Electrooxidation of ethylene glycol and glycerol on Pd-(Ni-Zn)/C anodes in direct alcohol fuel cells

Article abstract of DOI:10.1002/cssc.201200866

The electrooxidation of ethylene glycol (EG) and glycerol (G) has been studied: in alkaline media, in passive as well as active direct ethylene glycol fuel cells (DEGFCs), and in direct glycerol fuel cells (DGFCs) containing Pd-(Ni-Zn)/C as an anode electrocatalyst, that is, Pd nanoparticles supported on a Ni-Zn phase. For comparison, an anode electrocatalyst containing Pd nanoparticles (Pd/C) has been also investigated. The oxidation of EG and G has primarily been investigated in half cells. The results obtained have highlighted the excellent electrocatalytic activity of Pd-(Ni-Zn)/C in terms of peak current density, which is as high as 3300A g_(Pd)^-1 for EG and 2150A g_(Pd)^-1 for G. Membrane-electrode assemblies (MEA) have been fabricated using Pd-(Ni-Zn)/C anodes, proprietary Fe-Co/C cathodes, and Tokuyama A-201 anion-exchange membranes. The MEA performance has been evaluated in either passive or active cells fed with aqueous solutions of 5wt % EG and 5wt % G. In view of the peak-power densities obtained in the temperature range from 20 to 80 °C, at Pd loadings as low as 1mg cm ^-2 at the anode, these results show that Pd-(Ni-Zn)/C can be classified amongst the best performing electrocatalysts ever reported for EG and G oxidation. Copyright

Full text of DOI:10.1002/cssc.201200866

CHEMSUSCHEM
FULL PAPERS
DOI: 10.1002/cssc.201200866
Electrooxidation of Ethylene Glycol and Glycerol on
Pd-(Ni-Zn)/C Anodes in Direct Alcohol Fuel Cells
Andrea Marchionni,[a] Manuela Bevilacqua,[a] Claudio Bianchini,[a] Yan-Xin Chen,[a, b]
Jonathan Filippi,[a] Paolo Fornasiero,[b] Alessandro Lavacchi,[a] Hamish Miller,[a]
Lianqin Wang,[a, b] and Francesco Vizza*[a]
The electrooxidation of ethylene glycol (EG) and glycerol (G)
has been studied: in alkaline media, in passive as well as active
direct ethylene glycol fuel cells (DEGFCs), and in direct glycerol
fuel cells (DGFCs) containing Pd-(Ni-Zn)/C as an anode electro-
catalyst, that is, Pd nanoparticles supported on a Ni–Zn phase.
For comparison, an anode electrocatalyst containing Pd nano-
particles (Pd/C) has been also investigated. The oxidation of EG
and G has primarily been investigated in half cells. The results
obtained have highlighted the excellent electrocatalytic activity
of Pd-(Ni-Zn)/C in terms of peak current density, which is
Membrane-electrode assemblies (MEA) have been fabricated
using Pd-(Ni-Zn)/C anodes, proprietary Fe-Co/C cathodes, and
Tokuyama A-201 anion-exchange membranes. The MEA perfor-
mance has been evaluated in either passive or active cells fed
with aqueous solutions of 5 wt% EG and 5 wt% G. In view of
the peak-power densities obtained in the temperature range
from 20 to 808C, at Pd loadings as low as 1 mgcmÀ2 at the
anode, these results show that Pd-(Ni-Zn)/C can be classified
amongst the best performing electrocatalysts ever reported for
EG and G oxidation.
À1
À1
as high as 3300 Ag(Pd)
for EG and 2150 Ag(Pd)
for G.
Introduction
The selective conversion/production of chemicals from renew-
able resources with simultaneous production of energy is per-
haps one of the most desirable targets in sustainable chemis-
try.[1] The realization of such processes will provide energy with
no overall CO2 emissions and at
about the partial oxidation of the anolyte with high selectivity
and fast kinetics.
The selective conversion of ethanol into acetic acid has been
demonstrated in direct ethanol fuel cells (DEFCs) equipped
with anion-exchange membranes. In the most successful cases,
Pd metal was the major component of the anode electrocata-
lyst. Alcohols with a higher molecular weight than ethanol are
now arousing great interest as fuels in direct alcohol fuel cells
(DAFC) for various reasons, among which the relatively low
toxicity, high boiling point, high specific energy, and the ca-
pacity of some of them to be renewable. Included in this
group are ethylene glycol (EG) and glycerol (G).
the same time lead to the pro-
duction of industrially relevant
feedstocks. Indeed, starting from
renewable alcohols, a large varie-
ty of products such as aldehydes,
ketones, and carboxylic acids
might be obtained. Recent results
have shown that direct fuel cells
EG has a volumetric energy density of 5.9 KWhLÀ1 and can
be produced by heterogeneous hydrogenation of cellulose de-
rivatives.[2,3] G has a volumetric energy density of 6.3 KWhLÀ1
and is the by-product of biodiesel production. As such, it is in-
expensive (0.3 US$kgÀ1) and largely available (2.4 million
tonnes per year). Moreover, the increasing demand for methyl
esters as fuel additives is expected to increase the G produc-
tion, thus leading to further cost reduction.[4] It is also impor-
tant to stress that some EG and G oxidation products, such as
glycolic acid, oxalic acid, glyceric acid, tartronic acid, mesoxalic
acid, and 1,3 dihydro-acetone, are valuable fine chemicals, cur-
rently produced by costly and non-environmentally friendly
processes. Indeed, G oxidation reactions are generally per-
formed with stoichiometric oxidants, such as permanganate,
nitric acid or chromic acid, and are characterized by low yields
and selectivity.[5] Fermentation processes exhibit higher oxida-
tion selectivity, but have a low conversion rate. Alternatively,
one may apply heterogeneous catalytic processes with molec-
can be effectively employed to
convert alcohols into various oxy-
genates, providing at the same
time good power densities, ex-
COVER
ploiting the ability of certain
anode electrocatalysts to bring
[a] Dr. A. Marchionni, Dr. M. Bevilacqua, Dr. C. Bianchini, Dr. Y.-X. Chen,
Dr. J. Filippi, Dr. A. Lavacchi, Dr. H. Miller, Dr. L. Wang, Dr. F. Vizza
Istituto di Chimica dei Composti Organometallici
Consiglio Nazionale delle Ricerche
via Madonna del Piano 10, 50019 Sesto Fiorentino (Italy)
Fax: (+39)0555225203
E-mail: francesco.vizza@iccom.cnr.it
[b] Dr. Y.-X. Chen, Dr. P. Fornasiero, Dr. L. Wang
Department of Chemistry
ICCOM-CNR-Trieste
CENMAT, University of Trieste
via L. Giorgeri 1, 34127 Trieste (Italy)
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 518 – 528 518

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
ular oxygen as oxidant, but these processes are generally char-
acterized by harsh conditions and low selectivity, and as a con-
sequence low sustainability.[6–10]
ated in a DGFC that gave rise to a peak power density of
7 mWcmÀ2 at room temperature and 80 mWcmÀ2 at 808C.[24]
To the best of our knowledge, all known Pd- or Pt-based
electrocatalysts cannot promote the complete oxidation of EG
or G to CO2 in alkaline media. EG or G may undergo CÀC bond
scission with formation of CO2, yet this is invariably a minor re-
action pathway compared to the oxidation to carboxylates
which, in turn, is not completely selective. There is little doubt
that the nature, structure, and composition of the anode elec-
trocatalyst have a crucial effect on the oxidation in terms of
power output, Faradaic efficiency, and product selectivity.
