Highly efficient aerobic oxidation of alkenes over unsupported nanogold

Article abstract of DOI:10.1039/c0cc00664e

An octylsilane-stabilized colloidal dispersion of 2 nm crystalline gold nanoparticles is highly active and selective for the aerobic oxidations of stilbene and cyclohexene in methylcyclohexane.

Full text of DOI:10.1039/c0cc00664e

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Highly efficient aerobic oxidation of alkenes over unsupported nanogoldw
Malika Boualleg,a Kevin Guillois,b Bertrand Istria,b Laurence Burel,b
a
a
a
b
´
´
Laurent Veyre, Jean-Marie Basset,z Chloe Thieuleux* and Valerie Capsz*
Received 30th March 2010, Accepted 12th May 2010
First published as an Advance Article on the web 18th June 2010
DOI: 10.1039/c0cc00664e
An octylsilane-stabilized colloidal dispersion of 2 nm crystalline
gold nanoparticles is highly active and selective for the aerobic
oxidations of stilbene and cyclohexene in methylcyclohexane.
to be of particular interest since it can stabilize nanoparticles
of Ru or Pt with formation of metal colloids7,8 which, after
homogeneous dispersion within a silica matrix, showed no loss
of catalytic activity in hydrogenation of propene compared to
that of a reference catalyst made of supported ‘‘naked’’
nanoparticles.8 The concept of controlling the size of the
nanoparticles, via the addition of a silane ligand during the
process of crystal growth, was here successfully transferred to
gold due to the known affinity between alkylsilane groups and
Au surface atoms.9 In a typical procedure, 5 mL of an aqueous
solution of hydrogen tetrachloroaurate (49 mg, 0.14 mmol)
were mixed with 15 mL of a tetraoctylammonium bromide
solution in toluene (368 mg, 0.68 mmol). The two-phase
mixture was vigorously stirred until all the tetrachloroaurate
was transferred into the organic phase. 10 equivalents of the
stabilizing silane ligand (n-octylsilane, 20 mg, 1.4 mmol) were
further added to the reaction mixture. 5 mL of an aqueous
solution of sodium borohydride (64 mg, 1.7 mmol) were slowly
introduced during 10 min under stirring. After further stirring
for 3 h, the organic phase was separated and evaporated under
vacuum (10ꢀ3 mbar). For comparison with other gold colloids,
a similar procedurew was followed by using 10 equivalents of
octylthiol or no stabilizing agent other than the phase transfer
agent. Several purification steps (decantation, washings and
filtration) were required to yield small clean gold nanoparticles
(30% yield). After redispersion in MCH, the final colloidal
solutions are denoted by Au-SiC8, Au-SC8 and Au-N+(C8)4
when stabilized with n-octylsilane, n-octylthiol and tetraoctyl-
ammonium ligands, respectively.
Gold nanoparticles (NPs) have been shown to catalyze a
variety of oxidation reactions both in the gas and liquid
phase.1 In gas phase oxidations (e.g. CO oxidation), their
unique activity at low temperature critically relies on their
association with a reducible oxide support (e.g. TiO2) which is
assumed to trigger oxygen activation.2 In the aqueous phase,
unsupported gold sols can readily oxidize glucose with molecular
oxygen;3 however, the use of an activated carbon support is
essential to prevent particle aggregation and subsequent deacti-
vation of the catalyst. Recently, we have shown that supported
gold nanoparticles could selectively epoxidize trans-stilbene
with molecular oxygen, using methylcyclohexane (MCH) as
solvent and tert-butyl hydroperoxide (TBHP) as radical
initiator.4 In this particularly apolar medium, kinetics could
be drastically improved by designing a hydrophobic support.5
Here we show that both the rate of trans-stilbene aerobic
oxidation and the selectivity to trans-stilbene oxide can be
further improved by using a colloidal solution of 2 nm crystal-
line gold NPs stabilized by organo-silane ligands. Comparison
with a reference supported gold catalyst and other gold
colloids shows that, beyond improved mass-transfers, the
superiority of this system relies on the specific behavior of
the protecting silane precursor ligand under oxidative con-
ditions. It allows the resulting unsupported colloid to reach
full stilbene conversion and be the most active and selective
gold catalyst to date for selective aerobic oxidations of both
trans-stilbene and cyclohexene.
The gold loadings of these colloidal solutions (Table 1) vary
between 0.04 wt% for the solution stabilized by tetraoctyl-
ammonium ligands and 0.38 wt% for the solution containing
octylthiol ligands. In all cases, the average gold particle sizes as
determined by transmission electron microscopy (TEM)w are
all very similar, at about 2 nm, in a range traditionally
considered as active for catalysis (o10 nm, Table 1).2 The
narrowest size distribution is obtained for octylthiol-stabilized
solution while the colloidal solution containing the ammonium
The preparation of the colloidal solution of Au NPs
was achieved by reduction of a molecular complex of Au(III)
in a classical biphasic system6 using a tetraalkylammonium
bromide salt as phase transfer agent and n-octylsilane as
stabilizing ligand. This surface ligand has already been found
a
´
Universite de Lyon, Institut de Chimie de Lyon, UMR 5265
CNRS-CPE-UCBL (LC2P2), Equipe LCOMS, CPE Lyon,
43 bd du 11 Novembre 1918, 69616 Villeurbanne, France.
