Preparation, characterization, and kinetic evaluation of dendrimer-derived bimetallic Pt-Ru/SiO2 catalysts

Article abstract of DOI:10.1016/j.jcat.2009.11.021

A series of silica-supported Pt, Ru, and Pt-Ru catalysts has been synthesized using dendrimer-metal nanocomposite (DMN) precursors prepared by both co- and sequential complexation with metal salts. The catalysts have been characterized by several techniqu

Full text of DOI:10.1016/j.jcat.2009.11.021

Journal of Catalysis 269 (2010) 376–387
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Preparation, characterization, and kinetic evaluation of dendrimer-derived
bimetallic Pt–Ru/SiO2 catalysts
*
Dongxia Liu, Yaritza M. López-De Jesús, John R. Monnier, Christopher T. Williams
Department of Chemical Engineering, Swearingen Engineering Center, University of South Carolina, Columbia, SC 29208, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 July 2009
Revised 19 November 2009
Accepted 23 November 2009
Available online 31 December 2009
A series of silica-supported Pt, Ru, and Pt–Ru catalysts has been synthesized using dendrimer–metal
nanocomposite (DMN) precursors prepared by both co- and sequential complexation with metal salts.
The catalysts have been characterized by several techniques, including electron microscopy, tempera-
ture-programmed titration of adsorbed oxygen, and X-ray diffraction. Liquid-phase selective hydrogena-
tion of 3,4-epoxy-1-butene (EpB) was used as a probe reaction to evaluate their catalytic performance.
The bimetallic catalyst prepared by the co-complexation method exhibits a superior catalytic activity
compared to the sequential one, and is much more active than a conventional catalyst prepared by incip-
ient wetness. The activity enhancement is attributed to a bifunctional performance of the PtRu alloy sites
created, based on a strong correlation between turnover frequencies, and both the alloy compositions and
metal surface site distributions. In addition, the co-complexation catalyst is selective toward crotonalde-
hyde, suggesting that this reaction pathway is favored on the PtRu sites.
Keywords:
Platinum
Ruthenium
Bimetallic
Dendrimer
Nanoparticle
3,4-Epoxy-1-butene
Hydrogenation
Ó 2009 Elsevier Inc. All rights reserved.
1. Introduction
proach takes advantage of the fact that transition metal ions (e.g.,
Pt2+, Cu2+, and Pd2+) can coordinate with the interior amine groups
Over the years, bimetallic or multi-metallic catalysts have re-
placed many monometallic catalysts in industrial applications
due to the superior activities and/or selectivities that can be
achieved. Such enhanced performance can be contributed to a
combination of geometric, electronic, and/or bifunctional effects
[1–3]. Preparation of bimetallic or multi-metallic catalysts usually
involves co-impregnation or successive impregnation methods.
However, these traditional preparation methods often provide
unsatisfactory control over the dispersion of the metal atoms,
resulting in homogeneity problems such as metal segregation, sin-
tering, and surface enrichment of the ‘‘inactive” metal.
One promising method of preparing bi- or multi-metallic cata-
lysts with uniform composition is using dendrimers as templates
and stabilizers. Dendrimers [4,5] are mono-disperse, hyper-
branched polymers that emanate from a central core with repeti-
tive branching units. While possessing a very dense exterior, they
contain hollow pockets that can be ideal for use as nano-scale con-
tainers. The use of dendrimers as templates/stabilizers for the syn-
thesis of nanoparticles is a relatively new but active field. One of
the more successful applications along these lines has been the
synthesis of metal nanoparticles using poly-(amidoamine) (PA-
MAM) dendrimers. Originally pioneered by Crooks et al. [6], the ap-
of the dendrimer. Such ions may then be reduced to form dendri-
mer–metal nanocomposites (DMNs) that are stable for extended
periods of time and can exhibit interesting catalytic properties
[7]. The dendrimer can exert control over size and (in the case of
multiple metal ions) composition of the resulting nanoparticles
or clusters, which can allow for tuning of catalytic properties.
There are now many investigations exploring the use of DMNs
as precursors to synthesize supported catalysts [8–16], although
still relatively few examples for bimetallic Pt-based systems. For
example, Lang et al. [17] have examined the possibility of prepar-
ing dendrimer-templated, supported Pt–Au bimetallic nanoparti-
cles using the displacement reaction method. In this approach,
Cu0 nanoparticles complexed with dendrimer amine groups are ex-
changed with a mixture of Pt(II) and Au(III) in solution, thus form-
ing zero valent Pt–Au nanoparticles. Hoover et al. [18] synthesized
PAMAM dendrimer-templated Pt–Cu bimetallic nanoparticles sup-
ported on alumina that were tested as heterogeneous catalysts for
CO oxidation and toluene hydrogenation reactions. Xie and co-
workers [13] reported the synthesis of bimetallic Pt–Cu catalysts
supported on silica using dendrimer-templating methods. They ob-
served a synergistic catalytic effect on activity for hydrodechlori-
nation of 1,2-dichlorethane, which was attributed to bifunctional
bimetallic Pt–Cu sites.
Liquid-phase selective hydrogenation of the multi-functional
olefin 3,4-epoxy-1-butene (EpB) is an interesting probe reaction
* Corresponding author. Fax: +1 (803) 777 8265.
E-mail address: willia84@cec.sc.edu (C.T. Williams).
0021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.11.021

