Propene epoxidation with dioxygen catalyzed by gold clusters

Article abstract of DOI:10.1002/anie.200903011

Slze matters: Gold clusters (<2.0 nm), but not gold nanoparticles, deposited on alkaline-treated titanosilicalite-1 allow O₂ and H ₂O to react to give hydroperoxides (-OOH). These transfer to neighboring Ti sites to form Ti-OOH (see

Full text of DOI:10.1002/anie.200903011

Communications
DOI: 10.1002/anie.200903011
Heterogeneous Catalysis
Propene Epoxidation with Dioxygen Catalyzed by Gold Clusters**
Jiahui Huang, Tomoki Akita, Jꢀrꢀmy Faye, Tadahiro Fujitani, Takashi Takei, and
Masatake Haruta*
There are two major shifts that the chemical industry is going
to take in the 21st century: a gradual shift to renewable
biomass resources and a shift to reactions which produce little
waste. After polymerization, the oxidation of hydrocarbons
represents the second largest contributor to the total product
value in the chemical industry (18%).[1] The products, func-
tional organic compounds containing oxygen, such as epox-
ides, ketones, aldehydes, alcohols, and acids, are used for
producing polymers, surfactants, detergents, cosmetics, and
other products. If the selective oxidation of hydrocarbons
could be efficiently carried out by using dioxygen (O2) or air
alone instead of using oxidants, such as hazardous chlorine or
costly organic peroxides, it would contribute to the conversion
of current chemical production into a green and sustainable
industry.
In particular, if the selective oxidation of propene (C3H6)
with O2 to propene epoxide (PO) was realized, it might spark
considerable technological innovation in the chemical indus-
try. Indeed, this reaction has been regarded as the Holy Grail
in catalysis research[2] because a huge energy of 497 kJmolÀ1
is needed to dissociate O2 (Figure S1 in the Supporting
Information)[3] and, once atomically dissociated, oxygen tends
to be negatively charged (OÀ), and preferentially attacks the
propylene glycol. Its annual worldwide production has grown
at a rate above 4.0% and amounted to approximately
7.5 million tons in 2007.[4] Propene epoxide is currently
produced by the chlorohydrin process and several organic
peroxide processes.[5] The chlorohydrin process is accompa-
nied by the by-production of CaCl2 (2.2 tons per 1.0 ton of
propene epoxide) together with toxic chlorinated organic
compounds (several hundreds grams per 1.0 ton propene
epoxide), whereas the organic peroxide processes usually
suffer from the mismatch of market demand for the products,
because the demand for the co-products, such as tert-butyl
alcohol and styrene, is less than for propene epoxide.
Recently two new processes have been developed:
a
cumene hydroperoxide process in 2003[6] in Japan where the
co-product, cumyl alcohol, is recycled by reduction with
hydrogen (H2); and a hydrogen peroxide (H2O2) epoxidation
process in 2008[7] in Belgium using methanol as a solvent.
However, these new industrial processes are still based on
multi-staged liquid-phase reactions and require hydrogen.
In contrast to the simple industrial process for the gas-
phase epoxidation of symmetrical ethene (C2H4) with dioxy-
gen over a silver (Ag) catalyst, the use of silver catalysts in
propene epoxidation with O2 usually resulted in propene
epoxide selectivities of less than 50%,[8,9] while gold (Au)
sponge or Au/SiO2 was reported to yield acrolein as a main
product but no propene epoxide.[10] An alternative route
which needs no substantial energy is the reductive activation
of O2 with H2 (5.9 kJmolÀ1, Figure S1 in the Supporting
Information) under milder conditions. This process can be
regarded as an in situ H2O2 synthesis process. One of us
reported in 1998 the gas-phase C3H6 epoxidation with O2 and
H2 mixture over gold nanoparticles (NPs, 2.0–5.0 nm) depos-
ited on anatase TiO2.[11] Since then, the catalytic performances
have been appreciably improved by optimizing catalyst
formulations.[12–14] For example, gold nanoparticles deposited
on mesoporous titanium-silicate present a propene epoxide
selectivity of 91% at a C3H6 conversion of 8.5%.[12]
À
weakly bound allylic hydrogen atoms ( CH3) in propene
resulting in the production of acrolein (Figure S2 in the
Supporting Information).
