CO and H2O adsorption and reaction on Au(310)

Article abstract of DOI:10.1016/j.susc.2011.06.006

We have studied desorption of ¹³CO and H₂O and desorption and reaction of coadsorbed, ¹³CO and H₂O on Au(310). From the clean surface, CO desorbs mainly in, two peaks centered near 140 and 200 K. A complete analysis of desorption spectra, yields average binding energies of 21 ± 2 and 37 ± 4 kJ/mol, respectively. Additional desorption states are observed near 95 K and 110 K. Post-adsorption of H ₂O displaces part of CO pre-adsorbed at step sites, but does not lead to CO oxidation or significant shifts in binding energies. However, in combination with electron irradiation, ¹³CO₂ is formed during H₂O desorption. Results suggest that electron-induced decomposition products of H₂O are sheltered by hydration from direct reaction with CO.

Full text of DOI:10.1016/j.susc.2011.06.006

Surface Science 605 (2011) 1726–1731
Contents lists available at ScienceDirect
Surface Science
journal homepage: www.elsevier.com/ locate/susc
CO and H O adsorption and reaction on Au(310)
2
M.E. van Reijzen, M.A. van Spronsen, J.C. Docter, L.B.F. Juurlink ⁎
Leiden Institute of Chemistry, Leiden University, PO BOX 9502, 2300 RA, Leiden, The Netherlands
a r t i c l e i n f o
a b s t r a c t
We have studied desorption of 13CO and H O and desorption and reaction of coadsorbed, CO and H O on
13
Article history:
2
2
Received 7 April 2011
Accepted 3 June 2011
Available online 15 June 2011
Au(310). From the clean surface, CO desorbs mainly in, two peaks centered near 140 and 200 K. A complete
analysis of desorption spectra, yields average binding energies of 21± 2 and 37± 4 kJ/mol, respectively.
Additional desorption states are observed near 95 K and 110 K. Post-adsorption of H
pre-adsorbed at step sites, but does not lead to CO oxidation or significant shifts in binding energies. However, in
combination with electron irradiation, 13CO
is formed during H O desorption. Results suggest that electron-
induced decomposition products of H O are sheltered by hydration from direct reaction with CO.
© 2011 Elsevier B.V. All rights reserved.
2
O displaces part of CO
Keywords:
Gold
Carbon monoxide
Water
2
2
2
Surface chemical reaction
Stepped single crystal surface
1
. Introduction
2
Although hydroxyl formation from O+H O is observed on many
other transition metals [15–17] and may be expected to be part of the
explanation of gold's capability to oxidize CO at low temperatures, it is
not obvious that its occurrence on Au(111) reflects what happens on
real catalyst particles. First, such particles are only active in CO
oxidation when smaller than 5 nm [3] and at such diameters large
(111) domains are not abundant at the catalyst's surface. Second,
small particles contain many low coordinated sites, for example at the
Catalytic activity of gold has been a topic of research since the
960s [1]. However, it has only attracted significant attention since
1
Haruta's discovery of Au nanoparticles' high activity for CO oxidation
2]. Since this pioneering work, Au nanoparticles have been found to
catalyze many more reactions [3–6]. In parallel, surface science
studies have investigated the reaction mechanisms underlying gold's
remarkable reactivity using well-ordered Au single crystal surfaces.
These studies have been reviewed recently [7–10].
[
2
border between two facets, where the formation of OH from O+H O
may not occur to a large extent. For two stepped Pt surfaces, we have
recently shown that the (111) terraces, (110) steps and (100) steps
have strongly varying tendencies to producing OH from coadsorption
The origin of the high catalytic activity of Au nanoparticles toward
CO oxidation, and in particular the promotional effect of H O on the
2
oxidation rate, are still under debate [11–13]. Based upon catalytic
studies of supported Au particles, Daté and Haruta suggested direct
of O+H
In this article, we investigate the influence of low coordinated Au
atoms on coadsorbed CO and H O. We use the Au(310) surface, which
consists of 3-atom wide (100) terraces with monoatomic (110) steps.
The surface therefore provides 6-, 8- and 9-fold coordinated atoms, as
indicated in Fig. 1a, to interact with adsorbates. We use TPD, Auger
electron spectroscopy (AES), and low energy electron diffraction
(LEED) in our investigations of adsorption and desorption.
