A novel titanosilicate with MWW structure Catalytic properties in selective epoxidation of diallyl ether with hydrogen peroxide

Article abstract of DOI:10.1016/S0021-9517(04)00425-7

The catalytic activity and selectivity of Ti-MWW in the epoxidation of diallyl ether (DAE) with hydrogen peroxide to allyl glycidyl ether (AGE) and diglycidyl ether (DGE) have been studied by a comparison with those of TS-1, and the issues concerning the consecutive reaction and the selective production of AGE have been considered. Ti-MWW catalyzed the DAE epoxidation in the presence of aprotic solvents such as acetonitrile or acetone, and produced only minor levels of solvolysis products. Ti-MWW proved to be a reusable catalyst standing up to the Ti leaching and maintaining the catalytic activity and the product selectivity in the reaction-regeneration cycles. Studies with different solvents, Ti contents, reaction times, temperature, and catalyst amounts confirmed that the DAE epoxidation was a typical consecutive reaction with AGE as an intermediate product and DGE as a secondary one. The reaction rate for AGE formation was much faster than that for DGE, making the selective production of AGE possible by controlling the reaction up to a DAE conversion level of ca. 30%.

Full text of DOI:10.1016/S0021-9517(04)00425-7

Journal of Catalysis 228 (2004) 183–191
www.elsevier.com/locate/jcat
A novel titanosilicate with MWW structure
Catalytic properties in selective epoxidation of diallyl ether
with hydrogen peroxide
a,∗
a
a
b
Peng Wu , Yueming Liu , Minyuan He , Takashi Tatsumi
a
Shanghai Key Laboratory of Green Chemistry and Chemical Processes, Department of Chemistry, East China Normal University,
North Zhongshan Rd. 3663, Shanghai 200062, People’s Republic of China
b
Division of Materials Science & Chemical Engineering, Graduate School of Engineering, Yokohama National University, 79-5 Tokiwadai, Hodogaya-ku,
Yokohama 240-8501, Japan
Received 24 July 2004; revised 1 September 2004; accepted 2 September 2004
Available online 28 September 2004
Abstract
The catalytic activity and selectivity of Ti-MWW in the epoxidation of diallyl ether (DAE) with hydrogen peroxide to allyl glycidyl
ether (AGE) and diglycidyl ether (DGE) have been studied by a comparison with those of TS-1, and the issues concerning the consecutive
reaction and the selective production of AGE have been considered. Ti-MWW catalyzed the DAE epoxidation in the presence of aprotic
solvents such as acetonitrile or acetone, and produced only minor levels of solvolysis products. Ti-MWW proved to be a reusable catalyst
standing up to the Ti leaching and maintaining the catalytic activity and the product selectivity in the reaction–regeneration cycles. Studies
with different solvents, Ti contents, reaction times, temperature, and catalyst amounts confirmed that the DAE epoxidation was a typical
consecutive reaction with AGE as an intermediate product and DGE as a secondary one. The reaction rate for AGE formation was much
faster than that for DGE, making the selective production of AGE possible by controlling the reaction up to a DAE conversion level of
ca. 30%.

2004 Elsevier Inc. All rights reserved.
Keywords: Ti-MWW; TS-1; Titanosilicate; Diallyl ether; Allyl glycidyl ether; Diglycidyl ether; Hydrogen peroxide; Liquid-phase epoxidation
1
. Introduction
these processes with water as an only by-product has a great
impact on green and sustainable chemistry, and implies the
potential applications of titanosilicates in clean syntheses of
fine chemicals.
To solve the problems that medium pore TS-1 encounters
in bulky reactions, many others titanosilicates with larger
pores have also been developed thereafter by hydrothermal
synthesis or postsynthesis methods, for example, Ti-Beta
Titanosilicates containing tetrahedral Ti species in the
framework, being capable of activating hydrogen peroxide
under liquid-phase conditions, prove to be promising cata-
lysts for the selective oxidation of a variety of organic com-
pounds to corresponding oxygenated chemicals. The repre-
sentative titanosilicate has been TS-1 of the MFI structure,
which was reported two decades ago [1]. The discovery of
TS-1 has led to industrialized processes such as the hy-
droxylation of phenol to hydroquinone and catecol, and the
ammoximation of cyclohexanone to oxime. The success of
[
2,3], Ti-ZSM-12 [4], Ti-MOR [5], Ti-ITQ-7 [6], and Ti-
MCM-41 [7]. Particularly, Ti-Beta with 12-membered ring
MR) channels is a very attractive catalyst for the oxidation
(
of cyclic and branched alkenes and alkanes [2,3]. However,
they are not superior to TS-1 in the oxidation of small mole-
cules according to the intrinsic activity per Ti site and from
the viewpoint of industrialization in terms of catalytic ac-
*
Corresponding author. Fax: +86-21-5252-0411.
E-mail address: pwu@chem.ecnu.edu.cn (P. Wu).
