A molecular insight into propylene epoxidation on Cu/SiO2 catalysts using O2 as oxidant

Article abstract of DOI:10.1016/j.jcat.2009.09.017

We investigated the catalytic performances of Cu/SiO₂ with various promoters on the epoxidation of C₃H₆ by O₂. Propylene oxide (PO) is the only product at reaction temperatures below 500 K and the rate of PO for

Full text of DOI:10.1016/j.jcat.2009.09.017

Journal of Catalysis 268 (2009) 165–174
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
A molecular insight into propylene epoxidation on Cu/SiO2 catalysts
using O2 as oxidant
*
Weiguang Su, Shouguo Wang, Pinliang Ying, Zhaochi Feng, Can Li
State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, P.O. Box 110, Dalian 116023, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 16 June 2009
Revised 20 September 2009
Accepted 21 September 2009
Available online 21 October 2009
We investigated the catalytic performances of Cu/SiO2 with various promoters on the epoxidation of C3H6
by O2. Propylene oxide (PO) is the only product at reaction temperatures below 500 K and the rate of PO
formation can reach 2.5 mmol gÀ1 hÀ1 over the potassium acetate (KAc)-Cu/SiO2 catalyst. IR studies show
that both Cu0 and Cu+ in Cu/SiO2 can activate the C@C bond of C3H6, and
p
donation of C3H6 to Cu0 or Cu+
also results in a red shift of m
CO with respect to CO adsorbed on Cu0 or Cu+ ‘‘solo”. However, C3H6 is only
weakly adsorbed on Cu2+. Both Cu0 and Cu+ species show C3H6 epoxidation activity. It is proposed that the
isolated Cu species and small CuOx clusters with low valence state (Cu0 or Cu+) modified by KAc are the
main active sites for C3H6 epoxidation on Cu/SiO2 catalyst with O2 as the oxidant.
Ó 2009 Elsevier Inc. All rights reserved.
Keywords:
C3H6 epoxidation
Cu/SiO2
Propylene oxide
IR spectroscopy
Cu0 or Cu+ species
Potassium acetate (KAc) modification
1. Introduction
of COx [19]. Thus, the epoxidation of C3H6 using O2 remains one
of the most challenging goals in catalysis field.
Propylene oxide (PO) is one of the important building blocks in
chemical industry. Until now, the major conventional manufactur-
ing methods of PO are the chlorohydrin process and the Halcon
process. The chlorohydrin process is being phased out because of
environmental pollution, while the latter has the byproduct limita-
tion [1]. Therefore, it becomes urgent to establish a direct and
green epoxidation route for PO production. Many studies focused
on producing PO with novel methods, for example, the TS-1/H2O2
system [2,3] and in situ-generated hydrogen peroxide systems such
as gold supported on titania [4,5], as well as using N2O [6–8] as the
oxidant. However, the high cost of these oxidants or the low effi-
ciency of H2 utilization [5] limits the commercialization of these
processes. The most ideal process is the epoxidation of propylene
with molecular oxygen as the oxidant. The epoxidation of ethylene
by O2 has been commercialized for several decades using Ag-based
catalysts, but the epoxidation of propylene by O2 is not successful
[9]. Many research efforts have also been paid to Ag-based and Cu-
based catalysts in propylene epoxidation by O2, but PO selectivity
can hardly exceed 50% even at very low propylene conversion
[10–18]. The main reason for the low PO selectivity is that the
allylic hydrogen atoms react more easily with the nucleophilic
oxygen species resulting in the byproducts and final production
Through a series of surface chemistry studies on Ag and Cu
single-crystal surfaces under ultrahigh vacuum conditions, Cropley
et al. [20,21] and Lambert et al. [22] have suggested that Cu is more
selective than Ag for the epoxidation of higher alkenes with allylic
C–H bonds by preadsorbed oxygen atoms. The lower basicity of
the oxygen atoms on metallic Cu may favor its epoxidation selectiv-
ity [23]. On the other hand, Monnier and Hartley [24] have pointed
out that metallic Cu (Cu0) is difficult to function as a true epoxida-
tion catalyst because it may be readily oxidized under reaction con-
ditions into Cu2O or CuO, which is believed to be nonselective. In
fact, only very few studies to date have reported the epoxidation
of C3H6 by O2 with Cu-based samples as true catalysts [17,18,25–
28]. In our previous work, Lu et al. [17] reported a PO selectivity
of 43% at 0.19% C3H6 conversion over a NaCl-modified VCe1ÀxCux
oxide catalyst, with
a
maximum PO formation rate of
0.17 mmol gÀ1 hÀ1. NaCl-modified Cu/SiO2 showed a PO selectivity
of 44% at 0.16% C3H6 conversion and a maximum PO formation rate
of 0.19 mmol gÀ1 hÀ1 [18]. Vaughan et al. [25] showed that Cu/SiO2
prepared by a microemulsion technique without NaCl additive
could obtain 53% PO selectivity at 0.25% C3H6 conversion and a PO
formation rate of 0.014 mmol gÀ1 hÀ1 under their reaction condi-
tions. Recently, Chu et al. [26] have reported a halogen-free K+-mod-
ified CuOx/SBA-15 prepared by an impregnation method and Zhu
et al. [27] have reported a K+-modified CuOx/SiO2 prepared by a
sol–gel method, both of them could catalyze C3H6 epoxidation effi-
ciently especially under O2-rich conditions (PO2/PC3H6 = 98.8/2.5).
