The activity of ALD-prepared PtCo catalysts for ethanol oxidation in alkaline media

Article abstract of DOI:10.1016/j.jcat.2013.08.027

Controlled bimetallic catalyst materials can be obtained using atomic layer deposition (ALD) method. In this paper, this method was applied to prepare Pt, PtCo, and PtCoPt nanoparticle catalysts on carbon support. Their activity for ethanol oxidation was studied by various electrochemical methods and the dependency of the reaction on temperature and mass transfer was evaluated. In addition, FTIR analysis was performed to confirm the reaction products. The results showed that bimetallic PtCo enhances ethanol oxidation by increasing the amount of CO₂ produced; however, at elevated temperatures, the bimetallic PtCo catalyst lacked stability. Nevertheless, the ALD-prepared PtCoPt catalyst showed similar reaction mechanism as PtCo catalyst, but possess increased stability and superior performance for ethanol oxidation at elevated temperatures.

Full text of DOI:10.1016/j.jcat.2013.08.027

Journal of Catalysis 309 (2014) 38–48
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
The activity of ALD-prepared PtCo catalysts for ethanol oxidation in
alkaline media
b
c
a
d
c
b
A. Santasalo-Aarnio a, , E. Sairanen , R.M. Arán-Ais , M.C. Figueiredo , J. Hua , J.M. Feliu , J. Lehtonen ,
⇑
R. Karinen b, T. Kallio a
a Research Group of Fuel Cells, Aalto University, P.O. Box 16100, 00076 Aalto, Finland
b Industrial Chemistry Research Group, School of Chemical Technology, Aalto University, P.O. Box 16100, 00076 Aalto, Finland
c Instituto de Electroquimica, Universidad de Alicante, Apartado 99, 03080 Alicante, Spain
d NanoMaterials Group, Department of Applied Physics, Aalto University, Espoo, Finland
a r t i c l e i n f o
a b s t r a c t
Article history:
Controlled bimetallic catalyst materials can be obtained using atomic layer deposition (ALD) method. In
this paper, this method was applied to prepare Pt, PtCo, and PtCoPt nanoparticle catalysts on carbon sup-
port. Their activity for ethanol oxidation was studied by various electrochemical methods and the depen-
dency of the reaction on temperature and mass transfer was evaluated. In addition, FTIR analysis was
performed to confirm the reaction products. The results showed that bimetallic PtCo enhances ethanol
oxidation by increasing the amount of CO2 produced; however, at elevated temperatures, the bimetallic
PtCo catalyst lacked stability. Nevertheless, the ALD-prepared PtCoPt catalyst showed similar reaction
mechanism as PtCo catalyst, but possess increased stability and superior performance for ethanol oxida-
tion at elevated temperatures.
Received 14 June 2013
Revised 20 August 2013
Accepted 20 August 2013
Keywords:
Ethanol oxidation
ALD
PtCo catalyst
Temperature dependency
DEFC
Ó 2013 Elsevier Inc. All rights reserved.
Alkaline media
FTIR
Arrhenius plot
1. Introduction
therefore the power density obtained from the cell is moderate.
Secondly, only diluted alcohol solutions can be used due to the se-
Polymer electrolyte fuel cells (PEFC) have been studied inten-
sively for decades. However, real commercial breakthrough has
not been attained due to the expensive cell components including
the catalyst materials and the membrane used as an electrolyte.
For low power demand applications such as laptops, alcohols are
interesting fuels due to their high energy density and easy storage
compared to the gaseous hydrogen. From organic alcohols, mostly
methanol is currently used in fuel cell applications. However, it is
toxic, easily flammable and as a small molecule has a tendency to
permeate the electrolyte membrane reducing the overall effi-
ciency. Conversely, ethanol is non-toxic, can be produced from bio-
mass, and has both higher boiling point and theoretical energy
density [1] compared to methanol. In addition, as larger molecule
ethanol has reduced crossover rate through both proton and anion
exchange, membranes [2,3] making it an attractive fuel for direct
ethanol fuel cells (DEFC).
vere crossover of the fuel to the cathode leading mixed potentials
lowering the total performance. It has been known for decades that
alcohol oxidation kinetics is enhanced in alkaline media compared
to acidic [4–7], and therefore, lower amount of expensive noble
metal catalyst is needed reducing the cost of the total system
remarkably. The lack of anion exchange membrane materials has
limited the research in alkaline media; however, in recent years,
these membranes have been developed and demonstrated success-
fully in fuel cells [8–13].
The oxidation of ethanol can occur via a dual path mechanism
involving production of 4 or 12 electrons [5,14–16] that is illus-
trated in Scheme 1. According to the 4 eꢀ path, adsorbed ethanol
molecule is oxidized to acetaldehyde producing two protons and
electrons and can be further oxidized to acetic acid producing
two more protons and electrons. With this, 4 eꢀ path ethanol mol-
ecule is only partly oxidized and the final product is an organic acid
that is a problematic in fuel cell conditions because it effects the
local pH in the electrode layer. As can be seen in Scheme 1, the
other path includes the cleavage of CAC bond from either adsorbed
ethanol or acetaldehyde forming CO and CHx fragments which can
both further be oxidized to CO2. This path produces 12 protons and
Nevertheless, with all organic fuels, some challenges are faced:
Firstly, reaction kinetics of alcohol oxidation is sluggish and
⇑
Corresponding author.
E-mail address: annukka.aarnio@iki.fi (A. Santasalo-Aarnio).
0021-9517/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2013.08.027

A. Santasalo-Aarnio et al. / Journal of Catalysis 309 (2014) 38–48
39
surfaces, a FTIR analysis of the reaction intermediates and products
was performed further facilitating the analysis of the electrochem-
ical data. In addition, the activity for ethanol oxidation was evalu-
ated as a function of operation temperature providing important
knowledge for the catalyst activity in real fuel cell systems.
Scheme 1. The oxidation paths of ethanol.
2. Experimental
2.1. Catalyst preparation
electrons implying that from the same molecule, two to three
times higher current can be generated increasing significantly
the total efficiency of the system.
ALD catalysts were prepared in a commercial flow-type ALD
reactor (F-120, ASM Microchemistry) applying pressure of 0.5–
1 kPa. Carbon black was used as catalyst support (Vulcan XC72R,
Cabot GR-3875) and platinum acetylacetonate, Pt(acac)2 (Volatec
Oy) as well as cobalt acetyl acetonate, Co(acac)3 (Merck, 98%) were
applied as metal precursors. The catalyst preparation cycle con-
sisted of three steps: first, the catalyst support was dried in N2 flow
(AGA, 99.999%) at 180 °C for 5 h. Second, the metal precursor was
evaporated and fed on the support using N2 carrier gas at 180 °C for
6 h in reduced pressure. Third, the unreacted precursor was
flushed from the support with N2 and reactor was cooled down
to room temperature under N2 flow. Reaction cycles were repeated
to prepare bimetallic catalysts with additional Co or Pt cycles. A
commercial 20 wt% Pt on a carbon support (Alfa Aesar) was used
as a reference material.
