The positive role of cadmium in TS-1 catalyst for butadiene epoxidation

Article abstract of DOI:10.1016/j.molcata.2013.08.019

A series of Cd modified titanium silicalite 1 catalysts with different Cd content (xCd-TS-1, x = 1-15) were successfully prepared by ultrasound impregnation. Epoxidation of butadiene over these catalysts were investigated using hydrogen peroxide as oxidant, which indicated that Cd greatly improve the catalytic performance of TS-1 and the selectivity of epoxide. Various characterization methods including quantum chemical calculation were employed to explore the specific roles of Cd in promoting TS-1 catalytic activity. Theoretical calculation consistently suggested TiO bond were weakened owing to the introduction of Cd, which resulted in the structure of Cd-TS-1 becoming more relaxant. As a consequence, it is favorable to methanol solvent and H ₂O₂ interacting with the Ti active site to form five-member transition state during reaction. It was observed that catalysts modified with 1-5 wt% Cd presented both high catalytic activity and good reusability. The highest yield of 0.63 mol/L of vinyloxirane (VO) was obtained, while turnover number (TON, determined as the molar VO obtained per molar Ti atom) could reach to 1466.

Full text of DOI:10.1016/j.molcata.2013.08.019

Journal of Molecular Catalysis A: Chemical 379 (2013) 207–212
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
The positive role of cadmium in TS-1 catalyst for
butadiene epoxidation
Mei Wua,b, Huanling Song , Fang Wang , Lingjun Chou
a,∗
a
a,∗
a
State Key Laboratory for Oxo Synthesis and Selective Oxidation, Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences, Lanzhou 730000, PR
China
University of Chinese Academy of Sciences, Beijing 100049, PR China
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 17 June 2013
Received in revised form 19 August 2013
Accepted 20 August 2013
Available online 28 August 2013
A series of Cd modified titanium silicalite 1 catalysts with different Cd content (xCd-TS-1, x = 1–15) were
successfully prepared by ultrasound impregnation. Epoxidation of butadiene over these catalysts were
investigated using hydrogen peroxide as oxidant, which indicated that Cd greatly improve the catalytic
performance of TS-1 and the selectivity of epoxide. Various characterization methods including quantum
chemical calculation were employed to explore the specific roles of Cd in promoting TS-1 catalytic activity.
Theoretical calculation consistently suggested Ti O bond were weakened owing to the introduction of
Cd, which resulted in the structure of Cd-TS-1 becoming more relaxant. As a consequence, it is favorable
to methanol solvent and H2O2 interacting with the Ti active site to form five-member transition state
during reaction. It was observed that catalysts modified with 1–5 wt% Cd presented both high catalytic
activity and good reusability. The highest yield of 0.63 mol/L of vinyloxirane (VO) was obtained, while
turnover number (TON, determined as the molar VO obtained per molar Ti atom) could reach to 1466.
Keywords:
Cadmium
Epoxidation
Quantum chemical calculation
Titanium silicalite 1
Vinyloxirane
©
2013 Elsevier B.V. All rights reserved.
1
. Introduction
constraints of CH OH offers the best medium to the epoxidation of
3
BD [13].
Vinyloxirane (VO) synthesis has attracted more and more atten-
tions since its great value to produce high-volume industrial
and consumer products [1]. The direct epoxidation of butadiene
In order to further enhance catalytic reactivity, much attention
has been paid on the modification of the TS-1. Alkali and alkaline-
earth metals salts with weak basicity have been used as additives to
neutralize acid sites in the titanium silicalite, and epoxides selectiv-
ity increased markedly [9,15,16]. Arca et al. [17] pre-treated TS-1
with zinc salts, inducing a reduced Ti-OOH electrophilicity and a
higher propylene oxide selectivity towards the propylene epoxida-
tion. Also in our previous work [18,19], some incorporated metals
in the TS-1 skeletons improved the electrophilicity of active oxygen
in the five-member-ring intermediate, and significantly promoted
TS-1 catalytic activity for BD epoxidation. However, Zn treated TS-1
exhibited poorer activity for BD epoxidation although VO selectiv-
ity increasing slightly. Besides, metals V [20], Nb [21], Zr [22] and
Co [23] were also used to modify TS-1 for all kinds of oxidation
reactions.