* Corresponding author. Fax: +86 411 84694447.
E-mail address: canli@dicp.ac.cn (C. Li).
URL: http://www.canli.dicp.ac.cn (C. Li).
0021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.09.017

166
W. Su et al. / Journal of Catalysis 268 (2009) 165–174
The decline of PO selectivity with increasing C3H6 conversion was
much slow compared with other reported Cu-based catalysts. They
obtained a PO selectivity of 55% at 1.4% C3H6 conversion and a PO
formation rate of 2.2 mmol gÀ1 hÀ1 (TOF, ca. 18 hÀ1). In addition,
they proposed that Cu+ could function as the active site for the
epoxidation of C3H6 by O2 [27].
The Cu/SiO2 samples were pressed into self-supporting wafers
(ca. 10 mg/cm2) and mounted inside an IR cell with two CaF2 win-
dows for FT-IR spectroscopy. IR experiments were then performed
as follows:
(1) CO and C3H6 coadsorption on Cu+. Prior to the adsorption
measurements, the Cu/SiO2 samples were activated by evac-
uation at 673 K for 180 min, subsequently cooled to the
desired temperature. Previous literature results show that
evacuation at high temperature for Cu-containing samples
can result in the autoreduction of most Cu2+ to Cu+ [31].
(2) CO and C3H6 coadsorption on Cu0. The Cu/SiO2 samples in
the IR cell were reduced by H2 at 673 K for 120 min, and
evacuated at the same temperature for 90 min. After heat-
ing, the sample was cooled to room temperature (ca.
300 K), and the gas dosing experiments were carried out.
(3) NO and C3H6 coadsorption on Cu2+. The Cu/SiO2 samples
were heated to 673 K in O2 for 180 min, subsequently cooled
to room temperature (ca. 300 K) in O2. The O2 was evacuated
at room temperature for several minutes, and the adsorption
measurements were done.
Until now, although much effort has been made in the C3H6
epoxidation by O2 over Cu-based catalysts, there are many discrep-
ancies on the active sites for C3H6 epoxidation, and the effect of the
valence state and the size of Cu species on the catalytic perfor-
mances are not clear. In the present work, the catalytic perfor-
mances of Cu/SiO2 with different promoters for C3H6 epoxidation
by O2 were studied using the fixed bed catalytic reactor. The
adsorption and activation of C3H6 on Cu/SiO2 were examined by
IR spectroscopy using CO and NO as molecular probes. To the best
of our knowledge, this is the first time to correlate the results of
C3H6 epoxidation reaction with those of IR spectra of adsorbed
C3H6 on Cu/SiO2 catalysts. We found that C3H6 is strongly adsorbed
on Cu0 and Cu+ via its C@C double bond, and both Cu0 and Cu+ can
catalyze the epoxidation of C3H6 to form PO. Promoters especially
KAc can largely enhance PO selectivity and yield, while the adsorp-
tion of C3H6 on Cu species is nearly independent of the promoters
investigated in this work. Finally, the size effect of CuO nanoparti-
cles on the catalytic performances of C3H6 epoxidation was also
studied, and the catalyst with small CuO nanoparticles (ꢀ2.9 nm)
is found to be most beneficial to PO formation.
All infrared spectra were collected with a resolution of 4 cmÀ1
and 64 scans by a Fourier transform infrared spectrometer (Nicolet
NEXUS 470) with an MCT detector. All the spectra shown here
were in the absorbance mode, and their backgrounds were re-
corded before admitting the adsorbed gas under corresponding
experimental conditions of the spectra.
2. Experimental
2.1. Catalyst preparation
2.3. Catalyst testing
Cu/SiO2 catalyst (ꢁ9 wt%) was prepared by homogeneous depo-
sition-precipitation method [29,30]. The Cu/SiO2 catalyst was pre-
pared by a suspension of silica in water containing the appropriate
amounts of aqueous Cu(NO3)2 and urea, the molar ratio of
Cu(NO3)2/urea is 1/10. The initial pH of the well-stirred suspension
was adjusted to 2 by addition of nitric acid. The suspension was
subsequently heated to 363 K and kept at this temperature for
40 h. At 363 K the hydrolysis of urea results in a homogeneous
pH-rise causing precipitation of the copper ions. After completion
of the precipitation at a pH of about 7, the precursor was obtained
by filtration, and washed with distilled water. After drying for 60 h
at 393 K, the powdery sample was calcined at 723 K in air for 10 h
to obtain the Cu/SiO2. The promoter-modified Cu/SiO2 samples
were prepared by impregnation of the calcined Cu/SiO2 with an
aqueous solution of promoter followed by drying and calcination
at 723 K for 4 h. These samples are named as (promoter)-Cu/SiO2
(the molar ratio of promoter to Cu is about 0.4).
