Nanocrystalline gold supported on Fe-, Ti- and Ce-modified hexagonal mesoporous silica as a catalyst for the aerobic oxidative esterification of benzyl alcohol

Article abstract of DOI:10.1016/j.apcata.2011.02.028

The materials containing gold nanoparticles supported on pure and Ce, Ti, or Fe-modified hexagonal mesoporous silica were prepared and their structural and electronic properties were studied by XRD, TEM, XPS, and N₂ adsorption techniques. The materials were shown to be effective heterogeneous catalysts for the liquid-phase aerobic oxidation of benzyl alcohol in methanol solutions. It was found that the modifiers significantly improve the catalytic properties of supported gold particles and allow performing the selective one-pot oxidative esterification of benzyl alcohol resulting in methyl benzoate. The modified catalysts exhibited high activity (turnover frequencies of up to ca. 1000 h^-1), selectivity to methyl benzoate (up to 95%), and stability (turnover numbers of up to ca. 4300). Ce and Ti were found to be more effective promoters as compared with Fe in terms of catalyst stability.

Full text of DOI:10.1016/j.apcata.2011.02.028

Applied Catalysis A: General 397 (2011) 145–152
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Nanocrystalline gold supported on Fe-, Ti- and Ce-modified hexagonal
mesoporous silica as a catalyst for the aerobic oxidative esterification of benzyl
alcohol
Luciana A. Parreiraa, Nina Bogdanchikovab, Alexey Pestryakovc, T.A. Zepedab, Inga Tuzovskayac,d
,
M.H. Faríasb, Elena V. Gusevskayaa,∗
a Departamento de Química, Universidade Federal de Minas Gerais, 31270-901 Belo Horizonte, MG, Brazil
b Centro de Nanociencias y Nanotecnología, Universidad Nacional Autónoma de México, Ensenada, B.C. 22800, Mexico
c Tomsk Polytechnic University, Tomsk 634034, Russia
d University of Twente, 7500 AE Enschede, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
The materials containing gold nanoparticles supported on pure and Ce, Ti, or Fe-modified hexagonal
mesoporous silica were prepared and their structural and electronic properties were studied by XRD,
TEM, XPS, and N2 adsorption techniques. The materials were shown to be effective heterogeneous cata-
lysts for the liquid-phase aerobic oxidation of benzyl alcohol in methanol solutions. It was found that the
modifiers significantly improve the catalytic properties of supported gold particles and allow performing
the selective one-pot oxidative esterification of benzyl alcohol resulting in methyl benzoate. The mod-
ified catalysts exhibited high activity (turnover frequencies of up to ca. 1000 h−1), selectivity to methyl
benzoate (up to 95%), and stability (turnover numbers of up to ca. 4300). Ce and Ti were found to be more
effective promoters as compared with Fe in terms of catalyst stability.
Received 30 November 2010
Received in revised form 17 February 2011
Accepted 20 February 2011
Available online 24 February 2011
Keywords:
Oxidative esterification
Benzyl alcohol
Gold supported catalysts
Mesoporous metallosilicates
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
metal catalysts were reported to promote the oxidation of alcohols
with molecular oxygen as the final oxidant. These processes occur
Recent discoveries have demonstrated the catalytic activity of
nanoparticles of gold in various chemical transformations, such as
low temperature CO oxidation, hydrogen oxidation, hydrochlorina-
tion of acethylene, and some other processes [1–6]. These results
are truly remarkable considering that metallic gold is known to
be virtually non-reactive in a bulk form. The explosive progress of
nanoscience over the past two decades has stimulated a constant
growth of publications reporting the catalytic activity of nanosized
gold in organic reactions [4–6]. In particular, supported gold cat-
alysts were found to be promising substitutes for platinum group
metals in the aerobic oxidations of alcohols [7,8].
The oxidation of alcohols to carbonyl compounds is a vital
organic transformation in both laboratory and industrial synthetic
chemistry. It is considered as one of the most challenging reac-
tions in green chemistry [8] as many commonly used processes still
require stoichiometric amounts of expensive and/or toxic heavy
metal oxidants and, therefore, produce large amounts of waste.
Several homogeneous [9,10] and heterogeneous [11–14] transition
with high atom efficiency and give water as the only byproduct,
which is especially relevant to a green chemistry concept. Among
the catalysts reported for the aerobic oxidation of alcohols, materi-
als containing supported gold nanoparticles are considered as the
most promising ones due to their high activity and stability [4–8].
However, the oxidation of alcohols over gold catalysts under
mild conditions requires the presence of base (typically NaOH),
often in over-stoichiometric amounts [5]. In these reactions,
sodium salts of carboxylic acids rather than acids themselves are,
therefore, obtained as products. To avoid the use of a base in large
amounts it has been recently suggested to perform the reaction in
methanol solutions [15–20]. In these systems, methyl esters instead
of carboxylic acids are formed as final products. Thus, the base is
not consumed stoichiometrically and only catalytic amounts are
needed to promote the reaction.
Methyl esters are commercially valuable chemicals used as fla-
voring agents, solvents, diluents, extractants, etc. Methyl esters are
traditionally prepared by two-step procedures, which include the
synthesis of carboxylic acids or their derivatives, such as chlorides
and anhydrides, and further esterification of these compounds with
methanol. Most of the reported examples of the direct transforma-
tion of alcohols into methyl esters employ stoichiometric oxidants,
∗
Corresponding author. Tel.: +55 3134095741; fax: +55 3134095700.
E-mail address: elena@ufmg.br (E.V. Gusevskaya).