Propene epoxide is an important bulk chemical which is
used for producing mainly polyurethane foams and resins, and
[*] Dr. J. Huang, Prof. Dr. T. Takei, Prof. Dr. M. Haruta
Graduate School of Urban Environmental Sciences
Tokyo Metropolitan University (TMU)
1-1 Minami-osawa, Hachioji, Tokyo 192-0397 (Japan)
Fax: (+81)42-677-2851
E-mail: haruta-masatake@center.tmu.ac.jp
Dr. T. Akita
Research Institute for Ubiquitous Energy Devices, National Institute
of Advanced Industrial Science and Technology (AIST)
1-8-31 Midorigaoka, Ikeda, Osaka 563-8577 (Japan)
Herein, gold clusters with diameters below 2.0 nm were
deposited on alkaline-treated titanosilicalite-1 (TS-1) by solid
grinding (SG) with dimethyl gold(III) acetylacetonate.[15]
These gold-cluster catalysts exhibited, without any promoters,
high stability (for more than 24 h) as well as high catalytic
activity (C3H6 conversion about 8.8%) and propene epoxide
selectivities (about 82%) in C3H6 epoxidation with an O2 +
H2 mixture. Over 0.19 wt% Au/TS-1-K1 (herein alkaline-
treated TS-1 is denoted as TS-1-K1, where K stands for KOH
and 1 stands for 1.0 h, see Experimental Section for the
preparation of TS-1-K1) with a mean gold diameter of 1.8 nm,
a small amount of propene epoxide is produced even from a
mixture of C3H6 and O2 (Table S1 in the Supporting
Dr. J. Faye, Dr. T. Fujitani
Research Institute for Innovation in Sustainable Chemistry, AIST
16-1 Onogawa, Tsukuba, Ibaraki 305-8569 (Japan)
Dr. J. Huang, Dr. T. Akita, Dr. T. Fujitani, Prof. Dr. T. Takei,
Prof. Dr. M. Haruta
Japan Science and Technology Agency (JST), CREST
4-1-8 Hon-cho, Kawaguchi, Saitama 332-0012 (Japan)
[**] We are grateful to Prof. S. T. Oyama of Virginia Polytechnic Institute
and State University for his continuous cooperation, critical and
constructive discussion, and manuscript refinement.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200903011.
7862
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 7862 –7866

Angewandte
Chemie
Information). The addition of H2O vapor to the feed gas
markedly enhanced the catalytic performance to a C3H6
conversion of 0.88% and a propene epoxide selectivity of
52% (Table 1). In contrast, over 0.19 wt% Au/TS-1 (without
alkaline treatment) prepared by the solid-grinding method,
only large gold particles were observed with a mean diameter
of 4.6 nm and no propene epoxide was formed even when
H2O vapor was introduced to the feed gas (Table 1). Since
elemental analyses showed that both gold catalysts had
identical actual gold loading and Ti content, and a very
similar potassium content (Table S2 in the Supporting Infor-
mation), the main difference in propene epoxide production
could be ascribed to the difference in the diameter of gold
particles.
We further used TS-1 treated with NaOH or CsOH
instead of KOH as a support and deposited gold clusters over
them by solid grinding or deposition–precipitation (DP).
These gold/alkaline treated TS-1 could also catalyze C3H6
epoxidation with O2 in the presence of H2O, while catalytic
activity, selectivity, or stability was slightly inferior to that
over 0.19 wt% Au/TS-1-K1.
particles (> 2.0 nm) on TiO2 (Table 1), indicating that isolated
Ti sites are indispensable for gold clusters (< 2.0 nm) to
catalyze C3H6 epoxidation with O2 + H2O.
Lambert et al. have recently reported that even on inert
support materials, such as silicon dioxide (SiO2) and boron
nitride (BN), the well-defined size control of gold clusters
leads to the epoxidation of styrene with selectivity around
30%.[18] The diameter of gold particles is tunable by choosing
the calcination temperature for the gold catalysts according to
theoretical calculations[19] and experimental results.[20] After
propene oxidation reaction, the gold catalysts were examined
with HAADF-STEM. Different mean gold nanoparticle
diameters of 1.8 nm, 2.0 nm, and 3.5 nm were obtained by
controlled calcination at different temperatures (423 K,
573 K, and 673 K, respectively; Table S3 in the Supporting
Information). HAADF-STEM images showed that before
reaction, 92% of gold particles over 0.19 wt% Au/TS-1-K1
calcined at 423 K (with mean gold particle diameter of
1.6 nm) existed as clusters smaller than 2.0 nm and only a
small number of gold nanoparticles were observed between
2.0–4.5 nm (Figure 1a,c). After reaction at 473 K (Catalyst 1,
Table S3 in the Supporting Information), although the mean
diameter of the gold species slightly increased to 1.8 nm and
some gold nanoparticles larger than 5.0 nm appeared, 87% of
gold particles still existed as clusters (< 2.0 nm; Figure 1b,d).