The catalytic epoxidation of propylene was carried out in a mic-
roreactor system under atmospheric pressure. Oxygen was used as
oxidant. Before the reaction, the catalyst was pretreated with dif-
ferent gas (N2, H2, or O2) for 60 min at 673 K. Reaction products
were analyzed with online GC analysis. Two GCs were used: a
GC920 with an FID detector and AT-WAX capillary column
(50 m  0.32 mm i.d.  0.30
lm) and a GC7890II with a TCD detec-
tor and Porapak Q-packed column, both the AT-WAX and the Pora-
pak Q with 333 K as initiate temperature. The method of carbon
balance was used.
3. Results and discussion
3.1. UV–Vis, XRD, and TEM characterization of Cu/SiO2 catalysts
Fig. 1 shows the UV–Vis diffuse reflectance spectra of the Cu/
SiO2 samples. The SiO2 support shows a strong absorption band at
230 nm, this band may be due to the charge transfer of ligand to
minim impurity, such as Fe3+ in SiO2 [32]. For Cu/SiO2 catalyst,
several new bands appeared at 254, 302, and 760 nm are assigned
to the charge-transfer transitions of the ligand O2À to isolated
metal center Cu2+, and small CuO clusters, and the d–d transition
of CuO particles, respectively [33–35]. When the Cu/SiO2 sample
was modified with potassium acetate (KAc) or sodium acetate
(NaAc), the feature of the UV–Vis spectra is almost the same, sug-
gesting that the electronic absorption properties of Cu species is
hardly changed after the modification with K+ or Na+. However,
when the Cu/SiO2 sample was modified with KCl or NaCl, the
intensity of the band at 760 nm is greatly enhanced. The UV–
Vis spectra are very similar to those of the Cu/SiO2 (IMP573) sam-
ple which is expected to have large CuO particles. These results
suggest that the Cu species is agglomerated to form massive
CuO particles after Cu/SiO2 was modified with KCl or NaCl. Previ-
ous literature [36] showed that ClÀ could result in the agglomer-
ation of the Au species during the calcination process, so the
For comparison, a Cu/SiO2 sample with the same loading of Cu
was prepared by incipient wetness impregnation of the silica sup-
port with aqueous Cu(NO3)2 followed by drying in air (12 h, 393 K)
and air calcination (5 h, 573 K), this sample is identified as Cu/SiO2
(IMP573).
2.2. Catalyst characterization
Powder X-ray diffraction patterns were obtained on a Rigaku
MiniFlex diffractometer with a Cu Ka radiation source. Diffraction
patterns were collected from 5° to 90° at a speed of 5°/min. TEM
was taken on a JEM-2011 TEM for estimating the particle size
and morphology. UV–Vis diffuse reflectance spectra were recorded
on a JASCO V-550 UV–Vis spectrophotometer.
XPS spectra were obtained in a VG ESCALAB MK2 spectrometer,
and base pressure is 2 Â 10À8 Pa. Al K
a radiation was used as exci-
tation source and spectra were energy calibrated by taking Si2p
peak at 103.4 eV (BE).

W. Su et al. / Journal of Catalysis 268 (2009) 165–174
167
CuO is still about 3 nm. TEM observations show that the CuO spe-
cies is agglomerated from ꢁ3 to ꢁ15 nm after the modification by
KCl or NaCl (Fig. 3d and e). Accordingly, the results of UV–Vis spec-
tra, XRD patterns, and TEM images are well consistent with each
other.
254
302
757
f
g
e
3.2. Catalytic performances of Cu/SiO2 catalysts for C3H6 epoxidation
reaction
230
a
Fig. 4 displays the catalytic performances of C3H6 epoxidation
by O2 over the Cu/SiO2 catalysts. Before the reaction, the Cu/SiO2
catalysts were pretreated in N2 at 673 K. The results of XPS spectra
show that surface Cu+ species is nearly 90% for each catalyst under
this pretreatment (Supplementary material, Table S1 and Fig. S1).
