Energy Efficiency of Alkaline Direct Ethanol Fuel Cells Employing Nanostructured Palladium Electrocatalysts

Article abstract of DOI:10.1002/cctc.201500189

Carbon supported nanostructured palladium or palladium alloys are considered the best performing anode electrocatalysts currently employed in alkaline electrolyte membrane direct ethanol fuel cells (AEM-DEFCs). High initial peak power densities are generally obtained as Pd preferentially favors the selective oxidation of ethanol forming acetate thus avoiding strongly poisoning intermediates such as CO. However, few studies exist that investigate DEFC performance in terms of both energy efficiency and discharge energy density, as well as power density depending on the concentration of fuel. In this paper we have determined such parameters for room temperature air breathing AEM-DEFCs equipped with Pd based anodes, anion exchange membranes and FeCo/C cathode electrocatalysts. Combined with the optimization of the fuel composition a maximum energy efficiency of ≈7 % was obtained for this AEM-DEFC. Such a performance suggests that devices of this type are suitable for supplying low power applications such as small portable electronic devices.

Full text of DOI:10.1002/cctc.201500189

DOI: 10.1002/cctc.201500189
Full Papers
Energy Efficiency of Alkaline Direct Ethanol Fuel Cells
Employing Nanostructured Palladium Electrocatalysts
[
a, b]
[a]
[a]
[a]
Lianqin Wang,
Paolo Fornasiero,
Alessandro Lavacchi,* Manuela Bevilacqua, Marco Bellini,
[
a, b]
[a]
[a, c]
[a]
Jonathan Filippi, Massimo Innocenti,
Andrea Marchionni,
[a]
[a]
Hamish Andrew Miller, and Francesco Vizza*
Carbon supported nanostructured palladium or palladium
alloys are considered the best performing anode electrocata-
lysts currently employed in alkaline electrolyte membrane
direct ethanol fuel cells (AEM-DEFCs). High initial peak power
densities are generally obtained as Pd preferentially favors the
selective oxidation of ethanol forming acetate thus avoiding
strongly poisoning intermediates such as CO. However, few
studies exist that investigate DEFC performance in terms of
both energy efficiency and discharge energy density, as well as
power density depending on the concentration of fuel. In this
paper we have determined such parameters for room temper-
ature air breathing AEM-DEFCs equipped with Pd based
anodes, anion exchange membranes and FeCo/C cathode elec-
trocatalysts. Combined with the optimization of the fuel com-
position a maximum energy efficiency of ꢀ7% was obtained
for this AEM-DEFC. Such a performance suggests that devices
of this type are suitable for supplying low power applications
such as small portable electronic devices.
Introduction
As fuels, alcohols show a variety of advantages with respect to
hydrogen. First they are much easier to handle, store, and
transport due to the fact that they are liquid at room tempera-
ture. Furthermore, they have a high energy density, for exam-
dation. Consequently, high loadings of PtRu/C alloy electro-
catalysts are required to achieve sufficient performance. Addi-
tionally, concerns about the overall energy efficiency of DMFCs
[
4,5]
have also been recently raised.
À1
À1
ple, ethanol 24 MJL and methanol 15.6 MJL . In virtue of
these advantages, alcohols have been extensively tested in
fuel cells. Up till now direct methanol fuel cells (DMFCs) have
The substitution of methanol with ethanol in fuel cells is ad-
vantageous, as ethanol is far less toxic and volatile and it can
be obtained from biomass conversion through low energy pro-
cesses. Bio-ethanol energy convenience has been recently sup-
ported by a variety of papers on life cycle analysis (LCA) that
have shown that energy return of the invested energy higher
than 8 can be achieved when sugarcane or cellulose are used
[1–3]
been the subject of the most development.
DMFCs have
been technologically exploited for powering small portable
electronic devices, however the toxicity of methanol, together
with its flammability and high vapor pressure at room temper-
ature severely hampered full technological exploitation. More-
over, methanol oxidation occurs via pathways involving the
formation of adsorbed CO, which causes electrocatalyst poi-
soning and ultimately results in significant performance degra-
[
6]
as feedstock. Furthermore, the bio-ethanol industry is contin-
uously expanding. Projections suggest that the world global
production could surpass 100 million tonnes per year by
[
7]
2018.
