High performance and cost-effective direct methanol fuel cells: Fe-N-C methanol-tolerant oxygen reduction reaction catalysts

Article abstract of DOI:10.1002/cssc.201600583

Direct methanol fuel cells (DMFCs) offer great advantages for the supply of power with high efficiency and large energy density. The search for a cost-effective, active, stable and methanol-tolerant catalyst for the oxygen reduction reaction (ORR) is still a great challenge. In this work, platinum group metal-free (PGM-free) catalysts based on Fe-N-C are investigated in acidic medium. Post-treatment of the catalyst improves the ORR activity compared with previously published PGM-free formulations and shows an excellent tolerance to the presence of methanol. The feasibility for application in DMFC under a wide range of operating conditions is demonstrated, with a maximum power density of approximately 50 mW cm² and a negligible methanol crossover effect on the performance. A review of the most recent PGM-free cathode formulations for DMFC indicates that this formulation leads to the highest performance at a low membrane–electrode assembly (MEA) cost. Moreover, a 100 h durability test in DMFC shows suitable applicability, with a similar performance–time behavior compared to common MEAs based on Pt cathodes.

Full text of DOI:10.1002/cssc.201600583

DOI: 10.1002/cssc.201600583
Full Papers
High Performance and Cost-Effective Direct Methanol Fuel
Cells: Fe-N-C Methanol-Tolerant Oxygen Reduction
Reaction Catalysts
[
a]
[b]
[b]
[b]
David Sebastiꢀn, Alexey Serov, Kateryna Artyushkova, Jonathan Gordon,
[b]
[a]
[a]
Plamen Atanassov,* Antonino S. Aricꢁ, and Vincenzo Baglio*
Direct methanol fuel cells (DMFCs) offer great advantages for
the supply of power with high efficiency and large energy den-
sity. The search for a cost-effective, active, stable and metha-
nol-tolerant catalyst for the oxygen reduction reaction (ORR) is
still a great challenge. In this work, platinum group metal-free
range of operating conditions is demonstrated, with a maxi-
ꢀ
2
mum power density of approximately 50 mWcm and a negli-
gible methanol crossover effect on the performance. A review
of the most recent PGM-free cathode formulations for DMFC
indicates that this formulation leads to the highest per-
formance at a low membrane–electrode assembly (MEA) cost.
Moreover, a 100 h durability test in DMFC shows suitable ap-
plicability, with a similar performance–time behavior compared
to common MEAs based on Pt cathodes.
(PGM-free) catalysts based on Fe-N-C are investigated in acidic
medium. Post-treatment of the catalyst improves the ORR ac-
tivity compared with previously published PGM-free formula-
tions and shows an excellent tolerance to the presence of
methanol. The feasibility for application in DMFC under a wide
Introduction
Growing energy demand and environmental concerns have
driven the search for new technologies to produce green
energy because fossil fuels are a finite resource and their direct
ment of DMFCs is their low power density resulting from the
sluggish kinetics of the methanol oxidation reaction (MOR) and
oxygen reduction reaction (ORR). The most critical factors are
the high cost of the anode and cathode catalysts, which are
usually based on expensive and scarce platinum metal, and
the crossover effect, that is, the permeation of the alcohol
through the polymeric membrane, which creates a mixed po-
[
1]
combustion raises pollution levels. Fuel cells are a competitive
alternative for the supply of energy with high efficiency and
low emission levels by conversion of chemical energy into
electricity through electrochemical reactions. The fuel cell con-
figuration (i.e., fuel, electrolyte, start-up/shut-down speed)
used for a particular application depends on the purpose and
[6–9]
tential at the cathode and decreases the overall efficiency.
Although Pt presents the highest activity for the ORR, it is also
a good methanol dehydrogenation/oxidation catalyst, which
implies that if methanol and oxygen are present the selectivity
toward ORR is far lower than 100%.
[
2]
limitation of the operating temperature.
Direct alcohol fuel cells, in particular direct methanol fuel
cells (DMFCs), are considered a key enabling technology for
powering portable devices in a sustainable economy based on
The high cost of the cathode and the detrimental effect of
methanol crossover can be greatly reduced with a single solu-
tion: the use of platinum group metal (PGM)-free catalysts in
the cathode side. These are characterized by a much lower
[
3–5]
highly efficient and renewable sources.
DMFCs are very con-
venient for prolonged duration systems owing to the high
ꢀ
1 [5]
energy content of methanol (6100 mWhg ). The remaining
technical barrier that needs to be overcome for the develop-
[10–12]
cost than Pt
and by an extraordinarily high selectivity to-
wards the oxygen reduction (neither significant methanol poi-
soning nor oxidation). Many non-platinum formulations have
shown great activity compared to platinum catalysts, but the
performance, durability and stability are still very low in
[
a] Dr. D. Sebastiꢀn, Dr. A. S. Aricꢁ, Dr. V. Baglio
Istituto di Tecnologie Avanzate per l’Energia “Nicola Giordano” (ITAE)
Consiglio Nazionale delle Ricerche (CNR)
Via Salita S. Lucia sopra Contesse 5, 98126, Messina (Italy)
E-mail: baglio@itae.cnr.it
[13,14]
a single cell.
Until now, M-N-C catalysts (M=transition
metal such as Fe or Co) have shown great promise owing to
their good ORR activity and have been targeted for low cost
[
b] Dr. A. Serov, Dr. K. Artyushkova, Dr. J. Gordon, Prof. P. Atanassov
Department of Chemical and Biological Engineering and Center for Micro-
Engineered Materials
[15,16]
hydrogen-fuel cells.
The research interest for their applica-
Farris Engineering Center
tion in DMFCs has arisen recently. Investigations that have re-
University of New Mexico
Albuquerque, NM 87131 (USA)
E-mail: plamen@unm.edu
ported DMFCs can be divided into the following categories:
[17–23]
Fe-based catalysts mostly in acidic DMFCs
as well as a few
[22,24,25]
in alkaline type DMFCs,
and Co-based catalysts in
Supporting Information and the ORCID identification number(s) for the
author(s) of this article can be found under http://dx.doi.org/10.1002/
cssc.201600583.