Compared with Cu/SiO2, (KAc)–Cu/SiO2 and (NaAc)–Cu/SiO2 cata-
lysts show much higher PO selectivity. Among all promoters used
in the catalysts, the catalysts modified with KAc and NaAc show
the highest PO selectivity (100%) at the initial stage of C3H6 selec-
tive oxidation reaction. For the Cu/SiO2 catalyst, PO selectivity of
22% can be obtained only at C3H6 conversion of 0.64%. With the in-
crease of C3H6 conversion to >1% by raising the reaction tempera-
tures, the PO selectivity is dramatically decreased to <5%. When
NaCl was used as a promoter, PO selectivity can be elevated to
62% despite the lower C3H6 conversion. From the above results,
ClÀ ions can cause the agglomeration of the Cu species, and the de-
crease of the concentration of Cu+ may result in the lower C3H6
conversion. The reason for that the PO selectivity is decreased after
the modification by KCl compared with Cu/SiO2 catalyst is not clear
at the moment. When the Cu/SiO2 sample was modified by KAc or
NaAc, an interesting phenomenon is observed. PO is the only prod-
uct at lower C3H6 conversion. Thus, KAc or NaAc modification
switches the formation of acrolein to that of PO. With increasing
C3H6 conversion by elevating reaction temperatures, the PO selec-
tivity decreases monotonously. However, 40% PO selectivity can
still be sustained with C3H6 conversion of 1% on (KAc)–Cu/SiO2.
This result is comparable to the best results reported on Cu-based
catalysts for C3H6 epoxidation by O2. Furthermore, the decrease of
PO selectivity for (KAc)–Cu/SiO2 is obviously slower than that for
(NaAc)–Cu/SiO2. From Fig. 4, the effect of the promoters on both
PO selectivity and C3H6 conversion increases in the order:
KCl < NaCl < NaAc < KAc.
Fig. 5 summarizes C3H6 conversion and PO selectivity as a func-
tion of reaction temperature over the Cu/SiO2 catalysts which were
reduced by H2 at 673 K before reaction. Only Cu0 species is present
on catalyst surfaces according to the results of XPS spectra (Supple-
mentary material, Table S2 and Fig. S2). For Cu/SiO2, (KCl)–Cu/SiO2,
and (NaCl)–Cu/SiO2 catalysts, the highest PO selectivity hardly ex-
ceeds 50%, and more than 30% PO selectivity can be obtained at
C3H6 conversion <1%. The increase of C3H6 conversion to >1%
remarkably decreases PO selectivity to <10%. In the cases of
(KAc)–Cu/SiO2 and (NaAc)–Cu/SiO2 catalysts, PO selectivity is sig-
nificantly improved, and 100% PO selectivity is obtained at ꢁ0.2%
C3H6 conversion. These results are much better than those on Cu-
based catalysts reported previously [17,18,25,26]. The promoters
such as KAc and NaAc can greatly enhance the selectivity to PO.
For (KAc)–Cu/SiO2 and (NaAc)–Cu/SiO2 catalysts, PO selectivity
also decreases gradually with the increase of C3H6 conversion by
raising the reaction temperatures. Acrolein and COx are the main
oxidation products at higher reaction temperatures. It is worthy
to mention that 61.1% PO selectivity can still be sustained at
1.35% C3H6 conversion over (KAc)–Cu/SiO2 catalyst. It is clear that
(KAc)–Cu/SiO2 catalyst exhibits the highest PO selectivity at the
same level of C3H6 conversion among all the catalysts tested.
The above results show that irrespective of the pretreatment
atmosphere (in N2 or H2), KAc is the best modifier for PO formation
among all the modifiers investigated in this work. Comparing the
d
b
c
200 300 400 500 600 700 800 900
Wavelength(nm)
Fig. 1. UV–Vis diffuse reflectance spectra of the Cu/SiO2 samples. (a) SiO2, (b) Cu/
SiO2, (c) (KAc)–Cu/SiO2, (d) (NaAc)–Cu/SiO2, (e) (KCl)–Cu/SiO2, (f) (NaCl)–Cu/SiO2,
(g) Cu/SiO2 (IMP573).
agglomeration of CuO particles may be induced by the presence
of ClÀ ions.
The XRD patterns of Cu/SiO2 samples are displayed in Fig. 2. The
XRD pattern of Cu/SiO2 only shows a broad and weak peak cen-
tered at 22.5°, which is due to the diffraction of amorphous SiO2
powder. This indicates that CuO species is highly dispersed on
SiO2 support (the color of Cu/SiO2 is green). After a modification
of Cu/SiO2 with KAc or NaAc, the XRD patterns are still unaltered.
However, when Cu/SiO2 was modified with KCl or NaCl, except
for crystalline KCl (2h = 28.3° and 40.5°, JCPDS Number 75-0296)
or NaCl (31.7° and 45.4°, JCPDS Number 78-0751), sharp peaks at
2h = 35.3°, 38.4°, and 48.6° are observed. All these peaks are attrib-
uted to crystalline CuO (JCPDS Number 80-1917). The XRD pattern
of Cu/SiO2 (IMP573) also shows evident diffraction peaks of crys-
talline CuO. The results of UV–Vis spectra and XRD patterns of
Cu/SiO2 samples are in good agreement with each other. Both K+
and Na+ ions have little effect on the crystalline size of CuO, while
ClÀ ions can result in the agglomeration of the Cu species, forming
bulk CuO.