(
BD) provides an excellent pathway for VO preparing. Effec-
tive catalysis system for BD epoxidation has been thoroughly
studied. Ag/Al O [2–4], noble metals [5–7], transition metal
2
3
oxides [8], along with all kinds of titanium silicalite materials
such as TS-1, TS-2, Ti-␤, Ti-MCM-41/48, and Ti-MWW [9–12]
are commonly-used catalysts for olefin epoxidation. There into,
TS-1 exhibits remarkable high catalysis efficiency and molecu-
lar selectivity. In the report of Zhang et al. [13], BD epoxidation
was effectively catalyzed by TS-1. Oxidant and solvent make
many differences for various catalysts. TS-1 would perform high
activity and selectivity when H O2 is applied as oxidant. More-
2
over, H O is environmentally benign owing to H O as the
2
2
2
only by-product. A widely accepted five-membered cyclic struc-
ture (a well-defined Clerici-like cycle) mechanism demonstrates
that Ti centers in TS-1 are coordinated with the protic solvent
and stabilized through hydrogen bonding in the form of Ti-
peroxide complex [9,14]. Thus the small electrophilicity and steric
In this work, we have designed kinds of TS-1 catalysts modi-
fied by another inexpensive metal Cd for BD epoxidation. Catalysts
with different Cd content were prepared by impregnation method.
There into, 5Cd-TS-1 showed both excellent activity and durable
reusability. In addition, catalytic performance of Cd-TS-1 took good
advantage over other metals modifying TS-1 catalysts for the alkene
epoxidation, in which H O2 utilization was increased 18.1% com-
2
pared with TS-1. Characterizations by XRD, XPS, BET, FT-IR, UV–vis
and quantum chemical calculation were carefully performed, aim-
ing to explore the peculiar roles of Cd in TS-1.
∗ Corresponding author. Tel.: +86 931 4968066; fax: +86 931 4968129.
E-mail addresses: songhl@licp.cas.cn (H. Song), ljchou@licp.cas.cn (L. Chou).
1
381-1169/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2013.08.019

2
08
M. Wu et al. / Journal of Molecular Catalysis A: Chemical 379 (2013) 207–212
2
. Experimental
Castep code [24] implemented in Materials Studio 6.0 package of
Accelrys. The local density approximation (LDA) of CA-PZ [25,26]
with ultrasoft pseudopotential [27] was employed to include the
exchange-correlation energy in the total energy. The Kohn–Sham
one-electron states are expanded in a plane wave basis set up to a
cutoff energy of 300 eV. The criteria for energy and maximum force
2
.1. Materials and catalysts preparation
H O (40 wt%; local vender) concentrations were determined
2
2
by iodometrictitration prior to use. Titanium silicalite 1 (TS-1,
SiO /TiO = 40) with specific surface area 463.1 m /g was from
Shanghai Novel Chemical Technology Co. Ltd. Cadmium nitrate
2
−5
convergence used are 2.0 × 10 eV/atom and 0.05 eV/Å, respec-
2
2
tively. All geometries were fully optimized without any constraints.
A cluster-like complex, isolated from the TS-1 crystal, having two
(
Cd(NO ) ·4H O, Shanghai No. 2 Reagent Factory, China) and
3
2
2
methanol (CH OH, local vendor) were all A.R. grade. All chemicals
adjacent rings with 10- and 5-tetrahedral units and inserted in a
15 A˚ cubic cells was used to investigate the interaction between
3
were used without further purification.
TS-1 catalysts modified with different content of Cd (denoted
as xCd-TS-1, where x stands for CdO mass percent contents:
CdO and Ti-site.
x = mCdO/(mCdO + m
) × 100, x is in the range of 1–15), were pre-
2.3. Catalytic evaluation
TS−1
pared via impregnation method. The typical synthesis process was
as follows: 1.00 g of TS-1 was dispersed in distilled water, followed
by adding into designed amount of Cd(NO3)2 solution under vig-
orous stirring. The mixture was treated under rest for hours and
infrared drying until it changed to dry powder. Finally it was fur-
The epoxidation of BD was performed in a 100 mL stainless-steel
autoclave with the magnetic stirrer. Typically, 0.4 g xCd-TS-1 and
designed amount of H O2 were dispersed in 25.00 mL methanol
2
solvent followed by introducing 0.15 MPa BD. BD was in excess
amount during reaction process. The reaction continued for 60 min
◦
ther dried up overnight at 120 C. The output powder was calcined
◦
◦
at 550 C for 6 h.