In this respect, much research work is still needed to design
and develop improved catalysts for polyalcohol oxidation in
DAFCs.
The conversion of G or EG into oxygenated compounds in
DAFCs has been achieved on Pt-based anodes, yet none of
these catalysts has demonstrated the capacity to produce both
high power densities and high product selectivities. Recently,
Li and co-workers[11,12] have reported a maximum power densi-
ty of 71 mWcmÀ2 at 508C in a direct ethylene glycol fuel cell
(DEGFC) and a maximum power density of 124.5 mWcmÀ2 at
808C in a direct glycerol fuel cell (DGFC), both equipped with
an anion-exchange membrane, Pt/C anode, and Fe-Cu/C cath-
ode. In their elegant articles, these authors have provided evi-
dence for a good selectivity towards higher-value chemicals by
tuning the fuel cell operation voltage. Martemianow et al.[13]
have improved the electrode membrane assembly (MEA) fabri-
cation (Pt and Pt-based bimetallic anode) in a DGFC and se-
lected appropriate experimental conditions (i.e., fuel composi-
tion, fuel flow) to optimize the cell performance, so as to
obtain a power density of 24 mWcmÀ2 at 808C. Atanassov
et al.[14] have investigated the electrooxidation of EG and G on
Pt-based binary and ternary nanostructured catalysts synthe-
sized by spray pyrolysis. These authors were able to demon-
strate that the addition of Ru and Sn to Pt has a beneficial
effect on the oxidation of EG and G. Koper et al.[15] have ob-
Herein, we describe the cogeneration of electrical energy
and higher-value chemicals from the electrooxidation of EG
and G in both passive and active alkaline DAFCs in which Ni
foam anodes are coated with a Pd-(Ni-Zn)/C catalyst and the
cathodes are noble metal free, containing exclusively a FeÀCo/
C (C=Ketjen Black) catalyst. For comparative purposes, we
have also investigated the EG and G electrooxidation on
a nanostructured Pd/C catalyst prepared by chemical reduction
of H2PdCl4 adsorbed on C.[25]
tained a very high selectivity in the oxidation of G to dihydrox- Results and Discussion
yacetone in acidic media on a Pt/C catalyst in the presence of
Electrochemical oxidation of ethylene glycol and glycerol on
Pd(Ni-Zn)/C and Pd/C in half cells
Bi dissolved in the electrolytic solution.
Palladium-based electrocatalysts are the best materials for
the oxidation of G and EG. Shen and Xu were the first authors
to demonstrate that Pd adsorbed onto oxide/C materials yield
electrocatalysts for alcohol oxidation with a much higher cata-
lytic activity (i.e., onset oxidation potential and peak current
density) and electrochemical stability with respect to any other
known Pd/C or Pt/C catalyst under comparable experimental
conditions (C=Vulcan).[16–18] Out of all the electrocatalysts in-
vestigated, Pd-Co3O4 showed the highest activity for the elec-
trooxidation of EG and G.[19] The synergistic effect of palladium
alloyed to platinum has been observed by Coutanceau et al.
for the oxidation of EG in alkaline media in both, half-cell and
DAFC tests.[20,21] These authors have also studied the capacity
of Pd to depress CÀC bond cleavage and enhance the catalyst
lifetime, by investigating the EG oxidation on PtPd, PtPdBi, and
PtBi catalysts. On these electrocatalysts, EG is converted into
mixtures of glycolic acid, glyoxylic acid, oxalic acid, and formic
acid, but the relative composition depends upon the anode
catalyst employed. It has been proposed that Bi favors the ad-
sorption of OHÀ species and also affects the product distribu-
tion by depressing CÀC bond cleavage. A mechanism of EG ox-
idation has been proposed by Coutanceau.[22]
In a previous publication, we described the synthesis of a cata-
lytic material obtained by a redox trans-metalation process in-
volving the spontaneous deposition of palladium onto Ni–Zn
alloys supported on Vulcan XC-72. The characterization of the
resulting Pd-(Ni-Zn)/C material using high-resolution TEM
(HRTEM; see Figure 1), extended X-ray absorption fine struc-
ture (EXAFS), powder X-ray powder diffraction (XPRD), and X-
ray absorption near edge structure (XANES) techniques has
shown that its surface contains small (0.5–1 nm from EXAFS es-
timation and average diameter of 2.3 nm from HRTEM, as re-
ported in Figure 1), highly dispersed aggregates of Pd and Ni
clusters as well as single Pd sites, probably stabilized by the in-
teraction with oxygen atoms from Ni–O moieties.[25] As a refer-
ence material, nanostructured Pd/C was also synthesized by
the reduction of an aqueous solution of PdCl2/HCl with ethyl-
ene glycol in the presence of Vulcan XC-72. In Pd/C the parti-
cles are larger, less dispersed, and much less crystalline than
Pd-(Ni-Zn)/C. In the work presented here we investigated the
electrochemical performance of these catalysts in the electro-
oxidation of EG and G in half-cells as well as in passive and
active polymer electrolyte membrane fuel cells, with the aim
to achieve cogeneration of electricity and valuable chemicals.
The electrochemical activity of the nano-structured catalysts
Pd-(Ni-Zn)/C and Pd/C for EG and G oxidation was first investi-
gated by cyclic voltammetry (CV) at room temperature in de-
oxygenated 2m KOH solutions. A series of cyclic voltammo-
grams recorded at KOH concentrations ranging from 0.5 to 4m
show the generation of the highest current density in the
We have recently reported that the electrooxidation of EG
on Pd/C, Pd (Ni-Zn)/C anodes, and on a smooth Pd electrode is
dramatically influenced by the pH value of the anolyte as well
as the catalyst structure.[23] Recently, we also published data on
the activity of palladium nanoparticles supported on multi-
walled carbon nanotubes (Pd/MWNCT) towards the electrooxi-
dation of alcohols. The performance of Pd/MWNCT was evalu-
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 518 – 528 519

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Table 1. CV data for the electrooxidation of 5 wt% EG and G in 2m KOH
on Pd-(Ni-Zn)/C and Pd/C.
