The effect of the particle size on the kinetics of CO electrooxidation on high surface area Pt catalysts

Article abstract of DOI:10.1021/ja043602h

Using high-resolution transmission electron microscopy (TEM), infrared reflection-absorption spectroscopy (IRAS), and electrochemical (EC) measurements, platinum nanoparticles ranging in size from 1 to 30 nm are characterized and their catalytic activity for CO electrooxidation is evaluated. TEM analysis reveals that Pt crystallites are not perfect cubooctahedrons, and that large particles have rougher surfaces than small particles, which have some fairly smooth (111) facets. The importance of defect sites for the catalytic properties of nanoparticles is probed in IRAS experiments by monitoring how the vibrational frequencies of atop CO (ν_CO) as well as the concomitant development of dissolved CO ₂ are affected by the number of defects on the Pt nanoparticles. It is found that defects play a significant role in CO clustering on nanoparticles, causing CO to decrease/increase in local coverage, which yields to anomalous redshift/ blueshift ν_CO frequency deviations from the normal Stark-tuning behavior. The observed deviations are accompanied by CO₂ production, which increases by increasing the number of defects on the nanoparticles, that is, 1 ≤ 2 < 5 ? 30 nm. We suggest that the catalytic activity for CO adlayer oxidation is predominantly influenced by the ability of the surface to dissociate water and to form OH_ad on defect sites rather than by CO energetics. These results are complemented by chronoamperometric and rotating disk electrode (RDE) data. In contrast to CO stripping experiments, we found that in the backsweep of CO bulk oxidation, the activity increases with decreasing particle size, that is, with increasing oxophilicity of the particles.

Full text of DOI:10.1021/ja043602h

Published on Web 04/16/2005
The Effect of the Particle Size on the Kinetics of CO
Electrooxidation on High Surface Area Pt Catalysts
Matthias Arenz,*,†,§ Karl J. J. Mayrhofer,† Vojislav Stamenkovic,†
Berislav B. Blizanac,† Tada Tomoyuki,‡ Phil N. Ross,† and Nenad M. Markovic†
Contribution from the Materials Science DiVision, Lawrence Berkeley National Laboratory,
UniVersity of California, Berkeley, California 94720, and Tanaka Kikinzoku Kogyo K. K.
Cooperation (TKK), Tokyo, Japan
Received October 21, 2004; E-mail: matthias.arenz@ch.tum.de
Abstract: Using high-resolution transmission electron microscopy (TEM), infrared reflection-absorption
spectroscopy (IRAS), and electrochemical (EC) measurements, platinum nanoparticles ranging in size from
1 to 30 nm are characterized and their catalytic activity for CO electrooxidation is evaluated. TEM analysis
reveals that Pt crystallites are not perfect cubooctahedrons, and that large particles have “rougher” surfaces
than small particles, which have some fairly smooth (111) facets. The importance of “defect” sites for the
catalytic properties of nanoparticles is probed in IRAS experiments by monitoring how the vibrational
frequencies of atop CO (νCO) as well as the concomitant development of dissolved CO2 are affected by the
number of defects on the Pt nanoparticles. It is found that defects play a significant role in CO “clustering”
on nanoparticles, causing CO to decrease/increase in local coverage, which yields to anomalous redshift/
blueshift νCO frequency deviations from the normal Stark-tuning behavior. The observed deviations are
accompanied by CO2 production, which increases by increasing the number of defects on the nanoparticles,
that is, 1 e 2 < 5 , 30 nm. We suggest that the catalytic activity for CO adlayer oxidation is predominantly
influenced by the ability of the surface to dissociate water and to form OHad on defect sites rather than by
CO energetics. These results are complemented by chronoamperometric and rotating disk electrode (RDE)
data. In contrast to CO stripping experiments, we found that in the backsweep of CO bulk oxidation, the
activity increases with decreasing particle size, that is, with increasing oxophilicity of the particles.
1. Introduction
surface, then well-characterized single-crystal surfaces may be
used as reasonable models.3 Many research groups have used
this strategy to predict what the particle size might be, that is,
the variation of the reaction rate or selectivity with the
characteristic dimension of metallic clusters at the electrified
solid-liquid interface. For example, given the well-documented
sensitivity of the vibrational properties of adsorbed CO to its
chemical and electrostatic environment, infrared reflection-
absorption spectroscopy (IRAS) of CO adsorbed on ordered Pt
single crystals has been used to understand how the properties
of Pt nanoparticles depend on their size.4-6 Further insight into
the particle size effect has been obtained by electrochemical-
based experiments, for example, monitoring CO adlayer oxida-
tion, dubbed “CO stripping”, on Pt nanoparticles in the size
range of 1-10 nm.7-10
The size-dependent catalytic properties of metal nanoparticles
play a vital role in emerging technologies for environmental
and energy related applications, such as fuel cells. The
characteristic dimension of metal clusters employed in fuel cells
is in the size range of a few nanometers. Because the electronic
properties of metal aggregates in this size range do not differ
from those of the bulk metal,1,2 catalyst research of such
materials has focused on the variation of the surface structure
of metal aggregates with their size, that is, how the proportions
of different types of adsorption sites (facets) vary with dimension
for a given structure. There is no simple ideal structure that
will necessarily model all the aspects of nanoparticle catalysts,
particularly in the exact configuration as they are used in
electrolytic cells. However, if the equilibrium shape of a
nanoparticle, as is the case for the fcc metal Pt, is a cubo-
octahedral structure consisting of (111) and (100) facets bound
by edge atom rows that are like the topmost rows of the (110)
(3) Markovic, N. M.; Ross, J. Surf. Sci. Rep. 2002, 45, 117-229.
(4) Chang, S. C.; Roth, J. D.; Ho, Y.; Weaver, M. J. J. Electron Spectrosc.
Relat. Phenom. 1990, 54-55, 1185-1203.
(5) Lebedeva, N. P.; Rodes, A.; Feliu, J. M.; Koper, M. T. M.; van Santen, R.
A. J. Phys. Chem. B 2002, 106, 9863-9872.
(6) Markovic, N. M.; Ross, P. N. Electrochim. Acta 2000, 45, 4101-4115.
(7) Friedrich, K. A.; Henglein, F.; Stimming, U.; Unkauf, W. Colloids Surf.,
A 1998, 134, 193-206.
† University of California, Berkeley.
‡ TKK.
(8) Cherstiouk, O. V.; Simonov, P. A.; Savinova, E. R. Electrochim. Acta 2003,
48, 3851-3860.
§ Present address: Chair Physical Chemistry I, Chemsitry Department,
Technical University, Munich, Germany.
(1) Henry, C. R. Surf. Sci. Rep. 1998, 31, 231-325.
(2) Somorjai, G. A. Introduction to Surface Chemistry and Catalysis; John
Wiley & Sons: New York, 1993.
(9) Maillard, F.; Savinova, E. R.; Simonov, P. A.; Zaikovskii, V. I.; Stimming,
U. J. Phys. Chem. B 2004, 108, 17893-17904.
(10) Maillard, F.; Eikerling, M.; Cherstiouk, O. V.; Schreier, S.; Savinova, E.
R.; Stimming, U. Faraday Discuss. 2004, 357-377.
