Activity of carbon electrodes towards oxygen reduction in acid: A comparative study

Article abstract of DOI:10.1039/c1nj20612e

The oxygen reduction reaction (ORR) has been investigated on a variety of carbon substrates, in acidic conditions, including glassy carbon (GC), edge plane pyrolytic graphite (EPPG), basal plane pyrolytic graphite (BPPG) and boron doped diamond (BDD). Along with these macro-sized substrates, a selection of carbon nanotube modified BPPG electrodes have also been examined. These include single walled (SWCNT), bamboo multi walled (B-MWCNT) and hollow multi walled (H-MWCNT) carbon nanotube modified BPPG. Low potential responses for ORR are found for substrates with large quantities of edge plane sites, namely EPPG and MWCNT modified BPPG. Along with being analytically significant in respect of gas sensing, these results have implications for proton exchange membrane (PEM) fuel cell catalyst supports.

Full text of DOI:10.1039/c1nj20612e

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
NJC
Cite this: New J. Chem., 2011, 35, 2647–2652
www.rsc.org/njc
PAPER
Activity of carbon electrodes towards oxygen reduction in acid:
A comparative studyw
Matthew Gara and Richard G. Compton*
Received (in Montpellier, France) 12th July 2011, Accepted 22nd August 2011
DOI: 10.1039/c1nj20612e
The oxygen reduction reaction (ORR) has been investigated on a variety of carbon substrates, in
acidic conditions, including glassy carbon (GC), edge plane pyrolytic graphite (EPPG), basal
plane pyrolytic graphite (BPPG) and boron doped diamond (BDD). Along with these macro-sized
substrates, a selection of carbon nanotube modified BPPG electrodes have also been examined.
These include single walled (SWCNT), bamboo multi walled (B-MWCNT) and hollow multi
walled (H-MWCNT) carbon nanotube modified BPPG. Low potential responses for ORR are
found for substrates with large quantities of edge plane sites, namely EPPG and MWCNT
modified BPPG. Along with being analytically significant in respect of gas sensing, these results
have implications for proton exchange membrane (PEM) fuel cell catalyst supports.
Alkali
Introduction
HOꢀ2 þ H2O þ 2eꢀ ! 3OHꢀ
E0 ¼ 0:867 V
ð5Þ
ð6Þ
The reduction of oxygen in aqueous solutions can proceed by
two major pathways: A direct four electron pathway and a two
electron ‘peroxide’ pathway.1,2 These pathways are further
dependant on pH conditions. Both pathways are listed below
for both alkaline and acidic conditions.
Acid
H2O2 þ 2Hþ þ 2eꢀ ! 2H2O
E0 ¼ 1:77 V
It is worth noting that summation of the two sequential two
electron reductions (involving peroxide) leads to the respective
4 electron reduction (e.g. summation of equations 3 + 5 and
4 + 6 are equivalent to eqn (1) and (2) respectively.)
Direct 4-electron pathway
Alkali
O2 þ 2H2O þ 4eꢀ ! 4OHꢀ
E0 ¼ 0:401 V
E0 ¼ 1:229 V
ð1Þ
ð2Þ
Eqn (2) is the driving force behind proton exchange membrane
(PEM) fuel cells3,4 and is effected by a catalyst on the cathode.
This cathode catalyst is usually comprised of platinum nano-
particles (generally 3 nm in diameter) supported onto a finely
divided carbon support (generally 20 nm diameter.) Hydrogen
gas, oxidised on the anode, provides the protons required to
complete the reaction. These travel through a proton exchange
membrane, which is normally a perfluorinated ionomer; the
most common of which is Nafions.
Acid
O2 þ 4Hþ þ 4eꢀ ! 2H2O
Peroxide pathway
Alkali
Whilst the vast majority of oxygen reduction, on platinum,
proceeds via the four electron process in eqn (2) a very small
percentage follows the peroxide pathway shown in eqn (4).
This pathway produces hydrogen peroxide which has the
O2 þ H2O þ 2eꢀ ! HO2ꢀ þ OHꢀ
E0 ¼ ꢀ 0:065 V ð3Þ
Acid
ꢁ
potential to decompose into OH radicals. These radicals are
O2 þ 2Hþ þ 2eꢀ ! H2O2
E0 ¼ 0:670 V
ð4Þ
known to attack the carboxylic end groups (formed during the
polymerization process) of the perfluorinated, ionomer chain.5
Oxygen reduction on carbon usually proceeds by the more
common two electron, peroxide pathway.1 Whilst there is
much literature surrounding this process in alkali,6–8 there is
significantly less for the process in acidic solutions, relevant to
PEM fuel cells. Hence this work is directed solely at oxygen
reduction, on carbon, under acidic conditions (specifically
aqueous sulphuric acid.) Under these conditions oxygen
In certain circumstances the peroxide can be further
reduced, as shown in eqn (5) and (6)
Physical and Theoretical Chemistry Laboratory, University of Oxford,
South Parks Road, Oxford, OX1 3QZ, United Kingdom.