0021-9517/$ – see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2004.09.001

184
P. Wu et al. / Journal of Catalysis 228 (2004) 183–191
tivity and product selectivity. Ti-Beta and Ti-MCM-41 also
have a serious weakness of Ti leaching in aqueous solutions
because of the high concentration of hydrophilic silanol
groups on the defect sites and the external surface.
Table 1
a
Preparation of Ti-MWW catalysts with various Ti contents
b
No. Gel
Ti-containing precursor
Ti-MWW catalysts
c
Si/Ti Si/B Si/Ti
Si/B Si/Ti Si/B SSA
Ti-MWW, a novel titanosilicate recently developed with
the MWW structure (generally known as MCM-22), shows
unique catalytic properties in the epoxidation of various
alkenes with hydrogen peroxide [8]. Ti-MWW consists of
1
2
3
4
5
6
∞
100
70
50
30
0.75
0.75
0.75
0.75
0.75
0.75
∞
120
63
51
31
12
13
12
12
11
13
∞
146
76
64
46
102
34
41
36
49
54
611
628
622
624
624
556
1
2-MR side cups and two independent interlayer and in-
20
21
38
tralayer 10-MR channels, one of which contains the su-
percages of 7.1 × 18 Å dimension [9]. Taking advantages
of this structural diversity, Ti-MWW is not only active for
the epoxidation of linear alkenes but also for cyclic alkenes.
Especially, Ti-MWW catalyzes the epoxidation of linear
alkenes more actively and selectively than TS-1; the ti-
tanosilicate has been reported to show the highest intrinsic
activity so far [8d]. The unusual catalytic performance of
Ti-MWW makes it a promising catalyst for the oxidation of
functionalized alkenes such as allyl alcohol (AAL), in which
a
Crystallization: stirred at 403 K for 1 day, at 423 for 1 day, and then at
4
43 K for 5 days.
b
Acid treated and calcined. Acid treatment: solid to liquid (2 M HNO )
3
ratio of 1 g:20 mL; refluxed for 20 h. Calcination: at 823 K in air for 10 h.
c
2
−1
Specific surface area (Langmuir) in m g .
to the procedures reported previously [8b]. The synthesis
was carried out at molar compositions of 1 SiO2:0–0.05
TiO2:0.67 B2O3:1.4 SDA:19 H2O. Ti-containing MWW
lamellar precursor was obtained after the gel was crystal-
lized at 403–443 K. The precursor was then subjected to acid
treatment with the purpose of removing extraframework Ti
species and a part of framework boron as well. A subsequent
calcination on the acid-treated sample resulted in the objec-
tive Ti-MWW catalysts of the three-dimensional crystalline
structure. The detailed compositions are listed in Table 1.
TS-1 catalysts for control experiment were synthesized
from gels with Si/Ti molar ratios of 30, 50, 70, and 100
following the conventional procedures [11]. The products
were calcined to burn off organic species and further washed
with 1 M HCl solution at ambient temperature to reduce the
residual alkali cations contaminating the SDA solution of
tetrapropylammoniumhydroxide because alkali cations may
poison partially the Ti active sites [12]. Acid-washed TS-1
had Si/Ti ratios of 36, 57, 83, and 116, respectively.
9
5% AAL conversion and 99% selectivity to glycidol were
obtained [8e].
The success in AAL epoxidation encouraged us to search
for other applications of Ti-MWW in synthesizing oxy-
genated chemicals containing functional groups. Allyl gly-
cidyl ether (AGE) is an active monomer with high additional
value which is useful as a diluting agent in synthesizing
epoxy resins and alkyd resins, also used as an important sta-
bilization agent for resins and agrichemicals. AGE is mainly
produced from the dehydrohalogenation between allyl alco-
hol (AAL) and epichlorohydrin which is synthesized from a
stoichemical reaction of AAL and chlorine via the chloro-
hydrine [10]. This process lacks greenness as it coproduces
a large amount of unnecessary by-products. New synthesis
methods based on heterogeneous catalysts are required to
replace the old ones to improve the above drawback to the
AGE synthesis.
The zeolite structure was confirmed with X-ray diffrac-
tion (XRD) on an MAC Science MX-Labo diffractometer
To our best knowledge, there is still no open literature re-
porting the synthesis of AGE from the epoxidation of DAE
using titanosilicate catalysts. In this study, we have applied
Ti-MWW to the DAE epoxidation for the purpose of pro-
ducing AGE efficiently and selectively. The effects of sol-
vent and the reaction conditions on the DAE epoxidation
have been investigated by comparing Ti-MWW with con-
ventional TS-1. The issue concerning the consecutive reac-
tion involved in the DAE epoxidation has been considered,
to help identify how to produce AGE selectively against
diglycidyl ether.
