Structural evolution of bimetallic Pd-Ru catalysts in oxidative and reductive applications

Article abstract of DOI:10.1016/j.apcata.2015.06.025

Abstract Two types of bimetallic Pd-Ru catalysts with a 2:1 Ru:Pd molar ratio were prepared using a poly-(vinylpyrrolidone) stabilizer: one alloy structure with mixed-surface atoms and one core-shell structure with a Pd core and Ru shell, which were confirmed by a surface-probe reaction at mild conditions. In indan hydrogenolysis at 350 °C, inversion of the core-shell structure began with Pd atoms appearing on the surface of the particles. Both catalysts displayed distinctively different catalytic behavior and indicated the importance of structure control for this particular application within a studied time frame. For methane combustion over the 200-550 °C temperature range, both structures demonstrated identical activity, which was due to their structural evolution to one nanoparticle type with Pd-enriched shells, as evidenced by extended X-ray absorption fine structure.

Full text of DOI:10.1016/j.apcata.2015.06.025

Applied Catalysis A: General 502 (2015) 350–360
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Structural evolution of bimetallic Pd-Ru catalysts in oxidative and
reductive applications
a
b
a
a,∗
Jing Shen , Robert W.J. Scott , Robert E. Hayes , Natalia Semagina
a
Department of Chemical and Material Engineering, University of Alberta, 9107-116 St. Edmonton, Alberta T6G 2V4, Canada
Department of Chemistry, University of Saskatchewan, 110 Science Place, Saskatoon, Saskatchewan S7N 5C9, Canada
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 4 March 2015
Received in revised form 27 May 2015
Accepted 20 June 2015
Available online 26 June 2015
Two types of bimetallic Pd-Ru catalysts with a 2:1 Ru:Pd molar ratio were prepared using a poly-
(vinylpyrrolidone) stabilizer: one alloy structure with mixed-surface atoms and one core–shell structure
with a Pd core and Ru shell, which were confirmed by a surface-probe reaction at mild conditions. In
◦
indan hydrogenolysis at 350 C, inversion of the core–shell structure began with Pd atoms appearing on
the surface of the particles. Both catalysts displayed distinctively different catalytic behavior and indi-
cated the importance of structure control for this particular application within a studied time frame. For
Keywords:
Ring opening
EXAFS
Structure inversion
Ruthenium alloys
PVP
◦
methane combustion over the 200–550 C temperature range, both structures demonstrated identical
activity, which was due to their structural evolution to one nanoparticle type with Pd-enriched shells, as
evidenced by extended X-ray absorption fine structure.
© 2015 Elsevier B.V. All rights reserved.
1
. Introduction
Introducing a second metal component to supported catalysts is
In thermodynamically driven changes, achieving prolonged
catalyst stability is still feasible (especially in low-temperature
regions) because the structural evolution is often kinetically lim-
ited. The adsorbent-induced segregation [7,8], on the contrary,
may happen on the same time scale as the catalytic reaction and
requires lower temperatures and shorter times to modify the par-
ent bimetallic nanoparticles. For example, a ternary CuPdY alloy did
not segregate in a vacuum, but in the presence of oxygen, Y oxides
were formed on the surface, driven by preferential oxygen-induced
Y segregation [11]. Similarly, parent alloy Pt-Co nanoparticles
evolved into Pt(shell)/Pt-Co(core) particles upon annealing in CO
[7].
Studies of the particle structural evolution during pre-treatment
or catalysis are complicated not only by the limited availability of
operando and in situ catalyst characterization techniques but also
by the nature of the catalysts themselves. The traditional prepa-
ration methods that support impregnation with metal precursor
solutions followed by calcination do not allow the production
of well-defined uniform nanoparticles, so their structural evolu-
tion studies are complicated by the wide polydispersity of the
original nanoparticles. At present, the steadily increasing num-
ber of colloidal chemistry techniques to prepare structure- and
size-controlled metal nanoparticles [12–15] enable experimental
studies of the bimetallic nanoparticle rearrangement, including
structure inversion during synthesis or after thermal activation
and/or catalysis [16–18]. The structural changes depend on the
widely recognized as an efficient way to increase catalyst activity,
selectivity, and stability. Numerous large-scale industrial produc-
tions benefit from synergism phenomena in bimetallic catalysis
[
1–5]. Two metals may form intrinsic alloys with or without pref-
erential shell or core segregation of one of the metals, or individual
monometallic nanoparticles. Such structures, however, often are
not maintained during catalysis [6], leading to unstable catalyst
performanceand even deactivation. The changesare drivenbyther-
modynamics (the metal with lower surface energy segregates to
the surface of bimetallic nanoparticles) and/or chemically activated
selective metal diffusion (“adsorbent-induced segregation”) due to
different metal affinities to the reaction mixture components [7,8].
Thermodynamic considerations predict that nanoparticles with the
same metal ratios but initially different structures at high enough
temperatures transform into one most- thermodynamically-stable
configuration [9]. At different metal ratios, nanoparticles may be
stabilized in different core–shell or alloyed structures even at tem-
peratures above 1600 K [10].
∗ Corresponding author: Fax: +1 780 492 2881.
E-mail addresses: jshen@ualberta.ca (J. Shen), robert.scott@usask.ca
(
(
R.W.J. Scott), bob.hayes@ualberta.ca (R.E. Hayes), semagina@ualberta.ca
N. Semagina).
http://dx.doi.org/10.1016/j.apcata.2015.06.025
926-860X/© 2015 Elsevier B.V. All rights reserved.
0

J. Shen et al. / Applied Catalysis A: General 502 (2015) 350–360
351
Scheme 1. Structure-controlled synthesis of alloy and core–shell bimetallic nanoparticles.
atmosphere (for example, oxidizing or reducing) and may or may
not take place depending on the involved metals [15].
on ␥-Al O3 support with 5.8 nm pores. Monometallic Ru nanopar-
2
ticles were prepared using the ethylene glycol reduction method
[19] with some modifications [20]. At room temperature, 0.2 mmol
Ru(NO)(NO3)3 and PVP (2 mmol, PVP-to-Ru molar ratio = 10/1)
were dissolved in 200 mL EG in a 500 mL single-neck round-
bottom flask. The reaction temperature was increased from room
The objective of this work is to demonstrate the structural
changes of well-defined bimetallic nanoparticles in reductive and
oxidative catalytic reactions. We start with two different Pd-Ru
nanostructures synthesized with a protective stabilizer either via
simultaneous reduction of the metal precursors (“alloy” Ru2Pd1)
or via Ru atom deposition on the core of preformed Pd nanopar-
ticles (“core–shell” Pd(c)Ru(s) sample), as depicted in Scheme 1.
