Size effects in electronic and catalytic properties of unsupported palladium nanoparticles in electrooxidation of formic acid

Article abstract of DOI:10.1021/jp061690h

We report a combined X-ray photoelectron spectroscopy (XPS), cyclic voltammetry (CV), and chronoamperometry (CA) study of formic acid electrooxidation on unsupported palladium nanoparticle catalysts in the particle size range from 9 to 40 nm. The CV and C

Full text of DOI:10.1021/jp061690h

J. Phys. Chem. B 2006, 110, 13393-13398
13393
Size Effects in Electronic and Catalytic Properties of Unsupported Palladium Nanoparticles
in Electrooxidation of Formic Acid
Wei Ping Zhou,† Adam Lewera,†,‡ Robert Larsen,§ Rich I. Masel,§ Paul S. Bagus,| and
Andrzej Wieckowski*,†
Department of Chemistry, UniVersity of Illinois at Urbana-Champaign, 600 South Mathews AVenue,
Urbana, Illinois, 61801, Department of Chemistry, Warsaw UniVersity, Pasteura 1, Warsaw 02-093, Poland,
Department of Chemical and Biomolecular Engineering, UniVersity of Illinois at Urbana-Champaign,
600 South Mathews AVenue, Urbana, Illinois, 61801, and Department of Chemistry, UniVersity of North Texas,
Denton, Texas 76203-5070
ReceiVed: March 17, 2006; In Final Form: May 13, 2006
We report a combined X-ray photoelectron spectroscopy (XPS), cyclic voltammetry (CV), and chrono-
amperometry (CA) study of formic acid electrooxidation on unsupported palladium nanoparticle catalysts in
the particle size range from 9 to 40 nm. The CV and CA measurements show that the most active catalyst is
made of the smallest (9 and 11 nm) Pd nanoparticles. Besides the high reactivity, XPS data show that such
nanoparticles display the highest core-level binding energy (BE) shift and the highest valence band (VB)
center downshift with respect to the Fermi level. We believe therefore that we found a correlation between
formic acid oxidation current and BE and VB center shifts, which, in turn, can directly be related to the
electronic structure of palladium nanoparticles of different particle sizes. Clearly, such a trend using unsupported
catalysts has never been reported. According to the density functional theory of heterogeneous catalysis, and
mechanistic considerations, the observed shifts are caused by a weakening of the bond strength of the COOH
intermediate adsorption on the catalyst surface. This, in turn, results in the increase in the formic acid oxidation
rate to CO2 (and in the associated oxidation current). Overall, our measurements demonstrate the particle size
effect on the electronic properties of palladium that yields different catalytic activity in the HCOOH oxidation
reaction. Our work highlights the significance of the core-level binding energy and center of the d-band
shifts in electrocatalysis and underlines the value of the theory that connects the center of the d-band shifts
to catalytic reactivity.
Introduction
electronic effects in the formic acid electrooxidation process.
We present experimental evidence of the correlation between
the electronic properties of unsupported Pd nanoparticlessas
they vary as a function of their particle sizesand their catalytic
activity.