0926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2011.02.028

146
L.A. Parreira et al. / Applied Catalysis A: General 397 (2011) 145–152
such as molecular iodine [21] or hypochlorite [22]. Only a few cat-
alysts capable to promote the oxidative esterification of alcohols
with molecular oxygen as a final oxidant have been reported, with
all of them being based on gold nanoparticles, as far as we know
[15–20].
precursor) as described previously [28–30]. Dodecylamine was
employed as a surfactant and mesitylene as a swelling organic agent
[31]. The Ce-, Fe-, and Ti-modified HMS materials (denoted as HMS-
Ce, HMS-Fe, and HMS-Ti, respectively) were prepared by a direct
synthesis via co-precipitation using cerium(III) nitrate hexahy-
drate, iron(III) nitrate nonahydrate, and tetrabuthyl orthotitaniate
as Ce, Fe, and Ti precursors, respectively. The Si/M (M = Ce, Fe, Ti)
atomic ratio of 40 was used in the preparation of all materials. The
reaction products were filtered, washed with distilled water, and
dried at room temperature for 24 h and at 100 ◦C for 2 h. The tem-
plate was then removed by calcination at 550 ◦C for 3.5 h in air, at
In gold catalysis, the nature of the support is known to be a
crucial factor because the catalyst performance strongly depends
on the size and morphology of gold particles as well as on their
interaction with the support [5]. A wide variety of solids has been
tested in an attempt to stabilize gold nanoparticles and to influence
their catalytic properties. The aim of the present work is to explore
the application of hexagonal mesoporous silica (HMS) as a support
for the preparation of gold catalysts for the oxidation of alcohols.
HMS materials have a strong potential in catalysis due to their
high surface area and uniform mesoporous structure [23]. A high
hydrophobicity of HMS is a particularly advantageous property for
gold-catalyzed aerobic oxidations because these reactions produce
water and are usually performed in aqueous or alcoholic solutions.
HMS materials are prepared by a neutral S0I0 templating route
[24,25], in which primary amines are used as neutral organic surfac-
tants (S0). In the course of the material formation, a self-assembly
involves hydrogen-bonding interactions between S0 and a neutral
inorganic precursor (I0). Due to the absence of strong electrostatic
interactions, the organic phase can be completely removed from
HMS materials by solvent extraction preventing the partial degra-
dation of a mesoporous structure at calcination, as it often happens
with MCM-41 materials produced with ionic surfactants [26]. HMS
materials show a wormhole like pore structure and are less ordered
than MCM-41; however, they are more stable thermally and allow
faster diffusion of reactants due to shorter mesopores.
The techniques used for the synthesis of HMS opens the oppor-
tunity for the one-pot introduction of transition metal ions into the
material via either a direct inclusion into the silica framework or
the formation of individual highly dispersed metal oxides embed-
ded in the HMS matrix [26]. Such metal modified HMS materials
exhibit different redox and acidic properties and can show higher
stability and activity in catalytic processes. For example, it has been
reported that the incorporation of Ti into HMS enhances the activity
of Co–Mo/HMS catalysts in the hydrodesulfurization of dibenzoth-
iophene [27]. In the case of gold catalysts, it is especially valuable
that the modification of the HMS support can help to stabilize gold
nanoparticles preventing their aggregation. It has been recently
reported that Ce-containing Au/HMS materials are more active in
the oxidation of CO than their Ce-free counterpart, probably, due
to higher gold dispersion and higher redox ability of the support
[28]. In the liquid phase hydrogenation of biphenyl, Au/HMS-Ce
and Au/HMS-Fe catalysts have also exhibited higher stability to
sintering compared to the unmodified Au/HMS catalyst [29]. The
Au/HMS-Fe material has demonstrated, in addition, significantly
higher activity, probably, due to higher exposure of Au␦+ species
[29].
a heating rate of 2.5 ◦C min−1
.
2.2. Synthesis of Au(en)2Cl3
Ethylenediamine (0.45 mL) was slowly added to the solution of
HAuCl4·3H2O (1.0 g) in deionized water (10 mL). The transparent
brown solution was magnetically stirred for 30 min. Then, ethanol
(70 mL) was added to this solution resulting in the formation of a
white precipitate. The suspension was stirred for additional 20 min
and filtered. The solid was washed with ethanol and dried at 40 ◦C
in vacuum overnight.
2.3. Catalyst preparation
Au(en)2Cl3 (0.05 g) was dissolved in water (50 mL) and the pH
of the solution was adjusted to 10.0 by the addition of the aqueous
solution of NaOH (5.0 wt%). Then, HMS or HMS-M (1.0 g) was added.
The pH of the solution, which dropped after the addition of HMS,
was re-adjusted to ca. 10.0. The suspension was stirred at 60–70 ◦C
for 2 h and filtered. The solid was washed with distilled water and
dried in vacuum at 70 ◦C for 5 h. Finally, the catalysts were ther-
mally treated at 300 ◦C in air for 1.5 h to ensure a complete removal
of the organic template and then in hydrogen for 1.5 h.
2.4. Catalyst and support characterization
2.4.1. Chemical analysis
Metal contents were determined by inductively coupled plasma
atomic emission spectroscopy (ICP-AES) on a Perkin Elmer Optima
3300DV instrument. Solid samples were first digested in a mix-
ture of HF, HCl, and HNO3 in a microwave oven for 2 h. Then, the
solutions were diluted with deionized water and analyzed.
2.4.2. High resolution transmission electron microscopy (HRTEM)
HRTEM studies were carried out using a JEM 2100F microscope
operating with a 200 kV accelerating voltage. The samples were
ground into a fine powder and dispersed ultrasonically in hexane
at room temperature. Then, a drop of the suspension was put on
a lacey carbon-coated Cu grid. At least ten representative images
were taken for each sample. In order to obtain statistically reliable
information, the length of ca. 500 particles was measured. A parti-
cles size distribution was obtained by counting more than ca. 500
particles for each sample. Particle size diameters were calculated
from the following equation: davg = ꢀ(nidi)/ꢀni.
In the present work, we have prepared the materials contain-
ing gold nanoparticles supported on pure and metal modified HMS
(HMS-M, where M = Ce, Ti, Fe) and studied their catalytic behavior
in the aerobic oxidative esterification of benzyl alcohol. Benzyl alco-
hol was chosen as a test reactant for this reaction. To our knowledge,
the application of gold catalysts supported on HMS to the oxidation
of alcohols has not yet been studied.
2.4.3. X-ray diffractometry (XRD)
The catalysts were characterized by powder XRD according
to the step-scanning procedure (step size 0.02◦; 0.5 s) with a
computerized Seifert 3000 diffractometer, using Ni-filtered CuK␣
(ꢁ = 0.15406 nm) radiation and a PW 2200 Bragg-Brentano ꢂ/2ꢂ
goniometer equipped with a bent graphite monochromator and an
automatic slit. The assignment of crystalline phases was based on
the JPDS powder diffraction file cards.