[19,26,27]
[28]
acidic
or alkaline DMFCs. Other PGM-free formulations,
[29–34]
mainly chemically or physically modified carbon
or other
ChemSusChem 2016, 9, 1 – 11
1
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
[
35]
emerging materials for alkaline DMFCs, have also been stud-
ied. In general, these catalysts are promising, especially M-N-C
class, although further activity enhancement and especially
electrode optimization is still needed.
wards more negative values (i.e., higher overpotentials). Such
a current density plateau is known as the diffusion limiting cur-
rent and is shown by the Levich equation to depend on the
square root of the rotation rate (w) and the number of elec-
trons involved in the reaction (n), for a given electrolyte. The
difference in the limiting current density between a double
heat-treated (Fe-N -C-DHT) and triple heat-treated (Fe-N -C-
We have recently reported outstanding performances of an
iron aminoantipyrine derived catalyst at the cathode of
a DMFC with a low PtRu content at the anode and high meth-
x
x
ꢀ
2
[36]
anol molarity (30 mWcm with 10m methanol). It appears
that there is still room for further improvement on the basis of
the cathode catalytic activity while maintaining a low PtRu
content on the anode side. In the present work we investigate
the influence of post-treatment of the synthesized aminobenzi-
midazole-derived M-N-C catalyst on the electrochemical per-
formance of the fuel cells. It has been previously demonstrated
that the treatment temperature, duration and ramp rate signifi-
THT) catalyst at a fixed rotation rate indicates limiting diffusion
constraints ascribed to the catalytic layer on the RDE
ꢀ2
(600 mgcm ) as well as a small contribution from the higher
Helmholtz double layer capacitance of the THT sample.
From a catalysis point of view, the difference in the kinetic-
diffusion mixed region, that is, the intermediate region of the
curves [0.5–0.9 V vs. the reversible hydrogen electrode (RHE)] is
more interesting. The Fe-N -C-THT catalyst is clearly more
x
[
37]
cantly influence the catalytic activity. The rationale herein
has been to preliminary investigate the influence of heating
and leaching treatments of the Fe-N-C catalysts on the ORR
electrocatalytic activity in a rotating disk electrode (RDE) and
rotating ring disk electrode (RRDE) configurations, as common-
active than the Fe-N -C-DHT catalyst, as indicated by the 35 mV
x
shift. The enhancement of the ORR activity was attributed to
substantial improvement of the synthetic methodology. The
catalyst obtained after two heat-treatments (Fe-N -C-DHT) was
x
subjected to additional leaching in 4m HNO3 to: a) remove
iron nanoparticles covered with thin layers of graphitic carbon;
b) remove amorphous carbon, which is prone to be corroded
[
38–40]
ly reported in the literature.
Thereafter, the most active
formulation was further studied at the cathode of a DMFC
under high energy density conditions (high methanol concen-
tration) and cost-effectiveness (low PtRu loading at the anode).
High methanol-tolerance together with a high density of
active sites have resulted in a significant improvement of the
DMFC performance with respect to the state of the art.
during the fuel cell operation; and c) expose the FeꢀN active
x
sites buried by the amorphous carbon layers. A third heat
treatment (Fe-N -C-THT) in ammonia atmosphere was benefi-
x
cial not only for cleaning of the surface from partially oxidized
carbon-derived material, but also for creating additional FeꢀN
interactions. It should also decrease the oxygen content on the
catalyst surface (NH gas is a reducing agent at the treating
3
Results and Discussion
temperature), which is beneficial for the increase of the cata-
lyst hydrophobicity and for preventing the flooding of the
pores by water and permeated methanol. At low rotation
speed, a reduction peak appears at approximately 0.68–0.72 V
versus RHE before achieving the diffusion limiting current,
Fe-N -C (x=1–4) catalysts were prepared by the pyrolysis of
x
iron nitrate mixed with aminobenzimidazole using a modified
sacrificial support method (SSM) developed at the University of
New Mexico (UNM). The catalytic activity towards the ORR was
determined in acidic solution (0.5m H SO ) by the rotating disk
electrode (RDE) technique, as depicted in Figure 1. Both PGM-
free formulations show the typical sigmoidal wave, approach-
ing a constant current density when scanning the potential to-
which is especially evident for the Fe-N -C-THT catalyst. This
x
2
4
could be associated with the reduction of a redox couple
2
+
3+
[41]
(
such as Fe /Fe ) of the catalyst.
However, the stable
cyclic voltammograms in idle mode (Figure S1, Supporting In-
formation) show only a small contribution of such a redox
2
+
3+
Fe /Fe couple at approximately 0.4–0.65 V versus RHE, but
ꢀ
2
with much lower current densities (<0.2 mAcm ). Therefore,
it appears that the peak observed in the linear sweep voltam-
metry (LSV) at low w shown in Figure 1 does not correspond
2
+
3+
with the reduction of the Fe /Fe . Thus, this reduction
valley’ at about 0.7 V versus RHE can be attributed to transient
mass transport phenomena related to oxygen diffusivity in the
‘
[42]
catalytic layer, as reported for carbon-supported Pt catalysts.
Thus, the use of a thick catalytic layer produces this ‘valley’
effect owing to oxygen concentration onto the catalyst surface
higher than the oxygen saturation concentration in the elec-
trolyte. This effect was further confirmed by using lower scan
rates (Figure S2). The current related to the mentioned reduc-
ꢀ
1
tion ‘valley’ decreases if sweeping at 2 mVs and completely
ꢀ
1
disappears at 0.5 mVs , even with a low rotation speed
(
100 rpm).
Figure 1. RDE curves for Fe-N
-saturated 0.5m H SO
rected for series resistance nor double-layer capacitance).
x
-C-DHT (black) and Fe-N
, 5 mVs , room temperature, 600 mgcm (not cor-
x
-C-THT (red) catalysts.
ꢀ
1
ꢀ2
Koutecky–Levich (K–L) plots obtained from RDE experiments
O
2
2
4
are represented in Figure 2a. The K–L plots clearly show
&
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
2
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
1
ꢀ
=
2
of Figure 2a at a virtual w =0. The best kinetic currents are
obtained for the Fe-N -C-THT catalyst, which indicates a better
x
ORR turnover frequency.
As described earlier, additional leaching with 4m HNO3
should result in the removal of carbon-covered iron nanoparti-
cles. The Fe nanoparticles are well-known Fenton type radical
producers and facilitate hydrogen peroxide generation. The
latter process would affect the overall catalytic performance
and durability (peroxide and radicals attack ionomers and
membranes causing their degradation). RRDE studies were
used to confirm the effectiveness of the removal of carbon-
covered Fe nanoparticles. The calculation of the H O yield was
2
2
performed using Equation (1):
2
jr
%
H O ¼ 100 %
,
ð1Þ
2
2
Nj þ j
d
r
where j is the current density on the ring, N is the collection
r
efficiency of 40%, and j is the current density on the disk.
d
Figure 2b shows the H O2 production (%) calculated by
2
Equation (1) as a function of disk potential in the RRDE config-
uration. The variation of the H O yield with potential is similar
2
2
[45]
to previously published PGM-free catalysts. The Fe-N -C-THT
x
catalyst produces 40% less peroxide compared to Fe-N -C-DHT.
x
Such a decrease in peroxide production is a direct indication
ꢀ
of a partial shift of the ORR mechanism from a 2ꢃ2e path-
way, in which O is first reduced to H O followed by a second
2
2
2
ꢀ
step of electroreduction of H O to H O, to a direct 4e path-
2
2
2
[46]
way of O !H O.