Fig. 3 shows the TEM images of Cu/SiO2 samples. For Cu/SiO2
sample, the mean size of CuO nanoparticles is about 2.9 nm, which
is much smaller than that of 5 wt% CuOx–SiO2 (17 nm) prepared by
the sol–gel method reported by Zhu et al. [27]. After a modification
with K+ or Na+ ions, the TEM images reveal that the mean size of
Cu/SiO2(IMP573)
(NaCl)-Cu/SiO2
(KCl)-Cu/SiO2
(NaAc)-Cu/SiO2
(KAc)-Cu/SiO2
Cu/SiO2
SiO2
15 20 25 30 35 40 45 50 55 60 65 70 75 80
2θ (o)
Fig. 2. XRD patterns of Cu/SiO2 samples with and without promoter modification.

168
W. Su et al. / Journal of Catalysis 268 (2009) 165–174
Fig. 3. TEM images of the Cu/SiO2 samples. (a) Cu/SiO2, (b) (KAc)–Cu/SiO2, (c) (NaAc)–Cu/SiO2, (d) (KCl)–Cu/SiO2, and (e) (NaCl)–Cu/SiO2.
catalytic performances for C3H6 epoxidation over (KAc)–Cu/SiO2
catalyst treated in N2 and (KAc)–Cu/SiO2 reduced in H2, PO selec-
tivity for both catalysts is higher at lower C3H6 conversion. How-
ever, with the increase of C3H6 conversion, the decrease of PO
selectivity for the catalyst with H2 reduction is much slower than
that for that with N2 pretreatment. It seems that Cu0 species is
more favorable for PO formation than Cu+. Namely, the reduced
Cu species is more active for the epoxidation of C3H6 by O2
although the Cu species may be oxidized under reaction
conditions.
(Supplementary material, Table S3 and Fig. S3). All the Cu/SiO2 cat-
alysts have much low epoxidation activity except for (KAc)–Cu/
SiO2 catalyst. 19.1% PO selectivity is obtained only at 0.27% C3H6
conversion for (KAc)–Cu/SiO2 catalyst. Over all the Cu/SiO2 cata-
lysts, acrolein is the main partial oxidation product as well as a
portion of combustion product of COx. These results are quite dif-
ferent from those of Cu/SiO2 catalysts pretreated in N2 or H2. It
seems that the pretreatment of Cu/SiO2 catalysts by O2 is very
unfavorable for PO formation. The oxidized Cu species with high
valence state (Cu2+) always produce acrolein and COx.
Table 1 shows the catalytic performances of the Cu/SiO2 cata-
lysts which were oxidized by O2 at 673 K before reaction. Cu2+ spe-
cies is predominant in Cu/SiO2 catalysts in this instance
To gain further insight into the role of KAc, we compared the
catalytic performances of the Cu/SiO2 and (KAc)–Cu/SiO2 catalysts
at different reaction temperatures. Figs. 6 and 7 show the reaction

W. Su et al. / Journal of Catalysis 268 (2009) 165–174
169
8
a
a
6
5
4
3
2
1
0
Cu/SiO2
(KAc)-Cu/SiO2
6
(NaAc)-Cu/SiO2
(KCl)-Cu/SiO2
(NaCl)-Cu/SiO2
4
Cu/SiO2
(KAc)-Cu/SiO2
(NaAc)-Cu/SiO2
(KCl)-Cu/SiO2
(NaCl)-Cu/SiO2
2
0
450 500 550 600 650 700 750 800
Temperature (K)
450
500
550
600
650
700
750
Temperature (K)
b
b
100
80
60
40
20
0
100
Cu/SiO2
Cu/SiO2
(KAc)-Cu/SiO2
(NaAc)-Cu/SiO2
(KCl)-Cu/SiO2
(NaCl)-Cu/SiO2
(KAc)-Cu/SiO2
(NaAc)-Cu/SiO2
(KCl)-Cu/SiO2
(NaCl)-Cu/SiO2
80
60
40
20
0
450
500
550
600
650
700
750
450 500 550 600 650 700 750 800
Temperature (K)
Temperature (K)
Fig. 5. C3H6 conversion (a) and PO selectivity (b) as
a function of reaction
Fig. 4. C3H6 conversion (a) and PO selectivity (b) as
a function of reaction
temperature over the Cu/SiO2 catalysts pretreated by H2 at 673 K. Reaction
conditions: catalyst volume, 0.2 ml; C3H6:O2:N2 = 1:1:15; Space velocity =
temperature over the Cu/SiO2 catalysts pretreated by N2 at 673 K. Reaction
conditions: catalyst volume, 0.2 ml; C3H6:O2:N2 = 1:1:15; Space velocity =
30,000 hÀ1
.