Fuel
Catalyst
Jp
Sap
[mAmgPd
Eonset
[V]
Ep
[V]
À1
[mAcmÀ2
]
]
EG
G
Pd/C
Pd-(Ni-Zn)/C
Pd/C
64.0
84.2
55.5
65.3
2.76
3.33
2.10
2.15
0.42
0.39
0.41
0.42
0.83
0.90
0.95
0.92
Pd-(Ni-Zn)/C
Figure 2. Cyclic voltammograms (tenth cycle) of EG oxidation on Pd-(Ni-Zn)/
&
*
C ( ) and Pd/C ( )electrodes in 2m KOH and 5 wt% EG. Scan rate:
50 mVsÀ1
.
trocatalysts both show high activity towards the oxidation of
EG with an onset potential at approximately 0.4 V (versus RHE)
À1
and high specific current densities of 3.33 mAmg(Pd)
for
Figure 1. a) HRTEM histograms of particle volume distribution versus diame-
ter for Pd-(Ni-Zn)/C. The average size (d) and standard deviation (s) are re-
ported. b) HRTEM images of Pd-(Ni-Zn)/C at different magnifications; spots
of energy-dispersive X-ray (EDX) analysis are shown in circles; Pd/Ni relative
compositions as detected by EDX integrated peak areas: 1) Ni>Pd with Ni
(110) plane fringes; 2) Ni(111) plane fringes; 3) Pd/Ni=1:2; 4) Pd/Ni=1:1;
5) Pd with cuboctahedron crystal facets; 6) Ni; 7) Pd (110) plane fringes;
8) Pd>Ni; minor O contribution are also present in all cases. From Ref. [25].
À1
Pd-(Ni-Zn)/C and of 2.76 mAmg(Pd) for Pd/C. As shown by the
trace in Figure 2, the electrocatalyst Pd-(Ni-Zn)/C is more active
than Pd/C especially in terms of forward anodic peak potential
with the intensity controlled by the progressive coverage of
the two electrodes surfaces by PdO at high potentials.[25] The
reverse scans show two further anodic peaks owing to the oxi-
dation of fresh EG on the catalysts freed from PdO.[28] Accord-
ingly, Pd-(Ni-Zn)/C and Pd/C can be considered amongst the
best performing electrocatalysts ever reported for EG oxida-
tion. Figure 3 shows the cyclic voltammograms of the G oxida-
tion reaction on Pd/C and Pd-(Ni-Zn)/C electrodes. As expect-
ed, the oxidation of G starts at the potential corresponding to
the coverage of the electrode surface by OH groups.[23–25,29,30]
As shown by the traces in Figure 3, the two electrocatalysts
are characterized by rather similar activities, especially in terms
of forward and backward anodic peak potentials, with onset
range of 1–1.5m KOH. However, all the following electrochemi-
cal studies, either in half-cell or in monoplanar cell (passive or
active), were performed using a 2m KOH solution to keep the
concentration of the OHÀ group as constant as possible in all
experiments. The oxidation of EG and G in alkaline media con-
sumes OHÀ groups, irrespective of the oxidation level of the
substrate.[27] All the CV measurements were performed at
a sweep rate of 50 mVsÀ1 in solutions containing 5 wt% EG or
5 wt% G using glassy carbon electrodes with a Pd loading be-
tween 24 and 30 mgcmÀ2, corresponding to 3.5–4.0 mg of cata-
lyst ink. Relevant electrochemical parameters such as peak cur-
rent density (Jp), specific peak current density (Sap), and for-
ward anodic peak potential (Ep), are listed in Table 1 for the
fuels investigated.
potentials of around 0.41 V (versus RHE) and specific current
À1
densities higher than 2.0 mAmg(Pd)
(see Table 1), which,
though they are lower than those obtained for EG, are remark-
able results indeed. We are not aware of such a performance
in G oxidation in half cell experiments with any known electro-
catalyst (Table 2) especially at the very low metal loadings used
in this work. The only exception is a Pd/MWNCT anode cata-
lyst,[24] even if the data reported in Table 2 are not exactly com-
The cyclic voltammograms of EG oxidation on Pd/C and
Pd-(Ni-Zn)/C electrodes are shown in Figure 2.[23] The two elec-
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 518 – 528 520

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Pd-(Ni-Zn)/C[25] have unambiguously shown that the Ni-based
supports without Pd, though able to oxidize ethanol in alkaline
media, are not directly involved in the ethanol oxidation reac-
tion occurring at the largely negative potentials of the Pd-con-
taining electrocatalysts. It has been proposed that Ni increases
the amount of OHads groups on the catalyst surface, which
would favor the formation of the carboxylic acid by coupling
with acylads species. Adsorbed OHÀ species are required to re-
lease the carboxylic acid. Based on these considerations,
a mechanism for both the electrooxidation of EG to glycolate
and G to glycerate is shown in Scheme 1.
Passive oxygen-breathing DAFCs fuelled with EG or G con-
taining Pd-(Ni-Zn)/C or Pd/C-catalyzed anodes
Figure 3. Cyclic voltammograms (tenth cycle) of G oxidation on Pd-(Ni-Zn)/C
The MEAs used for the passive monoplanar DEGFCs were real-
ized (as described in the Experimental Section) with the anode
reaction catalyzed by Pd-(Ni-Zn)/
À1
&
*
( ) and Pd/C ( ) electrodes in 2m KOH and 5 wt% G. Scan rate: 50 mVs
.
C or Pd/C, Fe-Co/C as cathode,
Table 2. CV data for Pt and Pd-based anode catalysts for G electrooxidation.
and a Tokuyama A-201 anion ex-
À1
change-membrane.
Catalyst
Anolyte
Load
[mgcmÀ2
J [mAcmÀ2
]
Sa [mAmg(Pd)
peak
]
Ref.
]
peak
@0.5 V
@0.5 V
The potentiodynamic and
power density curves obtained
with 5 wt% EG in a 2m KOH so-
lution at room temperature (20–
228C) are reported in Figure 4.
In the curves it is demonstrated
that the cell with the Pd-(Ni-Zn)/
C electrocatalyst exhibits a con-
siderably higher open circuit
voltage (OCV) and peak power
density (24 mWcmÀ2 at 0.17 V)
compared to the cell containing
Pd-(NiZn)-C
Pd/C
Pd-NiO/C
Pt/C 40%
Pd/C 36%
5 wt% G; 2m KOH
5 wt% G; 2m KOH
1m G; 1m KOH
1m G; 1m KOH
0.1m G; 1m KOH
0.1m G; 1m KOH
0.1m G; 1m KOH
0.1m G; 1m KOH
0.1m G; 1m KOH
1m G; 1m KOH
5% G; 2m KOH
25
31.1
300
81.5
141.5
141.5
141.5
141.5
141.5
300
53.5
65.3
88.1
32.5
24.6
24.3
31.2
31.4
38.5
71.9
47.6
1.0
1.6
0.0
1.3
0
1.48
1.17
12.2
12.4
1.85
2.16
2.140
2.100
0.294
0.399
0.174
0.172
0.220
0.222
0.272
0.240
2.8
0.040
0.050
0
0.016
0
0.010
0.008
0.086
0.088
0.006
0.127
this work
this work
[16]
[12]
[22]
[22]
[22]
[22]
[22]
Pt/C 37%
Pd0.9Bi0.1/C 38%
Pt0.9Bi0.1/C 36%
Pt0.45Pd0.45Bi0.1/C 38%
Pd/CPAA 50%
Pd/MWCNT
[31]
[24]
17.0
parable because the experiments were not performed with the
same metal loading. Also in this case, as for the oxidation of
EG, the backward scan peak can be associated with the oxida-
tion of freshly chemisorbed G. Previous CV studies of ethanol
oxidation on Pd have shown that the decrease in current at
high voltages is mainly owed to
the Pd/C electrocatalyst as anode (12 mWcmÀ2 at 0.13 V).