The overall metal ratio is the same in both structures so that their
structural evolution is not affected by their composition [10]. We
further use these nanoparticles as catalysts for the hydrogenolysis
◦
temperature to the reflux point of EG (198 C), and maintained
◦
at 198 C for 3 h under stirring. Monometallic Pd nanoparticles
were synthesized by the one-step alcohol reduction method [21]
with some modifications [20]. A mixture containing 0.2 mmol
PdCl , PVP (4 mmol, PVP-to-Pd molar ratio = 20/1) and 170 mL of
2
ethanol/water (41 vol.% ethanol) was stirred and refluxed in a
500 mL single-neck round-bottom flask for 3 h under air.
◦
◦
of indan at 350 C and combustion of methane in air at up to 550 C.
In the former case, the structural evolution is minimal, and the
alloy and Pd(core)-Ru(shell) nanoparticles’ performance is distinc-
tively different. In methane combustion, both structures transform
into the very same nanoparticle structure with a Pd-enriched shell
and show similar catalytic performance. This comparative study
demonstrates the advantages and limitations of the structure-
controlled bimetallic nanoparticle synthesis for different catalytic
applications.
Bimetallic Ru-Pd nanoparticles with different surfaces were pre-
pared according to our previously published synthesis techniques
as shown in Scheme 1 [20]. For the Ru2Pd1 alloy, the monometal-
lic Pd synthetic procedure was followed using 0.133 mmol of Ru
and 0.067 mol Pd precursors. The Pd (core)-Ru (shell) nanoparticles
were synthesized by Teranishi and Miyake’s stepwise growth-
reaction method in ethanol/water [21] with some modifications
[20]. The Pd core colloids were prepared using the synthesis
method for monometallic Pd nanoparticles. The pre-synthesized
Pd seed colloidal solution (0.05 mmol Pd, 42.5 mL) and 0.1 mmol
Ru(NO)(NO3)3 were dissolved in 170 mL ethanol/water (41 vol.%
ethanol) at room temperature, without more fresh PVP addition.
The mixture was then stirred and refluxed for 3 h to allow the
formation of the Ru shell.
2
. Experimental
2.1. Materials
Ruthenium (III) nitrosyl nitrate (Ru(NO)(NO ) , Ru 31.3% min,
3
3
◦
The catalyst support, ␥-Al O , was dried in static air at 500 C
2
3
Alfa Aesar), palladium(II) chloride solution (PdCl , 5% w/v, Acros),
2
for 3 h. The PVP-stabilized monometallic Ru nanoparticles pre-
pared in EG were precipitated with acetone (acetone-to-colloidal
solution volumetric ratio = 4/1), and deposited on ␥-Al O by incip-
poly-(vinylpyrrolidone) (PVP) (MW: 40,000, Sigma–Aldrich),
reagent alcohol (ethanol, 95 vol.%, Fisher Scientific), ethylene glycol
2
3
(
EG, 99.8%, Sigma–Aldrich), gamma aluminum oxide (␥-Al O , 150
2 3
ient wetness impregnation. The ethanol/water colloidal solutions
of PVP-stabilized monometallic Pd and bimetallic Ru-Pd nanopar-
ticles were concentrated by vacuum evaporation of solvents,
followed by incipient wetness impregnation on the ␥-Al O sup-
mesh 58 Å pore size, Sigma–Aldrich), and acetone (99.7%, Fisher Sci-
entific) were used as received. MilliQ water was used throughout
the work.
2
3
◦
port. Finally, all catalysts were dried in static air at 60 C overnight.
The target total metal(s) loading was 0.3 wt.%.
2.2. Catalyst preparation
The actual loadings in the supported catalysts were determined
by neutron activation analysis as 0.250 wt.% Ru/␥-Al O ; 0.240 wt.%
All catalysts were prepared by synthesizing PVP-stabilized
–3 nm nanoparticles in colloidal solutions, followed by deposition
2
3
2

3
52
J. Shen et al. / Applied Catalysis A: General 502 (2015) 350–360
Pd/␥-Al O ; 0.154 wt.% Ru, 0.065 wt.% Pd in Ru2Pd1/␥-Al O ; and
sure drop in the gas burette was recorded. The initial reaction rate
was calculated from the slope of the hydrogen consumption graph
after complete dissolution of hydrogen in ethanol.
2
3
2
3
0.200 wt.% Ru, 0.072 wt.% Pd in Pd(c)Ru(s) catalysts. The Ru-to-Pd
molar ratios in the final catalysts are 2.4 and 2.8 for the Ru2Pd1 and
Pd(c)Ru(s) samples.
Extended X-ray absorption fine structure spectroscopy (EXAFS)
was performed at Canadian Light Source (Saskatchewan). X-ray
absorption spectra at the Pd K-edge and the Ru K-edge were
recorded at the HXMA beamline 061D-1 (energy range, 5–30 keV;
2.3. Catalyst characterization
−
4
The PVP-stabilized nanoparticles and/or pretreated catalysts
were characterized by transmission electron microscopy (TEM),
selected area electron diffraction (SAED), and high-resolution TEM
resolution, 1 × 10
ꢂE/E) at the Canadian Light Source (CLS,
2.9 GeV storage ring, ∼250 mA current). All samples were pressed
into pellets and measured in transmission mode at room tempera-
ture. Samples for EXAFS analyses were prepared with an expected
(
HRTEM) at 200 kV on a JEOL 2100 transmission electron micro-
scope (Cell Imaging Facility, University of Alberta). The mean
diameter and standard deviation of nanoparticles were calculated
by counting more than 200 particles from TEM images using ImageJ
software. X-ray powder diffraction (XRD) patterns were recorded
on a -Bruker AXS diffractometer (Department of Chemical and
Material Engineering, University of Alberta) with a Cu-K␣ radia-
tion source (ꢀ = 1.54059 Å) at 40 kV and 44 mA. Continuous X-ray
metal(s) loading of 2 wt.%. The Ru-Pd catalysts for EXAFS study were
◦
pre-calcined at 550 C for 16 h under air followed by H reduction at
2
◦
400 C, which is confirmed to be enough to reduce oxidized Ru-Pd
nanoparticles to metallic forms by our previous TPR analysis [20].
After the reduction the catalysts were kept at ambient tempera-
ture under air atmosphere for one week followed by pre-reduction
◦
in hydrogen at 350 C in situ before the EXAFS analyses. A double-
◦
◦
scans were carried out from 2ꢁ of 10 to 110 with a step width of
crystal Si(220) monochromator was employed for energy selection.
Higher harmonics were eliminated by detuning the double-crystal
Si(220) by using a Pt-coated 100 mm long KB mirror.
◦
◦
0
.05 and a scan speed of 2 /min. XRD peak identification and data
processing were performed using MDI Jade 9.0 software combined
with the ICDD database. The colloidal solutions were concentrated
by rotary evaporation of solvents under vacuum, followed by dry-
The IFEFFIT software package was used for EXAFS data process-
ing [22]. More details about obtaining EXAFS function can be found
in previous publications [23,24]. The EXAFS fitting was performed
in R-space using theoretical phase shifts and amplitudes generated
by FEFF. Lattice parameters and first shell coordination numbers of
12 for bulk fcc Pd and (or hcp Ru) were used to generate the ampli-
tude reduction factor for Pd (or Ru) by fitting Pd foil (or Ru foil).
