An efficient epoxidation of terminal aliphatic alkenes over heterogeneous catalysts: When solvent matters

Article abstract of DOI:10.1039/c5cy01639h

The epoxidation of unfunctionalized terminal aliphatic alkenes over heterogeneous catalysts is still a challenging task. Due to the tuning of a peculiar catalyst/oxidant/solvent combination, it was possible to attain good alkene conversions (73%) and excellent selectivity values (>98%) in the desired terminal 1,2-epoxide. Over the titanium-silica catalyst and in the presence of tert-butylhydroperoxide, the use of α,α,α-trifluorotoluene as an uncommon non-toxic solvent was the key factor for a marked enhancement of selectivity. The titanium-silica catalyst was efficiently recycled and reused after a gentle rinsing with fresh solvent.

Full text of DOI:10.1039/c5cy01639h

Catalysis
Science &
Technology
View Article Online
View Journal
PAPER
An efficient epoxidation of terminal aliphatic
alkenes over heterogeneous catalysts: when
Cite this: DOI: 10.1039/c5cy01639h
solvent matters†
C. Palumbo,a C. Tiozzo,a N. Ravasio,a R. Psaro,a F. Carniato,b C. Bisioab
a
*
and M. Guidotti
The epoxidation of unfunctionalized terminal aliphatic alkenes over heterogeneous catalysts is still a chal-
lenging task. Due to the tuning of a peculiar catalyst/oxidant/solvent combination, it was possible to attain
good alkene conversions (73%) and excellent selectivity values (>98%) in the desired terminal 1,2-epoxide.
Over the titanium–silica catalyst and in the presence of tert-butylhydroperoxide, the use of α,α,α-
trifluorotoluene as an uncommon non-toxic solvent was the key factor for a marked enhancement of se-
lectivity. The titanium–silica catalyst was efficiently recycled and reused after a gentle rinsing with fresh
solvent.
Received 26th September 2015,
Accepted 17th December 2015
DOI: 10.1039/c5cy01639h
www.rsc.org/catalysis
Unfortunately the use of TS-1, and of microporous materials
in general, is not compatible with safer, more selective, but
Introduction
The epoxidation of olefins plays
a
major role in the
bulkier peroxides, such as tert-butyl hydroperoxide (TBHP) or
cumyl hydroperoxide. Conversely, for bulkier and larger sub-
strates, much progress has been made, at least at laboratory
scale, in the selective heterogeneous catalytic epoxidation of
internal and cyclic alkenes.11–13 In this case, mesoporous cat-
alysts permit the successful use of bulky organic peroxides
under mild conditions.14,15 However, in the strategy of a bio-
based production of intermediates for the chemical industry,
growing attention has been paid to the selective epoxidation
of terminal linear alkenes since they can be conveniently
obtained from unsaturated fatty acid methyl esters (FAMEs)
of vegetable origin via olefin metathesis.16,17 In spite of the
performance observed on internal olefins, the selective epoxi-
dation of 1-alkenes still represents an ambitious
challenge.18–21 Terminal double bonds are, in fact, substan-
tially less active towards electrophilic oxidants.22–24 Long reac-
tion times and aggressive oxidants, such as peroxoacids, are
thus required. These systems imply the co-presence of free
carboxylic acids in the reaction mixture, leading to
rearrangement and oxirane ring-opening side products, cause
corrosion problems to the production plants and often re-
quire the use of chlorinated solvents and/or hazardous sol-
vents.25 At laboratory scale, many efforts have been made on
the sustainable epoxidation of terminal alkenes and some pa-
pers are reported in the recent literature, each of them focus-
ing on one single aspect of the issue.26–29 Previous experience
within our research team on the epoxidation of various sub-
strates, such as cyclic alkenes, unsaturated terpenes and un-
saturated fatty acid derivatives, over heterogeneous single-site
Ti- and Nb-containing silica solids,30 has prompted us to pay
valorisation of biomasses and fossil oil refinery by-products
since epoxides are essential and versatile intermediate build-
ing blocks for the pharmaceutical, flavour and polymer indus-
tries.1,2 The development of sustainable oxidative procedures
based on solid catalysts is therefore an object of extensive re-
search with the aim of reducing the environmental impact of
the epoxidation reaction.3 At industry level, propene is suc-
cessfully and selectively epoxidised over a mesoporous silyl-
ated titanium silicate catalyst, as in the styrene monomer
propylene oxide (SMPO) process, or over titanium-silicalite 1
(TS-1), as in the recent hydrogen peroxide propylene oxide
(HPPO) process.4,5 Owing to the use of a TiIJIV)-based oxide (in
the former) or zeotype (in the latter) with a marked hydropho-
bic character, in the presence of concentrated hydrogen per-
oxide as a source of oxygen, propylene oxide (PO) is obtained
with high selectivities with minor formation of secondary
products.3,5–8 The use of hydrogen peroxide at high concen-
trations gives rise to major safety and transportation issues,
and for this reason, the production facility of such an oxidant
has to be located just next to the epoxidation plant.9,10
a CNR-Istituto di Scienze e Tecnologie Molecolari, Via C. Golgi 19, 20133 Milano,
Italy. E-mail: m.guidotti@istm.cnr.it
b Dipartimento di Scienze e Innovazione Tecnologica and Nano-SISTEMI Interdisci-
plinary Centre, Università del Piemonte Orientale “A. Avogadro”, Viale Teresa
Michel 11, 15121 Alessandria, Italy
† Electronic supplementary information (ESI) available: N2 adsorption/desorp-
tion isotherms; pore size distribution analysis of the catalysts; X-ray
diffractograms; diffuse reflectance UV-vis spectroscopy; heterogeneity tests. See
DOI: 10.1039/c5cy01639h
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
special attention to the epoxidation of terminal aliphatic al-
kenes, promoting the use of heterogeneous catalytic proce-
dures, non-noxious solvents and selective oxidants. The inves-
tigation was performed aiming at finding the optimal
combination of catalyst, oxidant and solvent for the sustain-
able epoxidation of 1-alkenes. In particular, 1-octene was
identified as a diagnostic substrate due to the similarity with
PO (propylene oxide) of its oxidised product.31 Furthermore,
the use of α,α,α-trifluorotoluene, which was recently proved to
be an effective alternative to noxious reaction media,32 was
envisioned for the first time in the epoxidation of terminal
olefins.
