Selective synthesis of ethylene oxide through liquid-phase epoxidation of ethylene with titanosilicate/H2O2 catalytic systems

Article abstract of DOI:10.1016/j.apcata.2016.02.001

Liquid-phase epoxidation of ethylene to ethylene oxide (EO) with H₂O₂ over various titanosilicate catalysts like Ti-MWW, TS-1, Ti-MOR and Ti-Beta has been investigated. The effects of solvent, catalyst amount, reaction pressure, temperature and time on the catalytic performance of Ti-MWW have been studied in detail. Ti-MWW preferred acetonitrile as a solvent and showed the highest reactivity and EO selectivity among the titanosilicate catalysts investigated. Under optimized reaction conditions, Ti-MWW gave a EO selectivity high as 97.9% as well as a reasonable utilization efficiency of H₂O₂ of 77.7%. Ti-MWW was gradually deactivated after repeated use in ethylene epoxidation. High-temperature calcination easily recovered the catalytic activity of deactivated Ti-MWW after removing ethylene glycol (EG) and other heavy byproducts with high boiling points that were deposited inside micropores. The issues of molecular dimension and reactivity have also been considered by comparing the epoxidation of linear alkenes with different lengths (C₂ to C₆) between two representative titanosilicates Ti-MWW and TS-1. Ethylene, with the smallest dynamic diameter but containing electron-deficient C=C double bonds, was more difficult to be epoxidized than other alkenes with higher intrinsic activities.

Full text of DOI:10.1016/j.apcata.2016.02.001

Applied Catalysis A: General 515 (2016) 51–59
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Selective synthesis of ethylene oxide through liquid-phase
epoxidation of ethylene with titanosilicate/H2O2 catalytic systems
Xinqing Lua, Wen-Juan Zhoua,b, Haihong Wua, Armin Liebensb, Peng Wua,∗
a Shanghai Key Laboratory of Green Chemistry and Chemical Processes, School of Chemistry and Molecular Engineering, East China Normal University,
Shanghai 200062, China
b Eco-Efficient Products and Processes Laboratory (E2P2L), UMI 3464CNRS—Solvay, 3066 Jin Du Road, Xin Zhuang Ind. Zone, Shanghai 201108, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Liquid-phase epoxidation of ethylene to ethylene oxide (EO) with H2O2 over various titanosilicate cat-
alysts like Ti-MWW, TS-1, Ti-MOR and Ti-Beta has been investigated. The effects of solvent, catalyst
amount, reaction pressure, temperature and time on the catalytic performance of Ti-MWW have been
studied in detail. Ti-MWW preferred acetonitrile as a solvent and showed the highest reactivity and
EO selectivity among the titanosilicate catalysts investigated. Under optimized reaction conditions, Ti-
MWW gave a EO selectivity high as 97.9% as well as a reasonable utilization efficiency of H2O2 of 77.7%.
Ti-MWW was gradually deactivated after repeated use in ethylene epoxidation. High-temperature calci-
nation easily recovered the catalytic activity of deactivated Ti-MWW after removing ethylene glycol (EG)
and other heavy byproducts with high boiling points that were deposited inside micropores. The issues
of molecular dimension and reactivity have also been considered by comparing the epoxidation of linear
alkenes with different lengths (C2 to C6) between two representative titanosilicates Ti-MWW and TS-1.
Ethylene, with the smallest dynamic diameter but containing electron-deficient C C double bonds, was
more difficult to be epoxidized than other alkenes with higher intrinsic activities.
Received 26 December 2015
Received in revised form 29 January 2016
Accepted 1 February 2016
Available online 4 February 2016
Keywords:
Titanosilicate
Ti-MWW
TS-1
Ethylene oxide
Liquid-phase epoxidation
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
of a massive amount of chloride salts and the discharge of chlorine-
containing waste water. However, it definitely encounters a fatal
Ethylene oxide (EO), the simplest cyclic ether, is the third-largest
ethylene derivative after polyethylene (PE) and polyvinyl chloride
(PVC), with wide applications in the production of ethylene gly-
col (EG), nonionic surfactants, alcohol ether and other downstream
oxygenated chemicals. The EO demand has reached 20 million met-
ric tons in 2013 and is continually growing at an annual increasing
rate of 6–7% [1]. The discovery of EO goes back to 1859 when
Charles-Adolphe Wurtz conducted the reaction of 2-chloroethanol
and potassium hydroxide, based on which, the chlorohydrins pro-
cess was initially used to produce EO. However, suffering serious
problems of equipment corrosion and environmental pollution, it
was commercially replaced by greener vapor-phase oxidation pro-
cess in the 1960s. The vapor-phase oxidation employs supported
silver and oxygen as the catalyst and the oxidant, respectively. In
comparison to chlorohydrin technique, it avoids the by-production
disadvantage of complete burning of a part of ethylene and EO to
carbon dioxide at high reaction temperatures (200–260 ◦C). The EO
selectivity is still less than 90%, even though the Ag-based catalysts
have been greatly improved in last decades, e.g. by introducing into
the catalyst the promoters like Cl, Cs and Rh [2–6]. Moreover, in
order to inhibit the combustion of ethylene and EO, the ethylene
conversion per-pass must be maintained at a relatively low level
(4–8%), that leads to a limited EO productivity and a high energy-
consuming for recycling ethylene. The CO2 emission as by-product
accounts for about 3.4 million metric tons per year in this process,
ranking only second to ammonia synthesis. Thus, the gas-phase
oxidation process for EO production is arising big environmental
concerns.