2
2
The differences in the catalytic behavior were also analyzed
by Tafel plots, as shown in Figure 3. Very similar Tafel slopes of
Figure 2. (a) Koutecky–Levich (0.5 V vs. RHE) in RDE configuration and
b) H yield in RRDE configuration, for Fe-N -C-DHT and Fe-N -C-THT cata-
lysts in 0.5m H , 1600 rpm, scan rate 5 mVs , solution saturated with O
(
2
O
2
x
x
ꢀ
1
ꢀ1
2
SO
4
2
.
about 60 mVdec were observed for both catalysts, which is
a typical value also for Pt-based catalysts at low current densi-
[47]
ties. The difference between the Fe-N -C catalysts is thus at-
x
a linear behavior of the inverse of the current density (correct-
ed by the double layer capacitance) with the reciprocal of the
square root of the rotation speed. By applying the Levich
equation to the linear fitting of the experimental data we cal-
culated the transferred electron number (n) from the slope
considering the O concentration in saturated 0.5m H SO to
tributed to the exchange current density, j , calculated as the
0
intercept of the linear fittings with the x-axis at the equilibrium
potential for ORR (1.23 V vs. RHE). The exchange current densi-
ty is about two times higher for the THT catalyst compared
2
2
4
ꢀ
3
ꢀ1
be 1.26ꢃ10 moll , the diffusion coefficient of molecular O
2
ꢀ
5
2
ꢀ1
to be 1.93ꢃ10 cm s and the kinetic viscosity of the electro-
2
ꢀ1 [43,44]
lyte to be 0.01 cm s .
The n values calculated for the Fe-
N -C-DHT were in the range of 3.8–4.0 and 3.9–4.0 for the Fe-
x
N -C-THT catalyst in the same potential interval (0.50–0.65 V vs.
x
RHE), this is, in the mixed kinetic-diffusion-limited zone. Addi-
tional analysis of the number of electrons involved in the reac-
tion was performed using the experimental data obtained
using the RRDE configuration (Equation S1). The results show
that n is approximately 3.9 in a broad potential range (from 0
to 0.6 V vs. RHE). The data from RRDE and K–L analysis are
thus in perfect agreement, indicating that n is approximately
4. An n value of approximately 4 indicates that an efficient
mechanism in the electroreduction of oxygen to water pro-
ꢀ
ceeds through a 4e pathway. The difference between these
two PGM-free catalysts is thus related to the kinetic current, as
2 2 4
Figure 3. Tafel plots for the PGM-free formulations. O -saturated 0.5m H SO ,
room temperature, catalyst loading: 600 mgcm .
ꢀ
2
envisaged from the intercept of the fitted lines with the y-axis
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
3
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
with the DHT catalyst. This is presumably related to a larger
number of surface oxygen reduction active sites of the former,
which are formed during the additional leaching and third
heat treatment.
Table 1. Chemical speciation of Fe-N
derived from XPS analysis survey spectra.
x
-C-DHT and Fe-N
x
-C-THT catalysts
Sample
C1s [%]
O1s [%]
N1s [%]
Fe2p [%]
The physicochemical characterization of the Fe-N -C-THT cat-
Fe-N -C-DHT
87.1
91.2
11.1
5.5
1.6
3.1
0.09
0.13
x
x
x
Fe-N -C-THT
alyst was performed using microscopy techniques (SEM and
TEM), N physisorption (BET surface area) and surface chemical
2
composition analysis by X-ray photoelectron spectroscopy
(
XPS). As shown in Figure 4, the Fe-N -C-THT catalyst is highly
(11.1%), which decreases to 5.5% after the third heat treat-
ment (Table 1).
x
Interestingly, the Fe-N -C-THT sample has higher amounts of
x
surface nitrogen and iron, which might increase the density of
[46]
the active sites for ORR. The distribution of nitrogen and
iron species is similar in both samples, so the overall chemistry
of the Fe-N containing part of the electrocatalyst does not sig-
nificantly change after additional leaching and the third heat
[49]
treatment. Therefore, the removal of surface oxides, which is
accompanied with the increase of hydrophobic carbon (graph-
itic C and CꢀC) as evidenced by carbon speciation analysis re-
ported in Table 2, and the larger amount of atomic N and Fe
present after the third heat treatment results in higher activity
of the THT material.
Table 2. Carbon species distribution derived from C1s deconvolution of
the XPS spectra.
Sample
Graphitic C [%]
CꢀC [%]
CꢀN [%]
x y
C O [%]
Fe-N
Fe-N
x
-C-DHT
-C-THT
9.4
12.5
35.1
38.9
9.3
10.8
33.4
29.1
x
For the application of DMFCs, it is highly desirable for the
cathode catalyst to exhibit a high tolerance to the presence of
[50]
methanol. Unreacted methanol permeates through the poly-
meric membrane in the membrane–electrode assembly (MEA)
and reaches the cathode compartment. The adsorption of
methanolic moieties on the surface of Pt (the benchmark cath-
ode catalyst) causes a mixed potential, leading to a significant
loss of cell efficiency. The tolerance to methanol poisoning of
x
Figure 4. (a) SEM and (b) TEM images of the Fe-N -C-THT catalyst.
the Fe-N -C catalysts was studied in a half-cell configuration by
x
porous and exhibits the characteristic features of Fe-N-C cata-
adding methanol to the electrolyte solution and studying the
ORR response. Figure 5 depicts the variation of the ORR half-
[
48]
lysts made by SSM. Such a porous structure allows the ion-
omer to be integrated inside the catalyst layer, providing
oxygen access to the active sites as well as allowing the re-
moval of produced water. Notably, the leaching with 4m HNO3
resulted in the successful removal of iron nanoparticles cov-
ered with graphitic carbon layers (Figure 4b, top); the same
phenomenon was observed and discussed in detail in a previ-
wave potential (E =2)—that is, the potential at half the diffusion-
1
limiting current density—with the increase of methanol con-
centration using a logarithmic scale. The complete set of
curves are shown in Figure S2. For comparison, a commercial
Pt/C (40 wt% Pt, Johnson Matthey) is included in the figure. In-
terestingly, the decrease of E
1
with increasing methanol con-
=
2
[
49]
ously published work. The BET surface area was not changed
after the second leaching and the third heat treatment and
centration is only a few millivolts in both PGM-free catalysts,
which is related to their extraordinarily high selectivity to the
2
ꢀ1
[51]
was similar to Fe-N -C-DHT, 620 m g .