2
O [18].
reaction of H
the perimeter of nanoparticles as a possible means for H
reactivity [11]. Another suggested mechanism was the facilitation of
carbonate (CO ) dissociation by water, again accompanied by
2
O with O
2
yielding OH groups and an activated O atom at
2
2
O to increase
3
formation of hydroxyl groups. Recent combinations of density
functional theory (DFT) studies and molecular beam experiments
[12], near-edge X-ray absorption fine structure (NEXAFS) and infra-
red (IR) studies [13], and temperature programmed desorption (TPD)
and IR studies [14] using a Au(111) single crystal surface show
2. Experimental
unambiguously that H
2
O and surface-bound atomic O form OH
Experiments are carried out using a home-built ultrahigh vacuum
−10
groups. These OH groups readily oxidize CO at a surface temperature
of 77 K using a CO molecular beam [12]. The reaction is suggested to
proceed via an unstable HO–CO intermediate [13].
(UHV) system with a base pressure of 1×10
mbar during experi-
ments. The UHV chamber and manipulator are constructed for studies of
single crystal samples. The chamber is equipped with two quadrupole
mass spectrometers (QMS). One QMS (Baltzers, Prisma 200) protrudes
into the main chamber and is used for residual gas analysis and angle-
integrating TPD spectroscopy. The other QMS (UTI 100c) is differentially
pumped and probes desorption of molecules from the sample through a
3 mm diameter opening, positioned ~2 mm away from the face of the
⁎
Corresponding author. Tel.: +31 71 527 4221; fax: +31 84 7182999.
E-mail address: l.juurlink@chem.leidenuniv.nl (L.B.F. Juurlink).
0039-6028/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.susc.2011.06.006

M.E. van Reijzen et al. / Surface Science 605 (2011) 1726–1731
1727
thermocouple is electrically decoupled from the PID controller (Euro-
therm type 2416) that controls temperature ramps.
+
The Au(310) surface was daily cleaned by repetitive cycles of Ar
(
5 N) sputtering (5–10 min at 500 V and 2 μA) and 1–5 minute
annealing at 860 K. LEED and AES are periodically used to check
surface order and cleanliness. Fig. 1b shows a typical LEED image of
the cleaned surface, which reflects the exact same spot pattern as
observed previously for Ni(310) [20]. The ratio of spot row spacing to
spot splitting is consistent with the expected value from an
unreconstructed surface (1.58 measured vs. 1.58 calculated) [21].
The only clearly distinguishable peak in our AES spectra up to 1 kV is
the Au transition at 69 eV.
For CO adsorption studies, isotopically labeled 13CO (MSD iso-
topes, 99.7% 13C) was introduced using a leak valve with an elongated
nozzle inside the UHV chamber. A short distance between sample and
nozzle ensured localized deposition of CO. The sample was flashed to
~
280 K prior to CO dosing and kept at 89 K during dosing. Before
initiating the TPD ramp, the crystal was cooled to 82 K to ensure
linearity in the temperature ramp from 90 K upward. As CO has a high
sticking coefficient at the dosing temperature, only a modest increase
in background pressure is observed during dosing. Generally, we dose
CO with a measured increase in background pressure in the range of
−
10
−9
1
×10
order of 10
was limited.
to 1×10
mbar. As the background pressure is on the
−
10
mbar, reproducibility in dosing small amounts of CO
High purity water (Millipore, R=18.2 MΩ) was dosed onto the
crystal using a home-built capillary array doser [22]. This doser
ensures uniform exposure of the single crystal surface to H
significant increase in the background pressure at m/z=18 in
repetitive experiments. The H O was degassed by freeze–pump–
thaw cycles and kept in a container with He (6 N) at a total pressure of
.5 bar. The water container was kept at 40°C using a warm water
bath to ensure a constant partial pressure of water in the H O/He mix.
Monitoring the pressure rise in the UHV chamber, which results
mostly from co-dosed He, allows for reproducible H O coverages.
For coadsorption studies, the crystal was flashed to ~280 K, and CO
and H O were consecutively dosed at a maximum temperature of 89 K.