The amplitude reduction factors found from Pd and Ru foils were
0.828 and 0.773, respectively. Ru-Pd bimetallic systems were fitted
using bulk Pd lattice parameters, as XRD and SAED confirmed the fcc
structures of the bimetallic nanoparticles. Additionally, the lattice
spacings for the Ru-Pd bimetallic catalysts calculated from (1 1 1)
diffractions are very close to that of monometallic Pd nanoparticles,
which will be discussed later.
◦
ing in an oven at 60 C under air for a day. Nanoparticle powders
(
without further deposition on alumina support) were collected for
XRD analyses.
The actual loadings of Ru and Pd on ␥-Al O before and after pre-
2
3
treatment were determined by neutron activation analysis (NAA)
at Becquerel Laboratories (Maxxam Company, Ontario). Samples
were irradiated for 20 min in the Cd shielded, epithermal site of
the reactor core. Palladium and ruthenium were counted for 15 min
after 24 h decay on an Aptec CS13-A31C gamma detector.
CO chemisorption analyses were performed by dosing 3% CO/He
gas mixture at room temperature with an AutoChem 2950HP
instrument equipped with a quartz U-tube reactor and a thermal
conductivity detector (TCD). The volumetric flow rates of CO/He
loop gas and the He carrier gas were 25 mL/min. Prior to CO
2.4. Catalysis in reductive atmosphere: indan ring opening in
◦
◦
chemisorption experiments, the catalysts were calcined at 550 C
hydrogen, 350 C
in air for 16 h in a furnace. The precalcined catalysts were reduced
◦
in a flow of 10% H /Ar (25 mL/min) at 550 C for 1 h. After the
The developed monometallic and bimetallic Ru-Pd catalysts
were studied in a low-pressure indan ring opening, as described
in our previous study [20]. The PVP-stabilized catalysts were cal-
2
calcination-reduction pretreatment, the samples were purged with
argon for 30 min at 550 C and cooled to ambient temperature
◦
◦
◦
under inert atmosphere. The results were corrected by subtracting
the CO uptakes of alumina support.
Diffuse reflectance infrared spectra of the adsorbed CO (CO-
DRIFTS) were obtained using NEXUS 670 FT-IR fitted with a Smart
Diffuse Reflectance accessory. DRIFT spectra were recorded against
a KBr standard (256 scans, 4 cm resolution). Data processing was
performed with OMNIC software. The detailed experimental proce-
dures including catalyst pretreatment can be found in our previous
work [20]. Monometallic Pd and Ru and bimetallic Pd(c)Ru(s) sam-
cined at 200 C in static air, followed by in situ reduction at 375 C
in a flow of hydrogen (80 mL/min). The catalytic indan ring opening
◦
was carried out at an internal temperature of 350 C and 1 atm pres-
−
6
sure. Indan ((4.7 ± 0.6) × 10 mol/min flow rate) was fed into the
catalytic system by bubbling 120 mL/min hydrogen through indan
−1
◦
at a constant temperature of 10 C. The gas outlet from the reac-
tor was analyzed online using a Varian 430 gas chromatograph
equipped with FID. As reported previously [20], the ring opening
products are 2-ethyltoluene, n-propylbenzene, o-xylene, ethylben-
zene, toluene, benzene, and lights (mainly C and C ), in which only
◦
ples were calcined at 200 C for 1 h in air, and followed by reduction
in a flow of 10% H /Ar at 375 C for 30 min to simulate condi-
1
2
◦
2-ethyltoluene and n-propylbenzene are the desired ring opening
2
tions of pretreatment before the catalytic indan ring opening. Prior
to CO treatment, the sample was purged with Ar at 375 C, and
products with a naphthenic ring being cleaved once.
◦
cooled down to room temperature in Ar. Then 3% CO/He was passed
through the sample for 30 min, followed by degassing in an Ar
environment to remove the physically adsorbed CO.
2.5. Catalysis in oxidative atmosphere: methane combustion in
air, up to 550 C
◦
The hydrogenation of 2-methyl-3-buten-2-ol (MBE) to 2-
methylbutan-2-ol (MBA) that is catalyzed by Pd only was selected
to elucidate whether Pd atoms are present in the outermost layer
of bimetallic Ru-Pd nanoparticles, as described previously [20]. A
semibatch stainless reactor was filled with 0.04 M MBE in 200 mL
ethanol and 0.5 g of as-prepared catalyst. The MBE hydrogenation
Methane combustion was investigated over the developed Ru-
Pd nanocatalysts according to the previous study published by
Abbasi et al. [25]. A 20 inch long tubular reactor with an inner diam-
◦
eter of 3/8 inch was packed with calcined catalysts (550 C, 16 h in
static air) corresponding to 1.2 mg active Pd (and/or 3.1 ± 0.4 mg
Ru). Layers of quartz wool were placed at both ends of the cata-
lyst bed to hold the catalyst in place. The reactor was then placed
inside a furnace equipped with a temperature controller. The inter-
◦
reaction was carried out at 40 C, 0.45 MPa absolute pressure and
1
200 rpm stirring speed. During the reaction, the hydrogen pres-

J. Shen et al. / Applied Catalysis A: General 502 (2015) 350–360
353
nal reaction temperatures were measured by two thermocouples at
each end of the catalyst bed. The flow rates of the feed gases (10%
CH /N and air) were regulated by mass flow controllers (Math-
3. Results and discussion
3.1. Characterization of as-synthesized nanoparticles
4
2
eson and MKS). Methane (10% balanced in nitrogen, 8.5 mL/min)
and air (extra-dry, 200 mL/min) were pre-mixed and fed into the
catalytic system. The concentration of CH4 in the gas mixture was
We intentionally selected two different procedures for the
bimetallic catalyst preparation to obtain nanoparticles with differ-
ent structures, as outlined in Scheme 1. The ruthenium precursor,
without the palladium seeds present, could not be reduced in
ethanol, so no monometallic Ru nanoparticles could be obtained
in the ethanol solution used for syntheses. The presence of palla-
dium nanoparticles facilitated Ru3+ reduction by an autocatalytic
surface-growth mechanism, as suggested by Liu et al. for similar
Pt-Ru systems [27].
Fig. 1 shows HRTEM images and SAED patterns of PVP-stabilized
Pd(c)Ru(s) and Ru2Pd1 nanoparticles; the samples were prepared
from colloidal solutions without further treatment. Both materi-
als are single crystals with an fcc structure as follows from the
comparison of the observed ring-diffraction patterns with the sim-
ulated fcc pattern. No monometallic hcp Ru structures were found,
as expected from the synthetic procedure. Statistical analysis of
the nanoparticle size distribution based on TEM revealed the sizes
for monometallic Pd, monometallic Ru, Pd(c)Ru(s) and Ru2Pd1 sys-
tems to be 2.0 ± 0.5 nm, 2.0 ± 0.3 nm, 2.3 ± 0.7 nm and 2.6 ± 0.6 nm,
respectively, as reported in our previous work [20]. Larger sizes of
the bimetallic samples as compared to the monometallic forms sug-
gest the formation of intrinsic bimetallic structures. As-synthesized
Ru-Pd catalysts were in metallic form; however, a small degree of
Pd and Ru surface oxidation was possible after exposing to air for
few weeks (at room temperature), as shown in our previous XPS
analysis [28].