a Quantachrome Autosorb 1MP/TCD instrument. Prior to
analysis, the samples were outgassed at 373 K for 3 h (resid-
ual pressure lower than 10−6 torr). Apparent surface areas
were determined by using the Brunauer–Emmett–Teller equa-
tion in the relative pressure range from 0.01 to 0.1 P/P0. Pore
size distributions were obtained by applying the Barret–
Joyer–Halenda (BJH) approach applied to the desorption
branch for MCM-41 and Ti/MCM-41 samples and the NLDFT
method (N2 silica kernel based on a cylindrical pore model
applied to the desorption branch) for all other samples.
XRD patterns were obtained on an ARL XTRA48 diffrac-
tometer using CuKα radiation (λ = 1.54062 Å).
Diffuse reflectance UV-vis (DR UV-vis) spectra were
recorded at RT using a Perkin-Elmer Lambda 900 spectrome-
ter equipped with an integrating sphere accessory and using
a custom-made quartz cell. All samples were calcined at 773
K under dry air and, prior to the DR UV-vis analysis, dis-
persed in anhydrous BaSO4 (10 wt%) and treated in vacuo for
1 h at room temperature.
Experimental section
Chemicals
All substrates were used as purchased: 1-octene (98%, Al-
drich), trans-2-octene (97%, Aldrich), cis-2-octene (>95%,
TCI), cis-1,4-hexadiene (99%, Aldrich), 1,9-decadiene (98%,
TCI), 1-hexene (97%, Aldrich), TBHP (5.5 M decane solution
98%, Aldrich). For all of the reactions the substrate was
dissolved in purchased HPLC-grade solvent (Aldrich). The sol-
vents have been dried over 3A molecular sieves prior to use.
All reactants were used as purchased (Aldrich). The commer-
cially available Aeroperl granulated fumed silica (Evonik,
Aeroperl 300/30, 20–40 μm particle size) and Davisil silica
(Grace Davison, LC60A, 60–200 μm particle size) were used as
supports.
Catalytic reaction and product analysis
All catalysts were pretreated under air at 773 K and cooled at
room temperature under vacuum prior to use. The epoxida-
tion tests were carried out under inert atmosphere in a
round-bottom glass batch reactor in a preheated oil bath with
magnetic stirring (ca. 800 rpm). The substrate was dissolved
in 5 mL of solvent. tert-Butyl hydroperoxide (TBHP; 5.5 M in
decane) was used as oxidant, with an oxidant-to-substrate
molar ratio of 1.1/1. Decane (contained in the oxidizing TBHP
solution) was used as internal standard. Samples were col-
lected at periodic intervals and the catalytic performance was
computed on GC chromatograms (HP6890; HP-5 column, 30
Catalyst preparation
Three titanium-containing silicate materials have been stud-
ied as heterogeneous catalysts: Ti/SiO2-Aero, Ti/SiO2-Dav, and
Ti/MCM-41. The catalysts were prepared starting from the
corresponding siliceous support, i.e. Aeroperl granulated
fumed silica, Davisil silica or MCM-41, respectively, by the
following procedure. The silicate material was pretreated in
air at 573 K for 2 h and at 573 K for 2 h under dynamic vac-
uum (10−6 bar). Then it was cooled to room temperature un-
der vacuum and grafted via liquid phase grafting using a
titanocene dichloride solution in chloroform and triethyl-
amine as previously reported.33
1
m × 0.25 mm; FID and MS detectors) or H NMR. The prod-
uct molecular structure was confirmed by 1H NMR spectro-
scopy (Bruker DRX Advance 400 MHz), comparing peaks with
the ones reported in the literature. A standard deviation of
2%, 4%, and 4% has to be considered for conversion, se-
lectivity and yield, respectively. Promising catalytic systems
were tested multiple times with different batches of catalyst,
leading to comparable results.
Neither auto-oxidation nor support-catalysed epoxidation
was recorded with metal-free siliceous supports (conversion
about 5% with no detectable epoxide formation). After all cat-
alytic tests, the presence of residual TBHP was checked and
confirmed by iodometric assays.
Further experiments were performed in order to check the
competitive kinetics of 1-octene and trans-2-octene, following
the procedure reported above, with an oxidant/1-octene/trans-
2-octene molar ratio of 1.1/0.5/0.5.
In order to check the leaching of titanium active species,
the solid catalyst was removed from the liquid mixture by
centrifugation, and the resulting solution was tested for fur-
ther reaction.35
MCM-41 was synthesised with a pore wall thickness of
about 2 nm according to the procedure described in the
literature.34
Catalyst characterisation
The titanium content of the prepared samples was deter-
mined by inductively coupled plasma optical emission
spectroscopy (ICP-OES, ICAP 6300 Duo, Thermo Fisher Scien-
tific) after mineralisation of the samples in a microwave di-
gestion apparatus (Milestone MLS 1200; maximum power 500
W) with a mixture of hydrofluoric (aq. 40 wt%) and fuming
nitric acid.
N2 physisorption measurements were carried out at 77 K
in the relative pressure range from 1 × 10−6 to 1 P/P0 by using
In the tests for the catalyst reuses, the solid was separated
by filtration and washed with fresh reaction solvent and a
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
small amount of methanol (Fluka, HPLC grade). The filtered
solid was dried at 373 K, weighed and reused in a new test
without further pre-treatment.
restriction of the mesopores. The grafting methodology did
not lead to significant structural modifications of the meso-
porous solid (ESI,† Fig. S4).
