1°Revisão

Electrochimica Acta 52 (2007) 6997–7006

Carbon-supported ternary PtSnIr catalysts for direct

ethanol fuel cell
J. Ribeiro a,1 , D.M. dos Anjos a,c , K.B. Kokoh a,1 , C. Coutanceau a , J.-M. Léger a,1 ,
P. Olivi b , A.R. de Andrade b , G. Tremiliosi-Filho c,∗,1
a Equipe Electrocatalyse, UMR 6503 CNRS, Université de Poitiers, 40 avenue du Recteur
Pineau 86022 Poitiers Cedex, France
b Departamento de Quimica da Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto,

Universidade de São Paulo, Av. Bandeirantes, 3900, 14040-901 Ribeirão Preto, SP, Brazil
c Instituto de Quimica de São Carlos, Universidade de São Paulo, Caixa Postal 780,

13560-970 São Carlos, SP, Brazil

Received 23 February 2007; received in revised form 10 May 2007; accepted 10 May 2007
Available online 17 May 2007

Abstract
Binary PtIr, PtSn and ternary PtSnIr electrocatalysts were prepared by the Pechini–Adams modified method on carbon Vulcan XC-72, and these
materials were characterized by TEM and XRD. The XRD results showed that the electrocatalysts consisted of the Pt displaced phase, suggesting
the formation of solid solutions between the metals Pt/Ir and Pt/Sn. However, the increase in Sn loading promoted phase separation, with the
formation of peaks typical of cubic Pt3 Sn. The electrochemical investigation of these different electrode materials was carried out as a function
of the electrocatalyst composition, in a 0.5 mol dm−3 H2 SO4 solution, with either the presence or the absence of ethanol. Cyclic voltammetric
measurements and chronoamperometric results obtained at room temperature showed that PtSn/C and PtSnIr/C displayed better electrocatalytic
activity for ethanol electrooxidation compared to PtIr/C and Pt/C, mainly at low potentials. The oxidation process was also investigated by in situ
infrared reflectance spectroscopy, to identify the adsorbed species. Linearly adsorbed CO and CO2 were found, indicating that the cleavage of the
C C bond in the ethanol substrate occurred during the oxidation process. At 90 ◦ C, the Pt89 Sn11 /C and Pt68 Sn9 Ir23 /C electrocatalysts displayed
higher current and power performances as anode materials in a direct ethanol fuel cell (DEFC).
© 2007 Elsevier Ltd. All rights reserved.

Keywords: PtSnIr nanoparticles; Ethanol oxidation; Binary and ternary electrocatalysts; Direct ethanol fuel cell

1. Introduction Despite its attractive properties, finding a catalyst that allows a

complete ethanol oxidation to produce 12 electrons per molecule
Direct ethanol fuel cells (DEFCs) have attracted a great inter- still remains a great challenge. While Pt-based anodes are the
est as power sources in numerous applications, at low operating best catalysts for alcohol oxidation in acid medium, Sn is well
temperature. Among several alcohols which can be used in known for being potentially able to provide O-species for the
a direct fuel cell (DEFC), ethanol is the most promising one oxidation of CO produced during the dissociative adsorption
because it is a safe molecule comparatively to methanol. As a of ethanol on Pt active sites [1–11]. Depending on the method
liquid it is easier to be stored and handled when comparing to employed in the preparation of bimetallic PtSn catalysts, the
hydrogen. Moreover, ethanol has a high theoretical mass energy resulting material can have different electrocatalytic properties
(8 kWh kg−1 for ethanol versus 6.1 kWh kg−1 for methanol) [1]. concerning ethanol oxidation [1,7,12–18]. Recently, Jiang et al.
[12] reported the behavior of a PtSn alloy and PtSnOx particles
prepared by a modified polyol process. The authors showed that
∗ Corresponding author. Tel.: +55 16 33739933; fax: +55 16 33739952.
the lattice constant of Pt in the PtSn alloy was dilated because Sn
E-mail address: [email protected] (G. Tremiliosi-Filho). atoms penetrated into the Pt crystalline lattice. As for the PtSnOx
1 ISE member. nanoparticles, a little change in the spacing of the Pt(1 1 1)

0013-4686/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2007.05.017
6998 J. Ribeiro et al. / Electrochimica Acta 52 (2007) 6997–7006

