Selective hydrogen production from formic acid decomposition on Pd-Au bimetallic surfaces

Article abstract of DOI:10.1021/ja505192v

Pd-Au catalysts have shown exceptional performance for selective hydrogen production via HCOOH decomposition, a promising alternative to solve issues associated with hydrogen storage and distribution. In this study, we utilized temperature-programmed desorption (TPD) and reactive molecular beam scattering (RMBS) in an attempt to unravel the factors governing the catalytic properties of Pd-Au bimetallic surfaces for HCOOH decomposition. Our results show that Pd atoms at the Pd-Au surface are responsible for activating HCOOH molecules; however, the selectivity of the reaction is dictated by the identity of the surface metal atoms adjacent to the Pd atoms. Pd atoms that reside at Pd-Au interface sites tend to favor dehydrogenation of HCOOH, whereas Pd atoms in Pd(111)-like sites, which lack neighboring Au atoms, favor dehydration of HCOOH. These observations suggest that the reactivity and selectivity of HCOOH decomposition on Pd-Au catalysts can be tailored by controlling the arrangement of surface Pd and Au atoms. The findings in this study may prove informative for rational design of Pd-Au catalysts for associated reactions including selective HCOOH decomposition for hydrogen production and electro-oxidation of HCOOH in the direct formic acid fuel cell.

Full text of DOI:10.1021/ja505192v

Article
pubs.acs.org/JACS
Selective Hydrogen Production from Formic Acid Decomposition on
Pd−Au Bimetallic Surfaces
Wen-Yueh Yu,† Gregory M. Mullen,† David W. Flaherty,‡ and C. Buddie Mullins*,†
†McKetta Department of Chemical Engineering and Department of Chemistry, Center for Nano and Molecular Science and
Technology, Texas Materials Institute, and Center for Electrochemistry, University of Texas at Austin, Austin, Texas 78712, United
States
‡Department of Chemical & Biomolecular Engineering, University of Illinois at UrbanaChampaign, Urbana, Illinois 61801, United
States
S
* Supporting Information
ABSTRACT: Pd−Au catalysts have shown exceptional performance for
selective hydrogen production via HCOOH decomposition, a promising
alternative to solve issues associated with hydrogen storage and distribution. In
this study, we utilized temperature-programmed desorption (TPD) and
reactive molecular beam scattering (RMBS) in an attempt to unravel the
factors governing the catalytic properties of Pd−Au bimetallic surfaces for
HCOOH decomposition. Our results show that Pd atoms at the Pd−Au
surface are responsible for activating HCOOH molecules; however, the
selectivity of the reaction is dictated by the identity of the surface metal atoms
adjacent to the Pd atoms. Pd atoms that reside at Pd−Au interface sites tend to
favor dehydrogenation of HCOOH, whereas Pd atoms in Pd(111)-like sites, which lack neighboring Au atoms, favor dehydration
of HCOOH. These observations suggest that the reactivity and selectivity of HCOOH decomposition on Pd−Au catalysts can be
tailored by controlling the arrangement of surface Pd and Au atoms. The findings in this study may prove informative for rational
design of Pd−Au catalysts for associated reactions including selective HCOOH decomposition for hydrogen production and
electro-oxidation of HCOOH in the direct formic acid fuel cell.
catalysts on HCOOH decomposition remain topics of
debate.11−13,15,19,22,23 For instance, Pd−Au alloy catalysts
were reported to possess improved HCOOH dehydrogenation
activity and stability relative to monometallic Pd cata-
lysts.12,15,22 Such improvement has been attributed to a higher
resistance to CO poisoning due to alloying of Pd with Au.12,15
However, a recent study indicated that alloying of Pd with Au
leads to lower HCOOH decomposition activity, which was
suggested to result from a ligand effect as indicated by a slight
shift of the Pd 3d5/2 binding energy in X-ray photoelectron
spectroscopy (XPS).19 Apart from alloy catalysts, supported
Pd−Au catalysts with core−shell structures were also tested for
selective HCOOH decomposition;11,13,23 a higher hydrogen
yield was observed13,23 and attributed to the possible charge
transfer between Au and Pd (i.e., a ligand effect).23 However, it
has been shown that the charge transfer between Au and Pd is
insignificant owing to similar work functions between the Au
and the Pd,11 and a lower rate of HCOOH decomposition was
observed for the colloidal Au−Pd core−shell nanoparticles
relative to colloidal Pd nanoparticles.11 The inconsistencies
between these studies may be due to the vast diversity of
preparation methods, catalyst supports, and reaction conditions
employed in these studies. Indeed, classical heterogeneous
1. INTRODUCTION
Hydrogen is a promising energy carrier for electricity
generation in a fuel cell; however, hydrogen storage and
distribution remains a challenging issue. Recently, formic acid
(HCOOH) has been proposed as a potential liquid storage
medium capable of releasing H2 under mild conditions via
catalytic decomposition.1−5 For practical applications, suitable
catalysts are essential to facilitate HCOOH decomposition via
dehydrogenation (HCOOH → H2 + CO2) as opposed to
dehydration (HCOOH → H2O + CO). Recently, considerable
advances have been made in the selective dehydrogenation of
HCOOH at ambient and near-ambient temperatures using
homogeneous catalysts.6−9 Nevertheless, the separation issues
associated with homogeneous catalysts and requisite use of
organic solvents and additives hamper their practical
applications.10,11 Use of heterogeneous catalysts may circum-
vent these issues, but improved catalytic performance under
ambient conditions is still required.10,11
Pd is one of the most active catalysts for HCOOH
decomposition.11 Recently, heterogeneous catalysts containing
Pd have been extensively studied for H2 production via
HCOOH decomposition;11−23 among them, Pd−Au bimetallic
catalysts have received substantial attention due to their
promising performance in selective HCOOH dehydrogen-
ation.12−15,22,23 Yet, the key factors controlling the catalytic
performance (e.g., activity and selectivity) of Pd−Au bimetallic
Received: May 23, 2014
Published: July 14, 2014
© 2014 American Chemical Society
11070
dx.doi.org/10.1021/ja505192v | J. Am. Chem. Soc. 2014, 136, 11070−11078

Journal of the American Chemical Society
Article
Pd−Au model surfaces were prepared by depositing Pd atoms from
a homemade thermal evaporator onto the Au(111) surface at 77 K and
then annealing the surface to 500 K for 10 min under UHV conditions.
The deposition rate of Pd was calibrated with a quartz crystal
microbalance (QCM) controller (Maxtek Inc.) by assuming that the
thickness of 1 monolayer (ML) of Pd is equal to the diameter of a Pd
atom, which is 0.274 nm. Growth of the Pd overlayer on the Au(111)
surface at 77 K is believed to obey a layer-by-layer mechanism based
on a previous study36 in which a Pd film was grown on the Au(111)
surface at the higher temperature of 150 K. Upon annealing, some of
the surface Pd atoms diffuse into the subsurface of the Au(111)
sample, forming a Pd−Au alloy surface.34−38 AES spectra for the as-
deposited and annealed Pd/Au(111) surfaces are shown in the
Supporting Information (Figure S1).