The coordination geometry and chemical environment of
the metal sites dispersed on the different silica supports were
evaluated by DR-UV-vis spectroscopy (ESI,† Fig. S5).
Results and discussion
Catalyst characterisation
The DR-UV-vis spectrum of the Ti/MCM-41 catalyst
showed a main absorption centred at 235 nm with a slightly
pronounced shoulder at high wavelengths (280–300 nm). The
first band is attributed to the isolated tetrahedral titanium
sites located on the silica surface, whereas the second UV
component could be related to a small fraction of TiIJIV) sites
in expanded coordination (i.e. oligomeric species).36 Such
shoulder absorption became progressively evident in the Ti/
SiO2-Aero and Ti/SiO2-Dav samples. For the latest, the band
centred at 280 nm dominates the spectrum. In addition, the
absorption previously observed at 235 nm for Ti/MCM-41
shifted to 250 nm as a consequence of the modification of
the chemical environment of the Ti sites (Fig. S5†). The DR-
UV-vis spectrum of Ti/SiO2-Dav showed a very weak shoulder
at ca. 350 nm that could be assigned to the presence of extra-
framework polycondensed TiO2 species on the silica surface.
However, these species were far less abundant than the iso-
lated Ti sites.
First, a set of titanium-containing catalysts with different
characteristics was selected (Table 1). The metal centres were
added onto the silica supports following a liquid-phase
grafting procedure, starting from titanocene dichloride.11
Three different silica types with various textural features were
used: (i) a commercial granulated Aeroperl fumed silica by
Evonik (SiO2-Aero); (ii) a mesoporous non-ordered commer-
cial silica gel by Grace (SiO2-Dav); and (iii) an ordered meso-
porous purely siliceous MCM-41 (Table 1). The final content
of titanium in the catalysts was determined by inductively
coupled plasma atomic emission spectroscopy (ICP-AES) and
fell in the range 0.6–1.1 wt% Ti.
All the catalysts and the silica precursors were submitted
to N2 physisorption analysis to evaluate their specific surface
area and the pore size distribution. Aeroperl fumed silica and
the derived Ti-containing sample showed an undefined po-
rosity probably generated by the aggregation of the pyrogenic
non-porous particles, as expected from the factory specifica-
tions (ESI,† Fig. S1). Different textural characteristics were ob-
served for Ti-based non-ordered mesoporous silica (Ti/SiO2-
Dav) and the pristine sample. Both the samples showed iso-
therms of IUPAC “IV” type, with “H2” hysteresis loops typical
of solids with disordered structural pore size distributions
(ESI,† Fig. S2). The Ti deposition led to a reduction of the
specific surface area, from 580 m2 g−1 (for SiO2) to 474
m2 g−1. Conversely, the pore size distribution (falling in the
4–11 nm range) was not significantly affected by the presence
of the TiIJIV) centres.
The formation of oligomeric Ti species, responsible for
the presence of the band centred at ca. 280 nm, especially ev-
ident for Ti/SiO2-Aero and Ti/SiO2-Dav catalysts, could be as-
sociated to a lower specific surface area of the two silica sup-
ports with respect to the MCM-41 solid (see Table 1) and
thus to a higher density of TiIJIV) centres on the siliceous
support.37
Epoxidation of 1-octene
The first set of experiments concerned the selective epoxida-
tion of 1-octene to 1,2-epoxyoctane over Ti-grafted Aeroperl
granulated fumed silica as catalyst. Such a catalyst was se-
lected first, since, as reported in widely accepted literature,
titanosilicate systems with a marked hydrophobic character
are ideal catalysts for the epoxidation of simple aliphatic
alkenes,38–40 and Ti/SiO2-Aero, indeed, is the solid with a re-
markably low amount of hydroxyl moieties surrounding the
Ti centres (SiOH/Ti molar ratio = 0.26, obtained by TGA anal-
ysis as described in ref. 41). In previous works, acetonitrile
proved to be an ideal solvent for liquid-phase epoxidation re-
actions on Ti–silica catalytic systems.11 However, in the pres-
ent case, a good conversion (66% after 1 h), but poor selectiv-
ity for the desired epoxide, was obtained (Table 2, entry 1).
The gradual loss in selectivity along the reaction is caused by
the formation of a plethora of minor products (especially af-
ter 24 h) via epoxide opening, rearrangement and oligomeri-
zation reactions (heavy product with average molecular mass
>300 dalton). Among the most relevant secondary products,
it is worth mentioning 2-octanone and octanal, obtained via
acid-catalysed rearrangement of the epoxide, particularly at
longer reaction times; 2-octen-1-ol, that can be obtained via
Finally, the two ordered mesoporous MCM-41 solids, be-
fore and after functionalization with Ti, presented isotherms
of type “IV”. MCM-41 showed a surface area of 890 m2 g−1
and a narrow and regular pore size distribution centred at
2.5 nm (pore volume of 0.71 cm3 g−1) (ESI,† Fig. S3). The
metal grafting caused a decrease in both specific surface area
and pore size, reaching a surface area value of 713 m2 g−1
and pore diameter of 2.2 nm, as a consequence of the partial
Table 1 Textural properties, surface properties and elemental analysis of
the Ti-containing silica catalysts
Pore
SBET
diameter
[nm]
Pore volume
Ti content
[wt%]
Catalyst
[m2 g−1 a
]
[mL g−1
]
SiO2-Aero
Ti/SiO2-Aero
SiO2-Dav
Ti/SiO2-Dav
MCM-41
315
295
580
474
890
713
n.d.b
n.d.
4 ÷ 11
4 ÷ 11
2.5
n.d.
n.d.