2. Experimental
2.1. Support preparation
HMS was synthesized by a neutral S0I0 templating route
(S0—neutral primary amine surfactants; I0—neutral inorganic

L.A. Parreira et al. / Applied Catalysis A: General 397 (2011) 145–152
147
Table 1
Textural properties of the supports and Au/HMS catalysts.a
1000
Sample
SBET (m2 g−1
)
NSBET
Average pore
Vtotal (cm3 g−1
)
diameter (nm)
HMS
HMS-Ce
HMS-Fe
767
667
847
795
687
572
726
714
1
1
1
1
0.92
0.88
0.88
0.92
7.9
6.4
6.7
8.5
7.1
6.0
6.2
8.2
1.20
0.82
1.02
1.46
1.14
0.62
0.83
1.25
800
600
400
200
HMS-Ti
Au/HMS
Au/HMS-Ce
Au/HMS-Fe
Au/HMS-Ti
a
SBET: BET surface area; NSBET: normalized SBET; Vtotal: total pore volume.
Au/HMS-Ti
Au/HMS
2.4.4. X-ray photoelectron spectroscopy (XPS)
XPS measurements were performed on a VG Escalab 200R
electron spectrometer equipped with a hemispherical electron ana-
lyzer, using a MgK␣ (hꢃ = 1253.6 eV, 1 eV = 1.603 × 10−19 J) X-ray
source. After degassing at 10−6 mbar, the samples were transferred
to an ion-pumped analysis chamber, in which the residual pres-
sure was kept below 4 × 10−9 mbar during the data acquisition.
The binding energy (BE) of the C 1s peak at 284.5 eV was taken as
an internal standard. The accuracy of BE values was 0.1 eV. Peak
intensities were estimated by calculating the integral of each peak
after subtracting an S-shaped background and fitting the exper-
imental peak to a combination of Lorentzian/Gaussian lines of
variable proportions.
Au/HMS-Ce
Au/HMS-Fe
0
0,0
0,2
0,4
0,6
0,8
1,0
Relative Pressure / P/P0
Fig. 1. Nitrogen adsorption–desorption isotherms for the Au/HMS catalysts at
−196 ◦C.
tion of benzyl alcohol (1.25 M). Products were identified by GC–MS
on a Shimadzu QP2010-PLUS instrument operating at 70 eV.
2.4.5. N2 adsorption–desorption isotherms
The textural properties of supports and catalysts were deter-
mined from nitrogen adsorption–desorption isotherms (−196 ◦C)
recorded with a Micromeritics TriStar 3000 apparatus. Prior to
experiments, samples were degassed at 270 ◦C in vacuum for 5 h.
The volume of the adsorbed N2 was normalized to a standard
temperature and pressure. The specific areas of the samples were
calculated by applying the BET method to the nitrogen adsorp-
tion data within the 0.005–0.25 P/P0 range. An average pore
diameter was calculated by applying the Barret–Joyner–Halenda
method (BJH) to the adsorption and desorption branches of the
N2 isotherms. A cumulative pore volume was obtained from the
isotherms at P/P0 = 0.99.
3. Results and discussion
3.1. Characterization of the catalysts
BET surface areas, average pore sizes, and total pore volumes for
all prepared catalysts and pure supports are presented in Table 1.
The results show that the addition of modifiers only slightly affects
the structure of the HMS support. Moreover, in some cases the
changes fall within an experimental error and no general corre-
lations can be formulated. The presence of Ce decreases the surface
area of HMS; whereas for the HMS-Fe sample the area is higher
than for its unmodified counterpart. The influence of Ti on the sup-
port characteristics is almost negligible. The introduction of Ce and
Fe decreases the average pore size and total pore volume; on the
other hand, the introduction of Fe increases these characteristics.
In the presence of gold, surface areas of all supports slightly
decrease (ca. 10%), which is expected because of partial blocking of
the pore volume. The textural properties of the catalysts were eval-
uated from the nitrogen adsorption–desorption isotherms (Fig. 1).
All catalysts exhibit type IV isotherms and H1 hysteresis loops (in
the UPAC classification), which indicate the presence of textural
mesopores and uniform cylindrical pores with facile pore connec-
tivity. The Au/HMS and Au/HMS-Ti catalysts show two well defined
capillary condensation steps: the first one at P/P0 ≈ 0.45 indicates
the presence of framework mesoporosity, whereas the second one
at P/P0 ≈ 0.85 is related to the textural interparticle mesoporosity
or macroporosity. The adsorption isotherms of the Au/HMS-Ce and
Au/HMS-Fe samples show smaller sharpness indicating lower pore
uniformity in these materials.
In order to verify the evolution in the structural parameters
of the samples after the introduction of gold, the normalized SBET
values were calculated using the following equation:
SBET
of catalyst
NSBET
=
(1 − y) × SBET
of the support
where NSBET is normalized SBET and y is the weight fraction of the
gold phase.
2.5. Catalytic oxidation experiments
The reactions were carried out in a stainless steel reactor
equipped with a magnetic stirrer. In a typical run, a mixture of
benzyl alcohol (2.5 mmol), methanol (2 mL), K2CO3 (if any), and the
catalyst (7.5–15.0 mg, ca. 0.4–0.8 wt%) was transferred in the reac-
tor. The reactor was pressurized with oxygen to the total pressure
of 10 atm and placed in an oil bath; then, the solution was inten-
sively stirred at 110–130 ◦C for the reported time. The reactions
were followed by gas chromatography (GC) (Shimadzu 17 instru-
ment, Carbowax 20 M capillary column) using anisol as an internal
standard. At appropriate time intervals, stirring was stopped and
after catalyst settling aliquots were taken and analyzed by GC. To
ensure correct GC results, the aliquots were diluted fourfold with
methanol before the analysis because of the high initial concentra-
Chemical analysis data (Au, Ce, Fe, and Ti contents) obtained by
ICP and the values for the average size of gold particles and their sur-
face density obtained by HRTEM are presented in Table 2. It can be
seen that the modification of the HMS support by any of the tested
metals significantly improves the efficiency of the gold incorpo-

148
L.A. Parreira et al. / Applied Catalysis A: General 397 (2011) 145–152
Table 2
Au 4f
Chemical analysis and HRTEM data for the Au/HMS catalysts.
