Characteristics of adsorbed CO and CH3OH oxidation reactions for complex Pt/Ru catalyst systems

Article abstract of DOI:10.1016/j.electacta.2004.10.057

Pt/Ru powder catalysts of the same nominal Pt to Ru composition were prepared using a range of methods resulting in different catalyst properties. Two PtRu alloy catalysts were prepared, one of which has essentially the same surface and bulk Pt to Ru composition, while the second catalyst is surface enriched with Ru. Two powders consisting of non-alloyed Pt phases and surfaces enriched with Ru were also prepared. The oxidation state of the surface Ru of the latter two catalysts is mainly metallic Ru or Ru-oxides. The catalyst consisting of Ru-oxides was formed at 500 °C. Part of this catalyst was then reduced in a H₂ atmosphere under mild conditions, thus catalyst properties such as particle size are not changed, as they are locked in during previous high temperature treatment. The oxidation kinetics of adsorbed CO (CO_ads) and solution CH₃OH were studied and compared to the Ru ad-metal state and Pt to Ru site distribution of the as-prepared catalysts. The kinetics of the CO_ads oxidation reaction were observed to be slower for the catalyst containing Ru-oxides as opposed to mainly Ru metal. The CH₃OH oxidation activities measured per Pt surface area, i.e., the catalytic activities are better (by ca. seven times) for the alloy catalysts than the non-alloyed Pt/Ru catalysts. The latter two catalysts showed essentially the same catalytic CH₃OH oxidation activities, i.e., independent of the Ru ad-metal oxidation state of the as-prepared catalysts. Furthermore, it is shown that CO_ads oxidation experiments can be used to extract characteristics that allow the comparison of catalytic activities for the CO_ads oxidation reaction and Pt to Ru site distribution for complex catalyst systems. Crown Copyright

Full text of DOI:10.1016/j.electacta.2004.10.057

Electrochimica Acta 50 (2005) 2401–2414
Characteristics of adsorbed CO and CH3OH oxidation
reactions for complex Pt/Ru catalyst systems
C. Bock∗, M.-A. Blakely, B. MacDougall
National Research Council Canada, Montreal Road, Ottawa, Ont., Canada K1A 0R6
Received 22 July 2004; received in revised form 14 September 2004; accepted 11 October 2004
Available online 21 November 2004
Abstract
Pt/Ru powder catalysts of the same nominal Pt to Ru composition were prepared using a range of methods resulting in different catalyst
properties. Two PtRu alloy catalysts were prepared, one of which has essentially the same surface and bulk Pt to Ru composition, while
the second catalyst is surface enriched with Ru. Two powders consisting of non-alloyed Pt phases and surfaces enriched with Ru were also
prepared. The oxidation state of the surface Ru of the latter two catalysts is mainly metallic Ru or Ru-oxides. The catalyst consisting of
Ru-oxides was formed at 500 ◦C. Part of this catalyst was then reduced in a H2 atmosphere under “mild” conditions, thus catalyst properties
such as particle size are not changed, as they are locked in during previous high temperature treatment. The oxidation kinetics of adsorbed CO
(COads) and solution CH3OH were studied and compared to the Ru ad-metal state and Pt to Ru site distribution of the as-prepared catalysts.
The kinetics of the COads oxidation reaction were observed to be slower for the catalyst containing Ru-oxides as opposed to mainly Ru
metal. The CH3OH oxidation activities measured per Pt surface area, i.e., the catalytic activities are better (by ca. seven times) for the alloy
catalysts than the non-alloyed Pt/Ru catalysts. The latter two catalysts showed essentially the same catalytic CH3OH oxidation activities, i.e.,
independent of the Ru ad-metal oxidation state of the as-prepared catalysts. Furthermore, it is shown that COads oxidation experiments can be
used to extract characteristics that allow the comparison of catalytic activities for the COads oxidation reaction and Pt to Ru site distribution
for complex catalyst systems.
Crown Copyright © 2004 Published by Elsevier Ltd. All rights reserved.
Keywords: Fuel cells; Platinum; Ruthenium; Carbon monoxide; Methanol
Pt (CH3OH)ads = Pt (CO)ads + 4H+ + 4e−
Ru + H2O → Ru OH + H+ + 1e−
(2)
(3)
1. Introduction
Direct methanol and reformate fuel cells are emerging as
potential power sources for portable devices and transporta-
tion purposes. Among other obstacles, the performance of
the anode catalysts towards the oxidation of methanol and/or
CO to CO2 needs to be improved. Pt/Ru catalysts have shown
superior performance for the CH3OH and CO oxidation reac-
tions [1,2]. The beneficial catalytic effect has been attributed
to a bi-functional mechanism, as shown in the following
scheme for the case of the CH3OH oxidation reaction [2]:
Pt (CO)ads + Ru OH → Pt + Ru + CO2 + H+ + 1e−
(4)
As a result of the adsorption/dehydrogenation reaction of
CH3OH that takes place on Pt sites, an intermediate “CO”-
type species is formed. This species is believed to be oxidized
to CO2 with the assistance of OH species, which are formed
by the partial oxidation of H2O on Ru surface sites. Based
on this reaction scheme, it is clear that the ratio and distribu-
tion of Pt and Ru catalyst surface sites influences the CH3OH
oxidation kinetics [3–5]. In fact, prior work has shown that
catalystsconsistingofhomogenouslydistributedsurfacesites
Pt + CH3OH = Pt (CH3OH)ads
(1)
∗
Corresponding author. Tel.: +1 613 990 2252; fax: +1 613 941 2529.
E-mail address: christina.bock@nrc-cnrc.ca (C. Bock).
0013-4686/$ – see front matter Crown Copyright © 2004 Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2004.10.057

