Understanding the effect of postsynthesis ammonium treatment on the catalytic activity of Au/Ti-SBA-15 catalysts for the oxidation of propene

Article abstract of DOI:10.1016/j.jcat.2008.07.013

Postsynthesis ammonium treatment induces a substantial increase in the catalytic activity of Au/Ti-SBA-15 catalysts for the direct vapor-phase epoxidation of propylene using hydrogen and oxygen. The PO formation rate of a calcined Au/Ti-SBA-15 catalyst prepared by this method increased from 4.3 mg_PO h^-1 g^-1_cat to 37.2 mg_PO h^-1 g^-1_cat at 200 °C compared with a catalyst prepared in an identical manner without this treatment. The catalysts were characterized by XRD, N₂-sorption, UV-vis-NIR DRS, ²⁹Si MAS NMR, FT-IR spectroscopy, and TEM to gain insight into the relationship between ammonia treatment and catalyst activity. ²⁹Si MAS NMR measurements proved that the ammonium nitrate solution caused hydrolysis of {triple bond, long}Si{single bond}O{single bond}Si{triple bond, long} or {triple bond, long}Ti{single bond}O{single bond}Si{triple bond, long} bonds, resulting in a Ti-SBA-15 surface with a greater amount of surface hydroxyl groups. FT-IR measurements indicated the presence of amine species that favor the homogeneous deposition of Au nanoparticles. This was confirmed by TEM measurements showing greater metal dispersion for NH₄NO₃-treated Au/Ti-SBA-15 materials.

Full text of DOI:10.1016/j.jcat.2008.07.013

Journal of Catalysis 259 (2008) 43–53
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Understanding the effect of postsynthesis ammonium treatment on the catalytic
activity of Au/Ti-SBA-15 catalysts for the oxidation of propene
Elena Sacaliuc-Parvulescu a, Heiner Friedrich a, Regina Palkovits a, Bert M. Weckhuysen a,∗, T. Alexander Nijhuis b,∗
a
Inorganic Chemistry and Catalysis, Department of Chemistry, Faculty of Science, Utrecht University, Sorbonnelaan 16, 3584 CA Utrecht, The Netherlands
Eindhoven University of Technology, Chemical Reactor Engineering, P.O. Box 513, 5600 MB Eindhoven, The Netherlands
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Postsynthesis ammonium treatment induces a substantial increase in the catalytic activity of Au/Ti-
SBA-15 catalysts for the direct vapor-phase epoxidation of propylene using hydrogen and oxygen.
The PO formation rate of a calcined Au/Ti-SBA-15 catalyst prepared by this method increased from
Received 25 April 2008
Revised 20 July 2008
Accepted 21 July 2008
Available online 26 August 2008
−1
−1
−1
−1
◦
4.3 mgPO
h
g
to 37.2 mgPO
h
g
at 200 C compared with a catalyst prepared in an identical
cat
cat
manner without this treatment. The catalysts were characterized by XRD, N2-sorption, UV–vis–NIR DRS,
29Si MAS NMR, FT-IR spectroscopy, and TEM to gain insight into the relationship between ammonia
treatment and catalyst activity. 29Si MAS NMR measurements proved that the ammonium nitrate solution
caused hydrolysis of ≡Si–O–Si≡ or ≡Ti–O–Si≡ bonds, resulting in a Ti-SBA-15 surface with a greater
amount of surface hydroxyl groups. FT-IR measurements indicated the presence of amine species that
favor the homogeneous deposition of Au nanoparticles. This was confirmed by TEM measurements
showing greater metal dispersion for NH4NO3-treated Au/Ti-SBA-15 materials.
Keywords:
Propene epoxidation
Gold
SBA-15
Postsynthesis treatment
© 2008 Elsevier Inc. All rights reserved.
1. Introduction
thus could become a potential competitor for the new H2O2–PO
process, which involves 3 different reaction steps.
In many publications, it has been suggested that catalyst per-
formance depends on the nature of the support, site density and
size, and the oxidation state of Au nanoparticles [6,9–14]. Proper
choice of the support is critical. The approach to selecting the tita-
nium support was inspired by liquid-phase epoxidation reactions,
in which isolated titanium sites incorporated into different silica
matrix are the active species [15,16]. When gold is deposited on
TS-1, catalyst stability and PO yield improve when the reaction was
performed at high temperature (473 K) [17,18]. It was found that
the increase in epoxide yield also is dependent on the amount of
isolated titanium species. The amount of titanium in the silica ma-
trix must be carefully adjusted, because catalyst materials prepared
from supports with an Si/Ti molar ratio >500 are able to generate
propanal due to the increased acidic nature of the support, and
an excessively high titanium concentration leads to unwanted side
reactions and further oxidation [11]. Mesoporous silica supports,
such as MCM-41, SBA-15, and MCM-48 with titanium dispersed as
isolated species, have been found to be effective in the activity of
propene epoxidation due to their pore size distribution, pore vol-
ume, and high surface area [19–22]. The activity of Au supported
on Ti-MCM-48 exceeds that of Au supported on Ti-MCM-41, indi-
cating that the three-dimensional branched pore structure is more
suitable [21]. The drawback of MCM-41 and MCM-48 is that they
are generally hydrothermally unstable, but materials like SBA-15
are preferable for catalysis applications because of their thicker
pore walls (typically 3–9 nm) [23].
The gas-phase epoxidation of propene over various Au/Ti-
substituted catalysts is an intriguing scientific topic, not only be-
cause of the great industrial importance of this process, but also
because this reaction offers fundamental mechanistic insights into
Au-based catalysis [1–4]. By volume, propene oxide is among the
top 50 chemicals produced worldwide. It is an important building
block for the production of surfactants, polyurethane, and sol-
vents. The methods currently used for propene oxide synthesis
(i.e., the chlorhydrin and hydroperoxide processes) have important
disadvantages. The chlorhydrin process involves the production of
chlorinated side products and calcium chloride waste, whereas the
hydroperoxide process produces a stoichiometric amount of co-
product [5]. The association of highly dispersed gold nanoparticles
supported on titania (Au/TiO2) has been found to highly selective
epoxidise propene in the presence of hydrogen and oxygen under
mild conditions (323 K), but only at very low conversion (∼1%) [6].
Gold-based catalysts could become feasible for petrochemical in-
dustry if conversion (>10%) and hydrogen efficiency (>30%) could
be increased together with stability with time-on-stream, while
retaining high selectivity (>90%) [7,8]. This process is especially
attractive because of its simplicity (single reaction/reactor) and
Corresponding authors. Faxes: +31 30 251 1027, +31 40 244 6653.