84.3
Au/HMS-Fe
85.7
83.0
Catalyst
Au content
(wt%)
Ce, Fe, or Ti
Surface density
of Au particles
(parti-
Average Au
particle size
(nm)
contenta (wt%)
cles/1000 nm2)
Au/HMS
Au/HMS-Ce 2.94
Au/HMS-Fe
Au/HMS-Ti
2.19
4.4 0.3
5.1 0.2
6.5 0.3
5.7 0.2
5.4 0.2
5.9 0.5
4.1 0.4
4.8 0.3
9.6
4.1
2.3
Au/HMS-Ce
Au/HMS-Ti
2.87
2.96
a
Determined for the supports.
ration. In the Au/HMS-M samples, the Au content is higher than in
Au/HMS (ca. 2.9 wt% vs. 2.2 wt%). The density of gold particles on the
surface of the modified samples is also higher. TEM images show
that Au/HMS-Ce contains bigger gold particles as compared with
the unmodified Au/HMS sample, while in Au/HMS-Ti and, espe-
cially, in Au/HMS-Fe the particles are smaller. However, it should be
mentioned that the average size of gold nanoparticles in all samples
is within the region of ca. 4–6 nm and these differences between the
samples are not substantial.
Au/HMS
82
84
86
88
90
92
Thus, we can conclude that the effects of the modifier on the
structural properties of the materials do not follow general ten-
dencies. For example, the presence of Ce decreases the surface area
of the support and increases the average size of gold particles on
the catalyst surface; whereas Fe exerts an opposite influence. The
effect of Ti on the textural properties of the support is less pro-
nounced; however, slightly smaller gold particles can be obtained
on the HMS modified with Ti.
XRD spectra of the catalysts are presented in Fig. 2. All XRD
patterns show the presence of Au(0) microcrystallites, with no
defined peaks characteristic of Ce, Fe, and Ti oxides being observed.
Thus, the ions of modifiers seem to be essentially incorporated in
the silica framework of the HMS material so that the amounts of
extraframework metal oxides (if any) are small. The amorphous
behavior of all samples observed within the 20–30◦ 2ꢂ region is
due to the amorphous silica matrix.
It should be noted that both TEM and XRD methods do not allow
the observation of small gold clusters on the material surface. In our
previous studies [32], we have found that the formation of metal
nanoparticles on the surface is often accompanied by the forma-
tion of small metal clusters with a diameter of less than 1 nm.
A study of these small particles is a complicated task due to the
difficulty of their detection by standard structural methods, such
as XRD and TEM. As a result, the catalytic activity of only TEM-
and XRD-detectable gold nanoparticles is discussed in the major-
Binding Energy / eV
Fig. 3. XP spectra of Au 4f core level for the Au/HMS catalysts before the treatment
in H2.
ity of published studies, whereas the contribution of small clusters
is usually ignored. However, the presence of metal clusters and
their relative amounts can exert a significant effect on the catalytic
properties of supported metals [33,34].
The surface composition of the catalysts and chemical states of
surface gold species were investigated by XPS. The spectra were
registered for the catalysts before the treatment in H2 (after the
treatment at 300 ◦C in air for 1.5 h). These samples are denoted as
the oxidized catalysts. The XP spectra of Ce 3d, Fe 2p, and Ti 2p
core electron levels for the oxidized catalysts (not shown here) are
in agreement with those previously reported for the Au/HMS-M
(M = Ce, Fe, Ti) catalysts used in biphenyl hydrogenation [29]. The
XP spectra of the Ce 3d core level reveal the presence on the surface
of both Ce3+ and Ce4+ species (Ce3+/Ce4+ = 0.32). For the Au/HMS-
Fe sample, the Fe 2p3/2 BE at 710.9 eV and satellite line at 717.3 eV
indicates the presence of Fe3+ ions [35]. The XP spectrum of Ti 2p3/2
shows two components with BEs at 460 and 458.5 eV. The high
intensity component at 460.0 eV is due to the presence of Ti4+ ions
in a tetrahedral coordination, whereas the low intensity Ti 2p3/2
component at 458.5 eV is characteristic of Ti4+ ions in an octahedral
coordination [36].
The XP spectra of the Au 4f core electron levels for the oxidized
Au/HMS catalysts are presented in Fig. 3. A spectra analysis shows
that in all samples gold exists on the surface in at least two forms:
as Au0 or Au␦+ (BE = 84.3 eV) and as Au+ (BE = 85.7 eV). In addition,
a low BE signal at 83.0 eV can be observed in the spectra of all sam-
ples. The latter signal could be assigned to small clusters of metallic
gold formed on the surface which bear some negative charge due
to the electron transfer from the support (Au␦− species). The for-
mation of such negative states of gold was previously suggested
in [37–39]. On the surface of Au/HMS, the relative amount of the
cationic states of gold is much higher than that in the modified
samples, whereas the contribution of the low BE signal is negli-
gible (Fig. 3 and Table 3). On the other hand, in the spectra of all
modified samples, the relative intensity of the signal at 85.7 eV is
low, whereas the signal at 83.0 eV, assigned to small gold clusters,
is much more pronounced than in the unmodified sample (Fig. 3
and Table 3). Thus, it seems that gold supported on the pure HMS
is more susceptible to oxidation as compared to the modified HMS.
Au0
Au/HMS-Ce
Au/HMS-Fe
Au/HMS-Ti
Au/HMS
0
20
40
60
80
100
2θ
Fig. 2. XRD pattern for the Au/HMS catalysts.

L.A. Parreira et al. / Applied Catalysis A: General 397 (2011) 145–152
149
Table 3
Au/HMScatalyst
O2,methanol
Relative signal contributions in the XP spectra for the oxidized Au/HMS catalysts
(Fig. 3).
+
Ph-CH2OH
PhCHO
PhCO(OCH3)
2
3
1
BE (eV)
Au/HMS
Au/HMS-Ti
Au/HMS-Ce
Au/HMS-Fe
83.0
84.3
85.7
5%
33%
62%
29%
67%
4%
25%
57%
18%
33%
55%
12%
Scheme 1. Oxidation of benzyl alcohol 1 into benzaldehyde 2 and methyl benzoate
3.
adsorption of molecular oxygen. Thus, the difference in the oxida-
tive ability of gold could become of a crucial importance for the
performance of the catalyst in the oxidation of alcohols.