plane was observed. They correlated this structural difference to put in evidence the electrocatalytic activities of the prepared
between the catalysts to their electrocatalytic behavior in the catalysts.
hydrogen adsorption/desorption region; i.e., the ability of tin
oxides present in the vicinity of the Pt nanoparticles to convert 2. Experimental
the CO species into CO2 .
Many transition metals and their oxides such as RuOx , SnOx , 2.1. Catalyst preparation
VOx and MoOx have been investigated as possible Pt promot-
ers for alcohol oxidation [3,8,19–24]. Our previous works and Supported Pt-based catalysts were synthesized by the ther-
some reports in the literature showed that the mechanism of mal decomposition of a polymeric precursor [40–44]. The Pt
ethanol electrooxidation with Pt-based bimetallic electrocata- and Ir polymeric precursors (metal-resin) were separately pre-
lysts was strongly modified [3,6,7,12,15–18,25–34]. Ethanol pared, mixing citric acid (CA) (Merck) in ethylene glycol
oxidation occurs at lower potentials (<0.5 V versus RHE) (EG) (Merck) at 60–65 ◦ C. The metal precursor (H2 PtCl6 or
through adsorbed acetyl species, which leads to the forma- IrCl3 ·xH2 O Aldrich—0.1 mol dm−3 solution dissolved in iso-
tion of acetaldehyde and acetic acid as the main products. propanol) was then added, so that a CA:EG:M 1:4:0.25 molar
These final reaction products, including CO2 , and poisoning ratio (M = Pt or Ir) would be achieved. After total dissolution
intermediate species such as carbon monoxide were well iden- of the precursor salt, the temperature was raised to 90 ◦ C, and
tified by in situ or online spectroscopic techniques (FTIRS, the mixture was kept under vigorous stirring for 2–3 h. The tin
DEMS, respectively) and chromatographic analysis, although polymeric precursor (Sn-resin) was also prepared in a similar
the overall oxidation mechanism remains to be better elucidated way, but the molar ratio CA:EG:TC was 3:10:1, where TC is
[4,5,9,10,16,18]. tin citrate, prepared as described elsewhere [40,41]. Precursor
The aim of this work is to report preliminary results on mixtures were prepared for each nominal composition by dis-
the development of ternary PtSnIr/C catalysts prepared by the solving the appropriate amounts of Pt-resin (2.9 × 10−4 mol of
thermal decomposition of a polymeric precursor (DPP) and Pt/g of resin), Ir-resin (3.5 × 10−4 mol of Ir/g of resin) and Sn-
their application in ethanol electrooxidation. According to many resin (7.1 × 10−4 mol of Sn/g of resin). Carbon (Vulcan XC
reports, the optimum Sn content in PtSn catalysts for the best 72), which had been previously treated for 4 h at 400 ◦ C under
cell performance lies in the range 10–40%, depending on the nitrogen atmosphere, was added to the precursor mixture, to
DEFC operation temperature and working point. Under these obtain a catalyst loading of 40 wt.%. This mixture was finally
conditions, the addition of a third metal could contribute to dispersed in 2 mL ethanol by ultrasonication for 10 min. Ther-
decreasing the Pt loading, stabilizing the PtSn alloy and increas- mal treatments were carried out in a tubular oven under nitrogen
ing the ethanol oxidation efficiency. Chen et al. [35] showed that atmosphere. The temperature was firstly increased to 250 ◦ C at
the presence of an IrO2 thin layer largely increased the electro- a rate of 1 ◦ C/min and it was kept there for 1 h. Then, the tem-
catalytic activity of the Ti/IrO2 /Pt electrodes towards methanol perature was increased to 300 ◦ C at a rate of 15 ◦ C/min, and
oxidation, since not only the current density increased but the kept for another 1 h. Afterwards, the temperature was increased
onset potential of methanol electrooxidation also shifted over to 400 ◦ C at a rate of 30 ◦ C/min and kept there for 2 h. Finally,
200 mV downward compared to Ti/Pt nanoparticles. Iridium after being cooled down to room temperature, the electrocata-
oxide also exhibits an excellent corrosion resistance in acidic lysts were again kept in an oven under air atmosphere at 400 ◦ C
medium, good electrical conductivity, and high electrocatalytic for 1–2 h, to eliminate the excess carbon (compare the two last
activity towards oxygen evolution. IrO2 is widely used in the columns in Table 1).
preparation of mixed oxide electrodes and several reports have
mentioned the use of Ir and its oxides in fuel cells [36–39]. In this 2.2. Physicochemical characterizations of the catalysts
work, the electro-activity for ethanol oxidation of the PtSnIr/C
catalysts will be compared to that of Pt/C, PtIr/C, and PtSn/C XRD patterns of the catalysts were obtained with an
catalysts, and single cell DEFC performance will be reported, X-ray diffractometer (model D5005 Brucker AXS instrument)

Table 1
EDX analysis of Pt-based/C electrocatalysts with various bulk compositions prepared by the DPP process
Electrocatalyst Nominal composition (%) EDX (%) mnominal (mg) mexperimental (mg)

90:10 89:11 150 151.1 (39.7%a )

PtSn/C
80:20 79:21 300 326.5 (36.8%a )
98:2 98:2 100 109.9 (36.4%a )
PtIr/C
80:20 73:27 300 280.5 (42.8%a )
94:5:1 90:8:2 300 275.7 (43.5%a )
79:19:2 83:16:1 200 175.6 (45.6%a )
PtSnIr/C 81:18:1 74:25:1 300 260.1 (46.1%a )
73:25:2 67:29:4 300 271.6 (44.2%a )
70:10:20 68:9:23 300 280.0 (42.8%a )
a Final metallic charge.
 J. Ribeiro et al. / Electrochimica Acta 52 (2007) 6997–7006 6999

