Tungstenocene-grafted silica catalysts for the selective epoxidation of alkenes

Article abstract of DOI:10.1016/j.apcata.2019.05.027

Tungstenocene(IV) dichloride was successfully deposited and grafted on the surface of a commercially-available non-ordered silica support via either a liquid-phase or a dry impregnation approach. After a high-temperature calcination step, two W(VI)-grafted silica catalysts, with a metal loading around 1.2–1.7 wt.%, were obtained. They were fully characterized by physisorption and spectroscopic techniques, which evidenced well-dispersed tungsten oxide polyoxo cluster sites, for the catalyst prepared via liquid-phase grafting, and evenly dispersed larger monoclinic tungsten(VI) oxide aggregates, for the catalyst prepared via dry impregnation, respectively. Both W/SiO₂ solids showed moderate to good conversion values in the epoxidation of (R)-(+)-limonene and methyl oleate (up to 68%), in the presence of aqueous hydrogen peroxide, with good selectivity to the desired epoxides (63% and 78%, respectively). The heterogeneous character of the most interesting W/SiO₂ catalyst prepared via dry impregnation was confirmed by a hot centrifugation test and by an extended recovery and reuse of the solid in six catalytic runs.

Full text of DOI:10.1016/j.apcata.2019.05.027

Accepted Manuscript
Title: Tungstenocene-grafted silica catalysts for the selective
epoxidation of alkenes
Authors: Chiara Bisio, Alessandro Gallo, Rinaldo Psaro,
Cristina Tiozzo, Matteo Guidotti, Fabio Carniato
PII:
S0926-860X(19)30240-6
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2019.05.027
APCATA 17093
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
16 December 2018
12 May 2019
24 May 2019
Please cite this article as: Bisio C, Gallo A, Psaro R, Tiozzo C, Guidotti M, Carniato F,
Tungstenocene-grafted silica catalysts for the selective epoxidation of alkenes, Applied
Catalysis A, General (2019), https://doi.org/10.1016/j.apcata.2019.05.027
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Tungstenocene-grafted silica catalysts for the selective epoxidation of
alkenes
Chiara Bisio,a,b Alessandro Gallo,*c,d Rinaldo Psaro,a
a
a
b
Cristina Tiozzo , Matteo Guidotti* and Fabio Carniato
a
CNR-Institute of Molecular Sciences and Technologies, Via C. Golgi 19, Milan, Italy
b
University of Eastern Piedmont “A.Avogadro”, Viale T. Michel 11, Alessandria, Italy
c
SUNCAT Center for Interface Science and Catalysis, Department of Chemical Engineering, Stanford
University, Stanford, California 94305, United States
d
SUNCAT Center for Interface Science and Catalysis, SLAC National Accelerator Laboratory, Menlo Park,
California 94025, United States
(
*) corresponding authors: m.guidotti@istm.cnr.it ; agallo@slac.stanford.edu
dedicated to Prof. Guido Busca on the occasion of his 65th birthday
Graphical Abstract
1

Highlights





- Tungsten(VI)-silica heterogeneous catalysts for the epoxidation of alkenes
- Two novel preparation approaches from tungstenocene(IV) dichloride and commercial silica
- Extended physico-chemical characterisation of the surface W(IV) species
- Good performance as catalysts for the epoxidation of limonene and methyl oleate with H
- Promising robustness of the tungsten-grafted silica catalysts against metal leaching
2
O
2
ABSTRACT
Tungstenocene(IV) dichloride was successfully deposited and grafted on the surface of a commercially-
available non-ordered silica support via either a liquid-phase or a dry impregnation approach. After a high-
temperature calcination step, two W(VI)-grafted silica catalysts, with a metal loading around 1.2-1.7 wt.%,
were obtained. They were fully characterized by physisorption and spectroscopic techniques, which
evidenced well-dispersed tungsten oxide polyoxo cluster sites, for the catalyst prepared via liquid-phase
grafting, and evenly dispersed larger monoclinic tungsten(VI) oxide aggregates, for the catalyst prepared via
dry impregnation, respectively. Both W/SiO solids showed moderate to good conversion values in the
2
epoxidation of (R)-(+)-limonene and methyl oleate (up to 68%), in the presence of aqueous hydrogen
peroxide, with good selectivity to the desired epoxides (63% and 78%, respectively). The heterogeneous
2

character of the most interesting W/SiO catalyst prepared via dry impregnation was confirmed by a hot
2
centrifugation test and by an extended recovery and reuse of the solid in six catalytic runs.
Keywords: Epoxidation; hydrogen peroxide; heterogeneous catalyst; tungsten-silica catalyst;
tungstenocene dichloride; alkenes; XAS.
INTRODUCTION
In the panorama of single-site heterogeneous catalysts, the generation of active metal sites on the surface
of high-surface area inorganic oxides can be achieved by covalently grafting organometallic precursors onto
the solid support and by subsequent thermal treatment, in order to obtain a stable, active, easy-to-handle
and recyclable system [1,2]. The grafting approach allows the synthetic chemist to tune the metal loading,
optimise the surface distribution of active sites and their availability as well as select the most suitable oxidic
support in terms of specific surface area, porosity, morphology and hydrophobic/hydrophilic character.
In the field of heterogeneous oxidation catalysts, a very broad series of effective catalysts have been
prepared anchoring metallic precursors onto previously structured silica supports. Inorganic salts or
organometallic precursors of titanium, vanadium, tantalum, zirconium or niobium proved to be the most
suitable starting compounds and the active sites were mainly generated on mesoporous silica molecular
sieves [3,4,5,6,7]. Titanium-grafted silica catalysts are the first of this class of solids and have been
successfully applied to the selective transformation of alkenes into epoxides, (cyclo)alkanes into alcohols or
ketones, sulphides into sulphoxides, thiophenes into sulphones, alcohols into aldehydes or ketones, arenes
into phenols, phenols into benzoquinones, etc. [8,9]. Titanium-grafted silicate catalysts have shown
promising performance, especially in the presence of bulky and richly-functionalised substrates.
Nevertheless, the optimal performance of titanosilicalite-1 zeotype, TS-1, in selective oxidation is still
unparalleled, in terms of selectivity, activity and robustness of the catalyst structure [10]. Then, some
attention was attracted by vanadium, tantalum and zirconium-grafted species, which, on the contrary,
showed scarce stability to leaching in liquid-phase reaction media (especially for vanadium-based systems)
or limited results in terms of activity [11,12,13,14,15]. More recently, niobium-silicate catalysts displayed
interesting catalytic performance, in water-containing reaction media too, thanks to their enhanced
3