at 40 C. Remaining H O concentration was determined by
2
2
The regeneration of the 5Cd-TS-1 catalyst was carried out at
00 C for 4 h.
standard iodometrictitration. Chromatography–mass spectroscopy
(GC–MS 5973 equipment from Agilent Technology Company) was
used to confirm liquid products. VO was confirmed as the main
product, and main by-products were small amount of tetrahydro-2-
furanmethanol and 4-ethenyl-cyclohexene. Quantities of products
were performed on a gas chromatograph (SP-6800A GC) equipped
with a flame ionization detector and OV-1701 capillary column. VO
yields were determined relating to methanol via external standard
method based on the GC results, while concentrations of by-
products were too little to determine. VO turnover numbers (TON,
determined as the molar VO obtained per molar Ti atom, Eq. (1)),
and VO concentrations in the final reaction mixtures were utilized
to evaluate activities of different catalysts. Conversions and uti-
◦
5
2
.2. Physicochemical characterization and computational details
The X-ray powder diffraction analysis (XRD) measurements
were performed using an X’Pert Pro multipurpose diffractometer
PANalytical, Inc.) with Ni-filtered Cu K␣ radiation (0.15046 nm)
(
◦
◦
from 5.0 to 80.0 . Measurements were conducted using a voltage
◦
of 40 kV, current setting of 40 mA, step size of 0.02 , and count time
of 4 s.
◦
The N2 adsorption and desorption isotherms at −196 C were
recorded on an Autosorb-iQ analyzer (Quantachrome Instru-
◦
ments U.S.). Prior to the tests, samples were degassed at 200 C
lizations of H O2 were calculated by means of Eqs. (2) and (3),
2
for 4 h. The specific surface areas were calculated via the BET
method in the relative pressure range of 0.05–0.30; micropore
volumes were calculated using adsorption branches of nitrogen
adsorption–desorption isotherms by Saito–Foley (SF) methods.
respectively:
nVO
TON =
(1)
(2)
nTi
ꢀ
ꢁ
H -TPR was performed on a Zeton-Altamira instrument (AMI-
2
(
nH ) − (n
)
2O
2
H2O2
0
1
00) employing hydrogen as reducing agent. The samples (0.3 g)
H O conversion % = 100 ×
2
2
(
n
)
H2O2
0
were loaded in a U-shaped quartz reactor. Prior to the TPR mea-
◦
surements, samples were pretreated at 400 C for 30 min in flowing
ꢀ
ꢁ
He (50 mL/min) to remove any moisture and other impurities that
nVO
(nH2O2 )0
◦
H2O2 utilization % = 100 ×
(3)
might be present. After cooling the reactor to 20 C, a 5% H –He
2
(
50 mL/min) gas mixture was introduced. And then the catalyst was
◦
◦
heated to 900 C at a rate of 20 C/min and the hydrogen consump-
tion was measured using an AMETEK (LC-D-200 Dycor AMETEK)
mass spectrum.
In above equations, nVO stands for moles of VO yield after epoxi-
dation reaction; nTi stands for molar of Ti atom in catalyst; besides,
(nH2O2)0 and nH2O2 stand for the initial molar content and the
remaining molar content of H2O2 after reaction, respectively.
The oxidation state of Ti and Cd was characterized over a Thermo
Fisher Scientific K-Alpha X-ray photoelectron spectroscopy (XPS).
The xCd-TS-1 powder was pressed to self-supporting wafer prior
to analysis.
3. Results and discussion
Fourier transformed infrared spectra (FT-IR) of xCd-TS-1 sam-
ples were recorded on FT-IR spectrometer (Nicolet Nexus 870) with
a resolution of 4 cm and 64 scans in the region of 4000–500 cm .
3.1. Characterization results
−1
−1
3.1.1. XRD characterizations
Diffuse reflectance UV–vis spectra (DR UV–vis) was obtained on
Crystal structures of TS-1 catalysts with and without Cd were
studied by XRD (Fig. 1). All samples possess XRD diffraction pat-
terns identical to those of pure TS-1, confirming the presence of
orthorhombic MFI-type crystalline phase [28]. TS-1 samples retain
a high degree of crystalline even after Cd doping and calcination.
Within the experimental error, a painstaking analysis of xCd-TS-1
(x = 1–15) spectra do not disclose free CdO, and other possible Cd
oxidic species crystalline phases. This result indicates that CdO par-
ticles are highly dispersed on the surface of TS-1 [29]. Moreover, its
worth noting that the XRD peak intensities of TS-1 decrease with
a PE Lambda 650S spectrometer with BaSO as standard.