(
Cu-Kα). The chemical analyses were carried out by induc-
tively coupled plasma (ICP) on a Shimadzu ICPS-8000E
atomic emission spectrometer. The coordination states of Ti
species were characterized using UV–visible spectroscopy
on a JASCO-550 spectrophotometer using Spectralon as a
reference. Specific surface area was calculated using Lang-
muir plots from the N2 adsorption isotherms measured on a
Bel Japan BEL SORP 28 SA instrument.
2.2. Catalytic reactions
The epoxidation of diallyl ether (DAE) with hydrogen
peroxide was carried out under vigorous stirring in a 20-
mL glass flask connected to a condenser. In a typical run,
2
. Experimental
1
0 mmol of DAE, 5 mL of solvent, and a certain amount
2
.1. Synthesis of titanosilicate catalysts
Ti-MWW catalysts were hydrothermally synthesized us-
of catalyst were mixed in the flask and heated to desired
temperatures under agitation. Aqueous H2O2 (30 wt%) was
then added to the mixture to start the reaction. The products
were separated by filtration and analyzed on a gas chromato-
graph (Shimadzu GC-14B) using an OV-1 capillary column
ing boric acid as a crystallization-supporting agent and
piperidine as a structure-directing agent (SDA) according

P. Wu et al. / Journal of Catalysis 228 (2004) 183–191
185
of 50 m and an FID detector. The product amounts were
calculated in reference to cycloheptanone (generally 0.2 g)
which was used as a GC internal standard. The amount of
unconverted H2O2 was determined by the standard titration
method using a 0.1 M Ce(SO4)2 solution. The epoxidation
of allyl glycidyl ether was conducted under conditions sim-
ilar to those of DAE oxidation. The products were verified
using authentic chemicals commercially available or deter-
mined on a gas chromatograph-mass spectrometer (JEOL
DATUM-JMS-AX 500).
the as-synthesized precursors. A part of framework boron
was removed simultaneously (Table 1). After further calcin-
ing the acid-treated samples, XRD patterns indicated that Ti-
MWW catalysts with a three-dimensional crystalline struc-
ture were obtained as a result of the dehydroxylation occur-
ring between the layered sheets (XRD not shown). The high
specific surface areas measured by N2 adsorption were con-
sistent with the XRD patterns to verify that all the Ti-MWW
samples were of high crystallinity (Table 1).
Acid-treated and further calcined Ti-MWW contained
only isolated Ti species in the framework as evidenced by
the predominant band at 220 nm in their UV–visible spectra
when the Si/Ti ratio was higher than 46 (Fig. 1A). However,
a small amount of anatase-like extraframework Ti species,
showing a weak and broad band at 310–330 nm, remained
in Ti-MWW with an Si/Ti ratio of 38. A same tendency was
observed for TS-1; the sample with an Si/Ti of 36 showed
a similar band due to the anatase phase (Fig. 1B), which is
simply because that the amount of Ti added meets the limi-
tation of framework incorporation [1].
3
. Results and discussion
3
.1. Physicochemical properties of Ti-MWW and TS-1
catalysts
MWW-type zeolites, especially its aluminosilicates, are
generally synthesized using an SDA of hexamethyleneimine
HMI) [9]. On the other hand, a borosilicate of the MWW
(
structure, denoted as ERB-1, is synthesized possibly with
piperidine (PI) as an SDA [13]. We have succeeded for the
first time in preparing Ti-MWW using either PI or HMI as an
SDA when a crystallization-supporting agent of boric acid
coexists with Ti species in the synthetic gel [8b]. The crystal
size of hexagonal palettes, however, differs greatly between
PI and HMI. PI, leading to a smaller crystal size (gener-
ally 0.2–0.5 µm in length and 0.05–0.1 µm in thickness) and
then a more active catalyst in comparison to HMI was thus
adopted in the preparation of Ti-MWW in this study.
Ti-containing MWW precursor free of impurity and hav-
ing equally intensive XRD patterns was obtained when the
Si/Ti molar ratio was varied in the range of ∞–20. The Ti
species added in the gel were incorporated into the precur-
sor readily (Table 1). Nevertheless, some extraframework Ti
species, presumably introduced onto the layer surface of the
lamellar precursor, always coexist with tetrahedral frame-
work Ti species, and the former was easily transformed to
the anatase phase upon calcination. A selective removal of
these catalytically harmful species was then necessary. Tak-
ing advantage of previous findings [8b], the extraframework
Ti species were efficiently extracted by an acid treatment on
3.2. Effects of reaction conditions on epoxidation of diallyl
ether
3.2.1. Influence of solvents
The epoxidation of diallyl ether shows a relatively com-
plicated product distribution as DAE contains two C=C
bonds and undergoes decomposition partially during the re-
action, and furthermore the solvolysis reactions depending
on the solvents used are also involved in the reaction. After
identifying all the products by GC MASS and using authen-
tic chemicals, the reaction pathways were clarified as de-
picted in Fig. 2. The epoxidation of DAE produces the main
product, allyl glycidyl ether, which can be successively ox-
idized to diglycidyl ether (DGE). Acid-catalyzed reactions
also take place with AGE: (1) it is hydrolyzed at the ether
group to form allyl alcohol and glycidol; (2) its oxirane ring
undergoes the solvolysis to produce ring-opening products
when water (also existing in aqueous H2O2 solution), alco-
hol, and acetone are used as solvents. The reaction paths of
DAE first hydrolyzed to AAL and then oxidized to glycidol
are also possible. Theoretically, glycidol can be further hy-
Fig. 1. UV–visible spectra of Ti-MWW (A) and TS-1 (B) both with different Si/Ti ratios.