operating with the Cu K␣ radiation (λ = 1.5406 Å) generated at All solutions were prepared using 18.2 M cm water,
40 kV and 40 mA. The following parameters were kept the same produced and purified by a Millipore-Milli-Q system. The elec-
for all samples during the X-ray analysis: 2θ range = 20–90◦ , trolyte solution was purged with nitrogen (U quality) prior to
step = 0.03◦ , step time = 3 s, and total analysis time = 1.94 h. The the measurements, to expel oxygen. Nitrogen was also used
catalyst phase composition and the analysis of the relevant posi- to maintain an inert atmosphere over the solutions during the
tion to the K␣1 monochromatic radiation were obtained by fitting measurements.
the experimental angular range of interest to the pseudo-Voigt
function per crystalline peak with a computer refinement pro- 2.4. IR reflectance spectroscopic measurements
gram (Profile Plus Executable, Siemens AG). The crystallite size
values were obtained using the Debye-Scherrer equation [45]: Ten microliters of the electrocatalyst ink (1 mg of cata-
lyst + 275 ␮L ethanol + 15 ␮L Nafion® ) were deposited onto a
K(λ × (180◦ /π)) gold substrate previously polished with alumina (0.3 ␮m) to a
D=  ,
( β2 − S 2 ) cos θβ mirror-finish in order to obtain the best reflectivity. The result-
ing electrode was then dried in an oven at 75 ◦ C for 5 min. IR
where D is the apparent crystallite mean size, perpendicular reflectance spectra were recorded on a Bruker IFV 66v spec-
to the reflecting plane, K the shape factor (0.9 for spherical trometer with data acquisition techniques, which enabled the
crystallite), λ the radiation wavelength (1.5406 Å), S the performance of Single Potential Alteration Infrared Reflectance
instrumental line broadening (0.001◦ ), β the reflection width Spectroscopy (SPAIRS) experiments. The reflectivities were
at half-maximum intensity (FWHM), and θ β is the angle at the recorded at 50 mV intervals during the first voltammetric scan
maximum intensity. at a low sweep rate (1 mV s−1 ).
Cell parameters were calculated by a computer program
which took into account the unit cell parameters using the least- 2.5. Preparation of the catalysts and of the
squares method (U-Fit.exe v.1.3, 1992). Experimental 2θ and membrane/electrode assemblies (MEAs)
reflection parameters (h k l) were used as a first guess to obtain
the unit cell parameter. The membrane electrode assemblies (MEAs) were prepared
The catalyst powder was mixed with water and ethanol and by hot pressing a pretreated Nafion® 117 membrane placed
ultrasonically treated until complete ink homogenization was between an E-TEK cathode (2 mg cm−2 metal loading 40 wt.%)
achieved. The morphology of the different catalyst inks was and a home-made anode (2 mg cm−2 metal loading 40 wt.%)
then examined by Transmission Electron Microscopy (TEM) at 130 ◦ C for 90 s under a pressure of 35 kg cm−2 . The proce-
using a Philips CM 120 microscope equipped with an LaB6 dure of electrode fabrication has been described elsewhere [29].
filament. Energy dispersive X-rays analysis (EDX) using a The operating fuel cell performances were determined in a sin-
Leica-Zeiss LEO 440 model SEM coupled to an Oxford 7060 gle DEFC with an electrode geometric surface area of 5 cm2 ,
model analyzer also allowed us to estimate the nanoparticle using a Globe Tech test bench. The temperature was set to
composition. 90 ◦ C for the fuel cell and 95 ◦ C for the oxygen humidifier. The
ethanol and oxygen pressures were set to 1 and 3 bar, respec-
2.3. Electrochemical measurements tively. The E versus i and P versus i curves were recorded using a
high power potentiostat (Wenking model HP 88) interfaced with
Catalyst loads of 2 mg were used to prepare the anodes. They a PC.
were dispersed in a solution (100 ␮L) containing ethanol (95 ␮L)
and Nafion® (5 ␮L) (5 wt.% in aliphatic alcohol, Aldrich). The 3. Results and discussion
resulting ink was ultrasonically homogenized for 10 min. After
homogenization, it was deposited onto vitreous carbon previ- 3.1. Physicochemical characterizations of the catalysts
ously polished with alumina (0.5 ␮m) to a mirror-finish, and
then dried in an oven at 75 ◦ C for 10 min. EDX results for all the prepared catalysts are presented in
Electrochemical measurements were carried out in a conven- Table 1, which indicates that each metal is present in a composi-
tional three-electrode cell (30 mL), using an Autolab potentiostat tion close to the nominal values. However, analysis of the ternary
(PGSTAT30). A Reversible Hydrogen Electrode (RHE) and catalysts, except PtSnIr (70:10:20), reveals some heterogeneity
vitreous carbon with a 4 cm2 geometric surface area were in some regions.
used as reference and counter electrodes, respectively. The The TEM images and particle size distribution for all the
anode potential was taken versus the reference electrode, investigated electrocatalysts are shown in Figs. 1 and 2, respec-
which was separated from the solution by a Luggin-Haber tively, with particle sizes lying between 2 and 8 nm. Some large
capillary tip. The electrode activity was evaluated by cyclic agglomerates with sizes between 10 and 15 nm, which consist
voltammetry (CV, recorded in the 0.05–1.4 V versus RHE poten- of smaller particles with an average size close to 4 nm, are also
tial range, in 0.5 mol dm−3 H2 SO4 (Aldrich) or 0.5 mol dm−3 observed (Fig. 1B and C).
H2 SO4 + 1.0 mol dm−3 ethanol (Normapur) solution), and Fig. 3 shows the XRD patterns of the carbon-supported Pt,
chronoamperometry (in 0.5 mol dm−3 H2 SO4 + 1.0 mol dm−3 binary PtIr and PtSn (Fig. 3A), and ternary PtSnIr catalysts
ethanol solution at 400 mV for 2 h). (Fig. 3B). The main characteristic peaks of face-centered cubic
7000 J. Ribeiro et al. / Electrochimica Acta 52 (2007) 6997–7006

Fig. 1. TEM images of different electrocatalysts on carbon XC-72. (A) Pt79 Sn21 /C, (B) Pt73 Ir27 /C, (C) and (D) Pt68 Sn9 Ir23 /C.

(fcc) crystalline Pt (spatial group Fm-3m) [46] appears in all behavior suggests the formation of a solid solution between Pt
the XRD patterns, which shows that all the catalysts resemble and Ir.
the phase disordered structure (solid solution) with the reflec- XRD results for the different Pt-based/C electrocatalysts (lat-
tion planes (1 1 1), (2 0 0), (2 2 0), (3 1 1), and (2 2 2). Fig. 3A tice parameter, cell volume and apparent mean size D) are given
shows that the diffraction peaks of the binary PtSn electrocat- in Table 2. The introduction of Ir into the binary PtIr catalyst
alysts are shifted to lower 2θ. Moreover, the increase in Sn leads to a decrease in the lattice parameter, which is in agreement
loading leads to phase separation, responsible for the character- with the Vegard’s law for a true PtIr alloy [47]. On the other hand,
istic peaks attributed to cubic Pt3 Sn (spatial group Pm-3m) with an increase in Sn loading for both binary and ternary electrocat-
the reflection planes (1 1 1), (2 0 0), (2 2 0), (3 1 1) and (2 2 2) alysts leads to an increase in the lattice parameter up to a 25 at.%
[46]. This same behavior is observed for the ternary electro- Sn loading (from 3.907 to 3.929 Å), suggesting that there is a
catalysts (Fig. 3B), showing that the electrocatalysts containing straight dependence of the lattice parameter on Sn at.%, which
10 mol% Sn or more have separated phases. On the other hand, also follows Vegard’s law [48]. A complete phase separation is
the binary PtIr electrocatalysts shift the reflection planes to observed for the electrocatalysts with an Sn loading greater than
higher 2θ-values only, and no phase separation occurs. This 29 at.%. This is because the lattice parameter returns to values of
 J. Ribeiro et al. / Electrochimica Acta 52 (2007) 6997–7006 7001