After calcination of 0.19 wt% Au/TS-1-K1 at 573 K (Cata-
lyst 3, Table S3 in the Supporting Information), the mean
diameter of gold particles increased to 2.0 nm, while 70% of
them were still present as clusters. The catalytic performance
of Catalyst 3 was depressed only slightly in comparison with
Catalyst 1 (Table S3 in the Supporting Information). How-
ever, when calcined at 673 K (Catalyst 4, Table S3 in the
Supporting Information), 86% of the gold particles grew to
be nanoparticles (> 2.0 nm) with mean diameter of 3.5 nm,
and only small amounts of gold clusters remained. The fact
that Catalyst 4 produced only a little propene epoxide
indicates that gold clusters but not gold nanoparticles were
responsible for propene epoxide synthesis. Over Catalyst 5
calcined at 773 K (Table S3 in the Supporting Information),
no propene epoxide was produced.
In C3H6 epoxidation with O2 alone, two key conditions
appear to be of critical importance: isolated Ti sites and gold
clusters smaller than 2.0 nm. To find out whether isolated Ti
sites are needed or not, g-Al2O3 (specific surface area (SA):
178 m2 gÀ1), amorphous Al2O3 (SA: 193 m2 gÀ1), TiO2 (P-25,
SA: 46 m2 gÀ1), and nanoporous carbon (NPC, SA: 577 m2 gÀ1)
were selected as supports. Except for TiO2, on which the DP
method could deposit smaller gold nanoparticles and even
gold clusters, the SG method was used to deposit gold as
nanoparticles and clusters. To confirm the presence of gold
clusters (< 2.0 nm), high-angle annular dark-field scanning
transmission electron microscopy (HAADF-STEM) obser-
vation was also carried out. The mean diameter of the gold
particles ranged from 1.8 to 2.9 nm. It has recently been
reported that Au6–10 clusters supported on amorphous
[16]
[17]
Al2O3
and gold nanoparticles supported on TiO2
can
catalyze epoxidation of C3H6 with O2 + H2O. In our experi-
ments no propene epoxide but acrolein and/or CO2 were
produced over gold clusters smaller than 2.0 nm in diameter
on titanium-free supports, on TiO2, and over gold nano-
Table 1: C3H6 reaction with O2 in the presence of H2O over supported gold catalysts.[a]
Support
Au
Prep.
method
Pretreatment
atmosphere,[c]
temperature [K]
Mean Au
Conv. of
C3H6
[%]
Selectivity [%]
acrolein
loading
particle diameter
propene
epoxide
CO2
[wt%][b]
[nm][d]
titano-
silicalite
TS-1
TS-1-K1
TS-1-K1
g-Al2O3
Al2O3 (amorphous)
TiO2
TiO2
0.19
0.19
9.0[e]
0.86
0.19
1.2
SG
SG
DP
SG
SG
DP
DP
SG
H2, 423
H2, 423
H2, 473
air, 573
H2, 423
air, 573
air, 573
air, 573
4.6
1.8
–
0.07
0.88
0.47
0.15
0.15
0.03
0.05
0.48
0
52
57
0
0
0
34
12
7
41
54
0
66
36
36
44
metal
oxides
2.6
<2.0[f]
2.9[f]
1.8[11]
1.9
46
100
100
59
0.10
1.0
0
0
0
41
carbon
nanoporous carbon
[a] Reaction conÀd1itions: catalyst 0.30 g; reaction temperature 473 K (for Au/TiO2, 353 K); feed gas C3H6/O2/H2O/Ar=10/10/2/78; space velocity
4000 mLgÀ1cat. h . Data were taken under a steady state after at least 2.0 h duration. [b] Actual Au loadings obtained by inductively coupled plasma
(ICP) analysis. [c] H2 (10 vol%) was diluted by 90 vol% Ar. [d] The diameters of gold nanoparticles or clusters determined by HAADF-STEM
observation after catalytic tests. [e] Gold loading calculated from the gold content in the starting solution. [f] The diameters of gold nanoparticles or
clusters determined by TEM observation after catalytic tests. Over Au/amorphous Al2O3, because no Au particles could be found by TEM, the diameter
was estimated to be smaller than 2.0 nm. SG=solid grinding, DP=deposition–precipitation.
