Evidence for geometric effects in neopentane conversion on PdAu catalysts

Article abstract of DOI:10.1039/c4cy00846d

Silica-supported Pd and shell/core PdAu nanoparticles of a similar size were evaluated for neopentane conversion. Monometallic Pd exhibited poor neopentane isomerization selectivity in favor of high selectivity to primary and secondary hydrogenolysis products. Similarly sized PdAu catalysts of increasing Pd weight loading were synthesized to evaluate the effect of increasing Pd monolayers on neopentane conversion. All PdAu catalysts had neopentane conversion selectivity within the range of monometallic Pd catalysts from previous work (～5-30%). However, there was an inverse relationship between Pd weight loading and neopentane isomerization selectivity. The increase in isomerization selectivity did not correlate to a decrease in heats of adsorption as seen with monometallic Pd catalysts, but was correlated with the catalyst surface structure which suggests a geometric effect as the cause for changes in catalytic performance rather than an electronic effect.

Full text of DOI:10.1039/c4cy00846d

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Evidence for geometric effects in neopentane
conversion on PdAu catalysts
Cite this: Catal. Sci. Technol., 2014,
4, 4366
David J. Childers,a Neil M. Schweitzer,b Seyed Mehdi Kamali Shahri,c Robert M. Rioux,c
d
a
*
*
Jeffrey T. Miller and Randall J. Meyer
Silica-supported Pd and shell/core PdAu nanoparticles of a similar size were evaluated for neopentane
conversion. Monometallic Pd exhibited poor neopentane isomerization selectivity in favor of high
selectivity to primary and secondary hydrogenolysis products. Similarly sized PdAu catalysts of increasing
Pd weight loading were synthesized to evaluate the effect of increasing Pd monolayers on neopentane
conversion. All PdAu catalysts had neopentane conversion selectivity within the range of monometallic Pd
catalysts from previous work (~5–30%). However, there was an inverse relationship between Pd weight
loading and neopentane isomerization selectivity. The increase in isomerization selectivity did not correlate
to a decrease in heats of adsorption as seen with monometallic Pd catalysts, but was correlated with the
catalyst surface structure which suggests a geometric effect as the cause for changes in catalytic
performance rather than an electronic effect.
Received 1st July 2014,
Accepted 5th August 2014
DOI: 10.1039/c4cy00846d
www.rsc.org/catalysis
DFT simulations, Hammer and Norskov separated coordina-
tion and strain effects from ligand effects in their studies of
1. Introduction
Bimetallic catalysts often have selectivities and activities that
differ significantly from what would be expected from physi-
cal mixtures of the two metals.1–7 Alloying allows the surface
structure and composition to become tunable parameters of
the catalyst that are able to promote activity or selectivity to
desired products. Both ligand (electronic) effects and ensem-
ble (geometric) effects may be responsible for changes in
reactivity observed with alloys, but it is difficult to determine
experimentally which factor is dominant since it is a
challenge to isolate these effects. Electronic effects are
manifested as changes in the strength of surface–adsorbate
interactions due to modification of the electronic structure
when the two metals are alloyed. Ensemble effects arise when
reactions require a particular number or arrangement of
atoms to form an active site.6 Understanding the role of
electronic and geometric effects in alloy catalysts is compli-
cated because the two effects are often intertwined. Hammer
and Norskov have interpreted changes in reactivity of d-band
transition metals through the use of a single descriptor: the
d-band center of the metal. Adsorption of simple molecules
can be directly correlated to the d-band center and enable an
understanding of the observed catalytic activity. Through
bimetallic systems and demonstrated for a given adsorption
geometry linear correlations exist between the adsorbate
binding energy and the position of the d-band center.8,9
Greeley and Norskov demonstrated this simple d-band corre-
lation held for a large variety of alloys in their study of oxygen
adsorption.10 Experimentally separating the influence of
electronic and geometric effects in alloys is a difficult task
as the presence of a second metal not only contributes a
ligand effect from the change in hybridization between
neighboring metal atoms, but strain effects are often intro-
duced due to changes in metal–metal bond distances.11 In
addition, the presence of the second metal may change the
type of available sites at the surface (an ensemble effect).
We previously demonstrated the selectivity to isomeriza-
tion as opposed to hydrogenolysis in neopentane conversion
follows the same type of electronic structure correlation
described by Norskov et al., as the selectivity is linearly
related to the CO adsorption energy (presumably directly
governed by the energy of the d-band center).12 However, it is
not clear if the relationship between isomerization selectivity
and the electronic structure extends beyond monometallic
systems. PdAu provides an interesting test case for electronic
structure correlation since Au has previously been used as a
modifier for improved selectivity in hydrocarbon reforming.4
PdAu alloy catalysts have been studied for a wide variety
of reactions, including oxidation, hydrogen peroxide pro-
duction, hydrochlorination of acetylene and butadiene selec-
tive hydrogenation.13–18 Several different PdAu structural
a Department of Chemical Engineering, University of Illinois at Chicago, USA.
E-mail: rjm@uic.edu
b Center for Catalysis and Surface Science, Northwestern University, USA
c Department of Chemical Engineering, The Pennsylvania State University, USA
d Division of Chemical Sciences and Engineering, Argonne National Laboratory,
USA. E-mail: millerjt@anl.gov
4366 | Catal. Sci. Technol., 2014, 4, 4366–4377
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
configurations are possible depending on the method of
catalyst synthesis and metal loadings. Homogeneous alloys
with varying Pd and Au loadings have been synthesized by a
number of different methods,15,17,18 however, segregation of
either Au or Pd into core–shell geometry has also been well-
documented.16,19–22 Core–shell nanoparticles with Au-rich
shells have been synthesized by sequential deposition and
annealing.20 Pd(shell)/Au(core) nanoparticles has been
synthesized through a number of different methods: electro-
deposition,21 colloidal synthesis16,22 and seeded growth
(addition Pd shells to pre-made Au particles).19 Both Pd and
Au have been used as promoters to enhance the catalytic
properties of the other by decorating the surface.23,24
Dolle et al. studied the growth of Pd films on Au(110)
surfaces; at low Pd coverage (0–0.5 ML), gold preferentially
covers Pd atoms, but at higher coverage (0.5 to 3 ML), the
growth of Pd introduces a small tensile strain due to
the match with the Au lattice until ~9 ML when the lattice
parameter of the Pd lattice relaxes significantly.25 Li et al.
performed experiments on PdAu surfaces and found at low
palladium coverage, the surface restructures to form low-
coordination gold ensembles that block atop sites on
palladium.26 At high palladium coverage, CO adsorbs on
bridge sites, similar to pure palladium. Goodman's group has
investigated the structure of PdAu single crystal surfaces.13
The addition of Pd atoms to a Au single crystal surface led to
isolation of Pd atoms on the Au surface.20,27,28 Electronic
effects stem from charge transfer from gold to palladium
which shifts the d-band center of Pd away from the Fermi
level and weakens adsorbate bonding to Pd.29 This weakened
adsorbate bonding results in improved neopentane isomeri-
zation selectivity as has been observed in previous work.12
A number of theoretical studies have been conducted on
PdAu systems to analyze the observed changes in catalytic
behavior.13,26,30–33 Most findings have concluded a combina-
tion of geometric and electronic effects contribute to Au
altering the behavior of Pd. For example, Neurock et al.