Recently, the researchers at the Center for Environmentally
Beneficial Catalysis (CEBC) made a comprehensive economic and
environmental assessment for the EO production by shifting the
conventional vapor-phase process to a CO2 zero emission pro-
cess operated under mild liquid-phase conditions of 40 ◦C and
5 MPa [1,7]. The so-called CEBC process uses methyltrioxorhenium
homogeneous catalyst and environmental-friendly oxidant H2O2,
providing an excellent catalytic performance of almost 100% EO
∗
Corresponding author at: Shanghai Key Laboratory of Green Chemistry and
Chemical Processes, School of Chemistry and Molecular Engineering, East China
Normal University, North Zhongshan Rd. 3663, Shanghai 200062, China.
E-mail address: pwu@chem.ecnu.edu.cn (P. Wu).
http://dx.doi.org/10.1016/j.apcata.2016.02.001
0926-860X/© 2016 Elsevier B.V. All rights reserved.

52
X. Lu et al. / Applied Catalysis A: General 515 (2016) 51–59
Table 1
Physicochemical properties of various titanosilicates.
Catalyst
Structure Si/Tia Crystal sizeb (␮m) SAc (m2 g−1
)
Ti statesd
d
Ti-MWW MWW
44
50
51
38
0.6 × 0.6 × 0.1
0.2–0.3
0.2–0.4
∼1
545
527
572
614
Tetra.
Tetra.
Tetra.
Tetra.
TS-1
MFI
Ti-MOR
Ti-Beta
MOR
BEA*
c
b
a
a
Molar ratio determined by ICP analysis.
Evaluated by SEM.
b
c
Specific surface area (Langmuir) measured by N2 adsorption at −196 ◦C.
d
Evaluated with UV–vis spectroscopy. Tetra., tetrahedral Ti species.
5
10
15
20
25
30
35
selectivity and H2O2 efficient utilization. The preliminary economic
assessment on this CEBC process shows that the higher oxidant and
catalyst costs can be offset by its higher EO selectivity. Neverthe-
less, the low abundance (10−7%) of precise Re metal ($1400/1b)
and the reuse of homogeneous catalyst would become the main
bottlenecks for actual industrialization of bulk chemicals like EO
[8]. Therefore, developing alternative process based on heteroge-
neous catalysts useful for the liquid-phase ethylene epoxidation
with H2O2 is highly desirable. It may lead to a simple and eco-
efficient EO process of easy separation and recycle/regeneration.
Nb and W incorporated silica-based cubic mesoporous materials
(W-KIT-6 and Nb-KIT-6) have been investigated as heterogeneous
catalysts, showing a significant activity in the ethylene epoxidation
with H2O2. They are barely satisfactory due to inefficient decom-
position of H2O2 and metal leaching [9].
In view of the alkene epoxidations featured with low carbon
emission and high efficient utilization of carbon source, Hydrogen
Peroxide Propylene Oxide (HPPO) process has been commercial-
ized by Dow-BASF for the production of propylene oxide (PO) [10].
The HPPO process, based on the TS-1/H2O2 catalytic system, gives
water as the main byproduct. A small amount of alcohol ethers
due to the solvolysis of PO are also produced as TS-1 prefers protic
methanol as the solvent. Up to now, a large number of theoretical
studies have performed on the Hydrogen Peroxide Ethylene Oxide
(HPEO) process [11–14]. However, the experimental study on HPEO
process has not made a substantial progress, due to the relatively
low activity and selectivity achieved on conventional titanosili-
cates like TS-1, Ti-STT and Ti-CHA [15,16]. In early works, we once
reported that Ti-MWW was superior to TS-1 in HPPO process in
terms of PO selectivity and yield [17] as well as in the epoxidation
of other functional groups-containing alkenes with H2O2 [18–24].
Its catalytic properties and possible application to HPEO process
are still unknown.
2 Theta (degree)
Fig. 1. XRD patterns of Ti-MWW (a), TS-1 (b), Ti-MOR (c) and Ti-Beta (d).
following literature method [25]. The synthetic gel with a molar
composition of 1.0 SiO2: 0.05 TiO2: 1.4 PI: 0.67 B2O3: 19H2O was
hydrothermally crystallized at 170 ◦C for 7 days, then the powder
product obtained was refluxed in a 2 M HNO3 aqueous solution
for the purpose of removing extraframework Ti species and a part
of framework boron as well. TS-1 was hydrothermally synthesized
using tetrapropyl hydroxide (TPAOH), tetraethyl silicate (TEOS) and
tetrabutyl orthotitanate (TBOT) as SDA, Si and Ti sources, respec-
tively [26]. To remove extraframework Ti species and residual alkali
ions contaminated in TPAOH solution, the obtained TS-1 was fur-
ther washed with 1 M HCl solution before calcination at 550 ◦C
for 6 h in air. Ti-MOR was post-synthesized by the atom-planting
method between highly dealuminated mordenites and TiCl4 vapor
at elevated temperature [27–30]. Ti-Beta has synthesized in fluo-
ride medium according to the literature [31].
2.2. Characterization methods
The X-ray diffraction (XRD) patterns were recorded on a
Rigaku Ultima IV diffractometer using Ni-filtered Cu K␣ radia-
tion (␭ = 0.1541 nm) in a scanning range of 2␪ = 5–35 to confirm
the structure and crystallinity of the titanosilicates. The voltage
and current were 35 kV and 25 mA, respectively. Morphologies and
crystal sizes were examined by a Hitachi S-4800 scanning electron
microscope. The UV–vis spectra were collected on a PerkinElmer
UV–vis Lambda 35 spectrophotometer using BaSO4 as a reference.
The FT-IR spectra were recorded by a Nicolet Nexus 670 FT-IR spec-
trometer at a resolution of 2 cm−1 using a KBr technique. The bulk
Si/Ti ratios were determined by ICP-AES on IRIS Intrepid II XSP after
dissolving the titanosilicates in HF solution. The amount of acid
sites was determined by temperature-programmed desorption
of ammonia (NH3-TPD) with a Micrometrics tp-5080 equipment
equipped with a thermal conductivity detector (TCD). Typically,
0.1 g of sample was pretreated in helium stream (25 mL min−1) at
550 ◦C for 1 h. The adsorption of NH3 was carried out at 50 ◦C for
1 h. The sample was flushed with helium at 100 ◦C for 2 h to remove
physisorbed NH3 from the catalyst surface. The TPD profiles were
then recorded at a heating rate of 10 ◦C min−1 from 100 ◦C to 550 ◦C.