E-mail: richard.compton@chem.ox.ac.uk
w Electronic supplementary information (ESI) available. See DOI:
10.1039/c1nj20612e
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2011
New J. Chem., 2011, 35, 2647–2652 2647

View Article Online
but non-exhaustive, range of carbon substrates, which exclude
doped graphitic structures.
Experimental
Materials & instrumentation
All carbon nanotubes were acquired from Nanolab inc. Single
walled carbon nanotubes (Batch number P0340, containing
small amounts of Fe) bamboo multi-walled carbon nanotubes
(Batch number BPD30L520–020306, typical EDAX analysis
0.05% S, 0.2% Fe, 30 ꢂ 15 nm diameter, 5–20 mm length)
hollow multi-walled carbon nanotubes (Batch number
PD30L520–92206, typical EDAX analysis 0.05% S, 0.2% Fe,
30 ꢂ 15 nm diameter, 5–20 mm length) Basal and edge plane
pyrolytic graphite electrodes were cut from graphite obtained
from le Carbone. Each electrode was cut to 5 mm diameter and
mounted in PTFE. Glassy carbon electrodes were obtained
from BAS Technical, working diameter 3 mm. Boron doped
diamond was obtained from Windsor scientific cut to 5 mm
diameter and mounted in PEEK. Oxygen and Nitrogen (oxygen
free) were obtained from BOC gasses. All water used was
distilled before being passed through a Millipore system.
Chloroform was obtained from Fisher Scientific (Batch number
C/4920/17–1007754.) Sulphuric acid was of high purity Primar
grade, suitable for trace metal analysis (Fisher Scientific Batch
number S/9230/PB07–9895337.) All voltammograms were run
using a Metrohm Autolab, mAutolab II potentiostat.
Fig. 1 Schematic diagram showing the relative orientations of the
basal and edge planes within a graphitic structure.
reduction is known to progress according to eqn (4) with no
further reduction of the peroxide9 (eqn (6).)
Carbon electrodes come in a variety of forms,10,11 some of
which are briefly summarised in the subsequent text. Boron
doped diamond (BDD) is a hard material where the carbon
atoms are sp3 hybridised and arranged in a diamond structure,
with approximately one atom in a thousand replaced with
boron. This creates a hard, conducting, diamond like carbon
substrate.12 Ordered pyrolytic graphite contains ordered layers
of sp2 hybridised carbon stacked 3.35 A apart. Basal plane
pyrolytic graphite (BPPG) electrodes are cut so as to expose
the hexagonal plane of the graphite layer. Edge plane pyrolytic
graphite (EPPG) is cut perpendicular to BPPG, so as to expose
as may graphite sheet ends as possible. A schematic represen-
tation of graphite orientations can be seen in Fig. 1. Glassy
Carbon (GC) forms at high temperature and pressure and is
composed of interlinked strands of graphite forming a random
mesh of graphitic ribbons. Carbon nanotubes can be seen as a
graphite sheet that has been rolled up to make a tube. Single
walled nanotubes (SWCNTs) basically contain only one layer
of graphite. As the name suggests multi-walled nanotubes
contain several layers. If the tubes are placed one inside the
other this is termed a hollow multi-walled carbon nanotube
(H-MWCNT.) Bamboo multi-walled carbon nanotubes
(B-MWCNTs) are formed at an angle to the propagating axis
of the tube, these tubes are ‘stacked’ to form a bamboo type
structure, with the tube closed periodically along the axis.
Nanotubes based on chirality also exist, however separation
and purification difficulties place these outside the scope of
this study. Nitrogen doping of carbon nanotubes has been
shown to enhance the rate of oxygen reduction13,14 as has the
introduction of nitrogen containing species such as FeN4
porphyrin type moieties.15
Electrode preparation
EPPG electrodes were polished with 1 mm, 0.3 mm and 0.05 mm
alumina, with thorough rinsing, with water, and sonicating
(2 mins) after each polishing stage. BPPG electrodes were
sanded with P1000 abrasive paper, rinsed with water, repeatedly
cleaved with adhesive tape and rinsed again. BDD and GC
electrodes were polished with 1 mm and 0.1 mm diamond spray
with thorough rinsing, with water, and sonicating (2 mins)
after each polishing stage.