found that both surface coverage and metal–adsorbate bond
strength are influenced by geometric effects, since the
surface coverage and bond strength are correlated (bond
strength decreases as coverage increases).34 The bond strength
of ethylene to Pd is weakened by the addition of Au which
makes the sites more active for ethylene hydrogenation;
however, gold also reduces the binding energy of H by
130 kJ mol−1 which deters hydrogenation. For ethylene hydro-
genation, these two effects balance one another and the
result is a negligible effect on activity but there is an effect
on initial ethane dehydrogenation selectivity by increasing
the barrier for ethylidyne formation by 25 kJ mol−1. Although
charge transfer was observed (an electronic effect), it was not
significant and was therefore discounted as a cause for the
observed change in catalytic behavior. Mazzone et al. used
density functional theory to study the adsorption of CO on
PdAu (111) and (100) surfaces.29 Isolated Pd atoms on a Au
surface induces a local relaxation effect that causes shortened
Pd–Au bond distances compared to typical Au–Au bond
distances resulting in a stronger Pd–Au bond and weakened CO
binding energy by as much as 30 kJ mol−1. Garcia-Mota et al.
also used DFT to study the adsorption of CO on PdAu
surfaces.35 Their calculations showed that isolation of Pd on
Au surfaces is thermodynamically favorable and isolation of
Pd was stabilized by the presence of CO at high pressures
(>10−2 Torr).
Gold has been previously alloyed with Pt group metals to
study neopentane conversion. Schwank et al. performed
neopentane hydrogenolysis over Pt–Au/SiO2 and Pt–Sn/Al2O3
catalysts.36 Pt–Au samples did not have a significantly dif-
ferent selectivity compared to the monometallic samples.
Bonarowska et al. used PdAu catalysts to study neopentane
conversion.15 They synthesized two series of samples pre-
pared by direct redox and incipient wetness impregnation.
The reported dispersion of ~0.05 implies that the particles
were very large (20 nm). The monometallic Pd sample
showed a single-point isomerization selectivity of almost 40%
at low conversion (<1%), about a 15% increase from the
selectivity found in previous work.12 PdAu catalysts with lower
Au content (0.05–0.20 wt.%) showed either no change or a
slight decrease in isomerization selectivity whereas at a load-
ing of 0.40 wt.% Au, the isomerization selectivity increased to
almost 50% indicating a positive effect on selectivity.
Here we report on the synthesis, characterization and test-
ing of four PdAu catalysts supported on silica with different
Pd loadings. The physical and chemical effects of alloying
palladium with gold was investigated by electron microscopy,
X-ray energy dispersive spectroscopy, diffuse-reflectance
Fourier transform spectroscopy (DRIFTS), extended X-ray
absorption fine structure (EXAFS), chemisorption and adsorp-
tion calorimetry of CO. Neopentane hydrogenolysis was the
probe reaction used to evaluate changes in catalytic perfor-
mance of Pd upon alloying with Au.
2.0 Experimental methods
2.1 Catalyst synthesis
All samples were synthesized using sequential impregnation
of Au then Pd on silica (Davisil A60 silica gel from Sigma-
Aldrich, 200 m2 g−1 and 1.15 mL g−1 pore volume) using both
incipient wetness impregnation (IWI) and strong electrostatic
adsorption (SEA).37 During IWI, a support is contacted with
just enough metal-precursor solution to fill the pore volume.
All of the metal precursor solution contacts the surface due
to the incipient amount of liquid used; however, this syn-
thetic method typically leads to significant particle growth
during drying and pretreatment. In contrast, SEA uses an
excess amount of solution with a controlled pH to favor the
uptake of the metal precursor. The SEA method requires an
appropriately charged metal precursor based on the point of
zero charge (PZC) of an oxide support and the pH of the solu-
tion. The hydroxyl groups on the surface of the oxide support
will protonate or deprotonate depending on the pH of the
solution relative to the PZC of the support. Precursors can
then be selected that have the opposite charge as the
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 4366–4377 | 4367

View Article Online
Paper
Catalysis Science & Technology
support at a certain pH value and so the electrostatic interac-
tion between the support and the precursor leads to adsorp-
tion of the metal precursor to the surface.37 Weight loadings
were chosen to create catalysts with approximately one mono-
layer of Pd increasing to 4 Pd monolayers on the Au particle
surface. An estimate of the number of Pd monolayers was
determined assuming that one Pd atom covers a single Au
atom on the surface and Pd disperses uniformly on the sur-
face of the Au nanoparticles. Post treatment steps were evalu-
ated using EXAFS and XANES spectra to confirm reduction of
the metal.
lamp for 20 min. Images were taken using the High Angle
Angular Dark Field (HAADF) mode. XEDS measurements
were collected using an Oxford Instrument X-Max 80 mm2
SDD detector. A 1.3 Å probe size with a 143 pA beam current
was used to measure line scan data with a 1 s dwell time per
step. Data was plotted using the Oxford Instrument Aztec
program.
2.3 Extended X-ray Absorption Fine-Structure
Spectroscopy (EXAFS)
X-ray absorption spectroscopy (XAS) measurements at the
Pd K (24 350 eV), and Au L3 (11 919 eV) edges were made on
the bending magnet beamline of the Materials Research
Collaborative Access Team (MRCAT) at the Advanced Photon
Source (APS) at Argonne National Laboratory. Measurements
were made in transmission mode. A palladium or gold foil
spectrum was acquired through a third ion chamber simulta-
neously with each measurement for energy calibration.
Samples were prepared by grinding the catalysts into a
fine powder and pressing them into a metal cylinder sample
holder capable of holding up to six individual samples. The
sample holder is then placed in a quartz tube with ports on
each end to flow gases or isolate the sample after treatment.
The sample thickness was chosen to give a total absorbance
at the Au L3 or Pd K edge between 1–2 absorption lengths
and edge steps around 0.3–0.5. Samples were reduced at
300 °C in 4% H2/He mixture (50 cm3 min−1) at atmospheric
pressure. After reduction, the samples were purged with He
(100 cm3 min−1) and cooled to room temperature in He flow.
Trace oxidants in He were removed by passing through a
Matheson PUR-Gas Triple Purifier Cartridge containing a Cu trap.
All EXAFS spectra were obtained at room temperature in He.
0.52% Pd–3% Au & 1% Pd–3% Au. Gold was added to silica
using the SEA method. Silica (35 g) was added to an ammonia
solution (NH4OH (5 mL)/H2O (200 mL)). 1.7 g of HAuCl4 was
added to 50 mL of H2O, 2.5 mL ethylenediamine and 5 mL of
NH4OH which resulted in a dark orange-yellow solution with
a pH of 11. These solutions were combined and stirred for
10 min. followed by vacuum filtering of the solid catalyst,
which was subsequently washed 3 times with H2O. The sam-
ple was dried in a vacuum oven overnight at 50 °C. The sam-
ple was treated in flowing 4% H2/He at 150 °C. For the 0.52%
Pd sample, 0.55 g of 10% Pd(NH3)4(NO3)2 solution was then
added to 5 g of Au/SiO2 using IWI and dried at 125 °C over-
night and treated in flowing O2 at 225 °C for 3 h and reduced
at 200 °C. For the 1% Pd sample, 1.42 g of 10% Pd(NH3)4(NO3)2
was added to 1 mL NH4OH and 2.5 mL H2O solution and
then added dropwise to 5 g of Au/SiO2. The sample was dried
at 125 °C overnight followed by treatment in flowing O2 at
225 °C for 3 h and treated in flowing H2 at 200 °C for 30 min.