Temperature-programmed desorption (TPD) is a useful tool for
providing information concerning the interactions between the surface
(adsorbent) and the adsorbed species (adsorbate).39−41 In the TPD
measurement, a temperature ramp is applied to the surface and the
rate of desorbing molecules is measured by monitoring the increase in
evolving gas-phase species as a function of surface temperature using a
mass spectrometer. Analyses of a series of TPD spectra can provide
information such as the activation energy for desorption, the strength
of the adsorbate−adsorbent interaction, and the relative surface
coverage of adsorbate.39−41 The interactions (i.e., adsorption,
absorption, diffusion, and desorption) of hydrogen with Pd/Au(111)
surfaces have been studied previously using H2-TPD.34 In this study,
H2-TPD was conducted to characterize the annealed Pd/Au(111)
surfaces for qualitative and quasi-quantitative analyses of surface sites.
Hydrogen was dosed by impinging a molecular beam of H2 onto the
Pd−Au surface at a surface temperature of 77 K to yield a saturation
coverage. Afterward, the H-presaturated surface was heated to 500 K at
a rate of 1 K/s, while the signal for m/z+ = 2 (H2) was monitored by a
quadrupole mass spectrometer (QMS).
Application of molecular beam methods to well-defined model
catalysts enables the detailed investigation of the chemical kinetics of
surface reactions.42,43 By correlating the chemical kinetics with the
surface structure, mechanistic insights into the kinetic phenomena on
catalyst surfaces can be obtained.42,43 In this study, we used reactive
molecular beam scattering (RMBS) to acquire kinetic information (i.e.,
HCOOH decomposition rate and turnover frequency for hydrogen
production) on our Pd−Au surface. HCOOH-RMBS experiments
were conducted by impinging a molecular beam of HCOOH onto the
sample surface that was held at a temperature of 500 K. The flux of the
HCOOH beam was estimated to be ∼5 × 1014 molecules cm−2 s−1
using HCOOH-TPD spectra from the clean Au(111) surface (Figure
S2, Supporting Information) and assuming 1 ML of HCOOH equals
∼1.39 × 1015 molecules cm−2 (the surface atom density of Au(111)).
Prior to exposure to the sample surface, the HCOOH beam was first
impinged onto an inert stainless steel flag that was held in front of the
sample to establish baseline signals. Products and unreacted HCOOH
molecules were detected by monitoring the following QMS signals for
each species: HCOOH (m/z+ = 46, 44, 29, and 28), CO2 (m/z+ = 44
and 28), H2 (m/z+ = 2), CO (m/z+ = 28), and H2O (m/z+ = 18). The
amount of each species is proportional to the integral of its QMS
intensity. The integrals of QMS intensities for CO2 (m/z+ = 44) and
CO (m/z+ = 28) were corrected by considering the contribution from
the corresponding mass fragments of HCOOH and those from
HCOOH and CO2, respectively.
catalysts are generally complex and operate at high temper-
atures and pressures, increasing the level of difficulty to unravel
the factors governing the catalytic properties (e.g., reactivity
and selectivity). In model catalyst studies, reactions are
investigated on well-defined single-crystal surfaces under
ultrahigh vacuum (UHV) conditions, which enable thorough
surface characterization, precise control of reactant coverages,
and minimal environmental interference, thereby allowing the
study between surface structures and catalytic properties of
catalytic materials at the molecular level.24−29
In this study, the surface chemistry of HCOOH on Pd−Au
bimetallic model surfaces was investigated using temperature-
programmed desorption (TPD) and reactive molecular beam
scattering (RMBS) methods in an attempt to improve the
fundamental understanding of the selective decomposition of
HCOOH on Pd−Au catalysts. A variety of Pd−Au bimetallic
surfaces, prepared by varying the Pd coverage deposited on the
Au(111) surface followed by annealing, were characterized by
H2-TPD and tested via HCOOH-RMBS and HCOOH-TPD
under UHV conditions. Two types of surface sites, i.e., Pd−Au
interface sites and Pd(111)-like sites (sites that lack
neighboring Au atoms), were characterized qualitatively and
quasi-quantitatively for the Pd−Au surface by H2-TPD.
Reactivity tests based on HCOOH-RMBS demonstrate the
critical role of Pd−Au interface sites for selective hydrogen
production via catalytic HCOOH dehydrogenation. Although
increasing the fraction of Pd(111)-like sites on the surface
increased the rate of HCOOH decomposition, the selectivity
toward hydrogen production decreased because the Pd(111)-
like sites primarily catalyze dehydration of HCOOH. HCOOH-
TPD results reveal that reaction-limited desorption of CO2
occurred at a relatively low temperature, indicating that the
reaction intermediate(s) from HCOOH decomposition could
be readily decomposed on the Pd−Au surface. These
observations suggest that the exceptional performance of Pd−
Au catalysts for selective HCOOH decomposition could be
rationalized by the “ensemble” effect. The experimental results
shown in this study provide direct evidence that the selectivity
for HCOOH reaction pathways can be tailored by the
intermixing of Pd and Au atoms on catalytically active surfaces,
which is consistent with predictions from a recent theoretical
study.30
2. EXPERIMENTAL METHODS
All experiments in this study were conducted in a molecular beam
surface scattering apparatus with a base pressure of less than 1 × 10−10
Torr.31−35 Briefly, the apparatus contains an Auger electron
spectrometer (Physical Electronics 10−500), a quadrupole mass
spectrometer (Extrel C-50), a Fourier transform infrared spectrometer
(Bruker Tensor 27) combined with a mercury−cadmium−telluride
(MCT) detector (Infrared Associates), as well as nozzles and apertures
for generating two separate molecular beams.
HCOOH-TPD was carried out by adsorbing HCOOH onto the
sample surface at 77 K via a molecular beam followed by annealing to
500 K at a rate of 1 K/s. The QMS signals during HCOOH-TPD were
detected in a manner similar to that described for HCOOH-RMBS
experiments. The coverage of the HCOOH overlayer was determined
by TPD of HCOOH from the clean (Pd-free) Au(111) surface
(Figure S2, Supporting Information).
The Au(111) single-crystal sample is a circular disk (Princeton
Scientific, 12 mm in diameter × 2 mm thick) and held in place by a
Mo wire fitted around a groove cut into the side of the sample. This
wire is also used to resistively heat the sample and provide thermal
contact between the sample and a liquid nitrogen bath for cooling. The
temperature of the sample was measured with a K-type (Alumel−
Chromel) thermocouple placed into a small hole in the edge of the
disk-shaped sample. The Au(111) surface was periodically cleaned by
Ar ion bombardment (2 keV), carried out at room temperature,
followed by an anneal to 800 K. The cleanliness of the surface was
verified by Auger electron spectroscopy (AES) with a beam energy of
3 keV and emission current of 1.5 mA.