Direct formic acid fuel cell (DFAFC) is a convenient energy
source for micropower energy devices such as cell phones and/
or laptops and promises to be one of the first small fuel cells
on the market.1,2 Therefore, major efforts have been made to
demonstrate the DFAFC performance and, at the same time, to
advance understanding of mechanisms of formic acid electro-
oxidation processes.2-9 The studies reported below concentrate
on the formic acid oxidation process on unsupported nano-
particle Pd catalysts, that is, on the active DFAFC anode.1 Notice
that the use of nanoparticle materials in practical electrocatalysis
is essential as the nanoparticles guarantee sufficiently high
surface-to-bulk atomic ratio and utilization of the noble metal
catalyst components economically. As an additional bonus, the
nanoparticle electronic properties can be adjusted by controlling
the nanoparticle size.10
It is generally accepted that the oxidation of formic acid
occurs via the dual path mechanism:11,12 one path involving a
direct carbon dioxide formation and another an inhibiting
intermediate formation.13 At low potentials, adsorbed CO
produced from formic acid decomposition is the inhibiting
intermediate.12,14,15 Therefore, it is important to understand the
CO oxidation behavior at the Pd nanoparticles, as it will be
referred to below. At higher potentials, formate intermediate is
the poisoning species.13 However, no voltammetric signature
to the formate species has been assigned, and this essentially
XPS and electrochemistry study cannot refer to this intermedi-
ate.4,11,13 (We did comment on the role of the formate on formic
acid electrocatalysis previously.4)
In this report, we use electrochemistry and X-ray photo-
electron spectroscopy (XPS) to probe the particle size and
In present heterogeneous catalysis research, the density
functional theory (DFT) is frequently used to account quanti-
tatively for metal/alloy behaviors in reacting chemical molecules
to desired products.16-20 In particular, it has been shown
theoretically17-19 that a change in the adsorbate chemisorption
energy scales directly with the change in the metal center of
the d-band: lowering of the d-band center results in the decrease
* To whom correspondence should be addressed. E-mail: andrzej@
scs.uiuc.edu.
† Department of Chemistry, University of Illinois at Urbana-Champaign.
‡ Department of Chemistry, Warsaw University.
§ Department of Chemical and Biochemical Engineering, University of
Illinois at Urbana-Champaign.
| Department of Chemistry, University of North Texas.
10.1021/jp061690h CCC: $33.50 © 2006 American Chemical Society
Published on Web 06/20/2006

13394 J. Phys. Chem. B, Vol. 110, No. 27, 2006
Zhou et al.
pretreatment and XPS from the UHV cleaned Pd(111). Practi-
cally identical XPS spectral features in Figure 1 indicate a clean
nanoparticle surface system. Furthermore, our XPS for Pd metal
and for the Pd nanoparticles do not show any of the features
that are associated with formation of PdHx.33 The layer of the
immobilized Pd nanoparticles was thick enough to completely
attenuate the XPS Au substrate signal, which simplifies
interpretations of the Pd XPS results. If not stated otherwise,
the real surface area of the Pd nanoparticle surface was estimated
from the known sample weight multiplied by the value of the
sample surface area per gram.
As a reference for the XPS, a MaTeck disk Pd(111) crystal
(8 × 2 mm) was used. The crystal surface was cleaned by
multiple cycles of argon ion bombardment and subsequently
hot argon ion bombardment, and then the crystal was annealed
at high temperature (including annealing in oxygen atmosphere
at 10-8 Torr), until no impurities were detected by XPS.
Chemicals used were H2SO4 (GFS, double distilled from
Vycor) and HCOOH (Sigma-Aldrich, Fluka for HPLC) and
Millipore water. Ultrahigh purity quality argon and CO gases
were supplied by S. J. Smith Welding Supply. Electrochemical
measurements were carried out using the Autolab G30 poten-
tiostat in a conventional three electrodes electrochemical cell
with platinized platinum gauze as a counter electrode and Ag/
AgCl in 3 M NaCl as the reference electrode. All potentials are
quoted with respect to RHE, and the experiments were
performed at room temperature.
Figure 1. Representative XPS survey spectra for (a) completely
reduced Pd nanoparticles of the 9 nm size and (b) clean Pd(111).
Results and Discussion
in adsorption energy (or the surface bond strength) of the
adsorbate to the substrate. In gas-phase research, the correlation
between the metal d-band center, the core-level binding energy,
and the catalytic activity is now well-established.21-24 Verifica-
tion of similar correlations in the field of electrocatalysis begins
to emerge.25,26
Representative, high-resolution transmission electron micros-
copy (HRTEM) images of the Pd nanoparticles in use are
presented in Figure 2. The histograms of the particle size
distribution indicate that the average particle sizes are 9 ( 2.5,
11 ( 3, and 40 ( 15 nm. Figure 3 shows steady-state cyclic
voltammograms (CV) for formic acid oxidation for the three
nanoparticle samples in 0.5 M HCOOH in 0.5 M H2SO4, at 20
mV/s. The data indicate a low inhibition rate of formic acid
oxidation, which is consistent with previous studies with
conventional Pd electrodes.3,34 Figure 4 shows chronoampero-
metric current-time plots obtained during formic acid oxidation
in a 0.5 M HCOOH + 0.5 M H2SO4 solution using the three
Pd catalysts at 0.27 V (the same trend was found until E ) 0.5
V, that is, until the end of utility of Pd as the formic acid
oxidation catalyst). Clearly, the 9 and 11 nm Pd nanoparticles
display higher activity both at the experiment beginning and
after 1 h of reaction. The decay trends are quite similar: ca.