ORR. More specifically, the Fe-N -C-DHT and Fe-N -C-THT cat-
x
x
x
High resolution XPS spectra were obtained to analyze the
alysts show only an 18 mV and 10 mV decay of E =2, respectively,
1
chemical species on the surface of the Fe-N -C catalysts (not
when moving from 0.001m to 1m. In contrast, the Pt/C catalyst
shows a sharp decrease of half-wave potential (about 200 mV)
for concentrations higher than 0.01m and then a smoother
but not negligible shift towards more negative potentials,
x
shown). The elemental composition of the Fe-N -C-DHT and
x
Fe-N -C-THT electrocatalysts was obtained using XPS data. The
x
Fe-N -C-DHT sample has a significant amount of oxygen
x
&
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
4
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
1
=
2
Figure 5. Methanol concentration effect on the half-wave ORR potential (E )
in the O
2 2 4
-saturated 0.5m H SO electrolyte, recorded by LSV at 1600 rpm
ꢀ
1
and 5 mVs , and comparison with Pt/C (40 wt% Pt, Johnson Matthey).
which results in a total E decay of 296 mV in the same meth-
1
=
2
anol concentration interval (from 0.001 to 1m). A small part of
the current decay with increasing methanol concentration is
related to the change of the properties in the electrolyte solu-
tion with the addition of methanol, mainly oxygen saturation
concentration and oxygen diffusivity. This would explain the
1
0–20 mV decay in the Fe-N -C catalysts when moving from
x
methanol-free to 1m solution.
A performance analysis of the most active formulation, that
is, the Fe-N -C-THT catalyst, was performed in a single cell con-
x
figuration DMFC. The PGM-free catalyst was studied at the
cathode side of an MEA formed with a perfluorosulfonic acid
ꢄ
(
PFSA) membrane (Nafion 115) and a commercial PtRu black
ꢀ
2
catalyst (Johnson Matthey) at the anode side (1 mgPtRu cm ).
The results obtained with this kind of analysis can be consid-
ered more realistic than those obtained in a half-cell configura-
tion because of the occurrence of mass transport phenomena
such as oxygen diffusivity, electrode flooding or permeated
methanol local concentration, which are more complex to
study in a three-electrode configuration; therefore, it offers
a better indication of the performance of a DMFC in practical
applications.
DMFC polarization and power density curves obtained at
three different temperatures (30, 60 and 908C) and four meth-
anol concentrations (1, 2, 5 and 10m) are shown in Figure 6. At
low temperature (30–608C) the maximum power density in-
creases with methanol concentration. This may be attributed
to improved methanol transfer at higher methanol concentra-
Figure 6. Influence of methanol molarity on DMFC polarization (empty sym-
bols) and power density (filled symbols) curves at (a) 308C, (b) 608C and
ꢀ
2
(
c) 908C using the Fe-N
x
-C-THT catalyst at the cathode (4.5ꢁ0.2 mgcm ).
ꢄ
ꢀ2
Membrane: Nafion 115. Anode: 1 mgPtRu cm
.
tions, combined with the great tolerance of the Fe-N -C-THT
x
cathode to the permeated alcohol. Generally, Pt-cathode-based
MEAs show a sharp decrease in performance when feeding
denced by the increase of cell potential at high current densi-
ty.
[
36]
ꢀ1
with high methanol concentrations, whereby the detrimen-
tal crossover effect at the cathode is not compensated by the
better performance of the anode. This is not the case of the
Pure methanol contains 6100 mWhg but the supply of
water with the fuel is required for its oxidation (1 mol of water
per mol of methanol), so aqueous solutions are generally em-
ployed. The energy density of a 10m methanol solution is
MEA based on a Fe-N -C cathode. The cell potential experien-
x
ꢀ
1
ces mass transport constraints when feeding 1m methanol,
which is recovered by increasing the methanol molarity, as evi-
2106 mWhg
whereas
a
2m solution contains only
ꢀ
1
393 mWhg . Thus, a high methanol concentration is required
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
5
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
for the design of high energy density systems based on
DMFCs. In the polarization curves shown in Figure 6a and 6b,
it is remarkable that the cell potential does not vary with
methanol concentration up to a certain current density value
ꢀ
2
ꢀ2
(
30 mAcm at 308C or 60 mAcm at 608C). Indeed, the open
circuit potential (OCP) barely changes, which is directly related
to the methanol tolerance properties of the cathode. At high
current densities, the differences observed owing to the varia-
tion of methanol concentration are related to the mass trans-
port constraints at the anode side.
The polarization curves at 908C show a slightly different be-
havior with methanol concentration, as shown in Figure 6c. In
this case, there are no significant differences between the
curves obtained with methanol molarities in the 2–10m
region. The cell potential is almost independent of methanol
concentration from an OCP of 0.82–0.85 V up to approximately
ꢀ2
0
.3 V (135 mAcm ). At higher current densities only the curve
related to the lowest methanol molarity (1m) differs from the
others and shows a lower performance. It appears that under
high temperature and high methanol concentration conditions
the overall cell performance is determined by the cathode
rather than the anode. The maximum power density achieved
ꢀ
2
ꢀ2
is 48 mWcm at 225 mAcm and a cell potential of 0.22 V at
a methanol feed concentration of 2–10m to the anode side.
Figure 7 shows the variation of several cell operating param-
eters of interest in two working regimes, at different tempera-
tures, and as a function of the methanol concentration. The
comparison includes an MEA formed by a commercial Pt/C cat-
alyst at the cathode (40 wt% Pt, Johnson Matthey) and identi-
cal configuration of the anode and membrane. In the activa-
tion controlled region (Figure 7a), most of the contribution to
the cell behavior comes from the kinetic control of MOR and
ORR. It is remarkable that in all cases the cell potential is
higher for the PGM-free MEA (even at a low methanol concen-
tration of 1m). This indicates that, regardless of the tempera-
ture, the Fe-based catalyst is kinetically more active than the
benchmark Pt-based catalyst. From the RDE half-cell characteri-
zation (Figure 5) it was deduced that this condition occurs only
for local methanol concentrations above a certain value, at
which the adsorption of the alcohol on the Pt active sites neg-
atively affects the selectivity toward ORR, whereas the Fe-
based catalyst still offers a high selectivity (virtually close to
Figure 7. Effect of methanol concentration fed to the anode on (a) cell po-
ꢀ
2
tential in the activation controlled region (10 mAcm ) and (b) current densi-
ty at high power density regime (0.2 V). DMFCs are based on Fe-N -C-THT
x
(red squares) and Pt/C (blue circles) cathodes. Three different operating tem-
ꢄ
ꢀ2
peratures (30, 60 and 908C). Membrane: Nafion 115. Anode: 1 mgPtRu cm
.