4
TGA measurements were rendered on a NETZSCH STA 449F3
◦
thermogravimetric analyzer from room temperature to 900 C at
◦
the rate of 10 C/min.
Chemical analysis of prepared and used catalysts was done by
atomic absorption spectroscopy (AAS) on an ARL 3520 spectrome-
ter.
Ab initio quantum chemical calculations based on the frame-
work of density functional theory (DFT) were performed using

M. Wu et al. / Journal of Molecular Catalysis A: Chemical 379 (2013) 207–212
209
Fig. 2. H2-TPR profiles for xCd-TS-1(x = 1, 3, 5, 10, 15) catalysts.
Fig. 1. XRD patterns of various xCd-TS-1 (x = 0, 1, 3, 5, 10, 15) catalysts.
3.1.4. FT-IR characterizations
The FT-IR spectra of modified samples, which were exhibited in
the Cd content increasing from 1 to 15 wt%. This may be due to the
decrease of the relative amount of TS-1 in the samples as the Cd
loading amount increasing [22].
−
1
−1
Fig. 3(a), ranged from 4000 to 500 cm . Band at 1100 cm with a
−
1
shoulder at 1220 cm is attributed to Si O Si asymmetric stretch-
−
1
ing; band at 803 cm belongs to symmetric stretching/bending of
3.1.2. N2 absorption and desorption characterizations
The physicochemical properties of xCd-TS-1 (x = 0–15) catalysts
were summarized in Table 1. The surface area and micropore vol-
ume of TS-1 decrease with Cd loading increasing. As Cd content
2
increasing from 0 to 15 wt%, they decrease from 465.2 m /g and
3
2
3
0
.177 cm /g to 341.9 m /g and 0.131 cm /g, respectively. It indi-
cates that Cd oxidic species from the decomposition of cadmium
nitrate could block the pore channels and decrease the TS-1 sur-
face area. As a result, excess Cd loading may hinder mass transfer
of raw material and product, giving rise to a poor catalytic activity
of modified catalysts.
3
.1.3. H -TPR characterizations
2
As shown in Fig. 2, H -TPR was applied to investigate the inter-
2
actions between Cd species and TS-1 support. Only one reduction
peak presents in every H -TPR curve of various Cd modified sam-
2
ples. It is assigned to the reduction of Cd oxidic species on the
surface of TS-1 from the decomposition of cadimium nitrate. The
intensity of this reduction peak enhances with the increase of Cd
loading, and reaches the maximum for 15Cd-TS-1. As suggested in
reference information [30], the top of reduction peak of free CdO
◦
should be at 465 C. The higher reduction temperature of Cd oxidic
species supported on TS-1 was observed. Moreover, the reduction
◦
◦
temperature rised from 537 C toward 650 C with the Cd content
increasing from 1 to 15 wt%. It demonstrates that the strong inter-
action exsits between Cd oxidic spices and the TS-1 supporter, and
the interaction strength was greatly related to Cd loading.
Table 1
Physicochemical properties of xCd-TS-1(x = 0, 1, 3, 5, 10, 15) obtained by BET method.
2
3
Cd loading (%)
Surface area (m /g)
Micropore volume (cm /g)
0
1
3
5
0
5
465.2
445.4
419.0
398.3
371.8
341.9
0.18
0.17
0.16
0.15
0.14
0.13
Fig. 3. (a) FT-IR spectra of TS-1 modified with different contents of Cd (x = 0, 1, 3, 5,
1
1
−
1
1
5
0, 15); (b) FT-IR wavenumber decrease around 970 cm vs Cd loading (x = 0, 1, 3,
, 10, 15).

2
10
M. Wu et al. / Journal of Molecular Catalysis A: Chemical 379 (2013) 207–212
Fig. 5. The DR UV–vis spectra of TS-1 with different contents of Cd (x = 0, 1, 3, 5, 10,
5).
Fig. 4. The Ti 2p XPS spectra of the xCd-TS-1 (x = 0, 1, 5, 10) samples.
1
SiO Si bridges; band at 552 cm−1 is assigned to Si O Si rocking
31]. These bands reveal the silica matrix is in a MFI structure, which
is in good agreement with the XRD analytical result. Band around
the introduction of Cd. No characteristic peak at 458.5 eV, which
was widely believed as the sign of anatase Ti 2p3/2 out of TS-1
framework, presented in the XPS spectra of Cd-TS-1 samples.