The textural properties of the titanosilicates were determined by
N2 adsorption at −196 ◦C using a BEL SORP instrument after the
samples were degassed in vacuum at 300 ◦C for 6 h.
In this work, with the purpose to develop efficient HPEO process,
we have systematically studied the titanosilicate-catalyzed epox-
idation of ethylene with H2O2. By comparing four representative
titanosilicates of different topologies, Ti-MWW/H2O2/MeCN was
confirmed as the best reaction system with high reaction activity
and EO selectivity in HPEO process.
2. Experimental
2.1. Reagents and materials
Ethylene with a purity of 99.99% was procured from Shang-
hai Pujiang Special Gases Co., Ltd., China and hydrogen peroxide
(30 wt.%) was supplied by Sinopharm Chemical Reagent Co., Ltd.,
China. All other analytical reagents (MeOH, MeCN, acetone, tert-
butyl alcohol etc.) were commercially available and they are used
without further purification.
Four titanosilicates with different topologies have been
employed in the liquid-phase epoxidation of ethylene. Ti-MWW
was prepared using boric acid as a crystallization-supporting and
piperidine (PI) as a structure-directing agent (SDA) in two steps
2.3. Catalytic reactions
The selective epoxidation of ethylene to EO was carried out
in an autoclave reactor equipped with a 45 mL Teflon-inner. In a
typical run, 150 mg titanosilicate, 10 g MeCN and 10 mmol H2O2
(30 wt.%) were fed into the reactor. Ethylene was charged into
the autoclave to replace the air inside for three times. The pres-

X. Lu et al. / Applied Catalysis A: General 515 (2016) 51–59
53
Fig. 2. SEM images of Ti-MWW (a), TS-1 (b), Ti-MOR (c) and Ti-Beta (d).
Table 2
sure was then adjusted to a constant value of 2.5 MPa. After the
reaction was carried out under vigorous stirring at 40 ◦C for 1.5 h,
the reactor was cooled down by ice water and depressed slowly
before opening. For other n-alkenes (i.e. propylene, 1-butene, 1-
pentene and 1-hexene), the epoxidation was also carried out in the
autocalve reactor. The reaction products were analyzed by a gas
chromatograph (Shimadzu 2014, FID detector) using isopropanol
as an internal standard and the remaining amount of H2O2 after
reaction was determined by standard titration method with 0.05 M
Ce(SO4)2 solution. The products formed were further confirmed
by a GC–MS (Agilent 6890 series GC system, 5937 network mass
selective detector).
The ring-opening reactions of EO or PO over Ti-MWW and TS-1
were operated in the abovementioned autoclave reactor at 40 ◦C
for 1 h under N2 pressure of 2.5 MPa. With or without adding H2O2,
10 mmol epoxyalkane (EO or PO) and 150 mg catalyst (Ti-MWW
or TS-1) were added into 10 g different solvents (MeCN, MeOH or
H2O). The conversion of EO or PO was determined by GC analysis,
which was used to evaluate the ring-opening activity.
A comparison of epoxidation of ethylene over various titanosilicates in different
solventsa.
No. Catalyst
Products yieldb (%) Products distributionc (%) H2O2 (%)
EO
EG ME
others conv. eff.
d
d
1
2
3
4
Ti-MWW 44.3
98.8 1.2
–
–
54.4
88.0
20.4
5.9
81.4
98.9
78.3
30.9
TS-1
Ti-MOR
Ti-Beta
87.1
15.9
1.8
15.2 1.7 82.3 0.8
d
d
d
100.0
–
–
–
–
–
d
d
94.5 5.5
a
Reaction conditions: cat., 150 mg; ethylene, 2.5 MPa; H2O2, 10 mmol; solvent,
10 g; temp., 40 ◦C; time, 1.5 h. MeOH was used as the solvent for TS-1, whereas MeCN
was used for the reactions on other titanosilicate catalysts.
b
Including ethylene oxide and corresponding byproducts formed by hydrolysis
or solvolysis.
c
EO, ethylene oxide; EG, ethylene glycol; ME, methyl ether.
Not determined.
d
titanosilicates. The presence of the band at 960 cm−1 in IR spectra
(Fig. 3B) further confirmed that the Ti species was introduced into
the framework [33]. The ICP analysis showed that the Si/Ti ratios
were comparable for these four titanosilicates (Table 1). Their tex-
tural properties, measured by nitrogen adsorption technique, are
also summarized in Table 1. The specific surface areas (Langmuir)
were in the range of 527–614 m2 g−1. Thus, the four titanosilicates
were highly porous and crystalline materials with the isolated Ti
ions in framework, and qualitatively they are considered to qualify
the preconditions as the liquid-phase oxidation catalysts.
3. Results and discussion
3.1. Characterization of titanosilicate catalysts
The XRD patterns verified that the four titanosilicates were
highly crystalline materials free of impurity and possessed the
topologies of MWW, MFI, MOR and BEA* structures (Fig. 1). The
SEM images provided the morphologies of their crystals. Ti-MWW
crystals showed a uniformly platelet-shaped morphology with a
thickness of about 0.1 ␮m (Fig. 2a). The crystals of TS-1 and Ti-MOR
both were composed of the aggregate of nanocrystals (Fig. 2b and
c). Ti-Beta exhibited a morphology of truncated square bipyramid
(Fig. 2d), that was in accordance with the literature [32]. As shown
in UV–vis spectra (Fig. 3A), the predominant band at 220 nm was
ascribed to the tetrahedrally coordinated Ti4+ species in the zeo-
lite framework. Meanwhile, the band around 330 nm was nearly
negligible, indicating the absence of anatase-like phase in these
3.2. Epoxidation of ethylene
3.2.1. A comparison of ethylene epoxidation over various
titanosilicates
In order to choose a suitable catalyst for the HPEO process, the
titanosilicates with different topologies were investigated for the
epoxidation of ethylene, and the results are summarized in Table 2.