robustness against hydrolysis due to the more optimal geometry of Nb-O-Si bond angles compared to V-O-
Si [16]. The number of studies showing the successful application of Nb-containing systems in the presence
of hydrogen peroxide as an oxidant has therefore rapidly increasing accordingly, in the last years [17,18,19].
Grafted catalysts based on Group 6 organometallic precursors, on the other hand, have been somehow
neglected in this restless quest for the optimal metal-grafted silica catalyst for selective oxidations. This is
undoubtedly due to the scarce resistance to leaching of chromium- and molybdenum-containing systems
and their intrinsic higher activity when the metal sites are dissolved in homogeneous phase, rather than
deposited on the oxidic support [20,21]. In some cases, these catalysts did not show a genuine
heterogeneous nature, especially in the presence of aqueous hydrogen peroxide or when water-containing
reaction media were used [22]. With regard to tungsten, very good results have been recorded, in the last
decades, in the epoxidation of alkenes under homogeneous and/or biphasic conditions [23,24,25,26],
although these systems often show, as a major drawback, a difficult separation and recovery of the spent
catalyst. However, some recent works have paid attention to tungsten-based heterogeneous catalysts for
the selective oxidation of alkenes with hydrogen peroxide too [25,27,28,29,30,31,32,33,34,35]. Tungsten-
silica catalysts prepared by simple impregnation of the metal precursor onto porous siliceous molecular
sieves typically led to uneven surface WO domains, which are thus prone to further aggregation, activity
3
loss and scarce robustness [36,37]. For this reason, the most suitable strategy appeared to obtain isolated W
species grafted onto silica through W–O–Si covalent bonds [38]. Some of them displayed promising
performances in terms of activity and/or selectivity to desired epoxides. Other systems, on the contrary,
showed modest results either in terms of specific activity with values in the range of 10 h-1 (moles of
converted alkene per mole of W per hour) [39] or because of a remarkable loss of active metal in the reaction
medium [40,41,42,43,44,45]. In some successful cases, it was shown that, even though a non-negligible
leaching of tungsten was observed from the solid into the reaction mixture, such W species were not able to
act as homogeneous catalysts by themselves, hence confirming the heterogeneous behaviour of the studied
tungsten catalyst [32].
4

In this aim, in the present work, we describe a detailed spectroscopic study about the nature of tungsten
species deposited onto the mesoporous silica following two post-synthesis approaches: a liquid-phase
grafting and an alternative, less environmentally unfriendly, solventless synthesis protocol based on dry
impregnation, DI, of bis(cyclopentadienyl)tungsten(IV) dichloride, W(Cp)
2
Cl . Finally, the catalytic features of
2
the obtained sites were studied and compared in the liquid-phase epoxidation of two model alkene
substrates: an unsaturated terpene, limonene, and an unsaturated fatty acid methyl ester, methyl oleate.
EXPERIMENTAL DETAILS
Catalyst Preparation
Tungsten-silica catalysts were all prepared by grafting bis(cyclopentadienyl)tungsten(IV) dichloride
(W(Cp)
2 2
Cl ; 99% Strem Chemicals) onto a commercially available amorphous pure silica support (Grace
Davison, Davisil SiO
2
LC60A, 60-200 μm). The silica support was pre-treated under dry air at 500°C for 1 h,
then left at 500°C for 1 h in vacuo and finally cooled to room temperature in vacuo as well, in order to remove
any potential organic contaminant on the surface.
The W/SiO
previously adopted for Ti(Cp)
purity deionized water (MilliQ Academic, Merck Millipore, 18 MΩcm) for 2 h, dried at the rotary
evaporator and then pre-treated in air at 300°C for 1 h and in vacuo overnight at 300°C. W(Cp) Cl
2
-DI catalyst, obtained via dry impregnation, was prepared by adapting a method
2
Cl and Nb(Cp) Cl [18]. The silica support was hydrated with high-
2
2
2
2
2
was finely ground and mixed to the silica support, under inert atmosphere. The solid mixture was
continuously stirred overnight under steady vacuum at room temperature (20°C; in thermostatic bath).
The resulting light brown powder was calcined under dry oxygen at 500°C for 2 h in order to obtain
the final W(VI)-silica catalysts.
The W/SiO
previously described for titanium and niobium [46,47] and adapting it to
bis(cyclopentadienyl)tungsten(IV) dichloride. In this case, W(Cp) Cl was dissolved in anhydrous
CHCl (Prolabo, RP; dried over 3A molecular sieves) under argon. The resulting solution was added
2
–liq catalyst, prepared via liquid-phase grafting, was obtained modifying the method
2
2
3
to the pre-treated silica and stirred for 60 min at room temperature. Freshly distilled triethylamine
5

was then added to the suspension and left under stirring overnight. The solid was filtered on a Büchner
funnel, washed with fresh CHCl and dried for 3 h. It was calcined at 550°C under dry oxygen (80
3
-1
mL min ) for 6 h.
Pure commercial WO
modifications.
3
(Sigma-Aldrich, puriss. 99.9%) was used for comparison with no further
Catalyst Characterization
The tungsten content of the prepared samples was determined by inductively coupled plasma optical
emission spectroscopy (ICAP 6300 Duo, Thermo Fisher Scientific) after mineralization of the
samples in a microwave digestion apparatus (Milestone MLS 1200; maximum power 500 W) with a
mixture of hydrofluoric acid (aq. 40%) and fuming nitric acid.
Diffuse reflectance UV–Visible spectra were collected using a Perkin Elmer Lambda 900
spectrometer equipped with an integrating sphere accessory and using a hand-made quartz cell
-5
allowing analysis both under vacuum conditions (residual pressure 10 mbar) and under controlled
gas atmosphere. Prior the analysis, the samples were dispersed in anhydrous BaSO (10 wt.%).
4
-1
-1
Raman spectra were recorded in the 4000–50 cm range, using a resolution of 4 cm with 2000 scans,
using a RFS 100 Bruker FT-Raman spectrometer equipped with a 1064 nm wavelength excitation
laser. Laser power was set at 100 mW. Raman spectra were elaborated by using OPUS 5.0 software
(
Bruker Optics Company).
X-ray diffractograms (XRD) were collected on unoriented ground powders with a Thermo ARL
XTRA-048 diffractometer with a Cu Kα (k = 1.54 Å) radiation. Diffractograms were recorded at
‘
-1
room temperature with a step size of 0.02° and a rate of 1° 2 min .
physisorption measurements were carried out at 77 K in the relative pressure range from 1 x 10-6
N
2
to 1 P/P by using a Quantachrome Autosorb 1MP/TCD instrument. Prior to analysis, the samples
0
-6
were outgassed at 373 K for 3 h (residual pressure lower than 10 mbar). Apparent surface areas were
determined by using Brunauer–Emmett–Teller equation, in the relative pressure range from 0.01 to
6

0
.1 P/P
0
. Pore size distributions were obtained by applying the NLDFT method (N silica kernel based
2
on a cylindrical pore model applied to the desorption branch).
X-ray Absorption Spectroscopy
The XAS data for the W L
3
and L edges were collected at room temperature and under air at the
1
beamline 4-1 and 4-3 at the Stanford Synchrotron Radiation Lightsource (SSRL) operating with an
energy of 3 GeV and a current of 500 mA. The samples were packed inside the sample holder under
ambient atmosphere and the fluorescence data were collected by a Lytle detector. In order to improve
the signal to noise ration each run has been repeated at least four times.
The collected data were analyzed with Athena and Arthemis software programmes [48] in order to
obtain the coordination number (N), the distance of the scattering atoms (R) and the disorder DW
factor of the neighbour scattering atoms (v. Table S1; Supplementary Material).
Catalytic tests
The epoxidation reaction tests on (R)-(+)-limonene (97%, Sigma) and methyl oleate (99% Aldrich) were
carried out in a round-bottomed glass batch reactor, in an oil bath set at 90°C, with magnetic stirring (ca. 800
rpm) under dry nitrogen atmosphere. The W/SiO
2
catalysts were pre-treated in dry air at 500°C and cooled
was pre-treated either at 120°C in dry air for 1 h and
down to room temperature in vacuo prior to use. WO
3
then in vacuo for 1 h or at 500°C in dry air and cooled down to room temperature in vacuo prior to use.
The catalyst (100 mg) was dispersed in 5 mL of acetonitrile (Aldrich, HPLC grade), as solvent, in the presence
of 1 mmol of substrate (either limonene or methyl oleate) and 2 or 4 mmol of aqueous hydrogen peroxide
(
H
2
O ; 50% Aldrich), as oxidant (oxidant to substrate molar ratio of 2 : 1 or 4 : 1). Tert-butylhydroperoxide,
2
TBHP (Sigma-Aldrich; 5.5 M anhydrous solution in decane) was used in one test with an oxidant to substrate
molar ratio of 1.1 : 1. Samples were withdrawn from the reaction medium at regular intervals and analysed
by GC (Agilent 6890 Series; SP-5 column, 30 m × 0.25 mm; FID detector), by using mesitylene (puriss. ≥ 99%,
Fluka,) or methyl palmitate (≥ 99%, Sigma) as internal standards, for limonene or methyl oleate epoxidation,
respectively. Gas-chromatographic peaks were identified by comparison with peaks of genuine samples of
standard compounds and confirmed by means of GC-MS. For both limonene epoxide and methyl
7