9
10.1021/ja043602h CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 6819-6829
6819

A R T I C L E S
Arenz et al.
were carbon-supported Pt nanoparticles with mean diameters of 1-1.5,
2-3, and 5 nm, respectively (analysis by TKK). The fourth sample,
consisting of a nanostructured Pt film supported on crystalline organic
whiskers,20 was supplied by 3M Company (St. Paul, MN U.S.A.). The
particle size of the latter sample was roughly estimated to be about
30 nm based on the charge required to form a full monolayer of
adsorbed hydrogen (the so-called Hupd charge). In the following, the
catalysts will simply be denoted as 1, 2, 5, and 30 nm catalysts.
Electrochemical Measurements. The catalyst preparation has been
described previously.21 In short, the catalyst was dispersed ultrasonically
in ultrapure water, and 20 µL of the suspension was pipetted onto a
glassy carbon substrate (0.283 cm2 geometrical surface area) leading
to a Pt loading of 14 µgPt/cm2 for the carbon-supported catalysts
(TKK 1, 2, and 5 nm) and 42 µgPt/cm2 for the 3M (30 nm) sample.
The dried (Ar atmosphere) catalyst film was attached to the substrate
by a thin Nafion film. The thus prepared surface was then transferred
to the electrochemical cell protected by a drop of ultrapure water,
immersed under potential control at 0.05 V in argon-saturated solution,
and a cyclic voltammogram was recorded.
Two different procedures were used to form and oxidize a saturated
CO adlayer on the Pt nanoparticles. In the first procedure, denoted
hereafter as “oxide-annealing” (method I), before CO is adsorbed at
0.05 V, the catalyst was cycled in Ar-saturated solution (Air Products
5N5 purity) until a well-established cyclic voltammogram was observed.
After forming a saturated CO adlayer (holding at 0.05V for 10 min in
CO-saturated solution), the electrolyte was again purged with Ar
(30 min), that is, the CO stripping curves are recorded in CO-free
solution. In the second procedure, denoted hereafter as “CO-annealing”
(method II), before adsorbing CO at 0.05 V and replacing CO by Ar,
the electrode was “annealed” in CO-saturated solution by cycling of
the electrode potential between 0.05 < E < 1.2 V for 5 min. We used
this experimental approach because we recently found that the
CO-annealing pretreatment of Pt(111) may facilitate the removal of
surface irregularities, which after the flame annealing preparation
method are inherently present on the Pt(111) surface. As a consequence,
upon CO-annealing of the (111) surface, the domain size of the
p(2 × 2)-3CO structure, which is formed on Pt(111) at low potentials,
is significantly improved.3 As we demonstrate below, the catalytic and
spectroscopic properties of high surface area catalysts pretreated by
these two methods are completely different. Note, however, that the
pretreatment is reversible, that is, the particles do not agglomerate during
CO annealing. The stripping curves were recorded with a scan rate of
1 mV/s, and the upper potential limit was set to 1.0 V to guarantee the
complete oxidation of the CO adlayer. Before the potentiodynamic
measurements for the electrooxidation of CO gas dissolved in the
electrolyte (CO bulk) were performed, the electrolyte was saturated
by CO (1 atm) for 25 min while the electrode potential was held at
0.05 V. In chronoamperometric measurements, the electrooxidation rate
of a saturated CO adlayer was measured in CO-free solution at a
constant potential in order to further investigate the phenomena observed
in stripping voltammetry. For each experiment, CO was adsorbed at
50 mV for 10 min under 1600 rpm, and then the CO current transients
were recorded in Ar-saturated (15 min) solutions. The oxidation
potential was chosen to be either in the preoxidation potential region,
where depending on the electrode potential, the CO adlayer was
completely or partially oxidized (see Supporting Information), or
at/above the ignition potential, where CO was oxidized completely.
Subsequently, CO stripping voltammetry was recorded in order to assess
whether the oxidation of CO was complete. After several cycles between
0.05 and 1.1 V, the above potential-step experiment was repeated in
the absence of adsorbed CO to record the current transients due to the
Although this approach did provide some new insight into
the structure sensitivity of CO oxidation,7,9-16 in many cases,
this tactic has also showed significant weaknesses. Probably
the most notable example that the catalytic activities observed
on single crystal surfaces cannot be used as specific one-to-
one models for real nanoparticles has been CO monolayer
oxidation (CO stripping) on Pt nanoparticles in acid electrolytes.
In particular, it has been found that, although more oxophilic,
smaller particles are less active than larger particles.10 It was
proposed that the reason for this observation is a stronger
bonding of CO to the surface of smaller particles and a
concomitant decrease in CO diffusion.7,9,10,14,15
An interesting yet largely unexplored issue in structure
sensitivity is the importance of “defect” sites, which are
inherently present on Pt single crystals3,17 as well as on Pt
nanoparticles.18 For the onset of CO oxidation on Pt(hkl) (which
is observed in the so-called “pre-ignition” potential region), we
suggested that the kinetics are, in fact, determined predominately
by the availability of defect sites on which adsorption of
oxygenated species (hereafter denoted as OH) may occur.3,17
Two other pieces of evidence point to the importance of defects
in the surface chemistry of CO on Pt(hkl). The first is that
domains of the observed surface structures of CO on Pt(111)
are controlled by surface defects, that is, on less defected
surfaces, larger domains of the p(2 × 2)-3CO structure11 are
observed. The second is that the potential-dependent vibrational
behavior of CO (Stark-tuning) can be controlled by the number
of surface defects, that is, a less defected surface exhibits an
“anomalous” Stark-tuning slope.19 The importance of defects
in electrochemical reactions on Pt nanoparticles, however, has
rarely been discussed. One exception is the work by Sattler and
Ross18 who showed that Pt crystallites are not perfect cubo-
octahedrons and that real particles have rough surfaces, which
provide a variety of different sites for catalysis.
Given the structural as well as spectral richness of the Pt-
CO system, the examination of the effect of defects on Pt
nanoparticles on the CO surface chemistry is of interest. Such
a study of Pt nanoparticles in the size range of 1-30 nm is
reported here. We demonstrate how the particle size-dependent
number of defects in cubooctahedral particles affects the
vibrational properties of CO and the CO2 production during CO
oxidation. We also consider how the electrocatalytic activity
for CO oxidation changes with the experimental condition, such
as an excess of OHad or CO on the catalyst surface.
2. Experimental Section
Catalyst Samples. Four different Pt high surface area catalysts were
used in this study. Three samples, supplied by TKK (Tokyo, Japan),
(11) Lucas, C. A.; Markovic, N. M.; Ross, P. N. Surf. Sci. 1999, 425, L381-
L386.
(12) Markovic, N. M. The hydrogen electrode reaction and the electrooxidation
of CO and H2/CO mixtures on well-characterized Pt and Pt-bimetallic
surfaces. In Handbook of Fuel Cell Technology; Vielstich, W., Lamm, A.,
Gasteiger, H., Eds.; John Wiley & Sons Ltd.: New York, 2002.
(13) Markovic, N. M.; Grgur, B. N.; Lucas, C. A.; Ross, P. N. Surf. Sci. 1997,
384, L805-L814.
(14) Friedrich, K. A.; Henglein, F.; Stimming, U.; Unkauf, W. Electrochim.
Acta 2000, 45, 3283-3293.
(15) Friedrich, K. A.; Henglein, F.; Stimming, U.; Unkauf, W. Electrochim.
Acta 2001, 47, 689-694.
(16) Cherstiouk, O. V.; Simonov, P. A.; Zaikovskii, V. I.; Savinova, E. R. J.
Electroanal. Chem. 2003, 554, 241-251.