4
100 ppm. Methane oxidation reactions in this study were per-
formed as described in our previous work [26], except the catalysts
were tested at dry conditions. Methane oxidation was carried out at
a constant pressure of 30 psi. Ignition and extinction curves were
obtained by increasing and decreasing the reaction temperature
◦
◦
stepwise (50 C for each step; 60 C/min ramping rate), respec-
◦
tively. The ignition curves were initiated at 200 C and ended at
50 C, which refer to the lowest and highest combustion tempera-
◦
5
tures in the present work, respectively. The extinction curves were
performed in the opposite direction to investigate the catalytic
performance during cooling. Two ignition-extinction experiments
were performed to stabilize the catalysts. This was then followed
by a 40-hour thermal aging test in 4100 ppm methane by increas-
◦
◦
◦
ing reaction temperature to 550 C (60 C/min ramping rate) and
◦
then cooling the reactor to 375 C. The 550–375 C temperature
cycling was repeated 8 times (∼25 h); after this cycling, the cata-
◦
lyst was aged at 375 C for another 15 h. Finally, the catalytic test in
methane combustion was terminated by a third ignition-extinction
cycle to check the catalytic activities after 40 h thermal aging. Dur-
ing methane combustion, the temperature at each stage was held
for 30 min in ignition-extinction tests and 1 h in thermal aging
tests. The gas outlet from the reactor was analyzed online every
1
5 min using an Agilent HP-7890-A gas chromatograph equipped
with series TCD and FID. The reported curves in the results section
represent the initial 1st ignition-extinction, cycling and ignition-
extinction results after the cycling.
The crystal structures of the mono- and bimetallic nanoparticles
were further investigated by XRD (Fig. 2). Pd and Ru nanoparticles
Fig. 1. HRTEM images and SAED patterns of as-synthesized Pd(c)Ru(s) (a) and Ru2Pd1 (b) nanoparticles.

3
54
J. Shen et al. / Applied Catalysis A: General 502 (2015) 350–360
Fig. 3. Catalytic activities and product selectivities over mono- and bimetallic Ru-Pd
◦
catalysts in indan hydrogenolysis. Catalysts were pre-calcined at 200 C followed by
reduction in hydrogen at 375 C for polymer removal.
Fig. 2. XRD patterns of as-synthesized Ru, Pd, Pd(c)Ru(s) and Ru2Pd1 nanoparticles.
◦
display an fcc and hcp structure, respectively. Both bimetallic sam-
ples with the same Ru-to-Pd molar ratio of 2:1 reveal (1 1 1), (2 0 0),
to be active in many catalytic applications, such as hydrogenoly-
sis, methanation, oxidations, etc., [30]. However, ruthenium, with
its lower D-banD occupancy, is relatively inactive in hydrogena-
tions at low temperature and pressure as compared to Pd and
Pt, for example, in hydrogenations of 2-methyl-3-butyn-2-ol and
(
2 2 0) and (3 1 1) fcc peaks, which are consistent with the SAED pat-
terns (Fig. 1). This observation is similar to a previous study on the
structural change when alloying Pt with Ru [29]: Ru-Pt bimetallic
systems revealed fcc structures up to 75 mol% of Ru in the bimetal-
lic particles, as was shown by XRD and HRTEM [29]. The fcc (1 1 1)
lattice spacings of our bimetallic samples were 2.23 Å for both sam-
ples using Bragg’s law, higher than that of pure Ru (2.14 Å) and very
close to that of pure Pd (2.24 Å). The TEM, XRD and SAED observa-
tions are consistent with the proposed intrinsic bimetallic nature
of the synthesized nanoparticles.
1
,4-butynediol [31] or hydrogenation of o-chloronitrobenzene in
methanol at 303 K and 0.1 MPa [32], which is in line with our
observations. The Pd(c)Ru(s) catalyst also did not show any activity
confirming that there are no surface Pd atoms in the material and
Pd nanoparticles are covered by a Ru shell in the as-synthesized
nanoparticles.
For the Pd(c)-Ru(s) nanoparticle preparation, the palladium
seeds were formed first followed by ruthenium precursor addi-
tion to form a Ru shell. Indeed, the size change of the Pd seeds
from 2.0 ± 0.5 nm to 2.3 ± 0.7 nm after the Ru deposition indicates
the proposed structure is likely (no monometallic Ru nanoparti-
cles were observed, as shown above). The indicated sizes were
calculated based on length; the mass/volume average for the
monometallic Pd particles is 2.1 nm. 80% by mass of all Pd(core)-
Ru(shell) particles are larger than 2.6 nm diameter with an average
diameter (by mass) of 3.0 nm. Using crystal statistics for 2.1 nm Pd
core and 3.0 nm bimetallic particles [29], the predicted molar ratio
in the core–shell sample is 2.2:1 Ru to Pd, that is consistent with
the ratio used for synthesis and found by NAA in the final catalysts.
However, core–shell structure inversion may occur even during
catalyst synthesis due to the rearrangement of high-energy cor-
ner and edge atoms on ultra-small particles (<2 nm) [16]. To verify
whether Pd atoms are present on the surface of the as-synthesized
Pd(c)Ru(s) nanoparticles, a selective chemical probe surface reac-
tion was used. The three-phase liquid phase hydrogenation of
an allylic alcohol (2-methyl-3-buten-2-ol to 2-methylbutan-2-ol)
◦
3.2. Catalysis in a reductive atmosphere: hydrogenolysis at 350 C
The as-synthesized particles were deposited on ␥-Al O , and for
this particular reductive application, they were pretreated by calci-
2
3
◦
◦
nation at 200 C followed by reduction in hydrogen flow at 375 C.
This particular treatment for surface stabilizer removal resulted in
the maximized catalytic activity even in the presence of PVP resid-
uals on the surface, as we reported earlier [20]; higher calcination
temperatures caused sintering and/or undesired Pd/Ru rearrange-
ment on the nanoparticle surface. According to our previous TPR
analysis, Ru-Pd catalysts showed reduction peak at 80–85 C [20].
Thus, the in situ hydrogen reduction at 375 C is enough to provide
metallic surfaces for both Ru2Pd1 and Pd(c)Ru(s) catalysts, which
are the active sites for indan hydrogenolysis. The catalytic reac-
tion in a reductive (hydrogen) atmosphere was indan ring opening
(selective hydrogenolysis) carried out in a packed-bed reactor with
◦
◦
◦
gas feed at 350 C. The reaction serves as a model reaction for
fuel upgrading [33]. We already reported these results earlier and
details can be found in Ref. [20]. Here we provide a brief summary
with a focus on the nanoparticle surface structure only for a sake
of comparison with an oxidative application (following Section 3.3,
not reported previously).