3. RESULTS AND DISCUSSION
3.1. H2-TPD from Pd−Au Surfaces. In order to establish a
structure−activity relationship for HCOOH decomposition on
11071
dx.doi.org/10.1021/ja505192v | J. Am. Chem. Soc. 2014, 136, 11070−11078

Journal of the American Chemical Society
Article
the Pd−Au surface, a characterization technique that can
provide both qualitative and quantitative information on surface
properties is essential. Probe molecules such as CO and H2 are
commonly used to characterize the structure and composition
of catalytic surfaces.
surfaces with initial Pd coverages ranging from 0 to 4 ML. The
annealed 1 ML Pd/Au(111) surface exhibited a broad H2-TPD
peak centered at ∼208 K. Since this hydrogen desorption
temperature lies between those for Au(111) (∼108 K)53 and
Pd(111) surfaces (∼310−320 K),50,54 we speculate that this
desorption site consists of Au and Pd atoms and hydrogen
desorption likely occurs at the interface between Au and Pd
atoms.34 This H2-TPD feature is relatively broad because the
desorption temperature of hydrogen depends on the relative
number of Pd and Au atoms associated with the Pd−Au
interface sites during the desorption process. In general, a
higher hydrogen desorption temperature would be expected
when H atoms bind to more Pd atoms in the Pd−Au interface
sites since the bond strength of Pd−H is stronger than that of
Au−H. For the annealed 2 ML Pd/Au(111) surface, the peak
of hydrogen desorption from the Pd−Au interface sites (at
∼210 K) intensified and an additional shoulder at a higher
temperature (∼300 K) was observed. The desorption temper-
ature of this high-temperature shoulder is fairly close to the
desorption temperature of hydrogen from the Pd(111)
surface,50,54 which suggests formation of Pd(111)-like sites on
the annealed 2 ML Pd/Au(111) surface. Both hydrogen
desorption features (i.e., from the Pd−Au interface and
Pd(111)-like sites) grew in intensity on the Pd−Au surfaces
prepared with higher initial Pd coverages of 3 and 4 ML.
The integral of the peak area under each H2-TPD trace is
proportional to the amount of hydrogen that desorbed from
each Pd−Au surface.39−41 It is noted that no measurable H2
desorption was observed from the clean (Pd-free) Au(111)
surface under the same set of conditions. In other words, the
surface sites for hydrogen uptake are associated with surface Pd
atoms. Sykes and co-workers55,56 employed scanning-tunneling
microscopy (STM) to observe H adatoms on Pd/Au(111)
surfaces. STM imaging revealed that hydrogen molecules
dissociatively adsorb on surface Pd atoms, and no spillover
onto the Au terrace sites was observed.55,56 Therefore, the
integral peak area under each H2-TPD trace should also reflect
the relative number of surface Pd atoms for dissociative
adsorption of hydrogen. By integrating the peak for H2-TPD
spectra, the relative number of surface Pd atoms on each Pd−
Au surface was calculated (relative to that of the annealed 4 ML
Pd/Au(111) surface). As listed in Table 1, the relative number
of surface Pd atoms is 0.32, 0.63, 0.84, and 1 for the Pd−Au
surface with the initial Pd coverage of 1, 2, 3, and 4 ML,
respectively.
Adsorption of CO on the Pd−Au model surface has been
extensively studied using CO-RAIRS (reflection−absorption
infrared spectroscopy)35,38,44−47 and CO-TPD.35,44,47−49 Using
CO-RAIRS, qualitative information such as the type of
adsorption sites occupied by CO (e.g., atop sites, 2-fold or 3-
fold bridge sites) on the Pd−Au surface can be inferred by the
intramolecular CO stretch frequency35,38,44−47 due to varying
degrees of π-antibonding back-donation from the surface
electrons; however, it has been shown that quantitative analysis
using CO-RAIRS becomes difficult when contiguous Pd sites
are present on the Pd−Au surface, probably due to the
vibrational coupling effect that attenuates the IR intensity at
high surface CO coverages.45 Quantitative analysis based on
CO-TPD is complicated for the annealed Pd/Au(111) surfaces
as the CO desorption peak is fairly broad35 (likely as a result of
the superposition of multiple CO desorption peaks), making
the surface site assignment and peak deconvolution difficult.
The interaction of hydrogen with the Pd−Au model surface
has been investigated using H2-TPD.34,50−52 In a previous study
using H2-TPD we showed that the presence of contiguous Pd
atoms (characterized by CO-RARIS) on the annealed Pd/
Au(111) surfaces is crucial for dissociative adsorption of
hydrogen molecules at 77 K.34 Upon heating, hydrogen
adatoms recombinatively desorb from the Pd−Au surface.
According to the desorption temperature of hydrogen, two
distinct surface sites, i.e., Pd−Au interface and Pd(111)-like
sites (lacking adjacent Au atoms), were qualitatively assigned.34
Furthermore, a quasi-quantitative analysis is feasible by
performing peak deconvolution for the H2-TPD spectra.
Accordingly, here we utilized H2-TPD to characterize the
Pd−Au surfaces that were generated by annealing the as-
deposited Pd/Au(111) surfaces in UHV.
Figure 1 depicts the H2-TPD spectra for desorption of
saturation coverages of hydrogen from the annealed Pd−Au
Dispersion, defined as the fraction of atoms of a material
exposed to the surface, is an important indicator in
heterogeneous catalysis since catalytic reactions occur on
surfaces. For supported Pd catalysts, dispersion is generally
expressed as the ratio of the amount of hydrogen uptake
(determined from hydrogen chemisorption) to the total
amount of Pd. In this study, the relative dispersion was
computed by dividing the relative number of surface Pd atoms
(determined from H2-TPD) on each Pd−Au surface by its
initial Pd coverage (relative to that of the annealed 4 ML Pd/
Au(111) surface). The relative dispersion was found to
gradually decrease as the initial Pd coverage was increased
(Table 1), suggesting that a higher portion of Pd atoms was not
surface accessible when the Pd−Au surface was prepared with a
higher initial Pd coverage.
Figure 1. H2-TPD spectra from Pd−Au surfaces with initial Pd
coverages ranging from 0 to 4 ML. Pd−Au surfaces were prepared by
depositing Pd atoms onto the Au(111) surface at a surface
temperature (TS) of 77 K followed by annealing to 500 K in UHV.
A molecular beam of H2 was impinged onto each surface at TS = 77 K
to yield a saturation coverage of H adatoms. Heating rate was 1 K/s.