70% of the initial activity remains after 1 h of reaction. After
the 1 h, the current density from the 9 and 11 nm Pd
nanoparticles is similar, for example, 0.2 mA cm-2. However,
the current density for the 40 nm nanoparticles is lower and is
0.15 mA cm-2 (see also the data in Table 1).
Figure 5 shows results from voltammetric CO stripping
experiments using a 9 nm nanoparticle sample in 0.5 M H2SO4
and at 5mV/s, with the CO stripping peak at 0.89 V. The same
CO stripping voltammograms (and the same CO oxidation peak
potential) were found for the remaining Pd nanoparticles.
Clearly, the CO oxidation on nanoparticle Pd is sample size
independent. Since a dependence of the formic acid electro-
oxidation activity upon reduction in the particle size is found
(Figure 4), we conclude that the activity increase is not due to
facilitating the CO oxidation process but rather due to the
reaction leading to CO2 formation directly.11,12
Experimental Section
The electrochemical UHV-XPS (UHV ) ultrahigh vacuum)
instrument consisting of the main UHV chamber has been
described before.27,28 (This measuring configuration prevents
an exposure of samples to air upon the sample transfer between
UHV and an electrochemical cell.) An ESCA M-probe, high-
resolution multichannel hemispherical electron analyzer (Surface
Science Instruments) equipped with a monochromatic Al KR
line, operated at 110 W and a constant pass energy of 25 eV,
was used for all XPS measurements. The S-probe version of
the 1.36 ESCA software (Fison Instruments) was used, and all
XPS spectral peaks reported herein were fitted using a mixed
Gaussian-Lorentzian line shape and Shirley baselines. The
linearity of the BE scale of the detector was calibrated using
Au 4f7/2 (84.0 eV), Ag 3d5/2 (368.25 eV), and Cu 2p3/2 (932.68
eV).29 For the Pd 3d and valence band regions, the exact BE
corrections were determined by comparison with the clean
Pd(111) surface, assuming that the Pd 3d5/2 BE is 384.10 eV
and that the Pd Fermi edge is at 0.0 eV.30,31
Pd particles, 9 and 11 nm, were purchased from Sigma-
Aldrich and Pd black (40 nm) from Alfa Aesar. A transmission
electron microscopy (TEM) instrument, JEOL 2010F, was used
to confirm the average size of the nanoparticles. Prior to
performing oxidation of formic acid, Pd nanoparticles were
immobilized on a gold disk,27,32 electrochemically reduced at
0.10 V, and further stabilized at 0.5 V for 20 min. The
pretreatment produced a contamination free surface, as dem-
onstrated by the XPS measurements. Figure 1 shows a typical
XPS survey for 9 nm Pd nanoparticle surfaces after the
The clean and reduced Pd nanoparticle samples were trans-
ferred from the electrochemical cell to the XPS chamber

Palladium Nanoparticles in Formic Acid
J. Phys. Chem. B, Vol. 110, No. 27, 2006 13395
Figure 2. TEM images of the palladium nanoparticles used for this study, along with the statistical results from the analyses of many different
regions of the particles. Histograms of the particle size distribution yield average particle sizes of (a) 9 ( 2.5, (b) 11 ( 3, and (c) 40 ( 15 nm. A
20 nm scale bar is applied to images (a) and (b) and the 100 nm to image (c).
according to the procedure described in the Experimental
Section.27,28 The comparison of the core-level Pd 3d XPS spectra
shows different electron core-level binding energy (BE) among
the samples (the Pd 3d5/2 data in Figure 6 and in Table 1).