E-mail: thieuleux@cpe.fr; Tel: +33 4 72 43 27 95
Table 1 Gold loadings (ICP-OES elemental analysis) and average
gold particle diameters (size distributions obtained from TEM
analysis)w
b
´
Universite de Lyon, Institut de Chimie de Lyon, UMR CNRS 5256
(IRCELYON), 2 Avenue Albert Einstein, 69626 Villeurbanne,
France
w Electronic supplementary information (ESI) available: Experimental
protocols for characterizations and catalytic evaluations, TEM, FTIR,
XPS and catalytic data. See DOI: 10.1039/c0cc00664e
Average particle size
(standard deviation)/nm
o10 nm
410 nm
Catalyst
Gold loading (wt%)
z Present address: KAUST Catalysis Center (KCC), 4700 King
Abdullah University of Science and Technology, Thuwal 23955-
6900, Kingdom of Saudi Arabia. E-mail: valerie.caps@kaust.edu.sa;
Tel: +966 2 808 0304.
Au-SiC8
Au-SC8
0.13
0.38
0.04
2.1 (1.0)
1.8 (0.4)
2.1 (1.0)
—
—
Au-N+(C8)4
27 (10, 28%)
ꢁc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 5361–5363 | 5361

View Article Online
ligand also exhibits a population of particles with diameters
larger than 10 nm, hence showing a bimodal distribution of
sizesw with a second maximum at 27 nm. These results are
consistent with the relative strengths of the gold–ligand bond,
the reactants. As expected though, the tetraoctylammonium-
stabilized colloids rapidly deactivate: while 25% stilbene is
converted within the first 4 h of reaction at 80 1C, less than 5%
of the substrate is converted in the following 4 h, stilbene
conversion reaching a maximum of 50% in 72 h. This is
attributed to consumption of the active phase (2 nm gold
NPs) by agglomeration of the poorly-protected Au-N+(C8)4
under reaction conditions.3 On the other hand, the initial
stilbene oxidation rate observed over Au-SC8 is 4 to 5 times
lower but stilbene conversion steadily reaches 90% in 72 h.
This is related to the stronger bond between Au and S,9 which,
via steric effects, might hinder access to the active site while at
the same time maintaining the amount of 2 nm Au NP active
phase by preventing easy coalescence of the gold NPs. Yet, the
best catalytic results are obtained with unsupported Au-SiC8,
which achieves not only full stilbene conversion in 72 h (90%
conversion in 24 h), with an initial reaction rate one order of
magnitude higher than that observed over the reference cata-
lyst, but also the best epoxide selectivity ever reported for this
reaction (Fig. 2) of 90% at full conversion.
which is expected to increase in the following order: Au–N+
o
Au–Si o Au–S.9 In particular, the high population of gold
particles with diameters above 10 nm observed in Au-N+(C8)4
and the low gold content of this solution highlight the poor
stabilizing properties of the tetraoctylammonium ligand used
as the phase transfer agent. On the other hand, the strong
affinity between gold and sulfur is illustrated by the highest
gold content, lowest average particle size and narrowest size
distribution of Au-SC8.