In order to quantify the relative number of Pd−Au interface
and Pd(111)-like sites on each Pd−Au surface, the H2-TPD
spectra in Figure 1 were fitted by peak deconvolution, and the
11072
dx.doi.org/10.1021/ja505192v | J. Am. Chem. Soc. 2014, 136, 11070−11078

Journal of the American Chemical Society
Article
Table 1. Relative Number of Surface Pd Atoms, Relative Dispersion, Fraction of Pd−Au Interface Sites, HCOOH Conversion,
HCOOH Decomposition Rate, Turnover Frequency for H2 Production, and Relative H2/CO QMS Ratio on the Annealed Pd/
Au(111) Surfaces
initial
θPd
relative no. of
relative
HCOOH
TOF for H2
surface Pd atoms
dispersion
fraction of Pd−Au
conversion
HCOOH decomposition
production
Relative H2/CO
ac
,
ab
,
ab
,
b
c
c
bc
,
(ML)
(−)
0.32
(−)
interface sites (−)
(%)
rate (HCOOH cm−2 s−1
)
(H2 PdS s−1
)
QMS ratio (−)
−1
1
2
3
4
1.29
1
0.7
2.8
6.7
8.8
3.5 × 1012
1.4 × 1013
3.3 × 1013
4.4 × 1013
0.006
0.012
0.011
2.08
1.19
1.05
1
0.63
0.84
1
1.26
1.12
1
0.66
0.62
0.60
0.009
a
b
c
Relative to that of the annealed 4 ML Pd/Au(111) surface. Determined from H2-TPD spectra (Figure 1). Determined from HCOOH-RMBS
spectra (Figure 3).
corresponding peak integrals are displayed in Figure 2. As the
initial Pd coverage was increased, the amount of hydrogen
Figure 2. Integrals of H2 QMS signal intensity for H2-TPD spectra
shown in Figure 1 after peak deconvolution.
desorption from the Pd−Au interface and Pd(111)-like sites
both increased (Figure 2), indicating that the number of both
types of surface sites increased since the amount of hydrogen
desorption is proportional to the number of surface sites. The
fraction of Pd−Au interface sites on each Pd−Au surface was
calculated by the integral of hydrogen desorption from Pd−Au
interface sites divided by that of the total hydrogen desorption
from both Pd−Au interface and Pd(111)-like sites. The fraction
of Pd−Au interface sites is unity for the annealed 1 ML Pd/
Au(111) surface and reduces to ∼0.6 for surfaces with higher
initial Pd coverages, i.e., 2−4 ML (see Table 1).
3.2. HCOOH-RMBS on Pd−Au Surfaces. The reactivity of
HCOOH on Pd−Au surfaces was evaluated by HCOOH-
RMBS. A room-temperature beam of HCOOH vapor was
generated with a translational energy of ∼0.1 eV. The trapping
probability of HCOOH under these conditions will be very
close to unity regardless of the surface temperature. At elevated
surface temperatures trapped molecules will undergo a kinetic
competition between surface reaction and desorption.57−60
Figure 3 shows the HCOOH-RMBS results for annealed Pd/
Au(111) surfaces with various initial Pd coverages. The
HCOOH beam was first impinged onto the inert stainless
steel flag that was held in front of the sample for 5 s (from 30 to
35 s) to establish baseline signals and then impinged onto the
Pd−Au surface that was held at a surface temperature at 500 K
for 5 s (from 65 to 70 s) to complete a King and Wells
measurement.61,62
Figure 3. HCOOH-RMBS results for Pd−Au surfaces with various
initial Pd coverages. Pd−Au surfaces were prepared by depositing Pd
atoms onto the Au(111) surface at a surface temperature of 77 K
followed by annealing to 500 K in UHV. The surface was held at 500
K during HCOOH-RMBS measurements. HCOOH beam was
impinged onto the inert flag for 5 s (from 30 to 35 s) and then
onto the sample surface for 5 s (from 65 to 70 s). HCOOH flux was
∼5 × 1014 molecules cm−2 s−1. Panels a−d have the same ordinate
scale.
When the HCOOH molecular beam struck the inert flag,
QMS signals for m/z+ = 46, 44, and 28 were observed, in which
the m/z+ = 46 signal is the parent mass of HCOOH and the m/
z+ = 44 and 28 signals are mass fragments of HCOOH from
QMS ionizer dissociation (similar m/z+ ratios for masses 46, 44,
and 28 were observed when shooting the HCOOH beam
directly into QMS). The m/z+ = 18 signal is due to a water
impurity in the HCOOH32 as we observed this mass signal
when shooting the HCOOH beam directly into QMS as well.
As expected, no QMS signal for H2 (m/z+ = 2) was observed
when scattering the beam from the inert flag. After removing
the inert flag, the HCOOH beam was directed onto the
annealed 1 ML Pd/Au(111) surface, where a QMS signal for
11073
dx.doi.org/10.1021/ja505192v | J. Am. Chem. Soc. 2014, 136, 11070−11078

Journal of the American Chemical Society
Article
H2 production (m/z+ = 2) was observed (Figure 3a). The
emergence of the H2 QMS signal clearly shows that the
dehydrogenation reaction occurred on the 1 ML Pd/Au(111)
surface. Apart from H2 production, CO2 (the byproduct of
HCOOH dehydrogenation) and CO (the byproduct product
from HCOOH dehydration) also formed as indicated by
comparison of m/z+ = 44 and 28 signals for impingement of the
HCOOH beam on the inert flag and on the Pd−Au surface.
These results show that the presence of Pd adatoms on the
Pd−Au surface can facilitate HCOOH decomposition since no
reactivity was observed on the Au(111) surface under the same
testing conditions (Figure S3, Supporting Information). It is
noted that formation of H2O due to HCOOH dehydration was
difficult to observe in this system, which likely results from
efficient pumping of H2O by the cryogenic probe cooled by
liquid nitrogen.
Figure 3b−d shows the HCOOH-RMBS results for Pd−Au
surfaces with higher initial coverages of Pd (i.e., 2−4 ML). For
convenience of comparison, the integrals of QMS intensities for
each species produced (i.e., H2, CO2, and CO) during
HCOOH-RMBS (Figure 3) are summarized in Figure 4.
Figure 5. (a) Rate of HCOOH decomposition and (b) turnover
frequency for H2 production on Pd−Au surfaces during HCOOH-
RMBS experiments shown in Figure 3.
account, this sharp increase in the HCOOH decomposition
rate observed here is believed to be related to formation of
Pd(111)-like sites on the Pd−Au surface. In other words,
Pd(111)-like sites are more catalytically active for HCOOH
decomposition than Pd−Au interface sites.
The specific activity of Pd−Au surfaces for H2 production is
expressed in terms of turnover frequency (TOFH2) or the
number of H2 molecules produced per surface Pd site (PdS) per
unit time. In this study, the TOFH2 on Pd−Au surfaces was
computed from the integral of QMS signal intensities for H2
production from HCOOH-RMBS experiments and H2
desorption from TPD measurements (details associated with
these calculations can be found in the Supporting Information).