Values comparable or slightly higher (19, 22, and 28 mWcmÀ2)
have been reported by Countanceau et al., with DEGFCs (2m
EG in a 4 molcmÀ3 NaOH), containing Pt, PtBi, and PtPdBi
anode catalysts, respectively (overall metal loading
the formation of a PdO layer,
which blocks the adsorption of
the substrate.[28] The electroca-
talytic activity is restored in the
negative-going sweep once the
potential reaches the value at
which PdO is reduced to Pd.
At a fixed potential of 0.7 V,
current densities of 0.47 and
À1
0.33 mAmg(Pd) for Pd-(Ni-Zn)/C
and Pd/C, were measured (see
Figure 3). The difference of 40%
in power density may explain
the power output supplied by
the two electroctalysts in both
passive and active DGFCs inves-
tigated (vide infra). Previous
studies of ethanol oxidation on
Scheme 1. Proposed mechanism for the electrooxidation of EG to glycolate and G to glycerate on Pd in alkaline
media.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 518 – 528 521

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Figure 4. Potentiodynamic and power density curves at 258C for passive
Figure 5. Galvanostatic curves (at 20 mAcmÀ2) for the passive DEGFC con-
*
*
DEGFCs containing the Pd-(Ni-Zn)/C ( ) and Pd/C ( ) anode electrocatalysts
*
*
taining the Pd-(Ni-Zn)/C ( ) and Pd/C ( ) as anode electrocatalysts and
and fuelled with 2m KOH and 5 wt% EG solutions. The cathode was ex-
fuelled with 2m KOH and 5 wt% EG solutions.
posed to an oxygen atmosphere. Scan rate: 5 mVsÀ1
.
2 mgcmÀ2).[21] The power output for the cell with Pd-(Ni-Zn)/C
is higher than that with Pd/C in the whole potential range.
This indicates that the use of the (Ni-Zn)/C-containing electro-
catalyst leads to a significant increase in the fuel-cell efficiency.
This is confirmed by the energy released during the galvano-
static curves (see Figure 5). Galvanostatic experiments have
been performed at room temperature on the passive cells in
which the anode compartment was filled with 8.47 mmol of
EG. After a conditioning time of 1 h at the OCV, the circuit of
the DEGFC was closed and a constant current of 20 mAcmÀ2
was allowed to flow until the cell voltage reached zero. The so-
lution of fuel was not refueled during the experiment.
The energy released during the galvanostatic experiments
was 963 J for the cell with Pd-(Ni-Zn)/C and 381 J for the cell
with Pd/C. Accurate analytical quantitative and qualitative anal-
ysis of exhausts were performed by means of 13C{1H} NMR
spectroscopy and ionic chromatography, using reference mate-
rials and calibration curves.
Figure 6. 13C{1H} NMR spectra at room temperature of the anode solution of
a DEGFC with Pd-(Ni-Zn)/C after 10.2 h of a galvanostatic experiment at
20 mAcmÀ2
.
The DEGFC containing the Pd-(Ni-Zn)/C anode continued to
deliver constantly 20 mAcmÀ2 for 10.2 h, yielding glycolate
(55.4%), oxalate (37.6%), and carbonate (7.0%) with a total
conversion of 77.1% (Figure 6, Table 3). In contrast, the DEGFC
containing the Pd/C anode continued to deliver constantly
102 mA for 5.1 h, yielding glycolate (89.5%), oxalate (6.6%),
and carbonate (3.6%), with
Pd-(Ni-Zn)/C or Pd/C, independent CV experiments were per-
formed using solutions of potassium glycolate, potassium
glyoxylate, and potassium oxalate as anolytes in an alkaline
media.[23]
It was observed that the glycolate ion is much more difficult
to oxidize on Pd than EG. Slightly higher current densities
a
total conversion of 55.5%
(Table 3). Notably, no trace of
either glyoxaldehyde or hemice-
tal was detected at any stage of
the galvanostatic experiments.
The difference in product
Table 3. Product distribution of exhausts after galvanostatic experiments, total conversion, and released
energy.
Catalyst
Product selectivity [%]
Total con- Released
glycolate oxalate carbonate glycerate tartronate oxalate formate version [%] energy [J]
5 wt% EG; 2m KOH
Pd/C 89.5
Pd-(Ni-Zn)/C 55.4
distribution
provided
by
6.6
37.6
3.6
7
–
–
–
–
–
–
–
–
55.5
77.1
381
963
Pd-(Ni-Zn)/C and Pd/C seems to
be inconsistent with a similar
reaction mechanism. In our pre-
vious paper investigating the
electrooxidation of EG either on
5 wt% G; 2m KOH
Pd/C
Pd-(Ni-Zn)/C 4.2
11.8
13.6
14.1
28.8
24
15
27.2
28.4
28.1
13.6
14.1
2.4
2.4
42
55.5
366
578
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 518 – 528 522

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Scheme 2. Direct and sequential routes for the electrooxidation of EG.
were produced by glyoxylate, which can be an intermediate
along the conversion of EG and glycolate to oxalate, whereas
the oxalate was not oxidized to carbonate. Based on these
considerations, as well as on the distribution of the products
observed after the galvanostatic curves, a mechanism for elec-
trooxidation of EG is proposed in Scheme 2.
Pd/C is a selective electrocatalyst, yielding 89.5% of glyco-
late,
following
preferably
path a
in
Scheme 2.
A mixture of glycolate (>50%), oxalate (37.6%), and carbonate
(7.0%) are obtained with Pd-(Ni-Zn)/C (see Table 3). Carbonate
is produced by oxidation of glycolate (route c), whereas the
major part of oxalate seems to be produced by the direct oxi-
dation of EG (path b). It is reasonable to assume that the Ni–
Zn phase supporting the Pd nanoparticles, by virtue of nature
of both Ni and Zn, might favor the chelating adsorption of EG
(path b), leading preferably to the direct oxidation to oxa-
late.[23,31]
Figure 7. Potentiodynamic and power density curves at 258C for passive
*
*
DGFCs containing the Pd-(Ni-Zn)/C ( ) and Pd/C ( ) anode electrocatalysts
Polarization and power density curves obtained on passive
DGFCs, containing Pd-(Ni-Zn)/C or Pd/C anodes, Fe-Co/C catho-
des, and Tokuyama A-201 anion exchange-membranes, fuelled
with 5 wt% G in a 2m KOH solution and at room temperature
are reported in Figure 7. In this case, under comparable condi-
tions, the DGFC containing the Pd-(Ni-Zn)/C exhibits a higher
peak power density (16.5 mWcmÀ2) than the cell with the Pd/C
electrocatalyst (8.6 mWcmÀ2). To the best knowledge of the au-
thors this is the most efficient electrocatalyst reported so far
for the oxidation of G at room temperature. For a galvanostatic
experiment performed at 20 mAcmÀ2, without refueling, the
cell potential for Pd-(Ni-Zn)/C is higher than Pd/C (Figure 8).