0.97
0.81
0.71
1.05
—
0.59
—
1.14
—
0.65
Ti/MCM-41
2.2
a
Specific surface area (from BET analysis). b Not determined.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Table 2 The effect of solvent on the catalytic epoxidation of 1-octene
over Ti/SiO2-Aero with tert-butyl hydroperoxidea
ET
C 1 hc
[%]
S 1 hd
[%]
C 24 hc
[%]
S 24 hd
[%]
Entry Solvent (30)b
1
2
3
4
5
6
CH3CN 45.6
AcOEte 38.1
66
12
8
24
18
30
62
97
63
35
63
72
66
31
73
>98
94
65
>98
>98
98
71
31
Tol
33.9
38.5
41.3
39.4
TFT
DCE
TCE
<3
Fig. 1 Correlation between solvent polarity (expressed as Reichardt
parameter ET (30)) and alkene conversion after 24 h in the catalytic
epoxidation of 1-octene.
a
Reaction conditions: 100 mg catalyst; 1 mmol 1-octene; 5 mL sol-
vent; TBHP 1.1 mmol; 363 K oil bath; 24 h; stirring ca. 800 rpm; under
dry N2. Reichardt solvent parameter (in kcal mol−1). Alkene conver-
b
c
sion. d Selectivity for 1,2-epoxide. e Reaction performed at 358 K.
1,2-epoxyoctane. It means that, in that medium, further
rearrangement of the epoxide to secondary products is
disfavoured.
ring opening and rearrangement via a non-radical pathway;42
tert-butoxyoctanols, due to the reaction of epoxyoctane with
tert-butanol (by-product from TBHP); and minor amounts of
1,2-octanediol if traces of water are present.
After having chosen α,α,α-trifluorotoluene as a promising
solvent, the catalysts were compared in the epoxidation of
1-octene under the same conditions. The best conversion and
TON values were achieved over Ti/MCM-41 (Table 3, entry 3),
even though the selectivity for the desired epoxide lowered
markedly along the 24 h of reaction. Such behaviour is consis-
tent with previous observations in which Ti-grafted siliceous
MCM-41 systems (even with no aluminium centres) exhibited a
neat Lewis acid character that was exploited to obtain epoxide-
derived compounds via cascade reactions.32,45–47 Ti/SiO2-Dav led
to quite good conversions (73%) and excellent selectivity
(>98%) even after longer reaction times (Table 3, entry 2). How-
ever, considering that Ti/SiO2-Dav contains a higher loading of
active Ti species than the other catalysts (1.14 vs. 0.59 or 0.65
wt%), a specific test has been performed with a lower mass of
solid (Table 3, entry 5). A fully comparable TON value was
obtained (entries 2 and 5), showing that the specific activity
does not vary with the total mass of the catalyst (within this
mass/substrate range). After 24 h, however, a slightly lower con-
version of 1-octene was achieved.
The effect of the solvent was therefore studied by testing
polar and non-polar aprotic solvents: ethyl acetate, toluene
(Tol), α,α,α-trifluorotoluene (TFT), 1,2-dichloroethane (DCE)
and 1,1,2,2-tetrachloroethane (TCE). In DCE and TCE good
conversions were reached, but the selectivity was not satisfac-
tory (entries 5 and 6), as chlorinated solvents usually enhance
the acid character of TiIJIV)–silica catalysts and hence increase
remarkably the amount of undesired side products. In tolu-
ene, no side-products were detected by spectroscopic tech-
niques (selectivity >98%, entry 3), but such a high selectivity
value was related to very low conversion values, since the cat-
alyst was scarcely active. In ethyl acetate, which is a widely
used solvent for epoxidation reaction on titanosilicate sys-
tems, a good conversion was achieved after 24 h, but only
moderate selectivity was observed (entry 2).
To our delight, the use of α,α,α-trifluorotoluene led to the
1,2-epoxide in rather good conversion (63%) and very high se-
lectivity (94%) (entry 4). Interestingly, TFT, although
a
Ti/SiO2-Aero showed a good performance with an interme-
diate TON value between that of Ti/MCM-41 and Ti/SiO2-Dav,
albeit slightly lower in terms of both conversion and
fluorine-containing solvent, is attracting a growing interest as
a potential less hazardous alternative to dichloromethane, as
it possesses solvent properties that are comparable to those of
dichloromethane (DCM), but with lower volatility (b.p. 375 K
for TFT vs. 313 K for DCM) and a remarkably reduced toxic-
ity.32,43 The solvent effect on the reaction was rationalised by
means of a correlation between the solvent polarity (expressed
as Reichardt ET (30) parameter)44 and the conversion of
1-octene. A linear relationship of the 24 h conversions vs. ET
(30) was found with a very good correlation parameter (Fig. 1).
Such a result shows that the higher the polarity of the solvent,
the larger is the amount of 1-octene converted over the cata-
lyst. It also suggests that highly polar solvents are good reac-
tion media which favour the desorption of the newly formed
epoxide molecule from the catalyst surface and reduce the
surface deactivation of the solids. At the same time, TFT
was also the most selective solvent towards the formation of
Table
3 Performance of the titanium–silica catalysts in 1-octene
epoxidationa
C 1 hc
[%]
S 1 hd
[%]
C 24 hc
[%]
S 24 hd
[%]
Entry
Catalyst
TONe
1
2
3
4
5
Ti/SiO2-Aero
Ti/SiO2-Dav
Ti/MCM-41
Ti/SiO2-Davb
Ti/SiO2-Davf
24
33
39
35
20
98
>98
96
98
>98
19
14
29
15
15
63
73
94
76
60
94
>98
58
98
>98
a
Reaction conditions: 100 mg catalyst; 1 mmol 1-octene; 5 mL TFT;
1.1 mmol TBHP; 363 K oil bath 24 h; stirring ca. 800 rpm; dry N2.
b
c
d
Reaction performed at 383 K. Alkene conversion. Selectivity for
e
1,2-epoxide. Turnover number, computed after 1 h of reaction
[mmolconverted octene/mmoltotal Ti]. 54 mg catalyst.
f
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
selectivity than that of Ti/SiO2-Dav, especially at longer reac-
tion times. The attempt to increase yields by carrying out the
reaction in α,α,α-trifluorotoluene at higher temperature (383
K) did not lead to a significant improvement (Table 3, entry
4). However, the value of 75% of yield was attained more rap-
idly, in just 6 h. Interestingly, the increase in reaction tem-
perature did not enhance the acid-catalysed transformation
of the epoxide either and the conversion values were always
very good.
catalytic cycles, was 12%. Nevertheless, if one considers the
cumulative TON values at the end each test (Fig. 2), taking
into account the gradual diminution of actual catalyst mass
along the four runs, this number was almost constant and
even increased since the third run. Such behaviour shows
that the catalytically active centres do not suffer from deacti-
vation, likely due to the particularly mild recovery conditions.