Carbon nanotube modified electrodes were prepared by
drop-casting a dispersion onto a BPPG electrode. This was
done by placing 1 mg of the appropriate nanotubes into 5 ml
of chloroform and sonicating for 3 sessions of 5 min each,
leaving to cool in-between. 20 ml of this black dispersion
was then drop-cast onto a freshly prepared BPPG electrode
in 4, 5 ml aliquots, leaving the chloroform to evaporate
between each casting.
Methods
It is worth noting that the electrode kinetics of a redox couple
at a graphitic basal plane can be orders of magnitude slower
than those of the corresponding edge plane.1,16,17 Hence a small
quantity of basal planes on an edge plane electrode will make
little or no difference whereas a small amount of edge planes on
a basal plane electrode can potentially make a big difference16
to the observed electroactivity. Due to the fact that it is almost
impossible to make an electrode consisting purely of basal
planes, a certain amount of edge plane sites are to be expected
at step edges and grain boundaries, on a basal plane electrode.
The results from this paper will aim to characterise and
compare oxygen reduction, in acidic media, across a broad,
Voltammograms were run using a saturated calomel reference
electrode (SCE) and a platinum mesh counter electrode. Gasses
were bubbled through for 25 mins to achieve saturation. The
electrode to be tested was first placed into a 0.1 M solution of
H2SO4 saturated with oxygen. 3 blank voltammograms were
then run at 100 mV sꢀ1 to equilibrate the electrode and mitigate
any effects caused by trapped solvent, such as chloroform from
the carbon nanotube drop-casting.ꢀC1 yclic voltammograms were
then run at 10 mV sꢀ1, 20 mV s , 40 mV sꢀ1, 100 mV sꢀ1
,
200 mV sꢀ1, 400 mV sꢀ1, 800 mV sꢀ1, across the scan ranges
specified below, rinsing between each voltammogram.
c
2648 New J. Chem., 2011, 35, 2647–2652
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2011

View Article Online
The electrode was then removed and placed into a nitrogen
saturated solution of 0.1 M H2SO4 and cyclic voltammograms
were then run at 10 mV sꢀ1, 20 mV sꢀ1, 40 mV sꢀ1, 100 mV sꢀ1
,
200 mV sꢀ1, 400 mV sꢀ1, 800 mV sꢀ1, across the scan ranges
specified below, rinsing between each voltammogram.
For EPPG, GC, SWCNT, H-MWCNT and B-MWCNT
the potential was scanned from + 0.6 V to ꢀ0.8 V for BPPG
and BDD the potential was scanned from + 0.6 V to ꢀ1.0 V
and + 0.6 V to ꢀ1.3 V respectively
Results and discussion
The results for the investigation into oxygen reduction on
carbon substrates have been split into two main categories.
Results relating to the unmodified carbon, macroelectrodes
will be dealt with separately to the results concerning the
carbon nanotube modified BPPG electrodes.
Carbon Macroelectrodes
Cyclic voltammograms were measured for a range of scan
rates, in both O2 and N2 saturated H2SO4, detailed in the
experimental section. While the majority of experiments in O2
saturated solution gave a response for oxygen reduction, those
of the control experiments in N2 saturated solution (see ESIw)
gave very little response with most nothing except a response
for the breakdown of solvent. Fig. 2 summarises the results
acquired on unmodified carbon macroelectrode substrates, in
O2 saturated 0.1 M H2SO4. It is clear from Fig. 2 that BDD
gives little or no response to oxygen reduction. This agrees
with the work of Yano et al.18 who found that oxygen
reduction on BDD could only be initiated by pretreatment
at positive potentials greater than about + 1.4 V vs. Ag/AgCl.
This was attributed to sp2 hybridised carbon species becoming
oxidised and mediating oxygen reduction, with the most likely
location for the sp2 species being the grain boundaries of the,
sp3, diamond structure.
Fig. 2 Cyclic voltammogramss taken at 400 mV sꢀ1 using BPPG,
EPPG, GC and BDD electrodes in (a) O2 saturated 0.1 M H2SO4
solution and (b) N2 saturated 0.1 M H2SO4 solution. The potential
was swept in a negative direction from 0.6 vs. saturated calomel
electrode (SCE).
n is the number of electrons transferred, F is the Faraday
constant, A is the area of the electrode, [O2] is the oxygen
concentration, T is the absolute temperature, R is the gas
constant and n is the scan rate.