Over 6.5 h, the energy released was 578 J for the cell with the
Pd-(Ni-Zn)/C. The following product distribution was deter-
mined by NMR and ionic chromatography (IC) analysis: glycer-
ate (27.2%), tartronate (28.1%), glycolate (4.2%), oxalate
(14.1%), formate (2.4%), and carbonate (24.0%), with a total
conversion of 55.5% (Figure 9, Table 3). In contrast, the DGFC
equipped with Pd/C provided a current for 5.16 h, releasing
366 J of energy. About 42% of G was consumed to yield glyc-
erate (15%), tartronate (28.4%), glycolate (11.8%), oxalate
(13.6), formate (2.4%), and carbonate (28.8%). The absence of
a marked difference in selectivity of G oxidation on Pd(Ni-Zn)/
C and Pd/C indicates that, contrary to EG oxidation, the reac-
tion mechanism is not substantially different for the two cata-
and fuelled with 2m KOH and 5 wt% G solutions. The cathode was exposed
to an oxygen atmosphere. Scan rate: 5 mVsÀ1
.
Figure 8. Galvanostatic curves (at 20 mAcmÀ2) for the passive DGFC contain-
*
*
ing the Pd-(Ni-Zn)/C ( ) and Pd/C ( ) as anode electrocatalysts and fuelled
with 2m KOH and 5 wt% G solutions.
lysts studied. Most importantly, both the DEGFCs and DGFCs
described above were fully regenerated (same OCV and galva-
nostatic curves) upon replacement of the cell exhausts with
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 518 – 528 523

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
&
*
Figure 10. Cyclic voltammograms of Pd-(Ni-Zn)/C ( ) and Pd/C ( ) in 2m
KOH (empty symbols) or in 2m KOH+2 wt% tartronic acid (filled symbols).
Scan rate: 50 mVsÀ1
.
Figure 9. 13C{1H} NMR spectra at room temperature of the anode solution of
a DGFC with Pd-(Ni-Zn)/C after 6.5 h of a galvanostatic experiment at
20 mAcmÀ2
.
dized to glyceraldehyde, which in turn is quickly oxidized to
glycerate in a subsequent two-electron transfer step. Glycerate
is further oxidized to tartronate (path a2) and, by cleavage of
CÀC bond, into glycolate and formate (path a2 and b1). The
latter species leads to carbonate and glycolate is oxidized to
oxalate (path a4).[23,24,32–34] G is also oxidized directly to tartro-
nate (path b) by chelating adsorption on catalytic surfaces.
However, the oxidation of G yields significant amounts of car-
bonate. Since oxalate is very slowly oxidized on Pd-based elec-
trodes,[23,24,35–37] CO2 is prevalently a by-product of the oxida-
tion of either glycerate or glycolate (path a1 and a3). In an at-
tempt to understand whether
a fresh 2m KOH solution containing 5 wt% EG or G. This pro-
cedure was repeated four times with no apparent performance
decay.
The oxidation of G is more complex than that of EG, in
accord with the presence of three hydroxyl groups. Scheme 3
illustrates the reaction mechanism proposed for the G oxida-
tion on both Pd-(Ni-Zn)/C and Pd/C anode catalysts. In the
pathway of primary alcohol oxidation (path a), G is first oxi-
tartronate is oxidized to formate
and glycolate (path b1) on Pd-
(Ni-Zn)/C and Pd/C, a CV study
has been performed on an alka-
line solution of tartronate.
The corresponding cyclic vol-
tammograms are given in
Figure 10, which also shows
those recorded in a KOH solu-
tion with no substrate. Upon
addition of tartronic acid to the
latter solutions, a significant de-
crease in the peak associated
with PdO reduction in the re-
verse scan at 0.65–0.7 V is ob-
served together with an oxida-
tion wave at approximately
0.76 V attributable to the forma-
tion of PdO; the low specific
current excludes any detectable
oxidation of tartronate. Notably,
no trace of dihydroxyacetone,
hydroxypiruvate, or mesoxalate
was detected at any stage of
Scheme 3. Electrooxidation of G on Pd-(Ni-Zn)/C and Pd/C anode catalysts.
the galvanostatic experiments.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 518 – 528 524

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
This evidence allows us to rule out the secondary alcohols oxi-
dation (path c).
The oxidation of the secondary alcohols of G was reported
by Koper et al.[33] on a Pt electrode, but it was not observed on
an Au electrode. However, Li et al.[11,38,39] have demonstrated
that this can occur both in DGFCs containing nanostructured
anodic catalysts such as Pt/C (7% mesoxalic acid at 0.3 V, 2.0m
KOH+0.1m glycerol) and Au/C (17% glyceric acid, 39% tar-
tronic acid, 19% mesoxalic acid, 19% and oxalic acid at 0.3 V)
in which both the product selectivity and Faradaic efficiency
depends on the working voltages of the fuel cell. Moreover, ac-
cording to in situ infrared data obtained by Coutanceau
et al.,[4] dihydroxyacetone is detected on all monometallic
carbon supported catalysts (Pt/C, Pd/C, and Au/C), although
hydroxypyruvate was only detected on a Au catalyst. It was
proposed that an isomerization reaction of adsorbed dihydrox-
yacetone into adsorbed glyceraldehyde could occur with
higher kinetics on Pt and Pd surfaces than on Au. Using in situ
infrared spectroscopy the same absorption bands (i.e., the
same oxidation species) on Pd/C as on Pt/C were observed,
and it was concluded that the same mechanism was involved
on both catalysts. In both cases a band located at 1335 cmÀ1
and assigned to the presence of a dihydroxyacetone species
on the electrode seems to indicate that the carbon that bears
the secondary alcohol group can also be oxidized at the Pt
and Pd surface. Coutanceau reported that the oxidation reac-
tion of glycerol on a Pd0.5Ni0.5/C (material not in an alloy struc-
ture), is similar to that proposed on a Pd/C catalyst, but occur-
ring at faster kinetics. In contrast, several researchers have ob-
served that the presence of Ni as a co-catalyst enhances the
electrooxidation of alcohols in acidic or alkaline media.[25,34,40–44]
In all of our experiments no trace of the secondary alcohol
oxidation of G, such as dihydroxyacetone, or products of its
further oxidation, such as hydroxypyruvate or mesoxalate,
were observed. This result may seem unusual, but is consistent
with the fact that catalysts based on the same metal may have
different particle sizes, as well as shapes or crystallinity, which
could lead to different catalytic behavior owing to changes in
activity as well as in selectivity.