2402
C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
of 70–30 atomic percent (at.%) Pt to Ru ratio display the best
activity for the CH3OH oxidation reaction [5]. The chemical
state of the Ru ad-metal component has also been suggested
to influence the CH3OH oxidation kinetics significantly [6];
however, thereisdisagreementregardingthemostactiveform
of Ru. Bimetallic PtRu alloy catalysts have been reported to
show significantly poorer (ca. 250 times lower) activities for
the CH3OH oxidation reaction than catalysts consisting of
mixtures of Pt metal and various Ru-oxides [7]. Based on
these results, it was recommended that PtRu alloys should
not be used as catalysts for direct methanol fuel cells [6]. In
subsequent studies [7,8], it was suggested that Ru in its metal-
lic state is most active for the CH3OH oxidation reaction. The
authors of the first study [7] used commercially available un-
supported partially alloyed Pt/Ru catalysts. They based their
conclusions mainly on adsorbed CO (COads) to CO2 oxida-
tion studies for the as-received and H2 reduced catalysts, i.e.,
involving a limited number of catalysts. In the second study
[8], XPS was used to obtain the chemical state of the Ru im-
mediately before and after CH3OH oxidation studies were
carried out using Pt/Ru and Pt/Ru-oxide catalysts. This study
showed that Ru-oxides including RuO2 are at least partially
reduced to metallic Ru in the electrochemical environment; a
result supported by another report [9]. Other papers also ad-
dressed the issue of the influence of the Ru ad-metal state of
the as-prepared catalysts [10,11]. Different Pt/Ru-oxide cata-
lysts were formed within a broad and high temperature range
between 400 and 600 ◦C [10], while in another study an un-
supported PtRu alloy catalyst [11] was heat-treated at various
temperatures (up to 200 ◦C) and atmospheres to achieve dif-
ferent oxidation states of the Ru ad-metal of the as-prepared
catalysts. Unfortunately, treating and preparing Pt/Ru cata-
lysts at high temperatures typically alters a number of proper-
ties such as alloy composition, Pt to Ru surface distribution,
particle size as well as oxidation state [5], thus making it dif-
ficult to correlate catalytic activities with a particular catalyst
property.
In the present investigation, the influence of the ad-metal
state of the Ru of the as-prepared catalysts on the oxidation
reactions of COads and CH3OH is investigated using four
Pt/Ru powder catalysts. The influence of the Pt to Ru sur-
face site distribution on these reactions is also discussed. The
Pt/Ru powders are prepared so that different Ru oxidation
states and Pt to Ru surface site distributions are obtained. All
powder catalysts are characterized using X-ray diffraction
(XRD), X-ray photon spectroscopy (XPS), energy diffraction
spectroscopy (EDX) and field emission scanning electron mi-
croscopy (FESEM). Both the CH3OH and COads oxidation
kinetics are studied. Kinetic studies of the COads oxidation
reaction are carried out at low potentials where the OH nu-
cleation rate on Pt versus the OH nucleation rate on Ru sites
is small. The COads oxidation data are used to distinguish the
different catalysts with regard to their abilities such as OH
nucleation site formation as well as to gain insight into the
COads diffusion distance. The COads oxidation data are also
compared to the true catalytic activity, i.e., oxidation current
per Pt area of the individual catalysts toward the CH3OH
oxidation reaction.
2. Experimental
2.1. Catalyst powder preparation
Unsupported Pt and Pt/Ru powders were prepared for
this study. The latter consisted of bulk Pt to Ru ratios of
70–30 at.% that is equivalent to the optimal alloy ratio for
the CH3OH oxidation reaction [5]. The catalysts were pre-
pared, as follows: the Pt powder was prepared by reducing
2.03 g of PtCl4 (99.9% purity, Alfa & Aesar) dissolved in
150 mL of H2O with 75 mL of 0.2 M NaBH4 (Anachemia);
a PtRu alloy was formed by the simultaneous and rapid re-
duction [5] of 2.6767 g PtCl4 and 0.5603 g RuCl3 (99.99%
purity, Alfa & Aesar) dissolved in 150 mL of H2O with
75 mL of 0.2 M NaBH4. This catalyst powder is referred
to as PtRu (chemically reduced); a Pt/RuO2 (thermal) cat-
alyst was formed by mixing (using a mortar) 1.55 g of the
Pt powder with 0.5603 g of RuCl3. The RuCl3 was subse-
quently decomposed in air at 500 ◦C; a Pt/Ru (H2 reduced)
catalyst was formed by reducing the Pt/RuO2 (thermal) cat-
alyst at 100 ◦C in a H2 atmosphere for 1 h; a Pt/RuO2 (ball-
mill) catalyst was formed by ball-milling 2.83 g Pt pow-
der (−200 mesh, 99.98%, Alfa & Aesar) and 0.85 g RuO2
(Electronic Grade, Premion, 99.95%, Alfa & Aesar) using a
Spex2000 ball mill mixer for 40 h. Three grams of Al pow-
der (−40 + 325 mesh, 99.8%, Alfa & Aesar) and 0.2 g of NaF
(Anachemia) were also added to the mixture following a pro-
cedure described elsewhere [12]. The powders were mixed
in a tungsten carbide (WC) container (50 cm3, Spex) with
four WC balls (each ball with diameter of 7/16 in., Spex).
The ball-milled powder mixture was leached in 1000 mL of
1 M NaOH (EM Science, Merck, ACS grade) solution for
20 h to dissolve the aluminum. After preparation, the Pt and
Pt/Ru powder catalysts were washed excessively with H2O,
filtered and dried in an air oven at 80 ◦C. The PtCl4 and RuCl3
precursor salts were dried in an air oven at 135 ◦C prior to
their use.
2.2. Working electrode preparation
The catalyst powders were formed into electrodes by
sonicating 13 mg of a particular powder dispersed in solu-
tions consisting of 300 ␮L of a Nafion®117 solution (5 wt.%
Nafion dissolved in lower alcohols, Fluka) and 2 mL of H2O
for 30 min. Appropriate amounts (1–20 ␮L) of the suspen-
sions were pipetted onto ca. geometrical area of 0.35 cm2 Au
foil electrodes (99.9% Au, 0.1 mm thick, Goodfellow) form-
ing thin catalyst layers. The electrodes were dried in air and
room temperature. The Au foils were firmly attached to Au
wire electrodes and Au not covered with the catalyst powder
was carefully wrapped with Teflon tape.

C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
2403
2.3. Cells and electrodes
calstudies, thesolutionsweredeoxygenatedusinghighpurity
Argon gas. ACS grade chemicals and high resistivity 18 Mꢀ
water were used. The experiments were carried out at room
temperature.
Three compartment cells, in which the reference elec-
trode was separated from the working and counter elec-
trode compartment by a Luggin capillary, were employed
for the electrochemical studies. All potentials reported in
this paper are versus the reversible hydrogen electrode
(RHE), even though either a mercury sulfate electrode
(MSE = −0.68 V versus RHE [13]) or a saturated calomel
electrode (SCE = −0.24 V versus RHE [13]) were used for
the actual experiments. A large surface area Pt gauze served
as counter electrode.
3. Results and discussion
3.1. Characterization of the as-prepared Pt and Pt/Ru
catalysts
3.1.1. XRD characterization
Fig. 1 shows raw XRD spectra for the Pt based powders
prepared in this work. In all spectra, the diffraction peaks
of Si that was homogenously mixed with the Pt/Ru powders
and used as internal standard are clearly recognizable. The
entire spectra were used to obtain lattice parameter values
for the various phases as well as microstrain and particle size
estimations using the Topas software program. The data are
shown in Table 1. The following space groups were used for
the XRD data fitting: Pt: Fm-3m(225), Ru: P63/mmc(194)
RuO2: P42/mnm(136) and Al2O3: R-3c(167) [14]. In the
case of PtRu alloy formation, the lattice parameter values
estimated for the fcc Pt phase (aPt(fcc)) in nanometers were
used to calculate the amount of Ru dissolved in the Pt lattice
using a Vegard’s law relationship of:
2.4. Techniques and instrumentations
Electrochemical experiments were performed using So-
larotron SI 1287 electrochemical interface (Solarotron Group
Ltd.) driven by Corrware software program (Scribner, As-
soc.). A Scintag XDS2000 system was employed using a
Cu K␣ source to obtain XRD spectra of the catalysts. The
scanning angle 2θ extended from 20◦ to 80◦. The software
program Topas 2 (DIFFRACPLUS Topas, Bruker axis Inc.)
was employed to extract lattice parameter constants from the
experimental XRD spectra. The entire XRD spectra from 20◦
to 80◦ were employed to analyse the data. Si powder (typ-
ically 1–20 ␮m, 99.9985% purity, Alfa & Aesar) was used
as internal standard for the XRD analysis. XPS spectra were
obtained using a Kratos Axis Ultra spectrometer equipped
with a monochromatized Al K␣ source. The catalyst pow-
ders were attached to sticky Cu tape (3M cupper tape, So-
quelec, Montreal) for the XPS analyses. For each catalyst, a
survey spectrum was collected before high-resolution spec-
tra of the C 1s, O 1s, Pt 4f, Ru 3p and Ru 3d core level
regions were collected. Deconvolutions of the XPS spectra
were performed using a CasaXPS Version 2.1.34 (Neal Fair-
ley). A Hitachi S-4800 SEM and EDX were also employed.
The XRD, XPS and Field emission scanning electron mi-
croscopy characterizations were carried out on the unused,
i.e., as-prepared catalyst powders.
aPt(fcc) − 0.3778
Pt (at.%) =
(5)
1.5 × 10−4
This relationship has been determined in previous work
based on ab-initio calculations as well as on experimental
data [5].
2.5. CO adsorption (COads
)
CO was adsorbed onto the Pt based powder electrodes at
0.15 V by bubbling CO gas (Matheson purity, Matheson gas)
through the 0.5 M H2SO4 solution for 20 min. Solution CO
was subsequently removed by bubbling high purity Argon
gas (Air Products) for 30 min holding the potential at 0.15 V.
The potential was then either cycled starting at 0.15 V for two
complete oxidation/reduction cycles or stepped to a particular
potential value, as mentioned in the appropriate section of the
text.
Fig. 1. Slow scan XRD data for Pt and Pt/Ru powders as follows: (a) Pt,
(b) PtRu (chemically reduced), (c) Pt/Ru (H2 reduced), (d) Pt/RuO2 (ther-
mal) and (e) Pt/RuO2 (ball-mill) powder. Silicon powder was homogenously
mixed with the catalyst powders as internal standard. XRD spectra were
obtained between 2θ’s of 20◦ and 80◦ using a step size of 0.06◦ and an
acquisition time of 60 s.
2.6. Solutions
All CH3OH oxidation studies were carried out using 0.5 M
CH3OH + 0.5 M H2SO4 solutions. Prior to the electrochemi-