2% Pd–3.5% Au & 4% Pd–3.5% Au. Gold was added to
silica using the same SEA method described above. The
sample was dried in an oven overnight at 90 °C. The sample
was treated at 150 °C in 4% H2/He. For the 2% Pd sample,
2.78 g of 10% Pd(NH3)4(NO3)2 aqueous solution was added to
an ammonia solution (NH4OH (1 mL)/H2O (4 mL)) to achieve
a pH of 11 and then added dropwise to 5 g of Au/SiO2. The
sample was dried at 100 °C overnight and treated in flowing
O2 at 225 °C for 3 h and treated in flowing H2 at 225 °C
for 30 min. For the 4% Pd sample, the addition of Pd was
done in two steps. Initially, 2.84 g of 10% Pd(NH3)4(NO3)2
was dropwise to a solution of NH4OH (1 mL) and H2O (4 mL)
and 5 g of Au/SiO2. The sample was dried at 100 °C overnight
and then 2.75 g of precursor was added dropwise to an
ammonia solution (NH4OH (1 mL)/H2O (4 mL)). The sample
was again dried overnight at 100 °C and then treated in
flowing O2 at 225 °C for 3 h and treated in flowing H2 at
225 °C for 30 min.
2.4 Neopentane hydrogenolysis and isomerization
Neopentane hydrogenolysis and isomerization was studied
using 0.05–0.15 grams of catalyst diluted with 0.9 grams of
silica and loaded into a 0.5" O.D. quartz plug flow reactor.
Glass wool was used for the bottom 2 cm of the bed. A
0.5 cm long layer of silica was placed on top of the glass
wool, followed by the catalyst–silica mixture resulting in an
overall catalyst bed height of 3 cm. The reactor was purged
with He for 5 min before each run, and the catalyst was
reduced in 4% H2/He as the temperature was increased to
the reaction temperature, 271 2 °C. A K-type thermocouple
was inserted from the bottom of the reactor into the lower
portion of the catalyst bed. Once the reaction temperature
stabilized, the pre-mixed reactant feed gas consisting of
0.35% neopentane and 3.5% H2 balanced in He was passed
through the reactor. The flow rate of the feed gas was varied
from 25 to 100 cm3 min−1 to obtain differential conversions
(1–4%). Although the selectivity does not change drastically
with conversion, the selectivities reported are those obtained
by extrapolation to zero conversion.12 Each flow rate was run
for at least one hour to ensure steady-state conversion had
been reached. An Agilent 6890 N gas chromatograph (using a
2.2 STEM and X-ray Energy Dispersive Spectroscopy (XEDS)
The STEM spectra were obtained at UIC's Research Resources
Center Facility using the JEOL-ARM 200CF aberration
corrected microscope (70 pm spatial resolution and 300 meV
energy resolution). Samples were dispersed in isopropyl alco-
hol and sonicated for 20 min. A drop of the solution was
added to a holey-carbon copper grid and dried under a heat
4368 | Catal. Sci. Technol., 2014, 4, 4366–4377
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
J&W Scientific GS-Alumina column) with an FID detector was
used to analyze the products and was equipped with a back
pressure regulator on the outlet to maintain the system at a
constant pressure of 9 psig. Each experimental run was com-
pleted within six hours for consistency and multiple runs for
each catalyst were performed. No appreciable deactivation
was observed in any of the catalysts over this period of time.
The maximum error of any selectivity measurements was 6%,
with most of the data being reproducible within 2%. Turn-
over rates were calculated based upon the number of active
sites determined by the dispersion measured by CO chemi-
sorption to account for the fact that surface Au sites do not
adsorb CO at the experimental conditions employed and are
not active for neopentane conversion.38–40
factor determined from the DRIFTS linear-to-bridging ratios
for each catalyst. Linear bound CO accounts for one CO mol-
ecule per surface atom while bridge bound CO implies there
is one CO molecule per two surface atoms. The equation
used to determine the dispersion of Pd in the PdAu/SiO2 cata-
lysts is given below.
mol CO adsorbed
Dispersion 
 linear CO fraction 1



mol metal
+ bridge CO fraction  2



2.7 CO heats of adsorption
Determination of the initial heat of adsorption of carbon
monoxide on the monometallic Pd/SiO2 and four PdAu/SiO2
catalysts (60–70 mg) was conducted utilizing a Setaram
Sensys EVO differential scanning calorimetry interfaced with
a plug-flow reactor. The plug-flow reactor was connected to a
mass spectrometer. After reduction at 300 °C in 5.11% H2/Ar
(both gases 99.999%) for 2 h, the catalyst was cooled down to
35 °C in the same gas. The catalyst was exposed to a mixture
of 1% CO (99.999% research grade) in He (99.9999%,
research grade) pulsed into the 5% H2/Ar stream from a ten-
way switching value with a 1000 μL sample loop. The number
of moles of CO per pulse calculated from the ideal gas law
was ~4 × 10−4 mmol. Twenty pulses of carbon monoxide were
employed; the initial heat of adsorption was determined by
only considering injections in which the entire pulse of
carbon monoxide was consumed. The number of pulses
completely consumed depended on the content of Pd in the
catalyst.
2.5 Diffuse Reflectance Infrared Fourier Transform
Spectroscopy (DRIFTS)
Infrared spectra were obtained using a Thermo Scientific
Nicolet 6700 FTIR spectrometer equipped with a Harrick
Scientific Praying Mantis diffuse reflectance in situ cell at the
Northwestern Clean Catalysis (CleanCat) Core Facility. Sam-
ples were ground to a fine powder using a mortar and pestle,
and packed into the sample chamber to create a uniform sur-
face. The chamber was purged with Ar, and then the gas was
switched to 10% H2/N2 and the temperature was raised to
250 °C and held for 15 min. After reduction of the catalyst,
the gas was switched back to Ar and the temperature was
reduced to 25 °C. A background scan was then recorded,
which was averaged over 100 scans (2 minute observation
time) with 4 cm−1 resolution. The sample was then exposed
to 1.02% CO/N2 and another scan was taken once equilib-
rium was reached, at which point the flow was changed back
to Ar and a final scan was taken once the intensity of the
adsorbed CO peak was invariant with time. The linear-to-
bridge bound ratio reported here do not take into account
the differences in extinction coefficients between the adsorp-
tion sites and therefore are not quantitative, but do reflect
qualitative differences between catalysts.41
3.0 Results
3.1 Particle size analysis
STEM imaging was used to determine particle size. A repre-
sentative STEM image for the 2% Pd + 3.5% Au catalyst and
corresponding size distribution are shown in Fig. 1. Table 1
summarizes the particle sizes for all catalysts determined by
both STEM and chemisorption methods. The four catalysts
have similar particle sizes. Data for similar sized monometal-
lic Pd catalysts is taken from Childers et al. for comparison
with PdAu.12 Particle size has an effect on neopentane selectiv-
ity and activity; therefore, in order to isolate the effect of Au
alloying, it is critical that catalysts with varying composition
have similar particle size.12
Fig. 2 shows a line scan and the corresponding STEM
image for the 0.52% Pd + 3% Au and 1% Pd + 3% Au cata-
lysts, respectively. Line scans demonstrate an increased
concentration of Pd on the outer shells of the nanoparticles
while the Au is concentrated within the core. Analysis
performed on several particles per catalyst sample confirmed
that a majority of the nanoparticles have a Au-rich core and
Pd-rich shell structure.