Angew. Chem. Int. Ed. 2009, 48, 7862 –7866
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7863

Communications
Figure 1. a,b) HAADF-STEM images and c,d) diameter distribution of
gold particles (c: 535, d: 436 counts) for 0.19 wt% Au/TS-1-K1. (a,c):
before reaction of samples reduced in 10 vol% H2 balanced with Ar at
423 K; (b,d): after reaction for 21 h of samples reduced under the
same conditions as in (a,c). Reaction conditions: catalyst 0.30 g;
reaction temperature 473 K; feÀe1d gas C3H6/O2/H2O/Ar=10/10/2/78;
space velocity 4000 mLgÀ1cat. h (scale bar: 10 nm).
Figure 2. Enhancing effect of H2O on propene epoxidation with O2
over 0.10 wt% Au/TS-1-K1. Reaction conditions: catalyst 0.30 g; reac-
tion temperature 473 K; feed gas C3H6/O2/Ar=10/10À/180 or C3H6/O2/
H2O/Ar=10/10/2/78; space velocity 4000 mLgÀ1cat. h . Step 1: with-
out H2O; Step 2: H2O was added to the feed gas; Step 3: the addition
of H2O was stopped; Step 4: H2O was added to the feed gas again.
To investigate the promoting effect of H2O, Au/TS-1-K1
with a lower gold loading of 0.10 wt% was used, because H2O
can be produced in situ if C3H6 is oxidized to acrolein or CO2.
No propene epoxide was produced when only O2 was supplied
to the feed gas of C3H6 in Figure 2, Step 1. The addition of
H2O to the feed gas enhanced the formation of propene
epoxide to a selectivity of about 47% as well as the
conversion of C3H6 (Figure 2, Step 2). Since H2O was not
consumed and no product from H2O decomposition, such as
H2, was detected, H2O can be regarded as a gas-phase
promoter. When H2O was removed from the feed gas,
propene epoxide selectivity decreased quickly from 45% to
only 20% with an appreciably depressed C3H6 conversion of
about 0.15% (Figure 2, Step 3). The small amount of propene
epoxide formed can be ascribed to H2O physically adsorbed
in Step 2 and to H2O produced in situ. When H2O was
introduced to the feed gas again (Figure 2, Step 4), both
propene epoxide selectivity and C3H6 conversion were
recovered to the levels of Step 2 (Figure 2).
To answer the question of why H2O can enhance C3H6
epoxidation with O2 over Au/TS-1-K1 (Figure 2), in situ UV/
Vis spectroscopy was used to monitor this reaction. Before
in situ monitoring, UV/Vis spectra of TS-1-K1 were obtained
after contact with liquid H2O or aqueous H2O2 (Figure S3 in
the Supporting Information). The spectrum after contact with
liquid H2O displayed a shoulder peak centered at 265 nm,[21]
and the spectrum after contact with aqueous H2O2 showed a
broad band at 300–450 nm.[22] The difference spectra obtained
in a stream of O2 and H2O (Figure S4 in the Supporting
Information) showed that a band at 255 nm ascribed to H2O
adsorption and a band at 333 nm directly related to Ti-OOH
species[23] appeared, indicating the formation of hydroper-
oxides from O2 and H2O. As shown in Figure 3, the intensity
of this band gradually increased with time on stream until
about 100 min and then leveled off. The addition of C3H6 to
the feed gas resulted in a gradual decrease in the intensity of
this band, suggesting that Ti-OOH species were consumed by
C3H6 and this happened more quickly than the formation of
Ti-OOH species. Once water was removed from the feed gas,
the band intensity of the Ti-OOH species quickly decreased,
which strongly indicates that H2O was indeed required for the
formation of the Ti-OOH species.