1
epoxystearate, cis:trans ratios were determined by H-NMR analysis (Bruker DRX 300) of the 6 h reaction
-
1
mixture, after solvent removal in vacuo. A standard deviation of ± 2 %, ± 4 % and ± 2 h has to be considered
on average for the conversion, yield and specific activity values, respectively. All tests were replicated at least
two times. No significant oxidation reactivity was recorded in the absence of solid catalyst or in the presence
of the pure silica support (in the absence of tungsten loading). A maximum limonene conversion of ca. 5–6%
was observed after 6 h of reaction, but with no noteworthy formation of epoxide products. Analogously, a
maximum methyl oleate conversion of ca. 7% was recorded after 6 h, with very minor traces of the related
epoxide, methyl epoxystearate. After each test, the presence of residual hydrogen peroxide was checked and
confirmed by iodometric assays.
In order to check the possible leaching of tungsten species, the solid catalyst was removed from the liquid
mixture after 45 min by hot centrifugation (at the same temperature of the reaction mixture) and the filtered
liquid solution was tested for further reaction [49]. In the tests for the recovery of the catalyst, the solid was
separated by filtration and thoroughly washed with acetonitrile and then with methanol (Fluka, HPLC grade).
The filtered solid was dried gently at 110°C, weighed, activated again at 500°C under dry air and then reused
in a new catalytic cycle as described above. In one limonene epoxidation test, hydroquinone (Carlo Erba; p.a.)
was used as a free-radical scavenger in a quasi-equimolar amount (0.8 : 1 mol/mol) with respect to the total
tungsten content in the catalyst and added to the reaction mixture under the same conditions reported
above.
RESULTS AND DISCUSSION
Catalyst preparation
In the present work, at the best of our knowledge, two tungsten-containing silica catalysts were
prepared for the first time via deposition of tungstenocene(IV) dichloride over a cheap and
commercially available non-ordered mesoporous silica support. The selected commercial silica
2
-1
displays a relatively high specific surface area (605 m g ; Table 1), thus providing a sufficient
surface for the deposition of the organometallic precursor. Then, the silica support was first calcined
at 500°C in order to remove from the surface undesired organic impurities that might contaminate the
8

solid. However, such treatment can lead to an extended condensation of the surface silanol groups
and a reduction of the grafting points for the metallocene precursor. So, the support was first re-
hydrated with pure water and then treated at lower temperature, 300°C [50].
At the end the dry impregnation process, the freshly prepared W/SiO -DI sample (before final
2
calcination) showed a uniform greenish-brownish colour, suggesting an even deposition of the metal
species. Then, by means of the calcination under oxygen, the organic cyclopentadienyl ligands were
oxidised and removed. In that step, the catalytically active W(VI) centres were also obtained on the
silica surface. The amount of tungstenocene precursor added into the dry-impregnation mixture was
chosen to obtain a metal loading of ca. 2 wt.%, since 1 to 2 wt.% was observed to be, in general, an
optimal Ti, Nb or W loading for the epoxidation of a large variety of alkenes [51,32]. In addition, a
2
-1
metal loading of 2 wt.% on this kind of silica (with a specific surface area of 605 m g ; Table 1)
approximately corresponds to 0.10 tungsten atoms per square nanometer. This surface tungsten
-2
density is sufficiently lower than the value of ca. 0.3 W atoms nm , which is widely accepted as a
good approach to have an adequate metal loading and a homogeneous dispersion of the metal sites
on the porous silica support [52,53]. Beyond that limit undesirable large aggregates of WO species
x
may occur on the catalyst surface. Nevertheless, the actual metal content showed that only a part of
the expected metal was present (Table 1). A part of the tungsten precursor was lost, likely deposited
on the walls of the glass vessels where the dry impregnation process took place and/or aspired by the
vacuum pump during the evacuation procedure. Anyway, the elemental analysis of different aliquots
of catalyst powder, sampled randomly on the same batch, confirmed the even dispersion of the W
species on silica support.
At the same time, one W/SiO -liq catalyst was prepared via metallocene grafting in liquid phase and
2
used as a comparison with respect to the solid obtained via dry impregnation. It is worth highlighting,
however, that the dry impregnation route is conceptually simpler and does not require the use of a
hazardous and environmentally unfriendly solvent or of a base (i.e, chloroform and triethylamine,
respectively). Anyway, the scarce solubility of W(Cp)
2
Cl in chloroform led to a limited deposition
2
9

of the organometallic precursor and an amount lower than expected of tungsten was therefore grafted
onto the final solid (1.2 wt.% vs. expected 2 wt.%).
Physico-chemical characterization of W-containing silica catalysts
The grafting of bis(cyclopentadienyl)tungsten(IV) dichloride precursor in both W/SiO
2
-liq and
W/SiO -DI samples was monitored by diffuse reflectance UV-Visible spectroscopy (Fig. 1). The DR
2
UV-Visible spectra of the two samples before calcination clearly show a band at 230 nm assigned to
W–O charge-transfer transitions and other two absorptions around 270 nm and ca. 400 nm ascribed
to the transitions of the residual cyclopentadienyl groups attached to the metal site, in agreement with
literature data reported for similar catalysts [41,54]. Similar absorptions at 270 nm and in the 330-
4
70 nm range were observed in the spectrum of the pure bis(cyclopentadienyl)tungsten(IV)
dichloride precursor (Fig. S1; Supplementary Material). However, the wide broadening of these
bands in the spectrum of bis(cyclopentadienyl)tungsten(IV) dichloride may be tentatively ascribed to
hydrophobic molecular aggregation. The amount of tungsten deposited onto the silica surface in the
two samples was measured after calcination of the organic fraction, and was 1.18 for W/SiO
.70 for W/SiO -DI sample, respectively (Table 1).
The textural properties of the two catalysts were investigated by N
2
-liq and
1
2
2
physisorption analysis at 77 K
and the specific surface area (SSA) and pores volume and diameter were determined by Brunauer–
Emmett–Teller method and NLDFT approach. Both samples (prepared by liquid-phase grafting and
dry impregnation) showed N
2
physisorption isotherms of type IV with H2 hysteresis loop. These
isotherms are typically observed for porous materials characterized by an irregular pore size (Fig. 2).
2
-1
The specific surface area of the W/SiO –liq sample is ca. 603 m g and is comparable to the one of
2
the pristine siliceous support. This behaviour is in line with previous observations on niobium-
containing catalysts [18]. Such a result is a clear indication that the liquid grafting did not modify the
textural features of the starting material. Moreover, W/SiO
size distribution in the 2-10 nm range, with a relative pore volume of 1.38 cm g . Such unusual
increase of the specific pore volume with respect to the pristine SiO is probably due to a different
2
–liq catalyst displays a disordered pore
3
-1
2
1
0