Again, prior to initiating the TPD ramp, the crystal was cooled to 82 K.
2
O without
2
1
2
2
2
−
1
−1
Desorption traces were taken at rates of 0.9 Ks
and 1.8 Ks
.
TPD spectra were linear in the range of 90 K to 240 K and were
measured using the differentially pumped QMS with an ion energy of
7
0 V and emission current of 1 mA. TPD traces where recorded with
Labview 8.6 based, home-built software and a 12-bit ADAC converter
NI USB-6008).
(
3
. Results and discussion
Fig. 1. (a) Top and side views of the Au(310) surface with coordination number per Au
surface atom type indicated (b) color-inverted LEED image at 190 eV.
3.1. CO desorption from Au(310)
sample during TPD measurements. The apparatus also contains
equipment to perform LEED (VG RVL 900) and AES (Staib Instruments
ESA 100). Finally, it contains two directional dosers that allow for
localized gas dosing onto the sample.
Fig. 2a shows desorption traces of several initial coverages of 13CO.
The TPD traces are characterized by two main features located near
200 K (α) and 140 K (β ). The β feature appears prior to saturation of
1
1
α. For both features the peak desorption temperature shifts slightly
For the experiments described here, the sample is a 2 mm thick,
downward with increasing 13CO coverage. At the highest doses, a
1
0 mm diameter Au single crystal, cut and polished to within 0.1° of the
2 3
third (β ) and fourth (β ) feature appear at 110 and 95 K,
13
(
310) surface (purity 5 N, Surface Preparation Labs). It is welded at its
respectively. LEED images taken at various coverages of CO show
no order other than that of the bare surface (Fig. 1b). The intensity of
the spot pattern decreases with CO coverage. The lack of evidence for
a particular overlayer structure unfortunately does not allow us to
quantify CO adsorption beyond relative values.
back edge to a 1 mm wide U-shaped polycrystalline Au ribbon. The
ribbon is connected to a small copper block. This block is electrically
isolated by aluminum nitride plates from a larger copper arm that
connects the sample to the cryostat. The sample may be cooled to
approximately 77 K when supercooling the cryostat's LN
2
reservoir [19].
Our CO TPD spectra are in good agreement with results published by
Weststrate et al. for the same Au(310) surface [23]. They find desorption
features near 135 and 185 K in TPD spectra taken at a rate of 5 K s− . In
combination with XPS results, they conclude that both TPD features
result from step edge desorption. CO desorption from stepped surfaces
with 6- and 7-fold coordinated Au atoms is more often found near 190 K
For the experiments described here, the crystal is heated radiatively by a
filament (Osram, 250 W) mounted behind the sample, although
electron bombardment heating is also possible. The crystal temperature
is measured by a chromel/alumel thermocouple spot welded to the top
edge of the crystal in between the legs of the U-shaped gold ribbon. The
1

1728
M.E. van Reijzen et al. / Surface Science 605 (2011) 1726–1731
The Redhead analysis yields desorption energies, Edes, of approxi-
−
1
1
mately 54 and 37 kJ mol for the α and β features when assuming the
same frequency factor as used previously [23]. However, the complete
analysis yields different results. For α we have used the original data. To
analyze β , α must be subtracted, as well as β and β (when present).
1 2 3
As these features overlap and the line shapes are not known,
indisputable deconvolution is impossible. We have used Gaussian fits
as shown in the inset of Fig. 2a, realizing that our results for β
represent no more than a crude estimate of the desorption energy. In
Fig. 3 we present the results for α (solid symbols) and β (open
1
therefore
1
symbols). We have analyzed TPD spectra at multiple coverages relative
to saturation of each peak and included all results from spectra taken at
0
1
.9 K/s and 1.8 K/s. For α and β we find significant scatter in the
obtained values of Edes, but no significant drop with increasing relative
coverage. An unconstrained linear fit to the values is nearly flat, which is
consistent with only minor reductions of the peak desorption
temperatures with varying CO coverage (Fig. 2a). The desorption
energy averaged over the whole range of relative coverage for each
individual feature is indicated for α and β
1
as dashed lines. They are
found to be 37± 4 and 21± 2 kJ mol , respectively. We have not
analyzed β and β in a similar fashion as the extent of convolution of
−
1
2
3
these peaks prohibits determining desorption energies with reasonable
accuracy.
