (2009)
AldDH adsorbed on
graphite electrode
AlcOD/microperoxidase-
8 adsorbed on graphite
electrode
50 mM sodium acetate,
6 pH, 100 mM KCl,
2mMethanol
1.5 × 10
3
0.24 Ethanol as substrate for
both half-reactions.
Power decreases to half
initial value after 26 h
of operation
Ramanavicius et al.
(2008)
Page 9
Please cite this article in press as: Osman, M.H., et al., Biosens. Bioelectron. (2011), doi:10.1016/j.bios.2011.01.004
ARTICLE IN PRESS
G Model
BIOS 4303 1–16
M.H. Osman et al. / Biosensors and Bioelectronics xxx (2010) xxx–xxx 9
Table 1 (Continued)
Anode Cathode Electrolytes/membrane P
max
(cw cm
2
) V or j at P
max
OCV (V) Remarks Reference
4 layers of (CF/
poly-l-lysine/ GDH
/diaphorase/NADH/vitamin
K
3
/polyacrylic acid
sodium salt) (1 cm
2
each)
Air-breathing, 2 layer
of (CF sheet/
K
3
[Fe(CN)
6
]/PLL/ BOD)
(1 cm
2
each)
0.1 M, 7 pH, (PBS),
room temperature,
0.4 M glucose.
Electrodes stacked
with cellophane
membrane in a single
assembly.
1.45 0.3 V 0.8 SCC = 11 mA cm
2
. Sakai et al. (2009)
CDH /
polyvinylpyridine- Os
complex/SWCNTs on
graphite rods (3.05 mm
diam.) with
cross-linking
Pt –C (area anode
area)
PBS, 7.4 pH, 37
C,
0.1 M glucose,O
2
purged, non-quiescent
0.157 0.28 V 0.5 Tasca et al. (2008b)
PLL-
K
3
/diaphorase/GDH on
GC (0.07 cm
2
)
Poly-dimethylsiloxane
coated Pt cathode
PBS, 7 pH, 5 mM
glucose, 1 mM NAD
+
,
37
C
0.032 0.29 V 0.55 Current drops to half
initial value after 18 h
Togo et al. (2007)
AlcDH/ AldDH / NAD
+
/
modified Nafion on
polymethylene green
anode
BOD /modified Nafion
on 1 cm
2
CF paper.
Dried then soaked in
Ru(bpy)
3
2+
mediator
7.15 pH, PBS, 1 mM
ethanol, 1 mM NAD
+
,
room temperature
0.39, 0.83 (with Nafion membrane) 0.51, 0.68 (with Nafion) Power increases to a
maximum of
0.46 mW cm
2
then
rapidly drops after 20
days.
Topcagic and Minteer
(2006)
LDH /modified Nafion
with CaCl
2
on CF paper
(1 cm
2
)
Pt –C black Anolyte: 7.15 pH, PBS,
25 mM lactate.
Catholyte: 1 M NaCl,
dissolved O
2
,20
C.
Nafion Separator
0.022 0.85 Testing over 45 days
without any claimed
degradation in
performance
Treu and Minteer
(2008)
Porous
Si-functionalised
SWCNT-GOx
Porous
Si-functionalised
SWCNT-laccase
PBS, 4 mM glucose, air
bubbling, stirred. 5 mm
inter-electrode
distance
1.38 × 10
3
99 mV Lower power density
(0.35 × 19
3
mW cm
2
)
at higher voltage
(0.357 V) when
SWCNTs grown by
chemical deposition
followed by carboxyl
group attachment
rather then
electrophoretic
deposition of
pre-functionalised
SWCNTs.
Wang et al. (2009)
Covalently linked
SWCNT-NAD
+
deposited on classy
carbon. AlcDH
attached to NAD
+
through affinity, and
cross-linked
Thioanaline modified
BOD copolymerized
with thioanaline
capped Pt
nanoparticles on Au
electrode with
thioanaline monolayer
with crosslinking.
0.1 M PBS, 7 pH, 40 mM
ethanol l, O
2
saturated
0.2 0.55
V 0.62 Maximum power at
0.37 mA cm
2
Yan et al. (2009)
GOx / MWCNTs/Nafion
on carbon felt
(0.33 cm
2
)
Air-breathing Pt
cathode
100 mM glucose,
10 mM
1.4-benzoquinone.