As the Ti contents (or Si/Ti molar ratios) were comparable for four
catalysts, the yield of oxygenated products and EO selectivity are
considered to represent reliably their efficiency for this reaction.

54
X. Lu et al. / Applied Catalysis A: General 515 (2016) 51–59
220
960
A
B
d
c
b
a
330
d
c
b
a
1000
900
800
700
600
200
300
400
500
Wavelength (nm)
Wavenumber (cm-1)
Fig. 3. UV–vis spectra (A) and FT-IR spectra (B) of Ti-MWW (a), TS-1 (b), Ti-MOR (c) and Ti-Beta (d).
MeCN was chosen as the most appropriate solvent for Ti-MWW,
Ti-MOR and Ti-Beta, while MeOH was used as the most suitable
one for TS-1. Table 3 compares in detail the solvent effects on the
activity and EO selectivity between Ti-MWW and TS-1. Judging
from the products yield and H2O2 conversion, the catalytic activ-
ity decreased in the order of TS-1 > Ti-MWW > Ti-MOR > Ti-Beta.
Although TS-1 showed the highest H2O2 conversion and products
yield, the solvolysis of EO in protic solvent of MeOH lowered the
EO selectivity to a level of 15.2%. On the contrary, Ti-MWW gave an
extremely high EO selectivity (98.8%) in aprotic solvent of MeCN.
When shifting the solvent from MeOH to MeCN for TS-1, the EO
selectivity was enhanced greatly to reach 99.9%, but the products
yield was only 5.2% (Table 3). Thus, balancing the EO selectivity and
catalytic activity (oxygenated product yield and H2O2), Ti-MWW is
assumed to be the most effective catalyst among the investigated
titanosilicates in terms of developing a selective HPEO process.
To further investigate how the solvent characters influence the
EO selectivity, the hydrolysis or solvolysis of EO in various sol-
vents was carried out over Ti-MWW and TS-1 with or without
H2O2 (Fig. 4). The main products were ethylene glycol (EG) and
methyl ether. This reaction proceeded slightly even in the absence
of any catalysts and it was accelerated greatly by the presence of
titanosilicate catalysts. The ring-opening of EO depends on the con-
centration of weak acid sites derived from Si-OH, Ti-OH or Ti- O-O-H
in the titanosilicates [23]. TS-1 showed a higher EO conversion than
Ti-MWW when compared in the same solvent, indicating the for-
mer may contain more or stronger acid sites. We have measured
the NH3-TPD profiles of Ti-MWW and TS-1. Ti-MWW almost did
not show any chemical adsorption of ammonia, whereas TS-1 also
showed a NH3-TPD profile with an extremely low signal/noise ratio
(Fig. S1). This implies that both Ti-zeolites were characteristic of
very weak acidity. It was actually observed that the EO conver-
sion decreased in the order of MeOH > H2O > MeCN ≈ 0. The aprotic
solvent molecules of MeCN with a weak basity could retard the
ring-opening of EO. Moreover, the addition of H2O2 accelerated
the hydrolysis or solvolysis of EO. This further verified that carry-
ing out ethylene epoxidation in a favorable solvent of MeCN allows
Ti-MWW to achieve a high activity along with a high EO selectivity.
Based on the above results and the GC–MS analysis, we iden-
tified the hydrolysis and solvolysis products and confirmed the
reaction pathways in HPEO process with the absence or presence
of MeOH (Scheme 1). The main product was EO as a result of the
epoxidation of ethylene. As the hydration of EO can proceed in the
absence of any catalysts, it will occur along with the ethylene epox-
idation due to the presence of H2O from aqueous solution of H2O2
oxidant and its decomposition. The formed EG could further react
with EO to produce the corresponding polyols. When MeOH was
used as the solvent, the solvolysis of EO rather than the hydrolysis
would play a dominant role in the ring-opening reaction, produc-
ing the byproducts of corresponding methyl ethers. Similar results
have been reported previously for TS-1-catalyzed propylene epox-
idation [34–36].
3.2.2. Effect of solvents
Significant solvent effects are generally involved in
titanosilicate-catalyzed reactions. In this section, the solvent
effects on the Ti-MWW or TS-1-catalyzed HPEO processes were
investigated emphatically. The characters of solvents had a great
impact on not only the catalytic activity but also the product
selectivity. As shown in Table 3, Ti-MWW and TS-1 exhibited
significantly different solvent effects.
Table 3
A comparison of epoxidation of ethylene over Ti-MWW and TS-1 in different sol-
vents a
.
Catalyst Solvent Products yieldb (%) Product sel.c (%)
EO EG ME others conv. eff.
H2O2 (%)
d
d
MeCN
MeOH
Acetone 72.4
H2O 48.0
t-BuOH 59.0
44.3
7.0
98.8 1.2
–
–
54.4 81.4
7.5 93.2
92.4 78.4
86.7 55.4
82.7 71.4
Ti-
MWW
O
d
HO
24.1
–
74.2 1.7
O
d
OH
HO
82.6 0.4
72.3 13.7
82.0 2.0
–
–
–
17.0
14.0
16.0
H+
[H2O]
d
d
n
OH
n = 1 or 2
O
H+
[O]
Ti4+
O
d
d
TS-
1
MeCN
MeOH
Acetone 44.7
H2O
t-BuOH
5.2
87.0
99.9 0.1
–
–
5.2 99.2
88.0 98.9
76.7 58.3
44.1 45.0
9.8 45.4
H+
O
15.2 1.7 82.3 0.8
HO
d
31.1 1.4
20.3 69.7
82.1 1.9
–
–
–
67.5
10.0
16.0
[MeOH]
HO
OMe
d
d
H+
19.8
4.4
OMe
a
Reaction conditions: cat., 150 mg; ethylene, 2.5 MPa; H2O2, 10 mmol; solvent,
10 g; temp., 40 ◦C; time, 1.5 h.