Fig. 2. (a) TPD spectra of 13CO; (inset) example of TPD spectrum deconvolution with
original data (dotted) and fits (solid lines and areas); (b) peak integrals as function of
dose (for α open squares, for β solid squares, for total of α and β closed circles).
1 1
Table 1 compares our desorption energies to those published
previously for the same Au(310) surface by Weststrate et al., which
were based on a Redhead analysis of TPD data [23], and binding
energies found by Hussain et al. [31], who, from DFT calculations, find
the strongest adsorption site for CO to be the 6-fold coordinated Au
atom in the step edge. The CO binding energies in Table 1 for the Au
(310) surface with its 6-fold coordinated atoms show significant
discrepancy, the complete analysis yielding lower values. For
adsorption to 7-fold coordinated Au atoms in the (211) [24] and
and 140 K [23–25]. It is noteworthy that CO desorption from the Au
110)-(1×2) reconstructed surface shows the highest desorption
(
feature around 160 K, dropping continuously with increased CO dose
to well below 100 K, even though it also contains 7-fold coordinated Au
atoms [26]. Recent DFT-based calculations suggest that only desorption
from low coordinated Au atoms is expected above 100 K [27]. We
−
1
therefore continue our studies with the assumption that the α and β
1
(322) [25] surfaces, binding energies near 50 kJ mol
have been
features result from step edge desorption. The different peak temper-
atures may result, for example, from two different densities of CO
adsorbed at step sites or varying adsorption sites at the step edge.
Although any assignment without access to data from other surface
sensitive techniques is speculative, we note that the desorption
reported using Redhead analysis of TPD spectra. From DFT studies, CO
binding energies for 6-fold coordinated Au atoms have been reported
varying from ~35 [32] to ~73 kJ/mol [27,33]. For 7-fold coordinated
Au atoms it ranges from 29 [32] to 63 kJ mol [27]. It seems that
there is no consensus yet from either theoretical or experimental
studies on the binding energy for CO to low coordinated Au atoms.
Our results indicate that care must be taken in the type of analysis
used to extract desorption energies from experimental data.
−
1
temperature of β
27].
We have deconvoluted the TPD spectra using Gaussian line shapes.
3
seems consistent with desorption from terrace sites
[
Although this functional form does not properly reflect the time-
dependent desorption rate, it provides a reasonably accurate estimate
of peak areas, as shown by the example provided as an inset in Fig. 2a.
3.2. H O desorption from Au(310)
2
In Fig. 2b, we present the evolution of the α and β
1
peak areas as a
function of CO exposure. The solid line is a fit to the summed areas of
the α and β peaks using a Langmuir adsorption function for non-
dissociative adsorption. It includes data not shown here up to CO
Fig. 4 shows TPD spectra of water desorbing from the bare Au(310)
surface. It is noteworthy that we observe desorption in two separate
peaks. The earliest studies of water desorption from a gold single
1
3
1
13
−
7
exposure of 12 10
using the same functional form. The dashed line equals the difference
and nicely fits the β data. From the fits we can conclude that non-
dissociative Langmuirian adsorption describes build-up of CO with
exposure well, and that α and β are nearly identical in size. The β
and β features likely give rise to the apparent larger size of β in
mbar s. The dotted line is a fit to the data of α
1
13
1
2
3
1
comparison to the α feature in TPD spectra. Weststrate et al. based
their total adsorption curve on the C1s XPS intensity and fitted it with
a linear increase that saturates [23]. This method inherently assumes a
coverage-independent sticking probability when averaged over a
room temperature kinetic distribution. Our data suggests that this
assumption is not correct.
We further analyze our TPD data using a Redhead analysis [28] and
by applying a complete desorption rate analysis [29,30] based on the
1
Polanyi–Wigner equation, to α and β :
dθ
n
−ðEdes =RTÞ
Fig. 3. CO desorption energies as a function of relative coverage for α (solid symbols)
and β1 (open symbols). Dashed lines indicate averages.
r = − dt = νðθÞ ꢀ θ ꢀ e

M.E. van Reijzen et al. / Surface Science 605 (2011) 1726–1731
1729
Table 1
study of water adsorption on Au(321), which contains similar 6-fold
coordinated Au atoms in the step edge, also find these to be the most
Comparison of CO desorption energies for Au(310).