aggregation of the secondary larger silica particles/grains during the various liquid-phase grafting
treatments (dispersion in solvents, calcination, etc.). For these non-ordered mesoporous materials, the
porosity is partially linked to particles aggregation, that can be affected by the synthetic procedure
adopted for the catalyst preparation. A completely different scenario was observed for W/SiO -DI
2
2
-1
sample. In this case, a marked reduction of specific surface area to 465 m g was observed after the
impregnation process (Table 1). The differences in the textural properties of the two catalysts may be
associated to a different organization and distribution of the tungsten sites in the solids promoted by
the two synthetic procedures.
The coordination and chemical environment of the W(VI) sites obtained after calcination and
distributed on the silica surface were studied by DR UV-Vis spectroscopy (Fig. 3). The spectra of
both samples show two important bands at ca. 250 and 320 nm ascribed to metal-ligand charge
transfer in partially isolated sites or organized in tungsten oxides with nanometric dimensions with
different coordination geometry (tetrahedral and octahedral), formed during the deposition of
tungsten [55]. Further information about the nature and the coordination chemistry of these metal
sites can be obtained through XAS experiments (see below).
Moreover, the DR UV-Visible spectrum of W/SiO -DI catalyst (Fig. 3b) shows an additional wide
2
band centred at ca. 380 nm, not detectable on the sample prepared via liquid-phase grafting process.
This band is tentatively attributed to the formation of large tungsten oxide domains dispersed on the
surface. The presence of these metal oxide domains was also confirmed by other spectroscopic
(
Raman) and structural analyses (X-ray diffraction).
Transition metal oxides, as they are typically characterized by a high structural symmetry, are suitable
to be characterized by Raman spectroscopy. The Raman spectrum of commercial monoclinic WO
3
-1
sample exhibits some sharp peaks at 805, 715, 325 and 270 cm , assigned to octahedral tungsten
oxide units (Fig. 4) [55]. The first two peaks at high wavenumbers were also detected in the Raman
spectrum of W/SiO
2
-DI sample. The presence of these two peaks reveals the formation of octahedral
microdomains of tungsten oxide promoted by the calcination process (Fig. 4), in agreement with DR
1
1

UV-visible data. Interestingly, these two peaks were completely absent in the spectrum of W/SiO -
2
liq sample, in which the metal centres were isolated and distributed more homogeneously than in the
sample prepared via dry impregnation (Fig. 4).
The crystallographic phase of the tungsten oxide domains in the sample prepared via dry
impregnation was studied by X-ray diffraction technique too. The X-ray profiles of W/SiO
W/SiO -DI samples are reported in Figure 5, where they are compared with the XRD pattern of the
commercial WO sample. The diffractogram of the sample prepared by liquid grafting is
2
-liq and
2
3
characterized by a broad band at 22° 2θ, typical of amorphous silica (Fig. 5, curve a). Furthermore,
no signals typical of large tungsten oxide aggregates were present. These results are in agreement
with the DR UV-Visible spectroscopy data and they suggest that the W(VI) sites of W/SiO -liq were
2
well-distributed on the silica surface. In contrast, the diffractogram of the sample prepared by dry
impregnation (Fig. 5b) shows, along with the wide band centred at 22° 2θ, some weak reflections at
2
3.2°, 23.6°, 24.2° 2θ, which are attributed to the presence of larger tungsten oxide aggregates. By a
direct comparison with the X-ray profile of the commercial monoclinic WO sample, the micro-
domains of oxide distributed on the silica surface for the W/SiO -DI sample can be attributed to a
3
2
monoclinic structure. The set of data explained above indicates that the dry impregnation process,
although leading to a material with higher tungsten metal content, was also associated to the formation
of oxidic domains dispersed over the silica surface.
X-ray Absorption Spectroscopy characterization
The normalized XANES data for L
Supplementary Material). While the main edge features for the L
samples (Fig. 6a), confirming the presence of W(VI) on the silica support, differences were clear in
the pre-edge region (Fig. 6b). The pre-edge feature at ca. 12100 eV in the W L -edge is due to the 2s-
d dipole-forbidden transition for a pure octahedral geometry and its intensity increases with the
1
and L
3
edges are reported in Figure 6, 7, S2 and in table S1
(
1
-edge were similar for all the
1
5
degree of distortion around the W sites from pure octahedral to tetrahedral geometry, due to the loss
of the centre of inversion of symmetry and progressive mixing of empty p-d states [56]. The strong
1
2

pre-edge peak for the sample synthesized from liquid (Fig. 6b) is therefore related to a higher
distortion of the WO unit compared to W/SiO -DI, being the intensity of pre-edge peak of the latter
closer to WO . The pre-edge intensity for W/SiO -liq is however weaker than expected for well-
isolated WO unit [57], supporting the presence of mainly WO sites in all samples.
The position of the white line for the L -edge of the two samples (Fig.7a) was comparable to the one
of the WO standard, in agreement with the presence of W(VI). However, differences in the shape
were clearly visible, especially in the second derivative plot (Fig. 7b). The W L -edge is mainly due
6
2
3
2
4
6
3
3
3
to the dipole-allowed transition from the metal 2p3/2 orbitals to vacant d orbitals and its splitting is
related to the geometry of the W site. In fact, a larger splitting is expected for an octahedral site and
a smaller one for a tetrahedral one, as a result of the different energy splitting of the d orbitals due to
the different ligand field [58]. A splitting of 3.0 -4.0 eV is generally indicative for distorted WO
while a splitting of 1.8-2.0 eV is typical for WO moieties [59]. The second derivative of the white
line, reported in Figure 7b, clearly shows two distinct minima for WO and the W/SiO samples. The
energy separation of 3.1 eV and 3.9 eV for W/SiO -liq and for W/SiO -DI, respectively, prove the
presence of distorted octahedral WO units in both solids. The smaller value for the sample
synthesized via liquid-phase grafting is consistent with a higher distortion of the W sites as well (vide
supra L -edge).
The strong peaks between 1 Å and 2 Å shown in the FT of the EXAFS oscillation for the two samples
are the results of different W-O scattering paths in the WO unit (Fig. 8). While W/SiO -DI presented
only one broad peak, similar to WO , W/SiO -liq showed a splitting of the first shell into two different
peaks at shorter and a longer distance, reflecting the higher distortion of the WO unit [60]. The
quantitative results of the fitting reveal three different W-O bonds for W/SiO -liq (Table SI1): two
6
,
4
3
2
2
2
6
1
6
2
3
2
6
2
short bonds at 1.72 Å and 1.91 Å and a longer one at 2.26 Å. This long W-O bond is consistent with
the high distortion of the octahedral unit and is characteristic of hydrated tungsten oxide polyoxo
clusters species [61]. The results for W/SiO
2
-DI are instead comparable to WO
3
(W/SiO
2
-DI: 1.75 Å
and 2.04 Å; WO : 1.81 Å and 2.11Å) and in agreement with the presence of WO
3
3
domains for this
1
3