Hydration
or
Epoxidation
Condensation
b
Including ethylene oxide and corresponding byproducts formed by hydrolysis
Solvolysis
or solvolysis.
c
EO, ethylene oxide; EG, ethylene glycol; ME, methyl ether.
Not determined.
d
Scheme 1. Reaction pathways of ethylene epoxidation and ring-opening reaction.

X. Lu et al. / Applied Catalysis A: General 515 (2016) 51–59
55
100
80
60
40
20
0
100
B
A
80
60
40
20
0
MeCN
H2O
MeOH
H2O
MeCN
MeOH
Fig. 4. Ring-opening reaction of EO in different solvents with H2O2 (black) or without H2O2 (blank) over Ti-MWW (A) or TS-1 (B). Reaction conditions: Ti-MWW or TS-1,
150 mg; EO, 10 mmol; N2, 2.5 MPa; solvent, 10 g; H2O2 (if added), 10 mmol; temp., 40 ◦C; time, 1 h.
For both Ti-MWW and TS-1, the EO selectivity was relatively
low (<90%) in the solvents other than MeCN, indicating EO is chemi-
cally unstable, tending to undergo hydration or solvolysis reactions.
This is consistent with the results observed in the abovementioned
ring-opening reaction of EO (Fig. 4). Hence, MeCN is presumed to
be a more suitable solvent for Ti-MWW, although the H2O2 con-
version and the products yield in MeCN were not the highest. It
is of particular interests that Ti-MWW showed a higher reaction
activity in water than that in MeCN, although the H2O2 utilization
efficiency for the former was low as 55.4%. The effect of solvent
is one of the most complicated issues in the catalytic system of
titanosilicate/H2O2. It is related to the polarity of the solvent, the
hydrophilic/hydrophobic character of the titanosilicate surface, the
solubility of the substrates, and other factors. Ti-MWW prepared
from a layered precursor contains many defect sites such as silanol
groups as a result of incomplete dehydroxylation between the lay-
ers. Thus, the solvent of MeOH or H2O molecules adsorbed on the
surface silanol groups that would impose a steric hindrance on
the guest molecules (solvents, substrates and H2O2 molecules) and
limit their diffusion and approach to the Ti active site within the
zeolite channels. Thus, the reactivity of Ti-MWW is higher in H2O
than that of MeOH due to the smaller diameter of the former one,
although the ethylene solubility is much higher in MeOH compared
to H2O.
In the case of TS-1, MeOH was obviously the most efficient
solvent from the viewpoint of catalytic activity. Aprotic but basic
MeCN, a well-known unsuitable solvent for the epoxidation of sim-
ple alkenes on TS-1 catalyst, made its reaction activity somewhat
low in the epoxidation of ethylene. A similar phenomenon was also
observed for the solvent of t-BuOH.
The above results indicated that Ti-MWW may served as an
appropriate and efficient catalyst for the selective formation of EO
in the ethylene epoxidation when choosing MeCN as a solvent.
100
80
80
60
35 oC
35 oC
60
40 oC
40
20
40 oC
40
45 oC
45 oC
20
0
50 oC
50 oC
B
A
0
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
Time (h)
Time (h)
100
80
60
40
20
0
100
90
35 oC
40 oC
45 oC
50 oC
35 oC
40 oC
45 oC
50 oC
80
70
C
D
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
Time (h)
Time (h)
Fig. 5. Changes of products yield (A), EO selectivity (B), H2O2 conversion (C) and H2O2 utilization efficiency (D) with time on stream at different temperatures. Reaction
conditions: Ti-MWW, 150 mg; ethylene, 2.5 MPa; MeCN, 10 g; H2O2, 10 mmol.

56
X. Lu et al. / Applied Catalysis A: General 515 (2016) 51–59
100
80
60
40
20
0
100
TS-1 eff.
Ti-MWW eff.
Ti-MWW sel.
TS-1 yield
80
60
40
20
0
Ti-MWW yield
TS-1 sel.
TS-1 conv.
B
A
Ti-MWW conv.
0
1
2
3
4
5
0
1
2
3
4
5
Time (h)
Time (h)
Fig. 6. Dependence of products yield and EO selectivity (A), and H2O2 conversion and efficiency (B) with time on stream of Ti-MWW and TS-1. Reaction conditions: Ti-MWW
or TS-1, 150 mg; ethylene, 2.5 MPa; MeCN for Ti-MWW and MeOH for TS-1, 10 g; H2O2, 10 mmol; temp., 40 ◦C.
3.2.3. Effect of reaction time at different temperatures
3.2.4. Effect of reaction pressure and catalyst amount
The effect of reaction time on the performance of ethylene epox-
idation over Ti-MWW was investigated at 35, 40, 45 and 50 ◦C,
respectively (Fig. 5). Both the products yield and the H2O2 con-
version increased rapidly at the beginning of epoxidation reaction
(Fig. 5A and C), and then slowed down as H2O2 was gradually con-
sumed with prolonging the time on stream. The H2O2 utilization
efficiency decreased gradually with the reaction time (Fig. 5D), indi-
cating the aggravation of non-productive decomposition of H2O2
because of catalysts deactivation. The EO selectivity, however, was
always maintained at >95% (Fig. 5B). Fig. 5 shows that the epoxida-
tion of ethylene favored at lower temperatures (e.g. 35 ◦C), where
higher H2O2 utilization efficiency and EO selectivity were achiev-
able when compared at a comparable H2O2 conversion with the
reactions at higher temperature like 50 ◦C.