Most of the reported investigations on DEFCs have been
[
a] L. Q. Wang, A. Lavacchi, M. Bevilacqua, M. Bellini, P. Fornasiero, J. Filippi,
M. Innocenti, A. Marchionni, H. A. Miller, F. Vizza
ICCOM-CNR
concerned with the assessment of power density per-
[8–10]
formance,
the optimization of which depends largely on
the nature of the anode electrocatalyst. The focus on the
anode catalyst is indeed essential, as the activation potential
for ethanol oxidation is significant with respect to the overall
potential drop. Besides maximum power density and stability,
energy efficiency and delivered energy per fuel batch are criti-
cal. This is for two reasons: 1) highly efficient fuel cells mini-
mize volume and weight through reduction of the size of the
via Madonna del Piano 10, 50019, Sesto Fiorentino (FI) (Italy)
E-mail: alessandro.lavacchi@iccom.cnr.it
francesco.vizza@iccom.cnr.it
[
b] L. Q. Wang, P. Fornasiero
Department of Chemical and Pharmaceutical Sciences
University of Trieste
Via L. Giorgieri 1, 34127 Trieste (Italy)
[
c] M. Innocenti
Dipartimento di Chimica
Università di Firenze
Via della Lastruccia 3, 50019 Sesto Fiorentino (FI) (Italy)
[
11]
storage systems and 2) high fuel cell efficiency improves the
energy efficiency of whole systems, being the most important
factor in sustainability assessment.
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201500189.
In the present investigation we determine the energy per-
formance of alkaline direct ethanol fuel cells using palladium
based anode electrocatalysts. Palladium is very effective for
This publication is part of a Special Issue on Palladium Catalysis. To view
the complete issue, visit:
http://onlinelibrary.wiley.com/doi/10.1002/cctc.v7.14/issuetoc
[
12–14]
ethanol oxidation under alkaline conditions.
A recent
ChemCatChem 2015, 7, 2214 – 2221
2214
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
direct comparison of ethanol electrooxidation performance of
Pt/C and Pd/C in alkaline media demonstrated that Pd is supe-
rior for two main reasons; 1) the higher oxophilic nature of Pd
that promotes the adsorption of OH groups and 2) the relative-
Palladium based electrocatalysts have been proved to be
more effective for ethanol oxidation if used in conjunction
with promoting oxides such (e.g. CeO , SnO , Mn O , NiO, and
2
2
3
4
[
24–27]
TiO .
These oxides significantly lower the onset potential
2)
[15–17]
[14]
ly inert nature of Pd to CÀC bond cleavage
Pd is, conse-
for the oxidation reaction
with ceria being amongst the
[
25]
quently, more stable and less susceptible to poisoning in the
most utilized. The reason for such an enhancement can be
easily understood considering the first step of ethanol oxida-
tion on Pd in alkaline condition that consists in the formation
of the adsorbed hydroxide at the palladium surface [Eq. (1)].
Equation (1) is also the rate determining step of ethanol oxi-
[
18]
direct oxidation of ethanol to acetate (Scheme 1).
[
25]
dation in alkaline environment.
The addition of an oxide
such as ceria to Pd promotes the formation of the Pd surface
hydroxide thanks to a spillover of the primary oxide mecha-
[
28]
nism. In a previous investigation we have shown that thanks
to this mechanism ceria is able to double the energy efficiency
[
25]
of AEM- DEFCs.
Here we provide a detailed analysis of the energy per-
formance of alkaline DEFCs. The analysis has been performed
on AEM-DEFCs equipped with nanostructured palladium based
anode catalysts, Pd/C and Pd on a mixed CeO -C support (both
2
Scheme 1. Proposed mechanism for the selective conversion of ethanol into
acetate on Pd-based electrocatalysts in alkaline media.
with ꢀ5 wt% Pd). At the cathode a proprietary FeÀCo/C has
been employed. The analysis has been focused on the deter-
mination of the following three parameters:
Indeed, the number and diversity of Pd based compounds
that have been developed for alcohol electrooxidation under
alkaline conditions is large. DEFCs equipped with Pd anodes
1) maximum power density
2) delivered fuel energy density per single fuel load
3) energy efficiency
À2 [19–22]
can produce peak power densities of over 100 mWcm .
At a first look, incomplete oxidation of ethanol is an evident
drawback as only 4 of the possible 12 electrons are obtained.
Indeed, complete oxidation would certainly be beneficial as it
would allow, in principle, a better exploitation of the energy
content of ethanol. Nevertheless, the kinetics of complete oxi-
dation are very sluggish, ultimately resulting in high anodic
over potentials and the need for high loadings of expensive
electrocatalysts for example, PtSn alloys. Partial oxidation of
ethanol to acetate is much faster and occurs at high specific
We believe that such an analysis deserves special attention.
In the course of this paper we provide the first detailed investi-
gation of the energy performance of platinum free direct alco-
hol fuel cells.
Experimental Section
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
Cabot Corp., USA. The Fe-Co/C cathode supported on carbon cloth
[
14]
current densities even at reasonably low over potentials. The
advantage of combined stability, fast kinetics, and low overpo-
tential may well overcome the problems associated with in-
complete oxidation. In addition, the discovery that the acetate
ion can be converted to ethanol in a renewable way by elec-
troreduction, thus closing the fuel production/transformation
cycle, offers a further strong argument in favor of the use of
[22,29]
was prepared by a method reported elsewhere.