Figure 5b displays the calculated TOFH2 of each Pd−Au surface
as a function of the relative number of surface Pd atoms on the
Pd−Au surface. With the increase of the relative number of
Figure 4. Integrals of QMS signal intensities for H2, CO, and CO2
produced on annealed Pd/Au(111) surfaces with various initial Pd
coverages during HCOOH-RMBS (Figure 3).
The amount of H2, CO2, and CO generated on the Pd−Au
surface increased as the initial Pd coverage was increased,
suggestive of an activity enhancement for overall HCOOH
decomposition. By comparing the m/z+ = 46 signals obtained
from the HCOOH beam impinging on the inert flag and on the
sample surface, conversion of HCOOH on each surface can be
estimated. As listed in Table 1, the estimated single-collision
conversion of HCOOH on the Pd−Au surface increased from
0.7% to 8.8% as the initial Pd coverage was increased from 1 to
4 ML.
Figure 5a depicts the rate of HCOOH decomposition as a
function of the relative number of surface Pd atoms on Pd−Au
surfaces. The rate of HCOOH decomposition is the product of
the conversion of HCOOH and the flux of the HCOOH
molecular beam (∼5 × 1014 molecules cm−2 s−1) in HCOOH-
RMBS experiments. As illustrated in Figure 5a, the rate of
HCOOH decomposition increased as the relative number of
surface Pd atoms increased. A sharp increase in the rate of
HCOOH decomposition was observed when the relative
number of surface Pd atoms was larger than 0.32 (i.e., the
annealed 1 ML Pd/Au(111) surface). Taking the relative
number of Pd−Au interface and Pd(111)-like sites into
−1
surface Pd atoms, the TOFH2 first increased to 0.012 H2 PdS
s−1 and then slightly decreased to 0.009 H2 PdS s−1. Because
−1
the rate of HCOOH decomposition increased monotonically
(Figure 5a), the decrease in TOFH2 observed here suggests that
the activity of HCOOH dehydration was enhanced, causing the
selectivity for hydrogen production to be reduced as the relative
number of surface Pd atoms increased.
The selectivity for H2 production via HCOOH decom-
position depends directly on the fraction of Pd atoms that exist
at the Pd−Au interface or Pd(111)-like sites on the Pd−Au
surface. The selectivity for hydrogen production (or dehydro-
genation selectivity) can be expressed by the relative H2/CO or
CO2/CO QMS area ratios (relative to that of the annealed 4
ML Pd/Au(111) surface) in HCOOH-RMBS experiments.
Figure 6 shows the relative H2/CO QMS area ratios and the
11074
dx.doi.org/10.1021/ja505192v | J. Am. Chem. Soc. 2014, 136, 11070−11078

Journal of the American Chemical Society
Article
dissociation of COOH into CO and OH on the Pd6Au3 and
Pd3Au6 surfaces (∼1 eV) than on the Pd(111) and Pd ML
surfaces (∼0.5−0.6 eV).30 In other words, the dehydration
pathway is inhibited relative to the dehydrogenation pathway
on the Pd6Au3 and Pd3Au6 surfaces, which is consistent with
the importance of Pd−Au interface sites for HCOOH
dehydrogenation selectivity (Figure 6).
The catalytic properties of bimetallic catalysts are often
discussed in terms of the ligand (or electronic) effect and
ensemble (or geometric) effect.28,63,64 The ligand effect
describes electronic modifications caused by formation of
heterometallic bonding, which leads to a charge transfer
between the metals.28 As mentioned earlier, the ligand effect
has been suggested to explain the catalytic performance of Pd−
Au catalysts observed for HCOOH decomposition.19,23 Never-
theless, it is noted that the work functions for Au and Pd are
similar,11,38 and their electronegativities are identical.38
Furthermore, XPS spectra indicate that very little (<0.2 eV)
or no shifts occur in the binding energies of Pd(3d) and Au(4f)
or Pd Auger electrons on Pd−Au surfaces,36 which implies a
very limited charge transfer between Pd and Au in Pd−Au
alloys. The ensemble effect depicts that the atoms that exist in a
particular arrangement are required for facilitating a particular
catalytic process.28 The ensemble effect for HCOOH
decomposition on Pd−Au surfaces has been investigated
theoretically by DFT calculations.30 In this study, we have
experimentally shown that the fraction of Pd atoms that exist in
Pd−Au interface or Pd(111)-like sites has a significant
influence on the reactivity and selectivity of HCOOH
decomposition on the Pd−Au surface. Pd atoms in Pd(111)-
like sites, which lack neighboring Au atoms, were found to
exhibit a higher catalytic activity for HCOOH decomposition
(Figure 5a) via the undesired dehydration pathway, resulting in
a lower selectivity for hydrogen production. Pd atoms that exist
at the Pd−Au interface display a lower activity for overall
HCOOH decomposition but enhanced dehydrogenation
selectivity (Figure 6). These observations suggest that the
ensemble effect plays an important role in determining the
catalytic properties of Pd−Au catalysts for selective HCOOH
decomposition.
3.3. HCOOH-TPD from Pd−Au Surfaces. The inter-
actions of HCOOH with transition metal surfaces have been
extensively studied using HCOOH-TPD as reviewed by
Madix,65 Barteau,66 and Columbia and Thiel.67 CO2 desorption
from metal surfaces during HCOOH decomposition is typically
a reaction-limited process, i.e., CO2 desorbs immediately upon
formation.67 Thus, the catalytic activities of metal surfaces for
HCOOH decomposition are frequently characterized by the
desorption temperature of CO2, where lower temperatures for
CO2 desorption are indicative of higher activities for HCOOH
decomposition.65,66 The interactions of HCOOH with low-
index single-crystal surfaces of Pd have been investigated.68−73
It was reported that the temperature for CO2 desorption from
HCOOH decomposition is ∼240−260 K on the Pd(111)
surface,68,69 ∼237 K on the Pd(110) surface,70 and ∼180 K on
the Pd(100) surface.71 These results suggest that the reactivity
of Pd surfaces for HCOOH decomposition is sensitive to the
arrangement of surface Pd atoms and the (100) facet is the
most active among these structures.
Figure 6. (a) Relative H2/CO QMS area ratios measured from
HCOOH-RMBS experiments (Figure 3), and (b) fraction of Pd−Au
interface sites on Pd−Au surfaces determined by H2-TPD measure-
ments (Figure 1).
fraction of Pd−Au sites on Pd−Au surfaces as a function of the
relative number of surface Pd atoms (the trends of relative H2/
CO and CO2/CO QMS area ratios were almost identical,
Figure S4, Supporting Information).