Fig. 2. Particle size distribution of Pt-based/C electrocatalysts (40% metal loading) prepared by thermal decomposition of polymeric precursors.

pure Pt (a = 3.919 Å). This can also be confirmed by the lattice

parameter of the Pt3 Sn phase, which exhibits the same behavior
(from 3.958 to 3.985 Å). When Zhou et al. studied PtSn alloys
prepared by impregnation for 4 h, they also observed an increase
in the lattice parameters as a function of the Sn amount [49]. In
addition, Roman-Martinez et al. observed that the Pt/Sn ratio
plays an important role in the distribution of PtSn and Pt3 Sn
phase distribution in the electrocatalyst [50]. However, one can-
not rule out the effect of carbon and the third element (Ir atom
in the ternary electrocatalysts), which may explain why the lat-
tice parameters of the electrocatalysts are slightly different from
those reported for pure Pt and Pt3 Sn [46].
An apparent mean size D close to 10 nm was obtained for all
the electrocatalysts, except for Pt73 Ir27 /C, which has a D value
close to 5 nm. These results are not in agreement with the TEM-
data. This behavior can be explained by the fact that XRD data
describe well-defined crystalline particles, but not amorphous
ones. As a result, the average particle size is slightly different
from those obtained by TEM.

3.2. Electrochemical and in situ IR spectroscopic

characterizations

Fig. 4A shows the CV curves obtained for the different

electrocatalysts deposited on vitreous carbon in the absence
of ethanol. The voltammograms do not display a well-defined
hydrogen oxidation region (0.0–0.4 V versus RHE), conversely
to the case of pure platinum, and the current densities in the
double layer region (between 0.4 and 0.85 V versus RHE) are
larger than those on pure platinum. This behavior is character-
istic of binary and ternary electrocatalysts containing transition
metals [51,52]. The increase of the tin amount in the binary
PtSn/C electrocatalysts leads to a two-fold increase at least in
the anodic charge. Moreover, the beginning of the oxygen evo-
lution reaction (OER) occurs at lower potentials for Pt79 Sn21 /C
Fig. 3. XRD patterns of Pt-based electrocatalysts prepared by the DPP process compared to Pt89 Sn11 /C. The introduction of Ir into pure Pt/C
and dispersed on carbon Vulcan XC-72 (40 wt.% metal loading). (A) Pt, binary leads to an increase in the voltammetric charge. This behavior
PtIr, PtSn and (B) ternary PtSnIr electrocatalysts. can be explained by a better material dispersion on carbon and/or
7002 J. Ribeiro et al. / Electrochimica Acta 52 (2007) 6997–7006

Table 2
XRD results for different Pt-based/C electrocatalysts prepared by the DPP process

Electrocatalyst a (Å) V (Å3 ) D (nm)

(1 1 1) (2 0 0) (2 2 0) (3 1 1)

Pt100 /C 3.916 60.06 11.5 9.3 8.9 8.1

Pt98 Ir2 /C 3.902 59.41 12.3 11.2 11.0 10.9
Pt73 Ir27 /C 3.899 59.32 5.8 5.6 5.4 4.5
Pt89 Sn11 /C
Pt 3.907 59.66 9.5 8.3 8.3 7.6
Pt3 Sn 3.958 62.01 10.0 9.0 7.6 7.2
Pt79 Sn21 /C
Pt 3.924 60.40 11.2 9.9 9.0 8.3
Pt3 Sn 3.985 63.30 12.9 11.3 10.6 9.9
Pt90 Sn8 Ir2 /C 3.909 59.72 9.3 8.3 8.3 8.0
Pt83 Sn16 Ir1 /C
Pt 3.913 59.91 8.9 8.5 8.0 8.0
Pt3 Sn 3.966 62.40 9.1 8.0 9.1 8.1
Pt74 Sn25 Ir1 /C
Pt 3.929 60.66 10.7 9.4 8.9 8.7
Pt3 Sn 3.984 63.24 11.1 9.9 9.4 8.6
Pt67 Sn29 Ir4 /C
Pt 3.919 60.20 8.7 8.0 8.0 7.5
Pt3 Sn 3.979 62.99 10.1 9.1 8.4 8.6
Pt68 Sn9 Ir23 /C 3.918 60.14 8.9 8.2 8.2 8.1

Data: aPt = 3.920, aIr = 3.900 and aPt3 Sn = 4.001 [46].