The alkaline
solid electrolyte used in passive DAFCs was a anion-exchange
membrane A-201 obtained from Tokuyama Corporation (Japan). All
metal salts and reagents were purchased from Aldrich and used as
received. All the solutions were freshly prepared with doubly dis-
tilled deionized water.
[23]
Pd-based electrocatalysts for AEM-DEFC anodes.
The generally accepted mechanism for the oxidation of eth-
anol on Pd in alkaline media is described in Equations (1)–(5)
Synthesis of Pd-CeO /C anode catalyst: Vulcan XC-72 (4 g) was
2
added to a solution of Ce(NO ) ·6H O (5.31 g) in H O (250 mL). The
3
3
2
2
Pd þ OH Ð PdÀOH þ eÀ
0
À
mixture was kept under stirring for 30 min and sonicated for
0 min. After adjusting the pH to 12 with KOH, the resulting sus-
pension was stirred for 2 h. The product was separated by filtration
ð1Þ
ð2Þ
ads
3
0
Pd þ CH CH OH Ð Pd-ðCH CH OHÞ
3
2
3
2
ads
and washed with H O until neutral pH was obtained. The product
2
was dried at 658C and subsequently heated under air in a tube fur-
nace at 2508C for 2 h. Cooling to room temperature was undertak-
Pd-ðCH CH OHÞ þ 3 OH Ð Pd-ðCOCH Þ þ 3 H O þ 3 eÀ
À
3
2
ads
3
ads
2
ð3Þ
ð4Þ
ð5Þ
en under a flow of Ar. The yield of C-CeO was 6.24 g.
2
The C-CeO2 was suspended in water (500 mL), stirred vigorously
for 30 min and sonicated for 10 min. To this mixture, a solution of
K2PdCl4 (0.93 g, 2.84 mmol) in water (250 mL) was slowly added
(ꢀ3 h) under vigorous stirring, followed by an aqueous solution of
0
Pd-ðCOCH3Þads þ PdÀOHads Ð Pd-ðCH COOHÞ þ Pd
3
ads
À
0
À
Pd-ðCH COOHÞads þ OH Ð Pd þ CH COO þ H O
3
3
2
ChemCatChem 2015, 7, 2214 – 2221
www.chemcatchem.org
2215
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
2
1
m KOH (14.5 mL). Next, ethanol (100 mL) was added to the result-
a 5.13 cm nickel foam plate to obtain a metal loading of
À2
ing mixture, which was then heated at 808C for 30 min. The de-
1 mgcm . The same Pd loading was used for both the Pd/C and
sired product Pd-CeO /C was filtered off, washed several times
the Pd-CeO /C electro-catalysts. The MEA was obtained by me-
2
2
with distilled water to neutrality and finally dried under vacuum at
chanically pressing anode, membrane and cathode. Silicone-rubber
gaskets were employed to seal the system. Power density and gal-
vanostatic curves were determined on a ARBIN BT-2000 5A-4 chan-
nels potentiostat/galvanostat. All the experiments had been per-
formed at room temperature.
4
08C until constant weight (yield: 6.46 g). Pd content=4.6 wt%
(
ICP-AES).
Synthesis of Pd/C anode catalyst: Vulcan XC-72 (4.5 g) was sus-
pended in ethylene glycol (250 mL) and sonicated for 20 min in
a 500 mL three-neck round-bottomed flask. Then an aqueous solu-
Results and Discussion
tion of H O (50 mL) and ethylene glycol (50 mL) and HCl (37%,
2
6
mL) with dissolved PdCl (445 mg) were added drop by drop to
Anode electrocatalyst characterization
2
the resulting solution under stirring in a N stream. After adequate
2
All the tested catalysts have been analyzed for composition,
structure and morphology. ICP-EAS derived composition data
are reported in Table 1.
stirring, an alkaline solution of NaOH (5 g) in H O (10 mL) and eth-
2
ylene glycol (35 mL) was introduced in the reactor which then was
heated at 1258C for 3 h again under a N atmosphere. Then the
2
mixture was cool down to room temperature. The solid product
Information about the crystalline structure of the investigat-
ed anode electrocatalysts has been obtained by X-ray powder
diffraction (XRPD) (Figure 1). The carbon support shows the
was filtered off and washed with H O to neutral pH. The final prod-
2
uct was dried in vacuum oven at 408C. An amount of 4.47 g of cat-
alyst was obtained. Pd content=5.2 wt% (ICP-AES)
Transmission electron microscopy (TEM) was performed on a Philips
CM12 microscope at an accelerating voltage of 100 kV. The micro-
scope was equipped with an EDAX energy dispersive microanalysis
system.
Table 1. Chemical composition of the investigated catalysts.