In addition, we have compared the time course of ethylene
epoxidation between Ti-MWW and TS-1 with the purpose to fur-
ther investigate their catalytic performances (Fig. 6). From the
time-dependent courses of H2O2 conversion and products yield,
TS-1/MeOH was obviously more active than Ti-MWW/MeCN when
the reactions were conducted under the same reaction condi-
tions except for the solvent. Most of EO was transformed to EG
and methyl-ether in the catalytic system of TS-1/MeOH due to
hydrolysis and solvolysis whereas the ring-opening reaction of
EO was almost inert in the system of Ti-MWW/MeCN. Hence, Ti-
MWW/MeCN is more suitable and selective than TS-1/MeOH for
developing HPEO process.
The reaction pressure had a significant influence on the per-
formance of ethylene epoxidation over Ti-MWW. As shown in
Fig. 7, both H2O2 conversion and products yield were linearly pro-
portional to the reaction pressure. The pressure showed a more
significant effect on the ethylene expoxidation than that previously
observed on the propylene expoxidation [17,35], simply because
the range of the adjustable reaction pressure for ethylene is wider
than that of propylene due to its higher vapor pressure. Thus, more
ethylene would dissolve in the solvent (MeCN) with increasing
reaction pressure, which promoted the epoxidation of ethylene.
On the other hand, the EO selectivity and H2O2 efficiency hardly
varied with the reaction pressure.
Fig. 8 shows the effect of catalyst amount on the epoxidation
of ethylene with H2O2 in the Ti-MWW/MeCN system. The H2O2
conversion increased from 48.4% to 97.0% by increasing the cata-
lyst amount from 50 mg to 150 mg and by further increasing the
catalyst amount the conversion remained unchanged. The prod-
ucts yield increased gradually and reached the maximum at 150 mg
catalyst loading. Increasing the catalyst amount tended to lower
the H2O2 efficiency because of competition between epoxidation
and H2O2 decomposition on the Ti active sites. Therefore, there
was an optimum catalyst amount for the batchwise epoxidation of
ethylene.
3.2.5. Stability and reusability of Ti-MWW
The stability and reusability of a heterogeneous catalyst was
the essential factor for actual application in industrial processes.
We have checked the catalytic cycles of Ti-MWW in the epoxi-
dation of ethylene (Fig. 9). The used Ti-MWW catalyst was first
100
80
EO sel.
H2O2 eff.
80
60
40
60
40
Products yield
H2O2 conv.
20
A
B
20
0
0
1
2
3
4
0
1
2
3
4
Reaction pressure (MPa)
Reaction pressure (MPa)
Fig. 7. Changes of products yield and EO selectivity (A), and H2O2 conversion and H2O2 efficiency (B) with the reaction pressure. Reaction conditions: Ti-MWW, 150 mg;
MeCN, 10 g; H2O2, 10 mmol; temp., 40 ◦C; time, 1.5 h.

X. Lu et al. / Applied Catalysis A: General 515 (2016) 51–59
57
100
80
60
40
20
0
100
EO sel.
80
60
40
20
0
H2O2 eff.
B
Products yield
H2O2 conv.
A
0
50
100
150
200
250
0
50
100 150
200
250
Catalyst amount (mg)
Catalyst amount (mg)
Fig. 8. Dependence of products yield and EO selectivity (A), and H2O2 conversion and efficiency (B) on the amount of Ti-MWW. Reaction conditions: ethylene, 2.5 MPa;
MeCN, 10 g; H2O2, 10 mmol; temp., 40 ◦C; time, 4.5 h.
100
80
60
40
20
0
100
80
60
40
20
0
H2O2 conv.
A
B
EO sel.
H2O2 eff.
Further
calcined
Further
Products yield
Washed with acetone
calcined
Washed with acetone
0
2
4
6
8
10
12
0
2
4
6
8
10
12
Reuse number
Reuse number
Fig. 9. Changes of products yield and EO selectivity (A), and H2O2 conversion and H2O2 utilization efficiency (B) with the reaction-regeneration cycles on Ti-MWW. Reaction
conditions for the first run: Ti-MWW, 150 mg; ethylene, 2.5 MPa; MeCN, 10 g; H2O2, 10 mmol; temp., 40 ◦C; time, 4.5 h. The next catalytic runs proceed at a constant ratio of
catalyst-oxidant-solvent.
recycled by centrifugation and washing with acetone, and then it
was subjected to next catalytic runs. In the first three runs, both the
H2O2 conversion and the products yield dropped largely with the
reaction-regeneration cycles. The H2O2 conversion was maintained
at a relatively high level even after nine recycles. When the reused
catalyst was regenerated by washing with acetone and further cal-
cinated at 550 ◦C for 6 h, the activity of Ti-MWW was almost totally
recovered at the tenth run. As shown in Fig. 9B, H2O2 efficiency
decreased slightly with the reaction-regeneration cycles due to the
deactivation of the catalyst and it will be almost totally recovered
at the tenth run. The Ti content of reused Ti-MWW was close to
that of fresh one (Table 4), indicating that the leaching of Ti species
almost did not occur in repeated catalytic runs. The high specific
surface area verified that the regenerated Ti-MWW catalyst was
still a highly porous material. Furthermore, no significant differ-
ence was found in the XRD patterns and UV–vis spectra between the
fresh Ti-MWW and the recovered one (Fig. 10). The boron species
co-existing in the framework was decreased to an extremely level
220
B
A
c
330
c
b
b
a
a
5
10
15
20
25
30
35
200
300
400
500
2 Theta (degree)
Wavelength (nm)
Fig. 10. XRD patterns (A) and UV–vis spectra (B) of fresh Ti-MWW (a), recovered one washed with acetone and dried (b) and further calcined at 550 ◦C for 6 h (c) after
Ti-MWW was reused in the ethylene epoxidation for 8 times.