Equimolecular amounts (2.1 ꢂ 0.1 mmol of Au) of differently
concentrated methylcyclohexane solutions (B300 mg for
Au-SiC8, B103 mg for Au-SC8, B965 mg for Au-N+(C8)4)
are mixed with MCH (solvent, 20 mL/155 mmol) and the alkene
(substrate, 1 mmol). They are stirred (900 rpm) for catalytic
evaluation in air (1 atm) at 70–80 1C, in the presence of TBHP
(0.05 mmol, i.e. 5 mol%). Alkene conversion, product yields and
selectivities are then obtained by analyzing the reaction mixtures
with a Shimadzu GC-2014 gas chromatograph.w
All gold sols exhibit significant activity for the aerobic
epoxidation of trans-stilbene (Fig. 1). The epoxide yields
achieved (40–90%) are one order of magnitude higher than
the value expected from stoichiometric reaction with TBHP
(5 mol%), indicating that unsupported gold nanoparticles can
activate molecular oxygen from the air in the liquid phase,
without the assistance of any inorganic support. Actually, the
initial reaction rates observed over the colloidal solutions
are all higher than that obtained with the supported cata-
lyst, the Au/TiO2 reference from the World Gold Council
(Au/TiO2-WGC, 1.5 wt% Au, 3.5 ꢂ 0.9 nm).10 This can be
attributed (i) to the better mass transfer of the solubilized gold
crystallites and (ii) to the higher dispersion of gold atoms in
the gold colloids (60% vs. 36% in the reference catalystw); it
also shows that the ‘protected’ Au NPs are still accessible to
The octylsilane-stabilized colloidal solution of gold NPs was
thus investigated in more detail. Its hydrophobic character was
highlighted by injection of an aliquot of the colloidal solution
in a water/heptane mixture. This test led to the subsequent
pink coloration of the heptane phase exclusively,w indicating
that the alkylsilane ligands are still linked to the surface of the
gold particles, conferring on them their enhanced solubility in
the apolar catalytic reaction medium. The interaction between
gold NPs and silane ligands is supported by infra-red (IR)
spectroscopy. The IR spectrum of Au-SiC8 w indeed exhibits
all the bands present in the IR spectrum of free octylsilane w
between 2800–3000 cmꢀ1
,
characteristic of stretching
vibrations of –CH2 (nasym and nsym at 2925 ꢂ 10 cmꢀ1 and
2855 ꢂ 10 cmꢀ1 respectively) and –CH3 (nasym and nsym at
2960 ꢂ 10 cmꢀ1 and 2870 ꢂ 10 cmꢀ1 respectively). The nasym
deformation band of –CH3 can also be seen in both spectra
at 1463 cmꢀ1. Furthermore, the absence of the band at
2148 cmꢀ1 in the spectrum of Au-SiC8, characteristic of the
nSi–H vibration in the free silane ligand, is consistent with the
existence of RSi(n-C8H17) fragments grafted on the metal
Fig. 1 trans-stilbene conversion (squares) and epoxide yield (circles)
observed over Au-SiC8 (red), Au-SC8 (green), Au-N+(C8)4 (yellow)
and Au/TiO2-WGC (grey). Lines are merely a guide to the eye. Reaction
conditions: trans-stilbene (1 mmol), methylcyclohexane (20 mL),
catalyst (2.1 ꢂ 0.1 mmol Au), TBHP (0.05 mmol), air (1 atm), 80 1C,
900 rpm.
Fig. 2 Epoxide selectivity as a function of trans-stilbene conver-
sion over Au-SiC8 (red), Au-SC8 (green), Au-N+(C8)4 (yellow) and
Au/TiO2-WGC (grey).
ꢁc
This journal is The Royal Society of Chemistry 2010
5362 | Chem. Commun., 2010, 46, 5361–5363

View Article Online
surface, via loss of the three hydrogen atoms upon grafting of
the silane ligand.
However, XPS spectra of dried Au-SiC8 w stored in air show
a Si 2p3/2 core level binding energy (BE) at 102.2 eV, which is
significantly higher than the value found in silanes (o101 eV).11
It is consistent with the presence of siloxane, resulting from
oxidation of the silicon head-groups.12 Oxygenated species
resulting from RSiR oxidation have indeed been reported to
readily polymerize in a process catalyzed by gold.13 The
negative shift (ꢀ0.7 eV) in the Au 4f7/2 core level binding
energy of Au-SiC8 w (83.3 eV vs. 84.0 eV in bulk metallic gold)
suggests that these siloxane moieties are in close interaction
with gold NPs, making them electron-rich, through a charge
transfer from the polymer to the gold NP, similar to that in
Au/SiO2.5
Fig. 3 Cyclohexen-1-one selectivity as a function of cyclohexene
conversion over Au-SiC8 (red), Au-SC8 (green), Au-N+(C8)4 (yellow)
and Au/TiO2-WGC (grey). Reaction conditions: cyclohexene (1 mmol),
methylcyclohexane (20 mL), catalyst (2.0 mmol Au), TBHP (0.05 mmol),
air (1 atm), 70 1C, 900 rpm.
Hence the superiority of Au-SiC8 in aerobic oxidation
catalysis is attributed to the specific behavior of the protecting
agent towards oxidizing conditions. In addition to engineering
an optimized Au–Si interaction/coverage, it is possible that
‘‘surface’’ polymerization of the oxidized ligand into siloxane
provides active species in the immediate proximity of the Au
NP, which can trigger oxygen activation in the liquid phase,
according to the proposed radical mechanism,4 throughout the
reaction. Promotion of Au NP activity for low temperature
gas phase CO oxidation by nanooxides in close proximity to
gold ensuring oxygen activation, has also been described in
supported and unsupported Au–MOx nanocomposites.14
The superiority of Au-SiC8 is also evidenced in the selective
aerobic oxidation of cyclohexene, although to a lesser extent,
possibly due in part to the absence of water in this system.13b w
In this reaction, the presence of allylic hydrogen atoms in the
substrate molecule leads to allylic oxidation products, mainly
cyclohexen-1-one and cyclohexen-1-ol, which is consistent
with the radical mechanism proposed for these gold-catalyzed
oxidations.4 Although the overall profiles of the conversions
and yields vs. time w are lower and less distinctly different than
in the oxidation of stilbene, with maximum conversions at
72 h all being between 40 and 60%, the initial mass-specific
production rate of the main reaction product cyclohexen-1-one
obtained on Au-SiC8 is twice as high as that observed on the
Au/TiO2-WGC reference and 10 times higher than on Au-SC8.