D. Liu et al. / Journal of Catalysis 269 (2010) 376–387
377
to examine the kinetic performance of bimetallic catalysts. EpB has
both a C@C bond and an epoxy group, and in the presence of metal
catalysts and hydrogen can proceed via isomerization, hydrogena-
tion, and/or hydrogenolysis reactions to form a range of reaction
products. The reaction scheme for selective hydrogenation/hydrog-
enolysis of EpB is shown in Fig. 1. Several transition metals have
been tested as catalysts for EpB hydrogenation and/or hydrogenol-
ysis reactions [19–26]. Pt [19,23] and Rh [22,25] favor C@C bond
hydrogenation, with the former being less selective. Pd [19,21] is
especially active toward the rupture of the epoxy group to form
minimum, HPLC grade) was purchased from VWR. Oxygen (UHP),
helium (UHP), and hydrogen (UHP) were supplied by Airstar.
2.2. Catalyst preparation
The synthesis of dendrimer–metal nanocomposites was
adapted from the literature [6] and has been published before.
[12] A proper amount of 3.96 ꢀ 10ꢁ3 M RuCl3ꢂxH2O precursor solu-
tion was added under the N2 purging to G4OH dendrimer solution
with the concentration of 4.2 ꢀ 10ꢁ4 M to reach a molar ratio of
Ru3+ to G4OH of 40:1. The mixed solution was stirred for 3 days
at room temperature with N2 flowing to protect the Ru3+ ions from
oxidation by dissolved O2 and allow the complexation of Ru3+ ions
with the functional groups in G4OH dendrimer. Freshly made
0.1 M NaBH4 aqueous solution was then added to the dendri-
mer–Ru nanocomposites, yielding a mass ratio of NaBH4 to metal
of 1:1, to attempt to reduce the incorporated Ru3+ ions to Ru nano-
particles. The reduction procedure lasted for at least 30 min with
mild stirring until there were no more H2 bubbles generated. The
supported catalyst was made by standard wet impregnation of
the reduced dendrimer–Ru nanocomposite onto SiO2 to a nominal
Ru loading of 0.52 wt%. The extra water was removed under ambi-
ent conditions by evaporation of the stirring slurry for 3 days. The
same procedure was used to make dendrimer–Pt monometallic
nanocomposites, with the corresponding supported catalyst hav-
ing a nominal Pt loading of 1 wt%. The only difference was that
the concentration of H2PtCl6ꢂ6H2O aqueous precursor solution
was 2.05 ꢀ 10ꢁ3 M.
stable p-allylic intermediates. The deoxygenation reaction to form
1,3-butadiene is favored over Cu catalysts [20]. So far, the catalytic
performance of Ru has not been studied for EpB hydrogenation;
likewise, no bimetallic systems other than Pt–Ag [24] and Cu–Pd
[26] have been explored.
In this study, a family of dendrimer-derived Pt–Ru catalysts has
been prepared to illustrate dendrimer-templating effects on bime-
tallic structure and correlate physical and/or chemical properties of
the bimetallic catalysts with their kinetic behavior for EpB hydro-
genation. It was found that dendrimer-derived bimetallic catalysts
exhibited different metal nanoparticle sizes and distributions
when compared with conventionally prepared samples. In addi-
tion, Pt–Ru alloy sites are maximized using the dendrimer templat-
ing co-complexation method, resulting in an enhanced activity and
unusual selectivity to crotonaldehyde.
2. Materials and methods
Two synthesis routes were used to produce bimetallic Pt–Ru
DMNs, as shown in Fig. 2. The co-complexation method involved
adding an equal molar amount of the two metal precursor solu-
tions simultaneously to dendrimer solution. The resulting solution
was allowed to complex for 3 days under N2 purging and subse-
quently reduced with NaBH4. The sequential method involved first
mixing one metal precursor solution with the dendrimer solution,
followed by reduction after complexation was completed. Then,
the second metal precursor solution was added, whereupon com-
plexation with the second metal occurred, followed once again
by reduction with NaBH4. The first metal was Pt, while the second
was Ru. For both methods, the target ratio of each type of metal
atom to dendrimer was 20:1, thus making a total metal atom
loading of 40 atoms per dendrimer, as for the monometallics. The
dendrimer–Pt–Ru bimetallic nanocomposites made by both
2.1. Materials
Fourth generation hydroxyl-terminated poly(amido)amine (PA-
MAM) dendrimer (G4OH) (10 wt% in methanol solution) was ob-
tained from Aldrich. Prior to use, an aqueous solution was made
by evaporating methanol with flowing N2 and diluting the residue
to 4.2 ꢀ 10ꢁ4 M with deionized water. RuCl3ꢂxH2O (99.98 % Ru, Al-
drich), H2PtCl6ꢂ6H2O (99.9%-Pt, Strem Chemicals), and NaBH4
(granules, 99.995%, Aldrich) were used as received. Deionized
water (18 M
X cm Milli-Q) was used to prepare all the aqueous
solutions. Silica support (specific surface area = 67 m2/g, pore vol-
ume = 1.3 mL/g) was supplied by BASF Catalysts LLC. Crotonalde-
hyde (99.5%, Aldrich) and 3,4-epoxy-1-butene (EpB) (Eastman
Chemical Company) were used as received. Hexane (99.5% purity
O
H
2
n-butyraldehyde
O
crotonaldehyde
isomerization
H
2
H
2
isomerization
OH
crotyl alcohol
H
2
isomerization
H
2
O
H
2
OH
3-buten-1-ol
OH
n-butanol
3,4-epoxy-1-butene
H
2
O
butylene oxide
Fig. 1. Reaction scheme of selective hydrogenation/hydrogenolysis of 3,4-epoxy-1-butene.

378
D. Liu et al. / Journal of Catalysis 269 (2010) 376–387
EDS analysis, the incident probe beam was focused on a single
Sequential complexation
metallic particle and the X-ray fluorescence generated by this spe-
cific particle was detected. Quantitative analyses were based on
the characteristic X-ray fluorescence lines of Pt and Ru.
NaBH4
PtCl
n RuCl3
m
H2
6
2.4. X-ray diffraction
G4OHPtm(Ru2+
NaBH4
)
n
G4OHPtm
G4OH
n RuCl3
The crystallographic structures of powder catalysts were ob-
tained by X-ray diffraction (XRD) on a desktop X-ray diffractometer
m
PtCl
H2
6
(Rigaku MiniFlex II), using a Cu Ka-monochromatized radiation
source (k = 1.54045 Å) operating at 30 kV and 15 mA. The Bragg
angular region (2h) between 25° and 60° was explored with a slow
step size of 0.05°/min due to the extremely low metal loading
(ꢃ0.5 wt % of Pt and less than 0.2 wt % of Ru for all the three bimetal-
lic catalysts). The peaks were analyzed by JADE 8.5 software.
Co-complexation
NaBH4
G4OH(Pt2+)m(Ru2+
)
G4OHPtmRun
n
2.5. O2–H2. titration studies
Fig. 2. Schematic of Pt–Ru DMN precursor synthetic routes.
Temperature-programmed O2–H2 titration studies were con-
ducted using a gas flow system equipped with a quadruple mass
spectrometer as the analytical detector (Leybold Inficon, Model
type TSPTT300). All the samples were reduced in-situ in flowing
H2 (50 cc/min) at 300 °C for 2 h followed by He (50 cc/min) purging
at the same temperature for another 2 h to remove any residual H2.
The catalyst was then cooled to room temperature in flowing He. A
temperature-programmed oxidation (TPO) experiment was first
conducted in flowing 1% O2 in He at a ramp rate of 5 °C/min up
to 300 °C to dissociatively adsorb O2 onto exposed metal particle
surfaces. The sample was then cooled to room temperature in
the same gas flow, followed by Ar purging for 30 min to remove
the residual oxygen in the system. H2 titration measurements were
performed at room temperature as follows: pulses of 10% H2/Ar
(0.516 mL STP) were dosed over the catalysts until several succes-
sive mass spectral H2 peaks showed the same peak height. Since
water generated by this step will re-adsorb on the support surface
at 25 °C, an extra ‘‘water removal” stage was used to heat the sam-
ple in flowing pure Ar with a ramp of 5 °C/min up to 300 °C. The
‘‘dry” catalysts were then cooled in flowing Ar to room temperature
for temperature-programmed reduction which was executed as
follows: 1% H2/Ar was flowed through the sample at 50 cc/min
while the sample was heated to 300 °C at a constant ramp of
5 °C/min. The H2O+ signal (m/e = 18) was monitored during this
procedure.
co-complexation and sequential methods were impregnated onto
silica as described above, with the nominal loading of 0.5 wt% Pt
and 0.26 wt% Ru.
Conventional Pt and Ru catalysts were made by incipient wet-
ness method. A proper amount of 3.94 ꢀ 10ꢁ2 M of RuCl3ꢂxH2O pre-
cursor solution was added dropwise to SiO2 and the resulting
slurry was dried in an oven at 100 °C for 2 h before use. A conven-
tional Pt catalyst was made with exactly the same protocol. In the
case of conventional Pt–Ru bimetallic catalyst, equal volumes of
7.88 ꢀ 10ꢁ2 M of RuCl3ꢂxH2O and H2PtCl6ꢂ6H2O precursor solutions
were mixed together and then added to the SiO2. The nominal me-
tal loadings of all the conventional catalysts were the same as their
DMN-derived analogues.
All catalysts studied were activated under flowing H2 treatment
at 300 °C after a slow temperature ramp of 2 °C/min for 2 h fol-
lowed by cooling to room temperature. This activation protocol
was adapted for consistency from a previous study [12] which
showed that G4OH dendrimer template could be sufficiently
decomposed on an c-Al2O3 support to expose the metal nanoparti-
cles. An O2–H2 treatment was purposely avoided due to the severe
sintering of Ru particles under O2 at high temperatures [27–31].
2.3. Transmission electron microscopy and energy dispersive X-ray
spectroscopy
In addition to the above temperature-programmed O2–H2 titra-
tion measurements, the total amount of oxygen adsorbed on the
catalysts was confirmed by using a Micromeritics 2920 dynamic
chemisorption system. In this measurement, the catalyst was oxi-
dized in 1% O2/He up to 300 °C at a ramp of 5 °C/min (as above) and
cooled to 250 °C in the same gas flow. After purging with pure Ar
for 30 min to remove the residual O2, H2 doses were pulsed over
the catalyst to titrate the oxygen precovered on its surface until
several successive H2 peaks of equal area were observed.
High resolution transmission electron microscopy (HRTEM) was
conducted using a Philips CM 120 microscope (University of Poi-
tiers, Poitiers, France) operating at 120 kV. Supported Pt–Ru/SiO2
catalysts were finely crushed and dispersed in methanol (ultrason-
ically) before subsequent deposition and ambient drying onto a
copper grid coated with a carbon. Histograms of particle size distri-
bution were obtained by measuring at least 100 randomly selected
particles from at least 6 different micrographs for any sample ana-
lyzed. The volume-surface mean diameter was calculated from the
following equation:
2.6. Temperature-programmed desorption
,
N
N
X
X
D ¼
NiD3
NiD2
p;i
p;i
Temperature-programmed desorption (TPD) of crotonaldehyde
was performed using an Altamira AMI-200 system at Oak Ridge
National Laboratory as follows. After in-situ reduction in flowing
H2 at 300 °C for 2 h as described before, a 5% crotonaldehyde in
He gas mixture was passed over the catalyst surface at room tem-
perature. Adsorption of crotonaldehyde was considered complete
when the mass intensities for species of m/e = 39, 41, 42, 69, and
70, indicative of crotonaldehyde, were constant. A He gas flow of
100 mL/min was used for 60 min to purge the residual crotonalde-
hyde from the system. The temperature of the sample was then
i
i
where Ni is the number of particles with a diameter Dp,i
.
Energy dispersive X-ray spectroscopy (EDS) analysis was uti-
lized to measure the elemental composition of individual-sup-
ported bimetallic Pt–Ru nanoparticles with a JEOL 2100F 200 kV
FEG-STEM/TEM equipped with a CEOS Cs corrector on the illumina-
tion system. The geometrical aberrations were measured and con-
trolled to provide less than a
electron wave over the probe-defining aperture of 15.5 mrad. In
p/4-phase shift of the incoming