Monometallic Ru is highly active in indan hydrogenolysis,
unlike palladium, but the latter allows selective ring opening to
2-ethyltoluene without successive dealkylation. Ruthenium favors
deeper hydrogenolysis to o-xylene (Fig. 3). When the two metals
are alloyed together either into an Ru2Pd1 or Pd(c)Ru(s) struc-
◦
was performed at 40 C, 0.45 MPa absolute pressure and 1200 rpm
stirring speed. As-prepared monometallic and bimetallic Ru-Pd cat-
alysts were studied in MBE hydrogenation without any further
thermal treatment to minimize structural rearrangements so that
the as-synthesized structures can be evaluated. Only initial rates
were evaluated to avoid in-situ structural evolution. The Pd catalyst
initial activity was 2.7 ± 0.2 molH2/(molPd s) [20], while Ru catalyst
had relatively negligible MBE hydrogenation activity. Ru is known

J. Shen et al. / Applied Catalysis A: General 502 (2015) 350–360
355
Fig. 4. DRIFT spectra of CO adsorbed on Ru, Pd and Pd(c)Ru(s) catalysts after
◦
reductive atmosphere treatment. Catalysts were pre-calcined at 200 C followed
◦
by reduction in hydrogen at 375 C for polymer removal. Reprinted from [20] with
permission from Paragon.
Fig. 6. Methane combustion over mono- and bimetallic Ru-Pd catalysts, 4100 ppm
methane, 1.1 barg pressure, 1.2 mg Pd and/or 3.1 ± 0.4 mg Ru: (a) thermal aging
◦
◦
at 375–550 C (Ru is not shown because of <10% conversion at 375 C); (b)
ignition–extinction curves after 40 h thermal aging. Error bars represent one stan-
dard deviation from average for ignition–extinction curves (no significant hysteresis
◦
was observed). Catalysts were calcined at 550 C for 16 h under static air.
ture, they show intermediate behavior between the two metals
(with one exception) but are distinctively different from each other.
The exception is that the activity of Pd(c)Ru(s) even surpasses the
activity of monometallic Ru, indicating that active ruthenium is
more accessible to reactants being dispersed on the Pd core, versus
monometallic Ru particles with lower Ru dispersion. The Ru2Pd1
sample shows Ru-like activity but Pd-like selectivity, suggesting
the close contact between the two metals in one nanostructure: if
Ru is not promoted by Pd on the nanoparticle surface, the Ru2Pd1
will not show suppressed selectivity to o-xylene and enhanced
selectivity to 2-ethyltoluene as compared to monometallic Ru. The
Fig. 5. Pd 3d5/2 XPS of the spent monometallic Pd catalyst (after the reaction from
Fig. 6).

3
56
J. Shen et al. / Applied Catalysis A: General 502 (2015) 350–360
◦
Fig. 7. TEM images of as-synthesized nanoparticles, calcined catalysts (550 C, 16 h) and spent catalysts (after the reactions from Fig. 6).
Pd(c)Ru(s) selectivity is closer to the monometallic Ru selectivity,
indicating the shell enrichment with Ru. Fig. 3 also presents the
results of earlier reported alloyed Ru10Pd1 catalyst (molar ratio
of Ru to Pd = 10) only to show how the catalytic properties of
Pd(c)Ru(s) (2:1 molar ratio) are close to the Ru-enriched nanoparti-
cles with both Ru and Pd present on the surface. Thus, the catalytic
results in indan hydrogenolysis indicate that the two Ru-Pd cat-
alysts with the same molar ratio but initially different structures
paramount for this particular application: to achieve high single
cleavage selectivity to 2-ethyltoluene (Fig. 3), shell enrichment
with Ru is undesirable and the alloy structure preparation is pre-
ferred.
◦
3
.3. Catalysis in an oxidative atmosphere: combustion at 550 C
Prior to the methane combustion reactions, all catalysts were
(
alloy Ru2Pd1 and core–shell) still behave differently in the reac-
◦
calcined at 550 C for 16 h, which is an accepted de-greening pro-
cedure for automotive catalysts [25]. Since methane combustion
was carried out in oxygen-rich environment, the calcined Ru-Pd
catalysts were packed in the reactor without further reduction
by hydrogen. Methane oxidation follows Mars-van Krevelen redox
mechanism on site pairs consisting of adjacent Pd and PdO species
tion in accordance with their nanoparticle surface enrichment, but
the core–shell sample does undergo some structural evolution. If
initially, in low-temperature MBE hydrogenation the core–shell
sample displays purely Ru behavior, then at high-temperature
hydrogenolysis, Pd behavior also manifests, i.e., some Pd migration
to the nanoparticle surface occurred. This behavior is expected from
thermodynamics: Pd has lower heat of vaporization (380 kJ/mol) as
compared to Ru (580 kJ/mol), thus, it tends to segregate to the sur-
face to lower the surface free energy. CO-DRIFTS of the core–shell
catalyst (Fig. 4) confirms such structural evolution. Prior to CO-
[
34,35]. Metallic Pd dissociatively adsorbs CH4 producing H and
CHx species; while PdO is responsible for oxidation [34,36]. Thus,
according to the redox mechanism, the surface state of catalysts
should be partially oxidized during methane oxidation. This is
confirmed by XPS analysis over a spent Pd catalyst as shown in
Fig. 5: the Pd 3d5/2 peaks could be assigned to PdO (336.5 eV), PdOx
◦
DRIFTS analysis, Ru-Pd catalysts were calcined at 200 C in air and
◦
followed by in situ hydrogen reduction at 375 C to simulate the
(337.3 eV) and Pd (335.4 eV), indicating the coexistence of metallic
pretreatment procedures before indan RO. The low wavenumber
adsorption bands corresponding to bridged adsorbed CO mode on
Pd and Pd oxides.
Fig. 6 compares the activities of monometallic and bimetallic
catalysts in methane combustion during initial ignition-extinction,
−
1
−1
Ru (1946 cm ) and Pd (2026 and 1995 cm ) are significantly dif-
ferent for the core–shell structure: the bridged adsorption on Ru
◦
low-high temperature cycling (375 and 550 C), followed by
−
1
atoms is suppressed (low intensity of 1951 cm
peak) with no
ignition-extinction curve measurements as described in the exper-
imental section. The monometallic Pd catalyst is the most active
methane oxidation catalyst [34,37]. The monometallic Ru displays
bridged adsorption on Pd atoms, but a high-intensity new peak at
−
1
1
980 cm can be most likely ascribed to the bridged adsorption of
CO on adjacent Pd and Ru atoms, which is in line with our hypoth-
esis of the core–shell structure evolution into a surface containing
both elements.
◦
the lowest activity, with less than 10% conversion at 375 C; initial
activity of the bimetallic catalysts in the first ignition-extinction
cycle is lower than for the pure Pd (Fig. 6 (a)), but it improves
under the reaction conditions and surpasses the Pd activity after
To conclude, the catalytic results and CO-DRIFTS indicate
structural rearrangement of parent Pd(c)Ru(s) structure via Pd seg-
regation to the surface under indan hydrogenolysis conditions.