Element
Pd/C [wt%]
2
Pd-CeO /C [wt%]
C
Pd
CeO
94.8
5.2
–
62.0
4.6
33.4
X-ray powder diffraction (XRPD) data were acquired at room tem-
perature with a PANalytical X’PERT PRO diffractometer, employing
CuKa radiation (l=1.54187) and a PW3088/60-graded multilayer
parabolic X-ray mirror for Cu radiation. The spectra were acquired
in the 2q range from 5.0 to 120.08, using a continuous scan mode
with an acquisition step size of 2q=0.02638 and a counting time
of 49.5 s.
2
Electrochemical measurements with iR drop correction were per-
formed with PARSTAT 2273 potentiostat/galvanostat (Princeton Ap-
plied Research) in a three electrode cell. The working electrode,
2
Glassy Carbon (Sigradur G; 0.867 cm ), covered by the catalyst, was
loaded at the top end of the cylinder. The ink of catalyst was con-
sisted of Pd/C (22 mg) or Pd-CeO /C (22 mg) and 2-propanol (2 g).
2
The ink was sonicated for 120 min with a Branson 3200 bath. The
metal loading on each electrode was determined by weighting the
amount of ink deposited on the glassy carbon disk. An Ag/AgCl
electrode was used as reference, while counter electrode was con-
sisted of a platinum wire (0.5 mm diameter). All the solutions were
purged with N2 for 30 min before any electrochemical experi-
ments. The activity of ethanol oxidation was investigated in 2m
ethanol+2m KOH solution by cyclic voltammetry in the range of
À1
2
Figure 1. XRPD diffraction patterns of Pd/C and Pd-CeO /C.
0
–1.20 V. Reported CVs were acquired at a scan rate of 50 mVs
.
All the potentials reported are versus to reversible hydrogen elec-
trode (RHE).
presence of graphite domains. The width of the (002) peak of
graphite has been found to be large suggesting a small size of
the crystalline domain and the coexistence of amorphous do-
mains in both the examined samples. Metallic palladium has
The monoplanar cell tests were performed using an in-house con-
structed device composed of Plexiglas with gold coated steel cur-
[30]
rent collectors as described previously. The anode side fuel tank
was filled with 12 mL of static fuel solution for each experiment.
The cathode side was open to the ambient air. Membrane elec-
trode assemblies (MEAs) were realized using a nickel foam support-
ed anode containing a Pd catalyst, a commercial Tokuyama A-201
anion-exchange membrane and a cathode containing a FeÀCo/C
electro-catalyst supported on carbon cloth. Cathode electrodes
also been identified in both catalysts. In the Pd-CeO /C sample
2
the XRPD pattern shows that cerium is present prevalently in
IV
[25]
form of Ce oxide.
The catalyst morphology has also been investigated by TEM.
Images of the Pd/C catalyst reveal that the majority of the Pd
particles range between 2 and 10 nm, as shown in Figure 2a,b.
were prepared as described previously with a loading of about
À2 [22]
2
mgcat cm . The anode catalytic ink was prepared mixing the
catalyst (100 mg) with water (400 mg) and a quantity of aqueous
PTFE solution (10 wt%). The resulting ink was spread onto
reported in Figure 2c,d are the TEM images of the Pd-CeO
/C
2
catalyst. The contrast in these images is not sufficient to distin-
ChemCatChem 2015, 7, 2214 – 2221
www.chemcatchem.org
2216
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[
31]
tive PdÀO species. The cathodic peak C at 0.20 V has been
2
[
32]
assigned to the adsorption of hydrogen.
The CV of the Pd-CeO /C system in 2m KOH (Figure 3b)
2
shows a much more evident peak due to hydrogen desorption
(
A ). The peak at 0.37 V (A ) moving toward more positive po-
1 2
tentials, can be ascribed to the superimposition of the forma-
tion of palladium oxides and hydroxides with the oxidation of
adsorbed hydrogen. At more positive potentials only the Pd
oxidation is present. In the reverse scan the cathodic peak at
0
.78 V (C ) has been again assigned to the reduction of PdÀO,
1
whereas the broad cathodic region at potentials lower than
0
.55 V (C ) is attributed to the hydrogen uptake by both Pd
2
and ceria.
The CV experiments performed on the Pd/C electrode in 2m
ethanol+2m KOH electrolyte (Figure 3c) show an ethanol oxi-
dation peak (A ) at 0.78 V with an intensity controlled by the
1
Figure 2. Electrocatalyst morphology. Low and high magnification TEM mi-
crographs of a,b) Pd/C and c,d) Pd-CeO /C.
progressive coverage of the electrode surface by PdÀO. The
2
presence of this species has recently been observed by X-ray
[
33]
absorption (XAS and XANES) investigations.
At potentials
guish clearly between CeO and Pd particles, which prohibits
higher than 1 V, the anodic current is comparable to that ob-
served for the 2m KOH alone (Figure 3a), indicating that only
a negligible fraction of the catalyst sites remains active for eth-
anol oxidation. The reverse scan shows a further anodic peak
2
a determination of the Pd particle size distribution. The mor-
[
25]
phology of similar catalysts has been described previously.