Figure 11. Potentiodynamic (scan rate 5 mVsÀ1) and power density curves
*
*
for active DEGFCs containing the Pd-(Ni-Zn)/C ( ) and Pd/C ( ) anode elec-
trocatalysts fuelled with a 2m KOH and 5 wt% EG at 808C.
taining the Pd/C electrocatalyst as anode (65 mWcmÀ2 at
0.19 V).
The performance obtained with DGFCs using Pd-(Ni-Zn)/C
and Pd/C anodes and fuelled with a 2m KOH and 5 wt% G at
808C is reported in Figure 12. As previously observed for
DEGFCs, the cell containing the Pd-(Ni-Zn)/C anode exhibits
a higher peak power density (119 mWcmÀ2 at 0.25 V) than the
cell with the Pd/C electrocatalyst as anode (77 mWcmÀ2 at
0.17 V). Notably, at 808C the power output of the cell with Pd-
(Ni-Zn)/C increased nearly seven times, from 16.5 mWcmÀ2 at
208C to 119 mWcmÀ2 at 808C. Accordingly, Pd-(Ni-Zn)/C can
be classified as the best-performing Pd-based electrocatalyst
and amongst the best-performing electrocatalysts for EG and
G oxidation.
The higher power density value provided by Pd-(Ni-Zn) indi-
cates kinetics faster than that on Pd/C. Because of the non-al-
loyed structure of this electrocatalyst[25] only a bifunctional
mechanism between the Ni–Zn surface providing OHÀ and ab-
sorbed organic species from EG and G onto Pd nanoparticles
Active DEGFCs and DGFCs
The MEAs used for the active monoplanar DEGFCs and DGFCs
were realized (as described in the Experimental Section) with
the anode reaction catalyzed by Pd/(Ni-Zn)/C or Pd/C
(1 mg(Pd) cmÀ2), a Fe-Co/C cathode, and a Tokuyama A-201
membrane. The fuel (water solutions containing 2m KOH and
5 wt% EG or G) was delivered to the anode (4 mLminÀ1) and
the oxygen flow was regulated (200 mLminÀ1). Each cell tem-
perature was regulated at 25, 40, 60, or 808C. For simplifica-
tion, only the measurements performed at 808C are reported.
Figure 11 compares the potentiodynamic and power density
curves at 808C obtained with DEGFCs realized with the anodes
catalyzed by Pd-(Ni-Zn)/C and Pd/C. The cell containing the
Pd-(Ni-Zn)/C electrocatalyst exhibits a higher OCV (0.9 V) and
peak power density (95 mWcmÀ2 at 0.23 V) than the cell con-
Figure 12. Potentiodynamic (scan rate 5 mVsÀ1) and power density curves
*
*
for active DGFCs containing the Pd-(Ni-Zn)/C ( ) and Pd/C ( ) anode elec-
trocatalysts fuelled with a solution of 2m KOH and 5 wt% G at 808C.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 518 – 528 525

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Cabot Corp., USA. The Fe-Co/C cathode electrocatalyst [C=Ketjen
Black; inductively coupled plasma (ICP) analysis: 1.13 wt% Fe and
1.71 wt% Co] was prepared by a method reported elsewhere.[26]
The alkaline solid electrolyte used in both passive and active
DAFCs was a anion-exchange membrane A-201 (OHÀ type) ob-
tained from Tokuyama Corporation. All metal salts and reagents
were purchased from Aldrich and used as received. All the solu-
tions were freshly prepared with doubly distilled deionized water.
The quantitative analysis of the FC cell exhausts was obtained by
13C{1H}-NMR spectroscopy (Bruker Avance DRX-400) with chemical
shifts relative to external tetramethylsilane. The calibration curves
for the quantitative analysis were obtained using authentic sam-
ples of the various products in the appropriate range of concentra-
tions and using 1,4-dioxane as internal standard. IC was also used
to identify the oxidation products (Metrohm 761 Compact
equipped with a Metrosep Organic Acids column). The sonication
can be proposed to explain the catalytic improvement. A key
role should be played by the capacity of Ni and NiO to gener-
ate surface NiÀOH moieties at low potentials. Increasing the
amount of OHads on the catalyst surface would favor the forma-
tion of carboxylate species by coupling with Pd–acylads species,
as shown in Scheme 1. In turn, the greater oxophilicity of Ni
than Pd and the higher binding affinity of Ni towards the car-
boxylates ion may facilitate product desorption from the active
Pd sites, thus accounting for the high catalytic activity of
Pd-(Ni-Zn)/C.[25,42,45]
Conclusions
In this paper, examples of passive and active direct ethylene
glycol fuel cells and direct glycerol fuel cells containing Pd-(Ni-
Zn)/C as anode electrocatalysts (Pd nanoparticles supported
on a Ni–Zn phase) are reported. A comparison with an anode
electrocatalyst containing Pd nanoparticles (Pd/C) has also
been presented. The oxidation of ethylene glycol (EG) and
glycol (G) has primarily been investigated in half cells. The re-
sults obtained have highlighted the excellent electrocatalytic
activity of Pd-(Ni-Zn)/C in terms of peak current densities (as
process was performed with
instrument.
a Bandelin Sonopuls probe
Synthesis of electrocatalysts
The electrocatalysts Pd-(Ni-Zn)/C and Pd/C were prepared as previ-
ously described.[25]
À1
À1
high as 3300 Ag(Pd) for EG and 2150 Ag(Pd) for G).
Membrane-electrode assemblies (MEA) containing an anode
Pd-(Ni-Zn)/C, a Fe-Co/C cathode, and a Tokuyama A-201 anion-
exchange membrane have provided excellent results in mono-
planar fuel cells. The MEA performance has been evaluated in
passive and active cells fed with aqueous solutions of 5 wt%
EG or G. Based on the peak power densities obtained in the
temperature range from 20 to 808C and at Pd loadings at the
anode as low as 1 mgcmÀ2, to the authors’ best knowledge
Pd-(Ni-Zn)/C can be classified amongst the best performing
electrocatalysts for EG and G oxidation.
IC and 13C{1H} NMR spectroscopy have been employed to
analyze the anode exhausts of galvanostatic experiments. Pd/C
is a less active but more selective electrocatalyst than Pd-(Ni-
Zn)/C towards the EG oxidation, yielding 89.5% of glycolate.
A mixture of glycolate (>50%), oxalate (37.6%), and carbonate
(7%) was obtained with Pd-(Ni-Zn)/C. A much wider product
distribution, including glycolate, glycerate, tartronate, oxalate,
formate, and carbonate, was detected from the oxidation of
glycerol with both Pd/C and Pd-(Ni-Zn)/C. No trace of the sec-
ondary alcohol oxidation products of G such as dihydroxyace-
tone, hydroxypiruvate, or mesoxalate was observed.