The truly heterogeneous nature of the catalyst under these
conditions has been confirmed by hot separation tests, in
which the reaction mixture was centrifuged and the solid re-
moved.53 No sensible further reactivity was observed, thus
confirming that no leaching of active species took place dur-
ing the epoxidation reaction (ESI,† Fig. S6). Analogously, the
need for titanium as a catalytically-active centre was con-
firmed and no relevant 1-octene epoxidation was observed in
the absence of the catalyst or in the presence of the sole sili-
ceous support, i.e. Ti-free SiO2-Aero.
With regard to the oxidant, tert-butylhydroperoxide (TBHP)
was the reactant of choice. Organic hydroperoxides indeed
proved previously to be selective and stable oxidants in the
presence of Ti-containing mesoporous silicates and ensure
very high oxidant efficiencies (i.e., the amount of oxidised
products obtained per amount of consumed oxidant).3 Con-
versely, in the presence of aqueous hydrogen peroxide most
mesoporous titanium silica materials suffer from poor stabil-
ity, active metal leaching, low oxidant efficiencies and grad-
ual deactivation due to surface restructuring.48–50 In the pres-
ent work, iodometric assays on the final reaction mixtures
revealed that TBHP had never been the limiting agent and in
all cases the recorded TBHP efficiency was very high (>90%).
Under comparable conditions, in the presence of TBHP,
some reference titanosilicate catalysts, such as large-pore Ti-BEA
zeotype or improved nanocrystalline TS-1, have shown lower
yields of 1,2-epoxyoctane (i.e. 5% and 40%, respectively).51,54
As far as the recovery and reuse of the catalyst is
concerned, the solids were filtered and reused up to four
times (Fig. 2). Interestingly, no re-activation by pre-treatment
at high temperature was needed, whereas in previous reports
in which titanosilicate systems were used, a calcination step
(under air at T >673 K) was necessary to properly remove or-
ganic deposits from the catalyst surface and recover the pris-
tine catalytic activity.48,52 A thorough rinsing with fresh TFT
was sufficient to recover the pristine catalytic activity. Ti/
SiO2-Dav showed a good resistance to repeated recycles and
the overall decrease in epoxide yield after 24 h, along four
Epoxidation of various linear alkenes
The optimized reaction conditions determined above were
employed in the epoxidation of a set of terminal and internal
alkenes.
First, the internal alkene trans-2-octene was investigated
as a comparison with the terminal 1-octene. The use of
Ti/SiO2-Aero as catalyst, in α,α,α-trifluorotoluene and in the
presence of TBHP, led to a good conversion and moderate
selectivity for trans-2,3-epoxyoctane after 1 hour of reaction
(Table 4, entry 1). Ti/MCM-41 showed the highest activity
(61 TON) after 1 h with a complete conversion after 6 h
(Table 4, entry 3).
Remarkably, under the same conditions, an interesting re-
sult was achieved over Ti/SiO2-Dav. In fact, even though its
initial activity was the lowest (in terms of TON after 1 h), a
complete selectivity for trans-2,3-epoxide was maintained
along the following 24 h, leading to a quantitative conversion
of the substrate alkene. Lower selectivity values were ob-
served, instead, over Ti/SiO2-Aero and Ti/MCM-41. In this
case too, Ti/SiO2-Dav proved to be the most selective catalyst,
leading to a negligible amount of undesired secondary
products.
If one compares the results obtained under the best condi-
tions for 1-octene and trans-2-octene (Tables 3 vs. 4), the
Table 4 Catalytic epoxidation with TBHP of trans-2-octene epoxidation
over Ti-containing catalysts in TFTa
Entry
Catalyst
Cb [%]
Sc [%]
TONd
Ce [%]
Sf [%]
1
2
3
Ti/SiO2-Aero
Ti/SiO2-Dav
Ti/MCM-41
36
40
84
74
>98
>98
29
17
61
96
68
98
>98
>98g
75g
a
Fig.
2 Recovery and reuse of Ti/SiO2-Dav (four catalytic runs).
Reaction conditions: 100 mg catalyst; 1 mmol trans-2-octene; 5 mL
TFT; 1.1 mmol TBHP; 363 K oil bath 24 h; magnetic stirring (ca. 800
Reaction conditions for the first run: 150 mg catalyst; 1.5 mmol
1-octene; 7.5 mL TFT; 1.6 mmol TBHP; 363 K oil bath for 24 h; stirring
ca. 800 rpm; dry N2. The catalyst was filtered after each run and
reused after a simple rinsing with fresh solvent. The reagents were
scaled down as well to keep the same proportions.
b
c
rpm); dry N2. Conversion after 1 h. Selectivity for epoxide after 1
h. Turnover number, computed after 1 h reaction [mmolconverted
octene/mmoltotal Ti]. Conversion after 24 h. Selectivity for epoxide af-
ter 24 h. g Results after 6 h of reaction.