E
des (α)
E
des (β
1
)
2
favorable adsorption sites [38]. The binding energy for a single H O
−
1
−1
kJ mol
kJ mol
molecule is calculated to be 23 kJ/mol, whereas adsorption at the (111)
terrace is less than10 kJ/mol. The resemblance of the lower temperature
feature in our spectra to desorption from multilayers or clusters suggests
that water only forms 2- or 3-dimensional structures after adsorption at
step edges has nearly saturated. The zeroth-order desorption kinetics for
Current research
Red Head analysis
Current research
complete analysis
Weststrate et al.
Ref. [23]
54
37
37± 5
46± 6
70
21± 3
35± 4
–
2
H O bound to step edges is somewhat surprising, as it does not occur on,
Hussain et al.
Ref. [31]
for example, the (110) and (100) steps of Pt surfaces [37]. However, it is
not uncommon to observe zeroth-order desorption kinetics in water
desorption from (sub)monolayer coverages. It may result from, for
example, the coexistence of a two-phase system on the surface at
thermodynamic equilibrium [39].
The inset in Fig. 4 shows small changes in the desorption spectrum
that occur when electrons are allowed to impinge from the QMS
filament prior to and during the temperature ramp. Quiller et al. have
crystal surface were performed on Au(110) [34] and Au(111)[35]. For
Au(111), a single desorption feature appears near 150 K which
exhibits zeroth-order desorption kinetics. It was interpreted to
indicate formation of 3-dimensional clusters at low coverage and,
therefore, that this surface is hydrophobic [12,34]. STM studies have
shown that initial adsorption occurs in 2-dimensional islands [36].
Outka and Madix reported two very closely spaced peak temperatures
in TPD spectra of H O desorbing from Au(110) at 185 and 190 K [34].
2
These results are not consistent with related systems, such as Ag(110)
and Cu(110), and desorption temperatures near 200 K mostly occur
2
also noticed that H O on Au(111) is affected by electron impact [14].
In general, electrons may cause electron stimulated desorption (ESD)
of water and dissociative electron attachment (DEA), where the latter
has been used to generate OH groups on various metal surfaces that
do not dissociate water [16]. We have recently done the same to study
OH groups on Ni(111) [40]. We expect that the small changes
observed in the TPD spectra in the inset of Fig. 4 result from a minor
amount of water decomposition and, consequently, formation of O or
2
when OH–H O coadsorbed layers are formed from reaction with pre-
adsorbed O or from electron-induced dissociation [16,17].
Our TPD spectra suggest two different adsorption states for water.
The peak desorption temperature of the feature appearing at the lowest
OH within the H
attempted to identify H
2
O layer adsorbed to Au(310) surface. We have
formation during electron irradiation but
2
2
H O doses shifts from 158 K to 172 K. The overlapping leading edges
find no measurable increase in the partial pressure of m/z=2 during
indicate zeroth-order desorption kinetics. Prior to saturation, a second
desorption feature appears initially as a shoulder at ~165 K. This feature
electron irradiation.
does not saturate even at very large H
closely examining the TPD traces for these higher H
2
O doses (not shown here). When
O doses, we again
3.3. Desorption of coadsorbed CO and H
2
O from Au(310)
2
observe a series of overlapping leadingedges but startingnear 150 K and
exhibiting a steeper onset than those for the higher temperature feature.
As water is generally found to initially bind to defects and step edges on
other metals that do not dissociate water [17,37], it seems likely that the
higher temperature feature in our spectra results from an adsorption
state associated with the 6-fold coordinated Au atoms. A recent DFT
To study coadsorption of H O and CO, the Au(310) surface was
2
1
3
consecutively exposed to
corresponds to a coverage of approximately 25% of the α feature
shown in Fig. 2a. For the TPD traces shown in Fig. 5a (H O) and b (CO),
2
CO and H O. The amount of CO
2
the Au sample was kept at a negative bias of −49 V during dosing and
measuring to prevent electron-induced reactions. Traces for increas-
ing amounts of H
desorption of the pure adsorbates. First, with increasing dose of H
2
O show two interesting changes in comparison to
O,
2
Fig. 5. TPD spectra of (a) H
and varying H O coverage. The inset in (a) compares H
without (dashed) coadsorbed CO.