Au 4f
83.9
85.0
82.7
Au/HMS-Fe
3.2. Catalytic studies
The catalytic activity of the Au/HMS and Au/HMS-M (M = Ce, Ti,
Fe) samples in the oxidation of benzyl alcohol (1) was examined
in methanol solutions. In all experiments, these materials were
applied together with potassium carbonate as the base co-catalyst,
used in sub-stoichiometric amounts (usually 0.16 equiv.), under
the atmosphere of molecular oxygen. The results are presented in
Tables 5 and 6.
All gold containing samples effectively catalyzed the oxidation
of benzyl alcohol showing high activities, whereas in the presence
of the supports (HMS, HMS-Ce, HMS-Fe, or HMS-Ti) or without
any catalyst added the conversion of benzyl alcohol was negligible.
The reaction gave two major products, benzaldehyde 2 and methyl
benzoate 3 (Scheme 1), with a combined selectivity of 97–100% at
nearly complete conversions. Only trace amounts of benzaldehyde
dimethyl acetal and benzoic acid were detected in reaction solu-
tions. The distribution of the products depended strongly on the
nature of the doping metal in the HMS support and, for a given cat-
alyst, on the reaction time. However, in most of the runs methyl
ester 3 was a predominant product accounting for 60–90% of the
products at the end of the reaction.
Au/HMS-Ce
Au/HMS-Ti
Au/HMS
80
82
84
86
88
90
92
Binding Energy / eV
The reactions occurred with very low catalyst loadings
(0.4–0.8 wt%) in highly concentrated solutions of the substrate in
methanol, which allowed to achieve high concentrations of the
products in final mixtures (up to 20 wt%). Turnover numbers (TONs)
also reached high values (up to ca. 4300) reflecting the high stabil-
ity of the catalysts. It should be mentioned that in the oxidation of
alcohols over palladium systems, TONs rarely exceed a few hundred
units [40]. These features represent the important technological
advantages of the process. It is also important that the catalysts
are solid materials which are insoluble in the reaction mixture and
can be separated from the products by simple centrifugation or
filtration and re-used (see below).
The oxidation of benzyl alcohol in the presence of Au/HMS
occurred smoothly at 130 ◦C resulting in a 68% conversion for 2 h to
give almost exclusively aldehyde 2 and methyl ester 3 in compara-
ble amounts (Table 5, run 1). Keeping the mixture under stirring for
more than 2 h resulted in a further substrate conversion up to nearly
80%; however, then the reaction became stagnated. The selectivity
for the methyl ester increased to 60%, indicating that benzalde-
hyde was partially oxidized further to give the ester. On the other
hand, over the Au/HMS-Fe, Au/HMS-Ce, and Au/HMS-Ti catalysts
prepared from the modified supports, nearly complete conversions
were obtained for 4 h under the same conditions (Table 5, runs 2–4).
The results of these runs correspond to the TONs of nearly 2000
with respect to the total amounts of gold in the material. How-
ever, the real efficiency of the surface gold atoms is much higher,
because the dispersion of gold is obviously much less than 100%.
Although all three catalysts on the modified silica have shown sim-
ilar rates of the alcohol conversion under the conditions used in
runs 2–4 (Table 5), there is a significant difference in the product
distribution. The Au/HMS-Ce and Au/HMS-Ti materials promoted
the almost complete oxidation of not only the alcohol but also the
Fig. 4. XP spectra of Au 4f core level for the Au/HMS catalysts after the treatment in
H2.
The reduction of the catalysts by the treatment in H2 resulted
in the significant shifts of the signals in their XP spectra toward
the low BE region. The signals attributed to gold now appear at
82.7, 83.9, and 85.0 eV instead of 83.0, 84.3 and 85.7 eV (Fig. 4 and
Table 4). Although the XP spectra of all reduced samples are rather
similar (Fig. 4), it can be noted a significant increase in the contribu-
tion of the low-energy signal (82.7 eV) in the spectrum of Au/HMS
after the reduction (Table 3 vs. Table 4).
Although the XPS technique did not detect the presence of Au3+
ions in the samples, since a peak at BEs of about 86.7 eV was not
detected, we cannot discard their existence on the catalyst surface,
especially, in the oxidized catalysts considering that during the XPS
measurements the ions can be reduced by the X-ray beam.
Thus, the XPS study showed that the introduction of the modi-
fier in the HMS support changes significantly the redox properties
of supported gold as well as its electronic state. These differences
can influence the catalytic behavior of the sample, especially, under
oxidative conditions. In our work, the reduced samples, in which
the electronic state of supported gold seems to be similar, have
been used for the catalytic tests. However, under the reaction con-
ditions, gold species should be oxidized, at least partially, due to the
Table 4
Relative signal contributions in the XP spectra for the Au/HMS catalysts (Fig. 4).
BE (eV)
Au/HMS
Au/HMS-Ti
Au/HMS-Ce
Au/HMS-Fe
82.7
83.9
85.0
25%
50%
25%
35%
49%
16%
25%
59%
16%
30%
53%
17%

150
L.A. Parreira et al. / Applied Catalysis A: General 397 (2011) 145–152
Table 5
Au-catalyzed oxidation of benzyl alcohol (1) in methanol solutions.a
Run
Catalyst
Time (h)
Conversion (%)
Product selectivity (%)
Aldehyde 2
TONb
Ester 3
1
Au/HMS
2
4
2
4
2
4
2
4
2
4
68
76
88
94
87
96
89
98
80
95
48
38
48
40
14
12
20
10
31
18
51
60
50
57
86
88
79
90
68
82
1790
1640
2040
2110
1960
2
Au/HMS-Fe
Au/HMS-Ce
Au/HMS-Ti
Au/HMS-Ti
3
4c
5d
a
Conditions: Substrate (2.5 mmol); catalyst (15.0 mg; ca. 0.8 wt %; Au: ca. 0.08 mol%); K2CO3 (0.4 mmol); methanol (2.0 mL); 130 ◦C; 10 atm (O2). Conversion and selectivity
were determined by GC.
b
TON is calculated as a ratio between the amounts of 2 and 3 formed and the total amount of Au, considering that 3 is formed from 1 via 2 and both steps are catalyzed
by Au: TON = [n(2) + 2n(3)]/n (Au), where n represents the amount of the indicated compound or Au in moles.
c
After reaction, the catalyst was separated by centrifugation, the supernatant was added with a fresh portion of substrate and allowed to react on for 3 h, with no conversion
observed thereupon. The separated catalyst was reused without loss of activity and selectivity.
d
110 ◦C.
primarily formed aldehyde giving the ester. After the 4-h reactions
in runs 3 and 4, selectivities for ester 3 were nearly 90%. On the other
hand, the Au/HMS-Fe and Au/HMS samples produced the aldehyde
and the ester in comparable amounts (runs 2 and 1, Table 5). Thus,
gold nanoparticles supported on the unmodified or Fe-modified
silica are less effective catalysts for the oxidative esterification of
benzyl alcohol compared to the Ti- and Ce-modified counterparts.