Data of DMFC performance based on Pt/C cathode from ref. [36].
performances with respect to the Pt-based MEA when a high
methanol concentration is used. At an intermediate–high tem-
perature (60–908C) and methanol concentrations ꢂ5m, the
MEA based on the Pt/C cathode exhibits higher current densi-
ties. It is well-known that the methanol crossover effect de-
[52]
creases with current density. As a consequence, the gap be-
tween the PGM-free and the Pt formulations is reduced when
operating in the high current density regime. The differences
encountered may also be related to the different thicknesses
100%). In this zone, that is, at low current density, the cell po-
tential shows significantly different trends with methanol mo-
larity depending on the cathode catalyst, Fe-N -C-THT or Pt/C.
x
The PGM-free formulation almost maintains the cell potential
with increasing methanol concentration, with a decay of only
approximately 25 mV when moving from 1 to 10m. This decay
could be related to the oxygen transport variation at the cath-
ode side derived from the presence of a higher methanol con-
centration. Instead, the MEA based on the Pt/C cathode exhib-
its a decay in the order of 100–170 mV under identical working
conditions.
of the catalytic layer; approximately 220–250 mm for the Fe-N -
x
ꢀ2
ꢀ
2
C-THT (4.5 mgcm ) and 70–90 mm for the Pt/C (1 mgPtcm ).
As a result, the average O partial pressure is lower at the
2
active sites of the Fe-N -C cathode owing to a higher pressure
x
drop, leading to lower performance when demanding high
current. This effect is not that evident at low temperature
(308C), at which the performance of the MEA based on the
PGM-free cathode is higher. These results are promising for ap-
plications in low power devices working at near-ambient tem-
The performance analysis in the high power density region
[9]
(0.2 V, Figure 7b) shows different trends than those derived
perature such as small portable electronics. In summary, the
from the activation controlled zone. In this case, the Fe-N-C-
Fe-N -C-THT catalyst investigated in this work outperforms Pt/C
x
based MEA exhibits higher (30, 908C) or comparable (608C)
catalysts for application in high energy density DMFC systems
&
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
6
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
ꢀ
2
ꢀ2
(
>10m methanol molarity) and also under varying concentra-
58 mWcm at 608C by using 10 mgcm of an iron-based
[19]
tion at low temperature, such as air-breathing portable DMFCs.
Moreover, the analysis of the kinetic behavior indicates that
there is enormous potential to improve the performance by
improving the electrode structure.
electrocatalyst combined with a N-doped carbon aerogel.
ꢀ
2
ꢄ
They used 4 mg PtRu cm at the anode and Nafion 117 as
[34]
the polymeric membrane. Wan et al. reported a high per-
ꢀ
2
formance of 40 mWcm at 508C using a nitrogen-doped or-
dered mesoporous carbon at the cathode and a Pd catalyst at
the anode with an alkaline membrane. It must be noted here
that, in the latter work, the use of Pd represents an advantage
There have only been a few reports of the performance data
of PGM-free catalysts at the cathode of DMFCs, presumably be-
cause of the relatively recent research interest in this topic.
Figure 8 represent a summary of the latest reports of DMFC
[53,54]
since it has lower cost and larger abundance than Pt.
Ja-
[25]
[28]
narthanan et al. and Shen et al. have achieved good re-
sults using alkaline membranes operating at 808C, reaching
ꢀ
2
ꢀ2
5
2 mWcm with an Fe-N-C cathode and 40 mWcm with
a Co-N-C cathode, respectively. Our cost-effective acidic MEA
shows the best results in terms of performance-to-cost ratio.
Moreover, the Fe-N -C-THT cathode can be further optimized.
x
Presumably, enhancing the electrode structure and optimizing
the exposure of the active sites and diffusion of the reactants
through the open porosity could lead to less mass transport
losses at high current, and thus, improved performance.
Durability is also an important concern in the development
[55]
of new DMFCs.
Figure 9 shows a 100 h durability test at
Figure 8. Latest DMFC performance results using acidic (red circles) or alka-
line (blue stars) membrane and PGM-free catalysts at the cathode. For more
details, please consult Table S1. *Pd (not Pt) was used at the anode side in
Ref. [34]
performance using PGM-free formulations at the cathode. Fur-
ther details (e.g., electrode composition, cell operating condi-
tions) are provided in Table S1. The maximum power density is
represented against the content of Pt in the MEA. Two MEA
configurations have been differentiated according to the mem-
brane: acidic (mostly PFSA-based membranes) and alkaline
3
Figure 9. Stability test in DMFC at 0.3 V, 908C, 5m CH OH fed to the anode
ꢀ
2
and O (100%RH) to the cathode. Anode 1 mgcm PtRu black; membrane:
2
ꢄ
Nafion 115. Fe-N
1
x
-C-THT cathode 4.5ꢁ0.2 mgcmꢀ2; Pt/C cathode
2
.0ꢁ0.1 mgcmꢀ
.
(mostly poly(arylene ether ketone) (PAEK)-based membranes).
The major contribution to the total cost of the DMFC stack in
such an MEA configuration is the Pt content of the anode and
the polymeric membrane. At the anode side, PtRu and Pt cata-
lysts are generally used in acidic and alkaline MEAs, respective-
0.3 V, 908C and 5m methanol feed and fully humidified
oxygen, performed on the MEA with Fe-N -C-THT catalyst at
x
ꢀ
2
ly. Therefore, the PGM loading (mgcm ) at the anode has an
important impact on the total cost of the MEA. In practice,
moving towards the upper left side of the graph is mandatory
to achieve low cost targets and maintain a high power density.
In this regard, four regions have been identified in the figure
the cathode side. For comparison, a durability test performed
under the same operating conditions is also shown for an MEA
with Pt/C at the cathode side. A similar current–time (or
power–time) behavior was registered for both MEAs. In the
course of the experiments, unexpected shut-downs occurred
as indicated in the figure, which may have slightly influenced
the cell behavior. The performance variation with time is very
similar between the two MEAs, accounting for about 50%
steady-state performance decay in 100 h in the case of the Fe-
ꢀ
1
in terms of specific power (WgPt-MEA ). It must be pointed out
that the different cell operating conditions reported in the
cited works lead to considerable differences in terms of power
density, above all the cell temperature (most of them in the
range 60–908C) and composition of the fuel/oxidant streams
N -C-THT cathode and about 45% in the case of the Pt/C cath-
x
(most of them 2m methanol and pure oxygen). The highest
ode.