186
P. Wu et al. / Journal of Catalysis 228 (2004) 183–191
Fig. 2. Reaction pathways in the epoxidation of diallyl ether with hydrogen peroxide.
a
Table 2
A comparison of epoxidation of DAE between Ti-MWW and TS-1
Solvent
Ti-MWW
Conv.
TS-1
Product selectivity (mol%)
b
H O (mol%)
Conv.
(mol%)
Product selectivity (mol%)
b
H O (mol%)
2 2
2
2
c
c
(mol%)
AGE
DGE
Others
Conversion
Efficiency
AGE
DGE
Others
Conversion
Efficiency
MeCN
Acetone
Water
MeOH
EtOH
39.6
38.6
25.3
17.8
12.1
18.7
71.0
73.9
52.7
76.3
79.8
79.2
28.7
25.0
40.0
4.4
5.7
19.0
0.4
1.2
7.3
19.3
14.5
1.8
99.8
98.6
78.5
42.3
38.7
46.7
95.2
92.3
86.5
81.5
60.5
93.7
19.1
42.3
6.6
40.4
33.9
23.3
77.7
69.6
60.6
69.6
71.7
88.6
15.4
22.2
13.5
20.5
15.4
6.2
6.9
8.2
25.9
9.9
12.9
5.2
42.8
74.1
38.2
89.0
76.7
44.9
48.6
70.2
18.5
54.8
50.2
52.3
Dioxane
a
DAE, 10 mmol; H O /Ti-MWW (Si/Ti = 38) = 5 mmol/50 mg; H O /TS-1 (Si/Ti = 36) = 10 mmol/200 mg; solvent, 5 mL; temperature, 333 K;
2
2
2 2
time, 0.5 h for Ti-MWW, 1.5 h for TS-1.
b
Solvolysis products, etc.
c
H O efficiency = (AGE + 2DGE + others)/H O converted × 100%.
2
2
2 2
drolyzed to glycerol. Nevertheless, the amount of glycerol
produced was negligible because of a too low concentration
of glycidol. Thus, the hydrolysis of glycidol to glycerol is
omitted in Fig. 2. Both the hydrolysis and the solvolysis are
assumably related to the acidic character of titanosilicates
derived from the silanol groups and the additional bridg-
ing hydroxyls of trivalent cations coexisting with Ti in the
framework particularly in the case of boron-containing Ti-
MWW.
Significant solvent effects are generally involved in the
titanosilicate-catalyzed reactions. The characters of solvents
may have a great influence on the intrinsic activity of Ti
species but also the product distribution. Table 2 compares
the catalytic performance of Ti-MWW with that of conven-
tional TS-1 for the DAE epoxidation in various solvents. To
compare the product selectivity at a similar DAE conversion
Both Ti-MWW and TS-1 showed a clear but different
solvent effect. The most favorable solvents for Ti-MWW
were aprotic MeCN and acetone, in which comparable DAE
conversion (38.6–39.6%), similar AGE selectivity, and ex-
tremely high efficiency for H2O2 utilization were observed.
Although the by-products derived from the solvolysis were
very low in MeCN and acetone, DGE was formed at a selec-
tivity of 25–29%. The result that Ti-MWW favored aprotic
solvent of MeCN with weak basity in the present oxidation
of functionalized alkene is similar to the solvent effect on
the epoxidation of simple alkene previously reported [8d].
On the other hand, the epoxidation of DAE was greatly re-
tarded in the protic solvents of water and alcohols, and also
in the aprotic solvent of dioxane. Water has been verified
to be similarly efficient as MeCN for the epoxidation of
ALL over Ti-MWW [8e]. The difference observed between
DAE and AAL is probably due to their different solubility
in water, since the former is almost immiscible with water
while the later is totally soluble in water. The solvent ef-
fect involving water and alcohols was once assumed to be
relative to a stabilized five-membered cyclic intermediate
species (Scheme 1), formed by the coordination and hydro-
(
at least at the most high conversion in the most suitable
solvent), different reaction conditions were adopted between
Ti-MWW and TS-1, since the former is much more active.
As indicated in Table 2, higher levels of catalyst, H2O2, and
a longer reaction time were used for TS-1.

P. Wu et al. / Journal of Catalysis 228 (2004) 183–191
187
H2O2 efficiency when choosing MeCN or acetone as a sol-
vent.