As described above, Ti-OOH species can be formed over
Au/TS-1-K1 directly from O2 + H2O. However, if only TS-1-
K1 without gold clusters was monitored by in situ UV/Vis
spectroscopy, no Ti-OOH species could be detected, meaning
that gold clusters are indispensable for the formation of
hydroperoxides (-OOH species). Theoretical calculations also
confirmed that over gold clusters, hydroperoxides can be
produced directly from O2 + H2O.[24] Scheme 1 illustrates a
possible epoxidation route. First, O2 reacts with H2O over
gold cluster surfaces to produce -OOH species. Then -OOH
species reversibly transfer from gold cluster surfaces to
neighboring Ti sites to form Ti-OOH species, which are
responsible for subsequent C3H6 epoxidation. It is likely that
during the reaction OC radicals would be produced and further
7864 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 7862 –7866

Angewandte
Chemie
Experimental Section
Alkaline treatment of TS-1: Aqueous solution of 1.0m KOH (or
NaOH or CsOH) was added dropwise to the suspension of TS-1
(1.0 g) in H2O (100 mL) at 303 K under vigorous stirring until the
pH value of the suspension reached 12, it was then kept at 12 for 1.0 h
by adding a small amount of aqueous KOH dropwise. The suspension
was collected by filtration, washed five times with 2000 mL of H2O to
remove residual KOH and was dried at 373 K overnight. The alkaline-
treated TS-1 was denoted as TS-1-K1, where K stands for KOH and 1
stands for 1.0 h.
Deposition of gold nanoparticles and clusters by solid grinding
(SG): TS-1, alkaline-treated TS-1, g-Al2O3, amorphous Al2O3, and
nanoporous carbon (NPC) were used as supports. The powder of TS-
1, TS-1-K1, or amorphous Al2O3, and dimethyl gold(III) acetylaceto-
nate [Me2Au(acac)] which has a vapor pressure of 1.1 Pa at 298 K
were ground in an agate mortar in air for 15 min at room temperature,
followed by reduction in a stream of 10 vol% H2 in Ar at 423 K for
1.0 h. With g-Al2O3 and NPC as supports, the supports and [Me2Au-
(acac)] were ground by ball milling. [Me2Au(acac)], g-Al2O3, and
acetone were ground (350 rpm) at room temperature for 1.0 h, and
grinding of NPC was performed (300 rpm) for 5.0 h in the absence of
solvent. Final treatment of both gold catalysts was carried out by
calcination in air at 573 K for 4.0 h.
Spectroscopic measurements: In situ UV/Vis diffuse reflectance
spectra were collected using a large-compartment spectrometer
(Varian Cary 5000). About 40 mg of 0.19 wt% Au/TS-1-K1 (reduced
in 10 vol% H2 in Ar at 473 K) was loaded in a reaction chamber
(Harrick Scientific, model HVC-DRP) provided with a praying
mantis diffuse reflectance attachment (DRP-XXX). Spectra were
recorded at 473 K under atmospheric pressure in a stream of the
reactant gases at 20 mLminÀ1. Spectra were taken in the range of 200–
800 nm at a scan rate of 600 nmminÀ1, an averaging time of 0.1 s, and a
data interval of 1 nm. Spectra were referenced to 0.19 wt% Au/TS-1-
K1 treated at 473 K under Ar just before introduction of O2 + H2O
mixture (difference spectra). The presented spectra were obtained by
applying the Kubelka-Munk function, F(R), to the collected absorb-
ance data.
Figure 3. Intensity of in situ UV/Vis peak at 333 nm (assigned to
Ti-OOH) under different feed-gas compositions over 0.19 wt%
Au/TS-1-K1. After the catalyst sample was kept in Ar at 473 K for 1.0 h
(reference spectrum was taken), the feed gas was switched to
O2 +H2O mixture (O2/H2O/Ar=10/2/88), and then after 2.0 h, C3H6
was added (C3H6/O2/H2O/Ar=10/10/2/78). After another 2.0 h, H2O
was removed from the feed gas. Reaction conditions: catalyst about
40 mg; reaction temperatÀu1re 473 K; rate of flow 20 mLminÀ1; space
velocity 30000 mLgÀ1cat. h
.
Received: June 4, 2009
Published online: September 15, 2009
Scheme 1. A possible route for C3H6 epoxidation with O2 +H2O over
Au/TS-1-K1. PO=propene epoxide.
Keywords: clusters · epoxidation · gold ·
heterogeneous catalysis · propene
.
used to oxidize C3H6 to propene epoxide, acrolein, and CO2.
Accordingly, propene epoxide selectivity can be improved by
controlling the reactivity of OC radicals over the surfaces of
gold clusters.
[1] A. M. Thayer, Chem. Eng. News 1992, 70(10), 27.