Figure 6 shows that the relative QMS ratio of H2/CO, the
selectivity toward hydrogen production, decreased with
increasing relative number of surface Pd atoms. This decrease
in hydrogen selectivity correlates with a decrease in the fraction
of Pd atoms that exist as Pd−Au interface sites, suggesting that
HCOOH dehydrogenation occurs at Pd−Au interface sites on
the surface. The observations based on H2-TPD and HCOOH-
RMBS correlate the structure of Pd−Au surfaces with their
catalytic properties on HCOOH decomposition: (1) the
presence of Pd−Au interface sites on Pd−Au surfaces is crucial
for selective HCOOH dehydrogenation; (2) formation of
Pd(111)-like sites on Pd−Au surfaces enhances the overall
HCOOH decomposition at the expense of the dehydrogen-
ation selectivity. It is noted that high selectivity for HCOOH
dehydrogenation was reported in some studies employing
supported monometallic Pd catalysts.17,19,21 Since the catalytic
performance of supported monometallic metal catalysts often
results from an accumulative contribution from both metal
nanoparticles and the support used, the observed high
selectivity17,19,21 may be due to participation of the support
in the surface reaction or metal−support interactions.
Our experimental observations are conceptually consistent
with a recent theoretical study.30 Using density function theory
(DFT) methods, Yuan and Liu30 calculated the potential
energy profile for possible reaction pathways for HCOOH
decomposition on four different surfaces: the Pd(111) surface,
the Pd monolayer supported on the Au(111) surface
(abbreviated Pd ML), and two Pd-decorated Au(111) surfaces
(abbreviated Pd6Au3 and Pd3Au6). Energy barriers for
HCOOH activation to form formate (HCOO) (via O−H
bond cleavage) or carboxyl (COOH) intermediates (via C−H
bond cleavage) are lower on the Pd(111) and Pd ML surfaces
in comparison to those on the Pd6Au3 and Pd3Au6 surfaces.30
These results support our observations that the rate of the
HCOOH decomposition is enhanced by formation of Pd(111)-
like sites on the Pd−Au surface. They also found that the
energy barriers for further decomposition of HCOO or COOH
intermediate to CO2 and H are comparable for all surfaces
investigated; however, higher energy barriers are required for
Figure 7 shows the TPD spectra for various species following
adsorption of 2.6 ML of HCOOH on the clean Au(111)
surface and annealed 2 ML Pd/Au(111) surface.
11075
dx.doi.org/10.1021/ja505192v | J. Am. Chem. Soc. 2014, 136, 11070−11078

Journal of the American Chemical Society
Article
from the m/z+ = 46, 29, and 28 spectra. The pronounced signal
for m/z+ = 44 at ∼180 K shows that CO2 molecules evolved
from the Pd−Au surface during heating. CO2 evolution
observed here is a reaction-limited process rather than a
desorption-limited process as CO2 molecules desorb at much
lower temperatures from Au(111)75 and Pd(111) surfaces.76
Thus, CO2 desorption occurred immediately upon its
formation, which was due to decomposition of the reaction
intermediate(s) such as formate (HCOO) and carboxyl
(COOH).30 Unfortunately, we were unable to identify the
reactive intermediate(s) on the Pd−Au surface via RAIRS
(Figure S5, Supporting Information) due to the small surface
concentration of these intermediates which gave a low signal-
to-noise ratio. The temperature of CO2 evolution via HCOOH
decomposition from the Pd−Au surface (∼180 K) is coincident
with that from the Pd(100) surface,71 which suggests that the
Pd−Au surface is comparably active to the Pd(100) surface for
HCOOH decomposition during heating.
4. CONCLUSIONS
We conducted a model catalyst study to investigate the
reactivity of HCOOH on a variety of Pd−Au bimetallic surfaces
using temperature-programmed desorption and reactive mo-
lecular beam scattering. Our results reveal that Pd atoms are the
active component for HCOOH decomposition on the Pd−Au
alloy surface. The selectivity of the reaction is governed by the
type of the nearest neighbor atoms adjacent to the Pd atoms.
Pd atoms which lack adjacent Au atoms favor dehydration of
HCOOH, whereas Pd atoms that possess Au atoms as nearest
neighbors favor dehydrogenation of HCOOH, which is
desirable for efficient production of hydrogen. These
observations suggest that the reactivity and selectivity of
HCOOH decomposition on Pd−Au bimetallic catalysts could
be optimized by controlling the arrangement of Pd and Au
atoms on the surface through the ensemble effect. We believe
these findings will be beneficial for the future design of Pd−Au
bimetallic catalysts for associated reactions including selective
HCOOH decomposition for hydrogen production and electro-
oxidation of HCOOH in the direct formic acid fuel cell.77 We
hope our observations will inspire more advanced surface
characterization to further investigate the relationship between
the catalytic activities and the surface nature of Pd−Au surfaces.
Figure 7. TPD spectra of HCOOH from the (a) Au(111) and (b)
Pd−Au surfaces. The Pd−Au surface was prepared by depositing 2 ML
of Pd onto the Au(111) surface at a surface temperature (TS) of 77 K
followed by annealing to 500 K in ultrahigh vacuum. Each surface was
dosed with ∼2.6 ML of HCOOH at TS = 77 K. Heating rate was 1 K/
s. Panels a and b have the same ordinate scale.
As shown in Figure 7a, the desorption spectrum for the
parent mass of HCOOH (m/z+ = 46) from the Au(111)
surface displayed peaks at 154 and 180 K, which are due to
desorption of multilayer and monolayer HCOOH, respectively
(Figure S2, Supporting Information). We suggest that the
observed m/z+ = 44 and 28 signals originate from the mass
fragments of HCOOH rather than from formation of CO2 and
CO because of the following: (1) the desorption temperatures
of m/z+ = 44 and 28 signals line up with those of the m/z+ = 46
and 29 (the most intense mass fragment of HCOOH), and (2)
the ratios of integral intensities of these QMS signals are similar
to those obtained from the HCOOH beam impingement onto
the inert flag. No measurable H2 (m/z+ = 2) desorption was
observed during TPD from the HCOOH-precovered Au(111)
surface. As mentioned earlier, the observed signal for m/z+ = 18
was due to the water impurity in HCOOH. Results in Figure 7a
confirm that the Au(111) surface is inactive for HCOOH
decomposition during TPD measurements, which is also
consistent with observations from the HCOOH-RMBS experi-
ment on the clean Au(111) surface (Figure S3, Supporting
Information) and with reports that HCOOH interacts weakly
with the clean Au(111) surface using XPS and RAIRS.74
TPD spectra of HCOOH from the Pd−Au surface were
measured under the same set of conditions (Figure 7b).