3
.2.2. Influence of Ti content
A series of Ti-MWW with various Ti contents have been
prepared (Table 1) and their catalytic performance has been
investigated in DAE epoxidation using a suitable solvent of
MeCN (Fig. 3). The DAE conversion increased reasonably
with increasing Ti content, but reached a ceiling at a Ti con-
Scheme 1.
gen bonding of a protic molecule to the Ti–hydroperoxo
complex [14]. The dimensions of this intermediate species
would increase with increasing molecular size of ROH. This
should impose a steric hindrance on the substrate molecules
and limit their diffusion and approach to the Ti active sites
within the zeolite channels. Thus, the DAE conversion de-
creased slightly in the order H2O > MeOH > EtOH as
actually observed. Besides affecting the catalytic activity,
the protic alcohols caused the opening of the oxirane ring
on the acid sites, and gave a higher selectivity for solvoly-
sis products than aprotic solvents. For example, in the case
of DAE oxidation over Ti-MWW in MeOH, products other
than AGE and DGE consisted of 80% solvolysis product of
AGE by MeOH, 15% glycerin allyl ether, and 5% allyl alco-
hol and glycidol.
In the case of TS-1, acetone and MeOH were obviously
the most effective solvents from the viewpoint of DAE con-
version. Nevertheless, both acetone and MeOH promoted the
solvolysis to a relatively high level, resulting in the oxirane
ring-opening products over 8%. Aprotic but basic MeCN,
a well-known unsuitable solvent for the oxidation of sim-
ple alkenes on TS-1, showed somewhat lower conversion in
the epoxidation of DAE. EtOH retarded also the reactivity of
TS-1, which is simply assumed to be the result of forming
bulky and sterically restricted species as shown in Scheme 1.
Nonproductive decomposition of H2O2 was generally high
on TS-1, making the utilization efficiency of H2O2 lower in
comparison to Ti-MWW.
−1
tent of 0.43 mmol g (corresponding to Si/Ti = 38), which
is probably due to the presence of part of the anatase phase
as evidenced by UV–visible spectra (Fig. 1A). Moreover, the
amount of H2O2 was almost consumed at such a high Ti con-
tent. Too low H2O2 concentration also corresponded to the
leveling off of DAE conversion. The conversion of H2O2 si-
multaneously increased with Ti content, while its efficiency
for epoxidation was maintained over 93%. Concerning the
product selectivity, the selectivity to AGE decreased grad-
ually from 95 to 71% with increasing Ti content while the
DGE selectivity increased as a result of successively con-
suming AGE. The solvolysis products, however, was always
lower than 2% independent of the Ti content. Thus, from the
viewpoint of solvolysis, this indicates that MeCN is “inert”
enough to suppress the ring opening of epoxide in compari-
son to aprotic solvents such as MeOH and EtOH.
3
.2.3. Influence of reaction temperature
An obvious temperature effect on the epoxidation of DAE
was observed when the reaction was carried out in MeCN
Fig. 4). Ti-MWW showed an adequate catalytic activity
(
even around room temperature, indicating its high activity
at low temperatures. The DAE conversion reached a maxi-
mum of ca. 40% with increasing temperature up to 333 K
where H2O2 was almost consumed. The efficiency of H2O2
slightly decreased with increasing temperature as a result of
thermal decomposition, but the value was higher than 95%.
The AGE selectivity decreased while the DGE selectivity
increased with increasing temperature due to a successive
oxidation.
Considering that the epoxidation of DAE was conducted
under much weaker reaction conditions for Ti-MWW, the
above results prove that Ti-MWW is obviously more effec-
tive than TS-1 in catalytic activity, epoxide selectivity, and
Fig. 3. Dependence of DAE conversion, H O conversion, and efficiency (A), and product selectivity (B) on the Ti content of Ti-MWW. Reaction conditions:
2
2
Ti-MWW, 50 mg; DAE, 10 mmol; H O , 5 mmol; MeCN, 5 mL; temperature, 333 K; time, 0.5 h.
2
2

188
P. Wu et al. / Journal of Catalysis 228 (2004) 183–191
Fig. 4. Dependence of DAE conversion, H O conversion, and efficiency (A), and product selectivity (B) on the reaction temperature. Reaction conditions:
2
2
Ti-MWW (Si/Ti = 38), 50 mg; DAE, 10 mmol; H O , 5 mmol; MeCN, 5 mL; time, 0.5 h.
2
2
Fig. 5. Dependence of DAE conversion (A) and AGE selectivity (B) on the reaction time. Reaction conditions: Ti-MWW (Si/Ti = 38), 10–100 mg; DAE,
1
0 mmol; H O , 5 mmol; MeCN, 5 mL; temperature, 333 K.