2
O and (b) 13CO for coadsorption with a fixed CO coverage
O desorption with (solid) and
Fig. 4. H
2
O TPD spectra of various H
2
O doses. (inset) Spectra with 0.25 μA electron
2
2
impact up to 49 V for 0 (dotted) 120 s (dashed) and 600 s (solid).

1730
M.E. van Reijzen et al. / Surface Science 605 (2011) 1726–1731
the amount of CO desorbing in α is reduced while more CO desorbs
prior to H O. Second, in between α and β , we observe a third feature
at ~165 K that increases with H O exposure and occurs simulta-
neously with the lower temperature desorption feature of H O. It is
labeled γ in Fig. 5b. The first observation indicates that H O displaces
presence of both H
conclude that 13CO oxidation results from products produced by
electron impact on H O. We expect this to be surface-bound OH or O.
Finally, the very steep onset of CO production in Fig. 6 is
noteworthy. CO evolution suddenly starts at 164 K and continues in
parallel with H O desorption. This result is quite remarkable
2
O and electron impact. Therefore, we can safely
2
1
2
2
2
2
2
2
some CO from step edge desorption sites. It is consistent with our
interpretation of the high desorption feature in Fig. 4 as resulting from
2
considering that CO was found to be oxidized by OH at a surface
temperature of 77 K on Au(111) [12]. We consider two possible
origins for this large difference in temperature. First, we consider
possible differences in binding energies and activation barriers for CO
oxidation. Falsig et al. have used DFT calculations to investigate CO
oxidation on Au(111) and 12-atom Au clusters [41]. They find no
barrier for the CO+O reaction at the Au(111) terrace. When using Au
clusters, both the adsorption energy and transition state energy for
CO+O drop by ~1 eV, thus yielding no significant activation barrier
for CO oxidation at low coordinated Au atoms. Unless the stronger
CO–Au bond on Au(310) entirely prohibits CO diffusion below ~160 K
and both reactants are fully segregated at the surface, the reaction
temperature difference cannot be explained by the variations in
adsorption energies and activation barriers. Both of these require-
ments seem improbable as DFT-based calculations indicate small
differences between the binding energies at various adsorption sites
on Au(310) [31] and the γ feature in Fig. 5 strongly suggests
the step edge. Competition between H
e.g., lead to compression of CO adsorbed along the step edge, resulting
in the appearance of the β feature at such low CO coverages. The
unchanged desorption temperature of β tells us that there is little
interaction between CO appearing in this feature and H O. The second
observation, however, indicates that H O and CO do interact at other
locations and were likely bound very close to each other. Such H O–
was observed
2
O and CO for these sites may,
1
1
2
2
2
CO interaction does not result in CO oxidation, as no CO
in any of these TPD experiments.
2
A more subtle change in the H
the inset of Fig. 5a. We observe that pre-dosing CO affects the
adsorption states of H O. The leading edge of the H O TPD trace
initially tracks desorption of the more weakly bound H O, but shifts
O as CO desorption
2
O desorption spectra is shown in
2
2
2
toward the trace from the more strongly bound H
2
in γ stops.
Fig. 6 shows an example of how electron irradiation changes TPD
features for CO (Fig. 6a), H O (Fig. 6b) and CO (Fig. 6c). Dashed traces
2
2
2
interaction between H O and CO on the surface. Second, we consider
show desorption when the sample was kept at −49 V. For the solid
differences in the local O and OH environments. Results from
experiments by Ojifinni et al. were performed mostly at rather low
OH coverages created from reaction of isotopically labeled O with
traces, the sample was not biased during the TPD ramp and a 0.25 μA
2
current impinged onto the H O/CO/Au(310) surface through the hole
in the QMS's differentially pumped housing. We estimate that the e-
2
post-adsorbed H O. At 77 K, a molecular beam of CO reacted with
flux was ~2 electrons per Au atom, based on the aperture, the current
(part of) the OH and left some O and CO on the surface. In our
experiments, O or OH are most likely formed by electron irradiation
2
to the Au crystal and the time passed prior to H O desorption
(
~100 s). Note that the slight variation in the total amount of CO
within the hydrogen bonded network of H
dosed after CO adsorption and competes for step sites. As the CO
oxidation appears during H O desorption, we expect that O or OH only
reacts with CO when the hydrating H O network decomposes due to
O desorption.