However, there are still significant differences in behavior of the
Ru2Pd1 and Ru(c)Pd(s) sample in the reaction regardless of their
identical Pd-to-Ru molar ratio. The occurred structural rearrange-
ments thus did not result in the very same structure; the use
of initially applied different synthetic procedures (Scheme 1) is
2
5 h on stream (Fig. 6 (b) and (c)). Indeed, as seen for the cycling
◦
experiments (Fig. 6 (b)), 550 C temperature conditioning leads to
progressive deactivation of Pd, while the bimetallic catalysts do not
deactivate with cycling and even increase their activity. However,
as opposed to the previous case with a reductive atmosphere, both
Pd(c)Ru(s) and Ru2Pd1 catalysts display exactly the same behavior
for methane combustion during and after structural evolution.

J. Shen et al. / Applied Catalysis A: General 502 (2015) 350–360
357
As seen from the TEM images (Fig. 7) of the fresh, calcined
◦
(
550 C for 16 h), and spent monometallic Pd (after all the cycling
experiments) and Ru2Pd1 catalysts (after all the cycling exper-
iments as in Fig. 6), mono-Pd sintered to a large extent after
◦
exposure to 550 C, which is one of the known reasons for Pd cata-
lyst deactivation during methane combustion [34]. The bimetallic
catalyst possesses improved thermal stability, with no visible
agglomerates above 30 nm. The CO chemisorption showed the
CO/metal(s) molar ratios of 0.19, 0.15 and 0.33 for Pd, Ru and
Ru2Pd1 catalysts, respectively. Monometallic and bimetallic Ru-
Pd samples for CO chemisorption experiments were also calcined
◦
at 550 C for 16 h to simulate the pretreatment procedure before
methane combustion; this is then followed by in situ hydrogen
◦
reduction at 550 C for 1 h, as CO adsorption requires metallic sur-
faces of Pd and Ru. The higher thermal stability and surface area
of the bimetallic catalyst may partly explain the higher observed
methane combustion activity versus the monometallic Pd.
One of other feasible hypotheses of the bimetallic catalysts
activity improvement over the course of the reaction (Fig. 6) could
be volatilization of inactive Ru as oxides, leading to the progres-
sive increase in the Pd atom exposure. To check the hypothesis’s
validity, the Ru and Ru2Pd1 samples were analyzed by NAA after
◦
◦
two different calcination treatments: i) 550 C, 16 h, and ii) 550 C,
◦
1
6 h followed by 600 C, 72 h. Monometallic Ru content dropped
◦
from 0.28 ± 0.02 wt.% to 0.22 ± 0.01 wt.% after the 600 C treatment,
confirming the oxide volatility. However, when Ru was alloyed
with Pd in the Ru2Pd1 sample, the loading remained constant as
0
.17 ± 0.01 wt.%. This result is in line with the improved thermal
stability of the alloyed Ru-Pd catalysts (Fig. 7), and leads to a rejec-
tion of the hypothesis of increasing Pd exposure due to Ru oxides
volatilization. Monometallic Ru and bimetallic Ru2Pd1 catalysts
◦
treated at 600 C were used only to the study the possible Ru oxides
volatility, so they were not used in the catalytic tests.
Another reason for the bimetallic catalyst activity improvement
could be in situ nanoparticle reconstruction and Pd atom diffusion
to the surface driven by the minimization of surface free energy.
Indeed, the heat of vaporization of Pd (380 kJ/mol) is lower than
the one for Ru (580 kJ/mol), so the progressive nanoparticle shell
enrichment with Pd is expected during high-temperature appli-
cations. Such a hypothesis may also explain the same observed
activity of the Pd(c)Ru(s) and Ru2Pd1 bimetallic samples in the
methane combustion: regardless of the initial particle structures,
they both restructure to the most thermodynamically favourable
material. The variation in bimetallic activities with prolonged time
on stream could be related to the progressive change in surface
oxygen composition as Pd diffuses to the bimetallic surface. Dif-
ferent metals may display different combustion activities not only
because of their nature but also due to different metal-to-metal
oxide ratios. For example, PdO forms in low temperature range
Fig. 8. EXAFS spectra in k-space for monometallic Ru and Pd and bimetallic
Pd(c)Ru(s) and Ru2Pd1 nanoparticles in Pd K-edge (a) and Ru K-edge (b). Catalysts
◦
◦
were calcined at 550 C for 16 h followed by reduction in hydrogen at 400 C.
EXAFS of the high-temperature-treated Pd(c)Ru(s) and Ru2Pd1
samples was performed to verify their structures. The Ru-Pd cata-
lysts for EXAFS study were treated using the same procedure as in
◦
catalyst test that is pre-calcination at 550 C for 16 h under air; this
◦
was followed by H2 reduction at 400 C, which is confirmed to be
enough to reduce oxidized Ru-Pd nanoparticles to metallic forms
by our previous TPR analysis [20]. Prior to EXAFS analysis, sam-
ples were stored under ambient conditions for a week and then
reduced in hydrogen to avoid complications by the metal-oxygen
contributions when estimating coordination numbers. The reduc-
tive pre-treatment did not lead to the changes in methane oxidation
catalysis: both reduced and oxidized Ru2Pd1 samples showed
identical light-off temperature (temperature for 50% methane con-
◦
◦
(
300–l400 C), being stable up to 800 C, whereas, PtO decomposes
2
◦
at a much lower temperature (400 C) [37]. Since PtO2 is highly
unstable, significant amount of metallic Pt may still exist even after
calcination at 500 C in air [37], which explains the poor methane
◦
combustion activity of supported Pt catalyst under fuel-lean condi-
tions comparing to Pd metal. According to the Mars-van Krevelen
redox mechanism, an optimal surface oxidation state is required
for methane activation. In the case of bimetallic Ru-Pd system, Ru
◦
version) of 385 ± 10 C.
Fig. 8(a) and (b) show the Pd K-edge and Ru K-edge EXAFS spec-
tra in K-space for mono-Ru and Pd and bimetallic nanoparticles.
The amplitudes of the signal for Pd(c)Ru(s) and Ru2Pd1 nanopar-
ticles are significantly lower than those of mono-Pd nanoparticles
(Fig. 8 (a)), indicating lower coordination numbers around Pd in the
bimetallic particles [24]. The EXAFS oscillation patterns of bimetal-
lic nanoparticles at the Ru K-edge are similar in amplitude but differ
slightly in periodicity from the pure Ru sample (Fig. 8 (b)); this
could be ascribed to alloying between Pd and Ru in the bimetal-
lic structure. Similar changes in the EXAFS oscillation patterns as a
result of the formation of Au-Pd bonds have been reported by Liu
binds oxygen too strongly (the heat of formation for RuO is higher
2
than PdO), which can lead to a lower surface density of oxygen
vacancies for CH dissociative adsorption. Eventually, the active Pd
4
in bimetallic catalysts became more accessible being dispersed on
Ru core, which explains the increasing activities in both bimetallic
Ru-Pd catalysts. The Ru2Pd1 and Pd(c)Ru(s) catalysts should have
similar surface oxygen compositions during methane combustion,
because they both transformed into Ru(c)Pd(s) structure as was
shown by EXAFS analysis in the following.