30,000 hÀ1
.
formed IR studies of the adsorption and activation of C3H6 on Cu/
SiO2 catalyst with different valence states of Cu species. CO and
NO may be used as molecular probes to follow the electron-donat-
ing abilities of the adsorption sites [38–40]. The IR experimental
details were described in experimental section. Fig. 8 displays
the IR spectra of CO and C3H6 coadsorbed on Cu0 and Cu+ as well
as NO and C3H6 coadsorbed on Cu2+ in Cu/SiO2 catalyst. It is well
known that CO and NO are the most useful molecular probes for
monitoring the oxidation state of small Cu nanoparticles. CO is ad-
sorbed strongly on Cu+ or Cu0 sites, but it is hardly adsorbed on
Cu2+ sites at room temperature [41]. On the other hand, NO is pref-
erentially adsorbed on Cu2+ sites [42]. Thus, we are able to study
the adsorption and activation of C3H6 on different valence states
of Cu species with the two molecular probes by IR spectroscopy.
When CO is adsorbed on Cu0/SiO2 (H2 reduction), a broad and
strong band at 2103 cmÀ1 is observed, which is due to the C–O
stretching vibration of Cu0CO species [41]. After introducing C3H6
gas, the IR band is red shifted from 2103 to 2084 cmÀ1
results of the two catalysts pretreated in N2 and H2, respectively. In
both cases over Cu/SiO2, acrolein is the main partial oxidation
product even at very low C3H6 conversion. COx is the main product
(60–80%) at higher reaction temperatures. However, after KAc
modification, although C3H6 conversion is decreased, the selectiv-
ity to PO is increased significantly to 30–100% at lower reaction
temperatures and the selectivity to acrolein is in the ranges of
20–40%. At higher reaction temperatures, the selectivity to COx ex-
ceeds 50%, this may be due to the oxidation of the Cu species by O2
during the reaction process (Supplementary material, Fig. S5). Even
at the same reaction temperature, the selectivity to COx is de-
creased by about 25% compared with Cu/SiO2. Thus, KAc modifica-
tion switches the formation of a part of acrolein to that of PO at
lower reaction temperatures and suppresses the formation of COx
at higher reaction temperatures. According to the previous litera-
ture [27], we consider that KAc modification neutralizes the Lewis
acidity on Cu/SiO2 catalyst surface, which should contribute to the
inhibition of the isomerization of PO because it is known that
epoxides would easily undergo isomerization over an acidic cata-
lyst surface [37].
(D
m
= 19 cmÀ1) and the intensity of the IR band greatly decreases.
This result indicates that C3H6 donates its
electrons to Cu0, which
results in a stronger
-back donation of d electrons of Cu0 to anti-
bonding * orbitals of CO [43]. The decrease in intensity implies
p
p
p
3.3. IR spectra of adsorbed C3H6 on Cu/SiO2 with different valence
states of Cu species
that most CO molecules in Cu0CO forms are replaced by C3H6. A
weak band at 2126 cmÀ1 is also detected, which may be due to ad-
sorbed CO on Cu+ [41]. In the case of Cu+ (evacuation at 673 K, this
pretreatment can produce mostly Cu+ according to the literature
[31]), similar experimental results are obtained. Adsorbed CO
In order to figure out the role of different pretreatment methods
in the catalytic performances of C3H6 epoxidation, we also per-

170
W. Su et al. / Journal of Catalysis 268 (2009) 165–174
Table 1
Catalytic performances of C3H6 epoxidation over different Cu/SiO2 catalysts pretreated by O2 at 673 K.
Catalyst
T (K)
C3H6 conv. (%)
Selectivity (%)
PO
Acetaldehyde
Acrolein
COx
Blank
SiO2
Cu/SiO2
673
673
573
598
623
598
623
648
598
623
648
723
773
673
723
773
None
0.10
0.54
2.58
8.17
0.27
0.82
1.61
0.37
1.04
2.25
0.17
1.02
0.21
0.56
0.87
0
0
0
0
0
19.1
8.2
5.2
0
0
5.0
0
0
0
0
5.3
3.6
0
0
0
0
0
3.1
0
0
0
0
0
0
100
0
0
72.2
41.9
30.2
80.9
74.6
56.6
100
78.2
53.3
100
27.8
52.8
66.2
0
17.2
38.2
0
21.8
38.6
0
37.6
0
(KAc)–Cu/SiO2
(NaAc)–Cu/SiO2
(KCl)–Cu/SiO2
(NaCl)–Cu/SiO2
0
0
0
0
62.4
100
81.3
75.4
18.7
24.6
Reaction conditions: catalyst volume, 0.2 ml; C3H6:O2:N2 = 1:1:15; Space velocity = 30,000 hÀ1
.