[
−
1
9
70 cm , which is widely recognized as the stretching mode of
the [SiO ] tetrahedral bond with Ti atoms and the fingerprint of
3.1.6. DR UV–vis characterizations
4
framework titanium [22,32,33], appears in all samples. No obvi-
ous change in band intensity is disclosed with Cd amount rising.
All samples were analyzed by DR UV–vis spectroscopy (Fig. 5),
collected in the range of 200–850 nm. The local environment of
Ti and Cd in xCd-TS-1 (x = 1–15) were studied. Unmodified TS-1
was still employed as reference sample. To our best knowledge, the
absorption peak at 227 nm is originated from the electronic trans-
fer of the pꢀ–pꢀ transitions between titanium and oxygen in the
framework titanium species [34]. It is widely believed as a power-
ful evidence for the Ti to enter into TS-1 skeleton. Absorption peak
at 308 nm was identified as Ti4+ ions in an octahedral coordina-
tion with two water molecules in the coordination sphere or small
hydrated oligomeric TiOx species [35,36]. The absence of absorp-
tion of all samples in the range of 330–350 nm, which is due either
to anatase or to relatively big TixOy structures, further confirms no
anatase or any other extra-framework TiO2 phase presents in the
Cd-TS-1 samples.
−
1
Band migration is observed around 970 cm after the content of Cd
increasing. With Cd loading increasing from 0 to 15 wt%, it migrated
−1
−1
from 970.4 cm towards 967.7 cm (Fig. 3(b)). The downshift of
the stretching vibrational frequency indicated the strength of Ti O
bonds were weakened, which may ascribed to the effect of the
doped Cd.
3.1.5. XPS characterizations
Fig. 4 displays the XPS spectra of TS-1 samples with different Cd
content. The chemical state of titanium was elaborately explored.
The Ti 2p3/2 core-level spectra of reference standard TS-1 show
characteristic binding energy at 460.0 eV, which are assigned to
the framework Ti in tetrahedral coordination. For Cd modifying
catalysts, two peaks of binding energy at 460.3 and 459.1 eV were
clearly observed. The former one can be classified to framework
Ti site in the tetrahedral configuration [6]. The later one, which
3.1.7. Quantum chemical calculation
To evaluate the alteration of geometrical and electronic struc-
ture for TS-1 resulting from the modification of Cd, a quantum
chemical calculation has been carried out. The optimized geometry
of Cd-TS-1 was shown in Fig. 6(a). Table 2 exhibited the calculated
atomic charge for Ti, Cd and selected O atoms in TS-1, Cd-TS-1 and
CdO.
posses a 1.2 eV decrease relate to the reference tetrahedral Ti 2p3/2
,
is attributed to framework Ti center with higher coordination num-
ber. It can be supposed that the additive Cd coordinates with Ti
active site via O, thus makes the Ti coordination number increase.
So to that extent, framework [Ti O] should be weakened due to
Cd modification. It was in good agreement with FT-IR results that
As can be seen in Fig. 6(a), CdO modified TS-1 via the inter-
action of O5 with Ti active site. The binding energy (defined as
−
1
the peak around 970 cm
migrates to lower wavenumber after
Table 2
Calculated atomic charge for Ti, Cd and selected O atoms in optimized geometries.
Atoms
CdO
TS-1
Cd-TS-1
CH3OH
H2O2
I
II
O1
O2
O3
O4
O5
O6
O7
O8
Ti
–
–
–
–
−0.980
−1.030
−0.990
−0.890
−0.990
−0.820
–
–
–
–
–
–
–
–
–
–
–
−0.860
−1.050
−0.870
−0.950
−0.900
−0.680
−0.510
−0.430
1.690
−0.960
−0.930
−0.880
−0.950
–
−0.660
−0.540
−0.490
1.720
–
−0.970
−0.840
−0.970
−0.70
–
–
–
–
–
–
–
–
–
−0.800
–
–
–
–
–
–
−0.560
−0.560
1.670
–
1.600
1.370
–
–
Cd
0.70
1.350

M. Wu et al. / Journal of Molecular Catalysis A: Chemical 379 (2013) 207–212
211
Table 3
◦
Effect of Cd content on the BD epoxidation (Reaction conditions: 40 C, 0.15 MPa,
0 min, H2O2 (40 wt%) 0.71 mol/L, xCd-TS-1 0.40 g, CH3OH 25.00 mL.).