Tafel analyses (eqn (8)) on the EPPG results give an average
value for (n0 + a0) of 0.4; an example Tafel plot can be seen
in Fig. 4. Using this value along with [O2] = 1.3 mM19 and
n = 2, a diffusion coefficient of 2.1 ꢄ 10ꢀ5 cm2 sꢀ1 was
obtained for oxygen. This value shows good agreement with
the value of 2.0 ꢄ 10ꢀ5 cm2 sꢀ1 obtained from interpolating
data from reference 20.
The response obtained from glassy carbon towards oxygen
reduction occurred very close to the breakdown of solvent,
convoluting the signal. Although comparison with the corres-
ponding results for the nitrogen saturated solution clearly
indicate some response towards oxygen reduction. Signals
for glassy carbon have been corrected to account for the
difference in electrode area.
Fig. 2 also shows that the analytical response of EPPG is
better than that of any other electrode substrate. In particular
the current voltage curve for oxygen reduction on EPPG is
shifted around + 450 mV higher than for BPPG. This fits with
previous investigations which have found the electrode kinetics
on the basal plane to be far slower than those of edge plane
sites16 (possibly 7 orders of magnitude slower.) The results for
EPPG are shown in more detail in Fig. 3a. Fig. 3b shows the
0
0
@ ln I ðn þ an ÞF
¼
ð8Þ
@E
RT
Nanotube modified electrodes
´
corresponding Randles Sevcık plot of the data in Fig. 3a. The
equation used for this plot is shown in eqn (7).
All nanotubes were immobilised onto the surface using a
simple dropcasting technique using chloroform dispersions.
The hydrophobicity of all the carbon nanotube modified
electrodes was very apparent, with all the nanotube modified
electrodes forming beads of water rather than wetting the
surface. Again the results in N2 saturated solution (see ESIw)
showed little or no response except for the breakdown of solvent.
rffiffiffiffiffiffiffiffiffi
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
FuD
RT
0
0
Ipeak ¼ ꢀ0:496 ðn þ an ÞnFA½O2ꢃ
ð7Þ
where Ipeak is the peak current, n0 is the number of electrons
transferred before the rate determining electron transfer, a0n
is the electron transfer coefficient of the rate determining step,
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2011
New J. Chem., 2011, 35, 2647–2652 2649

View Article Online
voltage curves for hollow and bamboo carbon nanotubes, with
the curve for bamboo nanotubes shifted to higher potential.
This broadly agrees with the work of Matsubara et al.21 who
found a 312 mV difference in ‘onset potentials’. This difference
is likely to be attributable to the fact that the bamboo
structured nanotubes have a significantly larger proportion
of edge plane sites per unit length, compared with the hollow
structured nanotubes, where edge plane sites have a tendency
to appear only at the ends of the nanotube.
Of particular interest, and the most noticeable feature of
Fig. 5, is that only the hollow nanotube (H–MWCNT) gave an
analysable peak. Further analysis, which will be shown in
more detail later, reveals that this is partly due to the porous
nature of the H–MWCNT modified electrode. A full set of
results for the H–MWCNT modified electrode is shown in
Fig. 6.
Fig. 7a shows a log peak current against log scan rate plot
for calculated, theoretical peak currents and for experimental,
peak currents obtained using the H–MWCNT modified
electrode. Calculated peak currents were obtained from
eqn (7) using [O2] = 1.3 mM,19 (n0+a0) = 0.5, n = 2 and
D = 2.0 ꢄ 10.5 cm2 s
.
ꢀ1 20 If the electron transfer takes place in
the solution this plot should give a linear relationship, with a
Fig. 3 (a) Cyclic voltammograms taken at scan rates of 800, 400, 200,
100, 40, 200, 10 mV sꢀ1, on a 0.5 cm EPPG electrode in an O2
saturated 0.1 M H2SO4 solution. The potential was swept from 0.6 to
ꢀ0.8 V vs. saturated calomel electrode (SCE) (b) Corresponding
Randles Sevcik plot (using eqn (7)) for the data shown for EPPG.
Fig. 4 An example Tafel plot, showing the ‘Tafel region’, used to
calculate the value for (n0ꢀa) for the 400 mV sꢀ1 scan shown in Fig. 3a.