a 100
a
20
15
10
5
10
8
AA
PO
Acrolein
COx
100
80
60
40
20
0
AA
PO
Acrolein
COx
80
60
40
20
0
6
4
2
0
0
450
500
550
600
650
700
500
550
600
650
700
Temperature (K)
Temperature (K)
8
6
4
2
0
b
b 100
100
80
60
40
20
0
AA
PO
Acrolein
COx
AA
PO
Acrolein
COx
4
3
2
1
0
80
60
40
20
0
450
500
550
600
650
700
450
500
550
600
650
700
Temperature (K)
Temperature (K)
Fig. 7. Temperature dependences on catalytic performances of (a) Cu/SiO2 and (b)
(KAc)–Cu/SiO2 with H2 pretreatment for the oxidation of C3H6 by O2. Reaction
conditions: catalyst volume, 0.2 ml; C3H6:O2:N2 = 1:1:15; Space velocity =
Fig. 6. Temperature dependences on catalytic performances of (a) Cu/SiO2 and (b)
(KAc)–Cu/SiO2 with N2 pretreatment for the oxidation of C3H6 by O2. Reaction
conditions: catalyst volume, 0.2 ml; C3H6:O2:N2 = 1:1:15; Space velocity =
30,000 hÀ1
.
30,000 hÀ1
.
produces a sharp and strong peak at 2136 cmÀ1 which is assigned
to CO adsorbed on Cu+ [41]. The coadsorption of C3H6 results in a
sharp decrease in intensity of the band at 2136 cmÀ1, and a red
Fig. 9 shows the IR spectra of C3H6 adsorbed on Cu+, Cu2+, and
Cu0 in Cu/SiO2, respectively. For comparison, the IR spectra of gas
phase C3H6 and C3H6 physically adsorbed on SiO2 are also pre-
sented (the pressure of C3H6 is about 270 Pa). The IR spectra of ad-
sorbed C3H6 on Cu0 and Cu2+ are almost the same as those on SiO2
and gas phase C3H6. However, in the case of Cu+, two new bands at
1541 and 1456 cmÀ1 are observed. The band at 1541 cmÀ1 is a
characteristic feature due to C@C of C3H6 interaction with Cu+
shift of C–O band from 2136 to 2126 cmÀ1 = 10 cmÀ1). The rea-
(Dm
son is the same as that for Cu0. When NO is adsorbed on Cu2+, a rel-
atively weak band at 1890 cmÀ1 appears [42]. Unlike Cu0 and Cu+,
the band at 1890 cmÀ1 remains almost the same after introducing
C3H6, which shows that C3H6 is weakly adsorbed on Cu2+
.

W. Su et al. / Journal of Catalysis 268 (2009) 165–174
171
0.005
0.1
Cu0 CO
Cu0 CO+C3H6
Cu+ CO
Cu+
Cu2+
Cu0
Cu+ CO+C3H6
Cu2+ NO
Cu2+ NO+C3H6
2300 2200 2100 2000 1900 1800 1700
1900 1800 1700 1600 1500 1400 1300
Wavenumbers(cm-1
-1
)
Wavenumbers(cm
)
Fig. 8. IR spectra of CO and C3H6 coadsorbed on Cu0 and Cu+ as well as NO and C3H6
Fig. 10. IR spectra of adsorbed C3H6 on Cu+, Cu2+, and Cu0 in Cu/SiO2 catalyst after
an evacuation for 1 min. The residual pressure of C3H6 is about 70 Pa.
coadsorbed on Cu2+ in Cu/SiO2 catalyst.
by about 20 cmÀ1 in the presence of C3H6. We consider that C3H6 is
probably adsorbed parallely on Cu0 via its C@C bond, according to
the selection rules in an IR experiment, this adsorption mode of
C3H6 may not be detected by IR spectroscopy. The IR spectrum of
C3H6 adsorbed on Cu2+ is almost the same as that of gas phase
C3H6 and the IR bands almost disappear after the evacuation. Both
the IR results of C3H6 adsorption alone and NO and C3H6 coadsorp-
tion suggest that C3H6 is hardly adsorbed on Cu2+, namely, the C@C
[31]. The red shift of the mC@C is about 100 cmÀ1 relative to free
m
C@C in gas phase C3H6 (1640 cmÀ1), suggesting that the C@C bond
is evidently activated by Cu+. It is very possible that this is the ef-
fect of * antibonding
-back donation of d electrons of Cu+ ions to
orbitals of C3H6 [31,44].
p
p
The IR spectra shown in Fig. 10 are recorded upon the adsorp-
tion of C3H6 on Cu+, Cu2+, and Cu0 in Cu/SiO2 after an evacuation
for one minute (the residual pressure of C3H6 is about 70 Pa). When
C3H6 is adsorbed on Cu+, the IR bands at 1540, 1456, 1441, 1406,
and 1377 cmÀ1 ascribed to C@C interaction with Cu+, @CH2 scissor-
ing, @CH2 scissoring, @C–H in plane bending, and methyl symmet-
rical C–H bending, respectively, [31,38] are still clearly observed
after the evacuation, especially the former two bands. It demon-
strates that C3H6 is strongly adsorbed on Cu+ sites. The strength
of the adsorbed C3H6 on Cu+ is stronger than CO since it can replace
most CO from relatively stable Cu+CO complex (Fig. 8). However,
when C3H6 is adsorbed on Cu0, the result is very different. The
characteristic band of adsorbed C3H6 (1540 cmÀ1) is not observed.