6
Cd loading (%)
YVO (mol/L)a
CH2
O
(%)b
UH2
O
(%)c
2
TON
2
0
1
3
5
0
5
0.53
0.61
0.62
0.63
0.55
0.50
86.0
89.7
86.0
80.1
74.8
69.7
75.2
85.3
86.6
88.8
77.1
71.5
1238
1417
1447
1466
1271
1173
1
1
a
b
c
VO yield.
H2O2 conversion.
H2O2 utilization.
the energy difference between Cd-TS-1 and the free CdO, TS-1) for
CdO on TS-1 is as high as 4.92 eV, which combined with the short
Ti-O5 distance (1.792 A˚ ) indicating O5 has coordinated with Ti. It
coincided with the XPS results, in which the Ti 2p orbital binding
energy decreased from 460.3 eV to 459.1 eV, illustrating Ti coordi-
nation number increased. After Cd modification, the length of four
Ti O bonds were found lengthened to 1.820–2.001 A˚ from 1.776
to 1.795 A˚ . When comparing the charge distributions on Ti and O
atoms in TS-1 and Cd-TS-1, it is observed all of them got electrons
after the introduction of CdO, which made the repulsion interac-
tion between Ti and O atoms increase. Both of the lengthened Ti O
bonds and the reduced positive charge on Ti and O atoms suggested
the Ti O bonds were weakened and a more relaxed structure of Cd-
TS-1 was formed, which were in good agreement with the results
of FT-IR spectra. As a result, more chances are supposed to be pro-
vided for CH OH and H O to approach the Ti active site and hence
3
2
2
facilitate the formation of five-membered cyclic intermediate.
To confirm this conjecture and get more insight into the cataly-
sis mechanism, a further quantum chemical calculation was carried
out to investigate the interaction among Cd-TS-1, CH OH, and
3
H O , denoted as complex I in Fig. 6(b). For comparison, the inter-
2
2
action of CH OH and H O with TS-1 has also been studied, denoted
3
2
2
as complex II in Fig. 6(c). As the geometries shown in Fig. 6(b) and
c), it is obviously found that a five-membered cyclic complex was
formed between Ti, H O2 and CH OH. Moreover, the Ti-O6 dis-
(
2
3
tance between Ti and methanol, and Ti-O8 distance between Ti and
hydrogen peroxide, shifted to much shorter ones with the presence
of CdO in TS-1, indicating the interaction of CH OH and H O with
3
2
2
TS-1 became much stronger after the modification of CdO. Mulliken
population analysis about complex I (see Table 2) demonstrated the
charges on O7 and O8 increased from both −0.560 in free H O to
2
2
−
0.510 and −0.430, respectively, where O8 is more electron defi-
cient with respect to O7 and was supposed to react with C C of BD.
When compared complex I with II where TS-1 is not modified by
CdO, it is observed that charges on O7 and O8 of I are more positive,
implying the electrophility of O atoms in H O increases when TS-1
2
2
is modified by CdO, which is favorable to its electrophilic attack on
the double bond of BD.
3.2. Effect of Cd loading on BD epoxidation
Table 3 depicted the results of the BD epoxidation over xCd-TS-
1
. With the increase of the Cd content, VO yield, H O2 utilization
2
and TON effectively improved and reached maximum at 5% Cd
loading, although the H O conversion gradually decreased. How-
ever, higher content of Cd up to 15% resulted in some decline in
Fig. 6. (a) Optimized geometry of Cd-TS-1; (b) optimized geometry of five-
membered cyclic intermediate composited of CH3OH, H2O2 and CdO-TS-1; (c)
optimized geometry of five-membered cyclic intermediate composited of CH3OH,
H2O2 and TS-1.