2.6 CO chemisorption
The CO chemisorption measurements were conducted at the
Northwestern University Clean Catalysis (CleanCat) Core
Facility using an Altamira Instruments AMI-200. Catalysts
(0.05–0.2 g) were loaded into a U-shaped quartz reactor tube,
which was weighed before and after sample addition to
ensure an accurate weight measurement. The loaded tube
was then mounted in the furnace and the catalysts were
reduced in 4% H2/Ar at 300 °C for 2 h (10 °C min−1 ramp
rate) and then flushed for 30 min in He. Using a six-way
valve, 5% CO/He was then pulsed (595 μL loop volume) into
the system 15 times at 30 °C to ensure the surface was satu-
rated. Each pulse peak was integrated to find the volume of
CO remaining following adsorption. Surface saturation was
typically reached within 10 pulses. The dispersion determined
by CO chemisorption included the average stoichiometric
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 4366–4377 | 4369

View Article Online
Paper
Catalysis Science & Technology
Fig. 1 2% Pd–3.5% Au sample image and particle size distribution determined by STEM.
Table 1 Particle size determined by STEM and CO chemisorption
Sample
Particle size by STEM (nm)
Particle size by CO chemisorption (nm)
2% Pd_2.5 nm
2% Pd_3 nm
0.52% Pd + 3% Au
1% Pd + 3% Au
2% Pd + 3.5% Au
4% Pd + 3.5% Au
2.5 1.1
3.0 0.8
2.9 1.4
2.8 1.2
3.3 1.1
4.1 1.0
4.3 0.4
3.8 0.4
1.7 0.2
2.2 0.2
2.3 0.2
3.8 0.4
3.2 EXAFS
bond distances. Fig. 3b shows the Au L3-edge spectra of all
catalysts and the Au foil reference. The Au foil spectra show
two peaks at 2.45 and 2.95 Å, and these peaks shift to 2.37
and 3.00 Å in the PdAu catalysts, respectively.43,44 The spectra
show almost no variation from metallic Au except the peak at
2.45 Å is shifted lower by 0.075 Å and the peak at 2.95 Å is
shifted higher by 0.025 Å. The fitting parameters in Table 2
show that there are very few Au–Pd interactions and no
change to the bulk properties of Au in the bimetallic cata-
lysts. The Au–Au bond distance contraction from 2.88 to 2.86 Å
is not unexpected for smaller Au particles.22 This lack of
change indicates that nearly all gold present in the catalysts
are surrounded by other gold atoms.
Extended X-ray absorption fine structure (EXAFS) spectra
were collected for each catalyst following reduction at 300 °C
and fit to determine bond distances and coordination
numbers (CN). The EXAFS spectra for the Pd K-edge and the
Au L3-edge spectra are given in Fig. 3 and the fitting parame-
ters are summarized in Table 2. The reference Pd and Au foil
spectra (black dotted lines) are included for comparison. The
Pd foil EXAFS in Fig. 3a show two prominent peaks at 1.97
and 2.47 Å with an intensity ratio of 2 : 1, typical of metallic
Pd.42 The PdAu catalysts have small shifts in these peaks to
2.00 and 2.55 Å respectively; however, these shifts are not
unexpected due to slight changes in particle size amongst
samples. The 4% Pd–3.5% Au catalyst also has the 2 : 1 peak
intensity ratio typical of metallic Pd, while the other PdAu
catalysts show a change in that ratio to about 1.6 : 1. This
change indicates that there are Au atoms in contact with Pd
in the catalyst. The Pd–Pd bond distance in all of the cata-
lysts is 2.80 Å, an expansion by 0.05 Å from the metallic bond
distance of 2.75 Å.12 This expansion in the Pd–Pd bond dis-
tance could be the result of strain due to the epitaxial growth
of Pd over Au.25 Differences in the coordination number
between Pd–Pd and Pd–Au show the majority of the palla-
dium atoms in this catalyst are bonded to other palladium,
as would be expected for a core–shell configuration. The frac-
tion of Pd atoms which interact with Au decreases as the pal-
ladium loading increases. A Pd–Au bond distance of 2.82 Å is
approximately the average of the bulk Pd–Pd and bulk Au–Au
3.3 Diffuse Reflectance Infrared Fourier Transform
Spectroscopy (DRIFTS)
Characterization of CO adsorption on the PdAu/SiO2 catalyst
was conducted with DRIFTS to determine changes in CO
bonding to the catalyst surface with the addition of increas-
ing amounts of Pd. Peaks in the range of 1800–2000 cm−1 are
assigned to bridge bound CO; within this range, peaks
between 1800 and 1900 cm−1 are attributed to CO bound on
terrace sites while peaks between 1900 and 2000 cm−1 are
believed to be due to the presence of CO bound to corner
and edge sites.45,46 The linear bound region can also have
two peaks typically occurring around 2080 and 2050 cm−1
which Lear et al. have assigned to linear bound CO on corner
sites and edge sites respectively.45
4370 | Catal. Sci. Technol., 2014, 4, 4366–4377
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
Fig. 2 Linescan data for 0.52% Pd–3% Au catalyst (a) and 1% Pd–3% Au catalyst (b) Pd (red) and Au (blue).
Fig. 3 EXAFS spectra for (a) Pd K-edge and (b) Au L3-edge.
Fig. 4 shows the DRIFTS spectra for the PdAu catalysts as
well as monometallic Pd catalysts with a similar particle size.
The linear bound CO peak in each spectrum was normalized
to the largest linear peak to facilitate direct comparison
amongst catalysts. Palladium is known to typically favor bridge
bound CO over linear bound CO and alloying with gold caused
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 4366–4377 | 4371

View Article Online
Paper
Catalysis Science & Technology
Table 2 EXAFS fitting parameters for Pd K-edge and Au L3-edge
Catalyst
Pd edge energy (keV)
24.350
Scatter
Coordination number
Bond distance (Å)
0.52Pd–3Au
Pd–Pd
Pd–Au
Pd–Pd
Pd–Au
Pd–Pd
Pd–Au
Pd–Pd
Pd–Au
6.2
3.4
6.4
2.9
7.0
2.6
10.0
1.7
2.80
2.82
2.80
2.82
2.80
2.82
2.80
2.82
1Pd–3Au
24.350
24.350
24.350
2Pd–3.5Au
4Pd–3.5Au
Catalyst
Au edge energy (keV)
11.919
Scatter
Coordination number
Bond distance (Å)
0.52Pd–3Au
Au–Au
Au–Pd
Au–Au
Au–Pd
Au–Au
Au–Pd
Au–Au
Au–Pd
10.2
1.1
9.8
1.4
11.2
1.0
2.86
2.80
2.86
2.83
2.87
2.83
2.87
2.82
1Pd–3Au
11.919
11.919
11.919
2Pd–3.5Au
4Pd–3.5Au
11.3
1.1
bound CO on corners and edges, but Lear et al. have further
defined these two distinct peaks as bridge bound CO on
(100) facets (1970 cm−1) and CO bridge bound on (111) facets
(1930 cm−1).45 These PdAu catalysts have similar peak posi-
tions to the monometallic Pd (3 nm) catalyst, except that the
bridge bound CO peak on terrace sites is located at 1830 cm−1.