the formation of ultrafine particles. The latter phenomenon was its poisoning by strongly adsorbed intermediates such as COads .
being confirmed by XRD data, which give a D value close to However, the introduction of Sn and/or Ir into the binary and
5 nm. Comparing the binary and ternary electrocatalysts, the ternary electrocatalysts promotes the electrocatalytic activity. In
introduction of Ir into binary PtSn/C increases the voltammetric fact, as observed in Fig. 4B, the introduction of Sn and/or Ir leads
charge and shifts the OER onset potential upwards. Accord- to an increase in the electro-activity of the binary and ternary
ing to Juodkazis et al., the existence of reversible metal-oxide electrocatalysts compared to pure Pt/C. The synergistic effect
layers demonstrated experimentally for noble metals such as obtained with these elements can be explained by the activation
iridium, oxidized anodically in the range of potentials before O2 of interfacial water molecules at lower potentials than in the case
evolution, shows that nanometric layers of amorphous, hydrous of pure Pt due to the presence of preferential sites for OHads
oxides or oxides can be reversibly formed on the electrode sur- adsorption, as proposed in the case of PtRu for methanol elec-
face during the anodic process [52]. Although this is not clearly trooxidation [19,54]. The presence of OHads species in large
observed in the corresponding XRD patterns, the presence of Sn amounts is necessary for the complete oxidation of poisoning
and Ir oxides in small amounts and amorphous forms cannot be intermediates such as CHx and CO. Furthermore, when Tacconi
discarded, which can contribute to increasing the double layer et al. studied metallic Ir by FTIRS, they observed that ethanol
charge of the catalysts. adsorption led to the formation of linearly bonded CO on Ir.
Fig. 4B shows the representative cyclic voltammograms They also observed that oxidized states at lower potentials play
obtained with different electrocatalysts (Pt/C, Pt79 Sn21 /C and an activating role in ethanol electrooxidation [55].
Pt68 Ir23 Sn9 /C) deposited on vitreous carbon in the presence of Fig. 5 shows the anodic sweep curves in the presence of
1 mol dm−3 ethanol. At a first glance, the ethanol oxidation cur- 1 mol dm−3 ethanol recorded for the different electrocatalysts
rent at Pt-based catalysts begins at ∼0.25 V versus RHE. The in the potential range between 0.05 and 0.63 V versus RHE.
ethanol oxidation onset potential is about 200 mV lower than The binary and ternary electrocatalysts display better perfor-
that obtained with pure Pt/C electrocatalysts. Four oxidation mances than Pt/C for ethanol oxidation. Moreover, when the
peaks are also observed in the presence of ethanol, two dur- binary electrocatalysts are compared to the ternary ones in terms
ing the positive-going scan (at 0.7 and 1.1 V versus RHE) and of oxidation wave onset potential and achieved current densi-
two during the reverse scan (at 0.82 and 0.72 V versus RHE, ties, the latter catalysts give the best electrical performances.
respectively). Similar results were obtained by Avaca et al. when In the case of the ternary electrocatalysts, the best performance
they investigated ternary Pt-RuO2 -IrO2 /C catalysts for ethanol is achieved with lower tin atomic ratios (8 or 9%), which cor-
oxidation [53]. roborates previous results showing that low tin atomic ratios
As already known [1,53], pure Pt/C catalyst (Fig. 4B) does not (close to 10%) in bimetallic PtSn electrocatalysts lead to higher
behave as a very good anode for ethanol electrooxidation due to activities towards ethanol oxidation [26,56]. On the other hand,
 J. Ribeiro et al. / Electrochimica Acta 52 (2007) 6997–7006 7003

Fig. 6. Current vs. time plots for the electrooxidation of 1.0 mol dm−3 ethanol in
0.5 M H2 SO4 at 0.4 V vs. RHE on various Pt-based/C electrocatalysts. Current
values normalized by the Pt loading.

addition of iridium to platinum has little effect, whereas addition

of iridium to PtSn greatly enhances the electrocatalytic activity.
The role of iridium is thus still difficult to explain, but its pres-
ence may lead to a decrease in platinum poisoning by diluting
platinum sites, thus limiting the ethanol adsorption mode. This
should lead to the formation of strongly adsorbed CO species, as
recently proposed for the PtPd catalysts in the case of ethylene
glycol electrooxidation [57].
Long-term chronoamperometry experiments were carried out
at 400 mV for 2 h, to evaluate the electrocatalytic activity of
the catalysts and the poisoning of the active surface under
Fig. 4. Cyclic voltammograms for the 40 wt.% Pt-based/C electrodes in: (A) continuous operation conditions. Fig. 6 shows representative
N2 -purged 0.5 mol dm−3 H2 SO4 electrolyte at 20 mV s−1 and (B) - - - Pt100 /C chronoamperograms obtained for the different electrocatalysts
– - – Pt79 Sn21 /C and –– Pt68 Sn9 Ir23 /C electrodes in N2 -purged 0.5 mol dm−3 whose current densities were normalized by Pt mass. During the
H2 SO4 + 1.0 mol dm−3 ethanol electrolyte at 20 mV s−1 . Current values nor-
malized by the Pt loading.
first minutes, there is a sharp decrease in the current density, fol-
lowed by a slow decrease in the current values for longer time
periods. A steady-state current was not observed for all stud-
ied catalysts. In the sequence, the origin of the current decays
can be briefly explained. Firstly, under an applied potential of
400 mV, ethanol adsorbs on site which is initially covered by
a water molecule. After adsorption, ethanol can either dissoci-
ate, producing strongly adsorbed CO and CHx type residue, or
oxidized to acetaldehyde. Further oxidation requires the pres-
ence of oxygenated species on the surface, as OHads , to form
acetic acid or CO2 [58]. The dissociative adsorbed intermediates
formed at the beginning of the oxidation process, CO and CHx ,
are the main responsible for the initial and accentuated decay
of the current–time curves. This is due to their strong adsorp-
tion on the active sites causing the poisoning of the surface.
Additionally, surface instability of nanoparticles, due to surface
re-crystallization, surface metal segregation, particle agglom-
eration, etc., decreases slowly the number of active sites and
are responsible for the continuous and slow current–time decay
Fig. 5. Cyclic voltammetry of 40 wt.% Pt-based/C electrodes recorded in N2 - in the long operation time. Consequently, both effects should
purged 0.5 mol dm−3 H2 SO4 + 1.0 mol dm−3 ethanol electrolyte at 20 mV s−1 , be take into account for the non-achievement of a steady-state
considering only the anodic sweep. Current values normalized by the Pt loading. current in the chronoamperometric curves.
7004 J. Ribeiro et al. / Electrochimica Acta 52 (2007) 6997–7006

Fig. 7. SPAIR spectra of the species resulting from the adsorption and oxida-
tion of 0.2 M ethanol in 0.5 M H2 SO4 on a Pt68 Sn9 Ir23 /C electrode at different
potentials.