E-mail addresses: b.m.weckhuysen@uu.nl (B.M. Weckhuysen), t.a.nijhuis@tue.nl
*
(T.A. Nijhuis).
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.07.013

44
E. Sacaliuc-Parvulescu et al. / Journal of Catalysis 259 (2008) 43–53
The size and distribution of gold nanoparticles also are very im-
Au catalysts, designated Au/Ti-N-n and Au/Ti-H-30h, were pre-
pared by the deposition-precipitation method [14] using aqueous
HAuCl4 solution corresponding to 1 wt% Au and Na2CO3 (1 M) as
precipitant. In this method, HAuCl4 solution was dissolved in 50 ml
of deionised water, followed by the dispersion of 1 g of Ti-SBA-15
support. The solution mixture was stirred for 15 min before the
pH was adjusted to 9 with 1 M Na2CO3. The mixture was stirred
at room temperature for 4 h, with the pH maintained at 9. The
catalysts were washed with distilled water, dried in air overnight
at 60 C, and calcined at 400 C for 4 h. To understand the effect
of the ammonium treatment, an Au/Ti-G catalyst was prepared as
a reference.
An AuN-Ti-n catalyst was synthesized by changing the order of
Au and Ti incorporation. In this preparation procedure, calcined
SBA-15 material was treated with ammonium before Au deposi-
tion, after which titanium grafting was performed on the calcined
1% AuN material.
portant for their activity and selectivity in propene epoxidation
with hydrogen and oxygen [11,14,18,24,25]. The synthesis route,
like the way of reducing the gold, the pH of the solution, the
source of gold and the physical–chemical properties of the sup-
port, are greatly influenced by the size and distribution of the gold
nanoparticles. But only a portion of gold deposited on the sur-
face of the catalysts creates active Au–Ti epoxidation centers; thus,
it is very important to maximize the number of deposited gold
nanoparticles on the support and eventually their proximity to the
Ti species [11,14,26]. This can be influenced both by the method
of gold deposition and by modification of the support proper-
ties. Each support has a specific capacity for gold capture during
deposition–precipitation. To maximize the adsorption capacity, it
is very important to choose the proper type of the support or
to perform further pretreatments to obtain narrow well-dispersed
gold nanoparticles. Delgass et al. [14] showed that an NH4NO3
treatment of TS-1 material can enhance gold capture during the
deposition–precipitation procedure.
◦
◦
2.2. Catalyst characterization
In this work, we investigate the influence of postsynthesis am-
monium treatment of Ti-SBA-15 support on the morphology of
gold nanoparticles and their catalytic performance in the propene
epoxidation. By using various characterization methods, it is possi-
ble to put forward plausible explanations for the observed differ-
ence in catalytic performances.
The degree of crystallinity and morphology of the obtained Au–
Ti-SBA-15 materials was characterized using transmission electron
microscopy (TEM) and X-ray diffraction (XRD). The latter was per-
formed using a Bruker-AXS D8 Advance powder X-ray diffractome-
ter, equipped with automatic divergence slit, Våntec-1 detector,
and cobalt Kα1,2 (λ = 1.79026 Å) source.
TEM images were acquired in bright-field mode using a Tecnai
F20 electron microscope (FEI Company). The microscope, operated
at 200 keV, was equipped with a field emission gun, a twin objec-
tive lens, and a 1k × 1k Gatan Slow Scan CCD camera. TEM sample
preparation proceeded as follows. A small portion of dry powder
(∼15 mg) was placed in a 16-mm mold/die commonly used for
pelletization. The powder was then pressed to a peak weight of
approximately 12 tons (460 MPa), after which the pressure was
immediately released. The resulting pellet was ground up in a mor-
tar, and a small portion of the very fine powder was dispersed in
etanol by ultrasonication for 30 s. Finally, a few droplets of the
suspension were placed on a Quantifoil R2/1 lacey carbon support
film and dried under a lamp. The additional step of compressing
the SBA-15 powder before conventional TEM sample preparation
was found to be advantageous for imaging of the examined ma-
terial. As shown in earlier work, imaging of small metal (oxide)
particles in the pores of SBA-15 can be carried out by electron
tomography [29–31]. However, for conventional TEM, imaging of
small metal (oxide) particles through large or thick SBA-15 struc-
tures is difficult. In this case, very thin support structures are
advantageous. During pressing of the material, the SBA-15 parti-
cles, roughly 2000 × 500 × 500 nm in size, fracture, leaving much
smaller and thinner pieces intact. Very small Au particles are found
in the pores of theses thin pieces. Parts of the SBA-15 structure
collapse, resulting in disordered amorphous silica. Comparison of
pressed and unpressed material ensured that pressing did not af-
fect the gold particle.
The particle size distribution was estimated by measuring the
diameter of more than 100 particles in each sample using Analy-
sis software. The corresponding volume and surface area weighted
distributions were calculated directly from the particle diameters,
assuming a spherical particle shape. The specific surface area and
pore volume were determined by N2 sorption measurements us-
ing a Micromeritics ASAP 2400 instrument. Surface areas were
calculated using the BET model, with micropores and macropores
described by the Horvath–Kawazoe and BJH models, respectively.
UV–vis–NIR diffuse reflectance spectroscopy (DRS) measure-
ments were carried out at room temperature on a Varian Cary
500 instrument in the range of 200–2200 nm. This set-up was
equipped with a diffuse reflectance accessory, which was set to
2. Experimental
2.1. Catalyst preparation
Pure mesoporous silica SBA-15 material was hydrothermally
synthesized according to the procedure described by Zhao [27].
The reagent used were poly(ethylene glycol)-block-poly(propylene
glycol)-block-poly(ethylene glycol) (PEG–PPG–PEG) (average molec-
ular weight 5800, Aldrich), tetraethylortosilicate (Aldrich, 98%), HCl
(Merck, 37%), and distilled water. The gel solution was prepared by
dispersing 8 g of PEG–PPG–PEG in 60 g of distilled water, followed
◦
by 15 min of vigorous stirring at 40 C. The resultant solution was
mixed with 48 g HCl and 190 g H2O and stirred for another 17 h.
Finally, 21.2 g tetraethylortosilicate solution was added dropwise in
◦
the gel mixture at 40 C with stirring for 9 h before being placed
◦
in a Teflon autoclave at 100 C for 12 h. After the completion of
synthesis, the product was filtered, washed with distilled water,
and air-dried overnight at 60 C. The solid was calcined at 550 C
for 6 h.