sample. A second peak between 2.3 Å and 3.2 Å and a third one centred at 3.5 Å (non-phase shift
corrected distances) were present in the EXAFS of both WO and W/SiO -DI, confirming the similar
nature of the W site for the two samples. The first peak derives from multiple scattering within the
first shell (W-O-O) while the second one is related to W-W scattering path for corner-shared WO
Notably, the shorter distance of the W-W shell for W/SiO -liq evidences the presence of edge-shared
WO typical of octahedrally coordinated small oligomeric tungsten oxide species confirming the
quantitative EXAFS analysis of the W-O shell and the XANES analysis of L and L edges [62]. The
3
2
6
.
2
6
1
3
peak at 2.7 Å (non-phase shift corrected distance) for this sample may be assigned to W-Si scattering
path [63].
Interestingly, previous studies have already revealed the presence of both WO domains and hydrated
3
tungsten oxide polyoxo clusters under ambient conditions on the silica surface. Upon dehydration,
only the latter species decompose to isolated surface monoxo- and dioxo-isolated W species
[
64,65,66].
The combination of UV-visible and X-ray absorption spectroscopy studies revealed that the
coordination geometry and structure of tungsten sites are markedly different in the two catalysts and
are strongly dependent on the adopted synthetic procedure. Tentatively, we can conclude that well-
dispersed tungsten oxide polyoxo cluster sites are mainly distributed on the surface of sample
prepared via liquid-phase grafting, whereas monoclinic domains of tungsten oxide are clearly
detected on the surface of W/SiO sample.
2
Catalytic tests
The tungsten-grafted silica catalysts described above have been evaluated in the selective liquid-phase
epoxidation of two alkenes: limonene and methyl oleate.
1
4

O
O
W-SiO2
oxidant
+
+
a)
O
endo-cis endo-trans
limonene
epoxide
exo
limonene
epoxide
(+)-limonene
limonene
epoxide
O
O
O
W-SiO2
oxidant
b)
( )7
H CO
( )7
(
)7
H CO
( )7
3
3
methyl epoxystearate
methyl oleate
Scheme 1
By studying the epoxidation of limonene, it is possible to evaluate not only the activity and chemoselectivity
of the catalysts, but also their regio- and stereoselectivity (Scheme 1a). The performance of the two catalysts,
namely W/SiO
2
-liq and W/SiO
2
-DI, were compared to the one of a commercial batch of WO (Table 2).
3
All of the W-containing systems were active in the epoxidation of limonene, when aqueous hydrogen
peroxide was used as an oxidant, and moderate to good values were achieved, in terms of alkene conversion
(
up to 80%, after 6 h) and yield to epoxides (up to 54%, after 6 h). Clearly, specific activity values for WO
catalysts were very low, because of the very high (nominal) tungsten content in the bulk of pure tungsten
oxide samples. W/SiO -DI showed the best performance in terms of specific activity per W site, confirming
3
2
that the non-negligible presence of aggregated oxidic domains (as revealed by DR UV-Vis and XAS studies;
see above) is not a detrimental feature for the epoxidation reaction.
Limonene epoxides were produced with good selectivities (in the range 60 to 68%) and the formation of the
oxirane ring on the 1,2- and 8,9-C=C double bonds was predominant with respect to the formation of
carveols, which are conversely obtained via allylic oxidation of the endocyclic unsaturation, via a radical
oxidation pathway. Carveols and other allylic cyclic terpenic alcohols were indeed the major side products.
However, non-negligible amounts of menthenediols, derived from the water-mediated opening of the
oxirane rings, were observed as well. This behaviour suggests that over W-silica systems the concurrent
formation of oxidized products via homolytic oxidation of limonene by H
2
O -derived radicals played a less
2
1
5

relevant role than the heterolytic oxygen transfer from H
2
O to the alkene group. Conversely, when
2
anhydrous tert-butylhydroperoxide, TBHP, an oxidant that is particularly suitable for use under water-free
conditions on Ti-based systems [11,67], was used, limonene indeed underwent oxidation, but no selective
production of epoxides was detected. In this case, a very large number of minor side-products, generated by
radical species derived from TBHP decomposition, was observed. In terms of regioselectivity, the molar ratio
between endocyclic to exocyclic limonene epoxides (Scheme 1a) was ca. 80:20 over all the catalysts. This
result is fully consistent with the values previously observed over other Ti-silica systems and suggests that
the oxygen transfer preferentially takes place on the electron-rich endocyclic unsaturation [47,68,69]. Finally,
a similar cis to trans limonene epoxide molar ratio (ca. 80:20) was recorded for all of the W-containing
catalysts. Such preferential trend towards the cis epoxide suggests that over both grafted and bulk WO
3
systems the oxygen-transfer primarily occurs on the less hindered face of the cyclohexene ring, both methyl
and iso-propenyl groups (on C1 and C4 carbons, respectively) pointing towards the opposite face of the epoxy
ring (Scheme 1a). In one test, hydroquinone was used a radical scavenger in a quasi-equimolar amount with
respect to the total tungsten species in the W-SiO
2
catalyst. A modest slowdown of the epoxidation activity
was observed (specific activity after 1 h decreased to 24 h-1 and 21 h , for W/SiO
-1
-liq and W/SiO -DI,
2
2
respectively), but the reactivity was not suppressed. These experimental observations suggest that the role
of a radical-mediated pathway is small, although non-negligible, over these catalysts, under these conditions.
Since metal leaching is the main drawback W-silica catalysts suffer from, especially in the presence of
aqueous hydrogen peroxide [32,40,41,43], the actual heterogeneous nature of W/SiO
hot-centrifugation test and compared to the results obtained over commercial WO
tested conditions, the production of limonene epoxides did not stop when WO
reaction mixture, after 45 min, whereas W/SiO -DI showed a genuinely heterogeneous nature, with no
further conversion of limonene into the related epoxide after the solid was removed from the liquid phase.
From a visual point of view too, the reaction mixtures containing the WO sample rapidly turned into a
greenish-brownish colour (in few minutes), after the addition of the aqueous H . On the contrary, no
evident changes in colour were observed in the presence of the W/SiO -DI solid. This behaviour confirms that
no active soluble species leached out from W/SiO -DI, whereas most of the catalytic activity of the
2
-DI was examined in a
3
(Figure 9). Under the
3
was centrifuged out from
2
3
2
O
2
2
2
1
6