2
2
VO yield, conversion and utilization of H O2 and TON. Compared
2
with unpromoted TS-1 catalyst, both of H O2 utilization and TON
2
could increase 18% for Cd-TS-1 catalyst. The results were supe-
rior to Ni-TS-1 catalyst in which H O2 utilization enhanced 15%
2
relative to TS-1. Why Cd has this prominent feature? According
to experimental and theoretical results above, the introduction of

2
12
M. Wu et al. / Journal of Molecular Catalysis A: Chemical 379 (2013) 207–212
4. Conclusions
Cd modified TS-1 catalysts were employed to catalyze buta-
diene epoxidation for vinyloxirane. TS-1 catalytic activities were
significantly promoted with 1–5 wt% content of Cd. The special
modification effect was responsible for the coordination of addi-
tive Cd with Ti center via O atom. It leaded to the formation of a
more relaxed Cd-TS-1. As a result, more chances were provided for
CH OH and H O to approach the Ti active site. It had been testi-
3
2
2
fied that O of H O close to Ti in five-membered intermediate was
2
2
activated with promoted eletrophilicity by Cd modification. There-
fore, double bond electrophilic epoxidation of butadiene would be
facilely. 5Cd-TS-1, which presented promising reusability and high
catalytic activity, was turned out to be an ideal candidate catalyst
for butadiene epoxidation.
◦
Acknowledgement
Fig. 7. Reusability tests of 5Cd-TS-1 catalyst (Reaction conditions: 40 C; 0.15 MPa;
0 min; H2O2 (40 wt%) 0.71 mol/L; 5Cd-TS-1 0.40 g; CH3OH 25.00 mL.).
6
The authors sincerely acknowledge the financial support from
the National Natural Science of China (Grant No. 21133011,
Cd into TS-1 weakened Ti O bonds, which were facilely attacked
by CH OH and H O . As a result, it is propitious to form the five-
2
1203218).
3
2
2
membered reaction intermediate. Moreover, electrophility of H O
2
2
Appendix A. Supplementary data
in this intermediate was increased. By this token, active O was
favorable to be ectrophilic attacked by the double bond of BD. In a
word, the roles of Cd coordinating with Ti via oxygen atom played
in Ti active site could effectively promote the catalytic performance
of TS-1 for BD epoxidation. This positive effect strengthened with
Cd loading rising. However, as suggested in the XRD results, the rel-
ative TS-1 content in Cd-TS-1 decreased with Cd loading increasing.
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.molcata.2013.
0
8.019.
References
Also in the N adsorption and desorption results, excess content of
2
[
[
1] E. Doskocil, G. Mueller, J. Catal. 234 (2005) 143–150.
2] J.W. Medlin, M.A. Barteau, J.M. Vohs, J. Mol. Catal. A: Chem. 163 (2000)
1
Cd would block TS-1 pore channels and decrease the TS-1 surface
area. Hence, when Cd content was in excess of 5 wt%, TS-1 activity
would be suppressed.
29–145.
[3] H.L. Song, G. Chen, S.B. Li, et al., React. Kinet. Catal. Lett. 85 (2005) 45–49.
[
[
[
[
[
[
4] J. Monnier, J. Catal. 226 (2004) 321–333.
5] J.J. Bravo-Suarez, K.K. Bando, T. Fujitani, et al., J. Catal. 257 (2008) 32–42.
6] C. Shi, B. Zhu, M. Lin, et al., Eur. J. Inorg. Chem. 2009 (2009) 4433–4440.
7] V.R. Choudhary, D.K. Dumbre, Top. Catal. 52 (2009) 1677–1687.
8] V. Choudhary, R. Jha, P. Jana, Catal. Commun. 10 (2008) 205–207.
9] M.G. Clerici, P. Ingallina, J. Catal. 140 (1993) 71–83.
3.3. Deactivation and regeneration of 5Cd-TS-1 catalyst
The reusability of 5Cd-TS-1 catalyst was further tested. As Fig. 7
showed, the catalyst kept original activity during four runs: H O2
2
[
10] S. Krijnen, P. Sánchez, B.T.F. Jakobs, et al., Microporous Mesoporous Mater. 31
(1999) 163–173.
11] L. Wang, Y. Liu, H. Zhang, et al., Chin. J. Catal. 27 (2006) 656–658.
12] A. Wroblewska, A. Fajdek, J. Wajzberg, et al., J. Adv. Oxid Technol. 11 (2008)
conversions and VO yields maintained above 80% and 0.6 mol/L,
[
[
respectively. However, VO yield and H O utilization decreased in
2
2
the fifth repeated reaction. In order to clarify this phenomenon, the
468–476.
fresh and used catalysts were characterized by TGA, N adsorption
[13] X.M. Zhang, Z.R. Zhang, J.S. Suo, et al., Catal. Lett. 66 (2000) 175–179.
[14] X. Liu, X. Wang, X. Guo, et al., Catal. Today 93–95 (2004) 505–509.