58
X. Lu et al. / Applied Catalysis A: General 515 (2016) 51–59
Table 4
the diffusion problem inside zeolite micropores would emerge seri-
ously for large substrate and product molecules. Thus, Ti-MWW
and TS-1 both showed the maximum activity for propylene, and
their H2O2 conversion and products yield dropped greatly from
propylene to 1-hexene mostly due to the steric hindrance.
There was a big gap in the catalytic activity between ethylene
epoxidation and propylene epoxidation on both Ti-MWW and TS-1.
This result suggests that electron-deficient ethylene, although with
small molecular dimension and suffering low diffusion hindrance
by zeolite channels, is intrinsically inactive towards epoxidation.
Thus, for developing practical HPEO process, it is desirable to adopt
highly efficient titanosilicate catalysts capable of activating the
alkene molecules with a low intrinsic reactivity under mild con-
ditions and simultaneously suppressing the solvolysis or hydration
of EO.
Physicochemical and textural properties of fresh, used and regenerated Ti-MWW
catalysts.
No.
Catalyst
SAa
Ti contentb
B contentb
(mmol g−1
)
(m2 g−1
)
(mmol g−1
)
1
2
3
Ti-MWW-fresh
545
514
557
0.39
0.38
0.38
0.16
0.02
0.02
Ti-MWW-usedc
Ti-MWW-regenerationd
a
Specific surface area (Langmuir) measured by N2 adsorption at −196 ◦C.
Ti and B contents determined by ICP analysis.
b
c
Recovered one after Ti-MWW reused in the ethylene epoxidation for 8 times.
d
Regeneration: washed with acetone and dried at 120 ◦C, and further calcined at
550 ◦C in air for 6 h.
after the reuse (Table 4). They were unstable simply because of a too
small ionic radius in comparison to Si ions, which was in agreement
with the previously observed in the epoxidation of various alkenes
like allyl alcohol, diallyl ether and allyl chloride [20,21,23]. Consid-
ering the EG molecules, derived from the hydrolysis of EO, posses
a high boiling point, a partial deactivation observed for the regen-
erated catalyst by only acetone washing could be ascribed to the
organic compound deposition inside the channels and the covering
of Ti active sites. This kind of deactivation was reversible unless the
crystalline structure and the Ti sites were well maintained. Hence,
burning off the organic compound efficiently restored the catalytic
activity.
4. Conclusions
Ti-MWW is an effective catalyst for the selective oxida-
tion of ethylene with H2O2 to ethylene oxide. With suitable
hydrophilic/hydrophobic characters favoring MeCN as solvent, Ti-
MWW gives the H2O2 conversion and the ethylene oxide selectivity
both above 95% under optimized reaction conditions. TS-1 is even
more active than Ti-MWW for ethylene epoxidation, but it can-
not suppress the solvolysis of ethylene oxide because of preferring
MeOH solvent. Additionally, Ti-MWW proves to be superior to TS-
1 in the selective oxidation of other linear alkene (from propylene
to 1-hexene) with H2O2. No structural degradation and Ti leaching
occur during the repeated use of Ti-MWW. The formation of heavy
products like ethylene glycol may deactivate the Ti-MWW catalyst,
which can be recovered by calcination. This study indicates that
Ti-MWW is a promising catalyst for the selective synthesis of ethy-
lene oxide and highly active titanosilicate catalyst is still desirable
to develop more efficient HPEO process.
3.3. Comparison of Ti-MWW and TS-1 for the epoxidation of
different linear alkenes
Ethylene is a simple alkene with the smallest molecular dimen-
sion and carbon number. It is expected that ethylene may suffer
less diffusion problem in the liquid-phase epoxidation in compar-
ison to other linear alkenes. Thus, it is desirable to know how
the catalytic performance of epoxidation depends on the molec-
ular size of alkenes, which would be helpful to understand the
fundamental issues involved in developing HPEO process. Table 5
compares the catalytic behaviors in the epoxidation of different
linear alkenes between Ti-MWW and TS-1. The activity of ethy-
lene epoxidation of Ti-MWW/MeCN was inferior to TS-1/MeOH,
but the EO selectivity for the former (99.6%) was almost 7 times
of that for the latter (14.5%) (Table 5, No. 1). For larger alkenes
(C3 –C6 ), Ti-MWW/MeCN was superior to TS-1/MeOH not only
in epoxylalkane selectivity but also in reactivity. Benefited from the
electron-donating effects of the methyl groups, the electron den-
sity of the C C bond in propylene is higher than that of ethylene,
making the propylene molecules epoxidized more easily. With the
substituent groups on the C C bonds becomes longer or larger, the
intrinsic reactivity of alkenes tends to be more active. Meanwhile,
Acknowledgments
We gratefully acknowledge the National Natural Science Foun-
dation of China (21533002, 21373089, 21403071), the China
Postdoctoral Science Foundation (2013M541493), the PhD Pro-
grams Foundation of the Ministry of Education (2012007613000),
and the Shanghai Leading Academic Discipline Project (B409).
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.apcata.2016.02.
001.
References
Table 5
Comparison of Ti-MWW and TS-1 for the epoxidation of various linear alkenesa.
[1] M. Ghanta, B. Subramaniam, H. Lee, D.H. Busch, AIChE J. 59 (2013) 180–187.
[2] D. Torres, F.I.I. Ias, R.M. Lambert, J. Catal. 260 (2011) 380–383.
[3] T.C.R. Rocha, M. Havecker, A. Knop-Gericke, R. Schlogl, J. Catal. 312 (2014)
12–16.
[4] M.O. Ozbek, I. Onal, R.A. van Santen, ChemCatChem 5 (2013) 443–451.
[5] J.C. Dellamorte, J. Lauterbach, M.A. Barteau, Catal. Today 120 (2006) 182–185.