A occurring at 0.65 V, attributable to the oxidation of fresh
2
Half-cell characterization
ethanol on the catalyst surface freed from Pd oxide. The CV of
Cyclic voltammetry studies (Figure 3) have been performed for
the Pd-CeO /C electrode (Figure 3d) shows an ethanol oxida-
2
these catalysts Pd/C and Pd-CeO /C, respectively, in both 2m
KOH and 2m ethanol+2m KOH electrolytes. For the Pd/C cata-
lyst (Figure 3a) starting from the low potential region, a broad
tion peak culminating at 0.82 V with a current density of
2
À2
85.5 mAcm , which is higher than that measured for the Pd/C
À2
electrode (59.9 mAcm ), also with an onset potential for etha-
anodic peak (A ) attributable to oxidation of absorbed hydro-
nol oxidation considerably lower than for Pd/C (0.24 V). This
finding is in agreement with the data reported previously for
1
gen is observed. The oxidation of the palladium surface is ac-
companied with a gentle increase of the current density. On
[
25]
analogous systems. At a fixed potential of 0.5 V, current den-
À2
the reverse scan a well-defined cathodic peak C at 0.69 V can
sities of 19.7 and 3.5 mAcm for Pd-CeO /C and Pd/C were
1
2
be seen which is known to be related to the reduction of inac-
measured, respectively. The enhanced performance of the ceria
catalyst has been assigned to
the ability of ceria to promote
the formation of active PdÀOH
ads
[
25]
species.
Direct Ethanol Fuel Cell
performance
Fuel cell performance has been
evaluated employing the fuel
compositions
reported
in
Table 2. Each of the reported
fuel compositions has been
tested in two AEM-DEFCs
equipped respectively with the
Pd/C and the Pd-CeO /C anodes
2
with the same metal loading.
Performance evaluation has
been performed on the basis of
potentiodynamic and galvano-
À2
static (20 mAcm ) curves. From
the potentiodynamic curves we
have extracted the maximum
2
Figure 3. Cyclic voltammograms of Pd/C (&) and Pd-CeO /C (*) in a,b) 2m KOH and c,d) 2m Ethanol+2m KOH
electrolytes.
ChemCatChem 2015, 7, 2214 – 2221
www.chemcatchem.org
2217
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
superior for the system with Pd-CeO /C. This is particularly evi-
2
Table 2. Anode electrolyte composition. KOH and ethanol concentra-
tions.
dent in the low current region of the curve, where the poten-
tial gap between Pd-CeO /C and Pd/C is the most. This fact
2
Anode electrolyte composition
can be interpreted by looking at the CV data. Indeed, voltam-
metry shows a much lower onset potential for the ceria pro-
moted electrodes as compared to the bare Pd/C in agreement
Ethanol
KOH
2
4
4
4
6
m
m
m
m
m
2 m
2 m
4 m
6 m
6 m
[25]
with previous reports.
Shown in Figure 4b is a comparison of the galvanostatic
curves for each catalyst, using 2m KOH+2m ethanol fuel. At
À2
a constant current density of 20 mAcm the Pd-CeO /C cata-
2
lyst exhibits a considerably higher voltage than the cell with
Pd/C. As a consequence the energy-density released by
À3
power density, while from galvanostatic data the energy deliv-
ered in a single fuel batch and the energy efficiency have been
respectively obtained.
a single fuel batch resulted in 201.2 Jcm for the system em-
À3
ploying the Pd-CeO /C electro-catalyst and 112.3 Jcm for the
2
system using Pd/C. In terms of overall energy efficiency we cal-
The delivered energy from each test has been evaluated
using Equation (6), integrating the galvanostatic curve over the
time up till when a negligible cell potential was reached
culate a value of 6.57% for Pd-CeO /C, whereas the Pd/C
2
system reaches only 3.67%.
The influence of fuel composition on the polarization and
galvanostatic tests was investigated using both the Pd/C and
Pd/C-CeO2 catalysts. Firstly, performance was evaluated em-
ploying balanced concentrations of ethanol and KOH, namely
(
ꢀ0 V).
Z
E ¼ VðtÞdt
ð6Þ
2
, 4 and 6m. As expected the Pd/C-CeO anode was superior
2
in terms of peak power density, discharge volume energy den-
sity and energy efficiency for all fuel compositions studied
(Figure 5). Consequently, the discussion of results will be limit-
For which: I=current flowing through the cell, V(t)=time
dependent cell potential
The energy efficiency was determined using Equation (7):
ed to the mixed CeO -Carbon supported Pd catalyst. The peak
2
À2
power density reaches a maximum of 47 mWcm when the
R
tf
to
I
Vdt
1
ethanol and KOH concentration is 4m (Figure 5b). Galvanostat-
e ¼
ð7Þ
0
F
À2
DH
M
ic curves at 20 mAcm follow a different trend (Figure 5c).
complete
TOT
Indeed the best performance, at least in terms of energy deliv-
ered per single fuel batch, has been achieved with the most
concentrated electrolyte (6m ethanol+6m KOH) (Figures 5c–
d). The cell fuelled with 2m ethanol+2m KOH lasts 16.3 h and
releases 1347.6 J, while the same cell fed with 6m ethanol and
6m KOH released 3534.0 J in 40 h. The energy density that can
be extracted from a single fuel load benefits from the use of
a concentrated electrolyte even if high concentrations of etha-
nol and KOH can be detrimental in terms of maximum power
density.