2 2
3
.2.4. Influence of amount of Ti-MWW catalyst and
The reusability and regeneration of Ti-MWW in the DAE
epoxidation have been investigated by starting with a larger
reaction scale using 1.5 g Ti-MWW. The used Ti-MWW
catalyst was gathered by filtration and washed with ace-
tone, and then regenerated by calcination in air at 773 K for
5 h. Part of the regenerated sample was taken for chemi-
cal analyses, while the remaining catalyst was subjected to a
repeated epoxidation of DAE at the same ratio of catalyst–
substrate–solvent. In the second run, the used Ti-MWW cat-
alyst showed about 70% conversion of the fresh one when
the recovered sample was only acetone-washed and dried at
reaction time
Fig. 5 showed the time courses for the DAE epoxidation
with different amounts of Ti-MWW in the solvent of MeCN.
Reasonably, the more the catalyst used the higher the cat-
alytic activity. The DAE conversion increased rapidly within
3
0 min and then gradually with the time. For the 100 mg
Ti-MWW, no H2O2 was detected in the reaction mixture af-
ter the reaction for 30 min, while the H2O2 conversion was
3
is mainly due to a result of using up H2O2 for the former,
while it is probably due to the deactivation of pore blacking
by heavy organic species for the latter since the Ti species
of Ti-MWW were verified to be stable enough as shown in
Section 3.3. The selectivity to AGE deceased with longer
reaction times and increasing catalyst amounts (Fig. 5B) be-
cause of the secondary oxidation of AGE.
0% for 10 mg sample. Thus, the decrease of reaction rate
3
73 K. When the used catalyst was regenerated by the cal-
cination at 773 K, both the DAE conversion and the product
selectivity were restored and maintained for five reaction–
regeneration cycles (Fig. 6A). UV–visible spectroscopy con-
firmed that the used catalyst after five reaction–regeneration
cycles showed essentially the same band at 220 nm to the
fresh one (not shown). The decrease in the catalytic activity
of acetone-washed catalyst is thus assumed to be a partial
blocking of Ti sites from the substrate molecules as a result
of heavy organic compounds of high boiling points deposit-
ing within the zeolite channels, but not due to an irreversible
deactivation such as the change of Ti species coordination
states or the loss of Ti.
3
.3. Regeneration of Ti-MWW
The stability and reusability are among the important
properties of titanosilicates in addition to their catalytic ac-
tivity and product selectivity. The stability against Ti leach-
ing, a notable problem commonly encountered in the liquid-
phase reactions, directly determines whether a titanosilicate
is really applicable.
Chemical analyses further verified the reusability of Ti-
MWW since the leaching of Ti active species was negligible
during the reaction–regeneration cycles (Fig. 6B). Never-

P. Wu et al. / Journal of Catalysis 228 (2004) 183–191
189
Fig. 6. Changes of DAE conversion and product selectivity (A), and Ti and B contents (B) with the reaction–regeneration cycles. Reaction conditions: Ti-MWW
Si/Ti = 38)/DAE/H O /MeCN = 50 mg/10 mmol/5 mmol/5 mL; temperature, 333 K; time, 0.5 h. Regeneration: calcination in air at 773 K for 5 h.
(
2
2
Fig. 7. Reaction time dependencies of DAE conversion and product yield (A), and H O conversion and efficiency (B). Reaction conditions: Ti-MWW
2
2
(
Si/Ti = 38), 100 mg; DAE, 2 mmol; H O , 4 mmol; MeCN, 5 mL; temperature, 333 K.
2
2
theless, the boron species were gradually leached out of
the framework during the reaction probably because of its
small ionic radius compared with that of Si [15]. The re-
sults are very similar to those observed previously in the
epoxidation of simple alkenes and allyl alcohol [8d,8e]. The
loss of boron may decrease the framework electronegativ-
ity to change the electronic density in the vicinity of Ti
sites, a usual phenomenon reported for Al-containing Ti-
Beta [2] and Ti-MOR [5]. This may result in an increased
hydrophobicity of the framework, enhancing the catalytic
performance of framework Ti. In fact, the AGE conversion
increased slightly after reuse (Fig. 6A).
Generally, TS-1 is believed to be stable in the liquid-
phase oxidation of simple alkenes, while Ti-Beta and Ti-
MCM-41 are claimed to lack the stability of Ti species
under hydrothermal conditions because of a large amount
of hydrophilic silanol groups on the defect sites and/or the
external surface [16,17]. Even for TS-1, its Ti species are
leached in the epoxidation of crotyl alcohol, as the triol
by-products may chelate the Ti sites cleaving the Si–O–
Ti bonds [18]. The present epoxidation of DAE also gen-
erates minor amounts of products such as AAL, glycidol,
and glycerin allyl ether, etc. The stability of Ti-MWW,
however, was hardly affected, suggesting Ti-MWW has a
greater potential of being employed as a reusable catalyst
for the DAE epoxidation unless the used catalyst is sub-
jected to the regeneration by burning off the residual organic
species.