commercial WO
phase.
3
sample has to attributed to homogeneous tungsten species present and active in the liquid
The W/SiO -DI sample was then studied in a series of repeated catalytic cycles, up to six runs (Figure 10). It
2
is worth highlighting that an intermediate calcination step (in dry air at 500°C) was always necessary in order
to recover most of the pristine catalytic activity. The thermal regeneration step is indeed necessary to get rid
of the organic heavy deposit that might have fouled the active centres, thus hindering a proper recovery of
the catalytic activity [32]. Nevertheless, even after such regeneration step, a gradual loss in catalytic
performance was recorded along the six recycling runs, with a more remarkable diminution between the first
and the second run as well as between the fifth and sixth one. On the contrary, no remarkable changes in
epoxide selectivity were observed.
The amount of soluble W species in the liquid mixture after the hot centrifugation tests was quantified for
both solids by ICP-AES analysis. After the first recycle, that is usually the most prone to metal leaching, values
of 9 and 60 ppm of tungsten metal were found in the liquids that contained W/SiO
2
-liq and W/SiO -DI,
2
respectively. The release of metal species into solution from the sample obtained via dry impregnation
method was thus higher than the one prepared via liquid-phase grafting. Such leaching values are, on
average, lower than those reported in the previous literature, especially when polar protic solvents were
used as reaction media (37,40,70). This behaviour does not show a massive leaching of metal species into
the reaction mixture. From the spectroscopic point of view, the distribution of the isolated and organized
nanosized tungsten oxide domains appeared slightly different after six catalytic cycles. The UV-Visible band
at ca. 300-320 nm, assigned to the presence of small domains of WO
recycled catalyst (Fig. S3; Supplementary Material). Therefore, the gradual loss in the catalytic performance
of W/SiO -DI is likely to be attributed to the alteration of the tungsten species located on the silica surface
rather than to a massive leaching of catalytically active metal.
3
species, was more evident in the
2
W/SiO
2
-liq and W/SiO -DI were, in addition, studied in the epoxidation of methyl oleate, as a widely accepted
2
model substrate for unsaturated oleochemicals [71,72] (Scheme 1b). With the noteworthy exception of
1
7

heterogenised polyoxotungstate species [73,74], to our best knowledge, no tungsten-containing
heterogeneous silica catalysts have been used for the epoxidation of unsaturated oleochemicals so far.
The performance of the two catalysts were modest, in terms of yield (towards the desired methyl
epoxystearate) and specific activity, in comparison with other systems previously described in the literature
[75,76]. After 6 h of reaction, a maximum epoxide yield of ca. 50% was reached, the average values being
around 30% (Table 3). Under these conditions, allylic oxidation products, such as methyl hydroxyoleate, were
scarcely observed, the major side products being methyl dihydroxystearate derived from water-mediated
opening of the oxirane ring.
Thanks to a higher oxidant-to-alkene ratio (4:1 vs. 2:1 mol/mol), higher conversion and yield values were
achieved on both catalysts. Actually, when a 2:1 ratio only was used, hydrogen peroxide was entirely
consumed and no residual unreacted H
The oxidant efficiency (the amount of oxidised products obtained per amount of consumed H
average, quite scarce over these catalysts. Values always lower than 50% were indeed recorded due to the
unproductive decomposition of hydrogen peroxide into water and molecular oxygen. A larger excess of H
2
O
2
was detected by iodometric assay at the end of 6h of reaction.
2
O
2
) was, on
2
O
2
(4:1 molar ratio) was thus necessary to push further the epoxidation reaction. In terms of specific activity,
the best performance was obtained over W/SiO -liq with a four-fold excess of oxidant, whereas, if one
2
considers the yield towards the desired epoxide only, the best result was achieved over W/SiO
the case for the limonene epoxidation too.
2
-DI, as it was
1
Interestingly, by monitoring through H-NMR spectroscopy the cis-to-trans ratio in the formed methyl
epoxystearate at the end of the reaction, a noteworthy amount of the trans isomer was found (from 15% to
4
2% of the overall epoxide). Since the configuration of the parent methyl oleate is cis only, these values are
a clear clue that a radical-driven oxidation of the C=C bond occurred, with a consequent partial
recombination of the geometry around carbon-9 and carbon-10. In fact, if a free-radical epoxidation pathway
prevails over a heterolytic one, the formation of the trans isomer, the thermodynamically more stable form,
is predominant [77,78,79].
CONCLUSIONS
1
8

Tungstenocene(IV) dichloride proved to be a suitable precursor for the preparation of W(VI)-grafted silica
catalysts to be used in the epoxidation of alkenes. The catalyst preparation via a dry impregnation grafting
approach was simple, versatile and did not require the use of hazardous or environmentally unfriendly
solvents. The spectroscopic characterization of W/SiO
were fairly evenly distributed on the silica surface for both systems. However, the W/SiO
displayed the presence of larger monoclinic tungsten(VI) oxide aggregates, whereas the W/SiO
2
-liq and W/SiO
2
-DI samples evidenced that W(VI) sites
-DI catalyst
-liq sample
2
2
showed mainly well-dispersed tungsten oxide polyoxo cluster sites. Indeed, the chemical environment and
the coordination geometry of W(VI) sites were noticeably different in the two systems and were strongly
dependent on the synthetic protocol adopted for their preparation.
When these catalysts were tested in the epoxidation of (+)-limonene and methyl oleate, as model alkene
substrates, both catalysts were active in the presence of aqueous hydrogen peroxide and W/SiO
the best performance, on average, in terms of yield to desired epoxides. The genuine heterogeneous nature
of the grafted W/SiO -DI catalyst was successfully confirmed by a hot-centrifugation heterogeneity test and
2
-DI showed
2
the solid could be reused and recycled up to five times with a gradual loss in activity, before showing a clear
diminution in the performance at the sixth one. On the contrary, under the same reaction conditions, a
reference commercial WO
3
sample showed, at the first run, a remarkable leaching of homogeneously active
tungsten species into the reaction mixture.
The W/SiO -DI system was therefore a good compromise between an adequate interaction of the tungsten
2
sites with the silica surface, thus providing the catalyst with an satisfactory robustness against leaching, and
an even dispersion of the tungsten oxide domains over the support, thus providing the catalyst with a good
catalytic activity per metal site.
ACKNOWLEDGMENTS
Use of the Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory, is supported
by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under Contract No.
DE-AC02-76SF00515.
1
9

REFERENCES
References
[
1] M. K. Samantaray, E. Pump, A. Bendjeriou-Sedjerari, V. D’Elia, J. Pelletier, M. Guidotti, R.
Psaro, J. M. Basset, Chem. Soc. Rev., 47 (2018) 8403-8437.
[
[
[
2] D. Xuereb, R. Raja, Catal. Sci. Technol., 1 (2011) 517.
3] F. Quignard, A. Choplin, R. Teissier, J. Mol. Catal. A: Chem. 120 (1997) L27.
4] E. Gianotti, M. E. Raimondi, L. Marchese, G. Martra, T. Maschmeyer, J. M. Seddon, S.
Coluccia, Catal. Lett. 76 (2001) 21.
[
5] K. K. Kang, W. S. Ahn, J. Mol. Catal. A: Chem. 159 (2000) 403.
[
6] W. Adam, W. Malisch, K. J. Roschmann, C. R. Saha-Moeller, W. A. Schenk, J. Organomet.
Chem., 661 (2002) 3.
[
[
[
[
[
[
7] A. Held, P. Florczak, Catal. Today, 142 (2009) 329.
8] C. Perego, R. Millini, Chem. Soc. Rev. 42 (9) (2013) 3956–3976.
9] O. A. Kholdeeva, N. N. Trukhan, Russian Chem. Rev. 75(5) (2006) 411-432.
10] Modern Heterogeneous Oxidation Catalysis (Ed.: N. Mizuno), Wiley-VCH, Weinheim, 2009.
11] J. S. Reddy, P. Liu, A. Sayari, Appl. Catal. A: Gen., 151 (1997) 409.
12] Y. Deng, C. Lettmann, W. F. Maier, Appl. Catal. A: Gen., 214 (2001) 31.
[
13] M. Fadhli I. Khedher, J. M. Fraile, C. R. Chimie 20 (2017) 827-832.
[
[
[
[
14] N. E. Thornburg, A. B. Thompson, J. M. Notestein, ACS Catal., 5 (2015) 5077−5088.
15] X. Li, L. Zhang, H. Gao, Q. Chen, Russ. J. Phys. Chem. A, 90 (2016) 1545–1551.
16] M. Ziolek, Catal. Today, 78 (2003) 47.
17] F. Somma, P. Canton, G. Strukul, J. Catal. 229 (2005) 490.
[
18] C. Tiozzo, C. Bisio, F. Carniato, A. Gallo, S. L. Scott, R. Psaro, M. Guidotti, Phys. Chem.
Chem. Phys., 15 (2013) 13354.
[
(
19] S. K. Maiti, A. Ramanathan, W. H. Thompson, B. Subramaniam, Ind. Eng. Chem. Res. 56
2017) 1999−2007.
[
20] I. W. C. E. Arends, R. A. Sheldon, Appl. Catal. A: Gen., 212 (2001) 175-187.
[
21] F. Bigi, C. G. Piscopo, G. Predieri, G. Sartori, R. Scotti, R. Zanoni, R. Maggi, J. Mol. Catal. A:
Chem., 386 (2014) 108-113.
[
22] Q. Yang, C. Copꢀret, C. Li, J. M. Basset, New J. Chem., 27 (2003) 319−323.
[
23] C. Aubry, G. Chottard, N. Platzer, J. M. Bregeault, R. Thouvenot, F. Chauveau, C. Huet, H.
Ledon, Inorg. Chem. 30(23) (1991) 4409-4415.
[
24] M. Faiz Mukhtar Gunam Resul, A. M. López Fernández, A. Rehman, A. P. Harvey, React.
Chem. Eng., 3 (2018) 747-756.
2
0