Furthermore, the selectivity to cyclohexen-1-one is superior to
that observed with the other systems, over the whole range of
conversion (Fig. 3), confirming the superiority of Au-SiC8 for
selective aerobic oxidations.
Therefore, octylsilane-stabilized colloidal solutions of gold
NPs are expected to be highly efficient catalysts for liquid
phase radical oxidations in general.
The authors thank the French Research Agency for funding
project No. ANR-08-JCJC-0090-01 (ACTOGREEN), F. Morfin
for GC technical support, N. Cristin and P. Mascunan for
elemental analyses, P. Delichere for XPS.
Notes and references
1 C. Della Pina, E. Falletta, L. Prati and M. Rossi, Chem. Soc. Rev.,
2008, 37, 2077.
2 M. Haruta, Gold Bull., 2004, 37, 27 and references therein.
3 M. Comotti, C. Della Pina, R. Matarrese and M. Rossi, Angew.
Chem., Int. Ed., 2004, 43, 5812.
4 P. Lignier, F. Morfin, S. Mangematin, L. Massin, J.-L. Rousset
and V. Caps, Chem. Commun., 2007, 186; P. Lignier, F. Morfin,
L. Piccolo, J.-L. Rousset and V. Caps, Catal. Today, 2007, 122,
284; P. Lignier, S. Mangematin, F. Morfin, J.-L. Rousset and
V. Caps, Catal. Today, 2008, 138, 50.
5 D. Gajan, K. Guillois, P. Delichere, J.-M. Basset, J.-P. Candy,
V. Caps, C. Coperet, A. Lesage and L. Emsley, J. Am. Chem. Soc.,
´
2009, 131, 14667.
6 M. Brust, M. Walker, D. Bethell, D. J. Schiffrin and R. Whyman,
J. Chem. Soc., Chem. Commun., 1994, 801.
7 K. Pelzer, B. Laleu, F. Lefebvre, K. Philippot, B. Chaudret,
J.-P. Candy and J.-M. Basset, Chem. Mater., 2004, 16, 4937;
K. Pelzer, J.-P. Candy, G. Bergeret and J.-M. Basset, Eur. Phys.
J. D, 2007, 43, 197.
8 M. Boualleg, J.-M. Basset, J.-P. Candy, P. Delichere, K. Pelzer,
L. Veyre and C. Thieuleux, Chem. Mater., 2009, 21, 775.
9 B. L. V. Prasad, S. I. Stoeva, C. M. Sorensen and K. J. Klabunde,
Chem. Mater., 2003, 15, 935.
10 Gold reference catalysts, Gold Bull., 2003, 36, 24; http://www.
utilisegold.com/uses_applications/catalysis/reference_catalysts/.
11 R. C. Gray, J. C. Carver and D. M. Hercules, J. Electron Spectrosc.
Relat. Phenom., 1976, 8, 343.
12 K. S. Schneider, T. M. Owens, D. R. Fosnacht, B. G. Orr and
M. M. Banaszak Holl, ChemPhysChem, 2003, 4, 1111.
13 T. M. Owens, B. J. Ludwig, K. S. Schneider, D. R. Fosnacht,
B. G. Orr and M. M. Banaszak Holl, Langmuir, 2004, 20, 9636; B.
L. V. Prasad, S. I. Stoeva, C. M. Sorensen, V. Zaikovski and
K. J. Klabunde, J. Am. Chem. Soc., 2003, 125, 10488.
14 V. Caps, S. Arrii, F. Morfin, G. Bergeret and J.-L. Rousset,
Faraday Discuss., 2008, 138, 241.
In conclusion, we have presented a simple and reproducible
procedure for the preparation of a colloidal solution of small
crystalline gold NPs stabilized by silane ligands. Redispersion
in methylcyclohexane allows direct evaluation in the liquid
phase aerobic oxidations of trans-stilbene and cyclohexene, in
which they are found to be far more active and selective than
the reference supported catalyst, as well as ammonium- and
thiol-stabilized sols. These exceptional catalytic properties
(TOF E 80 hꢀ1) seem to come from the specific behavior of
the protecting ligand which, by turning into siloxane polymers
under oxidizing conditions, beyond geometric and electronic
effects, might supply the reaction with active intermediates.
ꢁc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 5361–5363 | 5363