(20) Debe, M. K. Novel catalysts, catalysts support and catalysts coated
membrane methods. In Handbook of Fuel Cells: Fundamentals, Technology
and Applications; Vielstich, W., Lamm, A., Gasteiger, H., Eds.; John Wiley
& Sons: New York, 2002; pp 576-589.
(21) Schmidt, T. J.; Gasteiger, H. A.; Stab, G. D.; Urban, P. M.; Kolb, D. M.;
Behm, R. J. J. Electrochem. Soc. 1998, 145, 2354-2358.
(17) Markovic, N. M.; Schmidt, T. J.; Grgur, B. N.; Gasteiger, H. A.; Behm, R.
J.; Ross, P. N. J. Phys. Chem. B 1999, 103, 8568-8577.
(18) Sattler, M. L.; Ross, P. N. Ultramicroscopy 1986, 20, 21-28.
(19) Stamenkovic, V.; Chou, K. C.; Somorjai, G. A.; Ross, P. N., Jr.; Markovic,
N. M. J. Phys. Chem. B 2005, 109, 678-680.
9
6820 J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005

Effect of Particle Size on Kinetics of CO Electrooxidation
A R T I C L E S
pseudocapacitance of the electrode, that is, the so-called double layer
charging curves. To assess the current transients that correspond
exclusively to CO oxidation, the double layer charging curves were
subtracted from the transients recorded in the first sets of experiments
(see Supporting Information).
All electrochemical measurements were conducted in a standard
three-compartment electrochemical cell. The reference electrode was
a saturated calomel electrode (SCE) separated by an electrolytic bridge
from the reference compartment. All potentials, however, are referenced
to the potential of the reversible hydrogen electrode (RHE) calibrated
from the hydrogen oxidation/reduction reaction measured in a rotating
ring disk configuration in the same electrolyte. As counter electrode, a
Pt mesh was used. The electrolyte was prepared using pyrolitically triply
distilled water and concentrated HClO4 (Aldrich, double distilled).
Infrared Spectroscopy. For measuring FTIR spectra on supported
nanoparticles, we applied a new methodology, which has been recently
described in detail in a separate paper.22 In short, as for the electro-
chemical measurements, the catalyst was dispersed ultrasonically in
water, and about 20 µL of the suspension was pipetted onto a heated
(∼130 °C) gold substrate (0.785 cm2 geometrical surface area), cooled
in an argon stream, and rinsed carefully with ultrapure water to remove
loosely bound catalyst particles. No further attachment to the substrate
was necessary. For the in situ FTIR measurements, a Nicolet Nexus
670 spectrometer was available equipped with a liquid N2-cooled MCT
detector. All IR measurements were performed in a spectroelectro-
chemical glass cell designed for an external reflection mode in a thin
layer configuration.23 The cell is coupled at its bottom with a CaF2
prism beveled at 60° from the prism base. Prior to each experiment,
the solution was saturated with argon. The electrode was immersed
into the electrolyte at a potential of 0.05 V and a CV recorded. For the
CO stripping measurements recorded in the spectroelectrochemical cell,
the above-described two different adsorption procedures for obtaining
a saturated CO adlayer, that is, oxide annealing and CO-annealing, were
applied. Subsequently, all CO was removed from the solution by
purging with argon for 20 min before pressing the sample onto the
prism. Starting at 0.05 V, the potential was scanned with a scan rate of
1 mV/s in positive direction while continually recording spectra. To
obtain a single spectrum, four interferometer scans were co-added. The
recording time was thus reduced to ca. 2.5 s per spectrum. The
resolution of the spectra was 4 cm-1, and p-polarized light was used.
Absorbance spectra were calculated as the ratio -log(R/R0), where R
and R0 are the reflectance values corresponding to the sample and
reference spectra, respectively. Reference spectra were recorded at either
0.95 V, where adsorbed CO (COad) is completely oxidized, or at 0.05
V, well before the onset of COad oxidation. The potential in the
spectroelectrochemical cell was controlled by an SCE separated by a
glass frit from the main compartment. All potentials, however, are
referenced to the RHE calibrated from the hydrogen evolution/oxidation
reaction in a separated cell using the same electrolyte.
Figure 1. Comparison of CO stripping curves in 0.1 M HClO4 solution
purged after CO adsorption with argon for 30 min at 0.05 V; T ∼ 20 °C,
scan rate 1 mV/s; dashed gray curves show the respective base voltammo-
gram (50 mV/s, currents multiplied by 1/50); the blue curves show the CO
stripping curves after oxide-annealing (method I, see text) and the red curves
after CO-annealing (method II, see text).
TEM images illustrate that the distribution of metal particles
on the carbon support is rather uniform.
Histograms of the particle size distribution (not shown), which
included analysis of several different regions of the catalysts,
established that the average particle size of Pt supported on
carbon is ca. 1, 2, and 5 nm. In agreement with Sattler and
Ross,9 lattice structure characterization obtained by high-
resolution electron microscopy (HRTEM) revealed that although
the basic shape of particles is cubooctahedral, the faces of the
crystal are quite “rough” and “rounded” (i.e., as demonstrated
in Figure 2b; the single crystal planes terminate in the surface
in an irregular way, creating steps; furthermore, there are
occasionally twinned particles which have an irregular surface).
The presence of such “irregularities” on larger Pt particles
(ca. ∼30 nm) is even more obvious. The HRTEM image,
depicted in Figure 2c, reveals the presence of defects, such as
grain bounders and/or twins. From the TEM analysis, it appears
that irregular surfaces of nanoparticles are the rule rather than
the exception and, consequently, may play an important role in
the observed particle size effect of the CO oxidation reaction.
Before we discuss the CO oxidation reaction, however, it
would be appropriate to introduce the observed cyclic volt-
3. Results and Discussion
3.1 Surface Characterization and Electrochemistry. Figure
1 shows representative low-magnification TEM images of Pt
catalysts supported on carbon (Figure 1a-c) and on crystalline
organic whiskers (Figure 1d). Although the nature of the support
for the 30 nm particles is different, the catalytic properties of
these particles are not affected by the nature of organic whiskers,
as evidenced by comparing the catalytic activity of the oxygen
reduction reaction on this catalyst and other bimetallic catalysts
supported on whiskers to corresponding bulk electrodes. The
(22) Stamenkovic, V.; Arenz, M.; Ross, P. N.; Markovic, N. M. J. Phys. Chem.
B. 2004, 108, 17915-17920.
(23) Iwasita, T.; Nart, F. C. Prog. Surf. Sci. 1997, 55, 271-340.
9
J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005 6821

A R T I C L E S
Arenz et al.
COad + OHad f CO2 + H+ + e-
(2)
which was previously proposed for extended surfaces;3,27 the
CO oxidation rate should increase in the same order as the
oxophilicity of the Pt particles. In line with previous findings,
however, Figure 1 unambiguously shows that the anodic CO
stripping peak from 30 nm particles occurs at lower potentials
compared to that of the particles with a size between 1 and
5 nm. A reason for this observation is still unclear, although
Friedrich et al.7,14,15 proposed that it originates from a stronger
bonding of CO to the surface of smaller particles and a
concomitant decrease in CO diffusion.