All PdAu catalysts lose the bridge bound CO peak on terrace
sites centered at 1830 cm−1 suggesting that gold breaks up
palladium ensembles. The resolution of two distinct peaks in
the bridge bound region and the loss of the peak around
1830 cm−1 are evidence for bimetallic formation.20,33,47 The
similarity of these spectra to the monometallic Pd spectra
suggests that any surface Au present does not adsorb CO.
Marx et al. found little contribution to the DRIFTS spectra
from CO adsorbed on Au sites on Pd/Au alloys.50
The linear-to-bridge ratios of adsorbed CO are calculated
by comparing the area of the different peaks (Table 3). All
PdAu catalysts have a higher linear-to-bridge ratio than
monometallic Pd catalysts with similar STEM-determined
particle size.12 Alloying with gold causes a decrease in bridge
bound CO, a further indication that palladium ensembles
break up. There is no apparent correlation between palladium
weight loading and linear-to-bridge ratio as the 1% Pd–3% Au
catalyst has the highest ratio. However, the addition of Pd
above 1 wt.% loading causes the linear-to-bridge ratio to
decrease to a value similar to monometallic Pd catalyst, con-
sistent with the presence of a contiguous overlayer of Pd on
the Pd(shell)/Au(core) nanoparticle. The change in the DRIFTS
spectra are another indication that palladium is altered by
alloying with gold and has been seen in literature.47,51
Fig. 4 DRIFTS spectra of the adsorption of carbon monoxide on silica
supported PdAu catalysts: 0.52% Pd–3% Au (green), 1% Pd–3% Au
(blue), 2% Pd–3.5% Au (red) and 4% Pd–3.5% Au (black) and
monometallic Pd catalysts of similar particle size (dotted lines): 2.5 nm
(orange) and 3 nm (purple). Intensity of linear bound CO peak at
2082 cm−1 normalized for each catalyst for comparison.
an increase in the linear-to-bridge ratio, suggesting division
of large Pd ensembles into smaller ones. All catalysts have a
peak centered at 2082 cm−1 which is attributed to linear
bound CO on isolated Pd atoms on the Au–Pd surface.20,47 CO
bound on Au surfaces shows a linear peak around 2110 cm−1
which is not seen in any of the catalysts.17 All catalysts pos-
sess a pair of peaks in the bridge bound region that shift
peak position depending on the Pd/Au composition. The
0.52% Pd–3% Au catalyst has a peak at 1955 cm−1 which is
assigned to bridge bound CO on corner and edge sites and
another peak at 1860 cm−1 which is assigned to bridge bound
CO on terrace sites.14,48,49 These peaks are in the same posi-
tions as the monometallic Pd (2.5 nm) catalyst and all other
catalysts have two bridge bound peaks centered at 1970 cm−1
and 1930 cm−1. Both of these peaks are in the range of bridge
3.4 Calorimetric determination of CO heat of adsorption on
PdAu/SiO2 catalysts
The initial heat of adsorption determined previously for a
pair of Pd/SiO2 catalysts with a mean TEM particle size of
4372 | Catal. Sci. Technol., 2014, 4, 4366–4377
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Table 3 Initial heat of CO adsorption values ( 10 kJ mol−1) in the presence of chemisorbed hydrogen and linear-to-bridge ratios from DRIFTS
Paper
Samplea
CO (initial) heat of adsorption on H-covered surfaceb (kJ mol−1 CO)
Linear to bridge ratio
2Pd/SiO2_2.5 nmc
0.52Pd–3Au/SiO2
1Pd–3Au/SiO2
2Pd–3.5Au/SiO2
4Pd–3.5Au/SiO2
92
104
97
99
90
0.20 : 1
0.59 : 1
0.94 : 1
0.37 : 1
0.25 : 1
a
Catalysts were initially reduced under 5.11% H2/Ar (both gases, 99.999% UHP) for 2 h at the specified reduction temperature, followed by
b
subsequent cooling in H2. Initial heat of CO adsorption determined by microcalorimetry on a reduced H-covered Pd or PdAu surface.
c
Catalyst from Ref. 12.
2.5–3 nm was ~130 kJ mol−1 12
.
The heat of CO adsorption on
Au in the Pd shell layer influences the energetics of H adsorp-
tion, but the reported initial heats of adsorption of CO are
approximately the same across all samples demonstrating Pd
dominates the interaction with CO.53–55 The results in Table 3
also show that increasing Pd wt.% decreased the heat of
adsorption to a value similar to monometallic Pd.
Pd was previously determined for the catalysts reduced in H2,
evacuated at the reduction temperature, followed by cooling
in vacuum to the adsorption temperature (35 °C). In the
previous work, the differential (initial) heat of adsorption
was determined using a volumetric adsorption instrument
(Micromeritics ASAP 2020C) interfaced with a Setaram Sensys
EVO differential scanning calorimeter. For the PdAu catalysts
studied in this work, calorimetric measurements of CO were
conducted in a plug-flow reactor interfaced to a mass spec-
trometer using the same Setaram Sensys EVO differential
scanning calorimeter. Attempts to measure the heat of
adsorption of CO on the PdAu catalysts in an inert environ-
ment were unsuccessful due to the large heat flow associated
with the oxidation of CO by the oxygen in He (UHP,
99.999%). The heat flow observed calorimetrically was accom-
panied with a nearly identical change in the CO2 (44 m/z)
signal in the mass spectrometer. All attempts to eliminate
the CO oxidation reaction were unsuccessful including utilizing
a CO/He mixture containing research grade He (99.9999%). In
order to reduce the oxidation of carbon monoxide, we intro-
duced pulses of CO into a 5% H2/Ar stream. The Ar contained
significantly less oxygen than both grades of He. Reduction of
the catalyst and cooling down to the adsorption temperature
in the H2/Ar ensured that Pd was covered with chemisorbed
hydrogen. Therefore, the CO heat of adsorption values
reported in Table 3 are those for H-covered Pd surfaces.
Chemisorbed hydrogen will lower the heat of adsorption of CO
compared to an adsorbate-free surface of Pd.52 For the mono-
metallic Pd sample, the heat of adsorption of CO decreased by
~40 kJ mol−1 in the presence of chemisorbed H. At this point,
we do not know to what extent the potential incorporation of
3.5 Neopentane hydrogenolysis and isomerization
Neopentane conversion was performed at 275 °C, 9 psig and
varying flow rates (25–100 cc min−1) in order to maintain dif-
ferential conversion (1–4%) behavior. Two reaction pathways
are possible for this reaction: hydrogenolysis and isomeriza-
tion. Isomerization products include isopentane (single isomer-
ization) and n-pentane if isopentane undergoes isomerization.
A single hydrogenolysis step results in the formation of iso-
butane and methane, and secondary hydrogenolysis reactions
produce additional methane and lighter hydrocarbons (pro-
pane and ethane). The selectivity values (extrapolated to 0%
conversion) and calculated turnover rates (TOR) based on Pd
surface concentration determined from CO chemisorption
are provided in Table 4. Since the activity of gold is several
orders of magnitude lower for neopentane hydrogenolysis
than Pd,38 CO chemisorption is required to determine the
surface concentration of palladium for the calculation of
nominal turnover rates. Products with non-zero selectivity at
0% conversion are considered primary products while products
with 0% selectivity in the limit of 0% conversion are consid-
ered secondary or higher order products.56 The isomerization
selectivity is defined as the selectivity to isopentane since no
n-pentane was observed, and hydrogenolysis products are <C5
molecules.