The binary PtSn/C and ternary PtSnIr/C electrocatalysts give

better results than the binary PtIr/C and pure Pt/C electrocata-
lysts. These latter catalysts display the same behavior, that is,
on both catalysts the platinum active sites are equally poisoned,
which does not confirm the proposed explanation for the role
of iridium. However, the better activity obtained for the ternary
electrocatalysts could be explained by the operation of a ben-
eficial synergistic effect between tin and iridium. Perhaps the
change in the crystalline structure promoted by the introduc-
tion of Ir and/or the increase in the number of exposed active
sites as observed by CV can explain this improvement. This may
indicate an increase in structural defects or roughness, making
the ternary electrocatalysts better candidates for such catalytic Fig. 8. Electrical performances of a 5 cm2 DEFC at 90 ◦ C using the Pt-based
process [59]. Furthermore, the addition of Ir to the PtSn alloy bi and trimetallic anodes (2 mg cm−2 catalyst loading, 40 wt.% catalyst on
electrocatalysts can lead to an increase in the surface oxophilic carbon). PO2 = 3 bar; PEtOH = 1 bar; [EtOH] = 2.0 mol dm−3 ; Nafion® 117
membrane. (A) Cell voltage against current density and (B) power density
character, thus increasing the Sn O bond strength and the acidity against current density.
of the Sn OH sites, which can also favor ethanol electrooxida-
tion at lower potentials [60]. On the other hand, the iridium atoms
may disturb the PtSn sites and orbital symmetries, thus affecting stretching and OH deformation modes from COOH (1370 and
the orbital spatial distribution and the ethanol electrooxidation 1280 cm−1 ) [29]. The IR spectroscopy studies carried out at
rate, as previously proposed for Ru atoms [61]. A quaternary low potentials give evidence of the presence of small amounts
PtRuOsIr catalyst prepared by arc plating was investigated by of intermediate species such as linearly bonded CO, which indi-
Gurau et al. [62]. They observed that an improvement in activity cates the higher catalyst resistance to CO poisoning due to the
was obtained by introduction of Ir into the PtRuOs catalyst, and presence of iridium.
this improvement was due to the ability of Ir to bind to C.
Fig. 7 shows the IR reflectance spectra obtained for ethanol 3.3. Fuel cell electrical performance
oxidation on Pt68 Sn9 Ir23 /Au. The characteristic band for linearly
bonded CO [29] at about 2050 cm−1 is clearly observed between Fig. 8 shows the cell performance of optimized binary (PtIr
0.2 and 0.6 V versus RHE, whereas that due to CO2 formation and PtSn) and ternary PtIrSn electrocatalysts used in the DEFC
located at 2344 cm−1 appears only at potentials higher than 0.5 V anode. For comparison, the DEFC performance of Pt/C elec-
versus RHE. The formation of CO2 is clearly related to the COads trocatalyst is also shown in Fig. 8. The open circuit voltage of
removal, which is the first reaction product of the C C bond the fuel cell containing either PtIr/C or Pt/C electrocatalysts is
cleavage. Moreover, the characteristic absorption bands which 0.40 and 0.35 V, respectively, while the corresponding values for
are present at 1280, 1717 and 2616 cm−1 are attributed, respec- ternary PtSnIr/C and binary PtSn/C electrocatalysts are about
tively, to the formation of products containing a carbonyl CO 0.7 V (Fig. 8A).
stretching mode due to COOH or CHO, a C O symmetric The DEFC performances of binary Pt89 Sn11 /C
stretching mode from adsorbed CH3 COO , and coupled C O (35 mW cm−2 ) are better than that of Pt68 Sn9 Ir23 /C
 J. Ribeiro et al. / Electrochimica Acta 52 (2007) 6997–7006 7005

Table 3
Maximum power densitiy normalized by platinum loading
Electrocatalyst EDX composition (%) Maximum i (mA cm−2 ) Maximum power density (mW cm−2 ) Normalized maximum power (W gPt −1 )

Pt 100 40 4 2
PtIr 73:27 60 8 5
89:11 100 35 20
PtSn
79:21 60 21 14
90:8:2 100 27 14
83:16:1 80 28 15
PtSnIr 74:25:1 40 11 7
67:29:4 80 22 15
68:9:23 100 31 22