The prepared catalyst was characterized by atomic emission
spectroscopy (AES) (ICP-AES Varian Liberty series II) to determine
the total metal loadings. A double-aberration corrected JEOL
2200FS (JEOL, Japan) high-resolution transmission electron micro-
scope (HRTEM) was used to characterize the particle size distribu-
tion. The microscope is equipped with a field emission gun (FEG)
operated at 200 kV and an energy dispersive X-ray (EDX) spec-
trometer for elemental analysis. A Gatan 4kx4k UltraScan 4000
CCD camera was employed for digital recording of HRTEM images.
Gatan Digitalmicrograph software was used for camera control and
image processing. The mean particle size was evaluated by analyz-
ing 150–250 particles.
Platinum is generally considered as the best monometallic cat-
alyst for the oxidation of small organic molecules due to its capa-
bility to catalyze CAH bond rapture at comparatively low
overpotential. Nevertheless, ethanol oxidation involves other reac-
tion steps such as CAC bond cleavage and CO or acetaldehyde oxi-
dation, and therefore, for total oxidation, high overpotentials are
required. On pure Pt catalyst, the ethanol oxidation follows mostly
the 4 eꢀ path [17] and therefore the demand for bimetallic catalyst
is evident. For instance, PtRu has been found to be an excellent
catalyst for methanol oxidation [18], because Ru allows water mol-
ecules to break forming OH groups on its surface at low overpoten-
tials that is needed for the total oxidation of methanol [19].
Unfortunately, this catalyst has not shown ability to break the
CAC bond and is, therefore, considered to be a good catalyst only
for molecules with single carbon. Despite the vast effort, a catalyst
material that could oxidize ethanol efficiently to CO2 in acidic
media has not been discovered. However, in alkaline media, the
reaction mechanisms are different, and in addition, many oxides
and alloys are stable at wider potential range. PtCo catalyst has
traditionally been studied for oxygen reduction reaction both in
acidic and alkaline media [20–22]. Recently, few studies with this
catalyst have been performed for alcohol oxidation [23–28] and
this bimetallic catalyst has shown decrease in onset potential
and increase in oxidation currents of alcohols especially at low
potentials. Nevertheless, most of these studies are made in acidic
media where the dissolution of Co is likely at 0.4–0.8 V [27] which
are relevant potentials for DEFC anode operation.
Unfortunately, the preparation of bimetallic catalyst is chal-
lenging: to ensure high catalytic activity, the metal particles should
be in near vicinity of each other and not as separate islands. With
traditional impregnation method, most commonly involving metal
salt reduction on carbon powder in a solvent, the metal distribu-
tion control is poor resulting in increased noble metal catalyst
loading that prevents the commercialization of the DEFC technol-
ogy. Conversely, with atomic layer deposition (ALD), the metal pre-
cursors are vaporized to gas phase and flowed through the porous
carbon support. This support is hydrophobic and therefore most of
the atoms rather attach to hydrophilic defects on the support.
When preparing bimetallic catalysts by the ALD method, a single
metal precursor forms metallic nanoparticles on support with the
first ALD cycles. After these cycles, the second metal is introduced
attaching to the first metal or to the carbon support. The nanopar-
ticle size and loading can be controlled by varying the ALD cycle
variables [29]. Overall, with the ALD method, very uniformly dis-
tributed monometallic [30–33] and bimetallic [34–36] catalyst
has been reported possessing high catalytic activity in comparison
with catalysts produced with traditional methods. However, no
ALD-prepared bimetallic catalyst for ethanol oxidation has been
previously reported.
2.2. Electrochemical measurements
The catalyst inks were prepared with 5 mg of the catalyst pow-
der that was dissolved in 200
following by the addition of 6
l
l
l n-methyl-2-pyrrolidone (Merck)
l of fumionÒ FAA-3 ionomer disper-
sion (FumaTech). The ink was left under mixing with a magnetic
stirrer for 2 h and sonicated with an ultrasonic bath for 40 min.
The surface of glassy carbon electrode was cleaned mechanically
by alumina slurry and ultrasonic bath. 3 ll droplet of the catalyst
ink was spread on the electrode and was dried for 1 h at 60 °C in
a vacuum oven. The electrochemical activity of the catalysts was
studied in a three electrode electrochemical cell presented in
Fig. 1. The previously mentioned glassy carbon electrode was used
as a working electrode (WE) attached with a rotating device (Pine
instruments), a large platinum wire was used as the counter elec-
trode (CE), and a reversible hydrogen electrode (RHE) was used as a
reference electrode (RE). Freshly prepared 0.1 M NaOH solution
(Merck) was used as electrolyte, and all the solutions were made
using MQ water (Millipore, 0.04 l
S cmꢀ1).
The surface potential of the working electrode was controlled
during the experiments to avoid unwanted anion adsorption, and
the temperature of the electrochemical cell was adjusted with a
water jacket controlled by CH cryostat/thermostat (Haake). Prior
to the experiments, the electrolyte was deaerated by purging with
N2 and the temperature was stabilized to 20 °C. During the mea-
surements, the gas stream was directed above the solution surface.
The experiments were performed by a potentiostat/galvanostat
In this study, ALD-prepared Pt and PtCo catalytic materials are
introduced and their activity toward ethanol oxidation in alkaline
media is studied with electrochemical methods including cyclic
voltammetry as well as long- and short potential step experiments.
To understand the reaction mechanism occurring on the catalyst

40
A. Santasalo-Aarnio et al. / Journal of Catalysis 309 (2014) 38–48
Fig. 1. The electrochemical cell with three electrodes: working electrode (WE), counter electrode, (CE) and reference electrode (RE).
PGSTAT100 Autolab system. Firstly,
a cyclic voltammogram
(50 mV sꢀ1) was obtained to ensure that the electrode preparation
has succeeded. The catalyst surface cleaning was done with CO
adsorption/oxidation method to avoid cycling to high potentials
that could cause Co dissolution and Pt particle agglomeration.
The electrolyte was purged with CO gas (Aga, 3.7), while the poten-
tial of the working electrode was cycled between 0.1 and 0.35 V to
observe the blocking of the Pt site. When no further decrease in
current was observed, the potential was fixed to 0.1 V and the
gas was changed to N2 and purged for 20 min to remove CO from
the electrolyte. Subsequently, a cyclic voltammogram with CO oxi-
dation was obtained with 10 mV sꢀ1. After these cyclic voltammo-
grams of the clean catalyst surface were obtained at 20 mV sꢀ1. All
potentials were scaled to reversible hydrogen electrode (RHE)
scale, and all currents were normalized by the electrochemical ac-
tive surface area obtained from surface area integrated from the
first and last scans during the CO adsorption in the potential range
from 0.1 to 0.35 V.
After the measurements in the pure electrolyte, ethanol (Altia,
p.a.) was added to the cell to obtain 1 mol dmꢀ3 ethanol solution.