3.4. Investigation into the consecutive reaction and control
of AGE selectivity
The epoxidation of DAE consists of many reaction path-
ways among which the consecutive oxidation of AGE to
DGE is the main side reaction (Table 2), which was con-
firmed in actual reactions (Figs. 3–5). From the viewpoint of
producing selectively AGE, which is demanded in industrial
and commercial applications, it is important to consider the
issue of consecutive oxidation and propose some ideas for
the selective production of AGE. To further confirm the issue
of consecutive oxidation, the epoxidation of DAE was car-
ried out under more extreme conditions, that is, at a higher
ratio of Ti-MWW and H2O2 to DAE. This would allow fur-
ther oxidation of DAE and AGE proceeding to a deep level.
The results are shown in Fig. 7. The DAE conversion in-
creased rapidly and reached about 65% after 90 min. The
AGE yield increased and reached a maximum of 30% at
20 min, and then gradually decreased with the reaction time.
The DGE yield, on the other hand, continually increased
with time. The results are clearly a time course pattern of
consecutive reaction, verifying that the formation of DGE is
at consumption of AGE.

190
P. Wu et al. / Journal of Catalysis 228 (2004) 183–191
Table 3
a
Epoxidation of DAE and AGE on Ti-MWW
Substrate
Catalyst
Conversion
TON
Product selectivity (mol%)
H O (mol%)
2 2
−1
b
(mg)
(mol%)
mol (mol-Ti)
AGE
DGE
Others
Conversion
Efficiency
GAE
GAE
AGE
AGE
10
100
10
8.5
35.9
1.6
213
111
35
86.1
69.0
10.5
28.0
99.9
98.9
3.5
2.4
0.1
1.1
18.7
93.9
3.4
95.4
94.6
85.5
78.6
c
n.a.
c
100
14.4
33
n.a.
31.5
a
Catalyst, Ti-MWW (Si/Ti = 38); substrate, 10 mmol; H O , 5 mmol; MeCN, 5 mL; temperature, 333 K; time, 0.5 h.
2
2
b
c
Solvolysis products, etc.
Not available.
Fig. 8. Dependence of product yields on the DAE conversion on Ti-MWW (A) and TS-1 (B) at 333 K. Representative reaction conditions: 50 mg and 0.5 h for
Ti-MWW; 200 mg and 1.5 h for TS-1.
To investigate the possibility of producing AGE selec-
tively, the DAE oxidation was compared with the AGE ox-
idation using different amounts of Ti-MWW (Table 3). The
AGE oxidation resulted in mainly DGE together with a small
amount of solvolysis by-products. Although the conversion
and turnover number (TON) were dependent on the catalyst
amount, it is obvious that the oxidation of DAE proceeds at
a much higher reaction rate than AGE oxidation. The higher
reaction rate of DAE assumably derives from its two C=C
bonds having a higher percentage reached by the Ti active
sites in comparison to AGE. The difference in hydrophilic-
ity/hydrophobicity between DAE and AGE could also be a
reason. As Ti-MWW containing hydroxyl defect sites is rel-
atively a hydrophilic catalyst [8e], the higher epoxidation
rate for DAE conversion relative to AGE conversion could
be because DAE is more hydrophobic and once it is ad-
sorbed in the catalyst the subsequent epoxidation occurs just
as quickly. Whereas for the less hydrophobic AGE the rate
is limited by adsorption onto to catalyst. This big difference
in oxidation rate between DAE and AGE would make the
selective production of AGE possible.
Fig. 8 summarizes the dependence of product yield on
the DAE conversion when the DAE epoxidation was carried
out over Ti-MWW and TS-1. All the data were gathered us-
ing the most suitable solvents, that is, in MeCN and acetone
for Ti-MWW and in MeOH and acetone for TS-1, while the
Ti content and reaction temperature were varied. It should
be noted that the amount of TS-1 was four times that of Ti-
MWW, and the reaction time was 2 h for TS-1 and 0.5 h for
Ti-MWW. The dotted line indicates the formation of AGE
at a selectivity of 100%. At a given DAE conversion, more
by-products were produced on TS-1 because it preferred sol-
vents not so “inert” to the solvolysis. Both Ti-MWW and
TS-1 showed similar profiles for the formation of AGE and
DGE; i.e., the AGE selectivity gradually decreased when
the reaction proceeded to a high level, due to the consec-
utive oxidation of AGE to DGE. The results suggest that
it is necessary to control the reaction level from the view-
point of producing AGE selectively on both Ti-MWW and
TS-1. Nevertheless, Ti-MWW is obviously superior to TS-1
because Ti-MWW produces less solvolysis products and re-
quires less harsh reaction conditions such as lower catalyst
amounts and shorter reaction times.