[
[
25] M. Amini, M. M. Haghdoost, M. Bagherzadeh, Coord. Chem. Rev., 268 (2014) 83-100.
26] K. Sato, M. Aoki, M. Ogawa, T. Hashimoto, D. Panyella, R. Noyori, Bull. Chem. Soc. Jpn.,
7
0(4) (1997) 905–915.
[
27] D. H. Koo, M. Kim, S. Chang, Org. Lett., 7(22) (2005) 5015-5018.
[
28] K. Kamata, K. Yonehara, Y. Sumida, K. Hirata, S. Nojima, N. Mizuno, Angew. Chem. Int.
Ed., 50 (2011) 12062 –12066.
[
29] C. Hammond, J. Straus, M. Righettoni, S. E. Pratsinis, I. Hermans, ACS Catal., 3 (2013) 321–
3
27.
[
30] T. R. Amarante, M. M. Antunes, A. A. Valente, F. A. Almeida Paz, M. Pillinger, I. S.
Gonꢁalves, Inorg. Chem., 54 (2015) 9690−9703
[
31] W. L. Dai, J. Ding, Q. Zhu, R. Gao, X. Yang, Catalysis, 28 (2016) 1–27.
[
32] N. Maksimchuk, I. Ivanchikova, O. Zalomaeva, Y. Chesalov, A. Shmakov, V. Zaikovskii, O.
Kholdeeva, Catalysts, 8 (2018) 95.
[
33] C. Y. Cheng, K. J. Lin, M. R. Prasad, S. J. Fu, S. Y. Chang, S. G. Shyu, H. S. Sheu, C. H.
Chen, C. H. Chuang, M. T. Lin, Catal. Commun. 8 (2007) 1060–1064.
[
[
34] E. G. Vieira, N. L. Dias Filho, Mater. Chem. Phys., 201 (2017) 262-270.
35] H. Chen, N. Li, M. Wang, X. Wang, Q. You, M. Sun, X. Ma, Patent CN 109513458, 2019.
[
36] R. Gao, W. L. Dai, X.Yang, H. Li, K. Fan, Appl. Catal., A, 332 (2007) 138.
[
37] M. Jin, Z. Guo, X. Ge, Z. Lv, React. Kinetics Mech. Catal. 125 (2018) 1139–1157.
[
[
38] J. Jarupatrakorn, M. P. Coles, D. Tilley, Chem. Mater., 17 (2005) 1818−1828.
39] R. Gao, X. Yang, W. L. Dai, Y. Le, H. Li, K. Fan, J. Catal. 256 (2008) 259–267.
[
(
40] R. Maheswari, M. P. Pachamuthu, A. Ramanathan, B. Subramaniam, Ind. Eng. Chem. Res. 53
2014) 18833–18839.
[
41] A. Ramanathan, R. Maheswari, B. P. Grady, D. S. Moore, D. H. Barich, B. Subramaniam,
Micropor. Mesopor. Mater., 175 (2013) 43–49.
[
(
42] A. Ramanathan, B. Subramaniam, D. Badloe, U. Hanefeld, R. Maheswari, J. Porous Mater. 19
2012) 961–968.
[
43] J. Tang, L. Wang, G. Liu, Y. Liu, Y. Hou, W. Zhang, W. R. Thiel, J. Mol. Catal. A: Chem.
3
13(1-2) (2009) 31–37.
[
(
44] X. L. Yang, W. L. Dai, H. Chen, J. H. Xu, Y. Cao, H. Li, K. Fan, Appl. Catal. Gen. A, 283
2005) 1-8.
[
45] G. Gelbard, T. Gauducheau, E. Vidal, V. I. Parvulescu, A. Crosman, V. M. Pop, J. Mol. Catal.
A: Chem. 182–183 (2002) 257–266.
[
[
46] T. Maschmeyer, F. Rey, G. Sankar, J.M. Thomas, Nature, 378 (1995) 159.
47] M. Guidotti, N. Ravasio, R. Psaro, G. Ferraris, G. Moretti, J. Catal., 214(2) (2003) 242.
2
1