Namely, the BE of Pd 3d from the 9 and 11 nm nanoparticles
is shifted by ca. +0.2 eV vs the 40 nm sample and the Pd bulk,
which is accompanied by the increase in the fwhm by 0.15 eV.
(All spectra and current densities were measured at least three
times, and the data in Table 1 are the average of three
independent measurements.)
the direction of the BE shift with particles size is consistent
with the shift in the VB center.21,22,37,38
The particle size effect on the metal electronic structure in
the particle size range investigated in this study was already
observed using metal nanoparticle nuclear magnetic resonance
(EC NMR).39 While dealing with Pt39 rather than with Pd, the
nanoparticle catalyst was unsupported as is the case here, and
the data informed us that the smaller the particle size, the smaller
the NMR frequency of “bulk” Pt. That is, the particle size had
a strong effect on the bulk Knight shift or on the “bulk” LDOS
at the Fermi level. We believe that this relates to the present
observation that the bulk electronic properties of the samples
measured by XPS are affected by the particle size (Figure 6).
More work is necessary to delineate possible correlations
between the Knight shift data and the present BE data. However,
the advantage in using large nanoparticles such as in this work
(ca. 10 nm and larger) is that final state changes in the relaxation
and screening of the core holes are not important since the
screening is expected to have converged to the bulk limit.40,41
Thus, our observed shifts are due to the initial state effect only.
Further, as we stated in the Experimental Section, we ruled out
the presence of impurities such as dissolved hydrogen or surface
carbon as being responsible for the observed BE shifts.
Therefore, we propose that the lattice strain is the origin of the
BE shifts to higher BE for the smallest Pd particles studied
(compared with the larger nanoparticles and the Pd single
crystal). However, the lattice strain does not have the same
physical origin as the 5-7% lattice strain observed for small
particles on inert supports,40,42-44 as discussed below.
Figure 7 shows the XPS valence band (VB) spectra from the
three Pd nanoparticle samples and from Pd(111). (The nano-
particle peak heights around EF are scaled to that of Pd(111).)
Notice that the broad peak associated with PdHx at 8 eV33 is
not found. The position of the center of the valence band is
given by ∫ N(ꢀ)ꢀ dꢀ/∫ N(ꢀ) dꢀ, where N(ꢀ) is the DOS or, in
our case, the XPS-intensity after background subtraction
(Experimental Section). To obtain the valence band centers, we
carried out the integration up to 10.0 eV BE with respect to EF.
The VB center for Pd(111) is 2.49 ( 0.06 and is in good
agreement with the value of 2.39 eV obtained from the gradient-
corrected DFT periodic slab calculations by Pallassana et al.35
This demonstrates the validity of our method. The largest
particles (40 nm) do not show a VB center change vs Pd(111).
However, the smallest nanoparticles (9 and 11 nm) exhibit the
VB center shift of ca. +0.2 eV vs the Pd bulk. Wertheim et
al.36 and also Mason et al.37 reported a trend in the VB-band
center shift with decreasing size of Pd clusters (vapor deposited
on amorphous carbon substrates) that we now confirm. Clearly,

13396 J. Phys. Chem. B, Vol. 110, No. 27, 2006
Zhou et al.