Table 4 TOR and product selectivity data for PdAu and Pd catalysts
Initial product distribution (%)
TOR
Catalyst
Dispersiona
(mol conv/metal site s−1
)
CH4
C2H6
C3H8
n-C4H10
i-C4H10
i-C5H12
0.52% Pd + 3% Au
1% Pd + 3% Au
2% Pd + 3.5% Au
4% Pd + 3.5% Au
2% Pd/SiO2_3 nm
2% Pd/SiO2_2.5 nm
0.59
0.45
0.43
0.26
0.33
0.40
6.2 × 10−4
3.0 × 10−4
5.8 × 10−4
1.3 × 10−3
7.7 × 10−4
1.0 × 10−3
38
43
43
45
40
49
2
3
0
0
0
1
3
6
4
3
1
5
0
0
1
4
1
2
33
29
33
36
34
33
26
20
19
12
24
9
a
Dispersion determined by CO chemisorption for PdAu catalysts and by STEM for Pd catalysts.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 4366–4377 | 4373

View Article Online
Paper
Catalysis Science & Technology
The isomerization selectivity for all of the PdAu/SiO2
enabling the number of Pd surface atoms to be determined.39,40
Increasing the weight loading of palladium also increases the
particle size determined by chemisorption since more of the
Au nanoparticle surface will be covered by Pd. The increase
in Pd surface composition results in a decrease in isomeriza-
tion selectivity to the extent that the 4% Pd–3.5% Au catalyst
performs almost identically to the monometallic Pd (2.5 nm)
catalyst. This behavior indicates gold must be in close con-
tact to the surface palladium to influence the catalysis.
catalysts falls within the range of the two monometallic Pd
catalysts (Table 4). There is a trend of increasing isomeri-
zation selectivity with decreased palladium weight loading.
Primary hydrogenolysis (i.e., increase in methane selectivity)
increases as the amount of Pd increases in the PdAu/SiO2 cat-
alysts, but the secondary hydrogenolysis product selectivity
doesn't show a consistent trend with Pd content in the bime-
tallic catalysts.
4.2 Comparison of geometric and electronic effects
4.0 Discussion
Fig. 5a shows the dispersion determined by chemisorption
for the pure Pd catalyst from our previous study combined
with the PdAu selectivity data.12 The monometallic Pd catalyst
behavior was determined to be governed by electronic effects
rather than changes in particle size as seen by the lack of a
strong correlation in Fig. 5a. The figure demonstrates the
addition of gold alters the behavior of palladium when it is
in direct contact with the surface palladium. A similar sized
monometallic palladium catalyst (by STEM) has an isomeriza-
tion selectivity of 3% compared to the 0.52% Pd–3% Au cata-
lyst which has a selectivity of 26%. Fig. 5b shows the heat of
adsorption behavior for both Pd and PdAu catalysts. As the
wt.% of Pd in the Pd/Au catalysts is increased, the measured
CO heat of adsorption approaches that of pure Pd. The 4%
Pd–3.5% Au catalyst has a heat of adsorption that fits well
with our correlation from previous work, suggesting that the
catalyst adsorbs CO more like Pd as the amount of added Pd
increases.12 Ward et al. found the heat of adsorption of CO
did not change significantly when Pd was alloyed with Au.39
The small changes were attributed to electronic effects
through electron transfer from Pd to Au; however, the differ-
ences in heats of adsorption were not larger than the stan-
dard deviation given in their data. Neurock et al. performed a
theoretical study of PdAu for ethylene hydrogenation and
found the electronic effect due to rehybridization was
4.1 Explanation of observed activity and selectivity of
PdAu alloys
XEDS measurements demonstrated particles had Pd-rich
shells and Au-rich cores. EXAFS data revealed palladium in
the bimetallic catalysts is under expansive strain compared to
monometallic palladium. The data from EXAFS demonstrates
the Pd–Pd bond length increases from 2.75 to 2.80 Å when
alloyed with Au. This type of lattice expansion has been
documented previously for Pd(shell)/Au(core) catalysts.19,22,57
The activity of the bimetallic catalysts also falls within the
monometallic Pd behavior for neopentane hydrogenolysis
established in our previous work.12 Although the selectivity is
not significantly improved compared to monometallic palla-
dium, the dependence of the selectivity with the effective
particle size has changed. An isomerization selectivity of
~25% requires a monometallic palladium particle of ~3 nm,
whereas the 0.52% Pd–3% Au catalyst with comparable isom-
erization selectivity has an effective particle size of 1.7 nm as
determined by CO chemisorption. Dispersion determined by
chemisorption is a more relevant determination of the num-
ber of active sites than STEM since Au has an almost negligi-
ble neopentane hydrogenolysis activity compared to Pd.38,58
In addition, CO does not bind to Au under the conditions of
the chemisorption experiment; the temperature is too high,
Fig. 5 a) Isomerization selectivity vs. particle size determined by CO chemisorption and b) initial CO heat of adsorption vs. isomerization
selectivity of monometallic Pd (black squares) and PdAu (red circles) catalysts.
4374 | Catal. Sci. Technol., 2014, 4, 4366–4377
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
negligible compared to the contribution from geometric
effects caused by the weakening of carbon and hydrogen
bonding to Au.34 Both figures suggest the gold must be close
to the surface to have an effect on the catalytic performance
and that the composition of the core of the particle is not
important if the shell is thick enough to be contiguous,
enabling the bimetallic particle to function as a monometallic
nanoparticle.
The expansion of the Pd–Pd lattice seen in EXAFS may
explain the changes in reactivity since the expansion may
influence the adsorption configuration of neopentane. As
suggested by Anderson and Avery, the adsorption of
neopentane may require multi-site attachment to contiguous
Pd atoms.59 Changes in the bond distance between Pd atoms
may affect the strength of neopentane adsorption to the sur-
face or the ability of the surface to perform C–H and C–C
bond cleaving reactions. The reduction in hydrogenolysis
product selectivity would explain the increase in isomeriza-
tion selectivity and the similarity of the turnover rate for
neopentane disappearance due to the loss of ensembles for
C–H and C–C bond cleavage. Although the performance of Pd
has not been improved significantly by the addition of Au,
two catalysts were created (the 3.0 nm Pd and the 0.52%
Pd–3% Au) with significantly different surface bonding char-
acteristics but very similar isomerization selectivity. This data
also highlights the influence of the near-surface composition
compared to the bulk properties of a nanoparticle on the cat-
alytic performance.60 The presence of gold alters the behavior
of palladium relative to a monometallic nanoparticle with a
similar effective diameter, however, the further the gold is
from the catalyst surface the effect on the catalyst perfor-
mance is lessened such that the 4% Pd–3.5% Au catalyst per-
forms like pure Pd.
Corner and edge sites typically bind adsorbates more
strongly and are more active for bond-breaking reactions
compared to terrace sites.61–63 To interrogate the types of
sites available, CO adsorption on these catalysts monitored
by DRIFTS has been used. The addition of gold decreases the
amount of bridge-bound CO as seen in Fig. 4, which could
indicate an increase in the fraction of low coordination cor-
ner and edge sites. From previous work, Pd catalysts with low
linear-to-bridge ratios resulted in higher isomerization selec-
tivity,12 but the PdAu catalysts do not follow this trend as the
0.52% Pd–3% Au and 1% Pd–3% Au catalysts had the highest
linear-to-bridge ratio and the highest isomerization selectivity
of all catalysts. Both these catalysts have an isomerization
selectivity similar to the largest Pd particles tested in our
previous work; however, the linear-to-bridge ratios are signifi-
cantly different. The best Pd catalysts in terms of isomeriza-
tion selectivity had a linear-to-bridge ratio of ~0.15 : 1 while
the PdAu catalysts with similar selectivity behavior have a
ratio of 0.59 : 1 (0.52% Pd–3% Au) and 0.94 : 1 (1% Pd–3% Au).12
Fig. 6 shows the isomerization and hydrogenolysis turnover
rates versus the linear-to-bridge ratio for the PdAu catalysts.