The line used for Tafel analysis can be seen as a dashed line with a
gradient of ꢀ14.62. This is an example of 7 such plots used to give an
average value for (n0ꢀa).
Fig. 5 Cyclic voltammograms taken at 400 mV sꢀ1, using SWCNT,
B–MWCNT and H–MWCNT modified BPPG electrodes, in (a) O2
saturated 0.1 M H2SO4 solution and (b) N2 saturated 0.1 M H2SO4
solution. The potential was swept in a negative direction from 0.6 vs.
saturated calomel electrode (SCE).
Results obtained using carbon nanotube modified BPPG
electrodes, in O2 saturated solution, are summarised in
Fig. 5. A ca. 320 mV difference was found between the current
c
2650 New J. Chem., 2011, 35, 2647–2652
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2011

View Article Online
data using the HMWCNT modified electrode gives a gradient
of 0.63, indicating that there is an additional contribution to
the current.
Further to Fig. 7a, Fig. 7b shows the difference in current
(DI), between the calculated peak currents and the ones
acquired for the H-MWCNT modified electrode, plotted
against scan rate. This graph shows a linear relationship which
intercepts very close to the origin, suggesting that thin layer
diffusion within the porous layer is contributing to the signal.
This is in keeping with previous studies on multiwalled carbon
nanotubes22,23 and other porous layer systems,24,25 which
show that as the thickness of a porous electrode increases,
the contribution from thin-layer diffusion also increases. This
current contribution from thin-layer diffusion was found to be
linearly dependant on scan rate.
Fig. 6 Cyclic voltammograms taken at scan rates of 800, 400, 200,
100, 40, 200, 10 mV sꢀ1, on a H–MWCNT modified BPPG electrode in
an O2 saturated 0.1 M H2SO4 solution. The potential was swept from
0.6 to ꢀ0.8 V vs. saturated calomel electrode (SCE.)
On the basis of the solubility of oxygen in water and its
diffusion coefficient the peaks shown in Fig. 3a for oxygen
reduction on EPPG are qualitatively consistent with a value of
n = 2 (see above) and hence the formation of hydrogen
peroxide (cf. eqn (4).) In comparison the peak current on
BPPG is a little reduced and occurs at higher overpotential.
Hence this signal likely does not reflect a decrease in the n
value but rather the fact that the electrolysis only occurs at
edge plane like defects on the BPPG surface (i.e. the electrode
is heterogeneous) with the basal plane terraces contributing
little to the current.1,16,17 The peak current response for
H–MWCNT is close to that seen for EPPG but quantitative
analysis is precluded because of the contribution of mass
transport within the porous H–MWCNT coat;22,23 that said
the closeness in size of the signal to that seen on the EPPG
electrode (of the same geometric area)suggests that n = 2
ꢂ0.2. The Single walled (SWCNT) and the bamboo multi
walled (B–MWCNT) carbon nanotube modified electrodes
show no well defined reduction signal (Fig. 5a) precluding
analysis of the number of electrons transferred.
Being comprised entirely of sp3 carbon atoms born
doped diamond does not possess any edge plane character
and it is clear to see from Fig. 2a that it gives a negligible
current response towards oxygen reduction in sulphuric
acid. Likewise the basal plane of graphite would also be
expected to show little/no edge plane character and whilst
some current response can be seen for BPPG, this can be
ascribed to the inevitable heterogeneity of the surface, with
some edge plane like defects being present. Being orientated
perpendicular to the basal plane (cf. Fig. 1) EPPG affords
the maximum density of edge plane sites per unit area; in
turn EPPG gives the best analytical response of all the sub-
strates tested. Glassy carbon is thought to have no well defined
edge plane sites. This is reflected in the results, with the
nature of the oxygen reduction signal being closer to BPPG
than EPPG.
In comparison to Hollow multiwalled carbon nanotubes
(H–MWCNT) bamboo multi walled carbon nanotubes
(B–MWCNT) contain many edge plane like sites along the
axis of the entire tube rather than purely at the ends. This
distinct difference in structure (the number of edge plane sites)
is reflected in significantly less overpotential being required to
enact oxygen reduction on bamboo compared to hollow multi
walled carbon nanotube modified electrodes.
Fig. 7 (a) log of the peak current (log Ipeak) plotted against log scan rate
(log n). The black line shows the calculated peak currents from eqn (7)
whereas the redline shows the peak currents obtained from H–MWCNT
electrodes. (b) The residual current (DI) plotted against scan rate(n).
gradient of 0.5, according to eqn (7). We can see from Fig. 7a
that while the data using calculated, theoretical, currents (for
semi-infinite diffusion) gives a gradient of 0.5, the experimental
c
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2011
New J. Chem., 2011, 35, 2647–2652 2651

View Article Online
7 B. Sljukic, C. E. Banks and R. G. Compton, J. Iranian Chem. Soc.,
2005, 2, 1–25.