TABLE 1: Steady-state Current Density for Formic Acid
Oxidation (measured at 1 h, see text), the Center of Valence
Band (VB), and the Core-Level Binding Energy (BE) of the
Pd 3d5/2 Level for Palladium
steady-state chronoamperometric current,
XPS peak of Pd 3d5/2, and center of VB band
for different sizes of Pd particles
steady-state
CA after 1 h
center of VB
(eV)
BE of Pd 3d5/2
(eV)
sample
(mA cm-2
)
Pd (111)a
not measured
0.15 ( 0.01
0.21 ( 0.02
0.21 ( 0.02
2.49 ( 0.06
2.49 ( 0.06
2.59 ( 0.06
2.64 ( 0.06
335.10 ( 0.03
335.12 ( 0.03
335.25 ( 0.03
335.33 ( 0.03
Pd (40 nm)b
Pd (11 nm)c
Pd (9 nm)d
a The data for Pd(111) as a reference for the XPS measurements.
b Pd black (particle size ) 40 nm). c Pd nanoparticles (11 nm). d Pd
nanoparticles (9 nm). The data were obtained from three independent
measurements and are given at a 95% confidence level.
Figure 3. Cyclic voltammetric curves taken in the positive direction
(solid lines) and the negative direction (long dashed lines) for (a) 9,
(b) 11, and (c) 40 nm Pd catalysts in 0.5 M HCOOH in 0.5 M H2SO4
solution at 20 mV/s.
Figure 5. Voltammetric CO stripping with Pd nanoparticles (9 nm)
in 0.5 M H2SO4. The scan rate was V ) 5 mV/s. Shown are the first
scan (black solid line) starting at 0.13 V and the second scan (red dot
line). The electrode was exposed to dissolved CO for 10 min at 0.13
V followed by Ar bubbling for 20 min.
chemistry. To make this correlation, we can bring to bear a
system that involves shifts of the BEs of metal nanoparticles
supported on inert, typically oxide substrates. Here, it is known
that the BEs of supported metal particles shift to lower BE with
increasing size36,40,45-48 by ∼1 eV. Although the range of sizes
of particles on inert supports are typically ∼1 nm,40 these
simpler systems are useful to establish the mechanisms relevant
for the BE shifts observed with the size of our present Pd
particles. Notice the key findings of Richter et al.40 to show
that two mechanisms contribute, about equally, to the ∼1 eV
BE shift. The first is the final state relaxation that screens the
core hole; this relaxation grows in magnitude as the particle
size increases. The second is an initial state effect arising from
the lattice strain that exists as bond distances are shorter in
smaller particles.42,43 Furthermore, it was shown that the lattice
strain affects the BEs primarily because the d-hybridization
changes with the lattice strain.40,44 We now use the concept of
hybridization to encompass all d-shell chemical bonding effects
that act to change the number and character of the d electrons
Figure 4. Formic acid oxidation current densities measured for three
sizes of Pd nanoparticles in 0.5 M HCOOH in 0.5 M H2SO4 at 0.27 V
vs RHE for 1 h reaction time. The current densities shown correspond
to Pd catalysts with the size of 9 nm (black line), 11 nm (red line), and
40 nm (blue line). The electrochemically active surface area of the Pd
nanoparticle film was estimated from the known sample weight
multiplied by the known sample surface area per gram. The plot was
constructed after a five adjacent points smoothing procedure was
applied.
It is important to correlate the shifts of the XPS BEs for
different particle sizes with the electronic structure of the
particles since this helps us to relate the XPS to the nanoparticle

Palladium Nanoparticles in Formic Acid
J. Phys. Chem. B, Vol. 110, No. 27, 2006 13397
Figure 7. Representative valence band spectra for three Pd nanoparticle
samples of the size of (a) 9, (b) 11, and (c) 40 nm and for (d) clean
Pd(111). The lines shown were smoothed using a five-points adjacent
average on the original lines. The peak heights around EF of the VB
spectra are scaled to that of Pd(111) (see text).