Pd/C: Vulcan XC-72 (5.94 g) was sonicated (20 min) in a three-
necked round-bottomed flask (500 mL) containing ethylene glycol
(250 mL). An aqueous solution (50 mL) obtained by treatment of
PdCl2 (0.6 g, 3.38 mmol) with 6 mL of HCl (37% w/w) was added
dropwise to the resulting dispersion with stirring. After addition of
the Pd salt, a solution of NaOH (5.1 g) in water (10 mL) was intro-
duced into the reactor, which was then heated (to 1408C) under
a nitrogen atmosphere (3 h). Afterwards, the reaction mixture was
allowed to cool to room temperature and the solid product was fil-
tered off under nitrogen atmosphere and washed with distilled
water to neutral pH. The final product was dried (408C, under
vacuum) to constant weight. The yield was 5.8 g. Inductively cou-
pled plasma atomic emission spectroscopy (ICP-AES) analysis re-
vealed a Pd content of 5.2 wt%.
Pd-(Ni-Zn)/C: A slurry of Zn powder (2.8 g, 43 mmol) and Vulcan
XC-72 (6.0 g) in H2O (40 mL) in a three-necked round-bottomed
flask (100 mL) was sonicated (20 min). The slurry was heated
(908C), and then a solution of NiCl2·6 H2O (5.06 g, 21.28 mmol) in
water (10 mL) was added under vigorous stirring (within 1–2 min).
Foaming was observed owing to gas evolution that gradually dis-
appeared (3–4 min). The resulting reaction mixture was stirred at
(RT, 30 min), then the solid product was filtered off and washed
portion-wise with hot water (508C; 200 mL). This solid was sus-
pended in an aqueous solution of NaOH (13 wt%; 300 mL), and
the suspension was stirred until no H2 gas evolution was evident
(40 min). The final product was separated from the reaction mix-
ture by filtration and washed with distilled H2O to neutral pH.
A small portion of this product was dried with a stream of nitrogen
and analyzed by ICP-AES (wt%; Ni 14.4, Zn 3.4). The solid product
was suspended in water (400 mL), and the resulting slurry was so-
nicated (30 min). A solution of K2PdCl4 (1 g, 3.01 mmol) in water
(250 mL) was slowly added (ca. 3 h) to this suspension under vigo-
rous stirring, and the reaction mixture was gently stirred overnight
at room temperature. The final product was filtered off, washed
with water (4ꢁ100 mL), and dried (408C under vacuum) to con-
stant weight. Yield: 6.0 g. ICP-AES (wt%): Pd 6.4, Ni 9.7, Zn 2.6.
The remarkable electrocatalytic activity of Pd-(Ni-Zn) can be
associated both with the high dispersion of the metal particles
and to the intrinsic properties of the Ni–Zn phase that should
increase the amount of OHads groups on the catalyst surface,
which are required for the formation of carboxylic acids.
Experimental Section
Materials and general instrumentation.
All manipulations, except as stated otherwise, were routinely per-
formed under a nitrogen atmosphere using standard airless tech-
nique. Carbon black (Vulcan XC-72 pellets) was purchased from
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 518 – 528 526

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Half-cell testing
Acknowledgements
A portion (about 45 mg) of each electrocatalyst was introduced
into a high-density polyethylene container (5 mL) together with
water (1.2 g), ethanol (99.8%, 0.70 g, Fluka) and a Tokuyama OHÀ-
type anion-exchange ionomer in alcohols solution (0.19 g). The re-
sulting suspension was sonicated (90 min) using a FALC ultrasonic
bath. Each suspension was freshly prepared just before carrying
out the experiment scheduled. The metal loading on each elec-
trode was determined by weighting the amount of the deposited
ink onto the glassy carbon electrode. Cyclic voltammetry (CV) ex-
periments were conducted on a Parstat 2273A potentiostat/galva-
nostat (Princeton Applied Research), using a three-electrode ar-
rangement with a saturated Ag/AgCl/KClsat reference electrode and
a platinum wire as counter electrode. The potential scale of the CV
curves was expressed according to the RHE.
The authors acknowledge the financial support from: Ing. Guido
Gay (Switzerland) project: ‘‘Conversion of CO2 in hydrocarbons
and oxygenated compounds’’; the MATTM (Italy) for the PIRODE
project no 94; the MSE for the PRIT project Industria 2015; the
MIUR (Italy) for the FIRB 2010 project RBFR10J4H7_002.
Keywords: electrochemistry
oxidation · palladium
· fuel cells · nanoparticles ·
[1] F. Alcaide, P. L. Cabot, E. Brillas, J. Power Sources 2006, 153, 47–60.
[2] N. Ji, T. Zhang, M. Y. Zheng, A. Q. Wang, H. Wang, X. D. Wang, J. G. G.
Chen, Angew. Chem. 2008, 120, 8638–8641; Angew. Chem. Int. Ed. 2008,
47, 8510–8513.
[3] M. Y. Zheng, A. Q. Wang, N. Ji, J. F. Pang, X. D. Wang, T. Zhang, Chem-
SusChem 2010, 3, 63–66.
[4] M. Simꢂes, S. Baranton, C. Coutanceau, Appl. Catal. B 2010, 93, 354–
362.
Fuel-Cell Testing
[5] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Green Chem. 2008,
10, 13–30.
[6] S. Carrettin, P. McMorn, P. Johnston, K. Griffin, C. J. Kiely, G. J. Hutchings,
Phys. Chem. Chem. Phys. 2003, 5, 1329–1336.
[7] L. Prati, P. Spontoni, A. Gaiassi, Top. Catal. 2009, 52, 288–296.
[8] D. Liang, J. Gao, J. H. Wang, P. Chen, Z. Y. Hou, X. M. Zheng, Catal.
Commun. 2009, 10, 1586–1590.
[9] A. Villa, G. M. Veith, L. Prati, Angew. Chem. 2010, 122, 4601–4604;
Angew. Chem. Int. Ed. 2010, 49, 4499–4502.
[10] S. Carrettin, P. McMorn, P. Johnston, K. Griffin, G. J. Hutchings, Chem.
Commun. 2002, 696–697.
[11] L. Xin, Z. Zhang, J. Qi, D. Chadderdon, W. Li, Appl. Catal. B 2012, 125,
85–94.
[12] Z. Zhang, L. Xin, W. Li, Appl. Catal. B 2012, 119–120, 40–48.
[13] A. Ilie, M. Simꢂes, S. Baranton, C. Coutanceau, S. Martemianov, J. Power
Sources 2011, 196, 4965–4971.
[14] A. Falase, M. Main, K. Garcia, A. Serov, C. Lau, P. Atanossov, Electrochim.
Acta 2012, 66, 295–301.
Passive breathing fuel cells. The MEAs (membrane electrode as-
sembly) for the passive monoplanar cells were realized using
a commercial Tokuyama A-201 anion-exchange membrane and
cathodes containing a proprietary Fe-Co/C electrocatalyst.[26] The
cell hardware was described previously.[24] The membrane was con-
ditioned in a saturated KOH solution (1 h) before assembling the
MEA. The anode was realized with a nickel foam plate (5.13 cm2)
onto which the appropriate amount of dense catalytic ink was de-
posited. This was prepared by introducing the catalyst (about
100 mg) into a polyethylene container (5 mL) together with water
(about 80 mg). The metal loading on each electrode
(ꢀ 1 mgcmÀ2), was determined by weighing the amount of ink de-
posited. The MEA was obtained by mechanically pressing the
anode, cathode, and membrane at approximately 40 bar. For a relia-
ble evaluation of the oxidation products, the anode compartment
was sealed under a nitrogen atmosphere while the cathode was
exposed to either air or oxygen with no significant variation in the
cell performance. This was evaluated by means of an ARBIN BT-
2000 5A 4 channels instrument. The fuel oxidation products were
qualitatively and quantitatively determined by IC and 13C{1H}-NMR
spectroscopy using appropriate calibration curves.