The neopentane hydrogenolysis rate decreases with increasing
linear bound CO while the isomerization rate is insensitive to
Fig. 6 Hydrogenolysis (blue circles) and isomerization (red squares)
turnover rate versus the linear to bridge ratio from DRIFTS of CO
adsorption.
changes in surface structure. As mentioned above, hydro-
genolysis is a well-known structure-sensitive reaction which
requires a minimum ensemble of two active sites.64–66 The
increase in linear bound CO suggests that large Pd ensem-
bles are broken up by Au. Although our previous work with
neopentane conversion suggests that the active site for
hydrogenolysis and isomerization are similar, apparently, the
presence of Au does not prevent formation of the transition
state for isomerization. This suggests that the hydrogenolysis
activity is altered by geometric effects although the overall
neopentane conversion activity is not greatly affected.
Fig. 7 shows the relationship between neopentane isomeri-
zation selectivity compared to the surface structure assessed
by DRIFTS. Decreasing bridge bound CO results in an
increase in isomerization selectivity but this effect appears to
reach a maximum at around 0.55 where the selectivity levels
off. This suggests that breaking up large Pd ensembles on
the surface is desirable, but limited as an approach to
increase isomerization selectivity.
Fig. 7 Linear-to-bridge ratio from DRIFTS of CO adsorption vs.
isomerization selectivity for PdAu catalysts.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 4366–4377 | 4375

View Article Online
Paper
Catalysis Science & Technology
4 K. Balakrishnan and J. Schwank, J. Catal., 1991, 132,
451–464.
5.0 Conclusion
Alloying palladium with gold on a silica support to form a
Pd(shell)–Au(core) nanoparticle improved the neopentane
isomerization selectivity compared to a silica-supported
monometallic palladium catalyst with a similar particle size.
This improvement was more significant when the Pd shell
was influenced by the underlying Au core (i.e., lower Pd
weight loading catalysts). As the amount of palladium added
to the catalyst increased, the catalytic behavior was similar to
monometallic Pd. The turnover rate for hydrogenolysis
decreased as the fraction of Au increased which correlated
with an increase CO linear-to-bridge ratio. Calorimetry
showed that the electronic effect of Au addition lessened with
increasing Pd wt.%. Although both electronic and geometric
effects have been observed, the changes in catalytic perfor-
mance correlate more with changes in geometry. The identity
of the metal atoms in the core of the particle appears to be
less significant to catalytic properties than the surface com-
position and metal atoms near the surface.
5 T. P. Wu, D. J. Childers, C. Gomez, A. M. Karim,
N. M. Schweitzer, A. J. Kropf, H. Wang, T. B. Bolin, Y. F. Hu,
L. Kovarik, R. J. Meyer and J. T. Miller, ACS Catal., 2012, 2,
2433–2443.
6 A. Gross, Top. Catal., 2006, 37, 29–39.
7 J. Sehested, K. E. Larsen, A. L. Kustov, A. M. Frey,
T. Johannessen, T. Bligaard, M. P. Andersson, J. K. Norskov
and C. H. Christensen, Top. Catal., 2007, 45, 9–13.
8 B. Hammer and J. K. Norskov, Surf. Sci., 1995, 343, 211–220.
9 B. Hammer and J. K. Norskov, Impact of Surface Science on
Catalysis, in Advances in Catalysis, ed. B. C. Gates and
H. Knozinger, 2000, vol. 45, pp. 71–129.
10 J. Kleis, J. Greeley, N. A. Romero, V. A. Morozov, H. Falsig,
A. H. Larsen, J. Lu, J. J. Mortensen, M. Dulak,
K. S. Thygesen, J. K. Norskov and K. W. Jacobsen, Catal.
Lett., 2011, 141, 1067–1071.
11 N. Schweitzer, H. L. Xin, E. Nikolla, J. T. Miller and S. Linic,
Top. Catal., 2010, 53, 348–356.
12 D. Childers, A. Saha, N. Schweitzer, R. M. Rioux, J. T. Miller
and R. J. Meyer, ACS Catal., 2013, 3, 2487–2496.
13 F. Gao and D. W. Goodman, Chem. Soc. Rev., 2012, 41,
8009–8020.
14 J. Rebelli, A. A. Rodriguez, S. G. Ma, C. T. Williams and
J. R. Monnier, Catal. Today, 2011, 160, 170–178.
15 M. Bonarowska, J. Pielaszek, W. Juszczyk and Z. Karpinski,
J. Catal., 2000, 195, 304–315.
16 P. Dash, T. Bond, C. Fowler, W. Hou, N. Coombs and
R. W. J. Scott, J. Phys. Chem. C, 2009, 113, 12719–12730.
17 N. El Kolli, L. Delannoy and C. Louis, J. Catal., 2013, 297,
79–92.
18 A. M. Venezia, L. F. Liotta, G. Pantaleo, V. La Parola,
G. Deganello, A. Beck, Z. Koppany, K. Frey, D. Horvath and
L. Guczi, Appl. Catal., A, 2003, 251, 359–368.
19 A. Sarkany, O. Geszti and G. Safran, Appl. Catal., A, 2008,
350, 157–163.
20 K. Luo, T. Wei, C. W. Yi, S. Axnanda and A. W. Goodman,
J. Phys. Chem. B, 2005, 109, 23517–23522.
21 S. W. T. Price, J. M. Rhodes, L. Calvillo and A. E. Russell,
J. Phys. Chem. C, 2013, 117, 24858–24865.
22 A. F. Lee, C. J. Baddeley, C. Hardacre, R. M. Ormerod,
R. M. Lambert, G. Schmid and H. West, J. Phys. Chem., 1995,
99, 6096–6102.
Acknowledgements
JTM and NS were supported as part of the Institute for
Atom-Efficient Chemical Transformations (IACT), an Energy
Frontier Research Center funded by the U.S. Department of
Energy, Office of Science, Office of Basic Energy Sciences.
R. J. M. and D. C. gratefully acknowledge funding for this
work from the National Science Foundation (CBET grant no.
0747646). Partial funding for DC was provided by the Chemi-
cal Sciences and Engineering Division at Argonne National
Laboratory and the Office of the Vice Chancellor for Research
at the University of Illinois at Chicago. S. M. K. S. and R. M. R.
acknowledge funding from the Department of Energy, Office
of Basic Energy Sciences, Chemical Sciences, Geosciences,
and Biosciences Division, Catalysis Sciences Program under
grant number DE-FG02-12ER16364. R. M. R. acknowledges
financial support provided through a 3M Non-Tenured
Faculty Grant (NTFG). The STEM work was performed at the
UIC Research Resource Center. Use of the Advanced Photon
Source was supported by the U.S. Department of Energy,
Office of Science, Office of Basic Energy Sciences, under
contract no. DE-AC02-06CH11357. MRCAT operations are
supported by the Department of Energy and the MRCAT mem-
ber institutions. The CleanCat Core facility acknowledges
funding from the Department of Energy (DE-FG02-03ER15457
and DE-AC02-06CH11357) used for the purchase of the
DRIFTS system and the AMI-200, respectively.