Figure 6. Representative XPS spectra for the Pd 3d core levels of the
completely reduced Pd nanoparticles with the size of (a) 9, (b) 11, and
(c) 40 nm and for (d) clean Pd(111). The maximum of the Pd 3d5/2
peak is used as the measure of the binding energy and is indicated by
the vertical lines. The Pd sample reduction was carried out by holding
the electrode at 0.1 V vs RHE for 20 min. See also Table 1.
reactivity. Norskov and co-workers17-19 have shown that
changes in the adsorbate chemisorption energies scale directly
with the changes in the center of the catalyst d-band: lowering
the d-band center (moving down and away from the Fermi level)
results in the decrease in the bond strength of adsorbates (or in
their chemisorption energy).
around the XPS ionized atom. Further, there is strong and
general evidence that relates the degree of hybridization to shifts
of the core-level BEs and to shifts of the d-band center.40,44,46,49
The key point is that the d-hybridization is known to affect the
XPS BEs for a wide range of cases.40,44,46,49 Furthermore, our
recent work has established the connection between the lattice
strain and the d-hybridization.40,44
The fact that the BE shifts depend on the extent of d-
hybridization is a very important observation relative to how
the BE shifts may reflect the chemical activity of the system.
This relationship occurs because d-hybridization which forms
stronger bonds between the metal atoms reduces the potential
of the metal to form strong bonds with adsorbed reactants. In
turn, this change of the adsorbate bond strength to a nanoparticle
is, as we discuss below, a key factor for the reactivity. We
invoke the lattice strain as the origin of the hybridization and
the change of reactivity because of the evidence that the strain
leading to a lattice contraction is consistent with an increase of
core-level BEs,40,41 exactly as we measure for our Pd particles;
see Table 1.
Notice that, for formic acid oxidation to CO2, we have the
following set of reactions to consider4,11,13
HCOOHbulk f 2H+ + 2e- + CO2
(1)
(2)
(3)
HCOOHads f (COOH)ads + H+ + e-
(COOH)ads
9
rds
8 CO2 + H+ + e-
We believe that the meaning of the experimental and
theoretical findings is as follows. The rate of formic acid
transformation to CO2 (reaction 1) is the highest at the surface
that is free of (COOH)ads. Since the smallest Pd nanoparticles
have the lowest d-band center, they bind the COOH intermediate
less strongly (vs larger nanoparticles) and reduce the surface
(COOH)ads coverage (reactions 2 and 3). If the limiting step of
reaction 3 is avoided because of the low COOH coverage, then
higher rates of the direct HCOOH decomposition process to
CO2 via reaction 1 are expected. Notice however that this
interpretation is limited to few nanoparticle sizes selected for
this project. If using a broader range of Pd nanoparticle sizes
will show the rate to maximize at an optimized particle size,50
interpretations involving a volcano behavior will have to be
invoked.26
To reiterate, the data in Table 1 show that the most active
catalyst for the formic acid oxidation is made of the smallest
Pd nanoparticles (the increase in the current density is from
0.15 to 0.21 mA cm-2). Such smallest Pd nanoparticles (9 and
11 nm) display, at the same time, the highest BE shift and the
highest valence band (VB) center downshift with respect to the
Fermi level. In other words, the center of the d-band is reduced
and the smallest nanoparticles display the highest formic acid
Conclusions
Our measurements demonstrate a Pd particle size effect on
the XPS spectra and on the rate of HCOOH oxidation reaction.

13398 J. Phys. Chem. B, Vol. 110, No. 27, 2006
Zhou et al.
(15) Corrigan, D. S.; Weaver, M. J. J. Electroanal. Chem. 1988, 241,
143.
(16) Hammer, B.; Norskov, J. K. AdV. Catal. 2000, 45, 71.
(17) Hammer, B.; Norskov, J. K. Surf. Sci. 1995, 343, 211.
(18) Ruban, A.; Hammer, B.; Stoltze, P.; Skriver, H. L.; Norskov, J.
K. J. Mol. Catal. A: Chem. 1997, 115, 421.
(19) Mavrikakis, M.; Hammer, B.; Norskov, J. K. Phys. ReV. Lett. 1998,
81, 2819.
(20) Greeley, J.; Norskov, J. K.; Mavrikakis, M. Annu. ReV. Phys. Chem.
2002, 53, 319.