D. Liu et al. / Journal of Catalysis 269 (2010) 376–387
379
ramped at 20 °C/min from room temperature to 500 °C in the same
He flow, followed by holding at this temperature for 30 min.
Desorption products were analyzed by mass spectrometer for the
same five species (m/e = 39, 41, 42, 69, and 70).
flush the residual liquid remaining in the sample dip tube back into
the reactor after sampling to ensure that no liquids were left in the
metal tubing between sample collections at different time inter-
vals. Freshly reduced catalyst (typically 300 mg) and 80 mL of hex-
ane (99.5% purity minimum, HPLC grade, VWR) were placed in the
reactor. The reactor was flushed three times with pure hydrogen at
room temperature before being pressurized to a constant H2 pres-
sure of 300 psig that was maintained during the reaction period.
After reaching the reaction temperature of 80 °C, the reaction
was initiated by starting the agitation (rotation speed = 1000 rpm)
2.7. Catalytic evaluation
Selective hydrogenation of 3,4-epoxy-1-butene (EpB) was per-
formed in a 100 mL, gold-coated stainless steel semi-batch auto-
clave reactor (Autoclave Engineers). A 4-port valve was used to
0.35
Conventional Pt
d = 4.6 nm
0.30
0.25
0.20
0.15
0.10
0.05
0.00
a
0
2
4
6
8
10
Diameter (nm)
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
Dendrimer Pt
d = 5.2 nm
b
0
2
4
6
8
10
Diameter (nm)
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
Conventional Ru
d = 4.0 nm
c
0
2
4
6
8
10
Diameter (nm)
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
Dendrimer Ru
d = 2.7 nm
d
0
2
4
6
8
10
Diameter (nm)
Fig. 3. HRTEM images and histograms of (a) conventional Pt/SiO2, (b) dendrimer-derived Pt/ SiO2, (c) conventional Ru/SiO2, and (d) dendrimer-derived Ru/SiO2.

380
D. Liu et al. / Journal of Catalysis 269 (2010) 376–387
and injecting 3,4-epoxy-1-butene (EpB) from a high-pressure li-
quid pump (Spectra-Physics, Isochrom LC pump). The liquid reac-
tion mixture was sampled at different reaction times and
analyzed by gas chromatography (HP 5890) using a 30 m long
Hewlett Packer DB-wax capillary column. The reaction was contin-
ued until there were no significant changes in the concentrations
of the reactor contents. A mole balance was used since all reaction
products either were isomers of EpB or had molecular weights that
only varied by a maximum of 4 amu (70 g/mol for n-butanol vs.
66 g/mol for EpB). Closure of the mole balance was always within
10%.
corresponding particle size distributions are plotted in the histo-
grams next to the HRTEM images. The conventional catalyst has
an asymmetric distribution with majority (85%) of the particles be-
tween 1 and 4.5 nm and the remaining 15% of particles spread out
between 4.5 and 8 nm. In contrast, the dendrimer-derived Pt
exhibits a more symmetric distribution, with ca. 50% of particles
between 1 and 4.5 nm. Conventional Ru (Fig. 3c) has ꢃ30% larger
volume-surface mean diameter (4.0 0.8 nm) and an obviously
broader particle size distribution than dendrimer-derived Ru
(Fig. 3d), with an average particle size of 2.7 0.6 nm.
There are two explanations for the discrepancy in the role of
dendrimer in influencing Pt and Ru particle size and size distribu-
tion. One possibility is the much higher mobility of Pt atoms during
the elevated temperature H2 thermal treatment on SiO2, which
causes severe sintering of Pt clusters. Indeed, it has been demon-
strated that sintering of metal nanoparticles takes place during
the deposition of dendrimer–metal nanocomposites (DMNs) onto
supports and in the subsequent elevated temperature H2 thermal
treatment for dendrimer shell removal. In particular, sintering of
3. Results and discussion
3.1. Particle size, distribution, and composition
HRTEM micrographs of monometallic Pt samples prepared by
conventional incipient wetness method and dendrimer-templating
method are shown in Fig. 3a and b, respectively. Analysis of these
images yields a volume-surface mean diameter of 4.6 1.5 nm for
conventional Pt and 5.2 1.3 nm for dendrimer-derived Pt. Their
c-Al2O3-supported Pt nanoclusters during the dendrimer thermal
removal step is substantially larger than that observed for Ru par-
ticles on the same support [8,15]. Thus, highly mobile Pt atoms/
0.35
cocomplexation PtRu
0.30
d = 4.2 nm
0.25
0.20
0.15
0.10
0.05
0.00
a
b
0
2
4
6
8
10
Diameter (nm)
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
sequential complexation PtRu
d = 3.4 nm
0
2
4
6
8
10
Diameter (nm)
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
conventional PtRu
d = 4.7 nm
c
0
2
4
6
8
10
Diameter (nm)
Fig. 4. HRTEM images and histograms of (a) dendrimer-derived PtRu/SiO2 by co-complexation method, (b) dendrimer-derived PtRu/SiO2 by sequential method, and (c) PtRu/
SiO2 by conventional incipient wetness method.