[2] M. McCoy, Chem. Eng. News 2001, 79(43), 19.
[3] M. Haruta, Nature 2005, 437, 1098.
[4] F. Nees, Heterogeneous catalysis at BASF (available at http://
basf.com/group/corporate/en/innovations/events-presentations/
catalysis/papers, 2008).
In conclusion, the gas-phase epoxidation of propene with
dioxygen can be catalyzed, in the presence of H2O, by gold
clusters (< 2.0 nm) deposited on alkaline treated TS-1. The
À1
À1
metal time yield of propene epoxide was 2.3 g
g
h
Au
with
PO
[5] S. T. Oyama, Mechanisms in Homogeneous and Heterogeneous
Epoxidation Catalysis, Elsevier B.V., Amsterdam, 2008.
[6] M. Ishino, J. Yamamoto, Shokubai 2006, 48, 511.
[7] A. Tullo, Chem. Eng. News 2004, 82(36), 15.
[8] J. Lu, J. J. Bravo-Suꢀrez, M. Haruta, S. T. Oyama, Appl. Catal. A
2006, 302, 283.
propene conversion near 1.0% and propene epoxide selec-
tivity above 50%. Although this performance is not suffi-
ciently high for an industrial application, it is better than those
of copper- and silver-based catalysts reported to date
(Figure S5 and Table S4 in the Supporting Information).
Therefore, our gold-based catalysts have greater potential for
future improvement if the size of the gold clusters, the
configuration of the gold clusters and isolated Ti sites, and the
reactivity of the active OC radicals towards propene epoxide
can be optimized. Accordingly, supported gold clusters may
be a significant stepping stone towards the development of
environmentally friendly chemical processes.
[9] G. Jin, G. Lu, Y. Guo, Y. Guo, J. Wang, W. Kong, X. Liu, J. Mol.
Catal. A 2005, 232, 165.
[10] N. W. Cant, W. K. Hall, J. Phys. Chem. 1971, 75, 2914.
[11] T. Hayashi, K. Tanaka, M. Haruta, J. Catal. 1998, 178, 566.
[12] B. Chowdhury, J. J. Bravo-Suꢀrez, M. Datꢁ, S. Tsubota, M.
Haruta, Angew. Chem. 2006, 118, 426; Angew. Chem. Int. Ed.
2006, 45, 412.
[13] L. Cumaranatunge, W. N. Delgass, J. Catal. 2005, 232, 38.
Angew. Chem. Int. Ed. 2009, 48, 7862 –7866
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7865

Communications
[14] T. A. Nijhuis, T. Visser, B. M. Weckhuysen, Angew. Chem. 2005,
[19] C. L. Cleveland, W. D. Luedtke, U. Landman, Phys. Rev. B 1999,
60, 5065.
[20] T. Castro, R. Reifenberger, E. Choi, R. P. Andres, Phys. Rev. B
1990, 42, 8548.
117, 1139; Angew. Chem. Int. Ed. 2005, 44, 1115.
[15] M. Ishida, M. Nagaoka, T. Akita, M. Haruta, Chem. Eur. J. 2008,
14, 8456.
[16] S. Lee, L. M. Molina, M. Lꢂpez, J. A. Alonso, B. Hammer, B.
Lee, S. Seifert, R. E. Winans, J. W. Elam, M. J. Pellin, S. Vajda,
Angew. Chem. 2009, 121, 1495; Angew. Chem. Int. Ed. 2009, 48,
1467.
[17] M. Ojeda, E. Iglesia, Chem. Commun. 2009, 352.
[18] M. Turner, V. B. Golovko, O. P. H. Vaughan, P. Abdulkin, A.
Berenguer-Murcia, M. S. Tikhov, B. F. G. Johnson, R. M. Lam-
bert, Nature 2008, 454, 981.
[21] M. C. Capel-Sanchez, J. M. Campos-Martin, J. L. G. Fierro, J.
Catal. 2003, 217, 195.
[22] S. C. Laha, R. Kumar, J. Catal. 2002, 208, 339.
[23] J. J. Bravo-Suꢀrez, K. K. Bando, J. Lu, M. Haruta, T. Fujitani,
S. T. Oyama, J. Phys. Chem. C 2008, 112, 1115.
[24] A. Bongiorno, U. Landman, Phys. Rev. Lett. 2005, 95, 106102.
7866 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 7862 –7866