(21) Ganduglia-Pirovano, M. V.; Natoli, V.; Cohen, M. H.; Kudrnovsky,
Due to the size of the nanoparticles in use, only the initial state
effect in the XPS measurements has to be considered. We could
then provide a line of evidence that the initial state effect results
from the lattice strain change that occurs as bond distances are
shorter in smaller nanoparticles. Therefore, the lattice strain
effect and its electronic-level consequences emerge as the key
driving force in the catalysis, at least in reacting HCOOH to
CO2. These electronic-level factors that are affected by the lattice
strain change are the core-level binding energies, the related
position of the center of the d-band, and the Knight sifts (or
local DOS) in the EC NMR measurements.39
J.; Turek, I. Phys. ReV. B 1996, 54, 8892.
(22) Hennig, D.; Ganduglia-Pirovano, M. V.; Scheffler, M. Phys. ReV.
B 1996, 53, 10344.
(23) Rodriguez, J. A. Surf. Sci. Rep. 1996, 24, 223.
(24) Rodriguez, J. A.; Goodman, D. W. Science 1992, 257, 897.
(25) Zhang, J. l.; Vukmirovic, M. B.; Xu, Y.; Mavrikakis, M.; Adizic,
R. R. Angew. Chem., Int. Ed. 2005, 44, 2132.
(26) Kibler, L. A.; el-Aziz, A. m.; Hoyer, R.; Kolb, D. M. Angew. Chem.,
Int. Ed. 2005, 44, 2080.
(27) Lewera, A.; Zhou, W. P.; Vericat, C.; Chung, J. H.; Haasch, R.;
Wieckowski, A.; Bagus, P. S. Electrochim. Acta 2006, 51, 3950.
(28) Vericat, C.; Wakisaka, M.; Haasch, R.; Bagus, P. S.; Wieckowski,
A. J. Solid-State Electrochem. 2004, 8, 794.
(29) Seah, M. P.; Gilmore, I. S.; Beamson, G. Surf. Interface Anal. 1998,
26, 642.
(30) Anthony, M. T.; Seah, M. P. Surf. Interface Anal. 1984, 6, 95.
(31) Wagner, C. D.; Naumkin, A. V.; Kraut-Vass, A.; Allison, J. W.;
Powell, C. J.; Rumble, J. R., Jr. NIST Standard Reference Database 20,
Version 3.4 (Web Version), http://srdata.nist.gov/xps/.
(32) Waszczuk, P.; Solla-Gullon, J.; Kim, H.-S.; Tong, Y. Y.; Montiel,
V.; Aldaz, A.; Wieckowski, A. J. Catal. 2001, 203, 1.
(33) Bennett, P. A.; Fuggle, J. C. Phys. ReV. B: Condens. Matter Mater.
Phys. 1982, 26, 6030.
To be more specific, we have found that the most active Pd
nanoparticle catalyst in formic aid electrooxidation is made of
the smallest particles. Such smallest particles display the highest
binding energy shift and the highest valence band center
downshift with respect to Fermi level, which is consistent with
enhancement of the d-band hybridization in small nanoparti-
cles.With the formic acid oxidation mechanism represented by
eqs 1-3 (see above), we interpret our data as follows: the lower
d-band center results in a decrease in adsorption energy of the
formate intermediate that enhances the rate of the HCOOH
molecule decomposition via the direct path. This conjecture is
consistent with predictions of the modern theory of heteroge-
neous catalysis concerning the center of the d-band shifts of
metal and alloy catalysts in reacting simple organic molecules
to products.
(34) Pavese, A.; Solis, V.; Giordano, M. C. J. Electroanal. Chem. 1988,
245, 145.
Acknowledgment. This work is supported by the National
Science Foundation Grant No. CHE 03-49999 and is also
supported by the U.S. Department of Energy under Award No.
DEGF-02-99-ER14993.
(35) Pallassana, V.; Neurock, M.; Hansen, L. B.; Hammer, B.; Norskov,
J. K. Phys. ReV. B 1999, 60, 6140.
(36) Wertheim, G. K.; DiCenzo, S. B.; Buchanan, D. N. E. Phys. ReV.
B 1986, 33, 5384.