2404
C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
Table 1
XRD characteristics for the as-prepared Pt and Pt/Ru catalyst powdersa
b
Catalyst
Phase
aPt(fcc) (nm)
Crystallite sizeb (nm)
Micro-strainb (%)
Comments
Pt
Pt fcc
Pt fcc
0.39231 6 × 10−5
25 ± 1
14 ± 1.5
1 ± 5 × 10−2
1 phase: Pt fcc (100 at.% Pt)
1 phase: PtRu alloy (70:30
at.% Pt:Ru ratio)
2 phases: Pt fcc (100 at.%
Pt) + RuO2 phase
PtRu (chemically reduced)
0.3886 1 × 10−5
2.5 ± 1.7 × 10−1
Pt/RuO2 (thermal)
Pt/Ru (H2 reduced)
Pt/RuO2 (ball-mill)
Pt fcc
0.3923 3 × 10−4
30 ± 5
30 ± 5
0.14 ± 3 × 10−2
0.3 ± 3 × 10−2
3.5 ± 1
a: 0.4503 3 × 10−4
RuO2
tetragonal
c: 0.3115 3 × 10−4
Pt fcc
0.39251 2 × 10−5
2 phases: Pt fcc (100 at.%
Pt) + Ru metal phase
a: 0.2698 4 × 10−4
Ru
hexagonal
c: 0.4407 9 × 10−4
Pt fcc
0.3883 1.6 × 10−4 5 ± 1
2 phases: PtRu-alloy (70:30
at.% Pt:Ru ratio) and
Al-oxide (Al2O3) phase
Al2O3
triginal
a: 0.486 3 × 10−3
c: 1.3 1.5 × 10−3
a
All catalysts powders have a nominal Pt to Ru ratio of 70–30 at.%, except for the Pt powder.
Values are estimated from the raw XRD data using the Topas software program.
b
The diffraction peak characteristics for Platinum are
clude a second triginal Al2O3 phase to obtain a good fit. The
presence of Al in the ball-mill catalyst powder was confirmed
by XPS and EDX analyses (see below). This indicates that
leaching in NaOH is insufficient to completely leach the Al
from this catalyst consistent with previous reports [12].
Table 1 also lists the crystallite size values estimated from
the Pt diffraction peaks using the Topas software program for
the five catalysts prepared in this work. The XRD data sug-
gest that the crystallite sizes for the catalyst powders used in
this work are different depending on the preparation method
used as expected. It is noteworthy that the XRD data suggest
the crystallite size of the Pt/RuO2 (thermal) and the Pt/Ru (H2
reduced) catalysts to essentially be the same. This indicates
that the reduction of the Pt/RuO2 (thermal) catalyst in the H2
atmosphere at the moderate temperature of 100 ◦C did not
affect the crystallite size of these two catalysts. This result
is not surprising, as the Pt/RuO2 (thermal) catalyst was pre-
pared at 500 ◦C, thus locking in physical properties such as
crystallite and particle size (the latter will be discussed using
FESEM images). Therefore, it appears reasonable to assume
that the reduction of the surface oxides are the sole changes
introduced by the low temperature H2 treatment.
clearly recognizable in the XRD spectra for all Pt based pow-
ders. The Pt(1 1 1) peak that is observed at 2θ values varying
between 39.8◦ and 40.2◦ (indicating different degrees of PtRu
alloy formation) is the highest intensity peak for all catalysts.
The diffraction peak for the Pt(1 1 1) phase for the Pt powder
catalyst is located at a 2θ value of ∼39.8◦. For this catalyst, an
aPt(fcc) value of 0.39231 ( 6 × 10−5) nm is estimated, con-
sistent with data typically observed for Pt [12]. For both the
Pt/Ru (H2 reduced) and the Pt/RuO2 (thermal) powders, the
positions of the Pt(1 1 1) diffraction peak, and hence also the
a
Pt(fcc) values are essentially the same as for the Pt only cata-
lyst (see Table 1). These results indicate that these two pow-
ders consist of Pt that is present as a non-alloyed phase. In
the case of the Pt/RuO2 (thermal) powder, diffraction peaks
assigned to rutile RuO2 are observed at 2θ values of ∼35.2◦
and ∼54.4◦ [14] while for the Pt/Ru (H2 reduced) powder, a
diffraction peak typical for hexagonal Ru(1 0 1) is observed
at a 2θ value of ∼44◦ [14]. In the case of the PtRu (chemically
reduced) powder, the position of the Pt(1 1 1) diffraction peak
is shifted to a higher 2θ value of ∼40.2◦. For this powder, an
aPt(fcc) value of 0.3886 ( 1 × 10−5) nm was estimated that
suggests that close to 30 at.% of Ru is dissolved in the Pt fcc
lattice. The fitting of the XRD raw data for this powder only
required the Pt fcc phase, i.e., the addition of a Ru phase was
not needed, thus being consistent with the view that Ru is
completely dissolved in the Pt fcc lattice. The XRD spectra
for the Pt/RuO2 (ball-mill) powder shows that the Pt(1 1 1)
diffraction peak is also shifted to a higher 2θ value of ∼40.2◦.
Essentially the same aPt(fcc) lattice parameter value of 0.3883
( 1.6 × 10−4) nm was estimated for the ball-mill catalyst as
for the PtRu (chemically reduced) powder indicating that a
PtRu alloy of 70–30 at.% Pt to Ru is formed for the two cat-
alysts. For the analysis of the XRD data of the ball-milled
catalyst, a Pt fcc phase was used. It was also necessary to in-
3.1.2. XPS characterization
Table 2 summarizes the XPS characteristics of the unused
Pt and Pt/Ru powder catalysts. The Pt to Ru (at.%) ratios
were also obtained from the XPS data. The mean free path
values for Pt 4f and Ru 3p are very similar, namely 1.684 and
1.536 nm, respectively [15]. Therefore, the XPS data reflect
the surface Pt to Ru ratios in the case of particles of uniform
Pt to Ru distribution. In the case of powders of catalyst parti-
cles consisting of non-uniform Pt and Ru concentrations, the
Pt to Ru ratios determined from XPS data may not present
a true measurement of the Pt to Ru surface ratios. This is
likely to the case, for Pt particles having surfaces enriched

C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
2405
Table 2
XPS characteristics for the as-prepared Pt and Pt/Ru catalyst powders
Catalyst
Species
Pt 4f7/2
Assignment
Binding energya (eV)
at.% Pt:Ru ratiob
70:30
at.% per speciesc
PtRu (chemically reduced)
Pt metal
PtO
71.3
72.2
74.1
462
52
19
29
69
31
Pt-chlorides
Ru metal
Intermediate between
Ru metal and RuO2
Pt metal
Ru 3p3/2
Pt 4f7/2
465.1
Pt/RuO2 (thermal)
71.2
72.4
74.2
36:64
45
10
45
75
PtO
PtO2
Ru 3p3/2
Intermediate between
Ru metal and RuO2
RuO2
462.8
466
25
Pt/Ru (H2 reduced)
Pt 4f7/2
Pt metal
PtO
Ru metal
Intermediate between
Ru metal and RuO2
Pt metal
PtO
PtO2
71.3
72.2
462
44:56
54:46
74
26
69
31
Ru 3p3/2
463.5
Pt/RuO2 (ball-milled)
Pt 4f7/2
71.3
72.6
74.3
462
58
24
18
65
35
Ru 3p3/2
Ru metal
Intermediate between
Ru metal and RuO2
465
a
The binding energy values were corrected for possible charging effects using the C 1s peak at 284.4 V [13].
At.% Pt to Ru ratio were estimated from the raw XPS signals of the Pt 4f and Ru 3p peaks using sensitivity factors of 5.575 and 2.043, respectively.
At.% ratio calculated for a particular species, i.e., Pt or Ru.
b
c
with Ru. The XPS data for the PtRu (chemically reduced)
catalyst have been discussed in previous work [5]. The fact
that in the case of this particular powder, the Pt to Ru ra-
tio estimated from the XPS data is the same as the Pt to
Ru bulk alloy ratio determined from the XRD data suggest
that catalyst particles of uniform Pt and Ru concentrations
are formed. This has been further supported by COads strip-
ping voltametric experiments [5], which are essentially the
same as found for well characterized bulk metal PtRu al-
loys of 70–30 at.% Pt to Ru surface ratio. In the case of the
Pt/RuO2 (thermal) and the Pt/Ru (H2 reduced) powders, the
Pt to Ru atomic ratio are clearly different from the nominal
70 to 30 at.% ratio. It is likely that the Pt to Ru atomic ratios
determined from the XPS data do not reflect the exact surface
composition for these two catalyst powders. The XPS data
suggest that the surfaces of these two catalysts are enriched
with Ru. This appears reasonable as the method used to pre-
pare these powders (impregnating Pt powders with the RuCl3
precursorsalt)isexpectedtoresultincatalystpowdersofnon-
uniform Pt to Ru composition with Ru preferably located on
the preformed Pt particle surface. The Pt to Ru atomic ra-
tios estimated from the XPS data of the Pt/RuO2 (ball-mill)
catalyst are also different from their nominal ratio, thus indi-
cating that the Pt and Ru components are not uniformly dis-
tributed for this catalyst. However, in the case of the latter, the
ball-milling is believed to improve the distribution between
the Pt to Ru sites. Furthermore, the XPS data also indicated
the presence of Al (ca. 30 at.% Al per Pt + Ru + Al) on this
catalyst surface.
The binding energy values and the correspondingly as-
signed species of the deconvoluted Pt 4f and Ru 3p core level
regions of the catalysts studied in this work are also shown
in Table 2. Small amounts of Pt-chlorides are found for the
PtRu (chemically reduced) catalyst. It should be noted that
the Pt-chlorides are believed to be located on the catalyst
surface. The actual amount of chlorides per catalyst weight
is expected to be very small, as Pt-chlorides were not de-
tected in the XRD and EDX spectra that reflect the bulk cat-
alyst properties. The XPS data for the Pt 4f region for the Pt
and Pt/Ru powders suggest Pt to be present as both Pt metal
and PtO. However, PtO2 is also detected for the Pt/RuO2
(thermal) powder. This is expected, as this catalyst was heat-
treated at 500 ◦C in air, thus resulting in the oxidation of
the catalyst surface. It should be noted that neither PtO nor
PtO2 phases are observed in the XRD spectra of the Pt/RuO2
(thermal) powder, i.e., the XRD data suggest the presence
of only a metallic Pt phase for this powder. Therefore, it is
believed that only the Pt surface is being oxidized to PtO
and PtO2. Furthermore, the electrochemical characteristics,
namely the cyclic voltamograms for a Pt powder heat-treated
at 500 ◦C in O2 for 1 h were the same as for polycrystalline Pt
metal, suggesting that the Pt-oxides are easily reduced elec-
trochemically. The data for the Ru 3p core level region show
that the form of Ru present on the catalyst surfaces depends
on the preparation procedure. The Ru on the surfaces for
the PtRu (chemically reduced), the Pt/Ru (H2 reduced) and
the Pt/RuO2 (ball-mill) catalysts is suggested to be mainly
Ru metal (between 60 and 80%) and an intermediate form