For which e is the energy efficiency for the complete oxida-
tion of ethanol to carbonate, I is the current (0.102 A), V(t) is
the transient potential in the constant current discharge ex-
0
periments, DH
is the enthalpy for ethanol oxidation to
complete
À1
F
carbonate (À1326 kJmol ) and M
is the total number
TOT
moles of fuel in the cell. Notably, Equation (7) is a comprehen-
sive efficiency definition, which account for the contribution of
thermodynamic, kinetics (activation overpotentials and ohmic
drops) and ability to convert the fuel (which account for the
faradaic conversion, and any form of fuel loss that may occur
in the cell and it is always less than 1).
Energy efficiency behaves differently. It drops from 6.57 to
3.41% passing from 2 to 6m fuels. This may be ascribed to
The performance of passive
AEM-DEFCs with 2m ethanol+
2
m KOH fuel, with Pd/C and Pd-
CeO /C anode catalysts is shown
2
in Figure 4. At first glance (Fig-
ure 4a) we may notice that both
catalysts show
a high OCV,
about 0.9 V due to the use of
the Fe-Co/C catalyst on the cath-
ode side which has no activity
[29]
toward ethanol oxidation.
After closing the circuit and
applying a variable load (Fig-
ure 4a) we can see that in the
full range of explored potentials
Figure 4. Comparison on the performance of the passive AEM-DEFCs containing the Pd/C (&) and Pd-CeO
2
/C (*)
anode catalysts and fueled with 2m ethanol+2m KOH. a) Polarization and power density curves and b) galvano-
the current density obtained is static curves with the constant current density of 20 mAcmÀ2
.
ChemCatChem 2015, 7, 2214 – 2221
www.chemcatchem.org
2218
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
À
periment OH is consumed at the same rate. Ulti-
mately alkaline AEM-DEFCs operate under a constant
and progressive reduction of the pH. On the side of
À
kinetics, OH adsorption plays a crucial role in the ox-
idation mechanism so much that at low overpoten-
tials the formation of adsorbed hydroxyl is the rate
determining step. Such aspects have been investigat-
ed in half cell experiments. Nevertheless, there is
À
a lack of information regarding the effect of OH
concentration on device performance. For this reason
we have performed a series of experiments (power
density and galvanostatic) where the ethanol concen-
À
tration has been kept constant (4m) varying OH
concentration between 2 and 6m (Supporting Infor-
mation, Figure S1). Shown in Figure 6 is an overview
of the obtained data. Peak power density (Figure 6a),
discharge energy density (Figure 6b) and energy effi-
ciency (Figure 6c) have been reported as a function
of the fuel composition for both Pd/C and Pd-CeO /C
2
anodes. These three parameters show different
trends indicating that care should be taken to opti-
mize the performance of AEM-DEFCs. Highest power
density can be obtained at suitable concentrations of
ethanol and KOH. For some applications the power
density is the overriding factor, but for portable
power sources, discharge capacity is the most relevant factor.
High concentrations of ethanol and KOH are required in order
to maximize discharge energy.
Figure 5. Effect of ethanol and KOH concentration on the performance of the air-breath-
ing AEM-DEFCs containing the Pd-CeO /C anode electrocatalyst. a) Polarization curves,
b) Power density curves, c) galvanostatic curves, and d) discharge energy curves.
2
two factors: 1) ethanol crossover and 2) incomplete conversion
of the fuel. Crossover accounts for ethanol losses driven by the
osmotic flow through the membrane and the successive evap-
oration from the cathode. Incomplete conversion of the fuel
may indeed be related to Pd catalyst poisoning. This hypothe-
sis is supported by a recent XAS investigation that has re-
vealed that, after a AEM-DEFC galvanostatic trial, palladium is
We have observed that the performance of Pd-CeO /C is
2
better for all the reported data, irrespectively of the fuel com-
position, indicating this electrocatalyst as the best possible
electrocatalyst for exploitation in devices. We also want to
mention that the energy efficiency are still rather low suggest-
ing that, at least for the moment AEM- DEFCs are best suited
for supplying low power densities to small portable micro-elec-
tronic devices.