The puzzling particle size effect is further complicated by
the observed difference in the CO stripping curves recorded on
oxide-annealed and CO-annealed surfaces. Figure 1 shows that
two different potential regions can be distinguished for both
surfaces: (i) the so-called pre-ignition potential region (0.05 <
E < 0.55 V), where only small currents are observed (especially
on the electrode prepared by the CO-annealing method); and
(ii) the potential region where sharp stripping peaks are observed
in the cyclic voltammogram (E > 0.55 V). Further inspection
of Figure 1 reveals that the CO stripping peak is shifted
positively on the CO-annealed surfaces, consistent with the
supposition that CO-annealed surfaces contain fewer irregulari-
ties than the respective surfaces that were never pretreated in
CO-saturated solution. Finally, the CO stripping curves in Figure
1 indicate that, independent from the preparation method, the
most active Pt catalyst (based on the position of the stripping
peaks) is the one consisting of 30 nm particles. Interestingly,
in contrast to the results in ref 10, the difference in the position
of the stripping peaks for catalysts in the size range of 1-5 nm
is subtle (at least for the sweep rate utilized in our experiments),
that is, the most positive peak is the one obtained for the 5 nm
particles, while those for the 1 and 2 nm particles are almost
identical. Clearly, as for extended surfaces, for high surface area
catalysts, it is very difficult to use the CO stripping peak as a
measure for the particle size effect in CO oxidation. As we
discuss below, to determine the onset potential for CO oxidation
on the nanoparticle samples, we utilize the high sensitivity of
IRAS to detect the O-C-O vibrational frequency of dissolved
CO2 as well as chronoamperometric measurements.
3.2 FTIR Measurements. In this section, the analysis and
discussion of the FTIR results will be divided into two parts.
First, we describe the vibrational properties of CO adsorbed on
Pt nanoparticles pretreated by the oxide-annealing method, and
then, we compare them to the results obtained for particles
prepared by the CO-annealing method.
3.2.1 Pt Nanoparticles Pretreated by the Oxide-Annealing
Method. In contrast to standard infrared spectral acquisition
tactics where the potential is increased in steps of 50 or
100 mV, here, the spectra were recorded every 2.5 s during a
positive potential sweep of 1 mV/s. A representative set of
spectra obtained by the latter tactics on the 5 nm catalyst is
summarized in Figure 3a (CO band) and b (CO2 band).
The spectra in Figure 3a consist of a main band around
2060 cm-1, which can be assigned to atop-bonded CO on Pt,28
and a less pronounced CO band around 1880 cm-1, which can
Figure 2. (a) Cyclic voltammograms of carbon-supported Pt catalyst
samples recorded in 0.1 M HClO4 solution; T ) 20 °C, scan rate 50 mV/s;
currents are normalized to the measured Pt surface area (Hupd charge after
double layer correction); HRTEM images of a carbon-supported Pt
nanoparticle (b), and the nanostructured Pt film supported on organic
whiskers (c).
ammograms for the Pt nanoparticles attached to carbon and
organic whiskers. The voltammograms for 1, 5, and 30 nm Pt
catalysts (Figure 2) show three characteristic potential regions:
the hydrogen adsorption potential region between 0.05 and
0.35 V is followed by the “double-layer” potential region, while
above ∼0.7 V, first OH adsorption and then oxide formation
can be observed. Due to different loadings of the Pt catalysts
as well as different contributions from the capacitance of the
support, the comparison of CV’s of high surface area catalysts
is not as straightforward as established for extended single
crystal surfaces.3 Therefore, the cyclic voltammograms are
2
“normalized” in terms of the specific surface area (µA/cmPt
)
obtained from the charge density for hydrogen adsorption/
desorption corrected for the charging of the double layer. With
this methodology, some trends in the adsorption/desorption
properties of these catalysts can be established. For OH
adsorption, the particle size dependence is less obvious;
however, the oxophilicity of the catalysts appears to increase
overall with increasing mean particle size. A small, yet clearly
discernible, positive shift in the OHad formation is observed by
a change in the particle size, that is, approximately 20 mV from
1 to 30 nm, while the onset for the 5 nm catalyst falls in
between. This effect is more pronounced on the reversible
sweep, where it is unambiguously shown that the peak for the
oxide reduction shifts toward more negative potentials for the
smaller particles, that is, about 130 mV for the 1 nm particles
compared to the 30 nm particles. In agreement with oxide
formation, the oxide reduction peak obtained for the 5 nm
particles lies just between the one for the 1 and 30 nm particles,
confirming previous findings that the oxophilicity of Pt particles
increases by decreasing the particle size.24-26 Assuming that
CO oxidation on small particles obeys the Langmuir-Hinshel-
wood reaction mechanism
2H2O f OHad + H3O+ + e-
(1)
(24) Takasu, Y.; Ohashi, N.; Zhang, X.-G.; Murakami, Y.; Minagawa, H.; Sato,
S.; Yahikozawa, K. Electrochim. Acta 1996, 41, 2595-2600.
(25) Mukerjee, S.; McBreen, J. J. Electroanal. Chem. 1998, 448, 163-171.
(26) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Appl. Catal.,
B 2005, 56, 9-35.
(27) Markovic, N. M.; Grgur, B. N.; Lucas, C. A.; Ross, P. N. J. Phys. Chem.
B 1999, 103, 487-495.
(28) Chang, S. C.; Weaver, M. J. Surf. Sci. 1990, 238, 142-162.
9
6822 J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005

Effect of Particle Size on Kinetics of CO Electrooxidation
A R T I C L E S
Figure 3. Representative CO stripping series of FTIR spectra (5 nm sample) recorded while scanning with 1 mV/s in argon-saturated HClO4, every fifth
spectra is depicted; (a) absorption band of atop and bridge-bonded CO; (b) CO2 band.
Figure 4. Summary of the results from the series of potential-dependent CO adsorption spectra for the different nanoparticle samples; measurements were
taken in Ar-saturated solution after cycling in Ar-saturated solution and subsequent CO adsorption at 50 mV (method I); the potential was scanned with
1 mV/s while continuously recording spectra; the first spectra of the respective catalysts are given in part (a); the upper part (b) shows the potential dependent
C-O stretch frequency of atop adsorbed CO; the lower part (c) compares the normalized peak area of CO2 dissolved into solution (2343 cm-1).
be assigned to bridge-bonded CO. The results obtained for the
different catalyst samples are summarized in Figure 4 (for the
sake of clarity, the results of the 2 nm sample are not shown
here; see Figure 6). In agreement with previous measure-
ments,7,15,29 at 0.05 V, νCO redshifts as the particle size decreases,
that is, from 2062 cm-1 on 30 nm particles to 2060, 2052, and
2049 cm-1 on 5, 2, and 1 nm particles, respectively.
The interpretation of the particle size-dependent C-O stretch-
ing frequency can be rationalized in terms of the chemical
interaction between CO and Pt (chemical shift) as well as
CO-CO lateral interactions (dipole-dipole coupling shift).7,30,31
In particular, the decreased vibrational frequencies on small
particles have been proposed to arise due to a stronger
interaction of CO with the high density of edge-corner
coordination in small nanoparticles, which is in line with detailed
theoretical density functional (DFT) calculations32,33 of CO
chemisorbed on low-coordination sites, compared to terrace Pt
sites, as well as a reduced dipole-dipole interaction of the CO
adlayer.34 It should be noted, however, that DFT calculations
show no general link between the metal-CO binding energy
(31) Severson, M. W.; Weaver, M. J. Langmuir 1998, 14, 5603-5611.
(32) Hammer, B.; Norskov, J. K. In Chemisorption and ReactiVity on Supported
Clusters and Thin Films; Lambert, R. M., Pacchioni, G., Eds.; Kluwer
Academic Publisher: Dordrecht, The Netherlands, 1997; p 285.