power density has been reported by Wei et al., achieving
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
7
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
images were obtained using Hitachi S-800 and JEOL 2010 EX in-
Conclusions
struments, respectively. The surface areas were measured by N ad-
2
sorption BET using a Micrometrics 2360 Gemini Analyzer. A four-
point BET analysis was performed using a saturation pressure of
An aminobenzimidazole-derived iron PGM-free catalyst was
physicochemically modified after synthesis by additional acid
leaching and heat treatment steps. The activity towards the
ORR was notably improved after the treatments, as evidenced
in the half-cell experiments in acidic electrolyte. The physico-
chemical characterization, including XPS, BET, SEM, and TEM,
indicated an increase of the density of Fe-N moieties as the
cause of the improved catalytic activity. Experiments using ro-
tating ring disk electrodes indicated a shift of the ORR mecha-
6
40 mmHg. The X-ray photoelectron spectroscopy (XPS) spectra
were acquired on a Kratos Axis DLD Ultra X-ray photoelectron
spectrometer using a monochromatic AlKa source operating at
1
50 W with no charge compensation. The base pressure was 2ꢃ
ꢀ10
10 Torr, and the operating pressure was approximately 2ꢃ
ꢀ9
10 Torr. Survey and high-resolution spectra were acquired at pass
energies of 80 and 20 eV, respectively. Data analysis and quantifica-
tion were performed using CasaXPS software. A linear background
subtraction was used for the quantification of the spectra. Sensitiv-
ity factors provided by the manufacturer were utilized. A 70%
Gaussian/30% Lorentzian line shape was utilized in the curve-fit of
the high resolution spectra.
ꢀ
ꢀ
nism from a 2ꢃ2e to a 4e pathway as a consequence of the
post-treatments of the Fe-N-C catalyst.
The treated catalyst also showed an exceptional tolerance to
the presence of methanol. This was further investigated in
a complete single cell fed with methanol, that is, a DMFC con-
figuration, using the treated PGM-free catalyst at the cathode
at a wide range of temperatures and methanol concentrations.
The results indicated an outstanding performance, even at
high methanol concentration (10m), owing to improved ORR
activity and high tolerance to the alcohol. Additionally, a 100 h
experiment at high temperature showed a similar current–time
Electrochemical characterization (RDE and RRDE)
The electrochemical characterization was performed in a three-
electrode cell at room temperature for the PGM-free catalysts (Fe-
N -C-DHT and Fe-N -C-THT) and a benchmark Pt/C catalyst (40 wt%
x
x
Pt, Johnson Matthey) for comparison purposes. For the activity
analyses, a rotating disk electrode (RDE) was used as the working
electrode, consisting of a thin film of the catalyst deposited on
behavior to
a membrane–electrode assembly based on
2
a glassy carbon disk (geometric area of 0.196 cm ). The H O yield
2
2
a benchmark Pt cathode. In summary, even with a relatively
low PtRu content at the anode side, the presented results can
be considered state-of-the-art for high-performance and cost-
effective DMFCs.
was determined using a rotating ring disk electrode (RRDE) consist-
2
ing of a glassy carbon disk (geometric area of 0.247 cm ) and a plat-
inum ring. The catalytic layer was obtained using the following
ꢀ1
recipe: first, we prepared a 3 mgmL ink by sonicating the cata-
ꢄ
lyst in isopropyl alcohol/water (3/1, v/v) solution and Nafion (Ion
Power, 5 wt%) using an appropriate catalyst-to-ionomer ratio.
Some drops of this ink were deposited onto the glassy carbon disk
up to the desired mass loading for an optimum characteriza-
Experimental Section
Synthesis and physicochemical characterization of Fe-N -C
[56,57]
ꢀ2
ꢄ
x
tion:
600 mgcm and 15 wt% Nafion . An aqueous 0.5m
catalysts
H SO solution was used as an electrolyte, the reference electrode
2
4
was a mercury/mercury sulfate electrode (HgjHg SO , sat. K SO )
2
4
2
4
A mixture of two types of commercial silica materials was used as
and a high surface Pt coiled wire was used as counter electrode.
Tests performed using a high surface area graphite rod confirmed
that the ORR curves were identical regardless of the counter elec-
trode (Figure S4), which excluded any effect of eventual Pt redepo-
sition from the counter electrode on the ORR electrocatalytic activi-
ty of the PGM-free catalysts. An Autolab potentiostat/galvanostat
was used to perform the RDE electrochemical experiments. A Pine
Instruments bipotentiostat was used for the RRDE experiments,
using 1.5 V versus RHE as the ring potential. Linear sweep voltam-
metry curves were obtained in the potentiostatic mode with
2
ꢀ1
a sacrificial support (LM150 with surface area of 150 m g and
2
ꢀ1
A90 with surface area of 90 m g ). Four grams of iron nitrate
Fe(NO ) ·9H O, Sigma–Aldrich) was dissolved in 10 mL of acetone.
(
3
3
2
In a separate beaker, 25 g of aminobenzimidazole (ABZIM) was dis-
persed in 100 mL of acetone. The blend of silica materials was
added to a solution of ABZIM and a colloidal solution was obtained
after ultrasonic treatment for 4 h in a low energy ultrasonic bath.
Iron nitrate solution was added to the SiO /ABZIM suspension
2
under vigorous stirring. Acetone was evaporated and the mixture
was ball-milled in a planetary ball mill at 400 rpm for 2 h. The
finely homogenized mixture of precursors was heat treated in an
inert atmosphere of ultrahigh purity (UHP) nitrogen at a flow rate
ꢀ1
a scan rate of 5 mVs and at rotation rates from 100–2500 rpm.
The tolerance of the catalysts to the presence of methanol was
evaluated in RDE by adding increasing aliquots of the alcohol to
the base electrolyte, saturated with oxygen, at concentrations of
ꢀ1
of 150 mLmin at 9458C for 60 min. The mixture of silica was re-
moved by means of 25 wt% of HF for 7 days. The powder was
washed with deionized water until neutral pH was achieved. To
remove the low soluble but volatile silica compounds, a second
0
.001–2m. The ORR response in the presence of methanol was
evaluated at a rotation speed of 1600 rpm.
treatment in ammonia atmosphere (10% of NH ) was performed at
3
9
758C for 45 min. The obtained material was denoted as Fe-N -C-
x
Fuel cell testing (DMFC)
DHT (DHT: double heat treated). An additional treatment with 4m
HNO was performed at room temperature for 48 h to remove iron
For single cell experiments, MEAs were prepared with the PGM-
3
nanoparticles coated with a graphitic layer (centers for Fenton
free Fe-N -C-THT catalyst at the cathode side. Cathode electrodes
x
type radical production). The remaining HNO material was washed
were obtained by spraying a catalytic ink on a commercial hydro-
phobic gas diffusion layer (GDL-LT, E-TEK). The catalyst ink was pre-
pared by sonicating the catalyst in an isopropyl alcohol/water mix-
3
and heat treated a third time in an atmosphere of 7 at% NH at
3
9
758C for 15 min. The as obtained catalyst was used in the present
ꢄ
study and denoted as Fe-N -C-THT (THT: triple heat treated). This
ture (3/1, v/v) and Nafion solution. The cathode electrode loading
x
ꢀ
2
ꢄ
latter catalyst was physicochemically characterized. SEM and TEM
for the Fe-N -C-THT was 4.5ꢁ0.2 mgcm and the Nafion content
x
&
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
8
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
[
58]
in the catalytic layer was 45 wt%. Anode electrodes based on
PtRu black (Pt:Ru 1:1, Johnson Matthey) were prepared according
[18] A. A. Serov, M. Min, G. Chai, S. Han, S. J. Seo, Y. Park, H. Kim, C. Kwak, J.