Namely, Au/HMS-Fe and Au/HMS seem to be less effective in the
second step of this process: the oxidation of the primarily formed
aldehyde to ester, at least under the conditions used in these runs.
Although it is unlikely that metallic gold could be soluble in
methanol, we have tested leaching of the active component from
the catalyst. To prove the absence of any contribution of homo-
geneous catalysis, the Au/HMS-Ti catalyst was removed from the
reaction system after the completion of the reaction in run 4
(Table 5), fresh substrate was added to the supernatant and the
reaction was allowed to proceed further. No activity was observed
in this experiment, which indicates that Au did not leach from silica
into the reaction medium under the conditions used. The catalyst
can be recycled. After run 4 (Table 5), the catalyst was re-used after
washing with hexane without any loss in its activity and selectivity.
The reactions in runs 1–4 (Table 5) were relatively fast showing
nearly 90% conversions for 2 h, when the first aliquots were taken
for the GC analysis. To compare the catalytic performance of dif-
ferent catalysts we decreased the reaction rate by decreasing the
temperature and then the catalyst amounts. At 110 ◦C, the reaction
was still too fast giving 80% conversion for 2 h (Table 5, run 5). In
this run, it can be clearly seen the increase in ester selectivity with
the reaction time. The result is consistent with a two-step mech-
anism involving the oxidation of alcohol to aldehyde and further
oxidation of aldehyde to ester (see below). The individual selectiv-
ities to ester 2 and aldehyde 3 were different at 110 and 130 ◦C at
almost the same conversions. It probably reflects a different effect
of temperature on the rates of two steps of the whole process. How-
ever, the combined selectivity for the aldehyde and ester in both
runs was ca. 100%.
The kinetic curve for the reaction at 110 ◦C with a half amount
of Au/HMS-Ti (Table 6, run 6) is shown in Fig. 5. The reaction was
almost completed in 6 h giving benzaldehyde and methyl benzoate
in a nearly quantitative combined yield. The ester was a predomi-
nant product at the end of the reaction (95% selectivity), whereas
at 42% conversion, it accounted only for ca. 50% of the mass bal-
Table 6
Au-catalyzed oxidation of benzyl alcohol (1) in methanol solutions.a
Run
Catalyst
Time (h)
Conversion (%)
Product selectivity (%)
Aldehyde 2
TONb
TOFc (h−1
)
Ester 3
6
7
8
9
Au/HMS-Ti
Au/HMS-Ce
Au/HMS-Fe
Au/HMS
1
4
6
1
4
6
1
4
6
1
4
6
6
6
43
91
96
43
94
98
42
69
70
23
58
60
58
<1
47
10
4
60
12
5
54
40
39
66
50
51
54
–
50
90
95
40
88
94
46
60
60
33
49
48
45
–
4260
4280
2520
950
995
905
2625
2490
–
744
760
–
10d
11e
Au/HMS
Au/HMS-Ti
a
Conditions: Substrate (2.5 mmol); catalyst (7.5 mg; ca. 0.4 wt%; Au: ca. 0.04 mol%); K2CO3 (0.4 mmol); methanol (2.0 mL); 110 ◦C; 10 atm (O2). Conversion and selectivity
were determined by GC.
b
TON is calculated as a ratio between the amounts of 2 and 3 formed and the total amount of Au, considering that 3 is formed from 1 via 2 and both steps are catalyzed
by Au: TON = [n(2) + 2n(3)]/n (Au), where n represents the amount of the indicated compound or Au in moles.
c
Initial rate of the substrate conversion per mol of Au (initial TOF).
K2CO3—0.6 mmol.
In the absence of K2CO3.
d
e

L.A. Parreira et al. / Applied Catalysis A: General 397 (2011) 145–152
151
ified samples (TOF > 900 h−1). In addition, the reaction becomes
stagnated at ca. 60% conversion (TON = 2625) giving the equal
amounts of the aldehyde and the ester (Table 6, run 9 and Fig. 5).
In attempts to complete the reaction and to clarify the origin of
the deactivation, we increased one-and-a-half times the amount
of base (Table 6, run 10 vs. run 9). However, the reaction stagna-
tion has been observed again at 60% conversion, indicating that
the supported gold particles are no more able to activate molecular
oxygen not because of the lack of base. It should be emphasized that
a small amount of base is really necessary for the reaction to pro-
ceed since in the absence of base no conversion of the alcohol has
been observed over Au/HMS-Ti (Table 6, run 11 vs. run 6). The base
co-catalyst is considered to be necessary for the first abstraction of
hydrogen from the substrate [5].
100
80
60
40
20
Au/HMS-Ce
Au/HMS-Ti
Au/HMS-Fe
Au/HMS
0
0
1
2
3
4
5
6
Thus, the results obtained reveal that the catalysts prepared
from the HMS modified by any of tested metals are more active in
the oxidative esterification of benzyl alcohol. The presence of the
doping metal ions in the HMS framework influences the catalytic
properties of the supported gold. The modifiers have changed the
morphological properties of the HMS supports as well as those of
the gold catalysts (Table 1). However, these changes are not signifi-
cant and no direct correlations with the catalyst behavior have been
observed. The differences in the average size of the gold particles
are also not significant (Table 2). Thus, the activation effect of the
doping metals should be associated with other factors, such as dop-
ing metal influence on the electronic state of gold, redox properties
of gold nanoparticles, and/or their aggregation tendency.