[6] W.J. Diao, C.D. DiGiulio, M.T. Schaal, S. Ma, J.R. Monnier, J. Catal. 322 (2015)
14–23.
[7] H. Lee, M. Ghanta, D.H. Busch, B. Subramaniam, Chem. Eng. Sci. 65 (2010)
128–134.
[8] M. Ghanta, T. Ruddy, D. Fahey, D. Busch, B. Subramaniam, Ind. Eng. Chem. Res.
52 (2013) 18–29.
No. Substrate
Ti-MWW (%)
TS-1 (%)
products
yieldb
H2O2
conv.
oxide
sel.
products
yieldb
H2O2
conv.
oxide
sel.
1
2
3
4
5
ethylenec
popylenec
1-butenec
6.6
94.9
64.3
8.1
98.9
68.0
39.1
29.1
99.6
99.5
99.8
99.5
99.3
13.0
89.3
48.5
22.4
15.6
13.2
91.6
52.0
30.4
20.2
14.5
82.8
92.2
87.6
92.4
1-pentened 35.5
1-hexened 26.9
a
MeCN as solvent for Ti-MWW and MeOH as solvent for TS-1.
Including ethylene oxide and corresponding byproducts formed by hydrolysis
[9] W.J. Yan, A. Ramanathan, M. Ghanta, B. Subramaniam, Catal. Sci. Technol. 4
(2014) 4433–4439.
b
or solvolysis.
[10] B. Subramaniam, G.R. Akien, Curr. Opin. Chem. Eng. 1 (2012) 336–341.
[11] G.N. Vayssilov, R.A. van Santen, J. Catal. 175 (1998) 170–174.
[12] E. Karlsen, K. Schöffel, Catal. Today 32 (1996) 107–114.
[13] M. Neurock, L.E. Manzer, Chem. Commun (1996) 1133–1134.
[14] H. Munakata, Y. Oumi, A. Miyamoto, J. Phys. Chem. B 105 (2001) 3493–3501.
c
Reaction conditions: cat., 50 mg; alkene, 0.2 MPa; H2O2, 15 mmol; solvent, 10 g;
temp., 40 ◦C; time, 2 h.
d
Reaction conditions: cat., 50 mg; alkene, 15 mmol; N2, 0.2 MPa; H2O2, 15 mmol;
solvent, 10 g; temp., 40 ◦C; time, 2 h.

X. Lu et al. / Applied Catalysis A: General 515 (2016) 51–59
59
[15] E.A. Eilertsen, F. Giordanino, C. Lamberti, S. Bordiga, A. Damin, F. Bonino, U.
[25] P. Wu, T. Tatsumi, T. Komatsu, T. Yashima, Chem. Lett (2000) 774–775.
[26] M. Taramasso, G. Perego, B. Notari, U.S. Patent 4410501 (1983).
[27] J.H. Ding, L. Xu, Y.J. Yu, H.H. Wu, S.J. Huang, Y.L. Yang, J. Wu, P. Wu, Catal. Sci.
Technol. 3 (2013) 2587–2595.
[28] L. Xu, J.H. Ding, Y.L. Yang, P. Wu, J. Catal. 309 (2014) 1–10.
[29] J.H. Ding, P. Wu, Appl. Catal. A 488 (2014) 86–95.
Olsbye, K.P. Lillerud, Chem. Commun (2011) 11867–11869.
[16] E.A. Eilertsen, S. Bordiga, C. Lamberti, A. Damin, F. Bonino, B. Arstad, S. Svelle,
U. Olsbye, K.P. Lillerud, ChemCatChem 3 (2011) 1869–1871.
[17] F. Song, Y.M. Liu, L.L. Wang, H.J. Zhang, M.Y. He, P. Wu, Stud. Surf. Sci. Catal.
170 (2007) 1236–1243.
[18] P. Wu, T. Tatsumi, T. Komatsu, T. Yashima, J. Catal. 202 (2001) 245–255.
[19] P. Wu, T. Tatsumi, J. Phys. Chem. B 106 (2002) 748–753.
[20] P. Wu, T. Tatsumi, J. Catal. 214 (2003) 317–326.
[21] P. Wu, Y.M. Liu, M.Y. He, T. Tatsumi, J. Catal. 228 (2004) 183–191.
[22] P. Wu, D. Nuntasri, Y.M. Liu, H.H. Wu, Y.W. Jiang, W.B. Fan, M.Y. He, T.
Tatsumi, Catal. Today 117 (2006) 199–205.
[23] L.L. Wang, Y.M. Liu, W. Xie, H.J. Zhang, H.H. Wu, Y.W. Jiang, M.Y. He, P. Wu, J.
Catal. 246 (2007) 205–214.
[24] H.H. Wu, Y.M. Liu, L.L. Wang, H.J. Zhang, M.Y. He, P. Wu, Appl. Catal. A 320
(2007) 173–180.
[30] Y.L. Yang, J.H. Ding, B.S. Wang, J. Wu, C. Zhao, G.H. Gao, P. Wu, J. Catal. 320
(2014) 160–169.
[31] T. Blasco, M. Camblor, A. Corma, P. Esteve, J. Guil, A. Martinez, J.
Perdigón-Melón, S. Valencia, J. Phys. Chem. B 102 (1998) 75–88.
[32] M. Camblor, A. Corma, S. Valencia, J. Mater. Chem. 8 (1998) 2137–2145.
[33] P. Wu, T. Komatsu, T. Yashima, J. Phys. Chem. 100 (1996) 10316–10322.
[34] G. Li, X.S. Wang, H.S. Yan, Y.Y. Chen, Q.S. Su, Appl. Catal. A 218 (2001) 31–38.
[35] X.S. Wang, X.W. Guo, G. Li, Catal. Today 74 (2002) 65–75.
[36] G. Li, X.S. Wang, H.S. Yan, Y.H. Liu, X.W. Liu, Appl. Catal. A 236 (2002) 1–7.