D. Liu et al. / Journal of Catalysis 269 (2010) 376–387
381
clusters may escape from the dendrimer interior and agglomerate
Pt (200)
Pt (111)
together to form larger particles. This in turn results in particle size
similar to its conventional counterpart. In contrast, less mobile
dendrimer-derived Ru particles are strongly bonded to the support,
yielding smaller particle size and narrower distribution. This
differing sintering behavior for Pt and Ru has implications for the
bimetallic catalyst synthesis. For example, Alerasool and co-
workers [32] also reported that bimetallic Pt–Ru/SiO2 catalysts
synthesized by co-impregnation of H2PtCl6ꢂ6H2O and RuCl3ꢂ3H2O
precursors were Pt surface-enriched. They attributed this to the
Pt precursor being more mobile on the silica surface than to the
Ru precursor in flowing H2 reduction treatment at 400 °C.
Conv PtRu
Seq-complex
PtRu
Fig. 4 shows HRTEM images and histograms of the three bime-
tallic Pt–Ru catalysts synthesized by (a) co-complexation, (b)
sequential, and (c) conventional incipient wetness methods. It is
apparent that the conventional Pt–Ru catalyst has a larger vol-
ume-surface mean diameter and broader size distribution than
the two dendrimer-derived Pt–Ru catalysts. This implies that
G4OH PAMAM dendrimer exerts a measure of control over particle
formation. Energy dispersive X-ray spectroscopy (EDS) was utilized
to obtain compositions of individual particles for the three bime-
tallic Pt–Ru catalysts. At least 10 particles with various sizes were
randomly selected from each catalyst. All particles examined in the
three catalysts are bimetallic, while monometallic particles con-
sisting of either Pt or Ru were not found. The average percentages
of Pt present in the catalysts are similar to 60% average elemental
composition obtained via atomic absorption spectroscopy (AAS),
albeit with some significant scatter.
Co-complex
PtRu
Dend Pt
35
40
45
50
2 θ (degree)
Fig. 5. X-ray diffractograms of silica supported (a) dendrimer-derived monome-
tallic Pt catalyst and Pt–Ru bimetallic catalysts prepared by (b) co-complexation, (c)
sequential complexation, and (d) conventional incipient wetness methods.
3.2. Nanoparticle bulk crystallinity
X-ray diffraction (XRD) patterns of the dendrimer-derived
monometallic Pt catalyst and the three bimetallic catalysts are
shown in Fig. 5. The conventional Pt catalyst has an identical
XRD pattern with its dendrimer-derived partner, and is therefore
not shown. Monometallic Ru catalysts, both conventional and den-
drimer-derived, did not exhibit any feature due to very low weight
loading (ꢃ0.5 wt %). It is well known that Pt has a face-centered cu-
bic (f.c.c.) structure. Therefore, the two prominent peaks in dendri-
mer-derived monometallic Pt at 2h = 39.78° and 46.08° can be
assigned to Pt (1 1 1) and Pt (2 0 0) phases, respectively. Comparing
the curves, a slight shift to higher Bragg angles and a broadening of
the peak width were observed for the Pt (1 1 1) peak in all three
bimetallic Pt–Ru catalysts, especially in the case of the co-com-
plexation sample. This indicates that PtRu alloys have been formed
by the substitution of Pt atoms in the lattice with smaller Ru
atoms, which causes a decrease in the dimension of the Pt unit cell
[33–37].
Pt (1 1 1) and (2 0 0) peak positions extracted from a four
parameter least squares fit to a Gaussian function with linear back-
ground subtraction are shown at the columns in Table 2. Based on
Bragg’s law, lattice constants of the Pt unit cell for the three bime-
tallic PtRu alloys can be calculated using both diffraction peaks and
the results are shown in the same table. According to the Vegard’s
law for a PtRu alloy, the value of the Pt lattice parameter must de-
crease when Ru content increases, and a linear relation exists (at
constant temperature) between the Pt crystal lattice parameter
and the Ru concentration in the bimetallic phase [38]. Vegard’s
law can be applied to unsupported alloys readily. However, it is dif-
ficult to evaluate the alloying constants for supported alloys [39].
One reason is that the lattice constant for supported Pt is shorter
than that for unsupported Pt due to Pt–support interactions. To
estimate the Ru atomic fraction value for supported alloys, an
assumption was made that the dependence of the lattice parame-
ter on Ru content for supported alloy is similar to that of the
unsupported alloy.
Based on the assumption above, Vegard’s law can be expressed
by
aPtRu ¼ ð1 ꢁ xRuÞ ꢂ aPt þ xRu ꢂ aRu
ð1Þ
where aPtRu, aPt and aRu are lattice constants for PtRu alloy, pure Pt
and pure Ru. xRu is Ru atomic fraction in the alloy. The equation
above can be further simplified as
aPtRu ¼ aPt ꢁ xRu ꢂ ðaPt ꢁ aRu
Þ
ð2Þ
where ðaPt ꢁ aRuÞ = 0.124 Å is a constant obtained from unsupported
PtRu alloy [35,39]. Such Ru atomic fraction estimates for the three
bimetallic samples are listed at right-most column in Table 2. The
co-complexation bimetallic sample has the highest alloying degree
with 27% Ru content in its PtRu alloy phase. In contrast, the conven-
tional sample has the lowest alloy degree with only 9% Ru, with the
sequential falling in between at 12%.
3.3. Surface site quantification
The type and number of surface sites on bimetallic nanoparti-
cles were probed by hydrogen titration of adsorbed oxygen. The
adsorbed oxygen can be titrated from Pt at low temperatures
(<100 °C), while significantly higher temperature is required for
Ru [40,41]. Therefore, the catalysts were examined by room tem-
perature and temperature-programmed O2–H2 titration to deter-
mine the temperature profile of the reduction reaction, and thus
the distribution of various types of surface sites.
Oxygen precovered on conventional Pt/SiO2 was titrated by
pulse dosing 10% H2/Ar at room temperature. The titration proce-
dure was considered to be completed when two successive H2
peaks exhibited identical heights. Fig. 6a shows the mass spec-
trometer Hþ2 signal (m/e = 2) with respect to time. After four injec-
tions, there was no more H2 consumption, indicating that the room