[33]
oxidized to inactive species. This oxidation depends on the
time that the catalyst spends at large positive overpotentials
(
anodic stress). This time is larger for high concentration etha-
nol fuels as compared to the low ethanol ones. Indeed, galva-
À2
nostatic (Figure 5c) curves at 20 mAcm last for 16 and 40 h
To this regard, encouraging results have been obtained by
2
for 2 and 6m fuels respectively. It is worth mentioning that
such a deactivation is, at least partially, reversible and that the
catalytic activity can be recovered by exchanging the exhaust
fuel batch with a fresh one.
assembling an air-breathing AEM-DEFC having a 1 cm MEA
composed of the Pd-CeO /C as anode, FeCo/C as cathode, and
2
Tokuyama A201 as anion exchange membrane. A solution of
6m Ethanol and 6m KOH was chosen as fuel while a constant
À
À2
It is widely accepted that the presence of OH plays a funda-
current density of 1 mAcm was selected to test the stability
mental role in the mechanism of ethanol electrooxidation in al-
of the device. As can be seen from Figure S2, (Supporting In-
formation) the cell ran for 40 days with a potential drop from
0.64 V to 5 V.
À
kaline conditions. The importance of the OH concentration re-
sides in its role in both thermodynamic and kinetic aspects of
ethanol electro-oxidation. In alkaline direct ethanol fuel cells
selective toward the formation of acetate five hydroxyl ions
are consumed for each oxidized ethanol molecule. This is es-
sential to guarantee the electro-neutrality according to the re-
action reported in Equation (8).
Conclusions
In summary, the performance of AEM-DEFCs equipped with
palladium based anodes have been investigated in terms of
maximum power density, discharge energy density per fuel
batch, and energy efficiency. All the parameters show a remark-
able dependence on the fuel composition. The maximize
power density was obtained with a balanced electrolyte con-
taining 4m ethanol and 4m KOH. Further increase in KOH con-
centration results in a decrease of the maximum power density
that we have also shown that this is scarcely dependent on
À
À
CH CH OH þ O þ OH ! CH COO þ 2 H O
ð8Þ
3
2
2
3
2
From Equation (8) we notice that the oxidation of ethanol
results in the formation of one acetate ion. Hence a hydroxyl
ion has to be consumed to balance the charge. A consequence
of this is the fact that as ethanol is consumed during the ex-
ChemCatChem 2015, 7, 2214 – 2221
www.chemcatchem.org
2219
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Acknowledgements
Authors acknowledge: the Financial support from MIUR (Italy),
FIRB 2010 project RBFR10J4H7, the Ente Cassa di Risparmio di
Firenze for the project HYDROLAB and International Centre of
Theoretical Physics (Programme for Training and Research in Ital-
ian Laboratories (TRIL) Trieste.
Keywords: electrocatalysts · fuel cells · microelectronics ·
nanocatalysis · palladium
[1] D. J. Kim, E. A. Cho, S. A. Hong, I. H. Oh, H. Y. Ha, J. Power Sources 2004,
1
30, 172–177.
[
[
2] Z. Guo, Y. Cao, J. Power Sources 2004, 132, 86–91.
3] F. Achmad, S. K. Kamarudin, W. R. W. Daud, E. H. Majlan, Appl. Energy
2011, 88, 1681–1689.
[
[
4] D. R. Chu, R. Z. Jiang, Electrochim. Acta 2006, 51, 5829–5835.
5] J. G. Liu, T. S. Zhao, R. Chen, C. W. Wong, Electrochem. Commun. 2005, 7,
288–294.
[
[
6] R. Hammerschlag, Environ. Sci. Technol. 2006, 40, 1744–1750.
7] Global Ethanol Supply to Exceed 100 Mln Tonnes in 2018, According to
Topical Report by Merchant Research & Consulting.
[
[
8] F. Vigier, C. Coutanceau, A. Perrard, E. M. Belgsir, C. Lamy, J. Appl. Electro-
chem. 2004, 34, 439–446.
9] W. J. Zhou, Z. H. Zhou, S. Q. Song, W. Z. Li, G. Q. Sun, P. Tsiakaras, Q. Xin,
Appl. Catal. B 2003, 46, 273–285.
[
[
10] E. Antolini, J. Power Sources 2007, 170, 1–12.
11] S. Srinivasan, O. A. Velev, A. Parthasarathy, D. J. Manko, A. J. Appleby, J.
Power Sources 1991, 36, 299–320.
[
[
12] R. W. Verjulio, F. Alcaide, G. Alvarez, N. Sabate, N. Torres-Herrero, J. P. Es-
quivel, J. Santander, J. Micromech. Microeng. 2013, 23, 115006–1150.
13] A. Morozan, B. Jousselme, S. Palacin, Energy Environ. Sci. 2011, 4, 1238–
1254.
[
[
14] C. Bianchini, P. K. Shen, Chem. Rev. 2009, 109, 4183–4206.
15] X. Fang, L. Q. Wang, P. K. Shen, G. F. Cui, C. Bianchini, J. Power Sources
2
010, 195, 1375–1378.