Significant attenuation of the desorption feature for the
HCOOH monolayer at ∼180 K indicates that HCOOH
decomposed on this surface. The small peak in the m/z+ = 28
signal at ∼450 K is due to desorption of CO adsorbed on the
surface from the background gas35 or from decomposition of
HCOOH. It is noted that the m/z+ = 44 spectrum is dissimilar
ASSOCIATED CONTENT
■
S
* Supporting Information
Additional information regarding AES characterization for the
as-deposited and annealed Pd/Au(111) surfaces, TPD of
HCOOH from the Au(111) surface, RMBS of HCOOH on
the Au(111) surface, derivation of equations for calculating
TOFH2, relative H2/CO and CO2/CO QMS area ratios for
RMBS of HCOOH on Pd−Au surfaces, RAIRS of HCOOH on
the Au(111) and annealed Pd/Au(111) surfaces; AES spectra
for the as-deposited and annealed Pd/Au(111) surface with
various initial Pd coverages; HCOOH-TPD spectra from the
Au(111) surface; HCOOH-RMBS results for the Au(111)
surface; relative H2/CO and CO2/CO QMS area ratios
measured from Pd−Au surfaces in HCOOH-RMBS experi-
ments; RAIRS spectra of HCOOH on the Au(111) and
annealed 2 ML Pd/Au(111) surfaces. This material is available
free of charge via the Internet at http://pubs.acs.org.
11076
dx.doi.org/10.1021/ja505192v | J. Am. Chem. Soc. 2014, 136, 11070−11078

Journal of the American Chemical Society
Article
(29) Pan, M.; Brush, A. J.; Pozun, Z. D.; Ham, H. C.; Yu, W.-Y.;
Henkelman, G.; Hwang, G. S.; Mullins, C. B. Chem. Soc. Rev. 2013, 42,
5002−5013.
(30) Yuan, D. W.; Liu, Z. R. J. Power Sources 2013, 224, 241−249.
(31) Flaherty, D. W.; Hahn, N. T.; Ferrer, D.; Engstrom, T. R.;
Tanaka, P. L.; Mullins, C. B. J. Phys. Chem. C 2009, 113, 12742−
12752.
AUTHOR INFORMATION
Corresponding Author
E-mail: mullins@che.utexas.edu
■
Notes
The authors declare no competing financial interest.
(32) Flaherty, D. W.; Berglund, S. P.; Mullins, C. B. J. Catal. 2010,
269, 33−43.
ACKNOWLEDGMENTS
■
We are thankful for the generous support of the Department of
Energy (DE-FG02-04ER15587) and the Welch Foundation
(Grant F-1436). G.M.M. thanks the National Science
Foundation for a Graduate Research Fellowship.
(33) Flaherty, D. W.; Yu, W.-Y.; Pozun, Z. D.; Henkelman, G.;
Mullins, C. B. J. Catal. 2011, 282, 278−288.
(34) Yu, W.-Y.; Mullen, G. M.; Mullins, C. B. J. Phys. Chem. C 2013,
117, 19535−19543.
(35) Yu, W.-Y.; Mullen, G. M.; Mullins, C. B. J. Phys. Chem. C 2014,
118, 2129−2137.
REFERENCES
■
(36) Koel, B. E.; Sellidj, A.; Paffett, M. T. Phys. Rev. B 1992, 46,
7846−7856.
(1) Enthaler, S. ChemSusChem 2008, 1, 801−804.
(2) Joo, F. ChemSusChem 2008, 1, 805−808.
(3) Enthaler, S.; von Langermann, J.; Schmidt, T. Energy Environ. Sci.
2010, 3, 1207−1217.
(37) Baddeley, C. J.; Ormerod, R. M.; Stephenson, A. W.; Lambert,
R. M. J. Phys. Chem. 1995, 99, 5146−5151.
(38) Chen, M. S.; Kumar, D.; Yi, C. W.; Goodman, D. W. Science
2005, 310, 291−293.
(4) Loges, B.; Boddien, A.; Gartner, F.; Junge, H.; Beller, M. Top.
Catal. 2010, 53, 902−914.
(39) Christmann, K. Introduction to Surface Physical Chemistry, In
Topics in Physical Chemistry; Springer: New York, 1991; pp 152−166.
(40) Masel, R. I. Principles of Adsorption and Reaction on Solid
Surfaces. In Wiley Series in Chemical Engineering; John Wiley & Sons,
Inc.: New York, 1996; pp 507−515.
(5) Johnson, T. C.; Morris, D. J.; Wills, M. Chem. Soc. Rev. 2010, 39,
81−88.
(6) Boddien, A.; Loges, B.; Gartner, F.; Torborg, C.; Fumino, K.;
Junge, H.; Ludwig, R.; Beller, M. J. Am. Chem. Soc. 2010, 132, 8924−
8934.
(41) Attard, G.; Barnes, C., Surfaces. Oxford Chemistry Primers 59;
Oxford University Press: Oxford, 1998; pp 71−75.
(42) Kleyn, A. W. Chem. Soc. Rev. 2003, 32, 87−95.
(43) Libuda, J.; Freund, H. J. Surf. Sci. Rep. 2005, 57, 157−298.
(44) Yi, C. W.; Luo, K.; Wei, T.; Goodman, D. W. J. Phys. Chem. B
2005, 109, 18535−18540.
(7) Boddien, A.; Gartner, F.; Jackstell, R.; Junge, H.; Spannenberg,
A.; Baumann, W.; Ludwig, R.; Beller, M. Angew. Chem., Int. Ed. 2010,
49, 8993−8996.
(8) Boddien, A.; Mellmann, D.; Gartner, F.; Jackstell, R.; Junge, H.;
Dyson, P. J.; Laurenczy, G.; Ludwig, R.; Beller, M. Science 2011, 333,
1733−1736.
(45) Wei, T.; Wang, J.; Goodman, D. W. J. Phys. Chem. C 2007, 111,
8781−8788.
(9) Hull, J. F.; Himeda, Y.; Wang, W. H.; Hashiguchi, B.; Periana, R.;
Szalda, D. J.; Muckerman, J. T.; Fujita, E. Nat. Chem. 2012, 4, 383−
388.
(46) Gao, F.; Wang, Y. L.; Goodman, D. W. J. Am. Chem. Soc. 2009,
131, 5734−5735.
(10) Boddien, A.; Junge, H. Nat. Nanotechnol. 2011, 6, 265−266.
(11) Tedsree, K.; Li, T.; Jones, S.; Chan, C. W. A.; Yu, K. M. K.;
Bagot, P. A. J.; Marquis, E. A.; Smith, G. D. W.; Tsang, S. C. E. Nat.
Nanotechnol. 2011, 6, 302−307.
(47) Li, Z.; Gao, F.; Furlong, O.; Tysoe, W. T. Surf. Sci. 2010, 604,
136−143.
(48) Sellidj, A.; Koel, B. E. Phys. Rev. B 1994, 49, 8367−8376.
(49) Baddeley, C. J.; Tikhov, M.; Hardacre, C.; Lomas, J. R.;
Lambert, R. M. J. Phys. Chem. 1996, 100, 2189−2194.