(37) Mason, M. G.; Gerenser, L. J.; Lee, S.-T. Phys. ReV. Lett. 1977,
39, 288.
References and Notes
(38) Citrin, P. H.; Wertheim, G. K. Phys. ReV. B 1983, 27, 3176.
(39) Babu, P. K.; Oldfield, E.; Wieckowski, A.; In Modern Aspects of
Electrochemistry; Vayenas, C. G., Conway, B. E., White, R. E., Gamboa-
Adelco, A. E., Eds.; Kluwer Academic Publishers: New York, 2003; p 1.
(40) Richter, B.; Kuhlenbeck, H.; Freund, H.-J.; Bagus, P. S. Phys.
ReV. Lett. 2004, 93, 026805.
(1) Rice, C.; Ha, S.; Masel, R. I.; Waszczuk, P.; Wieckowski, A.;
Barnard, T.; J. Power Sources 2002, 111, 83.
(2) Rice, C.; Ha, S.; Masel, R. I.; Wieckowski, A. J. Power Sources
2003, 115, 229.
(3) Baldauf, M.; Kolb, D. M. J. Phys. Chem. 1996, 100, 11375.
(4) Lu, G.-Q.; Crown, A.; Wieckowski, A. J. Phys. Chem. B 1999,
103, 9700.
(5) Waszczuk, P.; Barnard, T. M.; Rice, C.; Masel, R. I.; Wieckowski,
A. Electrochem. Commun. 2002, 4, 599.
(6) Zhao, M. C.; Rice, C.; Masel, R. I.; Waszczuk, P.; Wieckowski,
A. J. Electrochem. Soc. 2004, 151, A131.
(7) Ha, S.; Larsen, R.; Masel, R. I. J. Power Sources 2005, 144, 28.
(8) Zhu, Y. M.; Khan, Z.; Masel, R. I. J. Power Sources 2005, 139,
15.
(9) Larsen, R.; Zakzeski, J.; Masel, R. I. Electrochem. Solid-State Lett.
2005, 8, A291.
(10) Henry, C. R. Surf. Sci. Rep. 1998, 31, 231.
(41) Bagus, P. S.; Nelin, C. J.; Kay, E.; Parmigiani, F. J. Electron
Spectrosc. Relat. Phen. 1987, 43, C13.
(42) Klimenkov, M.; Nepijko, S.; Kuhlenbeck, H.; Baumer, M.; Schlogl,
R.; Freund, H.-J. Surf. Sci. 1997, 391, 27.
(43) Nepijko, S. A.; Klimenkov, M.; Adelt, M.; Kuhlenbeck, H.; Schlogl,
R.; Freund, H.-J. Langmuir 1999, 15, 5309.
(44) Bagus, P. S.; Wieckowski, A.; Freund, H.-J. Chem. Phys. Lett. 2005,
420, 42.
(45) Egelhoff, W. F., Jr. Surf. Sci. Rep. 1987, 6, 253.
(46) Mason, M. G. Phys. ReV. B 1983, 27, 748.
(47) Wertheim, G. K. Z. Phys. D: At., Mol. Clusters 1989, 12, 319.
(48) Wertheim, G. K. Z. Phys. B 1987, 66, 53.
(49) Bagus, P. S.; Illas, F.; Pacchioni, G.; Parmigiani, F. J. Electron
Spectrosc. Relat. Phenom. 1999, 100, 215.
(11) Capon, A.; Parsons, R. J. Electroanal. Chem. 1973, 45, 205.
(12) Vela, M. E.; Lezna, R. O.; De Tacconi, N. R.; Arvia, A. J.; Beden,
B.; Hahn, F.; Lamy, C. J. Electroanal. Chem. 1992, 323, 289.
(13) Samjeske, G.; Osawa, M. Angew. Chem., Int. Ed. 2005, 44, 5694.
(14) Kunimatsu, K.; Kita, H. J. Electroanal. Chem. 1987, 218, 155.
(50) Masel, R. I. Principles of adsorption and reaction on solid surfaces;
Wiley-Interscience: New York, 1996.