2406
C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
Fig. 2. FESEM images for (a) PtRu (chemically reduced) catalyst powder at a 50,000× magnification, (b) Pt/RuO2 (thermal) catalyst powder at a 40,000×
magnification, (c) Pt/Ru (H2 reduced) catalyst at a 40,000× magnification and (d) Pt/RuO2 (ball-mill) catalyst at a 10,000× magnification.
between the metallic and (+IV)-state (between 40 and 20%).
The XPS data for the Pt/RuO2 (thermal) catalyst suggest that
the Ru on the catalyst surface is present in its oxidized forms,
ca. 25% in the +IV state, i.e., as RuO2 and ca. 75% in its
intermediate form between the metallic and (+IV)-state. The
XPS data (consistent with the XRD results) strongly suggest
that the Ru-oxide powder is reduced to mainly Ru metal dur-
ing ball-milling. Ball-milling is a high energy process car-
ried out in an inert atmosphere and in the presence of Al
that could be oxidized to Al-oxide, while the Ru-oxides are
reduced. It has been reported that Mo-oxides are reduced to
Mo metal during ball-milling [16]. It is noteworthy that small
amounts of chlorides were found for Pt/RuO2 (thermal) cat-
alysts that were analyzed immediately after preparation, i.e.,
were not washed with H2O. Chlorides were not detected by
XPS analysis carried out for well-rinsed (with H2O) Pt/RuO2
(thermal) catalysts indicating that these water-soluble chlo-
rides are removed. This is consistent with previous re-
ports for thermally prepared Pt/RuO2 [10] and RuO2 [17]
electrodes.
catalyst morphology observed by FESEM for the thermal
and H2 reduced powders are seen to be essentially the same
(Fig. 2b and c). This is consistent with the view that reduc-
tion in the H2 atmosphere at the moderate temperature of
100 ◦C results in mainly the reduction of the surface oxides
that were formed at 500 ◦C. In both cases, the catalysts con-
sist of individual particles in the 30 to 160 nm size range, i.e.,
showing a slightly broader particle size distribution than the
PtRu (chemically reduced) catalyst. The surfaces of the indi-
vidual particles for the H2 reduced and the thermal catalysts
are also rougher than seen for the PtRu catalyst formed by
chemical reduction. Furthermore, the particles appear to be
fused together rather than grown from inside out as seen for
the PtRu (chemically reduced) catalyst reflecting the differ-
ences in the preparation methods. In contrast to the first three
catalysts, the morphology of the ball-mill catalyst (Fig. 2d) is
significantly different, namely less regular than for the alloy
catalyst produced by chemical reduction and the catalyst par-
ticles are significantly larger, namely in the 1–5 ␮m particle
size range. As expected, it is seen that ball-milling introduces
a large number of mechanical defects on the catalysts surface,
i.e., broken edges, corners, etc.
3.1.3. FESEM and EDX characteristics
EDX analysis suggested bulk ratios of 70–30 ( 5) at.% of
Pt to Ru for all four Pt/Ru catalysts, thus confirming that the
bulk and nominal Pt to Ru ratios are essentially the same for
these catalysts. The alloy, thermal and H2 reduced catalysts
consistedofsolelyPtandRu, whiletheEDXanalysesshowed
that the ball-mill catalyst also contains significant amounts
of Al, its composition being: 49:21:30 at.% of Pt:Ru:Al.
Fig. 2a–d show FESEM images for the alloy, thermal, H2
reduced and ball-milled Pt/Ru catalysts, respectively. The
PtRu catalyst (Fig. 2a) produced by the rapid chemical re-
duction method is seen to consist of well-ordered particles in
the nano-scale range (varying between ca. 20 and 120 nm in
diameter). Larger clumps of particles are recognizable that
seem to have been formed by growing from inside out. The

C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
2407
3.2. Pt surface area conversion factors extracted from
combined (COOH)2 and COads oxidation studies
the thermal catalyst, ca. 35% of the Ru surface sites are in
the oxidized form, which is not active for adsorption of CO
at this potential.
It has been shown in previous work that beneficial in-
formation about the Pt/Ru catalyst surface can be gained by
combining activation controlled (COOH)2 oxidation currents
and COads oxidation charge measurements [9]. The (COOH)2
oxidation reaction probes only the Pt surface, and can hence
be used to obtain the electro-active Pt area, APt, while CO
adsorbs on Pt as well as metallic Ru surface sites [9]. The
3.3. Summary of the catalyst properties: bulk and
surface characteristics
Based on the results presented above, the Pt/Ru catalysts
prepared in this work in their as-prepared states are reviewed
and summarized in this section. The PtRu (chemically re-
duced) catalyst powder appears to consist of a bulk PtRu al-
loy with a surface consisting of randomly distributed Pt to Ru
sites of 70–30 at.% composition, i.e., to have the same bulk
and surface compositions (see Section 3.3 for further experi-
mental support). The Pt/Ru (H2 reduced) catalyst is believed
to consist of non-alloyed Pt particles that are coated with Ru
islands. The surface is enriched in Ru and the Ru islands con-
sist of mainly Ru metal and easily reducible Ru-oxides. The
Pt/RuO2 (thermal) catalyst is believed to be very similar to the
Pt/Ru (H2 reduced) catalyst. However, the Ru islands of the
former mainly consist of RuO2 and RuO. The exact size and
shape of the Ru islands is not known. The Pt/RuO2 (ball-mill)
catalyst also consists of a bulk PtRu alloy of the same compo-
sition as the PtRu (chemically reduced catalyst). Its surface,
however, is enriched with Ru, i.e., possibly consisting of ran-
domly distributed Pt to Ru sites as well as some Ru islands.
The ball-mill catalyst also contains aluminum. Aluminum is
expected to not be active for the COads and CH3OH oxida-
tion reactions. In all cases, COads stripping voltammetry was
carried out before and after a set of oxidation experiments.
The COads stripping voltamograms were essentially the same
indicating that the catalyst surfaces were not altered during
the oxidation experiments.
fact is that the COads to CO2 oxidation charge QCO , in
ads
combination with the Pt area estimated using the (COOH)2
oxidation method, can be used to obtain information of the
number of CO molecules adsorbed on catalyst sites other than
Pt. This number is reflected in the conversion factor, fAPt , that
is defined, as follows [9]:
420(␮C cm−2
)
fA
=
× APt
(6)
Pt
QCO
ads
Since the Pt surface area, APt, is known from the (COOH)2
experiments, the anodic charge associated with COads oxida-
tion at Pt is easily calculated as 420 ␮C cm−2 times APt; di-
viding this value by the actual COads oxidation charge gives
the fraction of Pt sites versus the total number of sites, which
are active for CO adsorption. In previous work, the fA fac-
Pt
tors for the particular catalyst powders studied in this work
have been determined and are as follows: Pt powder: 1; PtRu
(chemically reduced) powder: 0.7; Pt/RuO2 (thermal): 0.55;
Pt/Ru (H2 reduced): 0.45 and Pt/RuO2 (ball-mill): 0.5. It has
been shown that the fAPt factors present a direct measurement
of the fraction of CO adsorbed on Pt sites, and in the case of
bi-metallic Pt/Ru catalysts, the value (1 − fAPt ) yields the
fraction of CO stripped from metallic Ru sites. A unity value
indicates that 100% of the adsorbed CO is stripped from Pt
sites, while a fAPt to factor of 0.7 indicates that 70% of the CO
adsorbs on Pt sites and 30% of the CO adsorbs on metallic Ru
sites. The fAPt factors for the four Pt/Ru catalysts used in this
work are less than unity. This indicates that at least some Ru
catalyst surface sites are present as Ru metal in the potential
range used for the adsorption of CO, i.e., 0.1 V. The data also
suggest that the surfaces of all catalysts, except for the PtRu
(chemically reduced) powder, are enriched in Ru, i.e., the Pt
to Ru surface ratio is higher than the 30 at.% nominal Ru
value. This is consistent with the XPS studies. Furthermore,
3.4. COads stripping voltammetry
COads stripping voltammetry can yield useful in situ elec-
trochemical information about Pt-based catalyst surfaces in
the electrochemical environment and has been employed
as a probe for the surface composition for PtRu alloys
[7]. Fig. 3a–e show typical COads stripping voltamograms
recorded at 10 mV s−1 for the Pt based powders prepared in
this work. The overall COads stripping charge (QCO ) esti-
mated for a particular catalyst powder is used to normalize
ads
the y-axis (i/QCO ) in these voltamograms to better compare
ads
the larger fA value for the thermal versus the H2 reduced
the data. Table 3 summarizes various characteristics extracted
from these voltamograms, namely the onset potential (Eon)
for the COads to CO2 oxidation reaction, and the peak width
of the COads stripping peaks measured at peak half height
(EFWHM). As mentioned above, the COads stripping volta-
mogram for the PtRu (chemically reduced) powder used in
this work is essentially the same as reported for a sputter
cleaned PtRu bulk alloy that is proposed to consist of ho-
mogenously distributed Pt and Ru surface sites at the atomic
level of 70–30 at.% Pt to Ru surface composition [18]. This
suggests that the very rapid and simultaneous reduction of the
Pt
catalyst suggests that a larger fraction of the surface Ru is
present in the metallic state at 0.1 V (where CO is adsorbed)
for the latter. This, in turn, indicates that the Ru-oxide present
on the surface of the thermal catalyst is not entirely reduced to
Ru metal. In fact, the fA factors for the H2 reduced and the
Pt
thermal catalyst suggest that 1.5 times less of the Ru surface
sites are in the reduced state at 0.1 V for the latter catalyst.
(This involves the plausible assumption that at 0.1 V, all Ru
surface sites are in the reduced metallic state for the H2 re-
duced catalyst.) The value of 1.5, in turn, suggests that for