2 2
O molecules. The H O was
desorption results from previously mentioned limited accuracy in
dosing small amounts of CO.
2
The most prominent difference observed in Fig. 6 is the 13CO
2
2
production when electron irradiation is used. The H
2
O desorption
H
2
trace is affected in the same way as shown in the inset of Fig. 4. This
change may be expected if electron-induced dissociation results in
4
. Summary
dissociation of (part of) the adsorbed H
an OH–H O network. Furthermore, we have verified that a decrease in
O exposure with a fixed consecutive total electron flux leads to a
2
O molecules and formation of
2
Using isotopic labeling, we have investigated desorption of CO and
2
O from Au(310). Applying both Redhead and complete analyses to
H
2
H
1
3
13
2 2
decrease in CO formation. Finally, CO is only observed in the
CO desorption spectra, we find strongly varying CO desorption
energies. Both are substantially lower than those predicted by DFT
calculations. TPD features for H
states on Au(310). We attribute the adsorption state with higher
binding energy to step edge bound H O. Changes in H O desorption
spectra when using electron irradiation suggest dissociation of H O.
This interpretation is corroborated by CO oxidation when we irradiate
coadsorbed CO and H O with electrons. Without electron irradiation,
no CO is formed. CO oxidation by fragments of H O on Au(310)
2
O desorption suggest two adsorption
2
2
2
2
2
2
occurs at much higher temperatures than on Au(111). The difference
is attributed to a hydration shell surrounding O or OH groups, making
CO oxidation impossible until the excess water has desorbed.
Acknowledgements
The authors kindly acknowledge the Leiden Institute of Chemistry
for funding and Prof. Dr. Ben Nieuwenhuys for generous donation of
equipment.
References
[
[
[
[
1] G.C. Bond, Gold Bulletin 5 (1972) 11.
2] M. Haruta, T. Kobayashi, N. Yamada, Chemical Letters 2 (1987) 405.
3] M. Haruta, Catalysis Today 36 (1997) 153.
4] G.C. Bond, D.T. Thompson, Catalysis Reviews Science and Engineering 41 (1999) 319.
Fig. 6. TPD spectra of (a) 13CO (b) H
without (dashed) and with (solid) electron irradiation.
O and (c) 13CO for coadsorbed 13CO and H
O
2 2
2
[5] R. Meyer, C. Lemire, Sh.K. Shaikhutdinov, H.-J. Freund, Gold Bulletin 37 (2004) 72.
[6] C.W. Corti, R.J. Holliday, D.T. Thompson, Topics in Catalysis 44 (2007) 331.

M.E. van Reijzen et al. / Surface Science 605 (2011) 1726–1731
1731
[
[
[
7] B.K. Min, C.M. Friend, Chemical Reviews 107 (2007) 2709.
8] J.L. Gong, C.B. Mullins, Accounts of Chemical Research 42 (2009) 1063.
9] S.A.C. Carabineiro, B.E. Nieuwenhuys, Gold Bulletin 42 (2009) 288.
[26] J.M. Gottfried, K.J. Schmidt, S.L.M. Schroeder, K. Christmann, Surface Science
536 (2003) 206.
[27] W.-L. Yim, T. Nowitzki, M. Necke, H. Schnars, P. Nickut, J. Biener, M.M. Biener, V.
Zielasek, K. Al-Shamery, T. Klüner, M. Bäumer, Journal of Chemical Physics C
111 (2007) 445.
[
[
10] S.A.C. Carabineiro, B.E. Nieuwenhuys, Gold Bulletin 43 (2010) 252.
11] M. Daté, M. Okumura, S. Tsubota, M. Haruta, Angewandte Chemie, International
Edition 43 (2004) 2129.