4. Conclusion
Ti-MWW catalyzes the epoxidation of DAE to AGE and
DGE with aqueous hydrogen peroxide more actively than
TS-1. Ti-MWW favors MeCN or acetone as a suitable sol-
vent while TS-1 prefers acetone and MeOH. The difference
in solvent effect results in less solvolysis products being
formed on Ti-MWW. Ti-MWW which is stable enough to
endure the Ti leaching proves to a regeneratable catalyst for
DAE epoxidation. The investigation under many reaction
conditions, such as Ti content, reaction time and tempera-
ture, catalyst amount, indicates that the DAE oxidation is a
typical consecutive reaction in which the reaction rate for
AGE formation is much faster than that for DGE. It is possi-
ble to produce valuable AGE selectively on both Ti-MWW

P. Wu et al. / Journal of Catalysis 228 (2004) 183–191
191
and TS-1 by controlling the reaction level carefully, but Ti-
MWW exhibits more benefits in catalytic activity using less
harsh reaction conditions.
[7] A. Corma, Chem. Rev. 97 (1997) 2373.
[8] (a) P. Wu, T. Tatsumi, T. Komatsu, T. Yashima, Chem. Lett. (2000)
774;
(
(
(
b) P. Wu, T. Tatsumi, T. Komatsu, T. Yashima, J. Phys. Chem. B 105
2001) 2897;
c) P. Wu, T. Tatsumi, Chem. Commun. (2001) 897;
Acknowledgments
(d) P. Wu, T. Tatsumi, T. Komatsu, T. Yashima, J. Catal. 202 (2001)
45;
e) P. Wu, T. Tatsumi, J. Catal. 214 (2003) 317.
9] (a) M.E. Leonowicz, J.A. Lawton, S.L. Lawton, M.K. Rubin, Sci-
ence 264 (1994) 1910;
2
(
M.H. thanks the financial support from the National Nat-
ural Science Foundation of China (Grant 20233030) and
Science and Technology Commission of Shanghai Munic-
ipality (03DJ14005).
[
(b) S.L. Lawton, M.E. Leonowicz, P.D. Partridge, P. Chu, M.K. Rubin,
Micropor. Mesopor. Mater. 23 (1998) 109.
[10] (a) H.A. Wittcoff, B.G. Reuben, in: Industrial Organic Chemicals, Wi-
ley, New York, 1996, p. 187;
(
b) Private communication with Showa Denko Co. Japan.
References
[
[
[
11] T. Tatsumi, K.A. Koyano, Y. Shimizu, Appl. Catal. A 200 (2000) 125.
12] C.B. Khouw, M.E. Davis, J. Catal. 151 (1995) 77.
13] R. Millini, G. Perego, W.O. Parker, G. Bellussi, L. Carluccio, Micro-
por. Mater. 4 (1995) 221.
[
[
1] G. Bellussi, M.S. Rigguto, Stud. Surf. Sci. Catal. 137 (2001) 911.
2] (a) M.A. Camblor, A. Corma, A. Martínez, J. Pérez-Pariente, J. Chem.
Soc. Chem. Commun. (1992) 589;
[
14] (a) G. Bellussi, A. Carati, M.G. Clerici, G. Maddinelli, R. Millini,
(
b) A. Corma, M.A. Camblor, P.A. Esteve, A. Martínez, J. Pérez-
Pariente, J. Catal. 145 (1994) 151;
c) A. Corma, P. Esteve, A. Martínez, S. Valencia, J. Catal. 152 (1995)
8.
3] (a) N. Jappar, Q.H. Xia, T. Tatsumi, J. Catal. 180 (1998) 132;
b) T. Tatsumi, N. Jappar, J. Phys. Chem. 102 (1998) 7126.
J. Catal. 133 (1992) 220;
(b) M.G. Clerici, P. Ingallina, J. Catal. 140 (1993) 71.
(
1
[
15] G.P. Heitmann, G. Dahlhoff, J.P.M. Niederer, W.F. Hölderich, J. Ca-
tal. 194 (2000) 122.
[
[
16] A. Carati, C. Flego, E. Previde Massara, R. Millini, L. Carluccio,
G. Bellussi, Micropor. Mesopor. Mater. 30 (1999) 137.
17] L.Y. Chen, G.K. Chuah, S. Jaenicke, Catal. Lett. 50 (1998) 107.
18] (a) L. Davies, P. McMorn, D. Bethell, P.C.B. Page, F. King, F.E. Han-
cock, G.J. Hutchings, Chem. Commun. (2000) 1807;
(
[4] A. Tuel, Zeolites 15 (1995) 236.
[
[
[5] (a) P. Wu, T. Komatsu, T. Yashima, J. Phys. Chem. 100 (1996) 10316;
(b) P. Wu, T. Komatsu, T. Yashima, J. Catal. 168 (1997) 400;
(c) P. Wu, T. Komatsu, T. Yashima, J. Phys. Chem. B 102 (1998) 9297.
(b) L. Davies, P. McMorn, D. Bethell, P.C.B. Page, F. King, F.E. Han-
[6] M. Diaz-Cabanas, L.A. Villaescusa, M.A. Camblor, Chem. Commun.
cock, G.J. Hutchings, Phys. Chem. Chem. Phys. 3 (2001) 632.
(2000) 761.