2408
C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
Fig. 3. COads stripping voltamograms recorded at 10 mV s−1 in 0.5 M H2SO4. The CO was adsorbed at 0.1 V for 15 min, subsequently, the solution CO was
removed by Argon bubbling for 30 min maintaining the potential at 0.1 V. The black line shows the first cycle, i.e., the COads stripping voltamogram, while
the grey line shows the second cycle that is equivalent to the background CV of the particular catalyst powder. (a) The data for the Pt powder, (b) for the PtRu
(chemically reduced), (c) for the Pt/RuO2 (thermal) powder, (d) for the Pt/Ru (H2 reduced) powder and (e) shows the same for the Pt/RuO2 (ball-mill) powder.
The current axes are normalized using the COads charge values extracted from the corresponding COads stripping voltamograms.

C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
2409
Table 3
COads oxidation kinetics is likely due to the different oxida-
tion states of the Ru ad-metal for these otherwise very similar
catalysts.
COads stripping characteristicsa for the Pt and Pt/Ru powder catalysts
Catalyst
Eon (V) COads
→
Epa (V) vs.
EFWHM
CO2 vs. RHE
RHE
(mV)
The sweep rate was observed to influence the COads strip-
ping characteristics for all four Pt/Ru powders. This is seen
in Fig. 4a and b that show the dependence of the Eon and
EFWHM values on the logarithms of the sweep rate for the
four Pt/Ru catalysts. Such a behavior is well known [19] and
indicates that the COads reaction is kinetically limited un-
der the selected experimental conditions. A shift in the Eon
value to more negative potential and a decrease in the EFWHM
values with decreasing sweep rate are observed for the four
Pt/Ru catalysts. Over the entire sweep rate range tested, the
lowest Eon values are observed for the ball-mill catalyst. This
indicates that this catalyst has very active and possibly a high
number of OH nucleation sites. The Pt/Ru (H2 reduced)
catalyst is also seen to have low Eon values that are in fact
lower than observed for the PtRu (chemically reduced) cat-
alyst at a particular sweep rate. However, the EFWHM values
are larger for the H2 reduced than for the chemically reduced
catalyst, indicating that the complete COads oxidation reac-
tion is slower for the former. This is attributed to the poorer
distribution of Pt to Ru surface sites of the H2 reduced cata-
lyst, which in turn is believed to result in longer COads dif-
fusion distances. Over the entire sweep rate range studied,
the Pt/RuO2 (thermal) catalyst is seen to clearly exhibit the
poorest COads oxidation kinetics, as reflected in the high Eon
as well as the large EFWHM values.
Pt
0.54b
0.42
0.69
0.49
70
80
5
5
PtRu (chemically
reduced)
Pt/RuO2 (thermal)
Pt/Ru (H2 reduced)
Pt/RuO2 (ball-mill)
0.45
0.43
0.39
0.66
0.51
0.48
320 20
110 10
100 20
a
COads stripping CVs were recorded at 10 mV s−1 in 0.5 M H2SO4.
Onset potential (Eon) estimated using the larger COads stripping peak.
b
Pt- and Ru-precursor salts results in alloyed catalysts consist-
ing of surfaces made of well-distributed Pt to Ru sites [5]. All
the Ru containing catalyst powders used in this work exhibit
at least a partial bi-functional catalytic effect. This is seen in
the lower onset potentials for the COads oxidation reactions
for the four Pt/Ru catalysts in comparison to that observed
for the Ru-free Pt powder catalyst. This beneficial effect is
assigned to the bi-functional mechanism, i.e., the formation
of “Ru OH” species at low potentials shown in Eqs. (1)–(4).
A partial beneficial catalytic effect is also observed for the
case for the Pt/RuO2 (thermal) powder catalyst, although the
COads oxidation peak for this powder is much broader ex-
panding into the more positive potential region than observed
for Pt. This indicates that the complete COads to CO2 oxida-
tion reaction for this particular catalyst is extremely slow.
The lowest onset potential for the COads to CO2 oxidation
reaction is observed for the Pt/RuO2 (ball-mill) catalyst and
the COads stripping peak is narrow. A narrow COads stripping
peak and low COads oxidation potential are also observed for
the Pt/Ru (H2 reduced) catalyst. These results suggest the H2
reduced catalyst has significantly better oxidation kinetics
for the complete COads to CO2 oxidation reaction than the
Pt/RuO2 (thermal) catalyst. This significant difference in the
3.5. Potentiostatic COads oxidation: current–time
transients
COads oxidation current-transients recorded at a constant
potential provide more detailed information about the COads
oxidation kinetics than COads stripping CVs. Current–time
Fig. 4. Dependence of (a) the onset potential (Eon) and (b) the EFWHM values of the COads stripping reaction, respectively, on the log of the sweep rate for the
Pt/Ru catalysts. (ꢀ) shows the data for the PtRu (chemically reduced) catalyst, (ꢀ) for the Pt/RuO2 (thermal) catalyst, (×) for the Pt/Ru (H2 reduced) and (
)
for the Pt/RuO2 (ball-mill) catalyst. The Eon values are reported vs. the RHE.