The Ti-SBA-15 grafted support, designated Ti-G, with a Si/Ti
molar ratio of 40, was synthesized using titanium isopropoxide
(Across, 98%) as a titanium source [28]. Titanium grafting was per-
formed inside a glove box in nitrogen environment. Titanium(IV)
isopropoxide was dispersed in 2 g of dry isopropyl alcohol (99.7%,
Merck) and mixed with 1 g of SBA-15 at ambient temperature. The
solid was dried at ambient temperature under flowing nitrogen. Fi-
◦
◦
◦
nally, the titanium-grafted material was calcined at 400 C for 4 h.
The postsynthesis ammonium treatment was carried using
NH4NO3 (1 M) solution [14]. Ammonium-treated Ti-SBA-15 sup-
ports were prepared by suspending 1 g of calcined Ti-SBA-15 ma-
terial in 60 g of NH4NO3 solution. The solution was heated at 80 C
under continuous stirring from 6 h to 30 h. The product thus ob-
◦
◦
tained was washed with 80 g of deionized water and dried at 60 C
overnight. The resulting samples were labeled Ti-N-n, with n rep-
resenting the ammonium treatment time. The ammonium-treated
Ti-SBA-15 supports were not calcined before the Au deposition.
A postsynthesis water treatment was also carried out, following
the same procedure as in ammonium treatment. The water-treated
Ti-SBA-15 support was designated Ti-H-30.

E. Sacaliuc-Parvulescu et al. / Journal of Catalysis 259 (2008) 43–53
45
(a)
(b)
Fig. 1. Catalytic performance of (a) Au-Ti-G and (b) Au/Ti-N-30 (conversion (2) and selectivity (1)).
Table 1
Overview of the catalytic performances of the different Au/Ti-SBA-15 catalyst materials under investigationa
Catalyst
Temperature
( C)
C (%)
C3H6
Selectivity, S (%)
H2 eff
(%)
◦
PO
CO2
Propanol
Propane
Other oxygenates
Au/Ti-N-6h
90
135
170
200
0.4
0.8
1.6
2.7
99.4
97.7
89.3
65
0
0
0.6
1.1
1.47
1.94
0
0
0
0
0
4.59
3.27
2.67
1.54
1.2
1.56
1.16
7.67
31.9
Au/Ti-N-15h
Au/Ti-N-30h
AuN-Ti-15h
90
135
170
200
0.5
0.9
1.9
3.2
97.4
92.7
82.9
58.4
0
0
1.87
1.7
2.42
0
0
0.73
1.93
2.85
2.63
3.83
2.81
2.26
1.4
3.67
0.69
0
11.14
38.97
90
135
170
200
0.7
1.4
2.8
4.6
98
0
1
1.29
1.32
1.73
2
0
0
0
0
0.71
1.98
2.47
1.32
5.46
3.63
2.22
1.38
95.7
74.6
46.4
21.2
50.28
90
135
170
200
0.4
0.7
0.8
1.9
98.7
96.5
89
0
0
1.3
3.5
10.94
26.65
0
0
0
0
0
0
0
0
12.44
8.47
4.5
0.06
33.65
39.7
0.92
Au-Ti-G
90
135
170
200
0
0
0
0
0
0
0
0
2.8
5.18
0
0
0
0
0
1.1
2
0
0.08
0.13
0.2
98.9
95.2
90.1
14.13
11.86
4.55
4.72
a
C—conversion, S—selectivity.
collect diffuse reflected light only. A baseline correction was per-
formed using a white Halon standard. 29Si MAS NMR measure-
ments were carried out with a Bruker Avance 500WB spectrometer
at a spinning rate of 10 kHz using π/4 pulses of 2.2 μs and a repe-
tition delay of 30 s. FT-IR spectra were recorded on a Perkin-Elmer
Spectrum One instrument on self-supported catalyst wafers placed
in an IR transmission cell equipped with CaF2 windows. For each
sample, an IR spectrum was recorded before and after drying in
perature 1 h, then cooled to the next reaction temperature in the
cycle. The performance was typically tested at 10–15 different tem-
peratures (including duplicates to determine for each catalyst the
deactivation pattern).
3. Results
Fig. 1 shows the catalytic activities of different Au/Ti-SBA-15
samples in the propene epoxidation reaction with O2 and H2.
The untreated Au/Ti-G catalyst showed relatively low activity un-
der the reaction conditions. This activity is typical for unmodified
Au/TS-1 catalyst under these conditions. But the Au/Ti-N-30h cat-
alyst with Au supported on NH4NO3-pretreated Ti-SBA-15 support
showed substantial activity for propene oxide formation. NH4NO3
treatment clearly exerted a pronounced effect on the catalytic per-
formance of the Au/Ti-N-30 catalyst. As the reaction temperature
was increased, the selectivity to propene oxide decreased, corre-
sponding to an increase in total oxidation products.
The conversion, selectivity, hydrogen efficiency, and product
distribution of the different Au/Ti-SBA-15 catalysts studied are
summarized in Table 1. All catalysts exhibited higher conversion
levels with increasing reaction temperature, but at the expense of
the selectivity toward propene oxide due to an increase in total ox-
◦
helium at 300 C.
2.3. Catalyst testing
A flow reactor was used to evaluate the catalytic performance
of the different catalysts prepared. The experiments were car-
ried out typically with 0.4 g of catalyst material and a flow of
50 Nml min−1. The gas mixture consisted of 10% oxygen, 10% hy-
drogen, and 10% propene in He. The analysis of the gas leaving the
reactor was carried out using an Interscience Compact GC system,
equipped with a Molsieve 5A and a Porabond Q column, each with
a thermal conductivity detector (TCD). Gas samples were analyzed
every 3 min. The experiments were carried out in cycles: 5 h at
reaction temperature with the reactant mixture, followed by a re-
generation cycle. In regeneration, 10% oxygen in helium was used,
◦
and the catalyst was heated at 300 C and maintained at that tem-

46
E. Sacaliuc-Parvulescu et al. / Journal of Catalysis 259 (2008) 43–53
Fig. 3. Small angle XRD patterns of: (a) SBA-15, (b) Ti-G, (c) AuN–Ti-15h, (d) Au/
Ti-N-6h, (e) Au/Ti-N-15h, and (f) Au/Ti-N-30h.
Fig. 2. Catalytic performance of Au-based catalyst materials tested at two differ-
◦
◦
ent temperatures (a) 130 C and (b) 200 C. The catalyst materials are: (Q) Au-Ti-G,
−1
(1) AuN–Ti-15h, (2) Au/Ti-N-6h and (P) Au/Ti-N-30h. 0.4 g catalyst, 50 Nml min
Fig. 4. Wide angle XRD patterns of: (a) Au/Ti-N-6h, (b) Au/Ti-N-15h, (c) Au/Ti-N-30h,
(d) Au-Ti-G, and (e) AuN-Ti-15h.
gas federate (10% H2, O2, propene), total pressure 1100 mbar.
idation to CO2. The water-treated Au/Ti-H-30h catalyst was inactive
in propene epoxidation reaction, with propane the main product.