[
[
48] B. Ravel, M. J. J. Newville, Synchrotron Radiat. 8 (2001) 322
49] M. Guidotti, B. Lazaro, Deactivation of Molecular Sieves in the Synthesis of Organic
Chemicals, in “Deactivation and regeneration of zeolite catalysts”, M. Guisnet, F. Ramoa-Ribeiro
(
[
[
Eds.), Imperial College Press, London, 2011, p. 303-334.
50] L. T. Zhuravlev, Langmuir, 3 (1987) 316.
51] L. J. Burcham, J. Datka, I. E. Wachs, J. Phys. Chem. B 103 (1999) 6015.
[
(
52] J. E. Herrera, J. H. Kwak, J. Z. Hu, Y. Wang, C. H. F. Peden, J. Macht, E. Iglesia, J. Catal. 239
2006) 200–211.
[
53] M. S. Morey, J. D. Bryan, S. Schwarz, G. D. Stucky, Chem. Mater., 12 (2000), 3435–3444.
[
54] E. Gianotti, C. Bisio, L. Marchese, M. Guidotti, N. Ravasio, R. Psaro, S. Coluccia, J. Phys.
Chem. C, 111 (2007) 5083-5089.
[
[
[
55] Z. Zhang, J. Suo, X. Zhang, S. Li, Appl. Catal. A: Gen., 179 (1999) 11-19.
56] S. Yamazoe,Y. Hitomi,T. Shishido, T. Tanaka, J. Phys. Chem. C, 112 (2008) 6869-6879.
57] J. A. Horsley, E. Wachs, J. M. Brown, G. H. Via, F. D. Hardcastlex, J. Phys. Chem., 91 (1987)
4
014-4020.
[
58] U. Jayarathne, P. Chandrasekaran, A. F. Greene, J. T. Mague, S. DeBeer, K. M. Lancaster, S.
Sproules, J. P. Donahue, Inorg. Chem., 53 (2014) 8230−8241.
[
59] S. Yamazoe, Y. Hitomi, T. Shishido, T. Tanaka, J. Phys. Chem. C, 112 (2008) 6869-6879.
[
60] V. L. Aksenov, M. V. Koval’chuk, A. Yu. Kuz’min, Yu. Purans, S. I. Tyutyunnikov,
Crystallography Reports, 51 (2006) 908–935.
61] L. Zhenfang, Z. Bisong, Zeit. Kristall. New Cryst. Struct., 223 (2008) 191-193.
62] X. Carrier, E. Marceau, H. Carabineiro, V. Rodrıguez-Gonzalez, M. Che, Phys. Chem. Chem.
Phys., 2009, 11, 7527–7539.
63] F. Hilbrig, H. E. Gobel, H. Knozinger, H. Schmelz and B. Lengeler, J. Phys. Chem., 1991, 95,
[
[
[
6
973–6978.
[
[
[
[
[
[
64] E. I. Ross-Medgaarden, I. E. Wachs, J. Phys. Chem. C, 111 (2007) 15089-15099.
65] S. Lwin, Y. Li, A. I. Frenkel, I. E. Wachs, ACS Catal., 6 (2016) 3061−3071.
66] E. L. Lee, I. E. Wachs, J. Phys. Chem. C, 111 (2007) 15089-15099.
67] A. Gallo, C. Tiozzo, R. Psaro, F. Carniato, M. Guidotti, J. Catal., 298 (2013) 77.
68] C. Cativiela, J.M. Fraile, J.I. Garcia, J.A. Mayoral, J. Mol. Catal. A 112 (1996) 259.
69] O. A. Kholdeeva, I. D. Ivanchikova, N. V. Maksimchuk, I. Y. Skobelev, Catal. Today, 2018,
DOI: 10.1016/j.cattod.2018.04.002
[
[
70] W. Yan, A. Ramanathan, M. Ghanta, B. Subramaniam, Catal. Sci. Technol. 2014, 4, 4433.
71] A. Köckritz, A. Martin, Eur. J. Lipid Sci. Technol. 110 (9) (2008) 812–824.
2
2

[
72] M. Guidotti, C. Palumbo, in Encyclopedia of Inorganic and Bioinorganic Chemistry, D. A.
Atwood (Ed.), John Wiley & Sons, Ltd., Chichester, UK (2016), 1-11.
[
9
73] E. Poli, R. De Sousa, F. Jerome, Y. Pouilloux, J. M. Clacens, Catal. Sci. Technol., 2 (2012)
10–914.
[
74] Y. Qiao, H. Li, L. Hua, L. Orzechowski, K. Yan, B. Feng, Z. Pan, N. Theyssen, W. Leitner, Z.
Hou, ChemPlusChem, 77 (2012) 1128 – 1138.
[
75] S. Dworakowska, C. Tiozzo, M. Niemczyk-Wrzeszcz, P. Michorczyk, N. Ravasio, R. Psaro, D.
Bogdał, M. Guidotti, J. Cleaner Prod., 166 (2017) 901-909.
[
(
76] M. Guidotti, R. Psaro, N. Ravasio, M. Sgobba, E. Gianotti, S. Grinberg, Catal. Lett., 122
2008) 53-56.
[
(
77] M. Guidotti, E. Gavrilova, A. Galarneau, B. Coq, R. Psaro, N. Ravasio, Green Chem., 13
2011) 1806-1811.
[
78] L. Vanoye, Z. Eddine Hamami, J. Wang, C. de Bellefon, P. Fongarland, A. Favre-Reguillon,
Eur. J. Lipid Sci. Technol., 119 (2017) 1600281.
[
79] S. M. Danov, O. A. Kazantsev, A. L. Esipovich, A. S. Belousov, A. E. Rogozhin, E. A.
Kanakov, Catal. Sci. Technol., 7 (2017) 3659–3675.
2
3

0.16
0.12
0.08
0.04
0.00
2
30
2
70
4
00
b
a
2
00
300
400
Wavelength [nm]
500
600
Figure 1. DR UV-Vis spectra collected at room temperature of cyclopentadienyl-tungsten species
grafted onto SiO
temperature.
2
for W/SiO
2
-liq (a) and W/SiO -DI (b), before calcination, recorded at room
2
0
0
0
0
.040
.032
.024
.016
1
000
a
b
8
6
4
2
00
00
00
00
0
a
b
0.008
0.000
0.0
0.2
0.4
0.6
0.8
1.0
20
40
60
80
100
120
140
Relative Pressure, P/P0
Pore diameter [A]
Figure 2. N
2
adsorption–desorption isotherms at 77 K (left) and pores size distribution obtained by
–liq (a) and W/SiO -DI (b) samples.
NLDFT (right) of W/SiO
2
2
2
4

0.30
0.24
0.18
0.12
0.06
0.00
2
50
b
a
3
20
3
80
2
00
300
400
Wavelength [nm]
500
600
Figure 3. DR UV-Vis spectra of W/SiO
2
-liq (a) and W/SiO -DI (b) after calcination, recorded at
2
room temperature.
8
05
0
.005
7
15
2
70
3
25
c
b
x4
a
1000
800
600
Raman shift [cm ]
400
200
-
1
Figure 4. Raman spectra of W/SiO
2
-liq (a), W/SiO
2
-DI (b) and commercial WO (c).
3
2
5

1
1
2000
0000
8
6
4
2
000
000
000
000
0
c
b
a
1
0
20
30
40
50
60
2THETA [deg]
Figure 5. X-ray patterns of W/SiO
2
-liq (a), W/SiO
2
-DI (c) and commercial WO (c).
3
(a)
(b)
Figure 6. W L -edge XANES region. (a) Main edge and (b) pre-edge region.
1
2
6

(a)
(b)
Figure 7. W L
3
-edge XANES region. (a) Normalized data and (b) second derivative.
2
7

Figure 8. Fourier Transform of W L -edge EXAFS data
3
Figure 9. Heterogeneity tests for the epoxidation of (R)-(+)-limonene. Limonene epoxide yield
profile following hot centrifugation, at 45 min, of commercial WO (red curve) and W/SiO -DI (blue
solid curve). Epoxide yield profile over W/SiO -DI (without hot centrifugation) is reported for
comparison (blue dotted curve). Reaction conditions: 5 mL CH CN solvent; 100 mg catalyst; 1.0
mmol limonene; 2.0 mmol H (aq. 50 wt.%); 90°C oil bath; batch reactor.
3
2
2
3
2
O
2
2
8

Figure 10. Epoxidation of (R)- (+)-limonene over W/SiO
bars) and 6 h (orange bars) in consecutive catalytic runs (1 to 6). 10 mL CH
catalyst; 2.0 mmol (+)-limonene; 4.0 mmol H (aq. 50 wt.%); 90°C oil bath; batch reactor. The
catalyst was washed and activated again at 500°C under dry air prior to each run.
2
-DI. Limonene conversion after 1 h (blue
3
CN solvent; 300 mg
2
O
2
2
9