[28] P.A. Redhead, Vacuum 12 (1962) 203.
[
[
[
12] R.A. Ojifinni, N.S. Froemming, J. Gong, M. Pan, T.S. Kim, J.M. White, G. Henkelman,
C.B. Mullins, Journal of the American Chemical Society 130 (2008) 6801.
13] S.D. Senanayake, D. Stacchiola, P. Liu, C.B. Mullins, J. Hrbek, J.A. Rodriguez, Journal
of Physical Chemistry C 113 (2009) 19536.
14] R.G. Quiller, T.A. Baker, X. Deng, M.E. Colling, B.K. Min, C.M. Friend, The Journal of
Chemical Physics 129 (2008) 064702.
[29] D.A. King, Surface Science 47 (1975) 384.
[30] A.M. de Jong, J.W. Niemantsverdriet, Surface Science 233 (1990) 355.
[31] A. Hussain, D.C. Ferré, J. Gracia, B.E. Nieuwenhuys, J.W. Niemantsverdriet, Surface
Science 603 (2009) 2734.
[32] N. Lopez, T.V.W. Janssens, B.S. Clausen, Y. Xu, M. Mavrikakis, T. Bligaard, J.K.
Nørskov, Journal of Catalysis 223 (2004) 232.
[
[
[
[
15] P.A. Thiel, T.E. Madey, Surface Science Reports 7 (1987) 211.
16] M.A. Henderson, Surface Science Reports 46 (2002) 5.
17] A. Hodgson, S. Haq, Surface Science Reports 64 (2009) 381.
18] M.J.T.C. van der Niet, A. en Dunnen, L.B.F. Juurlink, M.T.M. Koper, Angewandte
Chemie, International Edition 49 (2010) 6572.
[33] J.L.C. Fajín, M.N.D.S. Cordeiro, J.R.B. Gomes, Journal of Physical Chemistry C
112 (2008) 17291.
[34] D.A. Outka, R.J. Madix, Journal of the American Chemical Society 109 (1987) 1708.
[35] B.D. Kay, K.R. Lykke, J.R. Creighton, S.J. Ward, The Journal of Chemical Physics
91 (1989) 5120.
[
[
19] Y.T. Yates Jr., first, Experimental Innovations in Surface Science, AIP Press, 1997,
p. 572.
20] D. Fargues, P. Dolle, J.J. Ehrhardt, M. Sotto, J.C. Boulliard, Surface Science 274
[36] N. Ikemiya, A.A. Gewirth, Journal of the American Chemical Society 119 (1997)
9919.
[37] M.J.T.C. van der Niet, A. den Dunnen, L.B.F. Juurlink, M.T.M. Koper, The Journal of
Chemical Physics 132 (2010) 174705.
(1992) L495.
[
[
21] M.A. van Hove, G.A. Somorjai, Surface Science 92 (1980) 489.
22] Y.T. Yates Jr., Experimental Innovations in Surface Science, first, AIP Press, 1997,
p. 604.
[38] J.L. Fajín, M.N.D.S. Cordeiro, J.R.B. Gomes, Journal of Molecular Structure:
THEOCHEM 946 (2010) 51.
[39] J.L. Daschbach, B.M. Peden, R.S. Smith, B.D. Kay, The Journal of Chemical Physics
120 (2004) 1516.
[40] J. Shan, A.W. Kleyn, L.B.F. Juurlink, ChemPhysChem 10 (2009) 270.
[41] H. Falsig, B. Hvolbæk, I.S. Kristensen, T. Jiang, T. Bligaard, C.H. Christensen, J.K.
Nørskov, Angewandte Chemie, International Edition 47 (2008) 4835.
[
[
[
23] C.J. Weststrate, E. Lundgren, J.N. Andersen, E.D.L. Rienks, A.C. Gluhoi, J.W. Bakker,
I.M.N. Groot, B.E. Nieuwenhuys, Surface Science 603 (2009) 2152.
24] C. Ruggiero, P. Hollins, Journal of the Chemical Society—Faraday Transactions
9
2 (1996) 4829.
25] J. Kim, E. Samano, B.E. Koel, The Journal of Physical Chemistry. B 110 (2006) 17512.