d
e
f
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Encouraged by these results, a broader substrate scope
was tested (Table 5) in order to explore the potential of the
epoxidation method. The epoxidation of the shorter-chained
1-hexene led to a better conversion, with respect to 1-octene,
at equal reaction times (5 h; Table 5, entries 1 and 2) and
1,2-epoxyhexane was obtained with excellent selectivities as
well. Conversely, with longer alkenes, such as 1-decene, a
slower reactivity was recorded (Table 5, entry 3). Such a reac-
tivity order, 1-hexene > 1-octene > 1-decene, is fully consis-
tent with previous literature reports on microporous
titanosilicate molecular sieves.54
The reactivity of cis-2-octene was then evaluated in com-
parison with trans-2-octene. The cis isomer appeared to be
less reactive than the trans isomer, with 66% conversion af-
ter 6 h vs. 88%, respectively (Table 5, entries 6 and 7). Typi-
cally, for bulkier alkenes, cis isomers are more readily
epoxidised than their trans analogues. This is commonly at-
tributed to steric effects, the approach of cis double bonds
being less hindered than trans ones.55–57 In the case of lin-
ear alkenes, the formation of the trans-epoxide can be
favoured by thermodynamic factors (on homogeneous sys-
tems too),58,59 by different diffusion capabilities throughout
the catalyst pores for cis/trans isomers or by a favourable co-
operative interaction among substrate, catalyst and reaction
medium.60,61 In this particular case, the preferential reactiv-
ity of trans-2-octene can be due to a joint action of these var-
ious factors.
In terms of stereoselectivity, no isomerization at the CC
double bond has been observed for the internal olefins. In
fact, trans-2,3-epoxyalkanes and cis-2,3-epoxyalkanes were ex-
clusively detected starting from trans- or cis-alkenes, respec-
tively. This behaviour is a clear clue that the epoxidation re-
action with TBHP over these catalysts proceeds via a non-
radical (heterolytic) mechanism with retention of the configu-
ration.56 Conversely, when the epoxidation follows a radical-
based pathway, a free rotation of the intermediate radical
around the C–C bond can occur and a mixture of cis- and
trans-epoxides is found.
Fig.
3 Reaction conditions: (a) 100 mg Ti/SiO2-Dav; 0.5 mmol
1-octene; 0.5 mmol trans-2-octene; 5 mL TFT; 1.1 mmol TBHP; 363 K
oil bath 24 h; magnetic stirring (ca. 800 rpm); dry N2. (b) 100 mg Ti/
SiO2-Dav; 1 mmol 1-octene; 5 mL TFT; 1.1 mmol TBHP; 363 K oil bath
24 h; magnetic stirring (ca. 800 rpm); dry N2. (c) 100 mg Ti/SiO2-Dav; 1
mmol trans-2-octene; 5 mL TFT; 1.1 mmol TBHP; 363 K oil bath 24 h;
magnetic stirring (ca. 800 rpm); dry N2.
trend is fully consistent with the expected behaviour of an
electrophilic epoxidation of a terminal olefin with respect to
an internal one: a faster reactivity for the disubstituted CC
bond than for a monosubstituted one.
A direct comparison of the reactivity of the two substrates
has been performed with an epoxidation test on an equimo-
lar mixture of 1-octene and trans-2-octene (Fig. 3).
The conversion vs. time profiles of the intermolecular
competition test (0.5 mmol 1-octene + 0.5 mmol trans-2-
octene) were comparable to the corresponding curves of the
single-substrate reaction and no particular diminution of the
reactivity of 1-octene was observed. The lack of a marked
competition between the terminal and internal alkenes sug-
gests that the catalytically active TiIJIV) sites are equally acces-
sible to both substrates. Moreover, such results confirm that
the catalyst–oxidant–solvent system is suitable and applicable
to both linear terminal and internal alkenes.
Table 5 Catalytic epoxidation of linear alkenes over Ti/SiO2-Dav with TBHP in α,α,α-trifluorotoluenea
Entry
Olefin
TONb
22
17
7
Time [h]
Products
Cc [%]
68
Sd [%]
1
2
3
4
5
5
5
7
98
57
99
37
99
6
29
70 (a) 30 (b)
5
7
5
68
98 (a) 2 (b)
6
7
17
15
6
6
88
66
99
99
a
Reaction conditions: 100 mg catalyst Ti/SiO2-Dav; 1 mmol olefin; 5 mL α,α,α-trifluorotoluene; TBHP 1.1 mmol; 363 K oil bath; stirring ca. 800 rpm;
under dry N2. b Turnover number computed after 1 h of reaction [mmolconverted octene/mmoltotal Ti]. c Alkene conversion. d Selectivity for epoxide.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
The epoxidation of 1,9-decadiene, bearing two equally
substituted, terminal double bonds, led after 7 h to a mixture
of the mono- and di-epoxide at a molar ratio of 70/30, even
though at low conversion values (Table 5, entry 4). The ob-
served selectivity at such low conversion values suggests that
the epoxidation of the second CC bond on the intermedi-
ate monoepoxyalkene occurs before the original diene is fully
consumed. A similar behaviour was observed in the epoxida-
tion of diunsaturated fatty acid methyl esters (methyl linole-
ate) over a fully comparable epoxidation system (Ti–silica cat-
sample of Aeroperl silica was kindly provided by Evonik In-
dustries AG.
Notes and references
1 K. Bauer, D. Garbe and H. Surburg, in Common Fragrance
and Flavour Materials, Wiley-VCH, Weinheim, 2001.
2 S. T. Oyama, in Mechanisms in homogeneous and
heterogeneous epoxidation catalysis, ed. S. T. Oyama, Elsevier,
Amsterdam, 2008, pp. 5–12.
alyst with TBHP) and was ascribed to
hydrophilic/hydrophobic character between the reactant and
the catalyst surface.36
a
balance of
3 Liquid Phase Oxidation via Heterogeneous Catalysis: Organic
Synthesis and Industrial Applications, ed. M. G. Clerici and O.