(50) Behm, R. J. Z. Phys. Chem. 2009, 223, 9−36.
(51) Boscoboinik, J. A.; Calaza, F. C.; Garvey, M. T.; Tysoe, W. T. J.
Phys. Chem. C 2010, 114, 1875−1880.
(12) Zhou, X. C.; Huang, Y. J.; Xing, W.; Liu, C. P.; Liao, J. H.; Lu, T.
H. Chem. Commun. 2008, 3540−3542.
(13) Huang, Y. J.; Zhou, X. C.; Yin, M.; Liu, C. P.; Xing, W. Chem.
Mater. 2010, 22, 5122−5128.
(14) Zhou, X. C.; Huang, Y. J.; Liu, C. P.; Liao, J. H.; Lu, T. H.; Xing,
W. ChemSusChem 2010, 3, 1379−1382.
(52) Ogura, S.; Okada, M.; Fukutani, K. J. Phys. Chem. C 2013, 117,
9366−9371.
(15) Gu, X. J.; Lu, Z. H.; Jiang, H. L.; Akita, T.; Xu, Q. J. Am. Chem.
Soc. 2011, 133, 11822−11825.
(53) Pan, M.; Flaherty, D. W.; Mullins, C. B. J. Phys. Chem. Lett.
2011, 2, 1363−1367.
(16) Yadav, M.; Singh, A. K.; Tsumori, N.; Xu, Q. J. Mater. Chem.
2012, 22, 19146−19150.
(54) Gdowski, G. E.; Felter, T. E.; Stulen, R. H. Surf. Sci. 1987, 181,
L147−L155.
(17) Bulushev, D. A.; Beloshapkin, S.; Ross, J. R. H. Catal. Today
2010, 154, 7−12.
(55) Tierney, H. L.; Baber, A. E.; Kitchin, J. R.; Sykes, E. C. H. Phys.
Rev. Lett. 2009, 103, 246102.
(18) Bulushev, D. A.; Jia, L. J.; Beloshapkin, S.; Ross, J. R. H. Chem.
Commun. 2012, 48, 4184−4186.
(56) Baber, A. E.; Tierney, H. L.; Lawton, T. J.; Sykes, E. C. H.
ChemCatChem. 2011, 3, 607−614.
(57) Mullins, C. B.; Weinberg, W. H. J. Chem. Phys. 1990, 92, 3986−
3988.
(58) Rettner, C. T.; Mullins, C. B.; Bethune, D. S.; Auerbach, D. J.;
Schweizer, E. K.; Weinberg, W. H. J. Vac. Sci. Technol. A 1990, 8,
2699−2704.
(59) Mullins, C. B.; Rettner, C. T.; Auerbach, D. J.; Weinberg, W. H.
Chem. Phys. Lett. 1989, 163, 111−115.
(60) Sitz, G. O.; Mullins, C. B. J. Phys. Chem. B 2002, 106, 8349−
8353.
(19) Bulushev, D. A.; Beloshapkin, S.; Plyusnin, P. E.; Shubin, Y. V.;
Bukhtiyarov, V. I.; Korenev, S. V.; Ross, J. R. H. J. Catal. 2013, 299,
171−180.
(20) Zhao, Y.; Deng, L.; Tang, S. Y.; Lai, D. M.; Liao, B.; Fu, Y.; Guo,
Q. X. Energy Fuels 2011, 25, 3693−3697.
(21) O’Neill, B. J.; Gurbuz, E. I.; Dumesic, J. A. J. Catal. 2012, 290,
̈
̈
193−201.
(22) Metin, O.; Sun, X. L.; Sun, S. H. Nanoscale 2013, 5, 910−912.
(23) Wang, Z. L.; Yan, J. M.; Wang, H. L.; Ping, Y.; Jiang, Q. J. Mater.
Chem. A 2013, 1, 12721−12725.
(61) Davis, J. E.; Karseboom, S. G.; Nolan, P. D.; Mullins, C. B. J.
(24) Campbell, C. T. Annu. Rev. Phys. Chem. 1990, 41, 775−837.
(25) Somorjai, G. A. Chem. Rev. 1996, 96, 1223−1235.
(26) Weaver, J. F.; Carlsson, A. F.; Madix, R. J. Surf. Sci. Rep. 2003,
50, 107−199.
(27) Gong, J. L. Chem. Rev. 2012, 112, 2987−3054.
(28) Gao, F.; Goodman, D. W. Chem. Soc. Rev. 2012, 41, 8009−8020.
Chem. Phys. 1996, 105, 8362−8375.
(62) Wheeler, M. C.; Seets, D. C.; Mullins, C. B. J. Chem. Phys. 1996,
105, 1572−1583.
(63) Liu, P.; Norskov, J. K. Phys. Chem. Chem. Phys. 2001, 3, 3814−
3818.
11077
dx.doi.org/10.1021/ja505192v | J. Am. Chem. Soc. 2014, 136, 11070−11078

Journal of the American Chemical Society
Article
(64) Ham, H. C.; Hwang, G. S.; Han, J.; Nam, S. W.; Lim, T. H. J.
Phys. Chem. C 2010, 114, 14922−14928.
(65) Madix, R. J. Adv. Catal. 1980, 29, 1−53.
(66) Barteau, M. A. Catal. Lett. 1991, 8, 175−184.
(67) Columbia, M. R.; Thiel, P. A. J. Electroanal. Chem. 1994, 369, 1−
14.
(68) Davis, J. L.; Barteau, M. A. Surf. Sci. 1991, 256, 50−66.
(69) Jeroro, E.; Vohs, J. M. Catal. Lett. 2009, 130, 271−277.
(70) Aas, N.; Li, Y. X.; Bowker, M. J. Phys. Condens. Matter 1991, 3,
S281−S286.
(71) Sander, D.; Erley, W. J. Vac. Sci. Technol. A 1990, 8, 3357−3360.
(72) Jorgensen, S. W.; Madix, R. J. J. Am. Chem. Soc. 1988, 110, 397−
400.
(73) Solymosi, F.; Kovacs, I. Surf. Sci. 1991, 259, 95−108.
(74) Senanayake, S. D.; Stacchiola, D.; Liu, P.; Mullins, C. B.; Hrbek,
J.; Rodriguez, J. A. J. Phys. Chem. C 2009, 113, 19536−19544.
(75) Farkas, A. P.; Solymosi, F. J. Phys. Chem. C 2009, 113, 19930−
19936.
(76) Wambach, J.; Odorfer, G.; Freund, H. J.; Kuhlenbeck, H.;
Neumann, M. Surf. Sci. 1989, 209, 159−172.
(77) Larsen, R.; Ha, S.; Zakzeski, J.; Masel, R. I. J. Power Sources
2006, 157, 78−84.
11078
dx.doi.org/10.1021/ja505192v | J. Am. Chem. Soc. 2014, 136, 11070−11078