Another important finding is that the catalytic activity of the
Au/Ti-N-n materials was enhanced when the NH4NO3 treatment
of Ti/SBA-15 support was increased from 6 h to 30 h. Fig. 2 shows
the yields of propene oxide over different Au/Ti-SBA-15 catalysts as
a function of time on stream. The propene oxide yields were lower
for the Au–Ti-G and AuN–Ti-15h catalysts than for the Au/Ti-N-
30h catalyst. Small losses in the activity of Au/Ti-SBA-15 catalysts
for boosting the catalytic activity of NH4NO3-treated Au/TS-1 cata-
lysts. To gain more insight into this boosting effect, we thoroughly
characterized the different Au/Ti-SBA-15 catalyst materials under
investigation through various techniques.
3.1. X-ray diffraction
The powder XRD patterns of the calcined samples are shown
in Figs. 3 and 4. In the small-angle region of the XRD patterns
(Fig. 3), all of the samples exhibit three well-resolved diffrac-
tion peaks that can be indexed to (100), (110), and (200) reflec-
tions of the p6mm hexagonal space group, which matches with
the topology of SBA-15 structure [27]. After incorporation of Ti,
the SBA-15 structure was preserved, as shown by the presence
of intense reflection peaks of the Ti-SBA-15 sample. But the in-
tensities of the peaks decreased in the case of ammonium-treated
Au/Ti-SBA-15 samples, with this tendency even more pronounced
with increasing ammonium treatment time. These results may in-
dicate that the incorporation of Au and the increase in ammo-
nium treatment time from 6 h to 30 h decreased the long-range
order of the SBA-15 structure. XRD data from the high-angle re-
gion of the Au/Ti-SBA-15 samples are shown in Fig. 4. The ab-
sence of the gold peaks for the ammonium-treated Au/Ti-SBA-15
samples indicates that gold nanoparticles were highly dispersed
◦
at 200 C during 5 h time on stream were seen, although catalytic
activity could be restored almost completely after an activation
◦
procedure at 300 C. Comparing the activities of Au/Ti-SBA-15 cata-
lysts for which the order of Au and Ti incorporation differed shows
that the Au/Ti-N-15h material had greater catalytic activity than
the AuN–Ti-15h material. Propene oxide production was measured
◦
after 3–5 h of steady-state activity at 200 C. The Au-Ti-G cata-
−1
lyst had a rate of 4.3 mgPO
h
g
−1, whereas the NH4NO3 treated
cat
−1
Au/Ti-N-30h catalyst had a rate of 37.2 mgPO
h
g
−1. Delgass et al.
cat
[14] previously reported this substantially increased catalytic ac-
tivity of NH4NO3-treated catalyst materials for Au/TS-1 materials.
This group explained the observed differences in catalytic perfor-
mances in terms of increasing active site density. The reason for
the increased number of active sites due to NH4NO3 treatment
remains unclear, however; Delgass et al. suggested that the for-
mation of an Au-amine complex near Ti sites could be responsible

E. Sacaliuc-Parvulescu et al. / Journal of Catalysis 259 (2008) 43–53
47
Fig. 5. N2 adsorption–desorption isotherms of: (a) SBA-15, (b) Ti-G, (c) Ti-N-30h
calcined, (d) Ti-H-30h, (e) Au/Ti-H-30h, (f) Au/Ti-N-30h, and (g) Au-Ti-G.
Fig. 6. Pore size distributions derived from desorption branches of N2-sorption
isotherms of: (a) SBA-15, (b) Ti-G, (c) Ti-H-30h, (d) Ti-N-30h calcined, (e) Au/Ti-N-
6h, (f) Au/Ti-N-30h, and (g) Au-Ti-G.
through the silica matrix and that the average particles size was
too small to be detected by XRD. Interestingly, in the AuN-Ti-15h
sample, broad reflections of Au were seen, due to the formation
of larger Au nanoparticles inside the channels of the SBA-15 sup-
port.
[33]. Interestingly, the decrease in the pore volume was more pro-
nounced in the AuN–Ti-15h sample than in the Au/Ti-N-15h sam-
ple. This suggests that the order of Au and Ti incorporation also
has an influence on the textural properties of the resulting mate-
rial, likely due to the presence of larger Au nanoparticles in the
SBA-15 channels, as was seen in the XRD spectrum (Fig. 2).
3.2. N2 sorption measurements
Nitrogen adsorption–desorption isotherms of all synthesized
samples are displayed in Fig. 5. The calcined samples exhibit
type IV isotherms with H1-type hysteresis, typical of mesoporous
materials with one-dimensional cylindrical channels [32]. The
sharp inflection between the relative pressure P/Po 0.66 and 0.8
observed in the isotherm of SBA-15 correspond to the capillary
condensation within uniform mesopores. The sharpening of this
step reflects the uniform pore size distribution in the material. The
introduction of Ti induced a decrease in the BET surface area of
3.3. UV–vis–NIR diffuse reflectance spectroscopy
DRS in the UV–vis–NIR region can distinguish the presence of
framework and extra-framework titanium species in a silica matrix
[34,35]. Representative UV–vis–NIR DRS spectra of the samples are
shown in Fig. 7. In all DRS spectra of the titanium-containing SBA-
15 samples, an intense band centered at ca. 220 nm can be ob-
served, along with a shoulder at 260–270 nm. The band at 220 nm
is due to the charge-transfer process of isolated framework tita-
nium in tetrahedral coordination [34]. The shoulder at 270 nm has
been attributed to the presence of Ti in fivefold and sixfold coor-
dination, generated through the hydration of tetrahedrally coordi-
nated sites [36]. Note that the intensity of this shoulder is much
lower for the as-synthesized Ti-N-n supports than for the calcined
Ti-N-n supports. These results suggest that the Ti environment
for as-synthesized ammonium-treated Ti-SBA-15 supports differed
than after calcination. Moreover, the intensity of the shoulder at
270 nm decreased after NH4NO3 treatment, as seen by comparing
the Ti-G support with the calcined Ti-N-6h support (Fig. 7, spectra
a and e). The UV–vis–NIR DRS spectra cannot be used to estimate
the quantity of titanium in the sample, but the variation in the in-
tensity may reflect changes in the amount and type of species of Ti
−1
−1
SBA-15 material from 972 m2 g
to 738 m2 g
and a decrease
(Table 2). The BET
−1
−1
in pore volume from 1.3 m3 g
to 1 m3 g
surface area and pore volume of Ti-SBA-15 also were modified af-
ter the 6-h ammonium treatment and changed slightly when this
treatment was prolonged to 30 h. The decreased specific surface
area of the ammonium-treated Ti-SBA-15 supports likely is caused
by the modification of the Ti-SBA-15 structure.