A. Kholdeeva, Wiley, New Jersey, 2013.
Finally, even though the present catalyst–oxidant–solvent
combination had been optimized for the epoxidation of ali-
phatic terminal alkenes, cis-1,4-hexadiene was successfully
tested under the same conditions (Table 5, entry 5), reaching
68% conversion after 5 h, with a chemoselectivity of 98% to-
wards the 4,5-epoxy-1-hexene and a complete regioselectivity
towards the formation of the internal epoxide rather than the
terminal one. Such an intramolecular competition test is a
further confirmation that, as expected in the case of electro-
philic oxygen transfers, the internal double bond is more
readily epoxidised than the terminal one.
4 J. F. K. Buijink, J. J. M. Vlaanderen, M. Crocker and F. G. M.
Niele, Catal. Today, 2004, 199, 93–95.
5 T. A. Nijhuis, M. Makkee, J. A. Moulijn and B. M.
Weckhuysen, Ind. Eng. Chem. Res., 2006, 45, 3447–3459.
6 V. Russo, R. Tesser, E. Santacesaria and M. Di Serio, Ind.
Eng. Chem. Res., 2012, 52, 1168–1178.
7 R. A. Sheldon, Green Chem., 2007, 9, 1273–1283.
8 M. Ghanta, D. R. Fahey, D. H. Busch and B. Subramaniam,
ACS Sustainable Chem. Eng., 2013, 1, 268–277.
9 Sustainable Industrial Processes, ed. F. Cavani, G. Centi, S.
Perathoner and F. Trifirò, Wiley-VCH, Weinheim, 2009.
10 F. Cavani and J. H. Teles, ChemSusChem, 2009, 2, 508–534.
11 M. Guidotti, N. Ravasio, R. Psaro, G. Ferraris and G. Moretti,
J. Catal., 2003, 214, 242–250.
12 J. Přech, P. Eliášová, D. Aldhayan and M. Kubů, Catal.
Today, 2015, 243, 134–140.
13 N. Scotti, N. Ravasio, F. Zaccheria, R. Psaro and C.
Evangelisti, Chem. Commun., 2013, 49, 1957–1959.
14 A. Gallo, C. Tiozzo, R. Psaro, F. Carniato and M. Guidotti,
J. Catal., 2013, 298, 77–83.
15 F. Bigi, C. G. Piscopo, G. Predieri, G. Sartori, R. Scotti, R.
Zanoni and R. Maggi, J. Mol. Catal. A: Chem., 2014, 386,
108–113.
16 U. Biermann, W. Friedt, S. Lang, W. Lühs, G. Machmüller,
J. O. Metzger, M. Rüsch Gen. Klaas, H. J. Schäfer and M.
Schneider, Angew. Chem., Int. Ed., 2000, 39, 2206–2224.
17 M. A. R. Meier, J. O. Metzger and U. S. Schubert, Chem. Soc.
Rev., 2007, 36, 1788–1802.
Conclusions
A set of titanium-silica catalysts proved to be highly efficient
in the selective epoxidation of terminal linear alkenes. To our
best knowledge, such a result is a remarkable advancement
in the epoxidation of poorly activated terminal aliphatic al-
kenes over heterogeneous catalysts. Due to the careful selec-
tion of a peculiar combination of solid catalyst, oxidant and
solvent, i.e., Ti-grafted mesoporous non-ordered silica, tert-
butylhydroperoxide and α,α,α-trifluorotoluene, it has been
possible to obtain particularly promising yields (up to 73%)
and excellent selectivities (>98%) in the epoxidation of
1-octene. Such an unprecedented high selectivity for terminal
linear epoxides has to be attributed to the choice of α,α,α-
trifluorotoluene as an emerging interesting alternative to
chlorinated solvents and is due to the negligible acid-
catalysed formation of by-products from the newly obtained
epoxide. The catalytic system proved to be heterogeneous in
nature and fully recoverable and reusable without the need
for intermediate regeneration treatments under harsh condi-
tions. Such a catalyst–oxidant–solvent combination can be
successfully applied and extended to other linear terminal al-
kenes (as well as to internal olefins), displaying a fully com-
parable marked enhancement in selectivity for the terminal
epoxide desired product.
18 A. H. Hoveyda, Chem. Rev., 1993, 93, 1307–1370.
19 M. E. Judmaier, C. H. Sala, F. Belaj, M. Volpe and N. C.
Mösch-Zanetti, New J. Chem., 2013, 37, 2139–2149.
20 D. Serrano, R. Sanz, P. Pizarro and I. Moreno, Chem.
Commun., 2009, 1407–1409.
21 G. Grigoropoulou, J. H. Clark and J. A. Elings, Green Chem.,
2003, 5, 1–7.
22 B. Qi, X.-H. Lu, D. Zhou, Q.-H. Xia, Z.-R. Tang, S.-Y. Fang, T.
Pang and Y.-L. Dong, J. Mol. Catal. A: Chem., 2010, 322, 73–79.
23 B. S. Lane, M. Vogt, V. J. DeRose and K. Burgess, J. Am.
Chem. Soc., 2002, 124, 11946–11954.
24 J. M. Fraile, J. I. Garcia, J. A. Mayoral and E. Vispe, Appl.
Catal., A, 2004, 276, 113–122.
Acknowledgements
Financial support from Region Lombardy, through project
“SusChemLombardia”, is gratefully acknowledged. The
25 A. Mouret, L. Leclercq, A. Mühlbauer and V. Nardello-Rataj,
Green Chem., 2014, 16, 269–278.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
26 D. Banerjee, R. V. Jagadeesh, K. Junge, M.-M. Pohl, J.
Radnik, A. Brückner and M. Beller, Angew. Chem., Int. Ed.,
2014, 53, 4359–4363.
Solvents for Organic Synthesis and Fluorous Synthesis,
Modern Solvents in Organic Synthesis; Topics in Current
Chemistry, Springer, 1999, vol. 206, pp. 79–106.
27 K. Kamata, K. Yonehara, Y. Sumida, K. Hirata, S. Nojima
and N. Mizuno, Angew. Chem., Int. Ed., 2011, 50,
12062–12066.