To prevent the changes in the concentration, the gas inlet was first
directed into a bubbler bottle containing the same solution as the
cell and alcohol-saturated gas was directed to the experimental
cell. Cyclic voltammograms were recorded at 10 mV sꢀ1 rate with
0 and 1800 rpm rotation at various temperatures (0, 20, 40, and
60 °C). To further study the catalyst poisoning, chronoamperomet-
ric measurements were performed. These phenomena were stud-
ied with long (30 min) constant potential experiments at 0.5 V
potential that still is relevant for DEFC anodes [37] and with short
experiments of 5 min at various potentials.
where R and R0 are the reflectance corresponding to the single beam
spectra obtained at the sample and reference potentials, respec-
tively. All the spectroelectrochemical experiments were performed
at room temperature, with a RHE and a platinum wire used as
reference and counter electrodes, respectively. The attenuated total
reflection (ATR) spectra were collected using a SeZn hemicylindrical
window with an incident angle of 60°. One hundred scans were
collected with a resolution of 8 cmꢀ1. The reference spectra were
acquired with a 0.1 M NaOH solution.
3. Results and discussion
3.1. Physical characterization
ALD method was used to prepare a set of mono- and bimetallic
nanoparticle catalysts on a carbon support by varying the order of
metal depositions to obtain Pt, PtCo, and PtCoPt catalyst where
the catalyst name indicates the deposition order of the elements.
The metal contents of all the ALD-prepared catalyst were deter-
mined with AES, and these values are presented in Table 1 together
with the mean particle size obtained from HRTEM images (Fig. 2).
The obtained noble metal loadings are low 5–15 wt% as desired
for DEFC applications. The Pt (ALD) and PtCo (ALD) catalysts origi-
nate from the same catalyst batch, and they were prepared so that
after the ALD cycles of Pt, part of the catalyst material was removed
from the reactor. Subsequently, half of the Pt (ALD) material was re-
placed to the reactor and Co cycles were performed. For this reason,
Pt (ALD) and PtCo (ALD) catalysts have the same Pt/C ratio. It has
been previously shown that a noble metal layer on top of a less no-
ble metal decreases dissolution of the latter [27], and therefore,
three ALD depositions cycles in the order of Pt, Co, and Pt were
applied for one catalyst sample. Due to the sample amount require-
ments, PtCoPt (ALD) catalyst was prepared from a different batch
2.3. Reaction product analysis by FTIR
As reported elsewhere [38,39], spectroelectrochemical experi-
ments were performed with a Nicolet Magna 850 spectrometer
equipped with a MCT detector. The spectroelectrochemical cell
was provided with a prismatic CaF2 window beveled at 60°. Spec- Table 1
The catalysts metal loadings and the mean particle sizes analyzed by AES and TEM,
respectively.
tra shown are composed of 100 interferograms collected with a
resolution of 8 cmꢀ1 and p-polarized light. They are presented as
absorbance, according to
Pt (wt%)
Co (wt%)
Co/Pt ratio
Particle size (nm)
Pt (Com)
Pt (ALD)
PtCo (ALD)
PtCoPt (ALD)
20.00
13.97
13.78
8.06
–
–
1.38
0.79
–
–
0.10
0.10
2.30
2.00
2.64
1.68
R
R0
A ¼ ꢀ log
ð1Þ

A. Santasalo-Aarnio et al. / Journal of Catalysis 309 (2014) 38–48
41
Fig. 2. HRTEM images of the ALD-prepared Pt (A), PtCo (B), and PtCoPt (C) catalysts.
are considered as baselines for Pt (Com) and Pt (ALD).
With bimetallic catalyst, the hydrogen adsorption/desorption is
less defined and therefore only one clear hydrogen desorption peak
is observed in Fig. 3 with the positive-going scan around potential
of 0.25 V. In addition, the current for the double layer region in-
creased and a peak above 0.65 V appears, and these both phenom-
ena are associated with OHꢀ adsorption on Co or to CoOx
formation, whereas the corresponding CoOx reduction is observed
in the negative-going scan. Similar features have been reported
for PtCo/C catalyst in KOH electrolyte [20,23,44,45]. When an addi-
tional layer of Pt is added the peak at high potential and the CoOx
reduction peak are shifted to less positive potentials implying that
oxygenated species adsorb in a different way on this catalyst. The
electrochemical durability of various ALD catalysts including Pt
and Co was studied in acidic media elsewhere [40], and the results
showed no indication that the Co would had been lost from the cat-
alyst when the catalyst was cycled to 0.8 V. In alkaline media,
bimetallic catalysts are often more stable confirming that these
catalyst could be used also in DEFC anode conditions.
and therefore Pt loading of this catalyst is slightly lower. Neverthe-
less, it has been ensured that the Co/Pt ratio in this catalyst remains
the same as in the firstly made PtCo (ALD) catalyst (Table 1).
All of the particles are relatively small and in the same size
range (Table 1) confirming that the change in activity does not re-
sult only from difference in the particle size. In addition, Fig. 2
shows that the particles are well dispersed throughout the whole
carbon support. For bimetallic catalysts, EDX analysis on various
locations was also performed to verify that both metals are distrib-
uted evenly on the support material. The analysis confirmed that
the Co/Pt ratio was consistent with the AES results reported in
Table 1. For the ALD catalysts, more detailed physical characteriza-
tion is presented elsewhere [40].
3.2. Electrochemical characterization
The cyclic voltammograms of the cleaned catalyst surfaces in
alkaline media are presented in Fig. 3, and it can be seen that both
monometallic Pt catalysts have similar behavior: with the positive-
going scan, there is one clear peak around 0.27 V potential as well
as a shoulder around 0.36 V potential. They correspond to hydro-
gen desorption on Pt(110) and Pt(100), respectively, observed
more clearly on Pt nanoparticles without carbon support or iono-
mer [41]. Above 0.45 V, the OHꢀ adsorption on Pt surface initiates
followed by the formation of irreversible PtOx layer on both cata-
lyst [42]. Nevertheless, with the Pt (ALD) catalyst, both hydrogen
adsorption/desorption peaks and double layer current are higher
in comparison with the commercial catalyst due to the lower metal
loading [43]. It should be noted that the profiles are coincident if
the current values at the double layer region at potential of 0.5 V
3.3. Ethanol oxidation
The activity of the ALD-prepared catalysts was studied for eth-
anol oxidation in alkaline media first with cyclic voltammetry, and
the results obtained at 20 °C are presented in Fig. 4, for clarity only
the positive-going scans are shown. Fig. 4A and B presents the
scans for still and rotating (1800 rpm) working electrodes, respec-
tively, to facilitate the analysis of diffusion effects. The onset
potential of ethanol oxidation was determined in both cases (with
0 and 1800 rpm rotation) by detecting the potential where the cur-
rent was higher than the current caused by the hydrogen desorp-
tion in Fig.
3
(0.08 mA cmꢀ2), and the obtained values are
presented in Table 2. In addition, currents for ethanol oxidation
at the potentials of 0.5 and 0.8 V in Fig. 4 are also summarized in
Table 2.