Fig. 6 shows the BJH pore size distribution of the samples. The
Ti-SBA-15 supports exhibited a bimodal structure after the water
and ammonium treatment, both of which can cause hydrolysis of
≡Si–O–Si≡ or ≡Ti–O–Si≡ bonds, which may be associated with
the formation of smaller pores in the structure. The introduction of
Au inside the SBA-15 channels also induced a decrease in the pore
volume and BET surface area of the Ti-SBA-15 material, demon-
strated that some pore blocking occurred during Au incorporation
Table 2
Summary of the characterization results of the different SBA-15, Ti-SBA-15 and Au/Ti-SBA-15, samples under study
Catalyst
Si/Ti final
molar ratio
Au loading
(wt%)
BET surface area
Micropore
t-Plot area
Pore volume BJH
Pore size BJH
desorption (nm)
−1
−1
−1
−1
(cm3
g )
(m2
g
)
vol. (cm3
g
)
(m2
g
)
SBA-15
Ti-G
–
–
–
–
–
–
0.05
1
1
1
970
740
640
615
648
483
480
510
490
510
430
0.07
0.05
0.04
0.03
0.06
0.01
0.007
0.009
0.007
0.006
0.009
810
615
550
525
512
447
460
475
460
490
440
1.3
1
0.92
0.9
0.86
0.74
0.83
0.87
0.83
0.83
0.82
6.9
6.8
43.0
43.5
49.4
48.3
–
–
–
–
–
Ti-N-6h (calcined)
Ti-N-30h (calcined)
Ti-H-30h
Au/Ti-H-30h
Au/Ti-N-6h
Au/Ti-N-15h
Au/Ti-N-30h
Au–Ti-G
4.6; 6.9
4.6; 6.9
3.8; 6.3
3.8; 6.3
4.6; 6.9
4.6; 6.9
4.6; 6.9
6.6
0.1
1
AuN–Ti-15h
–
6.8

48
E. Sacaliuc-Parvulescu et al. / Journal of Catalysis 259 (2008) 43–53
Fig. 7. UV–vis–NIR DR spectra of Ti-SBA-15 supports: (a) Ti-G, (b) Ti-H-30h (c) Ti-N-
6h, (d) Ti-N-15h, (e) Ti-N-30h, (f) Ti-N-6h calcined, (g) Ti-N-30h calcined, and (h)
TiO2 anatase.
Fig. 9. 29Si MAS NMR spectra of: (a) SBA-15 calcined, (b) Ti-G calcined, (c) Ti-N-
6h as synthesized, (d) Ti-N-15h as synthesized, (e) Ti-N-30h as synthesized, (f) Au/
Ti-N-30h calcined, (g) Au-Ti-G calcined, and (h) AuN-Ti-15h calcined.
linked via siloxane bridges to four other silicon atoms, (≡SiO)4Si,
designated Q4. The two downfield resonance peaks at −97 and
−89 ppm are attributed to silicon in single silanols, (≡SiO)3SiOH,
designated Q3, and geminal silanols, (≡SiO)2Si(OH)2, designated
Q2 [40], respectively. The spectra of Ti-SBA-15 before (Fig. 9b) and
after treatment with NH4NO3 (Fig. 9, spectra c, d, and e) are con-
siderably different. A significant increase in signal intensity of the
peaks at −97 and −89 ppm occurred after the NH4NO3 treatment
(Fig. 9, spectra c, d, and e); however, the duration of this treatment
did not seem to be reflected in comparable spectra after 6, 15, or
30 h of treatment time. 29Si NMR spectra of Ti-SBA-15 supports
before and after water treatment are shown in Fig. 10. Again, an
increase in the intensity of Q2 and Q3 species can be seen, caused
in this case by the water treatment (Fig. 10b), which is similar to
the effect of the NH4NO3 treatment.
Fig. 8. UV–vis–NIR DR spectra of Au/Ti-SBA-15 samples of: (a) Au-Ti-G, (b) Au/Ti-H-
30h, (c) Au/Ti-N-6h, (d) Au /Ti-N-15h, (e) Au/Ti-N-30h, and (f) AuN-Ti-15h.
in Ti-SBA-15 after the ammonium treatment. These results confirm
that during NH4NO3 treatment, some extra-framework titanium
was removed from the surface of SBA-15, in good agreement with
the ICP-AES results. Note that the spectrum of the water-treated
Ti-H-30h sample (Fig. 7b) is similar to that of the Ti-G support
(Fig. 7a). The absence of a band at 330–340 nm indicates that no
bulk titania was formed during synthesis [36]. The presence of a
UV–vis absorption band at around 520 nm results from the plas-
mon resonance of nanosized metallic gold particles present in the
Au/Ti-SBA-15 samples (Fig. 8) [37]. The intensity of this band was
greater for the AuN/Ti-15h sample than for the Au/Ti-N-n samples,
confirming the presence of smaller Au particles in the Au/Ti-N-n
samples, as also demonstrated in the XRD pattern analysis.
The increased intensity of Q2 and Q3 species on pretreat-
ment of Ti-SBA-15 with 1 M solution of NH4NO3 or with water
can be associated with a partial hydrolysis of some of ≡Si–O–
Si≡ or ≡Ti–O–Si≡ bonds according to Scheme 1, resulting in in-
creased density of surface hydroxyl groups. The possibility that
these two resonance peaks also may contain contributions from
Si(3Si, 1Ti) sites and from Si(2Si, 1Ti, 1H) sites cannot be excluded,
however.
3.4. 29Si MAS NMR spectroscopy
29Si MAS NMR spectroscopy can provide information about the
nature and amount of silanol groups present in the Ti-SBA-15
mesostructure [38,39]. The 29Si NMR spectra for the calcined SBA-
15, Ti-SBA-15, Au-Ti-SBA-15, and as-synthesized Ti-N-n supports
are shown in Fig. 9. For the calcined SBA-15 and Ti-SBA-15 ma-
terials (Fig. 9, spectra a and b), the 29Si NMR spectra have an
intense resonance peak at about −107 ppm, associated with silicon
29Si MAS NMR spectra of the Au/Ti-SBA-15 samples (Fig. 9,
spectra f, g, and h) after gold deposition and calcination again ex-
hibited reduced intensity of the Q2 and Q3 signals, which can be
assigned to a decrease in surface hydroxyl groups due to conden-
sation during the calcination procedure.