2410
C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
Fig. 5. COads oxidation transients (i/QCO )–t plots recorded at 0.45 V for (a) the PtRu (chemically reduced) and Pt/RuO2 ball-mill catalysts and (b) the
ads
Pt/Ru(H2 reduced) and Pt/RuO2 (thermal) catalysts .The current axes are normalized using the COads charge extracted from the corresponding COads stripping
voltamograms.
transients for the COads oxidation reaction for Pt/Ru systems
have been reported previously in the literature [19,20]. In
the followingsections, experimentally observedcurrent–time
transients recorded at two particular potentials will be dis-
cussed. In this work, low potentials are selected at which the
Pt OH formation reaction can be ignored, i.e., only Ru sites
are believed to serve as OH nucleation sites. Subsequently,
the data will be analyzed to obtain information about the
OH nucleation rate and the surface diffusion of CO. Ex-
At 0.45 V, complete oxidation of all the CO molecules ad-
sorbed on the catalyst surfaces was found to take ca. 30 s for
the Pt/RuO2 (ball-mill) catalyst, ca. 70 and 90 s for the PtRu
(chemically reduced) and Pt/Ru (H2 reduced) catalysts, re-
spectively, and more than 30 min for the Pt/RuO2 (thermal)
catalyst. These data can be used to obtain an estimate for the
maximal surface diffusion distance of CO, defined as x, using
the square root approximation [21]:
√
x = Dt
(7)
perimental COads oxidation current–time ((i/QCO )–t) tran-
ads
sients are shown in Fig. 5a and b. The (i/QCO )–t transients
were recorded at 0.45 V and the currents are normalized using
In Eq. (7), t is the time and D the CO surface diffusion
coefficient. A D value of 4 × 10−14 cm2 s−1 is used here that
was estimated in previous work based on experimental COads
oxidation data for Pt/Ru catalysts [22]. The 30 s needed for
the complete oxidation of COads using the Pt/RuO2 (ball-
mill) catalyst suggest that the maximal CO surface diffusion
distance on this catalyst surface is ca. 11 nm. This is equiva-
lent to ca. 40 Pt atoms using a lattice constant of 0.3 nm. The
same calculations for the other three Pt/Ru catalysts suggest a
maximal CO surface diffusion distance of ca. 20 nm, equiv-
alent to ca. 65 Pt atoms for the PtRu (chemically reduced)
and the Pt/Ru (H2 reduced) catalysts and more than 90 nm,
i.e., more than 300 Pt atoms for the Pt/RuO2 (thermal) cat-
alyst. These are significant CO surface diffusion distances
for all catalysts, in particular for the ball-mill and chemically
reduced catalysts considering that they presumably consist
of surfaces with homogenously distributed Pt and Ru sites.
In the case of a catalyst surface that truly consists of atomic
level, homogenously distributed Pt to Ru sites of 70–30 at.%
composition, each Pt sites is located next to a Ru site. There-
fore, diffusion of COads to a Ru site is not needed if all Ru
sites are equally active as OH nucleation sites. The sug-
gested surface diffusion of CO over as many as 40 and 65 Pt
ads
the QCO values extracted from the corresponding CV data.
ads
Under these conditions, the COads reaction is seen to be the
most rapid using the Pt/RuO2 (ball-mill) catalyst, followed by
the PtRu (chemically reduced), the Pt/Ru (H2 reduced) and
finally the Pt/RuO2 (thermal) catalyst. The initial high current
rise seen in these transients is due to double layer charging.
This was confirmed by (i/QCO )–t transients recorded in
ads
the absence of CO, i.e., in the H2SO4 background solution.
The asymmetrical, i.e., tailing current decay indicates that
CO surface diffusion to OH nucleation sites is rate deter-
mining in the COads to CO2 oxidation reaction [20]. Such
a tailing current is observed for all four catalysts indicating
that COads surface diffusion plays a significant role in the
COads to CO2 oxidation reaction for all four Pt/Ru catalyst
systems studied here. A similar tailing current of compara-
ble time scale to the PtRu (chemically reduced) catalyst has
been reported in previous work for catalysts of “proposed”
homogenously distributed, at the atomic scale, Pt/Ru surface
sites of 70 and 30 at.% composition [19]. This further em-
phasizes the similarity of the previously studied system [20]
and the chemically reduced PtRu catalyst used in this work.

C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
2411
takes longer than 1 h. Therefore, the (i/QCO )–t transient for
ads
this catalyst is not shown in Fig. 6. The double layer charging
and the COads oxidation currents are clearly separated in the
transients recorded at 0.35 V. At this low potential, the ma-
jority of the Ru sites are believed to be in the reduced state,
i.e., Ru is in the metallic state for the ball-mill, chemically
reduced and the H2 reduced catalysts [9]. For both the chem-
ically reduced and H2 reduced catalysts, the 10 min at 0.35 V
used to record the transients shown in Fig. 6 are insufficient
to oxidize all of the COads molecules. This was confirmed by
subsequent COads stripping voltammetry that indicated that
during the 10 min at 0.35 V, 59 and 32% of the COads were
oxidized in the case of the chemically reduced and the H2
reduced catalysts, respectively. The fact that a larger number
of COads molecules are oxidized for the chemically reduced
than for the H2 reduced catalyst in the 10 min at 0.35 V sug-
gests again that the COads oxidation reaction is overall faster
for the chemically reduced than the poorly distributed Pt to
Ru surface sites of the H2 reduced catalyst.
Fig. 6. COads oxidation transients (i/QCOads )–t plots recorded at 0.35 V for
the PtRu (chemically reduced), the Pt/Ru(H2 reduced) and the Pt/RuO2
(ball-mill) catalyst. The currents are normalized using the corresponding
COads charge (QCO ) extracted from the COads stripping voltamogram
3.6. Analysis of the (i/QCO )–t transients: rate of OH
nucleation site formation
ads
ads
data recorded at 10 mV s−1
.
atoms for the two catalysts (that are believed to be made up of
surfaces of well-distributed Pt and Ru sites) suggests that not
all Ru sites are equally active for the COads oxidation and the
OH formation reaction. This is not unexpected considering
the fact that Ru is well known to show a range of energet-
ically different surface sites towards its oxidation reaction
[23]. Indeed, this results in the close to ideal super-capacitive
behavior of Ru and Ru-oxides [23].
The COads oxidation reaction is viewed to take place ac-
cording to the bi-functional mechanism as follows [2]:
k1
Ru + H2O −→ Ru OH + H+ + e−
(8)
(9)
k2
M − CO + Ru OH −→ M + Ru + CO2 + H+ + e−
In Eq. (9), M is either a Ru or Pt surface site. Accord-
ing to this scheme, “COads-free” Ru sites need to be present
to form Ru OH. This is of particular importance for sur-
faces with high COads coverage. As the COads oxidation re-
action proceeds, an increasingly higher number of Ru sites
are freed from adsorbed CO and become potential OH nu-
cleation sites. This results in an increase in the oxidation
current, which in fact is clearly observed in the experimental
An interesting difference is seen in the (i/QCO )–t tran-
ads
sients between the PtRu (chemically reduced) and the Pt/Ru
(H2 reduced) catalysts for short versus longer times of an-
odic polarization. During the initial few seconds, the COads
oxidation currents for the chemically reduced catalyst are
higher than for the H2 reduced catalyst indicating a faster re-
action. This is clearly seen in (i/QCO )–t transients recorded
ads
i/(QCO )–t transients. In order to oxidize COads to CO2, the
ads
at 0.35 V that are shown in Fig. 6 for the ball-mill, chemi-
cally reduced and H2 reduced catalysts. At this low potential,
an “induction” period (late raise and low magnitude of the
oxidation current) is observed, which is seen to be most pro-
nounced for the chemically reduced catalyst. In fact, the H2
reduced catalyst shows a higher COads oxidation current at
shorter times (during the initial ca. 100 s), while the rate is
lower (compared with the chemically reduced catalyst) at
longer times. This suggests that the H2 reduced catalyst has a
few OH nucleation sites that are more active than the chem-
ically reduced catalyst, which is consistent with the low Eon
values observed in the COads stripping voltamograms. The
initially more rapid formation of OH nucleation sites for
the H2 reduced catalyst, as compared to the chemically re-
duced catalyst, may be linked to the fact that the former has
“large” Ru-islands on its surface. The current intensity of the
Pt/RuO2 (thermal) catalyst at 0.35 V is too low to observe
a maximum and complete removal of COads at this potential
COads must be located close to a OH nucleation site. At a
high COads coverage and relatively low number of OH nu-
cleation sites, a large number of adsorbed CO molecules are
located in close vicinity to a nucleation site. However, as the
COads oxidation reaction proceeds, and if no new OH nu-
cleation sites are generated, CO needs to first diffuse across
the catalyst surface to be subsequently oxidized to CO2. This
COads surface diffusion process involves free catalyst surface
sites, M*, i.e.,
k0
M CO + M −→ M + M CO
(10)
∗
∗
Therefore, the complete COads to CO2 oxidation reaction
can be described as follows: initially, cylindrical CO con-
centration profiles develop around active OH nucleation
sites. The individual diffusion profiles eventually overlap
and rather complicated COads diffusion profiles can develop
depending on the Pt to Ru surface site distribution. If the