382
D. Liu et al. / Journal of Catalysis 269 (2010) 376–387
3.50E-009
3.00E-009
2.50E-009
2.00E-009
1.50E-009
1.00E-009
5.00E-010
0.00E+000
5.00E-011
b
a
4.00E-011
3.00E-011
2.00E-011
1.00E-011
0.00E+000
50
100
150
200
250
300
0
200
400
600
800
Time (sec)
Temperature (ºC)
3.50E-009
3.00E-009
2.50E-009
2.00E-009
1.50E-009
1.00E-009
5.00E-010
0.00E+000
5.00E-011
4.00E-011
3.00E-011
2.00E-011
1.00E-011
0.00E+000
c
d
50
100
150
200
250
300
0
100
200
Time (sec)
300
400
Temperature (ºC)
Fig. 6. Temperature-programmed H2 titration measurement on conventional Pt/SiO2 at (a) room temperature by pulse dosing of H2 and (b) as a function of temperature
during continuous flowing of H2; on conventional Ru/SiO2 at (c) room temperature by pulse dosing H2, and (d) as a function of temperature during continuous flowing of H2.
temperature titration was complete. However, only part of the ad-
sorbed oxygen atoms were titrated at room temperature and a
higher temperature is needed to remove them entirely. Fig. 6b
shows the H2O+ (m/e = 18) signal during a ramped heating of con-
ventional Pt/SiO2 in H2 from 25 to 300 °C, revealing a broad water
desorption peak at 80 °C. However, the total amount of oxygen ti-
trated from this step was much smaller than that titrated at room
temperature, as expected. Similar temperature-dependent H2 titra-
tion studies were accomplished on conventional Ru/SiO2. As shown
in Fig. 6c, H2 consumption is not observed indicating that adsorbed
oxygen on Ru is unable to be titrated at room temperature. Tem-
perature-programmed H2 titration of the same conventional Ru/
SiO2 from 25 to 300 °C in 1% H2/Ar is shown in Fig. 6d. Water evo-
lution was detected from 100 to 250 °C, with a maximum at 165 °C.
Therefore, adsorbed oxygen atoms on Ru sites are completely inac-
tive under 100 °C, indicating more stable adsorption on the oxo-
philic Ru (and perhaps some presence of more stable Ru oxide
species).
Temperature-programmed titration studies on the other cata-
lysts were also performed, and the results for all the catalysts are
plotted from 25 to 300 °C in Fig. 7. Oxygen atoms precovered on
the two monometallic Pt/SiO2 catalysts were mostly removed at
room temperature by pulse dosing 10% H2, and therefore relatively
flat lines are observed in Fig. 7. In contrast, Ru interacts with oxy-
gen atoms more strongly compared to Pt, resulting in large peaks
in the high temperature region. Water peak positions of the three
bimetallic catalysts sit in between the two monometallics, which
indicates that there are both Pt- and Ru-like sites on the bimetallic
surfaces.
sites, as shown by the dashed lines in Fig. 7. The center tempera-
tures of each type of site as well as their relative amounts are re-
ported in Table S1 (see Supplementary material). For instance,
analysis of the water desorption curves for the conventional Pt/
SiO2 catalyst reveals two peaks centered at 70 and 105 °C that
likely represent two types of Pt sites. Similarly, three types of Ru
surface sites centered at 132, 158, and 187 °C can be classified from
the conventional Ru/SiO2 catalyst. All these surface sites can be cat-
egorized into two regions based on their center temperatures: a
lower region from 25 to 110 °C and a higher region above 110 °C.
The former have been designated as ‘‘Pt-like” and the latter as
‘‘Ru-like”, and the areas under these curves have been used to esti-
mate (see Supplementary material) the number of Pt and Ru sur-
face atoms in each catalyst and the surface compositions
(Table 3). For comparison, bulk compositions obtained from atomic
absorption spectroscopy (AAS) are also listed in Table 3. It is impor-
tant to mention that surface compositions obtained as the result of
temperature-programmed O2 and H2 titration studies described
above would represent an average of all particles in the catalyst.
Nevertheless, the EDS results suggest that the large majority of
the particles are indeed bimetallic in nature.
It is not surprising to observe Pt surface enrichment in bulk
PtRu alloys due to the relatively lower enthalpy of vaporization
of Pt (510 kJ/mol) compared to Ru (595 kJ/mol) and negligible heat
of mixing of the two metals [42]. Gasteiger et al. [43] reported a
nearly constant 87% Pt surface composition for bulk Pt content
varying from 20% to 60%. In supported PtRu alloy, the surface
enrichment also depends on the metal–support interaction in addi-
tion to the enthalpy of vaporization of the two metals. Since SiO2 is
considerably acidic, when co-depositing two metal precursor solu-
tions (RuCl3 and H2PtCl6) onto the support without dendrimer
Water evolution curves for all the catalysts were fitted with sev-
eral Gaussian functions to distinguish different types of surface

D. Liu et al. / Journal of Catalysis 269 (2010) 376–387
383
Table 1
Metal particle size measured by HRTEM and O2 + H2 titration methods.
Conventional
Ru
Catalysts
Volume surface mean
Average particle
size from O2 + H2
titration (nm)
diametera from HRTEM (nm)
Conventional Pt
Conventional Ru
4.6 1.5
4.0 0.8
4.7 1.2
5.2 1.3
2.7 0.6
4.2 1.1
3.4 0.8
4.4
3.3
–
21.0
3.3
–
Dendrimer-derived
Ru
Conventional PtRu
Dendrimer-derived Pt
Dendrimer-derived Ru
Co-complexation PtRu
Seq. complexation PtRu
–
Conventional
PtRu
a
As described in the text.
3.4. Catalytic evaluation
Sequential complexation
PtRu
Before evaluating this family of catalysts for hydrogenation of
EpB, blank experiments were performed on the silica support to
estimate its influence on the reaction. Concentrations of all the
products increased very slowly, while the concentration of EpB de-
creased slowly, indicating only marginal activity for the silica sup-
port. This low activity is most likely due to the weak acid sites of
the silica that catalyze EpB isomerization reactions. However, over
the time scales used to evaluate the supported catalysts, the contri-
bution of SiO2 to the product formation rates was found to be neg-
ligible. The mole balance of the reaction is essentially constant
during the entire reaction period, indicating that all formed prod-
ucts were detected. In order to minimize internal mass transfer
limitations, all the catalysts were ground to fine powder before
use. Five stirring speeds of 100, 300, 700, 1000, and 1830 rpm were
examined for external mass transfer limitation study on a test Pt/
SiO2 catalyst. The reaction rates were essentially independent of
stirring rates over the entire range of stirring speeds. As a result,
1000 rpm was selected as the stirring rate for all catalytic runs. Fi-
nally, it has also been shown that EpB adsorbs strongly on most
metals through multiple chemical bonds with the surface
[19,23,47]. This strong adsorption has resulted in deactivation dur-
ing gas-phase reactions where EpB is either a reactant or a product
[24,26,48]. However, batch experiments involving multiple injec-
tions of EpB over the same catalyst showed no significant deterio-
ration in catalytic activity (see Fig. S1 in Supplementary material),
suggesting that this is likely not a problem in the present study.
Batch kinetic data were obtained for all catalysts during hydro-
genation of EpB. As an example, Fig. 8 shows the concentration pro-
file of the reactant and all the products plotted as a function of time
for (a) conventional Pt/SiO2 and (b) PtRu/SiO2 by co-complexation
method. In both plots, the concentration of EpB decreased rapidly,
and was completely consumed after a certain period of time. The
concentrations of crotyl alcohol, 3-buten-1-ol, and crotonaldehyde
passed through maxima indicating that they were reaction inter-
mediates. The reaction end products are butylene oxide (BO), n-
butanol, and n-butyraldehyde, whose concentrations monotoni-
cally increase and then stabilize with increasing reaction time.
The initial reaction rates of EpB were determined from such con-
centration-time data by taking the slope of the first two data points
at t = 0 and 5 min. Calculation of EpB turn over frequencies (TOFs)
was based on the total metal active surface sites (Pt-like + Ru-like)
as listed in Table 3. Dividing the initial reaction rate of each product
by the initial reaction rate of EpB yielded initial instantaneous
selectivities. Such selectivities were calculated individually for all
the six products, butylene oxide (BO), crotonaldehyde, 3-buten-1-
ol, n-butanol, n-butyraldehyde, and crotyl alcohol.
Co-complexation
PtRu
Conventional
Pt
Dendrimer-derived
Pt
50
100
150
200
250
300
Temperature [C]
Fig. 7. Temperature-programmed H2 titration over different catalysts.
template (as for conventional PtRu/SiO2), it exchanges protons for
cations (Ru3+) but not anions (PtCl26ꢁ) to form ligand bonds. There-
fore, the Pt precursor is more mobile than the Ru precursor on the
SiO2 surface during H2 reduction, and the particles that are surface-
enriched with Pt are formed. Such a Pt-enriched surface for con-
ventional PtRu bimetallic catalysts is in excellent agreement with
that in the reported data for similar catalysts [32,44–46]. In con-
trast, the surface compositions of the dendrimer-derived sequen-
tial complexation catalyst are almost identical to its bulk
composition. Finally, dendrimer-derived co-complexation sample
has a Ru-enriched surface with the surface composition of about
50–50% Pt–Ru active sites. This is consistent with the presence of
PtRu alloy as indicated by the XRD data, wherein the sample with
the higher Ru atomic fraction in its PtRu alloy phase has more sur-
face Ru active sites presented on its surface.
It is also interesting to compare the average particle sizes of
monometallic Pt and Ru catalysts calculated based on metal dis-
persions with those calculated from HRTEM data (Table 1). From
this comparison, it is clear that the results acquired from the two
methods are close for conventional monometallic catalysts. How-
ever, in the case of the dendrimer-derived Pt, the particle size esti-
mated from the titration method is much larger than that
determined from HRTEM. This discrepancy is attributed to residual
dendrimer fragments that cover a part of the metal surface sites,
thus leading to a larger calculated average particle size. Given that
the optimal treatment to remove PAMAM–OH from Pt–DMNs in-
volves a high temperature oxidation step [9,16], the single H2
treatment here is likely not enough for full removal. In contrast,
the H2 treatment is optimized for Ru–DMNs, which explains the
very good agreement for the Ru samples.
3.5. Catalytic activity
Table 4 lists the TOFs measured for the entire family of cata-
lysts. Both conventional and dendrimer-derived monometallic Ru