Conclusions
8 H. H. Yang and R. L. McCreery, J. Electrochem. Soc., 2000, 147,
3420–3428.
9 in Comprehensive Treatise of Electrochemistry, Vol 4: Electro-
chemical Materials Science, ed. B. Conway, J. Bockris,
E. Yeager, S. Khan and R. White, Plenum Press, New York,
1981, vol. 4, pp. 486–489.
10 C. E. Banks and R. G. Compton, Analyst, 2006, 131, 15–21.
11 R. L. McCreery, Chem. Rev., 2008, 108, 2646–2687.
12 R. G. Compton, J. S. Foord and F. Marken, Electroanalysis, 2003,
15, 1349–1363.
13 K. Gong, F. Du, Z. Xia, M. Durstock and L. Dai, Science, 2009,
323, 760–764.
From the results presented here it is possible to see that
graphitic structures give a better analytical response to oxygen
reduction than the cubic diamond structure of BDD. It is also
possible to see that, within those graphitic structures, the
substrates with a large number of edge plane sites have a
higher activity towards oxygen reduction and therefore a
higher tendency to produce H2O2 if used as a catalyst support
in a PEM fuel cell. It is therefore favourable to decrease the
number of edge plane sites in carbon catalyst supports for
PEM fuel cells in order to increase efficiency and reduce
membrane degradation from hydrogen peroxide attack.
14 N. Alexeyeva, E. Shulga, V. Kisand, I. Kink and K. Tammeveski,
J. Electroanal. Chem., 2010, 648, 169–175.
15 D. H. Lee, W. J. Lee, W. J. Lee, S. O. Kim and Y.-H. Kim, Phys.
Rev. Lett., 2011, 106, 175502.
Acknowledgements
16 C. E. Banks, T. J. Davies, G. G. Wildgoose and R. G. Compton,
Chem. Commun., 2005, 829–841.
Matthew Gara would like to thank EPSRC and Johnson
Matthey for an Industrial CASE award.
17 C. C. M. Neumann, C. Batchelor-McAuley, C. Downing and
R. G. Compton, Chem.–Eur. J., 2011, 17, 7320–7326.
18 T. Yano, E. Popa, D. A. Tryk, K. Hashimoto and A. Fujishima,
J. Electrochem. Soc., 1999, 146, 1081–1087.
19 T. N. Das, Ind. Eng. Chem. Res., 2005, 44, 1660–1664.
20 P. Han and D. M. Bartels, J. Phys. Chem., 1996, 100, 5597–5602.
21 K. Matsubara and K. Waki, Electrochim. Acta, 2010, 55,
9166–9173.
22 M. C. Henstridge, E. J. F. Dickinson, M. Aslanoglu, C. Batchelor-
McAuley and R. G. Compton, Sens. Actuators, B, 2010, 145,
417–427.
23 M. J. Sims, N. V. Rees, E. J. F. Dickinson and R. G. Compton,
Sens. Actuators, B, 2010, 144, 153–158.
24 I. Streeter, G. G. Wildgoose, L. D. Shao and R. G. Compton, Sens.
Actuators, B, 2008, 133, 462–466.
References
1 in Comprehensive Treatise of Electrochemistry, Vol 7 &: Kinetics
and Mechanisms of Electrode Processes, ed. B. Conway, J. Bockris,
E. Yeager, S. Khan and R. White, Plenum Press, New York, 1983,
vol. 7, pp. 301–398.
2 E. Yeager, Electrochim. Acta, 1984, 29, 1527–1537.
3 S. Litster and G. McLean, J. Power Sources, 2004, 130, 61–76.
4 J. Larminie and A. Dicks, Fuel Cell systems Explained, Wiley,
2nd edn, 2003.
5 D. E. Curtin, R. D. Lousenberg, T. J. Henry, P. C. Tangeman and
M. E. Tisack, J. Power Sources, 2004, 131, 41–48.
6 E. Yeager, J. Mol. Catal., 1986, 38, 5–25.
25 L. Xiao, G. G. Wildgoose and R. G. Compton, Sens. Actuators, B,
2009, 138, 524–531.
c
2652 New J. Chem., 2011, 35, 2647–2652
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2011