E. Sacaliuc-Parvulescu et al. / Journal of Catalysis 259 (2008) 43–53
49
Fig. 12. IR spectra of: (a) Ti-G, (b) Ti-N-30h, (c) Au/Ti-N-30h recorded after thermal
◦
treatment at 300 C.
Fig. 10. 29Si MAS NMR spectra of: (a) Ti-G calcined and (b) Ti-H-30 as synthesized.
Scheme 1. NH4NO3 treatment of Ti-SBA-15 material.
3.6. Transmission electron microscopy
To gain more insight into the effect of the postsynthesis
NH4NO3 treatment and water treatment, the Au/Ti-SBA-15 ma-
terials were analyzed by TEM. Fig. 13 presents TEM micrographs
of Au/Ti-SBA-15 samples with (Au/Ti-N-30h) and without (Au/Ti-
G) NH4NO3 treatment of the support before the gold deposition.
The corresponding number-, surface-, and volume-weighted par-
ticle size distributions [41] are given in Fig. 14. The TEM image
of the Au/Ti-G material (Fig. 13a) shows that very few Au par-
ticles were deposited on the support, in good agreement with
ICP-AES results. Furthermore, Au particles were not well dispersed
in the channels of Ti-SBA-15; some of them, mainly large parti-
cles, seemed to be formed on the outer surface of the support.
The Au particles in this sample had an average particle diame-
ter of 5 nm (number-weighted), 11 nm (surface-weighted), and
15 nm (volume-weighted). It was difficult to determine the exact
location of the gold nanoparticles from the TEM micrographs of
the water treated Au/Ti-H-30 sample (not shown), however, pre-
sumably because of the very low gold loading deposited on the
water-treated Ti-SBA-15 supports. The TEM image of the Au/Ti-N-
30h sample (Fig. 13b) clearly shows the presence only of small
gold nanoparticles. In addition, particles with a mean diameter of
2–3 nm (number weighted), 4 nm (surface weighted), and 5 nm
(volume weighted), were dispersed homogeneously in the Ti/SBA-
15 mesopores. The need for titanium in the SBA-15 structure to
ensure good Au dispersion/adsorption is clearly demonstrated by
the TEM micrograph of the AuN/Ti-15h sample (Fig. 13c), for which
Au was first deposited on an NH4NO3 pretreated SBA-15 and Ti
was then grafted. In this case, larger Au nanoparticles were formed,
comparable with those in the Au/Ti-N-30h sample in which the Ti-
SBA-15 support was pretreated before Au deposition. The average
diameter of Au nanoparticles for AuN/Ti-15 material was around
6 nm for number-weighted, 8 nm for surface-weighted, and 9 nm
for volume-weighted analyses. These finding clearly indicate that
the postsynthesis NH4NO3 treatment of the Ti-SBA-15 support had
a significant beneficial effect for obtaining smaller, more uniform
gold nanoparticles inside the SBA-15 channels.
Fig. 11. IR spectra of: (a) Ti-G, (b) Ti-N-30h, (c) Au/Ti-N-30h recorded at room tem-
perature.
3.5. FT-IR spectroscopy
Fig. 11 shows the IR spectra of Ti-G (a), Ti-N-30h (b), and
Au/Ti-N-30h (c) before the thermal treatment. A narrow band at
−1
−1
is as-
3750 cm
and a broad absorption band in the 3700–3300 cm
−1
range can be seen for all samples. The band at 3750 cm
signed to the OH stretching vibration of isolated silanol or geminal
silanol groups located at the surface of the inner walls of the SBA-
−1
15 channels [40]. The absorption band in the 1700–1600 cm
region can be attributed to the water deformation vibration. The
presence of NH4NO3 after postsynthesis treatment of the Ti-N-30h
sample (Fig. 11b) is confirmed by the presence of two broad ab-
−1
sorption bands at 1500–1250 cm
and 3600–3300 cm−1, which
+
can be assigned to the NH4 bending and stretching vibrations. Af-
◦
ter the thermal treatment at 300 C in helium (Fig. 12b), decompo-
sition of NH4NO3 occurred, and the spectra of all samples became
−1
similar. The band at 3750 cm
became sharper and more intense,
−1
−1
de-
whereas the bands at 3700–3300 cm
and 1700–1600 cm
creased in intensity during thermal treatment because of water
and NH4NO3 removal.

50
E. Sacaliuc-Parvulescu et al. / Journal of Catalysis 259 (2008) 43–53
(a)
(b)
(c)
Fig. 13. TEM micrographs of: (a) Au-Ti-G, (b) Au/Ti-N-30h, and (c) AuN–Ti-15h.
(a)
Fig. 14. Number, surface and volume weighted particle size distributions of: (a) Au-Ti-G, (b) Au/Ti-N-30h, and (c) AuN–Ti-15h.

E. Sacaliuc-Parvulescu et al. / Journal of Catalysis 259 (2008) 43–53
51
(b)
(c)
Fig. 14. (continued)
−
ually transformed into negatively charged Si–O , but in the 4.5–
4. Discussion
+
6 range, the Ti–OH groups were protonated as Ti–OH2 [44]. In
this pH range, the gold species in solutions were mainly anions
4.1. Proposed Au deposition mechanism in NH4NO3-pretreated
Ti-SBA-15
−
[AuClx(OH)4−x
]
(x = 1–4). This negatively charged species could
+
then be adsorbed on the protonated Ti–OH2 groups according to
Scheme 2. The electrostatic interaction will be stabilized by the for-
mation of an adsorbed hydroxochloro complex of Au (Scheme 2).