2412
C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
diffusion of surface adsorbed CO to a OH nucleation site is
rate determining in the COads to CO2 oxidation reaction, and
if no new or a very small number of new OH nucleation sites
are formed, a decrease in the oxidation current is observed.
This suggests that the increase in the oxidation current in the
tivity for the COads oxidation reaction than the H2 reduced
catalyst despite its better Pt to Ru surface site distribution.
This adds further support to the view that the Ru sites are
not equally active as OH nucleation sites. The origin of the
better COads oxidation kinetics for the ball-mill versus the
chemically reduced catalyst is not yet clear. In both cases,
the Pt to Ru surface sites are believed to be well distributed
and bulk PtRu alloys of similar compositions are formed. The
better COads oxidation kinetics, which are clearly linked to a
more rapid OH nucleation rate for the ball-mill versus the
chemically reduced catalyst, may be linked to the fact that the
Ru surface concentration of the ball-mill catalyst is higher.
Furthermore, ball-milling introduces mechanical stress and
the morphology of these two catalysts will be significantly
different. Therefore, a possibly higher number of surface de-
fect sites may be at least partly responsible for the observed
improved COads oxidation kinetics of the ball-mill catalyst.
It should be noted that electronic effects introduced by
alloying Pt and Ru can also influence the COads oxidation
kinetics. According to ab-initio calculations, alloying of Pt
with Ru weakens the Ru CO bond in the 0.1 eV range, while
increasing the concentration of Ru in the alloy lattice is pro-
posed to weaken the Pt CO bond substantially; differences
of up to 0.5 eV have been reported [24]. The fact that we
observe experimentally that the oxidation reaction of the CO
molecules adsorbed on presumably the Ru sites takes place
on the same time scale for the PtRu alloy and the two phase Pt
and Ru islands, i.e., the Pt/Ru (H2 reduced catalyst) could in-
dicate that the electronic influence of alloying on the Ru CO
bond has a smaller effect than predicted by ab-initio calcu-
lations that involve a number of assumptions. Statements re-
garding the influence of electronic effects of alloying on the
strength of Pt CO bond cannot be made using the catalysts
systems prepared and experiments carried out in this work.
i/(QCO )–t transients represents a measure of the sum of the
ads
formation rate and number of OH nucleation sites as well
as the oxidation rate of CO adsorbed on sites in direct contact
with an active OH nucleation site.
Such a COads oxidation mechanism suggests that the
COads oxidation transients can be used to extract several
valuable characteristic numbers. The initial rising slope of
the current–transient, d(i/QCO )/dt, represents the rate of
ads
the “ OH nucleation” site formation and oxidation of nearby
COads species. The time needed to reach the current max-
ima, t0, represents the period when OH nucleation sites are
formed at a measurable rate. The charge passed to reach this
maximum, Q0, represents a measurement of the number of
OH nucleation sites. It should be noted that the oxidation
of CO molecules located on nearby sites also contributes to
the Q0 values. Therefore, the percentage of the Q0 per to-
tal QCO value, i.e., the Q0 (%) value reflects the number
ads
of OH nucleation sites formed and the number of COads
molecules in close vicinity to such sites, and can be used
to compare different catalysts. One characteristic of a good
COads to CO2 oxidation catalyst is its capability to form a
high number of OH nucleation sites and oxidize a large
number of COads molecules within a short time period. This
is reflected in a large Q0 (%)/t0 ratio. Values for t0, Q0 (%)
and Q0 (%)/t0 are summarized in Table 4. The values are
extracted from the (i/QCO )–t transients at 0.45 and 0.35 V
ads
for the four Pt/Ru catalyst powders studied in this work. The
Q0 (%)/t0 values are seen to decay in the following order
at both potentials: Pt/RuO2 (ball-mill) ꢁ PtRu (chemically
reduced) > Pt/Ru (H2 reduced) ꢁ Pt/RuO2 (thermal). This
order represents the capability of a catalyst to form OH
nucleation sites and oxidize COads molecules, i.e., suggest-
ing that the ball-mill catalyst is superior to the chemically
and H2 reduced catalysts, and that the thermal catalyst has a
very poor capability to form OH nucleation sites. In fact, the
3.7. CH3OH oxidation studies
Fig. 7 shows Tafel plots, i.e., the potential (E) versus
the logarithms of the CH3OH oxidation current normal-
ized for the electro-active Pt area (jCH OH), obtained at
3
(i/QCO )–t transients recorded at 0.35 V for the thermal cat-
20 ◦C for the Pt and Pt/Ru powder catalysts studied in this
work. The CH3OH oxidation currents were obtained from
current–time transients recorded at a particular potential. In
all cases, a decay of the oxidation current was observed that
approached a pseudo steady-state value after several min-
utes. The CH3OH oxidation current values used for the plots
ads
alyst did not show a current maxima and the oxidation charge
measured during this period of 10 min was very small, thus
indicating a very poor capability of this catalyst to produce
OH nucleation sites at this potential. The PtRu (chemically
reduced) catalyst appears to display only a slightly better ac-
Table 4
Characteristic numbers for the COads oxidation reaction carried out at a particular potential for Pt/Ru catalysts
Catalyst
t0 (s−1
)
Q0 (%)
Q0/t0 (% s−1
)
t0 (s−1
)
Q0 (%)
Q0/t0 (% s−1
)
at 0.45 V
at 0.45 V
at 0.45 V
at 0.35 V
at 0.35 V
at 0.35 V
PtRu (chemically reduced)
Pt/RuO2 (thermal)
Pt/Ru (H2 reduced)
Pt/RuO2 (ball-mill)
5.5
5
8
13
3.5
17
10
2.4
0.7
2.1
8.3
260
n.a.a
300
80
11
0.04
n.a.a
0.03
0.2
n.a.a
10
1.2
14
Ratio vs. the corresponding t0 value found for the ball-mill catalyst.
a
No maxima observed.

C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
2413
essentially the same. They are ca. seven times higher than the
CH3OH oxidation activity for the H2 reduced and the thermal
catalysts over the entire potential range investigated in this
work. The activity of the H2 reduced and thermal catalysts
are essentially the same. The activity of the ball-mill and the
chemically reduced catalysts are ca. 30 times higher than for
the Pt catalyst, which is in good agreement with literature
data that showed a ca. 30 times increase in catalytic activity
for sputter-cleaned bulk PtRu alloy catalysts of 70–30 at.% Pt
to Ru composition over polycrystalline platinum [5]. Fig. 7
shows that all four Pt/Ru catalysts show better CH3OH ox-
idation activities than the Pt powder, thus indicating that all
four Pt/Ru catalysts exhibit a beneficial catalytic effect as
also seen for the COads oxidation reaction. However, the type
of Pt/Ru catalyst is seen to influence the CH3OH and COads
oxidation reactions differently. In fact, the CH3OH oxidation
activities can be placed in two groups; namely, essentially the
same activity for the ball-mill and chemically reduced cata-
lysts and essentially the same activity for the H2 reduced and
thermal catalysts. The clearly better CH3OH oxidation activ-
ity for the ball-mill and chemically reduced catalysts over the
H2 reduced and thermal catalysts is suggested to be due to the
better distribution of Pt to Ru sites on the surface of the first
two catalysts in the present work. In fact, it appears that the
CH3OH activity is determined by the optimal number of Pt
and nearby Ru sites that is clearly better for catalysts that are
alloyed and consist of surfaces made of well-distributed Pt to
Ru sites rather than the oxidation state of the Ru ad-metal of
the as-prepared catalyst. It should be noted that in this work,
the catalytic activities are compared to the Ru ad-metal state
of the as-prepared catalysts and cannot be correlated to the
most active ad-metal state of Ru during the reaction. It is very
well possible that Ru-oxides of these Pt/Ru catalysts are at
least partially reduced to Ru metal in the electrochemical en-
vironment, as has been indeed shown in previous work [8,9].
Fig. 7. CH3OH oxidation current (jCH OH)–potential (E) curves for (♦) Pt,
3
(ꢀ), PtRu (chemically reduced), ( ) Pt/RuO2 (ball-mill), (×) Pt/Ru (H2 re-
duced) and (ꢀ) Pt/RuO2 (thermal) catalysts. The CH3OH oxidation currents
are pseudo-steady state values that were extracted after applying a particular
potential for 20 min.
shown in Fig. 7 were extracted from the pseudo steady-state
values, i.e., 20 min after applying a particular potential. Prior
to the recording of each CH3OH oxidation current–time tran-
sients, the potential was held at 0.1 V for 30 s. This allowed
the recording of reproducible current–time transients, as dis-
cussed in detail elsewhere [5]. The amount of catalyst de-
posited on the Au substrate was adjusted depending on the po-
tential range and catalyst investigated. In all cases, electrodes
were prepared that resulted in CH3OH oxidation currents of
less than 10 mA. At such low currents, minimal amounts of
H2 are produced at the counter electrode, and hence a possible
interference of this with the anode reaction can be ruled out.
The dependence of the CH3OH oxidation current on the
potential is essentially the same for the four Pt/Ru catalysts.
For all four Pt/Ru powder catalysts, two regions with Tafel
4. Summary and conclusions
slopes of a ca. 120 mV per log (jCH OH) and an anomalous
A range of unsupported Pt/Ru catalyst powders of the
same nominal Pt to Ru ratios of 70–30 at.% were prepared
and characterized. A PtRu alloy powder catalyst that is
viewed to be of homogenous composition and the same bulk
and surface Pt and Ru concentration was made via a very
rapid chemical reduction method. Two non-homogenous
Pt/Ru powders were made that consist of a non-alloyed
Pt phase and a surface enriched with Ru. The first powder
was made by thermally decomposing RuCl3 on Pt powder
and is referred to as thermal catalyst. Part of this powder
was then reduced in a H2 atmosphere resulting in the
second non-homogenous Pt/Ru powder referred to as H2
reduced catalyst. The form of the Ru ad-metal of these two
as-prepared catalysts is different, i.e., RuO2 and lower, easily
reducible lower Ru-oxides for the thermal catalyst powder,
and mainly Ru metal and easily reducible lower Ru-oxides
for the H2 reduced catalyst powder. A Pt/Ru powder catalyst
3
slope of ca. 200 mV per log (jCH OH) are observed. The Tafel
3
slopes of ca. 120 mV per log (jCH OH) are observed at lower
3
potentials and indicate that an initial one electron electro-
chemical reaction is rate determining in the CH3OH oxida-
tion reaction. At these low potentials, the OH formation
reaction on Ru sites (i.e., Eq. (3)) is generally viewed to
be rate determining which is consistent with a 120 mV per
log (jCH OH) Tafel slope. The origin of the anomalous, larger
3
than 120 mV per log (jCH OH) slope is not understood and be-
3
yond the scope of this work. It is, however, in agreement with
previously reported values [25]. In Fig. 7, the CH3OH oxi-
dation currents are normalized for the Pt area using the fA
Pt
factors as discussed in a previous section of this work. There-
fore, the CH3OH oxidation data shown in Fig. 7 represent true
catalytic activities. It is seen that the CH3OH oxidation ac-
tivity for the chemically reduced and ball-mill catalysts are