It can be seen from the Fig. 4A that Pt (Com) catalyst has the
highest onset potential (0.56 V) similar to reported for Pt disk
and Pt nanoparticle catalyst in alkaline media [23,46,47]. The reac-
tion onset is reduced by 40 mV with the ALD-prepared Pt (Table 2),
indicating that the higher order of the catalyst surface [40] in-
creases the catalyst activity. In addition, the current produced in
the whole potential range is higher with the Pt (ALD) catalyst than
with Pt (Com). Nevertheless, the differences diminish when the
working electrode is rotated (Fig. 4B), suggesting that the reactant
diffusion to the stationary Pt (ALD) particles is more facile in com-
parison with the Pt (Com) particles. Similar observation has been
made earlier with the ALD-prepared Pd catalyst [30].
It is evident on the basis of Table 2 and Fig. 4 that the onset po-
tential of ethanol oxidation is clearly shifted to less-positive poten-
tials by introduction of Co into the catalyst as also reported
Fig. 3. Cyclic voltammograms of the cleaned catalyst surfaces in 0.1 mol dmꢀ3
NaOH electrolyte with 50 mV sꢀ1 scan rate.

42
A. Santasalo-Aarnio et al. / Journal of Catalysis 309 (2014) 38–48
Fig. 4. Positive-going scans of the cyclic voltammetry for ethanol oxidation in solution of 1 mol dmꢀ3 ethanol and 0.1 mol dmꢀ3 NaOH (A) without rotation of the working
electrode and (B) with 1800 rpm rotating, 10 mV sꢀ1 scan rate at 20 °C.
Table 2
The onset potential of ethanol oxidation and current values at 0.5 V potential, 20 °C.
EOnset
0 rpm (V)
EOnset
1800 rpm (V)
j at 0.5 V
1800 rpm
(mA cmꢀ2
j at 0.8 V
1800 rpm
(mA cmꢀ2
)
)
Pt (Com)
Pt (ALD)
PtCo (ALD)
PtCoPt (ALD)
0.56
0.52
0.40
0.43
0.52
0.49
0.39
0.42
0.06
0.10
0.17
0.17
0.91
0.81
0.79
1.11
previously [23], and this shift is independent of the diffusion
because it can be observed both with and without rotation. When
another Pt deposition cycle was performed on the PtCo catalyst,
the onset shifted toward more positive potentials; thus, it is still
clearly lower in comparison with pure Pt catalysts. As highlighted
in Table 2, the current density obtained on the bimetallic catalysts
at fuel cell relevant potential (0.5 V) is almost 3 times higher in
comparison with Pt (Com). This increase in activity with bimetallic
catalyst is in the same range as reported for this reaction on other
bimetallic catalyst at this potential (2–5 times increase)
[23,37,48,49]. At higher potentials, the PtCo (ALD) catalyst has low-
er performance than the monometallic Pt catalysts. Nevertheless,
when additional Pt layer was placed on the top of the PtCo (ALD)
catalyst, the performance remained superior in comparison with
both Pt catalysts. This implies that the last layer of Pt atoms on
the PtCoPt (ALD) catalyst is less prone to poisoning and simulta-
neously protecting the Co atoms still contributing with bi-func-
tional mechanism. With both PtCo catalysts, an oxidation peak
curve is visible below 0.8 V without rotation (Fig. 4A) indicating
that the diffusion of reactants or reaction products starts to control
at these potentials, though this is avoided with the rotation of the
working electrode (Fig. 4B) or with a constant reactant flow at the
DEFC anode.
Avoiding poisoning is essential for a catalytic material to be
used in fuel cell applications. Ethanol oxidation occurs on the an-
ode of a DEFC, and therefore, the potential of this electrode should
be as low as possible to increase the total cell voltage. At a high
voltage, the power obtained from the cell is not optimal and there-
fore a catalyst material that would oxidize efficiently ethanol at
low potentials is of interest. To study the suitability of these cata-
lysts for ethanol oxidation, potential step experiments of 30 min
were performed at a potential of 0.5 V, relevant to a DEFC anode.
The results obtained at 20 °C are presented in Fig. 5.
Fig. 5. Chronoamperometric curves for ethanol oxidation at 0.5 V potential in
solution of 1 mol dmꢀ3 ethanol and 0.1 M NaOH, with rotating working electrode
(1800 rpm) at 20 °C.
mechanism at 0.5 V potential is same; however, the current on Pt
(ALD) was clearly higher throughout the experiment due to the
more crystalline structure obtained by the ALD method [40]. With
the bimetallic catalysts, the initial current was significantly higher,
though the catalyst poisoning was more rapid than with the mono-
metallic catalysts which may indicate that the reaction yielding
poisoning species is enhanced. On PtCo (ALD), surface poisoning
is less severe probably due to higher amount of Co atoms on the
catalyst surface providing an oxygen source for the reaction inter-
mediates inhibiting catalyst surface poisoning. The values for cur-
rent densities obtained on PtCo (ALD) after 30 min of constant
potential are over twice higher than on Pt (Com) that is one of
the highest difference reported on bi- or trimetallic catalysts for
this reaction [23,37,46,48]. Overall, the behavior of both bimetallic
catalysts at this time scale differ in comparison with the pure Pt
catalysts, and therefore, more detailed analysis of the reaction
mechanism was performed with FTIR spectroscopy.
3.4. Reaction product analysis
In order to identify the species involved in the catalytic oxida-
tion of ethanol on the ALD catalysts, IR experiments were per-
formed. In the following spectra, the positive bands correspond to
the products formed during ethanol oxidation, while negative
bands arise due to the consumption of species present at the
reference potential. The contact of the electrodes with the ethanol
solution was performed at controlled potential (0.1 V) where,
apparently, no adsorption or reaction process occurs [50]. This po-
As can be seen in Fig. 5, the current behavior and current deg-
radation of both Pt catalysts is similar, suggesting that the reaction

A. Santasalo-Aarnio et al. / Journal of Catalysis 309 (2014) 38–48
43
tential was maintained until the electrode was pressed against the
CaF2 window. After collecting the reference spectrum, the potential
was stepped progressively to higher potentials up to 0.8 V and then
decreased stepwise again back to 0.1 V.
figurations (used in FTIR), the occurrence of local pH change is not
surprising and would lead to the presence of CO2 near the elec-
trode surface. This fact was also described for ethanol oxidation
in alkaline media on Pd surfaces [53].
It is known that some of the possible products of ethanol oxida-
tion can have vibrational band in a close range of wavenumbers.
Therefore, the spectra obtained with an ATR configuration for
acetate and acetaldehyde in a NaOH solution are presented in
Fig. 6. This IR analysis was performed to determine the vibration
frequencies of these species in alkaline solution. The NaOH solu-
tion was used as background meaning that the remaining bands
will result from acetaldehyde or acetic acid in this media. From
Fig. 6, it can be clearly observed that in contrast to acidic media
[16], in alkaline media, both species have different vibrational
bands. Acetate presents two major bands at 1553 and 1413 cmꢀ1
and a smaller one at 1357 cmꢀ1. For acetaldehyde, only one band
at 1715 cmꢀ1 is observed. These results allow a clear distinction
between the two products in the spectra obtained from potential
controlled experiments.