3
58
J. Shen et al. / Applied Catalysis A: General 502 (2015) 350–360
◦
Fig. 9. EXAFS single-shell fits in r-space for monometallic Pd and Ru and bimetallic Pd(c)Ru(s) and Ru2Pd1 nanoparticles. Catalysts were calcined at 550 C for 16 h followed
◦
by reduction in hydrogen at 400 C.
et al. [38]. The negligible change in amplitude of the Ru edge in the
bimetallic samples, combined with the decreasing amplitude on
the Pd edge, is strong evidence of Ru(c)Pd(s) formation after calci-
nation/reduction; i.e., an inversion of the original structure in the
case of Pd(c)Ru(s) particles.
Table 1. The coordination numbers for the bulk fcc and hcp struc-
tures are 12. As compared to the reference foils, monometallic Pd
and Ru nanoparticles have lower coordination numbers of 9.7 ± 0.7
(N
) and 10.6 ± 1.6 (NRu-Ru), respectively. These are indications
Pd-Pd
of the existence of large portions of atoms on the surface of the
nanoparticles [38]. The higher coordination number seen for Ru
after the high temperature treatment agrees with CO chemisorp-
tion/TEM evidence that Ru tends to sinter to a greater degree than
Pd.
For the bimetallic nanoparticles, NPd-M and NRu-M correspond
to the total coordination numbers around Pd and Ru, respectively,
where M refers both to Ru and Pd. The Pd–Pd and Pd-Ru con-
tributions cannot be well differentiated because of their similar
atomic numbers. For core–shell structures, the total coordina-
tion number for the core metal should be high, normally close
to 12, whereas the total coordination number for the shell metal
Fig. 9 shows high quality EXAFS R-space data with single-
shell theoretical fits for monometallic Ru and Pd and bimetallic
Pd(c)Ru(s) and Ru2Pd1 nanoparticles. The Pd and Ru experimen-
tal EXAFS data of the bimetallic samples were fit by the theoretical
phase-shift and amplitudes generated by FEFF [23,24]. The struc-
tural parameters of the mono-and bimetallic Ru-Pd nanoparticles
generated using IFEFFIT software package are shown in Table 1.
The shoulders observed on the Pd edge at low R values are most
likely due to the presence of small amounts of surface Pd oxi-
dation [24]. The EXAFS analyses were performed on the samples
exposed to air (at room temperature) for about 1 week; Pd sur-
face oxidation was possible; this was also observed by XPS in our
previous work [28]. The most useful information on the local struc-
ture obtained from EXAFS fits is coordination number, presented in
would be much smaller [38]. For our catalysts, after calcination
◦
in air at 550 C for 16 h, and subsequent reduction, the N
Pd-M
for Pd(c)Ru(s) and Ru2Pd1 samples were found to be 7.5 ± 0.7

J. Shen et al. / Applied Catalysis A: General 502 (2015) 350–360
359
Table 1
EXAFS fitting parameters.
0
s2 × 10 (Å )
3
2
Catalyst
Shell
N
Rj(Å)
ꢂE (eV)
R-factor
Pd foil
Ru foil
Pd
Ru
Pd(c)Ru(s)
Pd-Pd
Ru-Ru
Pd-Pd
Ru-Ru
Pd-M
Ru-M
Pd-M
Ru-M
12
12
9.7 (0.7)
10.6 (1.6)
7.5 (0.7)
11.3 (1.4)
8.0 (1.0)
11.5 (1.3)
2.741 (0.002)
2.671 (0.002)
2.735 (0.001)
2.670 (0.006)
2.713 (0.002)
2.672 (0.002)
2.699 (0.004)
2.673 (0.002)
2.3 (0.5)
2.9 (1.2)
−5.0 (0.5)
−7.5 (1.2)
−9.1 (0.7)
3.7 (1.0)
5.7 (0.4)
4.2 (0.9)
6.1 (0.4)
4.0 (0.8)
6.5 (0.7)
4.6 (0.7)
9.0 (1.0)
4.9 (0.6)
0.005
0.016
0.009
0.023
0.017
0.016
0.046
0.015
Ru2Pd1
−9.2 (1.0)
3.4 (0.9)
and 8.0 ± 1.0, respectively, while the total coordination numbers
around Ru atoms, NRu-M, were higher (11.3 ± 1.4 for Pd(c)Ru(s) and
adsorbent-induced structural evolution into one type of struc-
ture. The results for methane oxidation showed that nanoparticle
structural evolution is kinetically limited even at 550 C and thus
bimetallic catalysts to be evaluated under prolonged times on
stream.
◦
1
1.5 ± 1.3 for Ru2Pd1). This shows that both structures transformed
into a final Ru(c)Pd(s) structure. Only small amounts of Pd oxi-
dation are seen in these samples, and no Ru oxidation, which is
further evidence of the migration of Pd atoms to nanoparticle shells.
Note that the monometallic catalyst coordination numbers differ
significantly from those of the bimetallic samples. TEM evidence
shows much more moderate sintering in bimetallic nanoparticles
compared to their mono-Pd or Ru counterparts; which confirms
that the bimetallic nanoparticles do not segregate into much larger
Ru and small Pd monometallic particles. The structural inversion
of the parent Pd(c)Ru(s) particle occurred most likely due to the
thermodynamics-induced Pd segregation to the surface because
of its lower heat of vaporization. The fact that the shell becomes
enriched with Pd is in line with the catalytic results: the bimetal-
lic catalysts display dramatically enhanced activity over time as
compared to monometallic Pd (Fig. 6), which is consistent with Pd
diffusion to the surface over time.
Acknowledgements
Financial support from NSERC – Strategic program grant STPGP
4
30108–12 is acknowledged. We thank Becquerel Laboratory
of Maxxam Analytics (Ontario, Canada) for NAA analysis, Dr.
Xuehai Tan (University of Alberta) for his help with high reso-
lution TEM and SAED analyses, and Shiraz Merali (University of
Alberta) for performing XRD analyses. EXAFS work described in
this paper was performed at the Canadian Light Source, which
is supported by the Canadian Foundation for Innovation, Natural
Sciences and Engineering Research Council of Canada, the Univer-
sity of Saskatchewan, the Government of Saskatchewan, Western
Economic Diversification Canada, the National Research Council
Canada, and the Canadian Institutes of Health Research.
Thus, both samples’ methane combustion catalytic results
(
Fig. 6) and EXAFS evidence show that irrespective of the prepa-
ration method (Scheme 1) and differences in parent bimetallic
nanostructures (but recall their constant bulk metal ratio), both Pd-
Ru systems transform into one structure with Pd-enriched shells.