(33) Hammer, B.; Nielsen, O. H.; Norskov, J. K. Catal. Lett. 1997, 46, 31-35.
(34) Park, S.; Wasileski, S. A.; Weaver, M. J. J. Phys. Chem. B 2001, 105,
9719-9725.
(29) Rice, C.; Tong, Y.; Oldfield, E.; Wieckowski, A.; Hahn, F.; Gloaguen, F.;
Leger, J. M.; Lamy, C. J. Phys. Chem. B 2000, 104, 5803-5807.
(30) Severson, M. W.; Stuhlmann, C.; Villegas, I.; Weaver, M. J. J. Chem. Phys.
1995, 103, 9832-9843.
9
J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005 6823

A R T I C L E S
Arenz et al.
one and the correlation between the C-O vibrational frequency,
and the detailed structure of the nanoparticles is far from being
clear. To resolve this question, isotopic labeling measurements
are suitable36 and are currently performed on Pt single crystals
using our FTIR methodology.
Figure 4a also shows that the CO bands are rather symmetrical
with a bandwidth (fwhm) being almost independent of the
particle size (ca. 20 cm-1). This observation is at variance with
previous findings, namely, that small particles exhibit a lower
wavenumber “tail”.34,37 We recently suggested that the observed
“tailing” is most likely an artifact that can completely be avoided
by uniformly anchoring the metal nanoparticles onto the Au
substrate.22 Furthermore, Figure 4a shows that if the catalyst
layer is uniformly dispersed on Au, then the fwhm is indepen-
dent of the particle size.
The summary of the results from the series of potential-
dependent CO adsorption spectra and the corresponding CO2
production in Figure 4 serves to illustrate how the nanoparticle
size affects the vibrational properties of COad as well as the
catalytic properties of Pt. As exemplified in Figure 4, for the
30 nm Pt particles, three potential regions can be discerned from
the νCO versus E and the ICO versus E plots. In the potential
2
region (0.05 < E < ∼0.35 V), where the CO adlayer is stable
(no CO2 production in Figure 4c), a linear frequency shift
dν/dE (25 cm-1/V) is observed, consistent with the so-called
electrochemical Stark-tuning effect.38 Between ∼0.35 and
∼0.60 V, however, the Stark-tuning slope, dν/dE, changes its
sign, while simultaneously, a weak CO2 band is detected in the
spectra, indicating a slow but finite CO oxidation rate. In line
with our previous discussion of the Pt(hkl)-CO system,3 the
potential range where the CO oxidation rate is rather small, we
define as the pre-ignition potential region. Finally at E
∼ 0.60 V, the so-called ignition potential, we observe a fast
Figure 5. Summary of FTIR results of 2, 5, and 30 nm catalysts. In contrast
to Figure 4, the CO adlayer was formed by the CO-annealing procedure
(i.e., by cycling in CO-saturated solution for 5min, stopping at 0.05 V, and
purging with Ar in order to remove all CO from solution (method II)).
increase in ICO as well as a substantial change in dν/dE, directly
2
indicative of facile CO oxidation.
The spectral data of the 5 nm particles in Figure 4b reveal
some important differences as well as some similarities with
the 30 nm particles. As for larger particles, a linear positive
Stark-tuning slope of ∼25 cm-1/V is observed over the potential
region where no CO2 is detected, 0.05 < E < ∼0.35 V. At
E > 0.35 V, the slope starts to decrease, but in contrast to the
30 nm particles, it still remains positive all the way up to
0.7 V. The small change in the Stark-tuning slope is ac-
companied by the appearance of a relatively weak CO2 band.
Although the onset potential for CO2 production is the same
on the 5 and 30 nm particles, Figure 4c shows that at low
potentials, the integrated intensities for the CO2 band are much
lower for the 5 nm particles, suggesting that the rate of CO2
production is much smaller on the 5 nm particles than on the
30 nm particles. The confirmation that the oxidative removal
of COad on 30 nm particles is faster than that on 5 nm particles
is also obtained from the ignition potential region (the negative
dν/dE slope and extensive production of CO2), which is on the
5 nm catalyst shifted by ca. 50 mV toward higher potentials.
For very small Pt nanoparticles (1 and 2 nm), we observe an
anomalous behavior in the νCO-E frequency shift, which to our
knowledge for the case of nanoparticles, has not yet been
Figure 6. Comparison of CO adsorption/oxidation properties obtained with
different CO adsorption procedures, CO-annealing and oxide-annealing
(method I and II; for details, see text); blue colors: 30 nm sample; brown
colors: 2 nm sample.
and the C-O stretching frequency.35 On the basis of our
measurements, we cannot state which effect is the predominant
(36) Chang, S. C.; Roth, J. D.; Weaver, M. J. Surf. Sci. 1991, 244, 113-124.
(37) Park, S.; Tong, Y.; Wieckowski, A.; Weaver, M. J. Electrochem. Commun.
2001, 3, 509-513.
(35) Shubina, T. E.; Koper, M. T. M. Electrochim. Acta 2002, 47, 3621-3628.
(38) Bishop, D. M. J. Chem. Phys. 1993, 98, 3179-3184.
9
6824 J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005

Effect of Particle Size on Kinetics of CO Electrooxidation
A R T I C L E S
reported in the literature. For 1 nm particles, this anomalous
behavior is more pronounced than that for 2 nm particles, and
in the following, we will focus on the CO vibrational/catalytic
properties on 1 nm particles. Figure 4b shows that around
0.35 V, the Stark-tuning slope first slightly increases then while
still remaining positive, at E > 0.55 V, begins to decrease, and
finally at 0.7 V, where the rapid CO oxidation occurs, a negative
sign in dν/dE is observed. Interestingly, the blueshift deviations
in the νCO frequency are accompanied by a small yet clearly
discernible CO2 production, indicating that although the surface
is less covered by CO, the interaction energy of CO with the Pt
particles is weaker! Note that we were able to observe a similar
blueshift in the νCO frequency at the onset of CO oxidation on
Pt(111).19 For the Pt(111)-CO system, we suggested that the
initial CO oxidation is accompanied by adsorption of anions
from the supporting electrolyte. We also argued that co-adsorbed
anions may lead to a compression of CO molecules, resulting
in a higher local CO coverage thus yielding blueshifted atop
CO frequencies due to enhanced dipole-dipole coupling. Notice
that the phenomenon of island compression is observed in
segregated systems where the presence of one species causes
the other species to segregate into patches in which the local
density is higher than that observed when an equivalent amount
of particular species is adsorbed alone (see for example ref 39).
Extending this phenomenon to CO adsorbed on Pt nanoparticles,
the surprising blueshift in the νCO frequency for 1 nm particles
at E > 0.3 V may be a consequence of slow but continuous
interactions on larger particles). Further inspection of Figure
5a shows that for 30 nm particles, the “normal” Stark-tuning
behavior, with a slope of ∼25 cm-1/V, is observed from 0.05
to 0.65 V, that is, just prior to the occurrence of significant CO
oxidation, monitored by an extensive CO2 production (Figure
5b). For 5 nm particles, a linear Stark-tuning slope of
∼25 cm-1/V is also clearly observed between 0.05 and 0.7 V.
On both catalysts, a small CO2 production is observed at
potentials as low as 0.6 V (see the insert of Figure 5).