The results obtained under controlled potential conditions for
the ALD-prepared Pt and PtCo catalysts are presented in Fig. 7
where the sample potential of each spectrum is labeled, respec-
tively. Several bands can be observed in the spectra represented
in Fig. 7: a broad band between 2500 and 3000 cmꢀ1 was ob-
served with all the studied surfaces. This band corresponds to
the CAH region (2700–3000 cmꢀ1) and overlaps with other wide
band denoting OH from carboxyl group between 2500 and
3000 cmꢀ1. This group is expected to exist in ethanol as well as
in some of the most probable final products from its oxidation,
like acetic acid [51]. These bands become more intense with the
increase in the potential due to the higher oxidation rate. Other
observed bands are the following: at 2341 cmꢀ1 the CAO asym-
metric stretching from CO2 [51,52]; at 1715 cmꢀ1 the C@O
stretching of carbonyl groups from acetaldehyde (Fig. 7); the
bands at 1553 and 1413 cmꢀ1 from OACAO symmetric and asym-
metry stretching (respectively) from acetate (Fig. 7) and at
1357 cmꢀ1 the CH3 bending from acetate.
All the surfaces have common bands at 1715 cmꢀ1 from acetal-
dehyde and between 1560 and 1350 cmꢀ1 due to acetate suggest-
ing that both are reaction products in the studied catalysts
independently of the presence of Co on the surface. However, a
main difference can be observed between Pt and the two catalysts
containing Co in the region correspondent to CO2 production. It is
important to notice that CO is not observed on any of the surfaces
in the potential range studied. This fact does not mean that CO is
absent under these conditions but more likely that the CO oxida-
tion at these potentials is faster than its production and the corre-
sponding IR bands stay below the detection limit. Moreover, it is
well known that surface steps/defects enhance the rate of the CO
oxidation on platinum surfaces [54,15,55]. On Pt nanoparticles,
the CO oxidation to CO2 is expected to be fast due to the small size
of the crystalline sites and because the catalyst surface is formed in
majority by defects and steps [54]. The absence of CO in the spectra
obtained for ethanol oxidation on Pt-supported nanoparticles was
also reported previously by our group for other catalysts [40].
The CO2 band was only observed for the catalysts containing Co
suggesting that this metal plays a key role in the splitting of the
CAC bond that would lead to a more efficient total oxidation of
ethanol. In addition, it is important to remark that with PtCo
(ALD) (when Co is present), ethanol oxidation reaction starts ear-
lier as it was observed in the cyclic voltammetry (Fig. 4) by the
shift of the onset potential toward lower values and by the bands
from acetate and acetaldehyde that appear at potentials as low
as 0.4 V for this catalyst. These bands are also observed with the
PtCoPt (ALD) catalyst, but they are less intense and are totally ab-
sent with the pure Pt catalyst. These results indicate that on pure
Pt surface, the oxidation current observed is mainly contributed
from the 4 eꢀ path, while in the case of PtCo catalysts also the
12 eꢀ path is involved. Nevertheless, even on pure Pt catalyst,
the 12 eꢀ path should not be excluded because when small amount
of CO2 is produced it will be hydrolyzed to carbonate that cannot
be distinguish by IR because overlapping of the carbonate
(1390 cmꢀ1) and acetate bands. The FTIR results are summarized
in Fig. 8 that plots the band area for CO2 (2341 cmꢀ1), acetaldehyde
(1715 cmꢀ1), and acetate (1553 cmꢀ1) against the applied potential
for the three catalysts. The areas were obtained by fitting and
deconvolution of the spectra shown in Fig. 7.
The observation of a band corresponding to CO2 in alkaline
media is unexpected because in this media the produced CO2
would hydrolyze forming carbonate. However, during the ethanol
oxidation, H+ is produced neutralizing the NaOH species given rise
to changes in the local pH. In the particular case of thin layer con-
On the basis of Fig. 8, it is definitely concluded that with both
the bimetallic PtCo catalysts, CO2 is obtained as a final product
from ethanol oxidation. The fact that the amount of CO2 decreases
with additional Pt cycle supports the theory that some of this Pt is
deposited as a second layer on the pre-existent PtCo particles. This
surface Pt could be poisoned, thus leading to a decrease in the cat-
alyst activity. However, it should be remarked that this catalyst is
still more active than pure Pt. The higher activity can also be ob-
served by the higher integrated area for both acetaldehyde and
acetate products at PtCoPt (ALD), and these results being in agree-
ment with data from cyclic voltammetry (Fig. 4). At high poten-
tials, the current on the PtCo (ALD) catalyst is lower than on
catalyst with additional Pt cycle on the surface. It is also important
to mention that acetaldehyde production is higher for PtCo (ALD)
in the positive-going scan and it decreases very fast in the nega-
tive-going scan although the currents in the CV are still high. This
decrease can be tentatively explained by the high CO2 amount for
this catalyst, meaning that probably acetaldehyde is being further
oxidized to CO2 at the negative-going scan.
Fig. 6. ATR spectra of 0.1 mol dmꢀ3 acetate and 0.1 mol dmꢀ3 acetaldehyde in
0.1 mol dmꢀ3 NaOH, 100 scans and 8 cmꢀ1, p-polarized light. Reference spectra:
0.1 mol dmꢀ3 NaOH.

44
A. Santasalo-Aarnio et al. / Journal of Catalysis 309 (2014) 38–48
Fig. 7. Spectra obtained for ethanol oxidation on Pt (ALD), PtCo (ALD), and PtCoPt (ALD) catalysts in solution of 1 mol dmꢀ3 ethanol and 0.1 mol dmꢀ3 NaOH with 100 scans
and 8 cmꢀ1 of resolution. Potential of each spectrum is indicated in the Figure. Reference spectra were obtained at 0.1 V.
Fig. 8. Area of the bands for CO2 (2341 cmꢀ1), acetaldehyde (1715 cmꢀ1), and acetate (1553 cmꢀ1) for Pt (ALD), PtCo (ALD), and PtCoPt (ALD). Colored and open symbols refer
to values obtained with positive- and negative-going scans, respectively.
3.5. Temperature dependency
tive-going scans. This hysteresis mainly occurs when the catalyst
surface is blocked by poisoning reaction intermediates that are
only oxidized further at the highest potential of the cyclic voltam-
metry enabling the adsorption and oxidation of new reactants. In
on the catalysts is presented at this temperature range in Fig. 9.
the temperature-dependent experiments, this is especially ob-
The Fig. 9 shows that the ethanol oxidation is enhanced with
For portable DEFC applications, the operation temperature
could vary from 0 to 60 °C and the activity for ethanol oxidation
served on all the catalyst with cycles performed at 0 °C (Fig. 9)
due to the slow reaction kinetics. However, at this temperature,
the hysteresis is lower with the bimetallic catalyst, implying that
Co provides the opportunity for the intermediates to be oxidized
further at lower potential and thus decreasing the poisoning. Nev-
ertheless, with the bimetallic catalysts, the hysteresis is observed
at elevated temperatures where very high current densities are ob-
served. As it was seen in the chronoamperometric experiments
(Fig. 5), the bimetallic surfaces suffer more from severe poisoning
increasing temperature as expected. However, the shape of the vol-
tammogram at this potential region is not changed by tempera-
ture. The most important observation from Fig. 9 is that the Pt
(Com) catalyst has the lowest variation with the temperature but
with the ALD-prepared catalysts, the current increase with the
temperature seems to be more significant.