(31 mW cm−2 ), see Fig. 8B. However, when normalizing other hand, the presence of small amount of adsorbed CO also
the power density values of the DEFC with respect to indicated its easy removal of the surface of the catalysts through
the Pt loading, the electrical performance of the ternary CO2 formation. This shows a remarkable activity of these Pt-
Pt68 Sn9 Ir23 /C electrode material (22 W gPt −1 ) is markedly based catalysts for ethanol oxidation. The addition of iridium
improved, when comparing to the binary PtSn/C, PtIr/C and to the binary PtSn catalyst enhanced slightly the electrical per-
the other ternary PtSnIr/C electrocatalysts (Table 3). This formance for ethanol electrooxidation in a direct ethanol fuel
table shows that Pt68 Sn9 Ir23 /C displays better performances. cell.
When Pt90 Sn8 Ir2 /C electrocatalysts are used as anode there
is a remarkable decrease in the cell performance. Despite the Acknowledgements
higher activity of Pt90 Sn8 Ir2 /C at lower potentials as observed
in the CV and chronoamperometry experiments carried out at This work was mainly conducted within the framework of
room temperature (see Figs. 5 and 6), the same behavior is not a collaborative program CAPES/COFECUB (Comité Français
confirmed in the experiments realized under real operational d’Evaluation de la Coopération Universitaire avec le Brésil)
conditions. under grant no. 498/05. J. Ribeiro acknowledges CAPES
The DEFC performances obtained for the binary Pt89 Sn11 /C (Brazil) for the postdoc fellowship. G. Tremiliosi-Filho and A.R.
electrocatalyst (35 mW cm−2 ) is better than those reported by de Andrade acknowledge FAPESP and CNPq (Brazil).
Rousseau et al. [31], Antolini et al. [63] and Colmati et al.
[64]. This result demonstrates the efficiency of the preparation References
method. Moreover, comparing the Pt/C and PtIr/C electrocata-
lysts, the addition of tin increases the cell performances. Even [1] F. Vigier, C. Coutanceau, F. Hahn, E.M. Belgsir, C. Lamy, J. Electroanal.
under fuel cell working conditions, the best performance is Chem. 563 (2004) 81.
[2] N. Fujiwara, K.A. Friedrich, U. Stimming, J. Electroanal. Chem. 472 (1999)
achieved with lower tin atomic ratios (close to 10 at.%).
120.
[3] A.O. Neto, M.J. Giz, J. Perez, E.A. Ticianelli, E.R. Gonzalez, J. Elec-
4. Conclusions trochem. Soc. 149 (2002) A272.
[4] V.P. Santos, G. Tremiliosi-Filho, J. Electroanal. Chem. 554–555 (2003)
395.
This work shows that the DPP preparation route is very use- [5] M.G. Pereira, M.D. Jiménez, M.P. Elizalde, A. Manzo-Robledo, N. Alonso-
ful and leads to highly active electrocatalysts. Although the Vante, Electrochim. Acta 49 (2004) 3917.
presence of tin and iridium oxides in small amounts cannot [6] G.A. Camara, R.B. de Lima, T. Iwasita, Electrochem. Commun. 6 (2004)
be discarded (2θ = 26◦ , for example), these carbon-supported 812.
[7] W.J. Zhou, S.Q. Song, W.Z. Li, G.Q. Sun, Q. Xin, S. Kontou, K. Poulianitis,
Pt-based catalysts form mainly alloy nanoparticles which con-
P. Tsiakaras, Solid State Ionics 175 (2004) 797.
tain a large particles size in some zones. Cyclic voltammetry [8] C. Xu, P.K. Shen, X. Ji, R. Zeng, Y. Liu, Electrochem. Commun. 7 (2005)
and chronoamperometry results showed that the binary PtIr/C 1305.
and PtSn/C, and the ternary PtSnIr/C electrocatalysts pro- [9] G.A. Camara, R.B. de Lima, T. Iwasita, J. Electroanal. Chem. 585 (2005)
moted ethanol oxidation from potentials around 250 mV versus 128.
[10] G.A. Camara, T. Iwasita, J. Electroanal. Chem. 578 (2005) 315.
RHE. Cyclic voltammetry and chronoamperometry experiments
[11] M.R. Tarasevich, Z.R. Karichev, V.A. Bogdanovskaya, E.N. Lubnin, A.V.
showed that the binary PtSn/C and the ternary PtSnIr/C elec- Kapustin, Electrochem. Commun. 7 (2005) 141.
trocatalysts displayed the highest activity. Infrared reflectance [12] L. Jiang, G. Sun, S. Sun, J. Liu, S. Tang, H. Li, B. Zhou, Q. Xin, Electrochim.
spectroscopic measurements demonstrated that acetaldehyde Acta 50 (2005) 5384.
was produced. The presence of acetic acid is more important for [13] J.-M. Léger, Electrochim. Acta 50 (2005) 3123.
[14] S.Q. Song, W.J. Zhou, Z.H. Zhou, L.H. Jiang, G.Q. Sun, Q. Xin, V. Leon-
Sn and Ir containing electrocatalysts which activate the inter-
tidis, S. Kontou, P. Tsiakaras, Int. J. Hydrogen Energy 30 (2005) 995.
facial water necessary for its formation. Linearly adsorbed CO [15] W.J. Zhou, S.Q. Song, W.Z. Li, Z.H. Zhou, G.Q. Sun, Q. Xin, S. Dou-
was also seen at lower potentials, indicating that ethanol elec- vartzides, P. Tsiakaras, J. Power Sources 140 (2005) 50.
trooxidation led to more than four-electron conversion. On the [16] H. Wang, Z. Jusys, R.J. Behm, J. Power Sources 154 (2006) 351.
7006 J. Ribeiro et al. / Electrochimica Acta 52 (2007) 6997–7006