Intriguingly, on 2 nm particles, the anomalous blueshift deviation
from the normal Stark-tuning slope is observed over the potential
range of 0.55 < E < 0.7 V. Again, the observed deviation
coincides with a small yet clearly discernible CO2 production
(Figure 5b), confirming that the blueshift deviation is triggered
by the slow oxidation of CO. If we use ICO as a measure for
2
the catalytic activity, then the CO oxidation rate increases in
the order 30 > 5 > 2 nm.
3.2.3 Effect of Surface Irregularities. In Figure 6, we
compare the IRAS data for the two different pretreatments of
the nanoparticle catalysts. For the 30 nm particles, νCO on the
CO-annealed surface is blueshifted by ca. 5 cm-1 relative to
the surface pretreated by the oxide-annealing method. If
correlated to the CO bonding energy, the observed blueshift
would imply that the Pt-CO interaction is weaker with the
CO-annealed than with the oxide-annealed nanoparticles. The
CO2 production in Figure 6b, however, is higher on the latter
surface, showing that the C-O stretch of Pt-bound CO is no
indication for the catalytic activity of the CO adlayer.
Since we observed that surfaces containing more irregularities,
such as the oxide-annealed 30 nm particles, are more reactive,
we suggest that the oxidation rate of a CO monolayer (CO
stripping conditions) is controlled by the presence of catalytically
active OH which is formed on defect sites rather than by the
Pt-CO energetics, that is, the CO mobility. This is in agreement
with our findings that on Pt single crystal surfaces, the CO
oxidation rate is more influenced by the ability of the surface
to dissociate water and to form OHad on defect sites than by
CO energetics.17,42,43 Defect sites also play a role in CO
“clustering” on the 30 nm nanoparticles, causing CO to decrease
in local coverage on the oxide-annealed surface relative to the
CO-annealed surface. Consequently, the higher local CO
coverage on the latter surface yields to blueshifted νCO frequen-
cies due to enhanced dipole-dipole coupling, as shown in
Figure 6a.
CO oxidation which is accompanied by anion adsorption (either
- 40
ClO4
or trace levels of Cl- impurities41) and concomitant
CO compression on (111) facets. Furthermore, the close
similarity between the vibrational properties of CO adsorbed
on Pt single crystals and very small Pt nanoparticles is in line
with the TEM analysis, which showed that the surface of large
particles contains undistorted (111) facets in a smaller fraction
in contrast to the surface of small particles.
To conclude this section, we have shown that while the onset
potential for CO oxidation is almost independent of the particle
size, the rate of CO2 production strongly depends on the particle
size, that is, 1 e 2 < 5 , 30 nm. Closely following the
interpretation of the CO oxidation rate in the pre-ignition region
on extended Pt single crystal surfaces, in the next section, we
suggest that ICO is controlled by the presence of specific surface
2
irregularities (see section 3.1), which serve as nucleation center
for OH adsorption.
In this context, it is also worth mentioning that over the
potential range of 0.05 < E < 0.5 V, the νCO frequency on the
2 nm particles (as well as on the 1 nm particles) is independent
from the preparation procedure. This observation is consistent
with the TEM analysis showing that the “surface roughness”
of the cubooctahedrons increases with the particle size and, thus,
the vibrational properties of COad on smaller particles are less
affected by the preparation method. The importance of defects
is, however, observed at potentials E > 0.5 V, where a small
production of CO2 on CO-annealed particles is accompanied
by an increase in the C-O vibrational frequency. Notice that
the blueshift deviations from the normal Stark-tuning slope are
more pronounced on the CO-annealed, that is, the less irregular
3.2.2 Pt Nanoparticles Pretreated by the CO-Annealing
Method. For the purpose of our discussion here, we present
the νCO versus E plots for the 30, 5, and 2 nm particles.
As shown in Figure 5a and in agreement with the results
shown in Figure 4, by decreasing the particle size, the CO band
position shifts toward lower wavenumbers, that is, from 2067
to 2062 and 2052 cm-1 for 30, 5, and 2 nm particles,
respectively. As discussed in refs 7 and 15 and section 3.2.1,
the observed redshift can be rationalized by an interplay between
the chemical shift (stronger CO adsorption on smaller nano-
particles) and the dipole shift (enhanced CO-CO lateral
(39) McIntyre, B. J.; Salmeron, M.; Somorjai, G. A. Surf. Sci. 1995, 323, 189-
197.
(40) Ataka, K.; Yotsuyanagi, T.; Osawa, M. J. Phys. Chem. 1996, 100, 10664-
(42) Grgur, B. N.; Markovic, N. M.; Ross, P. N. J. Phys. Chem. B 1998, 102,
2494-2501.
(43) Markovic, N. M.; Lucas, C. A.; Grgur, B. N.; Ross, P. N. J. Phys. Chem.
B 1999, 103, 9616-9623.
10672.
(41) Arenz, M.; Stamenkovic, V.; Schmidt, T. J.; Wandelt, K.; Ross, P. N.;
Markovic, N. M. Surf. Sci. 2003, 523, 199-209.
9
J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005 6825

A R T I C L E S
Arenz et al.
2 nm particles. Most likely, the microscopic effect responsible
for the observed differences in the νCO frequency is the same
as the one proposed for the 30 nm particles; that is, the smaller
number of defects on the CO-annealed surface allows CO to
increase in local coverage accounting for the observed blueshift
in the νCO frequencies due to enhanced dipole-dipole coupling.
As for the 30 nm nanoparticles, the “high-frequency CO” on
the 2 nm nanoparticles is less reactive than the “low-frequency
CO”, which intuitively can be explained more easily by the
availability of defect sites for OH adsorption than by Pt-CO
energetics. Therefore, considering that surfaces of larger particles
are inherently rougher than those of smaller particles, the order
of reactivity for CO monolayer oxidation (CO stripping)
increases by increasing the size of the Pt nanoparticles. Along
the same lines, because CO-annealed surfaces contain fewer
defects than oxide-annealed surfaces, oxidative removal of CO
on particles with nominally the same size is enhanced on the
oxide-annealed surface.
3.3 Chronoamperometric Measurements. Additional in-
formation about the particle size effect on the oxidative removal
of CO is obtained from chronoamperometry experiments. The
same method has been used previously by Love et al.44 as well
as by Lebedeva et al.45 for studying the role of crystalline
“defects” on the rate of CO oxidation on Pt(hkl) and stepped Pt
single crystal surfaces, respectively. The authors found that an
increase in the step density for Pt(hkl) from the [11h0] zone
results in a systematic negative shift of the CO stripping peak,
consistent with the previous suggestion17 that low-coordination
sites may serve as active centers for OH adsorption and, thus,
oxidative removal of CO. The current-time transients for the
oxidation of a CO-saturated adlayer have been successfully
modeled by using the mean-field approximation of the L-H
mechanism, which implies fast diffusion of CO adsorbed on
the terraces of the Pt stepped single crystal surfaces. By contrast,
Maillard et al. suggested that for ideal cubooctahedral Pt
nanoparticles, the stepped single crystal surfaces could not be
used as an adequate model for understanding the shape of the
CO oxidation current transients.10 The model for the perfect
cubooctahedral particle suggested restricted CO mobility at Pt
nanoparticles below ca. 2 nm size and a transition toward fast
diffusion when the particle size exceeds 3 nm. Unfortunately,
the model did include neither the competitive adsorption of
bisulfate anions (which can control the adsorption of OH as
well as the mobility of co-adsorbed CO19) nor the existence of
irregularities on nanoparticles that, as for the extended single
crystal surfaces, may serve as the active centers for OH
adsorption. In the following, we use the chronoamperometric
method to get further insight into the effect of surface irregulari-
ties on the reactivity of Pt nanoparticles in CO monolayer
oxidation.