For ethanol oxidation, the currents obtained at the negative-
going scans are often higher than the ones obtained with the posi-

A. Santasalo-Aarnio et al. / Journal of Catalysis 309 (2014) 38–48
45
Fig. 9. Cyclic voltammograms of ethanol oxidation at various temperatures in solution of 1 mol dmꢀ3 ethanol and 0.1 mol dmꢀ3 NaOH. 10 mV sꢀ1 scan rate with rotating
working electrode (1800 rpm).
at all the temperatures supporting the previous assumption that
and the increase in the reaction rate with the temperature leads to
increased poisoning of the surface sites further increasing hystere-
sis. To compare the changes in the onset potentials with the tem-
perature, the potential values at the current of 0.08 mA cmꢀ2 (from
Fig. 9) are summarized in Table 3.
there are some Pt atoms covering the active surface sites in the
PtCoPt (ALD) catalyst. These results strongly support the claim that
with the ALD method, a preparation of structured bimetallic cata-
lysts having Pt on top of some Co sites can be obtained and that
these catalysts have superior performance compared to the tradi-
tionally prepared bimetallic catalysts. The poisoning behavior of
these catalysts has been further studied with short potentials step
experiments at various potentials because these experiments have
shown to have better correspondence with the results from direct
alcohol fuel cells [37]. The results are presented in Fig. 10 as a cur-
rent density values obtained after applying a constant potential for
300 s as a function of the applied potential.
It should be noticed that in Fig. 10, scales for all the tempera-
tures are the same, and therefore, at the lower temperatures, the
changes between catalysts seem less dramatic; nevertheless, they
are significant. At very low temperature (0 °C), there was hardly
any current produced below 0.5 V. However, at higher potentials,
the differences between the catalyst materials are remarkable:
the current density produced with the Pt (Com) catalyst at the
highest potential (0.8 V) is 1.5 times higher than that produced
by the Pt (ALD) or PtCoPt (ALD) catalyst and more than 3 times
higher produced by the PtCo (ALD) catalyst. This could indicate
that the annealing of the catalyst would improve its properties at
the lower temperatures.
The trends of the onset potentials as a function of temperature
in Table 3 are clear: firstly, the onset potential of all the catalysts
decrease with the increase in temperature due to the enhancement
of the reaction rate. Secondly, although the onset potential for
ethanol oxidation reaction over the pure Pt surfaces occurs at
0.56 V at 0 °C, the potential decrease as a function of temperature
is sharper with the Pt (ALD) catalyst, implying that at higher tem-
peratures, this reaction is enhanced, as expected. Moreover, the on-
set potential of ethanol oxidation is shifted to less positive values
at all the temperatures with the bimetallic catalysts reflecting that
the addition of Co provides OH groups on the catalyst surface en-
abling total oxidation of ethanol at lower potential. The shift in
the onset potential due to the Co addition with ALD process is
70–130 mV, depending on the temperature, whereas for the tradi-
tionally prepared Pt and PtCo catalysts, a shift of 20–70 mV has
been reported [23]. With additional Pt cycles on the bimetallic cat-
alyst, the onset potential shifts closer to values of pure Pt surfaces
Table 3
Around the room temperature, current was produced with the
bimetallic catalyst already at low potential region (<0.6 V) as also
observed earlier in Fig. 4. At this region, the Pt (ALD) catalyst
produced notable current when on Pt (Com) only minor currents
are observed below 0.6 V. Nevertheless, at higher potentials, more
poisoning reaction products were formed on bimetallic catalyst
surfaces and the current obtained after 300 s was lower than in
the case of the pure Pt catalysts (Fig. 10). The most likely poisons
The onset potential of ethanol oxidation at various temperatures.
EOnset at 0 °C EOnset at 20 °C EOnset at 40 °C EOnset at 60 °C
1800 rpm (V) 1800 rpm (V) 1800 rpm (V) 1800 rpm (V)
Pt (Com)
Pt (ALD)
PtCo (ALD)
0.56
0.56
0.43
0.52
0.49
0.39
0.42
0.49
0.44
0.37
0.41
0.46
0.41
0.32
0.37
PtCoPt (ALD) 0.44

46
A. Santasalo-Aarnio et al. / Journal of Catalysis 309 (2014) 38–48
Fig. 10. Currents obtained after 300 s of potential step at selected potentials for ethanol oxidation in solution of 1 mol dmꢀ3 ethanol and 0.1 mol dmꢀ3 NaOH. The lines are
added as a guide for the eye.
on a Pt-based catalyst at this potential region are acetaldehyde,
acetone, and CO of which two first were observed on all the studied
ysis was performed by using the cyclic voltammetry data instead of
catalyst surfaces with the FTIR analysis (Fig. 7). The bimetallic
potential step experiments, in which strong catalyst poisoning
surfaces seem to be poisoned faster that the pure Pt ones indicating
would affect the results. The currents were collected from the third
that they promote more the formation of poisoning species.
scan of the cyclic voltammetry curves at slow scan rate (10 mV sꢀ1
)
Observing CO2 only at the potential of 0.8 V on the bimetallic
surfaces suggests that there is also some formation of CO at these
lower potentials although it was not observed in the FTIR
experiments in Fig. 7 (likely because the amounts were under
the detection limit).
and from the negative-going scan, in order to minimize the effect
of anion and water adsorption as well as catalyst poisoning [56].
In addition, the values were taken from experiments where the
working electrode was rotated to ensure that there are no bubbles
covering the surface especially at the elevated temperatures and to
minimize the diffusion limitations. It was demonstrated earlier
that the activation energies do not depend on the WE rotation rate
[56]. For the analysis of the PtCo (ALD) catalyst, the values at 60 °C
When temperatures are further increased to 40 °C (Fig. 10), only
a slight increase in the performance on all the catalyst is observed
and this effect is most visible at the low potential region (below
0.6 V). The largest differences between the catalysts are observed
at the high temperature (60 °C): both monometallic Pt catalysts
perform similarly at lower potentials, however, at potentials higher
than 0.6 V the Pt (ALD) catalyst exhibits increased performance. At
60 °C, PtCoPt (ALD) shows superior performance throughout the
whole potential region confirming that this catalyst is not only sta-
ble but also active for ethanol oxidation at higher temperatures. In
contrast, the PtCo (ALD) catalyst shows lower activity for ethanol
oxidation at 60 °C than at 40 °C, and therefore, it cannot be totally
ruled out that at 60 °C, some dissolution of Co from the surface of
the PtCo (ALD) catalyst could occur. This fact underlines the impor-
tance of the protection of the bimetallic catalysts with additional Pt
cycles.