15] G. Li, X.S. Wang, H.S. Yan, et al., Appl. Catal., A: Gen. 218 (2001) 31–38.
16] C.V. Rode, U.N. Nehete, M.K. Dongare, Catal. Commun. 4 (2003) 365–369.
17] V. Arca, A. Boscolo Boscoletto, N. Fracasso, et al., J. Mol. Catal. A: Chem. 243
(2006) 264–277.
[18] M. Wu, L. Chou, H. Song, Catal. Lett. 142 (2012) 627–636.
19] M. Wu, L. Chou, H. Song, Chin. J. Catal. 34 (2013) 25–33.
20] S. Chen, S. Wang, X. Ma, et al., Chem. Commun. 47 (2011) 9345–9347.
[21] D. Prasetyoko, Z. Ramli, S. Endud, et al., Mater. Chem. Phys. 93 (2005)
43–449.
22] D. Prasetyoko, Z. Ramli, S. Endud, et al., J. Mol. Catal. A: Chem. 241 (2005)
18–125.
2
and desorption and AAS. TGA patterns of fresh or used 5Cd-TS-1
catalysts were shown in Fig. S1. No exothermic peak appeared for
the fresh catalyst as Fig. S1(a) shown. While after 6 runs reaction, a
small exothermic peak (Fig. S1(b)) accompany with the weight loss
[
[
[
◦
of 6.79% appeared around 420 C. It implied that organic compound,
[
[
which could deposit in the pore channel and overlay Ti active sites,
was absorbed by TS-1 supporter during reaction. It was confirmed
by the BET analysis results (Table S1), in which the surface area of
4
[
2
2
5
Cd-TS-1 catalyst dropped from 398.3 m /g to 319.2 m /g after six
1
repeated reaction, and also the micropore volume decreased from
[23] J. Jiang, R. Li, H.L. Wang, et al., Catal. Lett. 120 (2008) 221–228.
[24] M.D. Segall, P.J.D. Lindan, M.J. Probert, et al., J. Phys.-Condens. Matter 14 (2002)
3
3
0.152 cm /g to 0.126 cm /g. With regeneration runs increasing, the
2
717–2744.
25] D.M. Ceperley, B.J. Alder, Phys. Rev. Lett. 45 (1980) 566–569.
[26] J.P. Perdew, A. Zunger, Phys. Rev. B 23 (1981) 5048–5079.
residua of organic compound will be more and more difficult to
clear up. Deposited organic compound would reduce the mass
transfer efficiency for the reaction system, inducing the deactiva-
tion of the catalyst. In addition, Cd and Ti contents of the fresh
and used 5Cd-TS-1 catalysts were compared by AAS. As we can see
from Table S1, Cd content decreased from 3.39% in fresh catalyst
to 2.47% in used one, and 0.98 ppm of Cd was detected in the sixth
reacting solution. It suggested tiny Cd leached during the repeated
reactions. Whereas, the content of Ti, with lower than 0.5 ppm in
the sixth reacting solution, has no clear change after six repeated
reaction. It demonstrated that Ti loss was not the main reason for
catalyst deactivation.
[
[
[
[
27] D. Vanderbilt, Phys. Rev. B 41 (1990) 7892–7895.
28] T. Armaroli, F. Milella, B. Notari, et al., Top. Catal. 15 (2001) 63–71.
29] Q.H. Tang, Q.H. Zhang, H.L. Wu, et al., J. Catal. 230 (2005) 384–397.
[30] B.M. Abu-Zied, A.M. El-Awad, J. Mol. Catal. A: Chem. 176 (2001) 227–246.
[31] M. Capel-Sanchez, Appl. Catal. A: Gen. 246 (2003) 69–77.
[
32] S. Bordiga, A. Damin, F. Bonino, et al., J. Phys. Chem.
892–9905.
B 106 (2002)
9
[
33] G. Ricchiardi, A. Damin, S. Bordiga, et al., J. Am. Chem. Soc. 123 (2001)
11409–11419.
[
[
[
34] J. Zhuang, D. Ma, Z. Yan, et al., Appl. Catal. A: Gen. 258 (2004) 1–6.
35] X.S. Wang, X.W. Guo, G. Li, Catal. Today 74 (2002) 65–75.
36] H. Li, Q. Lei, X.M. Zhang, et al., Microporous Mesoporous Mater. 147 (2012)
110–116.