[17] J. Mann, N. Yao, A.B. Bocarsly, Langmuir 22 (2006) 10432. [44] R.D. Coteiro, F.S. Teruel, J. Ribeiro, A.R. de Andrade, J. Braz. Chem. Soc.
[18] L. Colmenares, H. Wang, Z. Jusys, L. Jiang, S. Yan, G.Q. Sun, R.J. Behm, 17 (2006) 771.
Electrochim. Acta 52 (2006) 221. [45] B.D. Cullity, Elements of X-Ray Diffraction, Addison-Wesley, San Fran-
[19] M. Goetz, H. Wendt, Electrochim. Acta 43 (1998) 3637. cisco, 1978, p. 102.
[20] A.O. Neto, J. Perez, W.T. Napporn, E.A. Ticianelli, E.R. Gonzalez, J. Braz. [46] Powder Diffraction File: 01-088-2342; 01-087-0646 and 00-035-1360,
Chem. Soc. 11 (2000) 39. Joint Committee on Powder Diffraction Standards, International Center
[21] C. Roth, M. Goetz, H. Fuess, J. Appl. Electrochem. 31 (2001) 793. for Diffraction Data, vol. PDF-2, Pennsylvania, USA, 2005.
[22] M.B. de Oliveira, L.P.R. Profeti, P. Olivi, Electrochem. Commun. 7 (2005) [47] F. Wu, H. Murakami, Y. Yamabe-Mitarai, H. Harada, H. Katayama, Y.
703. Yamamoto, Surf. Coat. Technol. 184 (2004) 24.
[23] S. Mylswamy, C.Y. Wang, R.S. Liu, J.-F. Lee, M.-J. Tang, J.-J. Lee, B.-J. [48] F. Colmati, E. Antolini, E.R. Gonzalez, Appl. Catal. B: Environ. 73 (2007)
Weng, Chem. Phys. Lett. 412 (2005) 444. 106.
[24] Y. Bai, J. Wu, J. Xi, J. Wang, W. Zhu, L. Chen, X. Qiu, Electrochem. [49] W. Zhou, Z. Zhou, S. Song, W. Li, G. Sun, P. Tsiakaras, Q. Xin, Appl.
Commun. 7 (2005) 1087. Catal. B: Environ. 46 (2003) 273.
[25] L.P.R. Profeti, F.C. Simões, P. Olivi, K.B. Kokoh, C. Coutanceau, J.-M. [50] C. Roman-Martinez, D. Cazorla-Amoros, H. Yamashita, S. de Miguel, O.A.
Léger, C. Lamy, J. Power Sources 158 (2006) 1195. Scelza, Langmuir 16 (2000) 1123.
[26] C. Lamy, S. Rousseau, E.M. Belgsir, C. Coutanceau, J.-M. Léger, Elec- [51] S. Tanaka, M. Umeda, H. Ojima, Y. Usui, O. Kimura, I. Uchida, J. Power
trochim. Acta 49 (2004) 3901. Sources 152 (2005) 34.
[27] L.X. Yang, R.G. Allen, K. Scott, P. Christenson, S. Roy, J. Power Sources [52] K. Juodkazis, J. Juodkazyte, T. Juodiene, V. Šukiene, I. Savickaja, Elec-
137 (2004) 257. trochim. Acta 51 (2006) 6159.
[28] M.H. Shao, R.R. Adzic, Electrochim. Acta 50 (2005) 2415. [53] M.L. Calegaro, H.B. Suffredini, S.A.S. Machado, L.A. Avaca, J. Power
[29] J.-M. Léger, S. Rousseau, C. Coutanceau, F. Hahn, C. Lamy, Electrochim. Sources 156 (2006) 300.
Acta 50 (2005) 5118. [54] M. Watanabe, S. Motoo, J. Electroanal. Chem. 60 (1975) 267.
[30] D.M. dos Anjos, K.B. Kokoh, J.-M. Léger, A.R. de Andrade, P. Olivi, G. [55] N.R. de Tacconi, R.O. Lezna, B. Beden, F. Hahn, C. Lamy, J. Electroanal.
Tremiliosi-Filho, J. Appl. Electrochem. 36 (2006) 1391. Chem. 379 (1994) 329.
[31] S. Rousseau, C. Coutanceau, C. Lamy, J.-M. Léger, J. Power Sources 158 [56] F. Vigier, C. Coutanceau, A. Perrard, E.M. Belgsir, C. Lamy, J. Appl.
(2006) 18. Electrochem. 34 (2004) 439.
[32] K. Taneda, Y. Yamazaki, Electrochim. Acta 52 (2006) 1627. [57] L. Demarconnay, S. Brimaud, C. Coutanceau, J.-M. Léger, J. Electroanal.
[33] Z.D. Wei, L.L. Li, Y.H. Luo, C. Yan, C.X. Sun, G.Z. Yin, P.K. Shen, J. Chem. 601 (2007) 169.
Phys. Chem. B 110 (2006) 26055. [58] X.H. Xia, H.-D. Liess, T. Iwasita, J. Electroanal. Chem. 437 (1997)
[34] Z. Liu, B. Guo, L. Hong, T.H. Lim, Electrochem. Commun. 8 (2006) 83. 233.
[35] A. Chen, D.J. La Russa, B. Miller, Langmuir 20 (2004) 9695. [59] G.A. Somorjai, J. Phys. Chem. 94 (1990) 1013.
[36] V.I. Birss, H. Andreas, I. Serebrennikova, H. Elzanowska, Electrochem. [60] G. Chen, D.A. Delafuente, S. Sarangapani, T.E. Mallouk, Catal. Today 67
Solid State Lett. 2 (1999) 326. (2001) 341.
[37] H. Tsaprailis, V.I. Birss, Electrochem. Solid State Lett. 7 (2004) A348. [61] R.B. Woodward, R. Hoffman, J. Am. Chem. Soc. 87 (1965) 395.
[38] S. Liao, K.-A. Holmes, H. Tsaprailis, V.I. Birss, J. Am. Chem. Soc. 128 [62] B. Gurau, R. Viswanathan, R. Liu, T.J. Lafrenz, K.L. Ley, E.S. Smotkin,
(2006) 3504. E. Reddington, A. Sapienza, B.C. Chan, T.E. Mallouk, S. Sarangapani, J.
[39] P. Sivakumar, V. Tricoli, Electrochem. Solid State Lett. 9 (2006) A167. Phys. Chem. B 102 (1998) 9997.
[40] M.P. Pechini, N. Adams, US Patent 3,330,697:1 (1967). [63] E. Antolini, F. Colmati, E.R. Gonzalez, Electrochem. Commun. 9 (2007)
[41] J.C. Forti, P. Olivi, A.R. de Andrade, Electrochim. Acta 47 (2001) 913. 398.
[42] A.J. Terezo, E.C. Pereira, Mater. Lett. 53 (2002) 339. [64] F. Colmati, E. Antolini, E.R. Gonzalez, J. Electrochem. Soc. 154 (2007)
[43] V. Mehta, J.S. Cooper, J. Power Sources 114 (2003) 32. B39.