Nafion/electrode
assembly
0.077 0.51 V 0.57 Zheng et al. (2008)
Values not explicitly reported, but estimated from graphical results.
Page 10
Please cite this article in press as: Osman, M.H., et al., Biosens. Bioelectron. (2011), doi:10.1016/j.bios.2011.01.004
ARTICLE IN PRESS
G Model
BIOS 4303 1–16
10 M.H. Osman et al. / Biosensors and Bioelectronics xxx (2010) xxx–xxx
the porosity; the chemical and mechanical stability of the gel; and494
negligible swelling (Lin and Brown, 1997; Wang, 1999).495
A method to co-immobilise the enzyme and mediator (designed496
to prevent mediator leaching) was developed by Heller and co-497
workers, who used soluble redox hydrogels to construct biosensors498
and, subsequently, miniature biofuel cells (Chen et al., 2001; Mano499
et al., 2002b; Soukharev et al., 2004; Heller, 2004, 2006). In this500
method the enzyme is complexed with a redox polyelectrolyte501
forming a water soluble adduct, which is cross-linked on the elec-502
trode surface. The cross-linked polymer swells on contact with503
water to form a hydrogel, to which the enzymes are covalently504
bound. The enzymes are electrically connected to the electrode505
by a redox network and are said to be ‘wired’; electron conduc-506
tion is predominantly controlled by collisional electron transfer507
between the reduced and oxidized (transition metal-based) redox508
centers tethered to the polymer backbone. Popular choices for509
the polymer backbone are polyvinylimidazole, polyallylamine and510
polyvinylpyridine (PVP) and the redox centres are typically osmium511
(Os) or ruthenium complexes (Gregg and Heller, 1991). Os com-512
plexes are particularly useful due to the ease with which the redox513
potential can be tuned by chemical modification of the complex
514
(Kavanagh et al., 2009). They can be tethered flexibly to poly-515
mer backbones, improving the electron transfer kinetics between516
the enzyme and electrode (Stoica et al., 2009). The mechanical517
strength of the hydrogels and the electron transfer rate can be518
improved by using spacers that connect the redox-active centers to519
the cross-linked networks. These spacers provide additional flexi-520
bility and improved collisional electron transfer. The lengths of the521
spacers are important; optimally between 8 and 15 atoms (Heller,522
2006; Mano et al., 2005; Mao et al., 2003). The redox potentials of523
the hydrogels are determined by the transition metal ion of their
524
complex and by its ligands, so they can be tailored to a specific525
enzyme/reaction combination (Heller, 2006; Kim et al., 2003). An526
example of a redox polymer structure, developed by Maoetal.527
(2003), is given in Fig. 3(a). The authors tethered a tris-dialkylated528
N,N
-biimidazole Os
2+/3+
complex to the backbone of a PVP poly-529
mer via 13-atom spacers. An order of magnitude increase in the530
apparent electron diffusion coefficient was observed when com-531
pared to a structure without spacers, used earlier by the authors.