384
D. Liu et al. / Journal of Catalysis 269 (2010) 376–387
250
200
150
100
50
800
400
Mole balance
1200
1000
800
600
400
200
0
Mole balance
a
350
EpB
b
BO
600
400
200
0
300
250
200
150
100
50
a
EpB
n-butanol
Crotyl alcohol
Crotyl alcohol
BO
n-butanol
n-butyraldehyde
3-buten-1-ol
n-butyraldehyde
Croton aldehyde
3-buten-1-ol
Croton aldehyde
0
0
0
20
40
60
80
100
120
0
20
40
60
80
100
120
Time (min)
Time (min)
Fig. 8. Concentration profile obtained during the EpB hydrogenation reaction on (a) conventional Pt/SiO2 and (b) dendrimer-derived co-complexation PtRu/SiO2 catalyst.
Table 2
Pt lattice constants and estimated Ru atomic fractions in PtRu alloys.
Catalysts
Pt (1 1 1) peak position (°)
Pt (2 0 0) peak position (°)
Lattice constant (Å)
Ru atomic fraction in PtRu alloy
Dendrimer-derived Pt
Conventional PtRu
Co-complexation PtRu
Seq. complexation PtRu
39.78
39.93
40.20
39.98
46.08
46.31
46.56
46.36
3.929 0.010
3.912 0.007
3.890 0.011
3.908 0.008
–
0.09
0.27
0.12
1400
1200
1000
800
600
400
200
0
readily at room temperature, supplying enough adsorbed H atoms
for rapid hydrogenation to occur.
Co-
Complex.
The EpB TOF on the dendrimer-derived Pt catalyst is about
twice that for conventional Pt, which has a majority (85%) of its
particles between 1 and 4.5 nm and the remaining 15% spread
out between 4.5 and 8 nm. In contrast, the dendrimer-derived Pt
has only 50% of its particles between 1 and 4.5 nm and the other
half larger than 4.5 nm (as shown in Fig. 3). A likely explanation
for this behavior is that EpB hydrogenation is a structure sensitive
reaction that requires a large number of highly coordinated crys-
talline sites (e.g., exposed Pt (1 1 1) planes) to facilitate the adsorp-
tion of both EpB and H atoms in close proximity to each other.
Since the dendrimer-derived Pt catalyst has a higher percentage
of large size particles compared to its conventional counterpart,
more of these crustal planes are exposed, thus resulting in higher
TOFs for EpB conversion.
Clearly, the dendrimer-derived PtRu bimetallic catalyst pre-
pared by the co-complexation method is the most active of the
three bimetallic catalysts examined. The catalyst prepared by
sequential complexation is a factor of two less active, and conven-
tionally prepared one is yet another factor of two less active. The
higher activity for the co-complexation catalyst can result from
(a) more exposed Pt monometallic sites that therefore yield higher
activity per combined surface metal site (i.e., Pt + Ru), (b) bimetal-
lic PtRu alloy sites have been created that have higher intrinsic
activity, or (c) a combination of the two hypothesis. It is easy to
rule out the first hypothesis by checking the exposed metal sites
Seq.
Complex.
Ru content in alloy phase
Ru surface composition
Con-
ventional
0
10
20
30
40
50
60
atomic % Ru
Fig. 9. EpB TOFs as a function of atomic Ru percentage in the PtRu alloys (circles)
and on the surface (squares) of co-complexation, sequential complexation, and
conventional PtRu bimetallic catalysts. The dashed lines are provided to illustrate
the trends.
catalysts are much less active than their Pt counterparts. This is
most likely linked to the relative inability of Ru to dissociatively
adsorb H2 at reaction temperatures of 80 °C. Consequently, higher
temperatures (>373 K) are typically used for reactions or processes
involving H2 dissociation. This is particularly the case for small par-
ticle sizes (<4 nm) such as those prepared in this study
[28,41,49,50]. In contrast, H2 dissociatively adsorbs on Pt surfaces

D. Liu et al. / Journal of Catalysis 269 (2010) 376–387
385
Table 3
Metal surface sites, surface composition, actual metal loading, and bulk composition.
Catalysts
Site amount (atoms/g cat) ꢀ10ꢁ18
Surface composition
Metal weight loadinga
Bulk compositiona
Pt-like sites
Ru-like sites
Total sites
Pt (%)
Ru (%)
Pt (%)
Ru (%)
Pt (%)
Ru (%)
Conventional Pt
Conventional Ru
7.5
–
6.6
2.1
–
–
7.5
12
100
–
76
100
–
–
0.95
–
0.53
1.19
–
–
100
–
63
100
–
–
12
2.2
–
9.9
3.2
1.9
100
24
–
100
49
33
0.5
0.16
–
0.42
0.19
0.18
100
37
–
100
40
38
Conventional PtRu
Dendrimer-derived Pt
Dendrimer-derived Ru
Co-complexation PtRu
Seq. complexation PtRu
8.8
2.1
9.9
6.6
5.9
3.4
3.9
51
67
0.56
0.57
60
62
a
Obtained by elemental analysis.
Table 4
EpB TOFs for all catalysts.
BO
crotonaldehyde
3buten1ol
Catalyst
EpB TOFs (minꢁ1
)
Conventional Pt
Dendrimer-derived Pt
Conventional Ru
Dendrimer-derived Ru
Conventional PtRu
Co-complexation PtRu
Sequential complexation PtRu
762 49
1486 70
nbutanol
nbutyraldehyde
crotylalcohol
21
29
0
0
179 11
1162 38
675 172
100
80
60
40
20
0
reported in Table 3. The co-complexation sample has neither the
highest number of Pt sites, nor the total metal exposed sites on
its surface. However, the second hypothesis appears to be sup-
ported by the XRD analysis (cf. Table 2). As shown in Fig. 9 (circles),
the initial EpB TOFs of the bimetallic catalysts exhibit a strong
dependence on the atomic Ru composition in their PtRu alloy
phases. The most active co-complexation sample has the highest
alloying degree with 33% Ru, while the lowest activity conven-
tional PtRu alloy contains only 12% Ru.
This ‘‘bimetallic site” supposition is further strengthened by
consideration of the relative distribution of Pt-like and Ru-like sur-
face sites in each catalyst (cf. Table 3). A plot of EpB TOFs vs. the Ru
surface composition is also shown in Fig. 9 (squares). It is clear that
there is also a strong relationship between reactivity and the atom-
ic surface composition. This trend suggests the existence of bime-
tallic surface PtRu bimetallic sites that activate EpB conversion.
Electronic effects between Pt and Ru are known to be small due
to their similar electronegativities [51]. Since Ru sites are essen-
tially inactive relative to Pt, Ru dilution of Pt ensembles would
be expected to depress the activity. However, more Ru atoms are
exposed on the surface in co-complexation case, which has the
highest EpB TOF. The enhanced activity of EpB conversion is there-
fore more likely the result of a bifunctional effect. Ru is oxophilic
[15,51], which will facilitate the adsorption of the epoxy group of
EpB molecule. The neighboring Pt site then could more readily
interact with the C@C bond of adsorbed EpB and other intermedi-
ates. Thus, the existence of such bimetallic sites, as is especially the
case with the co-complexation catalyst, increases the reactivity of
EpB.
Fig. 10. Initial instantaneous selectivities to all products during EpB hydrogenation
over various catalysts.
3.6. Selectivity
hydrogenation catalysts [52], significant selectivity to products other
than BO was observed. Furthermore, BO was formed by a parallel
pathway and is a stable end product, suggesting that a C@C bond
is required for adsorption of EpB-derived intermediates on Pt sur-
faces. The result is consistent with the results of Bartok et al. [19]
and Schaal et al. [24] , who have reported that EpB can adsorb on
Pt at both the C@C bond and the epoxide oxygen. Bartok reported
that BO was the main product at gas-phase reaction conditions of
The initial instantaneous selectivities to all the six products,
which are butylene oxide (BO) (red)1, crotonaldehyde (green), 3-bu-
ten-1-ol (blue), n-butanol (cyan), n-butyraldehyde (magenta), and
crotyl alcohol (yellow) for all catalysts are shown in the bar plots
of Fig. 10. For the monometallic Pt catalysts, which are active C@C
1
For interpretation of color in Fig. 10, the reader is referred to the web version of
this article.
301 K, with PH = 20 kPa. Initial instantaneous selectivity to crotonal-
2

386
D. Liu et al. / Journal of Catalysis 269 (2010) 376–387
dehyde was extremely low (2–3%), indicating either that crotonalde-
hyde was extremely reactive or that hydrogenation to the other
products proceeded through the 3-buten-1-ol route.
crotonaldehyde was performed at Oak Ridge National Laboratory’s
Center for Nanophase Materials Sciences with assistance from
Dr. Zili Wu, and was sponsored by the Scientific User Facilities
Division, Office of Basic Energy Sciences, US Department of Energy
and finally, two undergraduate students Michelle Casper and
Paul Wilkinson for assistance with the catalytic hydrogenation
experiments.
In general, the selectivities for the bimetallic catalysts are some-
what similar, although the PtRu/SiO2 catalyst produced by co-com-
plexation has a much higher initial selectivity to crotonaldehyde.
This suggests that isomerization of adsorbed EpB to form crotonal-
dehyde is favored, and that subsequent hydrogenation to other
products preferentially occurs through crotonaldehyde, as opposed
to 3-butene-1-ol. This particular catalyst does have a strong affin-
ity for crotonaldehyde adsorption, as shown by the results of cro-
tonaldehyde TPD studies (see Supplementary material, Fig. S2
and Table S2). In these studies, the co-complexation catalyst had
by far the highest uptake of crotonaldehyde as well as the stron-
gest interaction based on the temperatures of the desorption
peaks. Thus, the bimetallic sites produced by the co-complexation
method enhance the adsorption of crotonaldehyde, and thus facil-
itate the reaction of EpB by this isomerization route.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2009.11.021.
References
[1] J.H. Sinfelt, Acc. Chem. Res. 10 (1977) 15.
[2] J.H. Sinflet, Bimetallic Catalysts: Discoveries, Concepts, and Applications, John
Wiley & Sons, Inc., 1983.
[3] C.T. Campbell, Ann. Rev. Phys. Chem. 41 (1990) 775.
[4] D.A. Tomalia, A.M. Naylor, W.A.I. Goddard, Angew. Chem., Int. Ed. 29 (1990)
138.
4. Conclusions
[5] J.M. Frechet, Science 263 (1994) 1710.
[6] R.M. Crooks, B.I. Lemon III, L. Sun, L.K. Yeung, M. Zhao, Top. Curr. Chem. 212
(2001) 81.
[7] R.M. Crooks, M. Zhao, L. Sun, V. Chechik, L.K. Yeung, Acc. Chem. Res. 34 (2001)
181.
[8] O.S. Alexeev, A. Siani, G. Lafaye, C.T. Williams, H.J. Ploehn, M.D. Amiridis, J.
Phys. Chem. B 110 (2006) 24903.
[9] D.S. Deutsch, A. Siani, P.T. Fanson, H. Hirata, S. Matsumoto, C.T. Williams, M.D.
Amiridis, J. Phys. Chem. C 111 (2007) 4246.
[10] D.S. Deutsch, C.T. Williams, M.D. Amiridis, in: J. Regalbuto, (Ed.), Catalyst
Preparation, CRC Press LLC, Boca Raton, Fal, 2007, p. 209.
[11] G. Lafaye, A. Siani, P. Marecot, M.D. Amiridis, C.T. Williams, J. Phys. Chem. B
110 (2006) 7725.
[12] G. Lafaye, C.T. Williams, M.D. Amiridis, Catal. Lett. 96 (2004) 43.
[13] H. Xie, J.Y. Howe, V. Schwartz, J.R. Monnier, C.T. Williams, H.J. Ploehn, J. Catal.
259 (2008) 111.
[14] D. Liu, J. Gao, C.J. Murphy, C.T. Williams, J. Phys. Chem. B 108 (2004) 12911.
[15] G. Lafaye, C.T. Williams, O.S. Alexeev, M.D. Amiridis, Manuscript in draft.
[16] D.S. Deutsch, G. Lafaye, D. Liu, B. Chandler, C.T. Williams, M.D. Amiridis, Catal.
Lett. 97 (2004) 139.
[17] H. Lang, S. Maldonado, K.J. Stevenson, B.D. Chandler, J. Am. Chem. Soc. 126
(2004) 12949.
In conclusion, a series of silica-supported monometallic (Pt, Ru)
and bimetallic (PtRu) catalysts have been synthesized using both
dendrimer–metal nanocomposites (DMNs) and metal salt precur-
sors. They have been characterized by HRTEM, EDS, XRD, and
O2–H2 titration measurements. Histograms obtained from electron
microscopy reveal that the dendrimer-mediated synthesis yields
smaller particle sizes and narrower particle size distributions com-
pared with the conventional incipient wetness impregnation for
the monometallic Ru and Pt–Ru. In contrast, the Pt shows a larger
particle size distribution, consistent with the previous findings.
EDS measurements indicate that most (if not all) of the particles
are bimetallic in the bimetallic catalysts. While XRD patterns sug-
gest that PtRu alloy phase has been formed in all the three bimetal-
lic catalysts, analysis suggests that the co-complexation sample
has the highest Ru content (33%) in its alloy phase. In contrast, con-
ventional PtRu has the least Ru content (12%) in its alloy phase. In
concert with these findings, O2–H2 titration shows that the co-
complexation catalyst has a relatively Ru-enriched surface, while
conventional PtRu has a Pt-enriched surface.
[18] N.N. Hoover, B.J. Auten, B.D. Chandler, J. Phys. Chem. B 110 (2006) 8606.
[19] M. Bartok, A. Fasi, F. Notheisz, J. Catal. 175 (1998) 40.
[20] M. Bartók, A. Fási, F. Notheisz, J. Mol. Catal. A: Chem. 135 (1998) 307.
[21] I.S. Cho, B. Lee, H. Alper, Tetrahedron Lett. 36 (1995) 6009.
[22] H. Fujitsu, E. Matsumura, S. Shirahama, K. Takeshita, I. Mochida, J. Chem. Soc.
Perkin 1 (1982) 855.
Batch kinetics studies of liquid-phase hydrogenation of 3,4-
epoxy-1-butene revealed notable differences in the catalytic per-
formance across this family of catalysts. The two dendrimer-de-
rived PtRu bimetallic samples (especially the co-complexation
sample) are intrinsically more active than their conventional coun-
terpart. This difference is attributed to a bifunctional performance
of the PtRu alloy sites created, and the TOF values are found to cor-
relate strongly with both the alloy composition and the metal sur-
face site distribution. The crotonaldehyde reaction pathway is
favored on the co-complexation catalyst, in contrast to the other
catalysts that favor the 3-buten-1-ol route. Crotonaldehyde TPD re-
sults suggest that this difference is attributed to the relative affin-
ity of these catalysts toward this intermediate. Taken together,
these results indicate that the DMN methodology provides a signif-
icant measure of control over bimetallic catalyst properties for the
PtRu bimetallic system.
[23] A.S. Loh, S.W. Davis, J.W. Medlin, J. Am. Chem. Soc. 130 (2008) 5507.
[24] M.T. Schaal, A.C. Pickerell, C.T. Williams, J.R. Monnier, J. Catal. 254 (2008) 131.
[25] J.R. Monnier, B.D. Roberts, D.M. Hitch, US Patent 6,180 559, Eastman Chemical
Company, USA, 2001.
[26] M.T. Schaal, A.Y. Metcalf, J.H. Montoya, J.P. Wilkinson, C.C. Stork, C.T. Williams,
J.R. Monnier, Catal. Today 123 (2007) 142.
[27] G.C. Bond, J.C. Slaa, J. Mol. Catal. 89 (1994) 221.
[28] G.C. Bond, J.C. Slaa, J. Mol. Catal. A: Chem. 96 (1995) 163.
[29] C. Elmasides, D.I. Kondarides, W. Grunert, X.E. Verykios, J. Phys. Chem. B 103
(1999) 5227.
[30] V. Mazzieri, F. coloma-Pascual, A. Arcoya, P.C. L’Argentiere, N.C. Figoli, Appl.
Surf. Sci. 210 (2003) 222.
[31] J. Wellenbuscher, M. Muhler, W. Mahdi, U. Sauerlandt, J. Schutze, G. Ertl, R.
Schlogl, Catal. Lett. 25 (1994) 61.
[32] S. Alerasool, R.D. Gonzalez, J. Catal. 124 (1990) 204.
[33] A.S. Arico, P.L. Antonucci, E. Modica, V. Baglio, H. Kim, V. Antonucci,
Electrochim. Acta 47 (2002) 3723.
[34] C.T. Hsieh, Y.W. Chou, W.Y. Chen, J. Alloys Compd. 466 (2008) 233.
[35] J. Prabhuram, T.S. Zhao, Z.X. Liang, R. Chen, Electrochim. Acta 52 (2007) 2649.
[36] V. Radmilovic, H.A. Gasteiger, P.N. Ross Jr., J. Catal. 154 (1995) 98.
[37] M.-S. Loffler, H. Natter, R. Hempelmann, K. Wippermann, Electrochim. Acta 48
(2003) 3047.
Acknowledgments
[38] W. Vogel, P. Britz, H. Bonnemann, J. Rothe, J. Hormes, J. Phys. Chem. B 101
(1997) 11029.
[39] E. Antolini, F. Cardellini, J. Alloys Compd. 315 (2001) 118.
[40] G. Blanchard, H. Charcosset, Stud. Surf. Sci. Catal. 4 (1980) 515.
[41] H. Kubicka, React. Kinet. Catal. Lett. 5 (1976) 223.
[42] A.R. Miedema, Philips Tech. Rev. 36 (1976) 217.
[43] H.A.R. Gasteiger, P.N. Ross, E.J. Cairns, Surf. Sci. 293 (1993) 67.
[44] H. Miura, R.D. Gonzalez, J. Catal. 74 (1982) 216.
[45] H. Miura, T. Suzuki, Y. Ushikubo, K. Sugiyama, T. Matsuda, R.D. Gonzalez, J.
Catal. 85 (1984) 331.
This work was supported by a grant from the National Science
Foundation NIRT award (NSF-0103135). The authors would like to
thank Dr. Attilio Siani for assistance with the catalyst synthesis,
Dr. Snigdhamayee Praharaj for performing the XRD measurements,
and Dr. Gwendoline Lafaye for arranging the HRTEM imaging at
the University of Poitiers. Temperature-programmed desorption of

D. Liu et al. / Journal of Catalysis 269 (2010) 376–387
387
[46] H. Miura, H. Taguchi, K. Sugiyama, T. Matsuda, R.D. Gonzalez, J. Catal. 124
[50] D.O. Uner, M. Pruski, T.S. King, J. Catal. 156 (1995) 60.
(1990) 194.
[51] A.K. Shukla, A.S. Arico, K.M. El-Khatib, Appl. Surf. Sci. 137 (1999) 20.
[47] J.W. Medlin, M.A. Barteau, J.M. Vohs, J. Mol. Catal. A: Chem. 163 (2000) 129.
[48] J. Monnier, Stud. Surf. Sci. Catal. 110 (1997) 135.
[49] K. Lu, B.J. Tatarchuk, J. Catal. 106 (1987) 166.
[52] S. Nishimura, Handbook of Heterogeneous Catalytic Hydrogenation for
Organic Synthesis, John Wiley & Sons, Inc., 2001.