[15] Y. Kwon, Y. Birdja, I. Spanos, P. Rodriguez, M. T. Koper, ACS Catal. 2012,
2, 759–764.
[16] P. K. Shen, C. Xu, Electrochem. Commun. 2006, 8, 184–188.
[17] Z. Wang, F. Hu, P. K. Shen, Electrochem. Commun. 2006, 8, 1764–1768.
[18] C. W. Xu, P. K. Shen, Y. J. Liu, J. Power Sources 2007, 164, 527–553.
[19] F. P. Hu, G. F. Cui, Z. D. Wei, P. K. Shen, Electrochem. Commun. 2008, 10,
1303–1306.
Active breathing fuel cells. The active fuel cell system, purchased
from Scribner-Associates (USA) with a 25 cm2 fuel cell fixture and
an effective electrode area of 5 cm2, was modified in our laboratory
with titanium end plates to tolerate the alkaline conditions. The
MEA was fabricated by mechanically pressing at approximately
100 bar the anode (nickel foam), cathode (carbon cloth) and To-
kuyama A-201 alkaline exchange membrane. A dense anode ink
was prepared by mixing the powdered catalyst with an aqueous
dispersion of polytetrafluoroethylene (PTFE) to produce a binder
loading of 5 wt%. As a general procedure, an identical amount of
the resulting paste was spread onto two identical Ni-foam plates.
One of these was used almost immediately to fabricate the MEA,
the other was dried until constant weight for the quantitative de-
termination of the Pd loading (ꢀ1 mgcmÀ2 in all cases). The fuel
(water solution containing 5 wt% EG or G in 2m KOH) was deliv-
ered to the anode ( 4 mLminÀ1) and regulated at 25, 40, 60, and
808C, respectively. The oxygen flow (200 mLminÀ1)was regulated
at 25, 30, 40, and 608C, respectively (RH 100%). The temperature
of the cell under working conditions was controlled at 25, 40, 60,
and 808C, respectively, by a thermocouple positioned inside the
end plate at the cathode side.
[20] C. Coutanceau, L. Demarconnay, C. Lamy, J. M. Lꢃger, J. Power Sources
2006, 156, 14–19.
[21] L. Demarconnay, S. Brimaud, C. Coutanceau, J. M. Lꢃger, J. Electroanal.
Chem. 2007, 601, 169–180.
[22] M. Simꢂes, S. Baranton, C. Coutanceau, Appl. Catal. B 2011, 110, 40–49.
[23] V. Bambagioni, M. Bevilacqua, C. Bianchini, J. Filippi, A. Marchionni, F.
Vizza, L. Q. Wang, P. K. Shen, Fuel Cells 2010, 10, 582–590.
[24] V. Bambagioni, C. Bianchini, A. Marchionni, J. Filippi, F. Vizza, J. Teddy, P.
Serp, M. Zhiani, J. Power Sources 2009, 190, 241–251.
[25] V. Bambagioni, C. Bianchini, J. Filippi, W. Oberhauser, A. Marchionni, F.
Vizza, R. Psaro, L. Sordelli, M. L. Foresti, M. Innocenti, ChemSusChem
2009, 2, 99–112.
[26] V. Bambagioni, C. Bianchini, J. Filippi, A. Lavacchi, W. Oberhauser, A.
Marchionni, S. Moneti, F. Vizza, R. Psaro, V. Dal Santo, A. Gallo, S. Re-
cchia, L. Sordelli, J. Power Sources 2011, 196, 2519–2529.
[27] E. Santos, M. C. Giordano, Electrochim. Acta 1985, 30, 871–878.
[28] Z. X. Liang, T. S. Zhao, J. B. Xu, L. D. Zhu, Electrochim. Acta 2009, 54,
2203–2208.
[29] N. Dalbay, F. Kadirgan, J. Electroanal. Chem. 1990, 296, 559–569.
[30] C. Xu, Z. Tian, P. K. Shen, S. P. Jiang, Electrochim. Acta 2008, 53, 2610–
2618.
[31] L. Roquet, E. M. Belgsir, J.-M. Lꢃger, C. Lamy, Electrochim. Acta 1994, 39,
2387–2394.
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 518 – 528 527

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
[32] J.-L. Lin, J. Ren, N. Tian, Z.-Y. Zhan, S.-G. Sun, J. Electroanal. Chem. DOI:
10.1016/j.jelechem.2012.08.027.
[41] R. N. Singh, A. Singh, Anindita, Int. J. Hydrogen Energy 2009, 34, 2052–
2057.
[33] Y. Kwon, K. J. P. Schouten, M. T. M. Koper, ChemCatChem 2011, 3, 1176–
1185.
[34] C. Bianchini, P. K. Shen, Chem. Rev. 2009, 109, 4183–4206.
[35] H. T. Zheng, Y. Li, S. Chen, P. K. Shen, J. Power Sources 2006, 163, 371–
375.
[36] P. Bert, C. Bianchini, G. Giambastiani, A. Marchionni, A. Tampucci, F.
Vizza, PCT/EP 2008/055706.
[37] a) F. Hahn, B. Beden, F. Kadirgan, C. J. Lamy, J. Electroanal. Chem. 1987,
216, 169–180; b) P. A. Christensen, A. Hamnett, J. Electroanal. Chem.
1989, 260, 347–359.
[42] C. Bianchini, V. Bambagioni, J. Filippi, A. Marchionni, F. Vizza, P. Bert, A.
Tampucci, Electrochem. Commun. 2009, 11, 1077–1080.
[43] M. R. Tarasevich, Z. R. Karichev, V. A. Bogdanovskaya, E. N. Lubnin, A. V.
Kapustin, Electrochem. Commun. 2005, 7, 141–146.
[44] S. Sen Gupta, J. Datta, J. Power Sources 2005, 145, 124–132.
[45] V. Bambagioni, C. Bianchini, Y. Chen, J. Filippi, P. Fornasiero, M. Innocen-
ti, A. Lavacchi, A. Marchionni, W. Oberhauser, F. Vizza, ChemSusChem
2012, 5, 1266–1273.
[38] L. Xin, Z. Zhang, Z. Wang, W. Li, ChemCatChem 2012, 4, 1105–1114.
[39] Z. Zhang, L. Xin, W. Li, Int. J. Hydrogen Energy 2012, 37, 9393–9401.
[40] K. S. Kumar, P. Haridoss, S. K. Seshadri, Surf. Coat. Technol. 2008, 202,
1764–1770.
Received: November 12, 2012
Published online on February 12, 2013
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2013, 6, 518 – 528 528