The structural evolution was the most pronounced in the high-
temperature oxidative application versus the above-discussed
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.apcata.2015.06.
025
◦
indan hydrogenolysis at 350 C.
References
4
. Conclusions
[
[
1] B. Coq, F. Figueras, J. Mol. Catal. A. 173 (2001) 117–134.
2] M. Sankar, N. Dimitratos, P.J. Miedziak, P.P. Wells, C.J. Kiely, G.J. Hutchings,
Chem. Soc. Rev. 41 (2012) 8099–8139.
Pd-Ru bimetallic nanoparticles with the same bulk metal ratio
were synthesized using two different colloidal chemistry tech-
niques. One resulted in particles with mixed surface (alloy), and
another with Ru-shell particles. Several catalytic applications were
considered:
[
3] O.G. Ellert, M.V. Tsodikov, S.A. Nikolaev, V.M. Novotortsev, Russ. Chem. Rev.
83 (2014) 718–732.
[
4] A.K. Singh, Q. Xu, ChemCatChem 5 (2013) 652–676.
[5] D.M. Alonso, S.G. Wettstein, J.A. Dumesic, Chem. Soc. Rev. 41 (2012)
075–8098.
8
[
[
6] C. Cui, L. Gan, M. Heggen, S. Rudi, P. Strasser, Nat. Mater. 12 (2013) 765–771.
7] K.J.J. Mayrhofer, V. Juhart, K. Hartl, M. Hanzlik, M. Arenz, Angew. Chem. Int.
Ed. 48 (2009) 3529–3531.
8] B. Zhu, G. Thrimurthulu, L. Delannoy, C. Louis, C. Mottet, J. Creuze, B. Legrand,
H. Guesmi, J. Catal. 308 (2013) 272–281.
-
-
-
A low-temperature liquid-phase hydrogenation of an allylic alco-
hol showed that the presumably core–shell structure indeed
displays the properties of only the shell metal while the alloy sam-
ple displayed intermediate properties between the monometallic
counterparts.
In a high-temperature gas-phase hydrogenolysis, the core–shell
sample underwent some structural evolution with Pd segregation
to the surface, as seen from the catalytic behavior and CO-DRIFTS;
however, the two samples still exhibited distinctively different
behaviors.
[
[
9] H.B. Liu, U. Pal, J.A. Ascencio, J. Phys. Chem. C. 112 (2008) 19173–19177.
[10] R. Huang, Y.H. Wen, Z.Z. Zhu, S.G. Sun, J. Phys. Chem. C. 116 (2012) 8664–8671.
[
11] D.N. Tafen, J.B. Miller, N. Ö. Do gˇ an, J.P. Baltrus, P. Kondratyuk, Surf. Sci. 608
2013) 61–66.
12] J. Gu, Y.W. Zhang, F. Tao, Chem. Soc. Rev. 41 (2012) 8050–8065.
[13] N. Toshima, H. Yan, Y. Shiraishi, Metal Nanoclusters in Catalysis and Materials
Science: The Issue of Size Control, in: B. Corain, G. Schmid, N. Toshima (Eds.),
Elsevier B.V, Amesterdam, 2008, pp. 49–75.
14] A.U. Nilekar, S. Alayoglu, B. Eichhorn, M. Mavrikakis, J. Am. Chem. Soc. 132
(2010) 7418–7428.
[15] L. D’Souza, S. Sampath, Langmuir 16 (2000) 8510–8517.
(
[
[
A high-temperature combustion reaction revealed identical cat-
alytic properties of both samples, and EXAFS confirmed that both
parent nanoparticle types evolved into one structure with Pd-
enriched shells.
[
16] R.M. Anderson, L. Zhang, J.A. Loussaert, A.I. Frenkel, G. Henkelman, R.M.
Crooks, ACS Nano. 7 (2013) 9345–9353.
17] A.I. Frenkel, Chem. Soc. Rev. 41 (2012) 8163–8178.
[
[
18] S. Alayoglu, P. Zavalij, B. Eichhorn, Q. Wang, A.I. Frenkel, P. Chupas, ACS Nano.
3
(2009) 3127–3137.
The results show that the structure control in bimetallic
[
19] Y. Chen, K.Y. Liew, J. Li, Mater. Lett. 62 (2008) 1018–1021.
nanoparticle preparation is crucial for some applications while oth-
ers do not benefit because of thermodynamically and probably
[20] J. Shen, N. Semagina, ACS Catal. 4 (2014) 268–279.
[21] T. Teranishi, M. Miyake, Chem. Mater. 10 (1998) 594–600.

3
60
J. Shen et al. / Applied Catalysis A: General 502 (2015) 350–360
[
[
[
[
22] E.A. Stern, M. Newville, B. Ravel, Y. Yacoby, D. Haskel, Physica B 208–209
1995) 117–120.
23] P. Dash, T. Bond, C. Fowler, W. Hou, N. Coombs, R.W.J. Scott, J. Phys. Chem. C.
13 (2009) 12719–12730.
24] T. Balcha, J.R. Strobl, C. Fowler, P. Dash, R.W.J. Scott, ACS Catal. 1 (2011)
25–436.
25] R. Abbasi, L. Wu, S.E. Wanke, R.E. Hayes, Chem. Eng. Res. Des. 90 (2012)
930–1942.
26] J. Shen, R.E. Hayes, X. Wu, N. Semagina, ACS Catal. 5 (2015) 2916–2920.
27] M. Liu, J. Zhang, J. Liu, W.W. Yu, J. Catal. 278 (2011) 1–7.
28] J. Shen, X. Yin, D. Karpuzov, N. Semagina, Catal. Sci. Technol. 3 (2013) 208–221.
29] C. Pan, F. Dassenoy, M.J. Casanove, K. Philippot, C. Amiens, P. Lecante, A.
Mosset, B. Chaudret, J. Phys. Chem. B 103 (1999) 10098–10101.
[30] J.K. Norskov, T. Bligaard, J. Rossmeisl, C.H. Christensen, Nat. Chem. 1 (2009)
37–46.
[31] P. Rylander, Catalytic Hydrogenation Over Platinum Metals, Academic Press
Inc, New York, 1967.
[32] M. Liu, W. Yu, H. Liu, J. Zheng, J. Colloid Interface Sci. 214 (1999) 231–237.
[33] U. Nylen, L. Sassu, S. Melis, S. Järås, M. Boutonnet, Appl. Catal. A 299 (2006)
1–13.
[34] G. Centi, J. Mol. Catal. A 173 (2001) 287–312.
[35] K. Fujimoto, F.H. Ribeiro, M. Avalos-Borja, E. Iglesia, J. Catal. 179 (1998)
431–442.
(
1
4
1
[
[
[
[
[36] J.N. Carstens, S.C. Su, A.T. Bell, J. Catal. 176 (1998) 136–142.
[37] P. Gelin, M. Primet, Appl. Catal. B 39 (2002) 1–37.
[38] F. Liu, D. Wechsler, P. Zhang, Chem. Phys. Lett. 461 (2008) 254–259.