2414
C. Bock et al. / Electrochimica Acta 50 (2005) 2401–2414
was also prepared by ball-milling Pt and RuO2 resulting
in a bulk PtRu alloy that appears, however, to be surface
enriched in Ru. The surface Ru was found to be present as
Ru metal and lower, easily reducible Ru-oxides.
Acknowledgements
The authors thank G. Pleizier and D. Kingston (NRC, Ot-
tawa) for the XPS analysis and J. Margeson (NRC, Ottawa)
for the FESEM and EDX analyses as well as P. L’Abbe (NRC,
Ottawa) for the preparation of the electrochemical glass cells
used in this work. Financial support from the NRC fuel cell
program is also greatly acknowledged.
The catalysts were used to study the COads and CH3OH
oxidation reactions. Their catalytic activities were compared
to the properties of the as-prepared catalysts. The COads ox-
idation reaction was studied in a potential range where the
contribution of OH formation on Pt sites can be assumed to
be very small. It has been shown that several characteristics
can be extracted from COads oxidation experiments that allow
the comparison of these complex catalyst systems with regard
to OH nucleation rates and Pt to Ru site distributions. It was
found that there is no direct correlation between the capabil-
ities of these catalysts to oxidize COads and CH3OH. In the
case of the COads oxidation reaction, the presence of RuO2 in
combination with poorly distributed Pt to Ru sites on the cata-
lyst surface was found to result in a significantly slower COads
oxidation reaction. The formation of active OH species on
the Ru sites are needed for the COads to CO2 oxidation re-
action according to the bi-functional mechanism, Eqs. (8)
and (9) was very low for the catalyst containing RuO2 on
its surface. The presence of Ru in the metallic state on the
as-prepared catalyst surface clearly resulted in faster COads
oxidation reaction kinetics and higher OH nucleation rates.
The fastest COads oxidation kinetics were found for catalysts
made of well-distributed Pt to Ru surface sites. This is at-
tributed to the smaller surface diffusion distances needed for
the CO to diffuse to an active OH nucleation site on these
catalyst surfaces consisting of well-distributed Pt to Ru sites.
For all catalysts, asymmetrical COads oxidation current–time
transients were observed indicating that surface diffusion of
CO to an OH nucleation site is rate determining. The max-
imal COads diffusion lengths were estimated and long COads
diffusion distances were found, even for the catalyst surfaces
consisting of well-distributed Pt to Ru sites. The major fac-
tor determining the CH3OH oxidation activity was found to
be the Pt to Ru surface site distribution. Highest CH3OH
oxidation activities were found for the Pt/Ru catalysts that
consisted of well-distributed Pt to Ru surface sites, while the
oxidation state of the Ru ad-metal of the as-prepared cata-
lyst appears not to influence the catalytic CH3OH oxidation
activity. This suggests that an optimal arrangement of Pt and
nearby Ru sites is essential for the CH3OH oxidation reaction
rather than the Ru surface state of the as-prepared catalysts.
The observed kinetic differences for these catalysts to oxi-
dize COads and CH3OH may be partly related to the fact that
in the former case, CO is adsorbed on both Pt and Ru sites,
while CH3OH adsorbs from solution onto Pt sites only.
References
[1] J.O’.M. Bockris, H. Wroblowa, J. Electroanal. Chem. 7 (1964) 428.
[2] M. Watanabe, S. Motoo, J. Electroanal. Chem. 60 (1975) 267.
[3] H. Gasteiger, N. Markovic, P. Ross, E. Cairns, J. Electrochem. Soc.
141 (1994) 1795.
[4] H. Hoster, T. Iwasita, H. Baumgartner, W. Vielstich, Phys. Chem.
Chem. Phys. 3 (2001) 337.
[5] C. Bock, B. MacDougall, Y. LePage, J. Electrochem. Soc. 151 (2004)
A1269.
[6] J.W. Long, R.M. Stroud, K.E. Swider-Lyons, D.R. Rolison, J. Phys.
Chem. B 104 (2000) 9772.
[7] H.N. Dinh, X. Ren, F.H. Garzon, P. Zelenay, S. Gottesfeld, J. Elec-
troanal Chem. 491 (2000) 222.
[8] H. Kim, I.R. deMoraes, G. Tremiliosi, R. Haasch, A. Wieckowski,
Surf. Sci. 474 (2001) L203.
[9] C. Bock, B. MacDougall, J. Electrochem. Soc. 150 (2003) 377.
[10] K. Lasch, L. Joerisson, K.A. Friedrich, J. Garche, J Solid State
Electrochem. 7 (2003) 619.
[11] A.H.C. Sirk, J.M. Hill, S.K.Y. Kung, V.I. Birss, J. Phys. Chem. B
108 (2004) 689.
[12] M.C. Denis, P. Gouerec, D. Guay, J.P. Dodelet, G. Lalande, R.
Schulz, J. Appl. Electrochem. 30 (2000) 1243.
[13] A.J. Bard, L.R. Faulkner, Electrochemical Methods, Fundamentals
and Applications, John Wiley & Sons, 1980.
[14] X-ray Diffraction Database, Diffraction Management System for
Windows NT, Scintag Inc., Cupertina, California (1997).
[15] L.X. Yang, C. Bock, B. MacDougall, J. Park, J. Appl. Electrochem.
34 (2004) 427.
[16] G. Lalande, M.C. Denis, P. Gouerec, D. Guay, J.P. Dodelet, R.
Schulz, J. New Mater. Electrochem. Syst. 3 (2000) 185.
[17] T.A.F. Lassali, J.F.C. Boodts, S.C. DeCastro, R. Landers, S. Trasatti,
Electrochim. Acta 39 (1994) 95.
[18] H.A. Gasteiger, P.N. Ross, E.J. Cairns, J. Surf. Sci. 293 (1993) 67.
[19] H.A. Gasteiger, N. Markovic, P.N. Ross, E.J. Cairns, J. Phys. Chem.
98 (1994) 617.
[20] F. Maillard, M. Eikerling, O.V. Cherstiouk, S. Schreier, E. Savinova,
U. Stimming, Faraday Discuss. 125 (2004) 357.
[21] J.O’M. Bockris, A.K.N. Reddy, Modern Electrochemistry, vol. 1, A
Plenum/Rosetta ed., Plenum Press, New York, 1970.
[22] K.A. Friedrich, K.P. Geyzers, A. Marmann, U. Stimming, R. Vogel,
Zeitschrift fuer Physikalische Chemie 208 (1999) 137.
[23] S. Hadzi-Jordanov, H. Angerstein-Kozlowska, B.E. Conway, J. Elec-
troanal. Chem. 60 (1975) 359.
[24] T.E. Shubina, M.T.M. Koper, Electrochim. Acta 47 (2002) 3621.
[25] W. Chrzanowski, A. Wieckowski, Langmuir 14 (1998).