Appl. Electrochem. 2009, 39, 1509–1516.
[19] Y. Wei, C. Shengzhou, L. Weiming, Int. J. Hydrogen Energy 2012, 37,
942–945.
[59]
to the procedure described in a previous report. The catalytic
layer was composed of 85 wt% catalyst and 15 wt% Nafion ion-
omer spread onto a commercial gas diffusion layer (GDL-HT, E-
TEK). The noble metal (Pt+Ru) loading at the anode was
ꢄ
[
[
[
20] E. Negro, A. H. A. M. Videla, V. Baglio, A. S. Aricꢁ, S. Specchia, G. J. M.
Koper, Appl. Catal. B 2015, 166–167, 75–83.
21] J. Zhu, M. Xiao, C. Liu, J. Ge, J. St-Pierre, W. Xing, J. Mater. Chem. A
ꢀ
2
1
mgcm .
2
015, 3, 21451–21459.
22] M. Xiao, J. Zhu, L. Feng, C. Liu, W. Xing, Adv. Mater. 2015, 27, 2521–
527.
MEAs were assembled by a hot-pressing procedure at 1308C and
0 kgf cm for 10 min, and subsequently installed in a 5 cm fuel
cell test fixture (Fuel Cell Tech., Inc.). A Nafion 115 membrane
ꢀ
2
2
3
2
ꢄ
[23] Y. Hu, J. Zhu, Q. Lv, C. Liu, Q. Li, W. Xing, Electrochim. Acta 2015, 155,
335–340.
[24] J. Liu, X. Sun, P. Song, Y. Zhang, W. Xing, W. Xu, Adv. Mater. 2013, 25,
(
ꢃ130 mm thick) was used as the solid electrolyte; the anode load-
ꢀ2
ing was 1 mg PtRu cm (PtRu black, Johnson Matthey) for all the
MEAs. The cell hardware was connected to a Fuel Cell Tech., Inc.
test station. In case of single cell polarization experiments, aqueous
methanol (1–10m) was pre-heated at the same temperature of the
cell and fed to the anode chamber of the DMFC through a peristal-
tic pump; oxygen was also pre-heated at the same temperature of
the cell (fully humidified) and fed to the cathode side. The reactant
6879–6883.
[
[
[
[
25] R. Janarthanan, A. Serov, S. K. Pilli, D. A. Gamarra, P. Atanassov, M. R.
Hibbs, A. M. Herring, Electrochim. Acta 2015, 175, 202–208.
26] T. S. Olson, B. Blizanac, B. Piela, J. R. Davey, P. Zelenay, P. Atanassov, Fuel
Cells 2009, 9, 547–553.
27] B. Piela, T. S. Olson, P. Atanassov, P. Zelenay, Electrochim. Acta 2010, 55,
7615–7621.
ꢀ
1
flow rates were 2 and 100 mLmin for the methanol/water mix-
ture and oxygen streams, respectively. The cell temperature was
measured by a thermocouple embedded in the cathodic graphite
plate close to the MEA. Steady-state galvanostatic polarization ex-
periments in DMFC were performed with an Agilent electronic
load at various temperatures and methanol concentrations.
28] M. Shen, L.-R. Zheng, W. He, C. Ruan, C. Jiang, K. Ai, L. Lu, Nano Energy
2015, 17, 120–130.
[29] X. Sun, P. Song, Y. Zhang, C. Liu, W. Xu, W. Xing, Sci. Rep. 2013, 3, 2505.
[30] X. Sun, Y. Zhang, P. Song, J. Pan, L. Zhuang, W. Xu, W. Xing, ACS Catal.
2013, 3, 1726–1729.
[
31] S. Zhao, H. Yin, L. Du, L. He, K. Zhao, L. Chang, G. Yin, H. Zhao, S. Liu, Z.
Tang, ACS Nano 2014, 8, 12660–12668.
[
32] H. Zhong, J. Wang, Y. Zhang, W. Xu, W. Xing, D. Xu, Y. Zhang, X. Zhang,
Angew. Chem. Int. Ed. 2014, 53, 14235–14239; Angew. Chem. 2014, 126,
Acknowledgements
14459–14463.
[
[
[
[
[
33] I. Kruusenberg, S. Ratso, M. Vikkisk, P. Kanninen, T. Kallio, A. M. Kannan,
K. Tammeveski, J. Power Sources 2015, 281, 94–102.
34] K. Wan, G.-F. Long, M.-Y. Liu, L. Du, Z.-X. Liang, P. Tsiakaras, Appl. Catal. B
CNR-ITAE authors acknowledge the financial support of
PRIN2010-11 project ‘Advanced nanocomposite membranes and
innovative electrocatalysts for durable polymer electrolyte mem-
brane fuel cells’ (NAMED-PEM).
2
015, 165, 566–571.
35] M. Lei, J. Wang, J. R. Li, Y. G. Wang, H. L. Tang, W. J. Wang, Sci. Rep. 2014,
, 6013.
36] D. Sebastiꢀn, V. Baglio, A. S. Aricꢁ, A. Serov, P. Atanassov, Appl. Catal. B
016, 182, 297–305.
4
2
Keywords: direct methanol fuel cell
· electrocatalyst ·
37] A. Serov, M. H. Robson, B. Halevi, K. Artyushkova, P. Atanassov, Electro-
chem. Commun. 2012, 22, 53–56.
methanol tolerance · oxygen reduction reaction · post-
treatment
[38] F. Jaouen, V. Goellner, M. Lefꢅvre, J. Herranz, E. Proietti, J. P. Dodelet,
Electrochim. Acta 2013, 87, 619–628.
[
[
39] G. C.-K. Liu, J. R. Dahn, Appl. Catal. A 2008, 347, 43–49.
40] K. Wan, Z. Yu, X. Li, M. Liu, G. Yang, J. Piao, Z. Liang, ACS Catal. 2015, 5,
[
[
[
[
[
1] A. S. Aricꢁ, P. Bruce, B. Scrosati, J.-M. Tarascon, W. van Schalkwijk, Nat.
Mater. 2005, 4, 366–377.