According to the XPS data, the oxidized Au/HMS-M materials
contain larger amounts of small gold clusters (signals at 83.0 eV,
Fig. 3) than the unmodified Au/HMS sample. The modifiers seem
to favor the formation of small gold clusters under oxidative con-
ditions, at the expense of bigger gold nanoparticles. Gold atoms in
these clusters probably bear some negative effective charge due to
a strong interaction with the oxygen atoms of the modified surface
acting as Lewis basic sites. It is reasonable to expect that small gold
clusters are also active in the oxidation of alcohol and their relative
amounts should influence the catalytic behavior of the material. In
addition, higher susceptibility of gold on the unmodified surface
to oxidation could also affect the catalytic activity of the Au/HMS
sample.
Among the studied modifiers, Ce and Ti were found to be more
effective promoters as compared to Fe in terms of catalyst stabil-
ity (Table 6 and Fig. 5). The Au/HMS-Ce and Au/HMS-Ti materials
perform a deeper oxidation of benzyl alcohol resulting in methyl
benzoate as a virtually exclusive final product. On the other hand,
the reactions over Au/HMS-Fe and Au/HMS give benzaldehyde and
methyl benzoate in comparable amounts.
The mechanism proposed for the oxidative esterification of
primary alcohols involves the formation of hemiacetal as a key
intermediate as illustrated in Scheme 2 [16,19]. The alcohol is
first oxidized to give the aldehyde as an intermediate product.
The base co-catalyst seems to facilitate the abstraction of the first
hydrogen from the substrate [5]. A condensation reaction between
the aldehyde and alcohol results in a hemiacetal, whose further
dehydrogenation gives the ester. Small amounts of benzaldehyde
dimethyl acetal detected in the reaction solutions could be formed
from the hemiacetal due to its interaction with a second molecule of
methanol. Another minor product, benzoic acid, could be the result
Time / h
Fig. 5. Oxidation of benzyl alcohol over the gold catalysts. Conditions: substrate
(2.5 mmol), catalyst (0.4 wt%), K2CO3 (0.4 mmol), methanol (2.0 mL), 110 ◦C, 10 atm
(O2).
ance. A behavior of other materials under the conditions used in
run 6 (Table 6) has confirmed a remarkable effect of the doping
metal on both catalytic activity and stability of supported gold in
the oxidation of benzyl alcohol. The Au/HMS-Ce material has pro-
moted the oxidative esterification of alcohol with efficiency similar
to that of Au/HMS-Ti (Table 6, run 7 and Fig. 5). The data obtained
with Au/HMS-Ce and Au/HMS-Ti correspond to the TONs of more
than 4200 and to the turnover frequency (TOF) of ca. 1000 h−1 in
the 0–90% conversion range. These results compare well with most
of the reported catalysts and put the prepared materials among
the best gold catalysts developed for the oxidative methoxylation
of benzyl alcohol [16,18–20]. It should be mentioned that reaction
rates were not affected by changes in the intensity of stirring.
The reaction with the Au/HMS-Fe catalyst begins at almost the
same rate as with Au/HMS-Ti and Au/HMS-Ce (TOF = 905 h−1) and
shows a similar performance until ca. 50% conversion (Table 6, run
8 and Fig. 5). Thus, the initial activity of gold nanoparticles is similar
for all three catalysts regardless on the doping metal. However, the
reaction with Au/HMS-Fe becomes stagnated at ca. 70% conversion
after 3 h. Moreover, a product distribution is different from that for
Au/HMS-Ti and Au/HMS-Ce. Benzaldehyde accounts for ca. 40% of
the products after the 4-h reaction over Au/HMS-Fe and is not being
converted further to the ester at longer reaction times. This result
implies that the catalyst is losing its activity under the reaction
conditions. In other words, despite Au/HMS-Fe is also a highly effi-
cient catalyst for the oxidation of benzyl alcohol promoting more
than 2500 catalytic cycles, it is not stable to the same extent as
Au/HMS-Ce and Au/HMS-Ti and does not complete the reaction
under the conditions used in this run. It should be noted that in
this run the Au/HMS-Fe catalyst lost the activity in the oxidation of
both the alcohol and primarily formed aldehyde as the conversion
of both the substrate and the aldehyde became stagnated. Such
a behavior could be related to less stability of gold nanoparticles
in Au/HMS-Fe to aggregation under the reaction conditions due to
the weaker support–catalyst interaction as compared to Au/HMS-Ti
and Au/HMS-Ce.
The initial activity of the Au/HMS catalyst calculated per mol of
Au (initial TOF = 744 h−1) is significantly lower than that of the mod-
Au/HMS-M
CH3OH
Au/HMS-M
O2
OCH3
OH
OCH3
O
Ph-CH2OH
PhCHO
PhCH
PhC
O2
2
1
3
hemiacetal
Scheme 2. A proposed pathway for the oxidation of benzyl alcohol.

152
L.A. Parreira et al. / Applied Catalysis A: General 397 (2011) 145–152
of the dehydrogenation of a hydrate formed by the interaction of
the aldehyde with trace water instead of methanol. The oxidation
of primary alcohols to carboxylic acids over gold catalysts is usually
presented by a similar mechanism which involves the intermediate
formation of an aldehyde followed by the formation of a hydrate
and subsequent oxidation of the latter to a carboxylic acid [16].
[3] A.S.K. Hashmi, G.J. Hutchings, Angew. Chem. Int. Ed. 45 (2006) 7896–7936.
[4] C. Della Pina, E. Falletta, L. Prati, M. Rossi, Chem. Soc. Rev. 37 (2008) 2077–2095.
[5] G.J. Hutchings, Chem. Commun. (2008) 1148–1164.
[6] A. Corma, H. Garcia, Chem. Soc. Rev. 37 (2008) 2096–2126.
[7] L. Prati, M. Rossi, J. Catal. 176 (1998) 552–560.
[8] A. Abad, P. Concepcion, A. Corma, H. Garcia, Angew. Chem. Int. Ed. 44 (2005)
4066–4069.
[9] J. Muzart, Tetrahedron 59 (2003) 5789–5816.
[10] M.J. Schultz, M.S. Sigman, Tetrahedron 62 (2006) 8227–8241.
[11] K. Mori, T. Hara, T. Mizugaki, K. Ebitani, K. Kaneda, J. Am. Chem. Soc. 126 (2004)
10657–10666.