This complex will act as a nucleation center for the formation of
the Au(OH)3 species [45]. The formation of an Au–amine complex
next to the Ti sites was proposed by Delgass et al. [14] to ex-
plain the increased Au deposition on a TS-1 support pretreated
with 1 M NH4NO3. No experimental evidence for this complex
was given, but comparing the water-pretreated Ti-H-30h support
with the NH4NO3 pretreated Ti-N-n supports shows greater gold
adsorption on the latter. These findings that along with generat-
ing a greater density of hydroxyl groups on the Ti-SBA-15 support,
amine species also played a role in the homogeneous deposition of
Au. Therefore, we may assume that the presence of amine species
in the structure of the Ti-SBA-15 supports enhanced the homoge-
neous dispersion of gold in to the support. Amine species may be
concentrated near to the Lewis acidic Ti sites after postsynthesis
A high intake of gold on a support for TiO2 requires a high
number of available hydroxyl groups as adsorption centers for the
gold species [42]. As shown by our NMR findings, pretreatment
with NH4NO3 significantly increased in the number of hydroxyl
groups (Scheme 1). According to the ICP-AES measurements, Au
was completely deposited in the NH4NO3-pretreated Ti-SBA-15
support. In contrast, in the non-pretreated support, the amount
of Au deposited was only 10% of the available gold precursor in
the solution. The actual deposition mechanism should be simi-
lar to that proposed by Moreau et al. [43], who investigated Au
deposition on TiO2 by studying the effect of pH in the deposition-
precipitation process and the species formed at different pH val-
ues. When the Ti-SBA-15 support was immersed in the HAuCl4
solution, the pH was 2.6. The pH increased continuously up to
a value of 9 through the addition of Na2CO3 to obtain the de-
posited Au(OH)3. In this pH range, the Si–OH groups were grad-

52
E. Sacaliuc-Parvulescu et al. / Journal of Catalysis 259 (2008) 43–53
Scheme 2. Mechanism of HAuCl4 deposition on NH4NO3 pretreated Ti-SBA-15 at pH <6.
Scheme 3. Mechanism of HAuCl4 deposition on NH4NO3 pretreated Ti-SBA-15 at pH >6.
NH4NO3 treatment of the Ti-SBA-15 support, causing a more ho-
mogeneous deposition of gold nanoparticles in the proximity of
Ti sites. This was proven by the results obtained for the AuN/Ti-
30h sample when Au was deposited on a NH4NO3-treated SBA-15
and Ti was then added. In this latter case, gold was not homoge-
neously dispersed, because there was no Ti on the support that
could concentrate the amine species and facilitate the homoge-
neous deposition on Ti–OH species. At pH >6, the surface of the
Ti-SBA-15 support became negatively charged, due to deprotona-
tion of most of the hydroxyl groups. In solution, the predominant
of 2–3 nm, much smaller than that of the Au-G catalyst. Moreover,
the differences in volume-weighted mean diameter, from 5 nm
(Au/Ti-N-30h) to 15 nm (Au-G), clearly suggest that apart from
the low loading of Au present in Au/Ti-G catalyst, a large per-
centage of the Au atoms were catalytically inactive because they
were inaccessible due to the large Au nanoparticles. It is notewor-
thy that changes in the Au dispersion/particle size between these
types of catalysts were related to the different pretreatments of
the support with NH4NO3. As demonstrated earlier, NH4NO3 pre-
treatment led to better gold adsorption and dispersion due to the
increasing number of available adsorption sites at the Ti-SBA-15
surface. Examining the catalytic data in Table 1 for the Au/Ti-N-n
(n 6–30 h) pretreated with NH4NO3 for different times reveals the
greatest PO yield in the sample pretreated for 30 h (Au/Ti-N-30h).
Because all of these samples contained the same amount of gold,
the difference in activity can be attributed to better dispersion of
the gold particles on the support due to the increased number of
adsorption sites. UV–vis–NIR DRS spectra of the Ti-SBA-15 sup-
ports found titanium also as extra-framework octahedral species.
This type of titanium species, known to be responsible for side re-
actions in propene epoxidation, can be partially removed during
NH4NO3 pretreatment. Increasing the pretreatment time will then
further decrease the amount of these extra-framework Ti species,
resulting in increased PO yield.
−
−
gold species were AuCl(OH)3 and Au(OH)4 . At this pH, the already
coordinated AuCl(OH)2 species will react with some possible free
OH groups, resulting in the formation of an –O–AuCl(OH) complex
−
−
(Scheme 3). The AuCl(OH)3 and Au(OH)4 also can react with the
remaining free OH groups on the surface to produce the –O–Au
bonds (Scheme 3). Further increases in the pH can transform these
species into Au(OH)3. Previous studies of the deposition of gold on
TiO2 have shown decreased Au intake at pH 9. The presence of free
hydroxyl groups and amine species on the surface of the Ti-SBA-
15 will lead to enhanced adsorption of the gold species starting at
a low pH value when the hydroxyl groups are protonated. This in-
creased absorption will occur through electrostatic interaction with
anionic gold species, creating a good, uniform absorption of gold
from the solution. When the pH is increased from 6 to 9, all of
these species will be stabilized by chemical interaction with the
support (Scheme 3).
5. Conclusion
The enhanced activity of Au/Ti-N-n catalyst for the propene
epoxidation reaction compared with that of the Au-Ti-G catalyst
can be associated with the differences in size and dispersion of
the deposited Au nanoparticles in the SBA-15 pore structure. As
determined from detailed TEM measurements, the Au particles in
the Au/Ti-N-30h catalyst had a number-weighted mean diameter
The pretreatment of Ti-SBA-15 support material with NH4NO3
before gold immobilization was found to enhance both the gold
loading and dispersion. The resulting catalyst materials have en-
hanced activity in the direct epoxidation of propene with H2 and
O2 compared with untreated Au-Ti-SBA-15 materials. The main ef-

E. Sacaliuc-Parvulescu et al. / Journal of Catalysis 259 (2008) 43–53
53
A
fect of the NH4NO3 pretreatment was related to an increased num-
[20] B.S. Uphade, Y. Yamada, T. Akita, T. Nakamura, M. Haruta, Appl. Catal.
Gen. 215 (2001) 137.
[21] B.S. Uphade, T. Akita, T. Nakamura, M. Haruta, J. Catal. 209 (2002) 331.
[22] E. Sacaliuc, A.M. Beale, B.M. Weckhuysen, T.A. Nijhuis, J. Catal. 248 (2007)
235.
[23] D. Trong On, D. Desplamtier-Giscard, C. Danumah, S. Kaliaguine, Appl. Catal. A
Gen. 222 (2001) 299.
[24] C. Qi, T. Akita, M. Okumura, K. Kuraoka, M. Maruta, Appl. Catal. A Gen. 253
(2003) 75.
ber of gold adsorption sites at the Ti-SBA-15 surface. This can be
explained by the cooperative effect between an increasing number
of hydroxyl groups formed by a partial hydrolysis of the ≡Si–O–
Si≡ or ≡Ti–O–Si≡ bonds in the Ti-SBA-15 support material and
the concentration of amine species in the proximity of Ti sites.
Acknowledgments
[25] M. Haruta, Catal. Today 36 (1997) 153.
[26] N. Yap, R.P. Andres, W.N. Delgass, J. Catal. 226 (2004) 156.
[27] D. Zhao, J. Feng, Q. Huo, N. Melosh, B.F. Chmelka, G.D. Stucky, J. Am. Chem.