2] N. Radenahmad, A. Afif, P. I. Petra, S. M. H. Rahman, S.-G. Eriksson, A. K.
Azad, Renewable Sustainable Energy Rev. 2016, 57, 1347–1358.
3] Y. Chen, M. Bellini, M. Bevilacqua, P. Fornasiero, A. Lavacchi, H. A. Miller,
L. Wang, F. Vizza, ChemSusChem 2015, 8, 524–533.
4325–4332.
[
41] U. Tylus, Q. Jia, K. Strickland, N. Ramaswamy, A. Serov, P. Atanassov, S.
Mukerjee, J. Phys. Chem. C 2014, 118, 8999–9008.
[
[
42] J. Jiang, B. Yi, J. Electroanal. Chem. 2005, 577, 107–115.
43] N. M. Markovic, H. A. Gasteiger, P. N. Ross, J. Phys. Chem. 1995, 99,
4] S. K. Kamarudin, F. Achmad, W. R. W. Daud, Int. J. Hydrogen Energy 2009,
3
411–3415.
44] V. Stamenkovic, N. M. Markovic, P. N. Ross, J. Electroanal. Chem. 2001,
00, 44–51.
45] M. H. Robson, A. Serov, K. Artyushkova, P. Atanassov, Electrochim. Acta
013, 90, 656–665.
3
4, 6902–6916.
5] P. Joghee, J. N. Malik, S. Pylypenko, R. O’Hayre, MRS Energy Sustainable
015, 2, E3.
[
[
[
[
[
5
2
[
[
[
[
6] J. Han, H. Liu, J. Power Sources 2007, 164, 166–173.
2
7] C. Y. Du, T. S. Zhao, W. W. Yang, Electrochim. Acta 2007, 52, 5266–5271.
8] B. Gurau, E. S. Smotkin, J. Power Sources 2002, 112, 339–352.
9] D. S. Falc¼o, V. B. Oliveira, C. M. Rangel, A. M. F. R. Pinto, Renewable Sus-
tainable Energy Rev. 2014, 34, 58–70.
46] K. Artyushkova, A. Serov, S. Rojas-Carbonell, P. Atanassov, J. Phys. Chem.
C 2015, 119, 25917–25928.
47] D. B. Sepa, M. V. Vojnovic, A. Damjanovic, Electrochim. Acta 1981, 26,
781–793.
[
[
10] L. Li, S. Liu, A. Manthiram, Nano Energy 2015, 12, 852–860.
11] S. Liu, L. Li, H. S. Ahn, A. Manthiram, J. Mater. Chem. A 2015, 3, 11615–
48] A. Serov, K. Artyushkova, N. I. Andersen, S. Stariha, P. Atanassov, Electro-
chim. Acta 2015, 179, 154–160.
11623.
[
[
[
12] M. Jahan, Q. Bao, K. P. Loh, J. Am. Chem. Soc. 2012, 134, 6707–6713.
13] N. A. Karim, S. K. Kamarudin, Appl. Energy 2013, 103, 212–220.
14] G. Wu, A. Santandreu, W. Kellogg, S. Gupta, O. Ogoke, H. Zhang, H.-L.
Wang, L. Dai, Nano Energy 2016, DOI: 10.1016/j.nanoen.2015.12.032.
15] Z. Chen, D. Higgins, A. Yu, L. Zhang, J. Zhang, Energy Environ. Sci. 2011,
[49] D. Sebastiꢀn, A. Serov, K. Artyushkova, P. Atanassov, A. S. Aricꢁ, V.
Baglio, J. Power Sources 2016, 319, 235–246.
[50] E. Antolini, T. Lopes, E. R. Gonzalez, J. Alloys Compd. 2008, 461, 253–
262.
[51] T. S. Olson, S. Pylypenko, S. Kattel, P. Atanassov, B. Kiefer, J. Phys. Chem.
C 2010, 114, 15190–15195.
[52] A. S. Arico, V. Baglio, V. Antonucci, Direct Methanol Fuel Cells, Nova Sci-
ence Publishers, Inc., 2010.
[
[
[
4, 3167.
16] F. Jaouen, E. Proietti, M. Lefꢅvre, R. Chenitz, J.-P. Dodelet, G. Wu, H. T.
Chung, C. M. Johnston, P. Zelenay, Energy Environ. Sci. 2011, 4, 114–130.
17] S. Baranton, C. Coutanceau, J.-M. Lꢆger, C. Roux, P. Capron, Electrochim.
Acta 2005, 51, 517–525.
[53] A. S. Aricꢁ, A. Stassi, C. D’Urso, D. Sebastiꢀn, V. Baglio, Chem. Eur. J.
2014, 20, 10679–10684.
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
9
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
[
54] R. Carrera-Cerritos, V. Baglio, A. S. Aricꢁ, J. Ledesma-Garcꢇa, M. F. Sgroi,
D. Pullini, A. J. Pruna, D. B. Mataix, R. Fuentes-Ramꢇrez, L. G. Arriaga,
Appl. Catal. B 2014, 144, 554–560.
[58] A. Serov, K. Artyushkova, P. Atanassov, Adv. Energy Mater. 2014, 4,
1301735.
[59] V. Baglio, C. D’Urso, D. Sebastiꢀn, A. Stassi, A. S. Aricꢁ, Int. J. Hydrogen
Energy 2014, 39, 5399–5405.
[
[
[
55] A. Mehmood, M. A. Scibioh, J. Prabhuram, M.-G. An, H. Y. Ha, J. Power
Sources 2015, 297, 224–241.
56] A. Serov, U. Tylus, K. Artyushkova, S. Mukerjee, P. Atanassov, Appl. Catal.
B 2014, 150–151, 179–186.
57] A. Bonakdarpour, M. Lefevre, R. Yang, F. Jaouen, T. Dahn, J.-P. Dodelet,
J. R. Dahn, Electrochem. Solid-State Lett. 2008, 11, B105.
Received: May 2, 2016
Revised: May 20, 2016
Published online on && &&, 0000
&
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
10
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

FULL PAPERS
No nobles here! Platinum group metal-
free materials based on Fe-N-C are in-
vestigated as cost-effective, active,
stable and methanol-tolerant catalysts
for the oxygen reduction reaction (ORR)
in direct methanol fuel cells (DMFCs).
Outstanding performance of DMFCs is
observed, even at high methanol con-
centration (10m), owing to improved
ORR activity and high tolerance to the
alcohol.
D. Sebastiꢀn, A. Serov, K. Artyushkova,
J. Gordon, P. Atanassov,* A. S. Aricꢁ,
V. Baglio*
&& – &&
High Performance and Cost-Effective
Direct Methanol Fuel Cells: Fe-N-C
Methanol-Tolerant Oxygen Reduction
Reaction Catalysts
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
11
ꢂ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