3.6. Activation energies
To obtain more information of the temperature effects, Arrhe-
nius analysis of the electrochemical data from Fig. 3 was performed
(Supporting information, Figs. S1–S3) and the activation energies
(Ea) as a function of the applied potential are shown in Fig. 11. In
Fig. 11. The activation energies for ethanol oxidation at various potentials obtained
from the Arrhenius plots (Figs. S1–S3). The values base on the negative-going scans
of the cyclic voltammetry presented in Fig. 9 and the lines are added as a guide for
order to identify the limiting factors of ethanol oxidation, this anal- the eye.

A. Santasalo-Aarnio et al. / Journal of Catalysis 309 (2014) 38–48
47
were discarded to ensure that the values are describing ethanol
oxidation and not Co dissolution.
mainly produced by the 4 eꢀ path. The activation energies obtained
for the ALD catalyst showed that the PtCoPt (ALD) catalyst possess
more bimetallic-like properties than Pt-like properties, confirming
that the Co under the surface can still contribute via bi-functional
mechanism. Especially at the elevated temperatures, the stability
of the bimetallic catalyst remains as an issue even in an alkaline
media and therefore the protection of the bimetallic catalyst is
important. PtCoPt catalyst prepared by the ALD method showed
significant increase in stability at the high temperatures. The re-
sults clearly demonstrate that the bimetallic catalyst having three
deposition cycles alternating with Pt and Co show both increased
activity and durability in comparison with traditional bimetallic
catalysts.
Overall, Fig. 11 shows that the activation energies for ethanol
oxidation on the ALD-prepared catalysts are low (7–20 kJ molꢀ1
)
and in agreement with those reported for ethanol oxidation over
Pt nanoparticle catalysts in a NaOH solution [46,57,58]. However,
the previously reported values were less potential dependent that
might be due to the fact that the curves were analyzed by using
currents recorded at the positive-going scan of ethanol oxidation.
As pointed out by Cohen et al. [56] especially in alkaline media,
the difference between values obtained with positive- and
negative-going scans is significant, making the comparison of the
values demanding.
For the Pt (ALD) catalyst, the activation energies decrease with
the decrease in potential as expected for the negative-going scan
until 0.55 V is reached. At the lower potentials, the activation en-
ergy continues to increase, indicating that a reaction barrier is
faced that is most likely related to the availability of the OH-groups
on the catalyst surface. With the bimetallic catalysts, almost linear
activation energy dependency throughout the potential range is
shown. This result suggests that at this potential range, the reac-
tion mechanism remains the same as suggested the reaction prod-
uct distribution in Fig. 8. Even though the FTIR analysis showed
(Fig. 8) that less CO2 was produced over the PtCoPt (ALD) catalyst,
the Ea values obtained for PtCoPt (ALD) resemble more those of
PtCo (ALD) than Pt (ALD), suggesting that the reaction mechanism
is similar to the bimetallic PtCo (ALD) catalyst. This would support
the conclusion that the structural PtCoPt (ALD) catalyst produces
similar activity as bimetallic catalyst but has superior stability at
the high temperatures.
As discussed in Section 3.4, CO2 produced in an alkaline electro-
lyte will react to carbonate; however, the IR bands of carbonate
overlap with the acetate and therefore cannot be distinguish. On
the bimetallic catalyst, also traces of CO2 are observed associated
with a high production rate in the thin layer configuration where
all CO2 is not instantly carbonated. This implies that FTIR cannot
exclude CO2 production from ethanol oxidation in an alkaline med-
ia, and therefore, CO2 production also on the Pt catalyst cannot be
ruled out. The activation energies for ethanol oxidation on all of the
ALD-prepared catalysts (Fig. 11) seem to be similar above 0.55 V
that might suggests that ethanol is totally oxidized also on Pt
(ALD) even though the CO2 amounts produced are not high enough
to be detected with FTIR in an alkaline electrolyte.
Acknowledgements
The authors acknowledge Aalto University and Academy of Fin-
land for funding. Matti Aarnio is acknowledged for providing the
images of the cell configuration.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2013.08.027.
References
[1] C. Lamy, E.M. Belgsir, J.-M. Leger, J. Appl. Electrochem. 31 (2001) 799.
[2] A. Santasalo, T. Kallio, K. Kontturi, Platinum Metals Rev. 53 (2009) 58.
[3] J.R. Varcoe, R.C.T. Slade, E.L.H. Yee, S.D. Poynton, D.J. Driscoll, J. Power Sources
173 (2007) 194.
[4] C. Lamy, J.-M. Leger, S. Srinivasan, in: J.O’M. Bockris, B.E. Conway, R.E. White
(Eds.), Modern Aspects of Electrochemistry, vol. 32, Kluwer Academic/Plenum
Publishers, New York, 2001, p. 65.
[5] S.C.S. Lai, M.T.M. Koper, Phys. Chem. Chem. Phys. 11 (2009) 10446.
[6] J.S. Spendelow, A. Wieckowski, Phys. Chem. Chem. Phys. 9 (2007) 2654.
[7] G. Cui, S. Song, P.K. Shen, A. Kowal, C. Bianchini, J. Phys. Chem. C 113 (2009)
15639.
[8] A. Santasalo-Aarnio, S. Hietala, T. Rauhama, T. Kallio, J. Power Sources 196
(2011) 6153.
[9] H. Bunazawa, Y. Yamazaki, J. Power Sources 190 (2009) 210.
[10] J.R. Varcoe, R.C.T. Slade, Fuel Cells 5 (2005) 187.
[11] N. Fujiwara, Z. Siroma, S. Yamazaki, T. Ioroi, H. Senoh, K. Yasuda, J. Power
Sources 185 (2008) 621.
[12] Y.S. Li, T.S. Zhao, Z.X. Liang, J. Power Sources 190 (2009) 223.
[13] V. Bambagioni, C. Bianchini, A. Marchionni, J. Filippi, F. Vizza, J. Teddy, P. Serp,
M. Zhiani, J. Power Sources 190 (2009) 241.
[14] U. Schmiemann, U. Muller, H. Baltruschat, Electrochim. Acta 40 (1995) 99.
[15] J. Shin, W.J. Tornquist, C. Korzeniewski, C.S. Hoaglund, Surf. Sci. 364 (1996)
122.
[16] F. Colmati, G. Tremiliosi-Filho, E.R. Gonzalez, A. Berna, E. Herrero, J.M. Feliu,
Faraday Discuss. 140 (2009) 379.
4. Conclusion
[17] D. Bayer, S. Berenger, C. Cremers, J. Tubke, ECS Trans. 25 (2010) 95.
[18] O.A. Petrii, B.I. Podlovchenko, A.N. Frumkin, H. Lal, J. Electroanal. Chem. 10
(1965) 253.
[19] T. Frelink, W. Visscher, J.A.R. van Veen, Surf. Sci. 335 (1995) 353.
[20] C. Wang, N.M. Markovic, V.R. Stamenkovic, ACS Catal. 2 (2012) 891.
[21] S. Koh, C. Yu, P. Mani, R. Srivastava, P. Strasser, J. Power Sources 172 (2007) 50.
[22] M. Oezaslan, F. Hasche, P. Strasser, J. Electrochem. Soc. 159 (2012) B394.