23 M. S. Chen, D. Kumar, C. W. Yi and D. W. Goodman,
Science, 2005, 310, 291–293.
24 X. Yang, C. Huang, Z. Y. Fu, H. Y. Song, S. J. Liao, Y. L. Su,
L. Du and X. J. Li, Appl. Catal., B, 2013, 140, 419–425.
25 P. Dolle, R. Baudoing-Savois, M. De Santis, M. C. Saint-Lager,
M. Abel, J. C. Bertolini and P. Delichère, Surf. Sci., 2002, 518,
1–13.
References
1 J. A. Rodriguez and D. W. Goodman, J. Phys. Chem.,
1991, 95, 4196–4206.
26 Z. Li, F. Gao, O. Furlong and W. T. Tysoe, Surf. Sci., 2010,
604, 136–143.
2 J. A. Rodriguez and D. W. Goodman, Science, 1992, 257,
897–903.
3 J. H. Sinfelt, Sci. Am., 1985, 253, 90–98.
27 B. Zhu, G. Thrimurthulu, L. Delannoy, C. Louis, C. Mottet,
J. Creuze, B. Legrand and H. Guesmi, J. Catal., 2013, 308,
272–281.
4376 | Catal. Sci. Technol., 2014, 4, 4366–4377
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
28 F. Gao, Y. L. Wang and D. W. Goodman, J. Phys. Chem. C,
2009, 113, 14993–15000.
47 A. Hugon, L. Delannoy, J.-M. Krafft and C. Louis, J. Phys.
Chem. C, 2010, 114, 10823–10835.
29 G. Mazzone, I. Rivalta, N. Russo and E. Sicilia, J. Phys. Chem.
C, 2008, 112, 6073–6081.
48 J. B. Giorgi, T. Schroeder, M. Baumer and H. J. Freund, Surf.
Sci., 2002, 498, L71–L77.
30 H. L. Abbott, A. Aumer, Y. Lei, C. Asokan, R. J. Meyer,
M. Sterrer, S. Shaikhutdinov and H. J. Freund, J. Phys. Chem.
C, 2010, 114, 17099–17104.
31 V. Soto-Verdugo and H. Metiu, Surf. Sci., 2007, 601, 5332–5339.
32 D. Yuan, X. Gong and R. Wu, Phys. Rev. B: Condens. Matter
Mater. Phys., 2007, 75, 233401.
49 J. S. Bradley, E. W. Hill, S. Behal, C. Klein, B. Chaudret and
A. Duteil, Chem. Mater., 1992, 4, 1234–1239.
50 S. Marx, F. Krumeich and A. Baiker, J. Phys. Chem. C, 2011,
115, 8195–8205.
51 A. M. Venezia, V. La Parola, G. Deganello, B. Pawelec and
J. L. G. Fierro, J. Catal., 2003, 215, 317–325.
33 T. Wei, J. Wang and D. W. Goodman, J. Phys. Chem. C, 2007,
111, 8781–8788.
52 G. Rupprechter, M. Morkel, H. J. Freund and R. Hirschl,
Surf. Sci., 2004, 554, 43–59.
34 M. Neurock and D. H. Mei, Top. Catal., 2002, 20, 5–23.
35 M. Garcia-Mota and N. Lopez, Phys. Rev. B: Condens. Matter
Mater. Phys., 2010, 82, 075411.
36 K. Balakrishnan and J. Schwank, J. Catal., 1991, 132, 451–464.
37 J. R. Regalbuto, Catalyst Preparation: Science and Engineering,
CRC Press, 2006.
38 M. Boudart and L. D. Ptak, J. Catal., 1970, 16, 90–96.
39 T. Ward, L. Delannoy, R. Hahn, S. Kendell, C. J. Pursell,
C. Louis and B. D. Chandler, ACS Catal., 2013, 3, 2644–2653.
40 F. Menegazzo, M. Manzoli, A. Chiorino, F. Boccuzzi,
T. Tabakova, M. Signoretto, F. Pinna and N. Pernicone,
J. Catal., 2006, 237, 431–434.
41 M. A. Vannice and C. C. Twu, J. Chem. Phys., 1981, 75,
5944–5948.
42 L. Bollmann, J. L. Ratts, A. M. Joshi, W. D. Williams,
J. Pazmino, Y. V. Joshi, J. T. Miller, A. J. Kropf, W. N. Delgass
and F. H. Ribeiro, J. Catal., 2008, 257, 43–54.
43 J. T. Miller, A. J. Kropf, Y. Zha, J. R. Regalbuto, L. Delannoy,
C. Louis, E. Bus and J. A. van Bokhoven, J. Catal., 2006, 240,
222–234.
44 M. R. Knecht, M. G. Weir, A. I. Frenkel and R. M. Crooks,
Chem. Mater., 2008, 20, 1019–1028.
53 W. Y. Yu, G. M. Mullen and C. B. Mullins, J. Phys. Chem. C,
2014, 118, 2129–2137.
54 W. Y. Yu, G. M. Mullen and C. B. Mullins, J. Phys. Chem. C,
2013, 117, 19535–19543.
55 M. E. Bjorketun, G. S. Karlberg, J. Rossmeisl, I. Chorkendorff,
H. Wolfschmidt, U. Stimming and J. K. Norskov, Phys. Rev. B:
Condens. Matter Mater. Phys., 2011, 84, 045407.
56 N. A. Bhore, M. T. Klein and K. B. Bischoff, Ind. Eng. Chem.
Res., 1990, 29, 313–316.
57 T. Akita, T. Hiroki, S. Tanaka, T. Kojima, M. Kohyama,
A. Iwase and F. Hori, Catal. Today, 2008, 131, 90–97.
58 W. Juszczyk, D. Lomot, Z. Karpinski and J. Pielaszek, Catal.
Lett., 1995, 31, 37–45.
59 J. R. Anderson and N. R. Avery, J. Catal., 1966, 5, 446–463.
60 J. Greeley and M. Mavrikakis, Nat. Mater., 2004, 3, 810–815.
61 B. Hammer, Top. Catal., 2006, 37, 3–16.
62 S. Dahl, A. Logadottir, R. C. Egeberg, J. H. Larsen,
I. Chorkendorff, E. Tornqvist and J. K. Norskov, Phys. Rev.
Lett., 1999, 83, 1814–1817.
63 R. C. Egeberg, S. Dahl, A. Logadottir, J. H. Larsen,
J. K. Norskov and I. Chorkendorff, Surf. Sci., 2001, 491,
183–194.
45 T. Lear, R. Marshall, J. A. Lopez-Sanchez, S. D. Jackson,
T. M. Klapotke, M. Baumer, G. Rupprechter, H. J. Freund
and D. Lennon, J. Chem. Phys., 2006, 124.
46 J. L. Lu, B. S. Fu, M. C. Kung, G. M. Xiao, J. W. Elam,
H. H. Kung and P. C. Stair, Science, 2012, 335, 1205–1208.
64 D. W. Blakely and G. A. Somorjai, J. Catal., 1976, 42,
181–196.
65 D. W. Flaherty, D. D. Hibbitts, E. I. Gürbüz and E. Iglesia,
J. Catal., 2014, 311, 350–356.
66 J. H. Sinfelt, CRC Crit. Rev. Solid State Sci., 1973, 4, 311–332.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 4366–4377 | 4377