16] L. Ma, D. Chu, R. R. Chen, Int. J. Hydrogen Energy 2012, 37, 11185–
1194.
17] G. Cui, S. Song, P. K. Shen, A. Kowal, C. Bianchini, J. Phys. Chem. C. 2009,
13, 15639–15642.
18] F. Vigier, S. Rousseau, C. Coutanceau, J. M. Leger, C. Lamy, Top. Catal.
006, 40, 111 –121.
19] V. Bambagioni, M. Bevilacqua, J. Filippi, A. Marchionni, S. Moneti, F.
Vizza, C. Bianchini, Chemistry Today - Chim. Oggi 2010, 28, VII–X.
20] A. Brouzgou, A. Podias, P. Tsiakaras, J. Appl. Electrochem. 2013, 43, 119–
[
[
[
[
[
1
Figure 6. Maximum power density, discharge energy density and energy effi-
ciency of the passive AEM-DEFCs fueled with ethanol and KOH at different
concentration.
1
2
ethanol concentration. For the discharge energy density per
fuel batch we have shown that high KOH and ethanol concen-
tration are beneficial and that, in general, to better exploit the
energy content of the fuel it is better to operate with an
excess KOH concentration (e.g. ethanol 4m KOH 6m).
The energy efficiency for balanced electrolytes (i.e. 2m etha-
nol l+2m KOH, 4m ethanol+4m KOH, 6m ethanol+6m KOH)
decreases as the concentration increases. This effect has been
ascribed to a progressive reduction of the faradaic efficiency
produced by palladium catalyst oxidation. Further efforts are
needed to achieve stable palladium based electrocatalysts for
AEM-DEFCs operating in a basic environment. The use of
1
36.
[21] C. Bianchini, V. Bambagioni, J. Filippi, A. Marchionni, F. Vizza, P. Bert, A.
Tampucci, Electrochem. Commun. 2009, 11, 1077–1080.
22] Y. Chen, M. Bellini, M. Bevilacqua, P. Fornasiero, A. Lavacchi, H. A. Miller,
L. Wang, F. Vizza, ChemSusChem 2015, 8, 524–533.
23] K. J. J. Steinbusch, H. V. M. Hamelers, J. D. Schaap, C. Kampman, C. J. N.
Buisman, Environ. Sci. Technol. 2010, 44, 513–517.
24] A. Dutta, J. Datta, J. Mater. Chem. A 2014, 2, 3237–3250.
25] V. Bambagioni, C. Bianchini, Y. X. Chen, J. Filippi, P. Fornasiero, M. Inno-
centi, A. Lavacchi, A. Marchionni, W. Oberhauser, F. Vizza, ChemSusChem
[
[
[
[
2
012, 5, 1266–1273.
[26] Y. X. Chen, A. Lavacchi, S. P. Chen, F. di Benedetto, M. Bevilacqua, C.
Bianchini, P. Fornasiero, M. Innocenti, M. Marelli, W. Oberhauser, S. G.
Sun, F. Vizza, Angew. Chem. Int. Ed. 2012, 51, 8500–8504; Angew. Chem.
2
012, 124, 8628–8632.
a mixed carbon-CeO support does extend the stability of the
2
[27] H. M. Mao, L. L. Wang, P. P. Zhu, Q. J. Xu, Q. X. Li, Int. J. Hydrogen Energy
2014, 39, 17583–17588.
Pd catalyst under working conditions by promoting ethanol
electrooxidation at lower anode potentials. We believe that
these systems are ideal for powering portable devices that re-
quire low power densities over long periods such as microelec-
tronics.
[
28] Z. X. Liang, T. S. Zhao, J. B. Xu, L. D. Zhu, Electrochim. Acta 2009, 54,
203–2208.
2
[
29] 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.
ChemCatChem 2015, 7, 2214 – 2221
www.chemcatchem.org
2220
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[
[
[
30] V. Bambagioni, C. Bianchini, A. Marchionni, J. Filippi, F. Vizza, J. Teddy, P.
Serp, M. Zhiani, J. Power Sources 2009, 190, 241–251.
31] M. Grden, M. Lukaszewski, G. Jerkiewicz, A. Czerwinski, Electrochim. Acta
[33] L. Wang, A. Lavacchi, M. B. F. D’Acapito, F. D. Benedetto, M. Innocenti,
H. A. Miller, G. Montegrossi, C. Zafferoni, F. Vizza, Electrochim. Acta 2015,
DOI:10.1016/j.electacta.2015.02.026.
2008, 53, 7583–7598.
32] R. Pattabiraman, Appl. Catal. A 1997, 153, 9–20.
Received: February 26, 2015
Revised: April 15, 2015
Published online on June 18, 2015
ChemCatChem 2015, 7, 2214 – 2221
www.chemcatchem.org
2221
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim