Characterization and catalytic-hydrogenation behavior of SiO 2-embedded nanoscopic Pd, Au, and Pd-Au alloy colloids

Article abstract of DOI:10.1002/chem.200500971

Colloids embedded in a silica sol-gel matrix were prepared by using fully alloyed Pd-Au colloids, and pure Pd and Au colloids stabilized with tetraalkylammonium bromide following a modified sol-gel procedure with tetrahydrofuran (THF) as the solvent. Tetraethoxysilicate (TEOS) was used as the precursor for the silica support. The molar composition of the sol was TEOS/THF/H₂O/HCl = 1:3.5:4:0.05 for the bimetallic Pd-Au and TEOS/THF/H₂O/HCl = 1:4.5:4:0.02 for Pd and Au monometallic systems. After refluxing. the colloid was added as a 4.5 wt % solution in THF for Pd-Au. 10.2 wt % solution in THF for Pd and 8.4 wt % solution in THF for Au at room temperature. The gelation was carried out with vigorous stirring (4 days) under an Ar atmosphere. Following these procedures, bimetallic Pd-Au-SiCK catalysts with 0.6 and 1 wt % metal, and monometallic Pd- and Au-SiO₂ catalysts with 1 wt% metal were prepared. These materials were further treated following four different routes: 1) by simple drying, 2) in which the dried catalysts were calcined in air at 723 K and then reduced at the same temperature, 3) in which they were directly reduced in hydrogen at 723 K, and 4) in which the surfactant was extracted using an ethanol-heptane azeotropic mixture. The catalysts were characterized by nitrogen adsorption-desorption isotherms at 77 K, H₂ chemisorption measurements, solid-state ¹H, ¹³C ²⁹SiCP/MAS-NMR spectroscopy, powder X-ray diffraction (XRD), small angle X-ray scattering (SAXS), X-ray photoelectron spectroscopy (XPS), transmission electron microscopy (TEM), and ¹⁹⁷Au M?ssbauer spectroscopy. The physical characterization by a combination of these techniques has shown that the size and the structural characteristics of the Pd-Au colloid precursor are preserved when embedded in an SiO₂ matrix. Catalytic tests were carried out in selective hydrogenation of 3-hexyn-1-ol, cinnamaldehyde, and styrene. These data showed evidence that alloying Pd with Au in bimetallic colloids leads to enhanced activity and most importantly to improved selectivity. Also, the combination of the two metals resulted in catalysts that were very stable against poisoning, as was evidenced for the hydrogenation of styrene in the presence of thiophene.

Full text of DOI:10.1002/chem.200500971

FULL PAPER
DOI: 10.1002/chem.200500971
Characterization and Catalytic-Hydrogenation Behavior of SiO -Embedded
2
Nanoscopic Pd, Au, and Pd–Au Alloy Colloids
[
a]
[b]
[c]
[d]
Vasile I. Pârvulescu,* Viorica Pârvulescu, Uwe Endruschat, George Filoti,
[
d]
[e]
[f]
Friedrich E. Wagner, Christian Kübel, and Ryan Richards
Dedicated to Professor Helmut Bonnemann on the occasion of his 65th birthday
Abstract: Colloids embedded in a silica
sol–gel matrix were prepared by using
fully alloyed Pd–Au colloids, and pure
Pd and Au colloids stabilized with tet-
raalkylammonium bromide following a
modified sol–gel procedure with tetra-
hydrofuran (THF) as the solvent. Tet-
raethoxysilicate (TEOS) was used as
the precursor for the silica support.
The molar composition of the sol was
with 0.6 and 1 wt% metal, and mono-
metallic Pd– and Au–SiO2 catalysts
with 1 wt% metal were prepared.
These materials were further treated
following four different routes: 1) by
simple drying, 2) in which the dried
catalysts were calcined in air at 723 K
and then reduced at the same tempera-
ture, 3) in which they were directly re-
duced in hydrogen at 723 K, and 4) in
which the surfactant was extracted
using an ethanol–heptane azeotropic
mixture. The catalysts were character-
ized by nitrogen adsorption–desorption
isotherms at 77 K, H2 chemisorption
X-ray diffraction (XRD), small angle
X-ray scattering (SAXS), X-ray photo-
electron spectroscopy (XPS), transmis-
sion electron microscopy (TEM), and
1
97
Au Mçssbauer spectroscopy. The
physical characterization by a combina-
tion of these techniques has shown that
the size and the structural characteris-
tics of the Pd–Au colloid precursor are
preserved when embedded in an SiO2
matrix. Catalytic tests were carried out
in selective hydrogenation of 3-hexyn-
1-ol, cinnamaldehyde, and styrene.
These data showed evidence that alloy-
ing Pd with Au in bimetallic colloids
leads to enhanced activity and most im-
portantly to improved selectivity. Also,
the combination of the two metals re-
sulted in catalysts that were very stable
against poisoning, as was evidenced for
the hydrogenation of styrene in the
presence of thiophene.
TEOS/THF/H O/HCl=1:3.5:4:0.05 for
2
the bimetallic Pd–Au and TEOS/THF/
H O/HCl=1:4.5:4:0.02 for Pd and Au
2
monometallic systems. After refluxing,
the colloid was added as a 4.5 wt% so-
lution in THF for Pd–Au, 10.2 wt% so-
lution in THF for Pd and 8.4 wt% so-
lution in THF for Au at room tempera-
ture. The gelation was carried out with
vigorous stirring (4 days) under an Ar
atmosphere. Following these proce-
1
13
29
measurements, solid-state H, C, Si-
CP/MAS-NMR spectroscopy, powder
Keywords: alloys · colloids · hydro-
genation · nanostructures · sol–gel
processes
dures, bimetallic Pd–Au–SiO catalysts
2
[
a] Prof. V. I. Pârvulescu
[d] Dr. G. Filoti, Prof. F. E. Wagner
University of Bucharest, Faculty of Chemistry
Department of Chemical Technology and Catalysis
B-dul Regina Elisabeta 4–12
Technische Universität München, Physics Department E 15
James Franck Strasse 1
85747 Garching bei München (Germany)
Bucharest 70346 (Romania)
Fax : (+40)213-159-249
E-mail: v_parvulescu@chem.unibuc.ro
[
e] Prof. C. Kübel
Fraunhofer Institute for Manufacturing Technology and
Applied Materials Science, Wiener Strasse 12
28359 Bremen (Germany)
[
b] Dr. V. Pârvulescu
Institute of Physical Chemistry of Romanian Academy of Sciences
Splaiul Independentei 202
Bucharest (Romania)
[
f] Prof. R. Richards
School of Engineering and Science
International University Bremen, Campus Ring 8
28759 Bremen (Germany)
[
c] Dr. U. Endruschat
W.C. Heraeus GmbH & Co. KG, Chemicals Division
Heraeusstr. 12–14, 63450 Hanau (Germany)
Chem. Eur. J. 2006, 12, 2343 – 2357
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
2343

Introduction
Pd catalysts refer to systems prepared by means of a co-im-
pregnation of the metal salts on solid supports using an in-
[25–29]
Supported bimetallic catalysts have elicited a great deal of
interest, especially motivated by an observed increase of the
selectivity and stability in several catalytic processes of prac-
cipient wetness technique followed by calcination.
How-
ever, by using the usual preparation procedures some metal
[28,30,31]
sintering and segregation of the surface is evident.
Re-
[
1,2]
tical importance.
The element Pd is one of the most in-
duction at high temperature in molecular dihydrogen and in-
teraction of these metals with the support favor this pro-
vestigated in this capacity, but interest has mainly focused
towards increasing its activity and stability. Among the nu-
merous metals which were investigated in combination with
[32]
cess. During segregation the more noble metal constitutes
the core and the less noble metal the shell of a bilayered
cluster. Therefore, among the techniques proposed to pre-
pare supported homogeneous bimetallic Pd–Au catalysts
[3–12]
Pd, Au was found to exhibit a particular influence.
by itself supported on SiO , Al O , or MgO exhibits rather
Gold
2
2
3
[
13]
[33]
low catalytic activity.
The fact that its activity is lower
Diamy et al. proposed the afterglow of a dihydrogen mi-
than that of Group 8 metals was explained by the absence
of unpaired d electrons in its electronic structure; this ab-
sence prevents Au from chemisorbing dihydrogen or dioxy-
crowave plasma (2450 MHz), which contains active species
(mainly atomic hydrogen) and which makes it possible to
work at a kinetic temperature low enough to avoid the for-
mation of large particles. Deposition of the two metals onto
the zeolite support was carried out by successive exchange.
However, even the resulting system did not consist of al-
loyed metals.
[14]
gen at ambient temperature. The increase of the noble-
metal activity by adding Pd offered good opportunities to
investigate basic principles in catalysis as a function of geo-
[15–17]
metric and electronic effects.
Pd–Au catalysts have been demonstrated to be particular-
ly effective in a number of reactions, including hydrogena-
However, significantly supported Pd–Au homogeneous
alloy structures were prepared by either low or high-tem-
[34]
[3,4]
[35]
tion of organic compounds,
production of vinyl ace-
perature procedures.
In the low-temperature procedure
[4–8]
[8]
3+
tate,
and hydrodechlorination of chloro- and chlorofluorocar-
liquid-phase oxidation of glyoxal to glyoxalic acid,
silica was metallated with [Au(en)2] upon ion exchange.
The presence of a Pd–Au alloy together with monometallic
species and PdO clusters was reported by Guczi et al. in the
[9–12]
bons.
The addition of Au has been claimed to improve
[36,37]
both the catalytic performance and the stability against de-
activation of Pd, as in the case of sulfur. Boitiaux et al. have
shown that Pd promoted with Au is a suitable catalyst for
selective hydrogenation in the downstream treatment of
case of Pd–Au/TiO and Pd–Au/SiO catalysts.
Bonar-
2
2
owska et al. reported a new procedure for the preparation
[25]
of homogeneous Pd–Au bimetallic catalysts.
These au-
thors showed that more readily reducible gold ions might be
directly reduced by prereduced Pd species, resulting in inti-
mate contact of both metals. However, a vigorous debate is
ongoing as to whether the resulting Pd–Au particles on the
support using these preparation techniques are fully al-
[3]
naphtha cracking, reducing the formation of green oil.
Au and Pd bimetallic nanoparticles, either as alloys or
core/shell structures, have received significant attention be-
[18]
cause of their special catalytic properties. Toshima et al.
[25,28]
have described the catalytic activity and analyzed the struc-
ture of the poly(N-vinyl-2-pyrolidone)-protected Au/Pd bi-
metallic clusters prepared by simultaneous reduction of
HAuCl and PdCl in the presence of poly(N-vinyl-2-pyroli-
loyed.
A combined X-ray absorption near-edge struc-
ture (XANES) and extended X-ray absorption fine structure
[38]
(EXAFS) study by Couves and Meeham
emphasized
these doubts. For the system Pd–Au/MgF , Malinowski
4
2
2
[11]
done). Other groups have reported the formation of Au/Pd
bimetallic particles with a palladium-rich shell by a simulta-
et al. demonstrated the existence of two Pd–Au solid solu-
tions, Pd/Au 0.81:0.19 and Pd/Au 0.4:0.6, and that not all the
metal is reduced.
[19]
neous alcoholic-reduction method.
Successive alcoholic
reduction did not give the core/shell-structured products,
but instead “cluster-in-cluster” structured products or mix-
A further problem lies in the dispersion of the metal.
There is strong evidence that the addition of Au to Pd salts
results in a marked decrease of the dispersion of the bimet-
[20]
tures of the monometallic particles. Harada et al.
and
[21]
[31]
Mizukoshi et al. reported the preparation and structure of
Au/Pd bimetallic nanoparticles by sonochemical reduction
allic catalyst systems. An innovative procedure to prepare
bimetallic catalysts containing Au was first proposed by Tur-
[22]
[39,40]
of gold(iii) and palladium(ii) ions, Remita et al. used pulse
kevich et al.
They suggested that preformed colloids can
electron-beam irradiated on Au–Pd ion-mixed solutions,
be embedded in an oxidic matrix leading to very active and
stable systems. Afterwards, numerous reports about the syn-
thesis of Pd–Au bimetallic colloids using different reductants
and in the presence of different stabilizer ligands appeared
[23]
while Harpeness and Gedanken used the microwave-assisted
[24]
polyol method. Wu et al. described the synthesis of Au/Pd
bimetallic nanoparticles in reverse micelles of water/sodium
bis(2-ethylhexyl)sulfosuccinate (AOT)/isooctane by the cor-
eduction of HAuCl and H PdCl with hydrazine at 258C.
[41–46]
in the literature.
By using this approach, highly dis-
persed bimetallic catalyst precursors could be prepared. Fol-
4
2
4
[47]
To provide effective catalysts it has been suggested that
lowing the so-called “precursor concept”,
such pre-pre-
[24]
Au must be intimately mixed with Pd; however, the prep-
aration of a supported homogeneous Pd–Au bimetallic
system is a rather complicated problem due to the segrega-
pared nanometals, in which the particle size distribution,
composition, and structure is rather well defined, may be
deposited on various supports giving a new type of hetero-
[4]
[47–49]
tion tendency of these metals. Most reports on bimetallic
geneous, bimetallic catalyst.
2344
www.chemeurj.org
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 2343 – 2357

Colloids
FULL PAPER
The incorporation of preformed ligand-stabilized metal
pore surface by reacting OHgroups with alkoxylsilyl-substi-
[74–76]
clusters and colloids into heterogeneous catalyst systems is
currently of interest. Most of the studies have been done
using polymer- or dendrimer-stabilized metal colloids and
tuted, short diphosphane ligands.
However, these can
immobilize only very small clusters, which are then rather
sensitive to the reaction media, mostly to leaching. In addi-
tion, this technique is more versatile for monometallic clus-
ters.
[50–60]
several ways have been proposed to solve this problem.
Organic polymers are not very stable at high temperature
and the catalytic reaction in solution may be accompanied
by some swelling, which can cause mass-transfer diffusion
problems. Therefore the use of inorganic supports appears
to be more suitable.
The size of the metallic ensembles is another critical issue
[4]
related to supported bimetallic catalysts. The use of im-
pregnation techniques evidently leads to a dispersion in the
[47]
metal–particle size. The “precursor concept”,
however,
The incorporation of metal colloids into inorganic matri-
eliminates this hindrance. Therefore, the incorporation of
colloids with well-defined particle size may yield additional
information to highlight the effect of the ensemble size in
heterogeneous catalysis. Further structural characterization
of the supported bimetallic catalysts and especially the ho-
mogeneity of these catalysts is a complicated task.
[61–63]
ces has also been reported in the literature.
Sol–gel has
also been shown as an effective technique to embed nano-
[
64–66]
scale mono- and bimetallic noble metals.
Yu et al. re-
[67]
ported that metal particles of tailored size (3–5 nm) and
composition prepared by means of inverse microemulsion
were encapsulated by ultrathin coatings (<2.5 nm) of inor-
The aim of this work is to present evidence about the in-
ganic porous SiO aerogels covered with surface OHgroups.
tegrity of SiO -embedded Pd, Au, and Pd–Au highly alloyed
2
2
These composite materials were then anchored to conven-
tional solid supports (alumina, carbon) upon mixing. More
recently, Crooks and co-workers reported synthesis of TiO2-
supported Pd–Au bimetallic nanoparticle catalysts prepared
using dendrimer-encapsulated nanoparticles. For such pur-
poses the dendrimer-encapsulated nanoparticles were em-
colloids prepared by sol–gel embedding of presynthesized
colloids and their performances in the hydrogenation of sev-
eral substrates: cinnamaldehyde, 3-hexyn-1-ol, and styrene.
The catalysts were characterized by adsorption–desorption
isotherms of nitrogen at 77 K, H chemisorption, XRD,
2
1
97
SAXS, Au Mçssbauer spectroscopy, and XPS.
bedded in TiO by using a polymeric sol–gel procedure with
2
[68]
Ti(OiPr) as a precursor. Another attempt to entrap aque-
4
ous monometallic nanoparticles on the surface of microme-
ter-sized zeolite particles functionalized with amine groups
Results
[69]
was proposed by Mandal et al,
who bound Pt and Pd
Catalysts preparation: Scheme 1 depicts the synthesis of
nanoparticles at high surface coverage on 3-aminopropyltri-
methoxysilane (APTS)-functionalized Na–Y zeolites. All
these data confirm that the incorporation of bimetallic col-
loids is a complicated problem, because the components
may interact differently with the support. In a previous com-
munication, we reported the preparation of a heterogeneous
Pd–Au catalyst by embedding pre-prepared tetraalkylammo-
SiO -embedded nanoscopic Pd, Au, and Pd-Au alloy col-
2
loids. After the gelation has been fulfilled the materials
were treated following the four routes D, C, R, and E de-
scribed in Scheme 1. Actually, in addition to the chemical
composition of the colloid and its loading, these treatments
further differentiate the investigated catalysts. Therefore,
further characterization and catalytic-behavior discussions
nium-stabilized Pd–Au alloy particles in a solid SiO matrix
2
following
process.
a
modified sol–gel
Encapsulation of
[70]
nanometer-sized metals in solid
matrices was also investigated.
Ichikawa used both micro- and
mesoporous silica supports to
prepare nanowire metals. How-
ever, such a procedure limits
the catalytic performance of the
resulting composites, since most
of the pores are blocked. Also,
for bimetallic catalysts the con-
trol of alloying inside the pores
[71–73]
is rather small.
A new approach has been
proposed by the group of
Schmid. To preserve the organi-
zation of the colloids inside the
pores these authors proposed
Scheme 1. Synthesis of SiO
2
-embedded nanoscopic Pd, Au, and Pd–Au alloy colloids. Routes C, D, E, R;
functionalization of the interior TEOS/THF/H
2
O/HCl=1:3.5:4:0.05 for Pd–Au; TEOS/THF/H
2
O/HCl=1:4.5:4:0.02 for Pd and Au.
Chem. Eur. J. 2006, 12, 2343 – 2357
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
www.chemeurj.org
2345

V. I. Pârvulescu et al.
2
will refer mainly to these treat- Table 2. Hydrogen chemisorption data for SiO -embedded catalysts.
3
ꢀ1
[a]
2
ꢀ1
[a]
2
ꢀ1
ments, by nominating catalysts Catalyst
Chemisorbed H uptake [cm g
]
Dispersion [%] SM [m g
]
SM [m g
]
2
sample
sample
belonging to one of these
groups (D, C, R, or E).
C-0.6% (Pd-Au)-SiO
R-0.6% (Pd-Au)-SiO
D-0.6% (Pd-Au)-SiO
E-0.6% (Pd-Au)-SiO
C-1% (Pd-Au)-SiO
R-1% (Pd-Au)-SiO
2
0.1201
0.0887
0.1261
0.1279
0.1989
0.1661
0.2087
0.2148
0.2233
0.2145
0.2367
0.2536
0.0840
0.0734
0.0865
0.0869
12.2
9.0
0.25
0.19
0.27
0.28
0.42
0.35
0.44
0.45
0.47
0.45
0.50
0.55
0.24
0.21
0.25
0.25
25.4
18.7
26.7
27.7
42.1
35.1
44.1
45.3
47.2
45.3
50.0
55.6
24.8
21.5
25.6
25.6
2
2
12.8
13.1
12.1
10.1
12.7
13.1
10.6
10.1
11.2
12.0
7.4
2
Textural characterization: Ther-
mal activation (723 K) of such
2
2
hybrid materials resulted in a D-1% (Pd-Au)-SiO2
porous texture with a rather E-1% (Pd-Au)-SiO
narrow pore diameter in the
region of the mesoporous mate-
rials. Similar data were ob-
2
C-1% (Pd)-SiO
R-1% (Pd)-SiO
D-1% (Pd)-SiO
E-1% (Pd)-SiO
2
2
2
2
tained for dried samples after C-1% (Au)-SiO
2
R-1% (Au)-SiO
D-1% (Au)-SiO
E-1% (Au)-SiO
2
6.4
7.6
7.6
they were treated following a
procedure proposed by Hitz
2
[77]
2
and Prins. For all the samples
the Brunauer–Emmett–Teller
[a] SM=Metallic surface area.
(
BET) surface areas of series D
were higher than those of series
C and R. Extraction of the stabilizer compound (sample E)
led to even higher surface areas. Such behavior can be cor-
related to the space occupied by the stabilizer compound
For each series of catalyst the hydrogen uptake was in-
creased in the order: D~E>C>R. During the direct reduc-
tion of the catalysts, part of the protector template was de-
composed to coke which hindered hydrogen chemisorption,
while the samples in which the template was extracted ex-
hibited higher hydrogen uptake than the calcined or directly
reduced samples.
(
upon extraction of the stabilizer this space may result in ad-
ditional surface area in the case of sample E) and the
amount of carbonaceous species formed by thermal decom-
position and insufficiently removed (sample D). These data
are compiled in Table 1.
Dispersion of the embedded metal colloids was also
rather low. The bimetallic-embedded colloids with 0.6 and
1
.0 wt% metal exhibited almost the same dispersion, which
Table 1. Textural characteristics of SiO
2
-embedded colloids.
BET surface
proves the reproducibility of the preparation procedure.
Dispersion of the monometallic Pd colloid in the silica-gel
matrix was found to be lower than in the case of the bimet-
allic Pd–Au colloid in the same matrix. For the same metal
loading, the Au-containing silica-gel matrix exhibited the
lowest dispersion.
Catalyst
Pd/Au ratio
Average pore
size [nm]
2
ꢀ1
area [m g
]
C-0.6%(Pd-Au)-SiO
2
52 : 48
267
97
304
323
193
78
231
244
261
87
298
319
201
56
16.6
17.3
16.1
16.2
18.8
20.2
18.1
18.1
23.0
25.2
22.6
22.7
22.8
24.0
22.1
22.1
R-0.6%(Pd-Au)-SiO
D-0.6%(Pd-Au)-SiO
E-0.6%(Pd-Au)-SiO
C-1%(Pd-Au)-SiO
R-1%(Pd-Au)-SiO
D-1%(Pd-Au)-SiO
E-1%(Pd-Au)-SiO
C-1%(Pd)-SiO
R-1%(Pd)-SiO
D-1%(Pd)-SiO
E-1%(Pd)-SiO
C-1%(Au)-SiO
R-1%(Au)-SiO
D-1%(Au)-SiO
E-1%(Au)-SiO
2
52 : 48
52 : 48
52 : 48
52 : 48
52 : 48
52 : 48
52 : 48
–
–
–
–
–
–
–
–
2
2
2
2
X-ray diffraction (XRD): XRD patterns of the embedded
Au, Pd, and Pd–Au colloids show an intensity distribution
corresponding to silica and to metallic or bimetallic species.
Figure 1 presents the patterns for the calcined and reduced
samples (i.e., samples D). With the exception of the reflec-
tion lines of silica, these patterns are identical to those re-
corded for stabilized colloids. An intense diffraction peak at
2
2
2
2
2
2
2
2
2
259
273
2q=228 originates from calcined SiO , showing that the sup-
2
2
port is not totally amorphous. All peaks in the pattern re-
corded for the bimetallic system containing 1.0 wt% Pd–Au
in SiO2 correspond to a bimetallic Pd–Au alloy, and no
peaks can be attributed to the presence of monometallic Au
or Pd species. A peak-broadening analysis was performed
on the basis of full width at half maximum (FWHM) values
obtained from peak fitting by using a pseudo-Voigt, peak-
shape function. The instrument line broadening was ob-
tained by using a standard Si sample. The result indicates a
particle size of about 4.9 nm both for the Pd–Au and the Au
system, whereas the particles in the pure Pd system appear
to be larger (ca. 5.6 nm). From the lattice-parameter calcula-
tion for the Au colloid (cubic Fm3m, a=4.074 Š), Pd colloid
SAXS investigation of these materials indicated that
pores were randomly distributed. No pore organization has
been detected.
Hydrogen chemisorption data: Table 2 contains the H2
chemisorption data. The hydrogen uptake on Au colloids
was very small, in agreement with the literature data for
[
31]
Au. The presence of Au in Pd–Au colloids also led to a
decrease of the hydrogen uptake, relative to Pd-containing
samples.
2346
www.chemeurj.org
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 2343 – 2357

Colloids
FULL PAPER
Mçssbauer spectroscopy: Figure 2 shows typical Mossbauer
spectra recorded for the investigated catalysts. For the non-
calcined, embedded Pd–Au colloids, the lines (Figure 2 and
Table 3) are noticeable, but are not much broader than
those of the equivalent alloys obtained by traditional meth-
Table 3. Mçssbauer results for samples of Au and Au–Pd Alloy colloids
(
embedded in silica or unsupported) and for rolled foils of Au and Au–
[
a]
Pd alloys used as reference materials.
ꢀ
1
ꢀ1
ꢀ2
Sample
IS [mms
]
s [mms
]
A
tot
d [mgcm ]
D-1%(Au)-SiO
2
ꢀ1.227 (7)
ꢀ1.224 (5)
ꢀ1.268 (8)
ꢀ0.015 (7)
0.446 (16) 0.147 (1) 1974
R-1%(Au)-SiO
C-1%(Au)-SiO
D-1%(Pd-Au)-SiO
R-1%(Pd-Au)-SiO
C-1%(Pd-Au)-SiO
D-0.6%(Pd-Au)-SiO
R-0.6%(Pd-Au)-SiO
C-0.6%(Pd-Au)-SiO
2
0.258 (17) 0.140 (1)
0.295 (25) 0.112 (1)
851
975
2
2
0.400 (17) 0.257 (2) 1768
842
0.404 (10) 0.198 (1) 1432
0.466 (13) 0.166 (1) 1694
2
0.048 (12) 0.680 (19) 0.123 (1)
2
ꢀ0.038 (5)
Figure 1. XRD patterns of the D catalyst series: (*) SiO
2
; (¥) Pd; (#) Au;
2
0.182 (6)
(
+) Pd–Au.
2
0.180 (11) 0.473 (18) 0.160 (1) 1657
0.163 (8) 0.430 (19) 0.158 (1) 1638
2
Au-Pd colloid in THF ꢀ0.118 (7)
0.707 (11) 0.179 (1)
—
100
16.0
Au
ꢀ1.232 (4)
ꢀ0.465 (3)
0.0 fixed
0.386 (8)
0.279 (7)
0.378 (5)
0.295 (7)
0.644 (3)
0.267 (1)
0.439 (1)
0.280 (1)
0.403 (1)
Au0.60 Pd0.40 as rolled
(
5
a=3.886 Š), and Pd–Au alloy colloid (a=3.976 Š), a
2 wt% Au and 48 wt% Pd composition of the Pd–Au
Au0.60 Pd0.40 annealed ꢀ0.410 (2)
16.0
14.0
28.9
16.0
28.9
90
90
40
Au0.50 Pd0.50 as rolled
ꢀ0.275 (2)
system has been concluded.
Au0.50 Pd0.50 annealed ꢀ0.207 (2)
Embedding of the colloids, calcining followed by reduc-
tion with hydrogen, direct reduction with hydrogen, or ex-
traction of the stabilizer compound from the embedded cat-
alysts did not produce any change in the position or intensi-
ty of the reflection lines assigned to the metallic colloids as
compared with pure colloids.
Au0.40 Pd0.60 as rolled
Au0.40 Pd0.60 annealed
Au0.10 Pd0.90 as rolled
Au0.05 Pd0.95 as rolled
Au0.03 Pd0.97 as rolled
ꢀ0.013 (4)
0.054 (3)
0.802 (4)
0.905 (5)
0.916 (7)
0.403 (10) 0.202 (1)
0.299 (7) 0.483 (1)
0.197 (19) 0.562 (2)
0.158 (25) 0.157 (1)
0.143 (40) 0.074 (1)
[
[
[
b]
b]
b]
[
a] IS: Isomer shift with respect to a Au:Pd source; s: width of the Gaus-
sian distribution of line positions; Atot: total area under the absorption
line; d: absorber thickness. Last digit errors are given in parentheses;
these errors are the statistical ones only. Annealed samples were treated
in H
2
in similar conditions with the R catalyst series. [b] Thickness given
in mm.
ods. This represents a good indication that they are macro-
scopically homogeneous. Therefore, one can rule out any
surface segregation of the component elements or broaden-
ing due to small particle effects, most probably because the
colloids are rather small (ca. 3.9 nm, as derived from XRD
measurements). Treating the catalysts in hydrogen at 723 K
(
R series, Figure 2 and Table 3) produced an enlargement of
the line, this enlargement is comparable with values ob-
tained for nonembedded colloids. This effect apparently
could be attributed to a segregation tendency into more Pd-
and Au-rich fractions, perhaps with one metal accumulating
on the surface. However, the two metals did not become
separated after heating. Segregation would lead to Au being
surrounded by more Au and less Pd, which would give rise
to smaller mean isomer shifts. Therefore, it appears that this
enlargement is caused by a change in the colloid–support in-
teraction. Such a supposition is also supported by the spec-
tra recorded for the catalysts with 0.6% (Pd–Au) (Table 3),
in which more metal is supposed to be directly interacting
with silica. Simple calcination (more evident for the catalyst
in series C with 1% (Pd–Au)) is less influenced by this inter-
action than the reduced catalysts.
As mentioned above, the nonembedded Pd–Au colloids
(Figure 2) measured in frozen THF definitely have a broad-
Figure 2. Mçssbauer spectra of Au–Pd samples.
Chem. Eur. J. 2006, 12, 2343 – 2357
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
www.chemeurj.org
2347

V. I. Pârvulescu et al.
er line than that of the embedded ones. It could be derived
either from the smaller size of particles or from a random
distribution of Au and Pd in the investigated colloids.
Mçssbauer spectra of pure Au colloids after heating (both
in an air and hydrogen atmosphere) give narrower lines that
indicate either a cluster shape phenomenon or a colloid ag-
[78–80]
glomeration on annealing.
The results obtained for the mean isomer shifts and for
the variance of the Gaussian distribution of isomer shifts are
given in Table 3 for samples of Au, Au–Pd alloy colloids
and for the arc-melted alloys used as reference materials.
Electric-quadrupole interactions were not explicitly taken
into account in the fitting procedure, but one should keep in
mind that they should exist in random alloys, because the
distribution of Au and Pd have different radii breaks and
perfect local-cubic symmetry. Moreover, in small particles,
the cubic symmetry is evidently broken for surface atoms.
No differences have been observed in the Mçssbauer
spectra between samples D and E.
Figure 3. Dependence of the Mçssbauer isomer shift on the alloy compo-
sition
The as-rolled alloys Au Pd to Au Pd lie about
6
0
40
40
60
ꢀ1
0
.1 mms below the straight line connecting the data points
this would be that the colloids contain more than 50% Pd,
which is in total concordance with analytical results.
for pure Au and dilute Au in Pd. Treatment of the rolled
alloys in hydrogen at 723 K for 20 h leads to a slight increase
in the isomer shift (Table 3). Moreover, the lines become
somewhat narrower during this treatment, that is, s decreas-
es slightly. This indicates that the distribution of Au and Pd
in the vicinity of the Au atoms changes during the thermal
treatment in the sense that Au is surrounded by somewhat
more Pd and fewer Au atoms than Au in the as-rolled foils.
The distribution of Au and Pd in the vicinity of Au appa-
rently also becomes narrower.
Comparing the Mçssbauer spectra of the embedded Pd–
Au colloids with those of the rolled alloys shows that col-
loids contain lines slightly broader than those of the rolled
alloys. This is additional confirmation that they are macro-
scopically homogeneous and remain intact even after ther-
mal treatment.
Heating all these samples (rolled alloys and colloids) re-
sulted in an increase of the isomer shift, which may indicate
that indeed the number of Pd atoms in the vicinity of Au in-
creases by heating. However, as was mentioned before, the
increase in isomer shifts induced by heating cannot be at-
tributed to segregation. Segregation would lead to Au being
surrounded by more Au and less Pd, which in turn would
lead to a smaller mean isomer shift. It could only result
from a homogenization of the Au–Pd distribution, or from
colloid–surface effects like changes of the colloid shape or
of the colloid–support interactions.
X-ray photoelectron spectroscopy (XPS): Table 4 contains
the XPS results of the investigated catalysts. In all samples
the binding energy assigned to Pd and Au species corre-
0
[81]
0[82]
Figure 3 shows the isomer shifts as a function of the alloy
composition near the Au Pd
sponds to Pd
and Au
components. Calcination fol-
5
0
50
composition. The line in this
figure is a straight line connect-
ing the datum point for pure
Au with that of dilute alloys of
Au in Pd. The data for all Au–
Pd colloid samples is plotted at
Table 4. XPS characteristics of the investigated catalysts.
[a]
3
Catalyst
Binding Energy [eV]
XPS Pd/Au ratio
XPS M /Si ratio”10
Pd3d3/2
Au4d3/2
Si2p
C-0.6% (Pd-Au)-SiO
R-0.6% (Pd-Au)-SiO
D-0.6% (Pd-Au)-SiO
2
340.9
340.9
340.8
340.7
340.9
340.9
340.8
340.8
340.9
340.9
340.8
340.8
340.9
340.8
340.8
340.7
353.1
353.0
353.1
353.1
353.0
353.0
353.1
353.0
353.1
353.0
353.0
353.0
353.1
353.1
353.0
353.1
103.6
103.5
103.4
103.6
103.6
103.6
103.5
103.6
103.5
103.6
103.4
103.6
103.6
103.5
103.6
103.6
1.10
1.09
1.12
1.10
1.11
1.08
1.12
1.14
1.12
1.08
1.09
1.07
1.11
1.12
1.13
1.12
3.3
3.2
3.1
3.2
6.3
6.2
6.4
6.5
5.3
5.4
5.4
5.1
4.6
4.5
4.4
4.3
2
2
the composition Au Pd . The E-0.6% (Pd-Au)-SiO2
5
0
50
C-1% (Pd-Au)-SiO
R-1% (Pd-Au)-SiO
D-1% (Pd-Au)-SiO
E-1% (Pd-Au)-SiO
2
isomer shifts for all composi-
tions lie above that for the
rolled and thermally treated
2
2
2
Au Pd
alloy by up to C-1% (Pd)-SiO2
50
50
ꢀ1
0
.25 mms . This can be inter- R-1% (Pd)-SiO
2
D-1% (Pd)-SiO
E-1% (Pd)-SiO
C-1% (Au)-SiO
R-1% (Au)-SiO
2
preted in two ways: 1) as a shift
arising from small particle ef-
fects or 2) as a shift caused by
Au in the particles surrounded D-1% (Au)-SiO2
on average by more than 50%
2
2
2
E-1% (Au)-SiO
2
Pd. The obvious explanation for [a] M=Pd, Au or (Pd+Au).
2348
www.chemeurj.org
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 2343 – 2357

Colloids
FULL PAPER
Figure 4. HAADF-STEM overview together with a more detailed image showing the distribution of the Pd and Au particles as bright spots (some of
them indicated by arrows) in the SiO matrix. The scalebars represent 200 and 100 nm in the left- and right-hand images, respectively.
2
lowed by reduction or direct reduction of these samples did
not cause any change in the oxidation state of these ele-
ments, indicating very good stability of the embedded col-
loids. The values determined for the Pd/Au ratios, which
were indeed in a very narrow interval for all these samples,
irrespective of the treatment they were exposed to, are also
very important. The ratios given in Table 4 actually corre-
spond to atomic Pd/Au ratios between 53:47 and 52:48,
which is the same ratio as that determined by Mçssbauer
and XRD measurements. The stability of the embedded col-
loids was also confirmed by the XPS atomic Pd, Au, or (Pd–
Au)/Si ratios. The values presented in Table 4 also confirm
the fact that different treatments do not affect the colloid in-
tegrity and dispersion.
trum of the D samples, 1% Pd–Au/SiO . It contains the typ-
ical bands assigned to the NR4 stabilizer (bands at 0.649,
1.348, and 7.664 ppm) and a band located at 2.820 ppm due
2
+
to a ꢀCHꢀOꢀ group, a remnant from the solvent during the
sol–gel synthesis. Heating the catalysts in hydrogen and air
led to a different situation. Air-calcined catalysts produced
materials in which no clear NMR band assigned to carbona-
ceous materials could be detected, while for the R series an
advanced decomposition of the stabilizer template was
1
3
indeed observed. The C-CP/MAS NMR spectrum present-
ed in Figure 7 for the R series, 1% Pd–Au/SiO2 catalyst
demonstrates such behavior. The existence of the various
carbon bonds in these catalysts seems indicative that the
template was not completely removed.
2
9
The Si CP/MAS NMR spectra of the the C series, 1%
Pd–SiO , 1% Au–SiO , and 1% Pd–Au/SiO catalysts are
TEM analysis: Grinding the bulk C-1% Pd–Au/SiO result-
2
2
2
2
ed in fragments with a typical diameter of 0.2–2 mm. High-
angle annular dark-field scanning transmission electron mi-
croscopy (HAADF–STEM) imaging of these fragments re-
vealed the distribution of the Au and Pd nanoparticles as
given in Figure 8. These spectra show the characteristic
2
9
structure for SiꢀO bonds. A Si MAS analysis using the
“block-decay” procedure indicated that the population of
4
the Q species slightly decreases for the Pd catalysts, which
bright spots in the SiO matrix. The size distribution and
might be in line with the increased size of the Pd colloids.
The differences between the R, C, and D series are very
small. Although the Q species were also detected, albeit at
2
particle density varied quite a bit throughout the sample,
but typically nanoparticles with a diameter of 5–30 nm were
observed (Figure 4). The larger particles correspond to su-
perficial aggregation of the colloids during calcination.
Energy dispersive X-ray (EDX) analysis of these particles
reveals the presence of Pd and Au in addition to the SiO2
matrix material. The STEM–EDX line scan (Figure 5) shows
that the nanoparticle consists of Pd and Au, which form a
uniform alloy within the resolution limit of the EDX analy-
sis. Quantification of the averaged line scan using the Pd-L
and Au-L lines reveals a ratio of 48 wt% Au and 52 wt%
Pd, which is within the accuracy of the EDX quantification,
in good agreement with the nominal 50:50 alloy composition.
2
a low concentration, these data confirm a high degree of
silica polymerization. With respect to the published data
3
4
,the Q /Q ratio, compared with published data proved to be
consistent with the incorporation of the colloids into the
[83]
29
silica network. On the basis of the same Si MAS analysis
it might be concluded that the so-called “T-surrounding spe-
cies” are absent.
Catalytic tests, hydrogenation of styrene: Figure 9 shows the
hydrogenation of styrene (Scheme 2) on the investigated
catalysts. Samples D and E exhibited the highest activity in
this reaction, which is in accordance with the exposed metal-
lic surface. Calcining in air or even worse in hydrogen led to
an important decrease of the activity (samples C and R).
Cross-polarization magic-angle-spinning (CP/MAS) NMR
1
spectroscopy: Figure 6 shows the H-CP/MAS NMR spec-
Chem. Eur. J. 2006, 12, 2343 – 2357
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
www.chemeurj.org
2349

V. I. Pârvulescu et al.
Figure 5. HAADF-STEM/EDX line profile showing the Au:Pd composition of a selected nanoparticle. Within the accuracy of the measurement, the
composition does not change significantly throughout the particle.
1
3
2
Figure 7. C-CP/MAS NMR spectrum of R 1% Pd–Au/SiO .
1
2
Figure 6. H-CP/MAS NMR spectrum of D Pd–Au/SiO .
activity. The chemoselectivity was 100% to ethylbenzene for
all of the catalysts investigated.
The activity of samples D and E was very similar for all
the investigated catalysts, indicating that for dried catalysts
the presence of the stabilizers did not diminish the accessi-
bility of the reactants. In both series, embedded Pd colloids
exhibited higher activity than the Pd–Au alloys. Therefore,
addition of Au in these catalysts leads to a decrease of the
Hydrogenation of styrene in the presence of thiophene:
Figure 10 shows variation of the activity for the hydrogena-
tion of styrene in the presence of thiophene during several
runs on the D catalysts. The poisoning with thiophene im-
2350
www.chemeurj.org
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 2343 – 2357

Colloids
FULL PAPER
than mixed Pd–Au colloids. After six runs, the activity of Pd
catalysts approximated the activity of Pd–Au catalysts. It
was thus observed that alloying of the two noble metals en-
hances the stability against poisoning with sulfur, which is
indeed remarkable.
This behavior occurred irrespective of the way in which
the catalysts were pretreated. Figure 11 shows variation of
the activity in the presence of thiophene for the C catalyst
series. In all the cases the poisoning with sulfur affected
only the activity and not the selectivity.
Hydrogenation of 3-hexyn-1-ol: Figure 12 shows the behav-
ior of the investigated catalysts in the hydrogenation of 3-
hexyn-1-ol (Scheme 3). For the embedded colloids tested,
the Pd–Au mixed colloids were the most effective, indicating
the positive effect of alloying in this case. In terms of activi-
ty, the differences between the catalysts with 0.6 and
1
.0 wt% were very small. The activity varied in the order
E~D>C>R, which is in accordance with the accessibility
to the metal surface.
The chemoselectivity towards hexenol reached 100% in
the case of embedded Au colloids (Figure 13). The other
catalysts also yielded hexenol and hexanol. However, even
for these catalysts, the chemoselectivity to hexenol exceeded
9
>
5%. A remarkably high regioselectivity (cis selectivity
90%) was obtained for all catalysts. Again, the best re-
sults were obtained on Au and bimetallic Pd–Au catalysts,
which exhibited a nearly complete cis selectivity. The addi-
tion of gold enhanced the regioselectivity relative to that of
embedded Pd/SiO colloids.
2
Hydrogenation of cinnamaldehyde: Figure 14 shows the var-
iation of the activity in hydrogenation of cinnamaldehyde
(
Scheme 4) on catalysts with different pretreatments. These
data refer to the transformation of cinnamaldehyde into cin-
namyl alcohol, 3-phenylpropanal, and 3-phenyl-1-propanol.
The activity of the investigated catalysts in this reaction was
smaller, when compared to that determined for hydrogena-
tion of 3-hexyn-1-ol or styrene. However, these data con-
firmed the results obtained in the hydrogenation of those
molecules, with the D and E catalyst series being more
active than the catalysts obtained at higher temperatures.
The selectivity to cinnamyl alcohol is the most important
parameter in this reaction. Figures 15 and 16 present the
variation of the selectivity for two series of catalysts,
namely, C and D. These data correspond to a moderate se-
lectivity in the desired product (ca. 50%). The way in which
the catalysts were pretreated brought about only minor
changes. On monometallic Pd- and Au-embedded colloids,
3
-phenylpropanal and 3-phenyl-1-propanol were found to be
2
9
the predominant reaction products. Alloying of the two
nanometer-sized metals in colloids led to a very important
enhancement of the selectivity, with respect to the monome-
tallic ones, with cinnamyl alcohol being the major product.
The solvent plays a very important role in this reaction
because the alcohols used leads to acetals. Thus, for the re-
actions carried out in ethanol, the formation of 1,1-dieth-
Figure 8. Si CP/MAS NMR spectrum of a) C-1% Au-SiO
2
catalysts
(Q :Q :Q =51:42:7), and
2
catalysts (Q :Q :Q =50:41:9).
4
3
2
4
3
2
(
Q :Q :Q =52:43:5), b) C-1% Pd–Au/SiO
2
4
3
2
c) C-1% Pd-SiO
pacted the activity of these catalysts differently; embedded
Pd and Au colloids being much more dramatically affected
Chem. Eur. J. 2006, 12, 2343 – 2357
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
www.chemeurj.org
2351

V. I. Pârvulescu et al.
oxy-3-phenylpropane occurred with conversions over 25%.
It is worth noting that the conversion of cinnamaldehyde to
1
,1-diethoxy-3-phenylpropane was even higher for embed-
ded colloids than for pure silica, also prepared by means of
the sol–gel route. In the presence of isopropanol as a sol-
vent, the formation of acetals
Scheme 2. Hydrogenation of styrene.
was less than 10%.
Discussion
Both Mçssbauer and XRD
analyses of the free Pd–Au col-
loid evidenced a fully alloyed
material consisting of 52% Pd
and 48% Au. These values cor-
respond closely to those found
by chemical and EDX analysis.
Embedding of the Pd–Au col-
loids in the sol–gel matrix does
not alter the Mçssbauer spec-
trum. The similarities between
the isomer shifts in the free Pd–
Au colloid and in the embed-
ded sample confirmed that no
modification of the colloid
structure had occurred during
the sol–gel embedding process.
The integrity of the incorporat-
ed Pd–Au alloy particles re-
mained virtually untouched as
evidenced by XPS. In addition,
these measurements confirmed
the preservation of the oxida-
tion state irrespective of the
treatment the catalysts were ex-
posed to. No band assigned to
any oxidized species has been
detected in these analyses.
Figure 9. Comparative activity in the hydrogenation of styrene of the sol–gel embedded catalysts as a function
of the pretreatment and colloid composition (10% conversion, 50 mg catalyst, 353 K, 20 bar H ).
2
Figure 10. Evolution of activity for the D catalyst series in the presence of 100 ppm thiophene (50 mg catalyst,
53 K, 20 bar H , 30 min).
Preservation of the colloids
was also confirmed by TEM
analysis. TEM measurements
3
2
(
Figures 4 and 5) have con-
firmed a narrow size distribu-
tion of the incorporated, bimet-
allic colloid (about 5.0 nm).
Embedding of these colloids
led to macroporous, solid mate-
rials with intermediate (ca.
2
ꢀ1
3
00 m g ) surface areas. The
pore diameter and the porous
texture could be associated
with the characteristics of the
surfactant used, that is, rather
large stabilizers. SAXS has
shown that the channels are
randomly distributed and no
Figure 11. Evolution of the activity for the C catalyst series in the presence of 100 ppm thiophene (50 mg cata-
lyst, 353 K, 20 bar H , 30 min).
2
2352
www.chemeurj.org
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 2343 – 2357

Colloids
FULL PAPER
hydrogen, the decomposition is
not complete, as indicated from
1
3
C-CP/MAS NMR spectra.
The surface areas of the R cata-
lysts were low in comparison
with the other catalysts, which
may be an indication that the
pores are partially blocked by
coke deposits.
Alloying Pd with Au has long
been expected to impart an im-
provement of the catalysts per-
[3–12]
formances.
These results
confirm that assumption. In ad-
dition, the uniformity of the
metal-particle size in the cata-
Figure 12. The activity in hydrogenation of 3-hexyn-1-ol of the catalysts differently pretreated (conversion
1
2
5%, 6.1 mL of 3-hexyn-1-ol, 50 mg of catalyst (1 wt-% metal incorporated), 20 bar H , 293 K).
Scheme 3. Hydrogenation of 3-hexyn-1-ol.
Figure 13. The chemoselectivity to 3-hexen-1-ol (^,~,*,*) and regioselec-
tivity for cis-3-hexen-1-ol (^,~,
,*) on the catalysts differently pretreated
((^,^-1%(Pd); ~,~- 0.6%(Pd); &,
- 1%(Pd-Au); *,*-1%(Au); conver-
sion 15%, 6.1 mL of 3-hexyn-1-ol, 50 mg of catalyst (1 wt% metal incor-
porated), 20 bar H , 293 K).
tubes (such as those found in MCM materials) or other or-
ganized pore architectures were present here. However, it is
expected that around the colloids well-defined cavities exist
&
&
2
(
this is pretty speculative and should be verified by some
scientific evidence).
CP/MAS NMR analysis indi-
cated that the D series corre-
sponded to catalysts in which
the stabilizer compound is pre-
served without any damages.
Silica is polymerized to a high
degree in these catalysts, al-
though by comparison with
pure silica prepared by a sol–
gel procedure, the content in
3
the Q species and the existence
2
of Q ones account for a lower
degree of polymerization. Heat-
ing of these catalysts either in a
hydrogen atmosphere or in air
followed by reduction in hydro-
gen led to an expected decom-
position of the stabilizer com-
pounds. However, in the case of
the catalysts treated directly in Figure 14. Variation of the activity as a function of the catalysts pretreatment (solvent isopropanol).
Chem. Eur. J. 2006, 12, 2343 – 2357
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
www.chemeurj.org
2353

V. I. Pârvulescu et al.
sponds to the more active catalysts, will be discussed further.
The results for the E series were very close to those for the
D series, but the E series requires an additional step, that is,
the extraction of the stabilizer ligand. The performances of
the C and R series decreased in this order. These results
may be explained, both by blocking of the pores by carbona-
ceous deposits (mostly in the case of the R series), but also
by a blocking of the colloid surfaces by these deposits. The
very good concordance of the activity of the catalysts with 1
and 0.6% (Pd–Au) confirmed the absence of any structure-
sensitive effects.
Scheme 4. Hydrogenation of cinnamaldehyde.
Hydrogenation of pure sty-
rene occurred with the higher
activity on Pd colloids, while
the addition of Au decreased
the activity of the embedded
colloids. Pure Au catalysts ex-
hibited a very poor activity.
However, reactions carried out
in the presence of thiophene in-
dicated that Pd–Au catalysts
have a high level of tolerance
to the poison, which might be
of a great practical importance.
It is not the case with Pd col-
loids, for which a continuous
decrease of the activity in the
Figure 15. The variation of selectivity to cinnamalcohool, 3-phenylpropanal, and 3-phenyl-1-propanol for the C
series of catalysts.
presence of thiophene was ob-
served. Hydrogenation of 3-
hexyn-1-ol and cinnamaldehyde
indicated Pd–Au colloids as the
most active catalysts in the ab-
sence of any poison.
The selectivity of these cata-
lysts is possibly more important
in the investigated reactions.
Hydrogenation of styrene oc-
curred with a total chemoselec-
tivity to ethylbenzene irrespec-
tive of the catalyst. No hydro-
genation of the benzyl ring was
detected. For the other mole-
cules examined in hydrogena-
tion reactions the chemical
composition of the colloids was
an important factor. Thus, in
the hydrogenation of 3-hexyn-
1
-ol the interesting compound
Figure 16. The variation of selectivity to cinnamyl alcohol, 3-phenylpropanal, and 3-phenyl-1-propanol for the
D series of catalysts.
is cis-3-hexen-1-ol. Au colloids
were the most chemoselective.
The very high regioselectivity
lysts prepared in this study eliminated any comment about
the structure effects. The behavior of these catalysts in the
investigated reactions differed according to the nature of
the substrate molecule. However, since the activation of the
catalysts imparts a strong influence on the performances of
these catalysts only the results of the D series, which corre-
of these catalysts (cis selectivity >90%) and the fact that
bimetallic Pd–Au catalysts exhibited a near complete cis se-
lectivity is also very important. These data clearly indicated
that the addition of gold enhanced the regioselectivity, when
compared with embedded Pd/SiO colloids.
2
2354
www.chemeurj.org
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 2343 – 2357

Colloids
FULL PAPER
Hydrogenation of cinnamaldehyde represented the case
decomposition in the presence of air, all steps were carried out under an
Ar atmosphere. The resulting gel was dried at 383 K by using a ramp of
in which a cooperative effect for these colloids was ob-
served. As previously mentioned, the important product of
this reaction is cinnamyl alcohol. Hydrogenation on both Pd
and Au colloids led to a very poor selectivity for this com-
pound. Alloying these metals into Pd–Au colloids resulted
in an enhanced selectivity, when compared to the monome-
tallic catalysts.
ꢀ
1
0
7
.12 Kmin . A portion of these materials was then calcined under air at
23 K with a ramp of 0.3 Kmin . After cooling, the samples were re-
ꢀ1
duced in a flow of pure hydrogen at 723 K using the same ramp. Reduc-
3
ꢀ1
tion was carried out with a H
2
flow of 30 cm min . In some dried sam-
ples the surfactant was extracted using the procedure of Hitz and
[
77]
Prins.
Accordingly, the dried parent material (5 g) was stirred in an
ethanol heptane azeotropic mixture (200 mL) for 1 h at 351 K, followed
by filtration and washing.
These data confirmed the fact that gold by itself, irrespec-
tive of the support (SiO , Al O , or MgO) exhibits a rather
In conclusion, four different routes were used to obtain the catalysts. The
materials that were calcined in air and then reduced were denoted as C,
those directly reduced in hydrogen as R, and those that were simply
dried as D. Additionally, samples in which the surfactant was extracted
using an ethanol heptane azeotropic mixture were denoted as E. Follow-
2
2
3
[13]
low catalytic hydrogenation activity.
However, recent
studies demonstrated that nanometer-sized gold particles
may act as active catalysts in selective hydrogenation of al-
ing these procedures bimetallic Pd–Au/SiO
metal, and monometallic Pd and Au–SiO
2
catalysts with 0.6 and 1 wt%
catalysts with 1 wt% metal
[
84–87]
lylic compounds.
In the present study, although the size
2
[84]
of gold particles was close to that reported by Claus et al.
were prepared. The metal loading in these catalysts was confirmed by
ICP–AES spectroscopy.
both the activity and selectivity in the hydrogenation of cin-
namaldehyde were very low. In this particular case, the addi-
tion to Pd was found to lead to an enhanced catalytic activi-
ty and selectivity, which might be interpreted in terms of
electronic effects. In the case of hydrogenation of cinnamal-
dehyde and crotonaldehyde, the effect of the addition of a
second metal such as Sn was interpreted as follows: the spe-
cies created by tin in close proximity to the noble metal acti-
Catalysts characterization: Characterisation of the Pd–Au colloids, both
in the free and in the embedded state, was first performed using TEM,
EDX, and atomic absorption spectroscopy (AAS). In addition, nitrogen
adsorption–desorption curves at 77 K, H
2
-chemisorption measurements,
1
13
29
solid-state H, C, and Si-CP/MAS-NMR spectroscopy, XRD, SAXS,
1
97
XPS, TEM, and Au Mçssbauer spectroscopy were applied.
The sorption isotherms of N at 77 K were obtained with a Micromeritics
2
ASAP 2000 apparatus after outgassing the samples at 393 K for 24 h
under vacuum 0.00001 mbar.
[88,89]
vate the C=O group.
A similar behavior can be expect-
H
2
-chemisorption measurements were carried out using a Micromeritics
ASAP 2010C apparatus. The samples were activated under vacuum at
03 K for 36 h. Afterwards, a hydrogen flow was passed initially at 308 K
for 15 min, then the temperature was increased to 403 K at a heating rate
ed in this case as well. Hydrogen-chemisorption data are
also in line with these results.
4
ꢀ
1
of 5 Kmin and maintained for 2 h. After reduction, the samples were
purged with a helium flow at 403 K for 2 h and then at 308 K for another
Conclusion
3
0 min. The amount of chemisorbed hydrogen was measured at 308 K by
the desorption method after equilibration for 45 min in 400 mbar of ad-
sorbate. The total hydrogen uptake was determined by extrapolating the
The data presented in this paper show evidence that alloying
of Pd with Au on a nanometer scale leads to improvements
in activity in hydrogenation reactions and have a positive
effect on the relevant chemoselectivity. The alloying of the
two metals may improve catalyst tolerance to poisoning by
sulfur compounds. The way in which the colloid-embedded
catalysts were pretreated was crucial for the activity of these
catalysts. High temperatures resulted in decomposition of
the stabilizer ligand, which caused blockage of the colloid
surface, and thus deactivation of the catalysts.
linear portion of the adsorption isotherm to zero pressure. Reversible H
2
sorption was measured by outgassing at 0.00005 mbar at the adsorption
temperature and running a second isotherm. The difference between the
total and reversible uptakes was ascribed to irreversible hydrogen
uptake. The metal dispersion, surface area, and particle size were deter-
mined from the irreversible uptake, assuming a H/metal stoichiometry of
1
. For the Pd–Au colloids, the calculation of dispersion was made consid-
ering Equations (1)–(3), in which V is the hydrogen uptake, SF is the cal-
culated stoichiometry factor, weight is the percentage of the sample
weight, and A is the effective area of one active metal atom.
i
m
1
00 % Â 100 %
V Â SF
D ¼
 %
ð1Þ
weightPd
Watomic Pd
%weightAu
þ Watomic Au
2
2414
Experimental Section
2
3
6
:023 Â 10
ð2Þ
ð3Þ
Ms ¼
 V  SF  A
m
Catalysts preparation: The Pd, Au, and Pd/Au (50% Au, 50% Pd) col-
22412
loid precursors were prepared following a procedure described else-
where. Embedding of these colloids was carried out by using tetrae-
thoxysilicate (TEOS) as the precursor for the silica support. Since the
[
70]
1
00 %
Msm ¼ Ms Â
%
weightPd þ %weightAu
4 8
NR C H19- stabilized Pd–Au colloids are completely insoluble and even
decompose at elevated temperatures in alcoholic solutions, the normal
sol–gel procedure required modification, thus THF was used as the sol-
vent. To have comparable preparations, the same procedure was used for
monometallic Pd and Au colloids. The molar composition of the sol was
The XPS spectra were recorded by using a SSI X probe FISONS spec-
trometer (SSX-100/206) with monochromated AlKa radiation. The spec-
trometer energy scale was calibrated using the Au4f7/2 peak (binding
energy 84.0 eV). For calculation of the binding energies, the C1s peak of
the C-(C,H) component of adventitious carbon at 284.8 eV was used as
an internal standard. To limit reoxidation, the reduced samples were
transferred from the reduction setup to the XPS apparatus under isooc-
tane. The peaks assigned to Pd3d3/2, Au4d3/2, Si2p, and O1s levels were
analyzed.
TEOS/THF/H
2
O/HCl=1:3.5:4:0.05 for the bimetallic Pd–Au and TEOS/
THF/H O/HCl=1:4.5:4:0.02 for Pd and Au monometallic systems. The
2
above mixture was refluxed at 343 K for 2 d, after which the colloid was
added as a 4.5 wt-% solution in THF for Pd–Au, 10.2 wt% solution in
THF for Pd, and 8.4 wt% solution in THF for Au at room temperature.
The gelation was carried out under vigorous stirring for 4 d. To avoid any
Chem. Eur. J. 2006, 12, 2343 – 2357
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
www.chemeurj.org
2355

V. I. Pârvulescu et al.
XRD patterns were recorded with a D5000 Kristalloflex diffractometer
from Siemens using CuKa radiation (l=1.5418 Š). Data acquisition was
done in the 2q=2–658 range, with a scan step of 0.038.
Catalytic tests were carried out in selective hydrogenation of 3-hexyn-1-
ol, which gives the corresponding cis-hexenol derivative (a valuable per-
[
46]
fume ingredient),
of cinnamaldehyde to cinnamyl alcohol, which is
[
94]
used in the field of flavor and fragrances, and of styrene, which is im-
SAXS experiments were performed with an evacuated Kratky compact
camera mounted on a Siemens rotating copper anode. A take-off angle
of 38 was used for these experiments. The power of the generator was set
to 200 mA”30 kV; selection of CuKa radiation was achieved with the
help of a Ni filter, completed by electronic discrimination. A 1D propor-
tional counter (mBraun OED50m) placed at a distance of 24.5 cm from
the sample was used to record the scattering patterns. The channel width
of the detector amounted to about 71 mm, close to the resolution of the
detector. Precise values of channel width and sample-to-detector distan-
ces were obtained by measuring calibration standards, namely rat-tail col-
lagen and crystalline monodisperse oligomers with the repeating unit (-F-
O-F-O-F-CO-), in which F represents a para-linked aromatic moeity. The
beam shape P(h,v) was measured in the plane of the detector in the hori-
zontal direction with 0.658 cm as a standard deviation. In the vertical di-
rection (v), the beam was very sharp (about 300 mm FWHM), and was
thus considered as a delta function. The powders were introduced into
Mark glass capillaries 1 mm in diameter. Acquisition times of 1800 s were
used throughout. The scattering of an empty capillary was also recorded
to obtain the so-called parasitic scattering. The parasitic scattering, scaled
by sample transmittance, was subtracted from sample scattering to obtain
the parasitic-corrected scattering.
[
95]
portant for the treatment of the gasoline from pyrolysis. Under stan-
dard test conditions, 3-hexyn-1-ol (6.1 mL) was converted without solvent
in a stirred 50 mL stainless-steel autoclave using a catalyst (50 mg, 1 wt-
%
metal incorporated) at 20 bar hydrogen and 293 K. Experiments con-
cerning hydrogenation of styrene were carried out in the same type of
autoclave using the same volume of reactant and the same amount of cat-
alyst at 20 bar hydrogen but at 353 K. Additionally, several runs were
performed in which styrene (6.1 mL) and thiophene (100 ppm) were
treated under similar experimental conditions. After each run, the cata-
lyst was separated and reused in a new cycle. Hydrogenation of cinna-
maldehyde was carried out in the same type of autoclave by using the
substrate (1.1 mL) dissolved in ethanol or isopropanol (5 mL) under
2
0 bar hydrogen pressure and at 353 K. The reaction products were ana-
lysed by GC using a Varian Star 3400CX gas chromatograph equipped
with a Chrompack 7530 capillary column (WCOT fused silica column
with CP-Sil PONA CB stationary phase) and a Hewlett Packard 5890 ser-
ies II gas chromatograph, using a WCOT Fused Silica CP-Wax 58
(
FFAP) CB column, of 50 m length. The reaction products were identi-
fied using the same GC coupled with a Finnigan MAT SSQ7000 mass
1
13
spectrometer and by Hand
C NMR spectroscopy with a Varian
1
Gemini 300BB instrument, operating at 300 MHz for Hand 75 MHz for
1
97
13
The Au Mçssbauer measurements were performed with sources and
absorbers cooled to 4.2 K in a liquid He bath cryostat. The sources were
moved with a sinusoidal velocity waveform. An intrinsic Ge detector was
C.
1
97
197
used to detect the 77 keV gamma rays of Au. The sources of Pt
1
96
(T1/2 =19 h) were produced by irradiating isotopically enriched Pt (ca.
00 mg) in a thermal neutron flux of about 2 ” 10 cm s at the Munich
1
3
2 ꢀ1
2
Acknowledgements
Research Reactor for about 20 h. For comparison with samples of Au–Pd
colloids, a number of Au–Pd alloys of different compositions were pre-
pared by arc melting in an argon atmosphere and subsequent rolling into
foils of appropriate thickness. The Mçssbauer spectra of all samples con-
sisted of a single, more or less broadened line. To these experimental
spectra, appropriate line shapes were fit by using transmission integral
Vasile I. Pârvulescu thanks CERES, Project 4–140, and Ryan Richards
thanks IUB for financial support.
[
1] J. H. Sinfelt, G. H. Via, F. W. Lytle, Catal. Rev. Sci. Eng. 1984, 26,
1–140.
2] G. C. Bond, Chem. Soc. Rev. 1991, 20, 441–475.
[
90,91]
calculations.
It was found that line shapes corresponding to a Gaussi-
8
an distribution of isomer shifts gave virtually perfect fits to the data. The
emission line of the source and the distributed line width of the absorbers
were assumed to be Lorentzian lines with a natural width of G=
[
[3] J. P. Boitiaux, J. Cosyns, M. Derrien, G. Leger, Hydrocarbon Process.
1985, 64, 51–59.
ꢀ
1 [92]
0
.95 mms
.
Since the line broadening due to finite absorber thickness
[4] G. C. Bond, A. F. Rawle, J. Mol. Catal. A 1996, 109, 261–271.
[5] B. M. Chapnick, P. W. Paustian, E. Klainer, P. D. Joiner, A. L.
Hyman, P. J. Kadowitz, UK Patent 1246015, 1971 to Bayer.
[6] G. Scarfe, W. Kronig, US Patent 3775342, 1973 to Bayer.
[7] P. Wirtz, K. Woerner, F. Wunder, K. Eichler, G. Roscher, European
Patent 519436, 1992 to Hoechst.
was taken into account by the transmission, integral fits, the variance s of
the Gaussian distribution of line positions in the absorbers is expected to
be due only to the inhomogeneous line broadening in the alloy samples
arising from the peculiar coordination and distribution of Au and Pd
atoms in the vicinity of the Mçssbauer atoms.
[
[
8] S. Hermans, M. Devillers, Catal. Lett. 2005, 99, 55–64.
9] S. Morikawa, S. Samejima, M. Yositake, S. Tatsematsu, European
Patent 0347830, 1989 for Asahi Glass Co.
The MAS NMR experiments were performed at room temperature on a
Bruker ASX100 spectrometer (2.33T) operating at a Larmor frequency
2
9
of 19.89 MHz for Si with a 4 mm probehead, and of 400.33 and
1
13
[10] V. N. M. Rao, US Patent 5447896, 1995 to du Pont de Nemours. A.
Malinowski.
[11] A. Malinowski, W. Juszczyk, J. Pielaszek, M. Bonarowska, M. Woj-
ciechowska, Z. Karpinski, Chem. Commun. 1999, 685–686.
[12] M. Nutt, J. Hughes, M. Wong, Environ. Sci. Technol. 2005, 39, 1346–
1353.
13] S. Galvagno, G. Parravano, J. Catal. 1979, 57, 272–286.
14] J.-C. Legrand, A.-M. Diamy, G. Riahi, Z. Randriamanantenasoa, M.
Polisset-Thfoin, J. Fraissard, Catal. Today 2004, 89, 177–182.
15] S. Galvagno, G. Parravano, J. Catal. 1978, 55, 178–190.
16] J. W. Sachtler, G. A. Somorjai, J. Catal. 1983, 81, 77–94.
17] A. Sachdev, J. Schwank, J. Catal. 1989, 120, 353–369.
[18] N. Toshima, M. Harada, Y. Yamazaki, K. Asakura, J. Phys. Chem.
1992, 96, 9927–9933.
[19] H. Liu, G. Mao, M. Meng, J. Mol. Catal. 1992, 74, 275–284.
[20] M. Harada, K. Asakura, N. Toshima, J. Phys. Chem. 1993, 97, 5103–
5114.
1
00.67 MHz with a 7 mm probehead for Hand
C, respectively. The
MAS NMR spectra were obtained at 5 kHz spinning rate by using a
1
single-pulse excitation (p/10). For Hspectra 32 transients were added
1
3
1
with a recycling delay of 1 s. C spectra were acquired with HTPPM de-
2
9
coupling; 3000 scans were accumulated. Si spectra were acquired with a
s recycle delay. Hand C chemical shifts were referenced to TMS. The
Si chemical shifts were referenced relative to Si(CH ) . The simulations
3 4
were performed using the DMfit program.
1
13
6
[
[
2
9
[
93]
The electron microscopic analysis was performed by using a FEI Tecnai
F20 ST instrument operating at 200 kV in STEM mode equipped with an
EDAX r-TEM Si(Li) EDX detector with a super ultra-thin window (S-
UTW). HAADF–STEM images and drift-corrected STEM–EDX line
scans were acquired with a nominal spot size of 1 nm. The line scans
were analyzed by using TIA software version 2.0.202. Quantification of
the EDX spectra was done after background correction without thickness
correction within TIA. The samples were prepared for TEM analysis by
grinding the bulk material resulting in a fine powder, which was distribut-
ed on 200 mesh, carbon-coated copper grids.
[
[
[
[21] Y. Mizukoshi, T. Fujimoto, Y. Nagata, R. Oshima, Y. Maeda, J.
Phys. Chem. B 2000, 104, 6028–6032.
2356
www.chemeurj.org
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 2343 – 2357

Colloids
FULL PAPER
[
22] H. Remita, A. Etcheberry, J. Belloni, J. Phys. Chem. B 2003, 107,
1–36.
[60] R. W. J. Scott, O. M. Wilson, S.-K. Oh, E. A. Kenik, R. M. Crooks, J.
Am. Chem. Soc. 2004, 126, 15583–15591.
3
[
[
[
23] R. Harpeness, A. Gedanken, Langmuir 2004, 20, 3431–3434.
24] M. L. Wu, D. H. Chen, T. C. Huang, Langmuir 2001, 17, 3877–3883.
25] M. Bonarowska, J. Pielaszek, W. Juszczyk, Z. Karpinski, J. Catal.
[61] G. Ihlein, B. Junges, U. Jungas, F. Laeri, F. Schüth, U. Vietze, Appl.
Organomet. Chem. 1998, 12, 305–314.
[62] M. T. Reetz, M. Dugal, Catal. Lett. 1999, 58, 207–212.
[63] S. Tao, Z. Zhan, G. Meng, J. Mater. Sci. Lett. 1999, 18, 707–710.
[64] E. Lindner, T. Schneller, F. Auer, H. A. Mayer, Angew. Chem. 1999,
111, 2288–2309; Angew. Chem. Int. Ed. 1999, 38, 2155–2174.
[65] F. Schweyer, P. Braunstein, C. Estournes, J. Guille, H. Kessler, J.-L.
Paillaud, J. Rose, Chem. Commun. 2000, 1271–1272.
[66] P. Braunstein, R. Devenish, P. Gallezot, B. T. Heaton, C. J. Hum-
phreys, J. Kervennal, S. Mulley, M. Ries, Angew. Chem. 1988, 100,
972–973; Angew. Chem. Int. Ed. Engl. 1988, 27, 927–929.
[67] K. M. K. Yu, C. M. Y. Yeung, D. Thompsett, S. C. Tsang, J. Phys.
Chem. B 2003, 107, 4515–4526.
2
000, 195, 304–315.
[
[
26] E. L. Kugler, M. Boudart, J. Catal. 1979, 59, 201–210.
27] W. Juszczyk, Z. Karpinski, J. Pielaszek, Z. Paal, J. Catal. 1993, 143,
5
83–593.
[
28] W. Juszczyk, Z. Karpinski, D. Lomot, J. Pielaszek, J. W. Sobczak, J.
Catal. 1995, 151, 67–76.
[
[
[
29] R. J. Wu, T. Y. Chou, C. T. Yeh, Appl. Catal. B 1995, 6, 105–116.
30] Z. Karpinski, Adv. Catal. 1991, 37, 45–100.
31] A. Malinowski, W. Juszczyk, D. Lomot, J. Pielaszek, Z. Karpinski,
Polish. J. Chem. 1995, 69, 308–315.
[
32] M. Boudart, G. Meitzner, in EXAFS and Near Edge Structure III
[68] R. W. J. Scott, C. Sivadinarayana, O. M. Wilson, Z. Yan, D. W.
Goodman, R. M. Crooks, J. Am. Chem. Soc. 2005, 127, 1380–1381.
[69] S. Mandal, D. Roy, R. V. Chaudhari, M. Sastry, Chem. Mater. 2004,
16, 3714–3724.
[70] H. Bçnnemann, U. Endruschat, B. Tesche, A. Rufꢃnska, C. W. Leh-
mann, F. E. Wagner, G. Filoti, V. Pârvulescu, V. I. Pârvulescu, Eur. J.
Inorg. Chem. 2000, 819–822.
[71] M. Sasaki, M. Osada, N. Sugimoto, S. Inagaki, Y. Fukushima, A. Fu-
kuoka, M. Ichikawa, Microporous Mesoporous Mater. 1998, 21, 597–
606.
(Eds.: K. O. Hodgson, B. Hedman, J. E. Penner-Halm), Springer,
Berlin, 1984, p. 217.
[
33] A.-M. Diamy, Z. Randriamanantenasoa, J.-C. Legrand, M. Polisset-
Thfoin, J. Fraissard, Chem. Phys. Lett. 1997, 269, 327–332.
34] R. J. Davis, M. Boudart, J. Phys. Chem. 1994, 98, 5471–5477.
35] S. H. Inami, H. Wise, J. Catal. 1972, 26, 92–96.
[
[
[
36] L. Guczi, A. Beck, A. Horvꢁth, Zs. Koppꢁny, G. Stefler, K. Frey, I.
Sajꢂ, O. Geszti, D. Bazin, J. Lynch, J. Mol. Catal. A 2003, 204/205,
5
45–552.
[
37] A. M. Venezia, L. F. Liotta, G. Pantaleo, V. La Parola, G. Deganello,
A. Beck, Zs. Koppꢁny, K. Frey, D. Horvꢁth, L. Guczi, Appl. Catal.
A 2003, 251, 359–368.
[72] M. Ichikawa, in Metal Clusters in Chemistry (Eds.: P. Braunstein,
L.A. Oro, P. R. Raithby), Wiley-VCH, Weinheim, 1999, p. 1273.
[73] A. Fukuoka, H. Araki, Y. Sakamoto, S. Inagaki, Y. Fukushima, M.
Ichikawa, Inorg. Chim. Acta 2003, 350, 371–378.
[74] T. Hanaoka, H.-P. Kormann, M. Kroll, T. Sawitowski, G. Schmid,
Eur. J. Inorg. Chem. 1998, 807–812.
[75] T. Hanaoka, A. Heilmann, M. Kroll, H.-P. Kormann, T. Sawitowski,
G. Schmid, P. Jutzi, A. Klipp, U. Kreibig, R. Neuendorf, Appl. Orga-
nomet. Chem. 1998, 12, 367–373.
[38] J. W. Couves, P. Meeham, Phys. B 1995, 208–209, 665–667.
[39] J. Turkevich, G. Kim, Science 1970, 169, 873–879.
[40] New Horizons in Catalysis: R. S. Miner, Jr., S. Namba, J. Turkevich,
Stud. Surf. Sci. Catal. 1981, 7, 160–167.
[
41] G. Schmid, A. Lehnert, J.-O. Malm, J.-O. Bovin, Angew. Chem.
1
991, 103, 852–854; Angew. Chem. Int. Ed. Engl. 1991, 8, 1015–
1
016.
[76] P. Braunstein, H.-P. Kormann, W. Meyer-Zaika, R. Pugin, G.
Schmid, Chem. Eur. J. 2000, 6, 4637–4646.
[77] S. Hitz, R. Prins, J. Catal. 1997, 168, 194–206.
[
[
[
[
[
[
42] M. Harada, K. Asakura, N. Toshima, J. Phys. Chem. 1993, 97, 5103–
114.
5
43] N. Toshima, Fine Particles Science and Technology (Ed.: E.Peliz-
zetti), Kluwer Academic, 1996, p. 371.
44] Y. Mizukoshi, K. Okitsu, Y. Maeda, T. A. Yamamoto, R. Oshima, Y.
Nagota, J. Phys. Chem.B 1997, 101, 7033–7037.
[78] Y. I. Lam, M. Boudart, J. Catal. 1977, 50, 530–540.
[79] M. P. A. Viegers, J. M. Trooster, Phys. Rev. B 1977, 15, 72–83.
[80] L. Steviano, S. Santucci, L. Lozzi, S. Colagero, F. E. Wagner, J. Non-
Cryst. Solids 1998, 232–234, 644–649.
[81] J. M. Tura, P. Regull, L. Victori, M. Dolors de Castellar, Surf. Inter-
face Anal. 1988, 11, 447–449.
[82] N. H. Turner, A. M. Single, Surf. Interface Anal. 1990, 15, 215–222.
[83] J. T. Payne, D. A. Reuschle, K. A. Mauritz, Polym. Prepr. 1997, 38,
247–248.
[84] P. Claus, A. Bruckner, C. Mohr, H. Hofmeister, J. Am. Chem. Soc.
2000, 122, 11430–11439.
45] G. Schmid, H. West, H. Mehles, A. Lehnert, Inorg. Chem. 1997, 36,
8
91–895.
46] H. Bçnnemann, W. Brijoux, A. Schulze Tilling, K. Siepen, Top.
Catal. 1998, 4, 217–227.
47] H. Bçnnemann, W. Brijoux, Advanced Catalysts and Nanostructured
Materials (Ed.: W. R. Moser), Academic Press, San Diego, 1996,
p. 165.
[
48] H. Bçnnemann, R. Brinkmann, W. Brijoux, E. Dinjus, T. Joussen, B.
Korall, Angew. Chem. 1991, 103, 1344–1346; Angew. Chem. Int. Ed.
Engl. 1991, 10, 1312–1314.
[85] C. Mohr, H. Hofmeister, P. Claus, J. Catal. 2003, 213, 86–94.
[86] J. E. Bailie, G. J. Hutchings, Chem. Commun. 1999, 2151–2152.
[87] C. Milone, R. Ingoglia, M. L. Tropeano, G. Neri, S. Galvagno, Chem.
Commun. 2003, 868–869.
[88] J. L. Margitfalvi, A. Tompos, I. Kolosova, J. Valyon, J. Catal. 1998,
174, 246–249.
[89] P. Gallezot, D. Richard, Catal. Rev. Sci. Eng. 1998, 40, 81–126.
[90] G. Lang, Nucl. Instr. Methods 1963, 24, 425–431.
[91] T. E. Cranshaw, J. Phys. E 1974, 7, 497–505.
[
49] Preparation of Catalysts IV: J. B. Michel, J. T. Schwartz, Stud. Surf.
Sci Catal. 1987, 31, 669–678.
[50] Y. Nakamura, H. Hirai, Chem. Lett. 1986, 1197–1200.
[51] H. Hirai, M. Ohtaki, M. Komiyama, Chem. Lett. 1986, 269–272.
[52] Y. Nakao, K. Kaeriyama, J. Colloid Interface Sci. 1989, 131, 186–
1
91.
[
[
53] Y. Wang, H. Liu, Y. Huang, Polym. Adv. Technol. 1996, 7, 634–639.
54] Q. Wang, H. Liu, Y. Wang, J. Colloid Interface Sci. 1997, 190, 380–
[92] G. K. Shenoy, B. D. Dunlap, Mossbauer Isomer Shifts (Eds.: G. K.
Shenoy, F. E. Wagner), North Holland, Amsterdam, 1978, p. 869.
[93] D. Massiot, F. Fayon, M. Capron, I. King, S. Le Calvꢄ, B. Alonso,
J. O. Durand, B. Bujoli, Z. Gan, G. Hoatson, Magn. Reson. Chem.
2002, 40, 70–76.
3
86.
[
[
55] Y. Wang, H. Liu, Y. Y. Jiang, Chem. Commun. 1998, 1878–1879.
56] W. Yu, M. Liu, H. Liu, X. An, Z. Liu, X. Ma, J. Mol. Catal. A 1999,
1
42, 201–211.
57] B. Corain, K. Jerabek, P. Centomo, P. Canton, Angew. Chem. 2004,
16, 977–980; Angew. Chem. Int. Ed. 2004, 43, 959–962.
[94] V. I. Parvulescu, V. Parvulescu, S. Kaliaguine, U. Endrushat, B.
Tesche, H. Bonnemann, in Catalysis of Organic Reactions (Ed.:
M. E. Ford), Marcel Dekker, 2001, p. 301.
[
[
[
1
58] B. Corain, C. Burato, P. Centomo, S. Lora, W. Meyer-Zaika, G.
Schmid, J. Mol. Catal. A 2005, 225, 189–195.
59] R. W. J. Scott, A. K. Datye, R. M. Crooks, J. Am. Chem. Soc. 2003,
[95] V. I. Pârvulescu, V. Pârvulescu, D. Macovei, L. Frunza, J. Chem. Soc.
Faraday Trans. 1997, 93, 1827–1835.
Received: August 9, 2005
1
25, 3708–3709.
Published online: December 27, 2005
Chem. Eur. J. 2006, 12, 2343 – 2357
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
www.chemeurj.org
2357