Figure 7. Chronoamperometric measurements of CO adlayer oxidation
on 1 nm particles in 0.1 M HClO4; CO was adsorbed at 50 mV, and the
final potential was chosen in the pre-ignition (a) as well as in the ignition
potential region (b); the insert indicates the respective potentials in the
corresponding CO stripping voltammetry.
the ignition potential, and curves 4 and 5 are characteristic
representatives for the transients observed within the ignition
potential region. These examples characterize the general pattern
of induction periods, that is, the time required for a system to
pass from an inactive state (charging of the double layer) to
active state (the onset of CO oxidation), as well as the shape of
current transients. As expected, the induction time is longer (by
orders of magnitude) in the preoxidation region (Figure 7a) than
at or above the ignition potentials (Figure 7b), consistent with
the proposition that the rate of oxidative removal of CO in
stripping experiments depends on the number of sites on which
OH can be adsorbed. The same argument can be used to
rationalize the observation that the time required to reach the
maximum current, as well as to completely oxidize CO, is
shorter after stepping to more positive potentials.
Even though the same potential-step experiments were
conducted for 2, 5, and 30 nm Pt nanoparticles, only selected
results for CO monolayer oxidation on 2 and 30 nm particles
will be presented and compared with the results for the 1 nm
particles. Figure 8 summarizes representative transients for CO
oxidation on the three different nanoparticles, along with the
charging curve of the “double layer” (dotted curves), that is,
after stepping from 0.05 V to the same final potentials but on
CO-free surfaces. Figure 8 a′-c′ shows the difference between
the transient for CO oxidation and the charging of the double
layer, that is, the true transients for CO oxidation. Clearly, the
particle size generates quite significant changes in the CO
transient curves. (i) The induction period decreases by increasing
the particle size, consistent with the FTIR results that larger
particles are more active for CO oxidation. (ii) While on the 1
nm particles a well-defined single maximum (characteristic for
3.3.1 Potentiostatic Conditions. Figure 7 shows the transient
currents for CO oxidation on 1 nm Pt particles as the potential
is stepped from 0.05 V to five more positive potentials that are
designated on a corresponding cyclic voltammogram (see insert).
Curves 1 and 2 represent the current transients for CO oxidation
in the preoxidation region; curve 3 shows a potential step at
(44) Love, B.; Lipkowski, J. Effect of surface crystallography on electrocatalytic
oxidation of carbon monoxide on Pt electrodes. In Molecular Phenomena
at Electrode Surfaces; Soriaga, M. P., Ed.; American Chemical Society:
Washington, DC, 1988; pp 484-496.
(45) Lebedeva, N. P.; Koper, M. T. M.; Feliu, J. M.; van Santen, R. A. J. Phys.
Chem. B 2002, 106, 12938-12947.
9
6826 J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005

Effect of Particle Size on Kinetics of CO Electrooxidation
A R T I C L E S
Figure 8. Chronoamperometric measurements of (a) 1, (b) 2, and (c) 30 nm particles, as recorded current-time transients (left-hand side), transients after
correction of the double charging/discharging transients (the right-hand side); initial potential of 0.05 V, final potential as indicated.
current transients on Pt single crystals46) is observed, on the
2 nm particles, current transients can be deconvoluted into two
peaks, one that appears within 5 s and the other that is developed
after 10 s (0.780V), suggesting that 1 nm particles are more
“ideal” cubooctahedrons than are the 2 nm particles. (iii) The
oxidation of CO on 30 nm particles is so fast that no apparent
current maximum is observed in the current transients. Note
that because the CO oxidation on the 30 nm particles is much
faster than on the other three, the final positive potential on the
former catalyst is at 0.63 and 0.655 V.
On the basis of these observations, and in light of our
discussion in the previous section, it becomes obvious that the
puzzling size effects on the reactivity of Pt nanoparticles in CO
monolayer oxidation (CO stripping) can be explained by
considering how the fraction of irregularities (defects), surface
atoms in (111), (100) facets, and edge-corner sites may effect
the rate of the oxidative removal of CO from a real nanoparticle.
In our previous report for CO oxidation on Pt(111),17 we argued
that on the surface, which is completely covered by CO, the
initial adsorption of OH (i.e., the initial CO oxidation) takes
place at defects in the (111) surface. This is confirmed from
the results published later for the CO stripping on the Pt stepped
single crystal surfaces.45-47 We also suggested that in acid
solutions there is a strong competition for the active sites
between the OH and anions from the supporting electrolyte.
Initially, on the surface fully covered by CO, the adsorption of
relatively small OH ions will be predominant because of
geometric considerations, even though at low overpotentials
(0.05 < E < 0.7 V), specific adsorption of relatively large anions
would be thermodynamically more favorable. As the CO
oxidation proceeds, however, the anion adsorption will be
possible on CO-free Pt sites. Consequently, anions can block
the further adsorption of OH that, in turn, will affect the rate of
CO oxidation and the observed tailing. In addition to the
blocking effect, the adsorbed anions may also affect the diffusion
In Figure 9, we compare the chronoamperometric data for
CO oxidation on the 1 nm catalyst pretreated either by the oxide-
annealing procedure (red curves) or by the CO-annealed
procedure (blue curves). In these experiments, by changing the
number of irregularities while keeping the same size of the
particles, the intention is to study the effect of the density of
active sites for OH adsorption on the induction period and
current relaxation during CO oxidation. As expected, the
induction period increases by decreasing the number of ir-
regularities on the CO-annealed surface. In addition, the time
required to reach the current maximum is shorter for the oxide-
annealed than for the CO-annealed surfaces, in agreement with
our discussion above that the initial oxidation rate depends on
the number of active sites for OH adsorption (in the Supporting
Information, we also show the results for 2 nm).
(46) Lebedeva, N. P.; Koper, M. T. M.; Feliu, J. M.; van Santen, R. A. J.
Electroanal. Chem. 2002, 524, 242-251.
(47) Lebedeva, N. P.; Koper, M. T. M.; Herrero, E.; Feliu, J. M.; van Santen,
R. A. J. Electroanal. Chem. 2000, 487, 37-44.
9
J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005 6827

A R T I C L E S
Arenz et al.
Figure 10. Rotating disk electrode measurements of CO bulk oxidation
on 1, 5, and 30 nm catalysts in 0.1 M HClO4 solution constantly purged
with CO; CO adsorption at 0.05 V, T ∼ 20 °C, scan rate 1 mV/s, rotation
rate 1600 rpm; positive sweep shown only for 30 nm (others are identical),
positive limit 1.0 V; insert: logarithmic plot of specific activity for
COb-ox at 0.8V obtained at cathodic sweep versus particle size.
a nonreactive oxygen species,50 in a manner similar to that
observed by Ostermaier et al.51 for ammonia oxidation over Pt.
In electrocatalysis, it has also been found that under different
experimental conditions, either a nonreactive oxide (E >
1.1 V) or a reactive oxide (termed as OHad; E < 1.1 V) can
build up on the surface. To test how the build-up of reactiVe
oxygen affects CO oxidation on Pt nanoparticles in an electro-
chemical environment, polarization curves were recorded upon
sweep reversal at 1.0 V, where the surface is predominantly
covered by reactive oxygenated species. The results in Figure
10 clearly show that if the electrode is predominately covered
by oxide, the CO oxidation reaction is demanding and the
activity of Pt nanoparticles increases in opposite order to that
observed in CO stripping experiments (30 < 5 < 2 < 1 nm).