4. Conclusions
[12] K. Yamaguchi, N. Mizuno, Angew. Chem. Int. Ed. 41 (2002) 4538–4542.
[13] T. Mallat, A. Baiker, Catal. Today 19 (1994) 247–283.
[14] T. Nishimura, N. Kakiuchi, M. Inoue, S. Uemura, Chem. Commun. (2000)
1245–1246.
[15] T. Hayashi, T. Inagaki, N. Itayama, H. Baba, Catal. Today 117 (2006) 210–213.
[16] I.S. Nielsen, E. Taarning, K. Egeblad, R. Madsen, C.H. Christensen, Catal. Lett. 116
(2007) 35–40.
[17] E. Taarning, A. Theilgaard Madsen, J.M. Marchetti, K. Egeblad, C.H. Christensen,
Green Chem. 10 (2008) 408–414.
[18] S.K. Klitgaard, A.T. DeLa Riva, S. Helveg, R.M. Werchmeister, C.H. Christensen,
Catal. Lett. 126 (2008) 213–217.
[19] F.-Z. Su, J. Ni, H. Sun, Y. Cao, H.-Y. He, K.-N. Fan, Chem. Eur. J. 14 (2008)
7131–7135.
[20] R.L. Oliveira, P.K. Kiyohara, L.M. Rossi, Green Chem. 11 (2009) 1366–1370.
[21] N. Moria, H. Togo, Tetrahedron 61 (2005) 5915–5925.
[22] C.E. McDonald, L.E. Nice, A.W. Shaw, N.B. Nestor, Tetrahedron Lett. 34 (1993)
2741–2744.
Nanosized gold supported on HMS is an effective catalyst for
the liquid-phase aerobic oxidation of benzyl alcohol. The modifica-
tion of the support with Ce, Ti, or Fe ions improves significantly the
catalytic performance of the materials. Ce and Ti are more effective
promoters as compared to Fe in terms of catalyst stability. Although
the modifier changes the morphological properties of the mate-
rials, no direct correlations with the catalyst behavior have been
observed. The Au/HMS-Ce and Au/HMS-Ti materials efficiently per-
form the one-pot aerobic oxidative esterification of benzyl alcohol
exhibiting high activity, selectivity to methyl benzoate and stabil-
ity. The significant practical advantages of the process are the use
of easily recoverable heterogeneous catalysts in very low loadings
and molecular oxygen as a final oxidant as well as the possibility
to obtain the desired product in high final concentrations. Large
pore HMS catalysts could be especially important for the conver-
sion of bulky molecules in the fine chemicals industry. Therefore,
our further studies will be targeted toward the application of the
Au/HMS catalysts for the oxidation of bulky biomass-based alcohols
within our ongoing project aimed at adding value to the natural
ingredients of renewable essential oils.
[23] A. Corma, Chem. Rev. 97 (1997) 2373–2419.
[24] P.T. Tanev, M. Chibwe, T.J. Pinnavaia, Nature 368 (1994) 321–323.
[25] P.T. Tanev, T.J. Pinnavaia, Science 267 (1995) 865–867.
[26] A. Tuel, Micropor. Mesopor. Mater. 27 (1999) 151–169.
[27] T.A. Zepeda, B. Pawelec, J.L.G. Fierro, T. Halachev, Appl. Catal. B 71 (2007)
223–236.
[28] J.A. Hernandez, S. Gómez, B. Pawelec, T.A. Zepeda, Appl. Catal. B 89 (2009)
128–136.
[29] P. Castan˜o, T.A. Zepeda, B. Pawelec, M. Makkee, J.L.G. Fierro, J. Catal. 267 (2009)
30–39.
[30] T.A. Zepeda, B. Pawelec, J.L.G. Fierro, A. Montesinos, A. Olivas, S. Fuentes, T.
Halachev, Micropor. Mesopor. Mater. 111 (2008) 493–506.
[31] B. Pawelec, P. Castan˜o, T.A. Zepeda, Appl. Surf. Sci. 254 (2008) 4092–4102.
[32] N. Bogdanchikova, A. Pestryakov, M.H. Farias, J.A. Diaz, M. Avalos, J. Navarrete,
Solid State Sci. 7 (2008) 908–914.
[33] A. Simakov, I. Tuzovskaya, A. Pestryakov, N. Bogdanchikova, V. Gurin, M. Avalos,
M.H. Farías, Appl. Catal. A 331 (2007) 121–128.
[34] I. Tuzovskaya, N. Bogdanchikova, A. Simakov, V. Gurin, A. Pestryakov, M. Avalos,
M.H. Farías, Chem. Phys. 338 (2007) 23–32.
[35] C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, G.E. Muilenberg, Handbook
of X-ray Photoelectron Spectroscopy, Perkin Elmer Co., Eden Prairie, MN, 1979,
p. 183.
[36] M.C. Capel-Sanchez, J.M. Campos-Martin, J.L.G. Fierro, M.P. de Frutos, A.P. Polo,
Chem. Commun. (2000) 855–856.
Acknowledgments
The authors would like to express their gratitude to E. Flores,
J.A. Peralta, P. Casillas, I. Gradilla, M. Sainz, C. Gonzalez, M. Vega,
and J. Palomares for their valuable technical assistance with the
experimental work. This research was supported by CNPq, CAPES,
FAPEMIG and INCT-Catálise (Brazil), by CONACYT project No 79062,
PAPIT-UNAM IN100908 (Mexico), and by RFBR grant 09-03-00347-
а (Russia).
[37] F. Boccuzzi, A. Chiorino, M. Manzoli, Surf. Sci. 454 (2000) 942–946.
[38] F. Boccuzzi, A. Chiorino, M. Manzoli, D. Andreeva, T. Tabakova, J. Catal. 188
(1999) 176–185.
[39] T. Tabakova, F. Boccuzzi, M. Manzoli, D. Andreeva, Appl. Catal. A 252 (2003)
385–397.
References
[1] M. Haruta, N. Yamada, T. Kobayashi, S. Iijima, J. Catal. 115 (1989) 301–309.
[2] G.C. Bond, C. Louis, D.T. Thompson, Catalysis by Gold, Imperial College Press,
London, 2006, 366.
[40] A.C. Bueno, J.A. Gonc¸ alves, E.V. Gusevskaya, Appl. Catal. A 329 (2007) 1–6.