Soc. 120 (1998) 26024.
[28] X. Gao, S.R. Bare, J.L.G. Fierro, M.A. Banares, I.E. Wachs, J. Phys. Chem. B 102
(1998) 5653.
[29] A.H. Janssen, C.-M. Yang, Y. Wang, F. Schüth, A.J. Koster, K.P. de Jong, J. Phys.
Chem. B 107 (2003) 10552.
[30] H. Friedrich, J.R.A. Sietsma, P.E. de Jongh, A.J. Verkleij, K.P. de Jong, J. Am. Chem.
Soc. 129 (2007) 33.
Financial support was provided by STW/NWO-VIDI (to T.A.N.),
NWO-VICI (to B.M.W.), and NRSC-C (to B.M.W.). The authors thank
C. van de Spek for the TEM measurements and A. Mens for the
N2-sorption measurement.
References
[1] T.A. Nijhuis, T. Visser, B.M. Weckhuysen, Angew. Chem. Int. Ed. 44 (2005) 1115;
T.A. Nijhuis, T. Visser, B.M. Weckhuysen, J. Phys. Chem. 109 (2005) 19309;
T.A. Nijhuis, B.M. Weckhuysen, Chem. Commun. (2005) 6002.
[2] T.A. Nijhuis, B.M. Weckhuysen, Catal. Today 117 (2006) 84.
[3] B. Chowdhury, J.J. Bravo-Suárez, N. Mimura, J. Lu, K.K. Bando, S. Tsubota, M.
Haruta, J. Phys. Chem. C 110 (2006) 22995.
[4] T.A. Nijhuis, M. Makkee, J.A. Moulijn, B.M. Weckhuysen, Ind. Eng. Chem. Res. 45
(2006) 3447.
[5] K. Weissermel, H.J. Arpe, in: Industrial Organic Chemistry, second ed., VCH,
New York, 1993, p. 141, 264.
[6] T. Hayashi, K. Tanaka, M. Haruta, J. Catal. 178 (1998) 566.
[7] A. Zwijnenburg, M. Makkee, J.A. Moulijn, Appl. Catal. A Gen. 270 (2004) 49.
[8] B. Chowdhury, J.J. Bravo-Suárez, M. Daté, S. Tsubota, M. Haruta, Angew. Chem.
Int. Ed. 45 (2006) 412.
[9] M. Haruta, M. Date, Appl. Catal. A Gen. 222 (2001) 427.
[10] A.K. Sinha, S. Seelan, S. Tsubota, M. Haruta, Top. Catal. 29 (2004) 95.
[11] B. Taylor, J. Lauterbach, W.N. Delgass, Appl. Catal. A Gen. 291 (2005) 188.
[12] A.K. Sinha, S. Seelan, T. Akita, S. Tsubota, M. Haruta, Appl. Catal. A Gen. 240
(2003) 243.
[13] A.K. Sinha, S. Seelan, T. Akita, S. Tsubota, M. Haruta, Catal. Lett. 85 (2003) 3.
[14] L. Cumaranatunge, W.N. Delgass, J. Catal. 232 (2005) 38.
[15] A. Corma, M.T. Navarro, J. Pérez-Pariente, J. Chem. Soc. Chem. Commun. (1994)
147.
[31] U. Ziese, K.P. de Jong, Appl. Catal. A Gen. 260 (2004) 71.
[32] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscow, R.A. Pierotti, J. Rouquerol,
T. Siemieniewska, Pure Appl. Chem. 57 (1985) 603.
[33] P. Van Der Voort, P.I. Ravikovitch, K.P. De Jong, M. Benjelloun, E.A. Van Bavel,
H. Janssen, A.V. Neimark, B.M. Weckhuysen, E.F. Vansant, J. Phys. Chem. B 106
(2002) 5873;
P. Van Der Voort, P.I. Ravikovitch, K.P. de Jong, A.V. Neimark, A.H. Janssen, M.
Benjelloun, E. Van Bavel, P. Cool, B.M. Weckhuysen, E.F. Vansant, Chem. Com-
mun. (2002) 1010.
[34] G. Ricchiardi, A. Damin, S. Bordiga, C. Lamberti, G. Spano, F. Rivetti, A. Zecchina,
J. Am. Chem. Soc. 121 (2001) 11409.
[35] S. Bordiga, S. Coluccia, C. Lamberti, L. Marchese, A. Zecchina, F. Boscherini, F.
Buffa, F. Genoni, G. Leofanti, G. Petrini, G. Vlaic, J. Phys. Chem. 98 (1994) 4125.
[36] D. Trong On, S. Nguyen, V. Hulea, E. Dumitriu, S. Kaliaguine, Microporous Meso-
porous Mater. 57 (2003) 169.
[37] M.C. Daniel, D. Astruc, Chem. Rev. 104 (2004) 239.
[38] C.A. Fyfe, G.C. Gobbi, G.J. Kennedy, J. Phys. Chem. 89 (1985) 277.
[39] Z. Luan, H. He, W. Zhou, J. Klinowski, J. Chem. Soc., Faraday Trans. 94 (1998)
979.
[40] L. Marchese, E. Gianotti, V. Dellarocca, T. Maschmeyer, F. Rey, S. Coluccia, J.M.
Thomas, Phys. Chem. Chem. Phys. 1 (1999) 585.
[41] R.J. Matyi, L.H. Schwartz, J.B. Butt, Catal. Rev. Sci. Eng. 29 (1987) 41.
[42] R. Zanella, L. Delannoy, C. Louis, Appl. Catal. A Gen. 291 (2005) 62.
[43] F. Moreau, G.C. Bond, A.O. Taylor, J. Catal. 231 (2005) 105.
[44] P. Inego, G. Aprile, M.D. Serio, D. Gazzoli, E. Sactacesaria, Appl. Catal. A Gen. 178
(1999) 97.
[16] S. Gontier, A. Tuel, J. Catal. 157 (1995) 124.
[17] T.A. Nijhuis, B.J. Huizinga, M. Makkee, J.A. Moulijn, Ind. Eng. Chem. Res. 38
(1999) 884.
[18] E.E. Stangland, K.B. Stavens, R.P. Andres, W.N. Delgass, J. Catal. 191 (2000) 332.
[19] B.S. Uphade, M. Okumura, S. Tsubota, M. Haruta, Appl. Catal. A Gen. 190 (2000)
43.
[45] G.C. Bond, C. Louis, D.T. Thompson, in: G.J. Hutchings (Ed.), Catalysis by Gold,
in: Catal. Sci. Ser., vol. 6, Imperial College Press, London, 2006.