Clearly, in the case of the CO oxidation reaction, depending
upon the reaction conditions, small particles are able to exhibit
higher or lower specific activity than larger ones. For example,
because the isosteric heat of CO adsorption on Pt(hkl) is
relatively insensitive to the surface structure, it is reasonable to
suggest that if the surface is completely covered by CO (high-
pressure conditions), the Pt-CO energetics will play a negligible
role in the kinetics of CO oxidation.
Figure 9. Comparison of chronoamperometric measurements on the oxide
and CO-annealed surface of the 1 nm sample; insert: double layer corrected
current-time transients; initial potential of 0.05 V, final potentials as
indicated.
as well as the clustering of CO, as we discussed in the FTIR
section. Closely following this discussion, it is plausible that
the observed tailing observed at Pt nanoparticles below 2 nm10
is affected rather by the influence of anions on CO diffusion
than by the energetics of the Pt-CO interaction. To conclude,
in the future, the modeling of CO oxidation current transients
would require the specific adsorption of anions to be taken into
consideration, as well.
3.4. RDE Measurements of CO Bulk Oxidation. Although
CO stripping curves and chronoamperometric measurements
provide valuable information on the oxidative removal of CO,
in fuel cell applications, the anode catalyst experiences a
continuous supply of CO to the surface, which can be simulated
in RDE measurements. We found for CO bulk measurements,
that is, RDE measurements in CO-saturated solution, that if the
surface is completely covered by CO, the difference in catalytic
activity between the four Pt catalysts is negligible in the positive
going sweep. Interestingly, in gas-phase catalysis, McCarthy
et al.48 found that under high CO partial pressure conditions
(equivalent to our experimental conditions when surface is
completely covered by CO), the specific rate is “facile”
(insensitive to the particle size) but “demanding” (structure
sensitive49) for low CO partial pressure. Under the latter
conditions, the specific rate is higher for low area (sintered) Pt.
It has been suggested that the observed deactivation with
decreasing particle size stems most likely from the build-up of
By contrast, the number of defects on which the initial
adsorption of reactive OH will occur is very important, so that
the rate of oxidative CO removal in CO stripping experiments
increases with the availability of those sites, that is, 1 < 2 < 5
< 30 nm. However, under experimental conditions in which
the surface is predominantly covered by OHad (oxide), the
reversed order in activity is obtained, consistent with the fact
that the heat of oxide formation increases by increasing the
number of low-coordinated surface atoms (see voltammetry in
Figure 2). Given that at the adsorbate-free nanoparticle the
fraction of active edge-corner sites is larger than the defect
sites and considering that the surface distribution of the former
(48) McCarthy, E.; Zahradnik, J.; Kuczynski, G. C.; Carberry, J. J. J. Catal.
1975, 39, 29-35.
(49) Boudart, M. AdV. Catal. 1969, 20, 153.
(50) Cant, N. W. J. Catal. 1980, 62, 173-175.
(51) Ostermaier, J. J.; Katzer, J. R.; Manogue, W. H. J. Catal. 1974, 33, 457-
473.
9
6828 J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005

Effect of Particle Size on Kinetics of CO Electrooxidation
A R T I C L E S
sites increases by decreasing the particle size,52 then small
particles are more oxophilic. Thus, at high overpotentials, they
provide more active sites for the oxidative removal of adsorbed
CO. Therefore, depending on the experimental conditions, the
ratio between the availability of “defects” versus “ideal” sites
is changing, from highly available defect sites on an electrode
fully covered by CO, which is determining the induction period
and the shape of current transients in the CO stripping
experiments, to a high fraction of available regular facets and
edge-corner sites at the surface predominately covered by
oxide, which determines the rate of CO bulk oxidation.
depending on the particle size (d), the onset of CO2 production
is accompanied either by redshift (d > 5 nm) or by anomalous
blueshift (d < 2 nm) deviations from the normal Stark-tuning
slope. The interpretation of the anomalous dνCO/dE behavior is
discussed in relation to similar changes observed at the Pt(111)-
solution interface, that is, in terms of slow CO oxidation, co-
adsorption of anions, and concomitant CO compression into
small islands on fairly smooth (111) facets.
(3) Chronoamperometric measurements indicate that the
observed tailing in the transients on small Pt nanoparticles is
rather due to the influence of co-adsorbed anions on CO surface
diffusion than by the energetics of the Pt-CO interaction. In
the future, the modeling of CO oxidation current transients
would require the specific adsorption of anions and irregular
particle shapes to be taken into consideration, as well.
4. Conclusion
Platinum particles ranging in size from 1 to 30 nm were
characterized and evaluated for their catalytic activity for CO
electrooxidation, using high-resolution transmission electron
microscopy, infrared reflection-absorption spectroscopy, and
electrochemical measurements. The results can be summarized
as follows:
(1) TEM analysis shows that the Pt crystallites are, in general,
not perfect cubooctahedrons, and that the particles provide a
variety of low-coordination sites for adsorption and catalysis.
We found that large particles are quite “rough”, containing
irregular steps and occasionally twinned particles. On the other
hand, the smallest particles show mostly regular facets and
uniform edge-corner sites as expected from perfect cubo-
octahedrons.
(2) The surface sensitivity for the oxidative removal of
adsorbed CO was probed in IRAS experiments by monitoring
the potential dependence of the atop CO vibrational frequency
and the concomitant development of asymmetric O-C-O
stretch of dissolved CO2. It was found that while the onset
potential of CO oxidation is almost independent of the particle
size, the rate of CO2 production is strongly dependent on the
particle size (i.e., 1 e 2 < 5 , 30 nm). We suggest that the
oxidative removal of CO is mainly controlled by the number
of defects, which in an ideal cubooctahedral particle may serve
as an active center for OH adsorption. It was also found that,
(4) In contrast to CO adlayer stripping, we found that the
oxidation kinetics of dissolved CO on an OHad-covered surface
increase by decreasing the particle size, that is, more oxophilic
small particles are more active for bulk CO oxidation than are
larger particles. Therefore, the observed particle size effect
strongly depends on the experimental conditions.
Acknowledgment. This work was supported by the Director,
Office of Science, Office of Basic Energy Sciences, Division
of Materials Sciences, U.S. Department of Energy under
Contract No. DE-AC03-76SF00098. We thank Dr. V. Rad-
milovic from the National Center for Electron Microscopy for
his expertise in analyzing the TEM images, as well as Dr. R.
Atanasoski and Dr. M.K. Debe from the 3M Corporation for
supplying one of the catalyst samples. K.M. acknowledges the
Austrian BMBWK for a Ph.D. scholarship. M.A. is grateful for
a Feodor Lynen fellowship from the German Alexander von
Humboldt foundation.
Supporting Information Available: Figures showing the
current-time transients and a comparison of the chronoampero-
metric measurements of the oxide CO-annealed surface of the
2 nm sample. This material is available free of charge via the
Internet at http://pubs.acs.org.
(52) Kinoshita, K. Electrochemical Oxygen Technology; John Wiley & Sons:
New York, 1992.
JA043602H
9
J. AM. CHEM. SOC. VOL. 127, NO. 18, 2005 6829