[23] C. Xu, Y. Su, L. Tan, J. Zhang, S. Chen, S.P. Jiang, Electrochim. Acta 54 (2009)
6322.
[24] H. Zhao, J. Dong, S. Xing, Y. Li, J. Shen, J. Xu, Int. J. Hydrogen Energy 36 (2011)
9551.
[25] H. Huang, Y. Fan, X. Wang, Electrochim. Acta 80 (2012) 118.
[26] P. Hernandez-Fernandez, M. Montiel, P. Ocon, J.L.G. Fierro, H. Wang, H.D.
Abruña, S. Rojas, J. Power Sources 195 (2010) 7959.
[27] X. Zhang, H. Wang, J. Key, V. Linkov, S. Ji, X. Wang, Z. Lei, R. Wang, J.
Electrochem. Soc. 159 (2012) B270.
[28] H. Qiu, F. Zou, Appl. Mater. Interfaces 4 (2012) 1404.
[29] S. Haukka, T. Suntola, Int. Sci. 5 (1997) 119.
[30] E. Rikkinen, A. Santasalo-Aarnio, S. Airaksinen, M. Borghei, V. Viitanen, J.
Sainio, E.I. Kauppinen, T. Kallio, A.O.I. Krause, J. Phys. Chem. C 115 (2011)
23067.
[31] T. Shu, S.-J. Liao, C.-T. Hsieh, A.K. Roy, Y.-Y. Liu, D.-Y. Tzou, W.-Y. Chen,
Electrochim. Acta 75 (2012) 101.
In this paper, Pt, PtCo, and PtCoPt catalysts were prepared by
ALD and their activities for ethanol oxidation at a wide range of
temperatures are evaluated. At the room temperature, the Pt
(ALD) catalyst showed increased activity per Pt surface area in
comparison with commercial Pt (Com) catalyst and this can be
associated with a more carefully controlled synthesis resulting in
more crystalline structure. Introduction of Co ALD cycles to form
bimetallic catalyst enhanced ethanol oxidation, which could be
demonstrated by a decrease in the onset potential and the increase
in oxidation current at low overpotentials; however, the poisoning
of the catalyst surfaces was accelerated. When an additional cycle
of Pt was introduced on the bimetallic catalyst, the onset potential
shifted toward values more Pt like; thus, the current densities on
PtCoPt (ALD) were further increased.
The FTIR reaction product analysis confirmed that with both
bimetallic PtCo catalysts, CO2 is detected as an additional final
product; however, it was not observed with the pure Pt catalyst.
This suggests that the addition of Co enhances the mechanism to-
ward to 12 eꢀ path, while on a pure Pt surface, the current is
[32] J. Li, X. Liang, D.M. King, Y.-B. Jiang, A.W. Weimer, Appl. Catal. B: Environ. 97
(2010) 220.

48
A. Santasalo-Aarnio et al. / Journal of Catalysis 309 (2014) 38–48
[33] Y.-C. Hsued, C.-C. Wang, C.-C. Kei, Y.-H. Lin, C. Liu, T.-P. Perng, J. Catal. 294
(2012) 63.
[34] S.T. Christensen, H. Feng, J.L. Libera, N. Guo, J.T. Miller, P.C. Stair, J.W. Elam,
Nano Lett. 10 (2010) 3047.
[44] Q. He, S. Mukerjee, Electrochim. Acta 55 (2010) 1709.
[45] H. Schulenburg, J. Durst, E. Muller, A. Wokaun, G.G. Scherer, J. Electroanal.
Chem. 642 (2010) 52.
[46] J. Datta, A. Dutta, S. Mukherjee, J. Phys. Chem. C 115 (2011) 15324.
[47] D. Bayer, S. Berenger, M. Joos, C. Cremers, J. Tubke, Int. J. Hydrogen Energy 35
(2010) 12660.
[35] X. Jiang, T.M. Gur, F.B. Prinz, S.F. Bent, Chem. Matter. 22 (2010) 3024.
[36] Y. Lei, B. Liu, R.J. Lobo-Lapidus, T. Wu, H. Feng, X. Xia, A.U. Mane, J.L. Libera, J.P.
Greeley, J.T. Miller, J.W. Elam, Chem. Mater. 24 (2012) 3525.
[37] A. Santasalo-Aarnio, S. Tuomi, K. Jalkainen, K. Kontturi, T. Kallio, Electrochim.
Acta 87 (2013) 730.
[38] M. Figueiredo, V. Climent, J.M. Feliu, Electrocatalysis 2 (2011) 255.
[39] M. Figueiredo, A. Santasalo-Aarnio, F. Vidal-Iglesias, J. Solla-Gullón, J.M. Feliu,
K. Kontturi, T. Kallio, Appl. Catal. B: Environ. 140–141 (2013) 378.
[40] E. Sairanen, M. Figueiredo, R. Karinen, A. Santasalo-Aarnio, H. Jiang, J. Sainio, T.
Kallio, J. Lehtonen, Appl. Catal. B: Environ. (2013) (submitted for publication).
[41] F.J. Vidal-Iglesias, R.M. Arán-Ais, J. Solla-Gullon, E. Herrero, J.M. Feliu, ACS
Catal. 2 (2012) 901.
[48] P.-C. Su, H.-S. Chen, T.-Y. Chen, C.-W. Liu, C.-H. Lee, J.-F. Lee, T.-S. Chan, K.-W.
Wang, Int. J. Hydrogen Energy 38 (2013) 4474.
[49] S.Y. Shen, T.S. Zhao, J.B. Xu, Int. J. Hydrogen Energy 35 (2010) 12911.
[50] S.C.S. Lai, M.T.M. Koper, Faraday Discuss. 140 (2009) 399.
[51] X.H. Xia, H.-D. Liess, T. Iwasita, J. Electroanal. Chem. 437 (1997) 233.
[52] T. Iwasita, B. Rasch, E. Cattaneo, W. Vielstich, Electrochim. Acta 34 (1989)
1073.
[53] Z. Zhou, Q. Wang, J. Lin, N. Tian, S. Sun, Electrochim. Acta 55 (2010) 7995.
[54] Q. Chen, J. Solla-Gullón, S. Sun, J.M. Feliu, Electrochim. Acta 55 (2010) 7982.
[55] D. Tarnowski, C. Korzeniewski, J. Phys. Chem. B 101 (1997) 253.
[56] J.L. Cohen, D.J. Volpe, H.D. Abruña, Phys. Chem. Chem. Phys. 9 (2007) 49.
[57] S.S. Mahapatra, A. Dutta, J. Datta, Electrochim. Acta 55 (2010) 9097.
[42] A.V. Tripkovic, K.Dj. Popovic, J.D. Lovic, V.M. Jovanovic, A. Kowal, J. Electroanal.
Chem. 572 (2004) 119.
[43] S. Chumillas, C. Buso-Rogero, J. Solla-Gullon, F.J. Vidal-Iglesias, E. Herrero, J.M.
Feliu, Electrochem. Commun. 13 (2011) 1194.
[58] A.V. Tripkovic, S. Strbac, K.Dj. Popovic, Electrochem. Commun.
484.
5 (2003)