532
Furthermore, oxidation of glucose was found to occur at potentials
533
close to the reversible potential of the FAD/FADH
2
centers of the534
enzyme.535
Realising DET using GOx is difficult due to the deeply embed-536
ded nature of the active FAD sites. The same applies for PQQ and537
heme containing enzymes. In an attempt to overcome this issue,538
Willner and co-workers introduced a method based on reconsti-
539
tuting apo-enzymes on functionalised electrodes (Willner et al.,540
1996, 1998a,b; Xiao et al., 2003) (see Fig. 3(b) for an illustration).541
In one example, gold nanoparticles were linked to a gold electrode542
by a dithiol bridge, while amino-FAD was linked to the particles543
(Xiao et al., 2003). The FAD cofactor units were extracted from544
GOx to give the apo-enzyme, which was reconstituted on the FAD-545
functionalised particles. The gold nanoparticles were seen to act
546
as electron relays between the FAD redox site and the electrode.547
Similarly, PQQ-dependent GDH was electrically wired by the recon-548
stitution of apo-GDH on PQQ-functionalised nanoparticles (Zayats549
et al., 2005). Patolsky et al. (2004) instead reconstituted apo-GOx on
550
FAD units linked to the ends of single wall carbon nanotubes (SWC-551
NTs) assembled on a gold electrode, motivated by the efficient DET552
between SWCNTs and absorbed GOx redox active sites (Guiseppi-
553
Elie et al., 2002; Cai and Chen, 2004). The authors deduced that the
554
SWCNTs behaved as electrical contacts between the active site of
555
the enzyme and the electrode. Such electrodes, using either single-556
walled or multi-walled CNTs (Ivnitski et al., 2006; Zhang et al.,557
2003) display good stability and sensitivity (Cai and Chen, 2004;
558
Wang et al., 2003; Lin et al., 2004).559
A summary of recent key developments in EFCs are presented 560
in Table 1. They are discussed in detail in the sequel. 561
7.1. Physical immobilisation of enzymes and mediators 562
Tasca et al. (2008a) investigated the direct electron transfer 563
(DET) capabilities of different CDHs adsorbed on a graphite elec- 564
trode in the presence or absence of SWCNTs. SWCNTs were found 565
to increase the electrocatalytic current, the onset of which was 566
shifted to more negative potentials. CDH is composed of a large 567
flavin-associated domain and a smaller heme-binding domain that 568
allows direct electron transfer to the electrode. A membrane-less 569
fuel cell was constructed using Phanerochaete sordida CDH coad- 570
sorbed with SWCNTs, together with a Pt/C cathode. A solution with 571
0.1 M, 4.5 pH citrate buffer containing O
2
and 5 mM lactose was 572
used. The open circuit voltage was 590 mV and a maximum power 573
density of 0.032 mW cm
2
at 430 mV was obtained. CDH was also 574
adsorbed on graphite in a fuel cell with laccase immobilized in a 575
polymer (Coman et al., 2008). The power density (0.005 mW cm
2
) 576
was lower than the SWCNT system of Tasca et al. (2008a), but the 577
cell voltages achieved were slightly higher.
578
In a similar fashion to CDH, FDH contains a heme group that 579
should in principle allow direct electron transfer to the electrode. 580
Previous investigations, however, were not successful in achieving 581
practical currents. Kamitaka et al. (2007) immobilized FDH from 582
Gluconobacter sp. by adsorption on a Ketjen’s black (KB) modified 583
carbon-paper anode that was capable of 4 mA cm
2
. Combined with 584
a laccase biocatalyst from Trametes sp. adsorbed on a carbon aero- 585
gel cathode, a membrane-less bio-fuel cell was constructed and 586
operated at room temperature in an O
2
saturated, 5 pH McIlvaine 587
buffer containing 200 mM fructose. Under stirred conditions, to
588
alleviate the O
2
mass transfer limitation, a maximum power den- 589
sity of 0.85 mW cm
2
at 410 mV was obtained and the open circuit 590
voltage was recorded as 790 mV. The power output decreased to 591
63% of the maximum after 12 h of continuous operation. Under 592
low power conditions, 4 cells connected in series continuously 593
powered a small light-emitting diode for 60 days. Since the cell 594
was operated at low power output for two months, the short life- 595
time was unlikely to be due to enzyme desorption but rather to
596
a loss of activity during continuous operation at a high current
597
density. 598
Though stirring is not usually desired in real applications, it 599
is often used in systems to improve mass transport and increase 600
the power output. Katz et al. (2005) investigated the effects of a 601
constant magnetic field applied parallel to the electrode surface in 602
surface-confined bio-electrocatalytic systems. In the two systems
603
of GOx-FAD-PQQ and lactate dehydrogenase (LDH)/NAD
+
-PQQ, in 604
which the current was limited by mass transport, it was found that 605
the current increased by a factor of three when a magnetic field 606
(0.92 T) was applied. This improvement was brought on by a mag- 607
netohydrodynamic effect, engendering a magnetic force on the ions 608
in solution, and thus decreasing the hydrodynamic layer thickness 609
and increasing the current density.
610
Glucose dehydrogenase (GDH) was later </