28 A. Corma, P. Esteve, A. Martinez and S. Valencia, J. Catal.,
1995, 152, 18–24.
29 Y. Sawada, K. Matsumoto and T. Katsuki, Angew. Chem.,
2007, 119, 4643–4645; K.-P. Ho, W.-L. Wong, K.-M. Lam,
C.-P. Lai, T. H. Chan and K.-Y. Wong, Chem. – Eur. J.,
2008, 14, 7988–7996.
30 C. Tiozzo, C. Palumbo, R. Psaro, C. Bisio, F. Carniato, A.
Gervasini, P. Carniti and M. Guidotti, Inorg. Chim. Acta,
2015, 431, 190–196.
44 C. Reichardt, Chem. Rev., 1994, 94, 2319–2358.
45 A. Bhaumik and T. J. Tatsumi, J. Catal., 1999, 182, 349–356.
46 M. Guidotti, G. Moretti, R. Psaro and N. Ravasio, Chem.
Commun., 2000, 1789–1790.
47 M. N. Timofeeva, O. Kholdeeva, S. H. Jhung and J. S. Chang,
Appl. Catal., A, 2008, 345, 195–200.
48 E. Gianotti, C. Bisio, L. Marchese, M. Guidotti, N. Ravasio, R.
Psaro and S. Coluccia, J. Phys. Chem. C, 2007, 111, 5083–5089.
49 N. N. Trukhan, V. N. Romannikov, A. N. Shmakova, M. P.
Vanina, E. A. Paukshtis, V. I. Bukhtiyarov, V. V. Kriventsov,
I. Yu Danilov and O. A. Kholdeeva, Microporous Mesoporous
Mater., 2003, 59, 73–84.
31 Y. Pérez, S. Bázquez, M. Fajardo, P. de Frutos and I. del
Hierro, J. Organomet. Chem., 2013, 741–742, 102–108.
32 J. M. Fraile, N. García, J. A. Mayoral, F. G. Santomauro and
M. Guidotti, ACS Catal., 2015, 5, 3552–3561.
50 A. Hagen, K. Schueler and F. Roessner, Microporous
Mesoporous Mater., 2002, 51, 23–33.
51 D. Serrano, R. Sanz, P. Pizarro and I. Moreno, Chem.
Commun., 2009, 1407–1409.
33 M. Guidotti, A. Conti, A. Fusi, N. Ravasio and R. Psaro,
J. Mol. Catal. A: Chem., 2002, 182–183, 151–156.
34 K. Suzuki, K. Ikari and H. Imai, J. Am. Chem. Soc., 2004, 126,
462–463.
52 O. A. Kholdeeva and N. N. Trukhan, Russ. Chem. Rev.,
2006, 75, 411–432.
53 R. A. Sheldon, M. Wallau, I. W. C. E. Arends and U.
Schuchardt, Acc. Chem. Res., 1998, 31, 485–493.
35 H. E. B. Lempers and R. A. Sheldon, J. Catal., 1998, 175,
62–69.
54 A. Corma, P. Esteve, A. Martinez and S. Valencia, J. Catal.,
1995, 152, 18–24.
36 M. E. Raimondi, E. Gianotti, L. Marchese, G. Martra, T.
Maschmeyer, J. M. Seddon and S. Coluccia, J. Phys. Chem. B,
2000, 104, 7102–7109.
37 F. Carniato, C. Bisio, L. Sordelli, E. Gavrilova and M.
Guidotti, Inorg. Chim. Acta, 2012, 380, 244–251.
38 S. Klein and W. F. Maier, Angew. Chem., Int. Ed. Engl.,
1996, 35, 2230–2233.
55 Y. Itoi, M. Inoue and S. Enomoto, Bull. Chem. Soc. Jpn.,
1986, 59, 3941–3943.
56 R. D. Oldroyd, J. M. Thomas, T. Maschmeyer, P. A. MacFaul,
D. W. Snelgrove, K. U. Ingold and D. D. M. Wayner, Angew.
Chem., Int. Ed. Engl., 1996, 35, 2787–2790.
57 C. Cativiela, J. M. Fraile, J. I. Garcia and J. A. Mayoral, J. Mol.
Catal. A: Chem., 1996, 112, 259–267.
39 T. Tatsumi, K. A. Koyano and N. Igarashi, Chem. Commun.,
1998, 325–326.
40 K. Lin, L. Wang, F. Meng, Z. Sun, Q. Yang, Y. Cui, D. Jiang
and F.-S. Xiao, J. Catal., 2005, 235, 423–427.
41 M. Guidotti, N. Ravasio, R. Psaro, E. Gianotti, L. Marchese
and S. Coluccia, J. Mol. Catal. A: Chem., 2006, 250, 218–225.
42 D. P. Serrano, R. van Grieken, J. A. Melero and A. Garcia,
Appl. Catal., A, 2007, 319, 171–180.
43 J. J. Maul, P. J. Ostrowski, G. A. Ublacker, B. Linclau and
D. P. Curran, Benzotrifluoride and Derivatives: Useful
58 R. Zwaans, J. H. Van Lenthe and D. H. W. Den Boer, in
Theoretical Aspects of Homogeneous Catalysis, Catalysis by
Metal Complexes, ed. P. W. N. W. van Leeuwen, K.
Morokuma and J. H. van Lenthe, Dordrecht, Kluwer, 1995,
vol. 18, p. 210.
59 A. Berkessel, T. Gunther, Q. Wang and J. M. Neudorfl,
Angew. Chem., Int. Ed., 2013, 52, 8467–8471.
60 P. Wu and T. Tatsumi, J. Phys. Chem. B, 2002, 106, 748–753.
61 M. Guidotti, N. Ravasio, R. Psaro, E. Gianotti, L. Marchese
and S. Coluccia, Green Chem., 2003, 5, 421–424.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015