Only weak bands attributed to gas phase and physisorption of C3H6
remain after the evacuation. It seems that C3H6 is hardly adsorbed
on Cu0, but from the IR results of the coadsorption between CO and
C3H6 on Cu0 (Fig. 8), we can see adsorbed C3H6 on Cu0 is so strong
that it can replace most CO from Cu0CO forms, and the charge
transfer between C3H6 and Cu0 also exists since mC–O is red shifted
bond of C3H6 cannot be activated by Cu2+
.
From the reaction results for the oxidation of C3H6 by O2, we
find that both Cu0 and Cu+ species in Cu/SiO2 catalyst show C3H6
epoxidation activity, and it seems that Cu0 species is more favor-
able for PO formation. Cu2+ species in Cu/SiO2 hardly produces
PO, but mainly produces acrolein. As we know, the activation of
the C@C bond of C3H6 is one of the most important steps to gener-
ate PO. The result of IR spectrum shown in Fig. 10 indicates that the
interaction of C3H6 with Cu+ can greatly weaken the C@C bond: the
IR band shifts from 1637 cmÀ1 (position typical of free C3H6 mole-
cules) to about 1540 cmÀ1
(D
m
ꢀ 100 cmÀ1). In Fig. 10 we have not
observed the adsorbed C3H6 species on Cu0, however, the interac-
tion between C3H6 and Cu0 still exists because most adsorbed CO
molecules on Cu0 are replaced by C3H6 (Fig. 8). The IR results from
the coadsorption of CO and C3H6 show that
Cu+ or Cu0 can result in a red shift of mCO with respect to CO ad-
sorbed on Cu+ or Cu0 ‘‘solo”, too [43]. In addition,
donation of
p donation of C3H6 to
p
C3H6 to Cu0 is stronger than that to Cu+ (20 vs. 10 cmÀ1). Namely,
the activation of the C@C bond on Cu0 is stronger than that on Cu+.
While in the case of Cu2+, C3H6 is hardly adsorbed and the C@C
bond cannot be activated. Thus, these IR results are in good agree-
ment with those of the fixed bed catalytic reactor experiments. In a
word, the more the activation of the C@C bond of C3H6, the more
the favorableness for PO formation.
0.02
Cu+
We also performed the IR experiments for the coadsorption of
CO or NO and C3H6 on promoters-modified Cu/SiO2 catalysts with
different valence states of Cu species (spectra not shown). These
results are very similar to those of Cu/SiO2 without modification,
and only the intensity of the adsorbed C3H6 bands is slightly differ-
ent. This may be due to the various concentrations of Cu species
after modification by promoters. It indicates that the adsorption
behavior of C3H6 on Cu0, Cu+, and Cu2+ in Cu/SiO2 is nearly indepen-
dent of the promoters (such as KAc, NaAc, KCl, and NaCl). These
promoters have little effect on the adsorption and activation of
the C@C bond of C3H6 on Cu/SiO2 catalysts, irrespective of the va-
lence states of Cu species. Thus, the role of the promoters in the
catalytic performances of the Cu/SiO2 catalysts has not been ex-
plained by the IR experiments.
Cu2+
Cu0
SiO2
gas
1900 1800 1700 1600 1500 1400 1300
Wavenumbers(cm-1)
Fig. 9. IR spectra of adsorbed C3H6 on Cu+, Cu2+, and Cu0 in Cu/SiO2 catalyst,
respectively. For comparison, IR spectra of physisorbed and gas phase C3H6 are also
presented. The pressure of C3H6 is about 270 Pa.

172
W. Su et al. / Journal of Catalysis 268 (2009) 165–174
a
b
c
d
Fig. 11. TEM images of Cu/SiO2 samples calcined at different temperatures with respective particle size distributions. (a) 723 K (2.9 nm), (b) 848 K (5.9 nm), (c) 983 K
(8.5 nm), and (d) 1073 K (11.5 nm).
From the above discussion, the catalytic performances on Cu/
SiO2 and promoters-modified Cu/SiO2 catalysts are very different,
especially for (KAc)–Cu/SiO2 and (NaAc)–Cu/SiO2 catalysts during
the low reaction temperatures of the oxidation reaction. However,
the IR results show that the activation of the C@C bond of C3H6 is
independent of the promoters investigated in this work irrespec-

W. Su et al. / Journal of Catalysis 268 (2009) 165–174
173
tive of Cu0 or Cu+ species. We know that the other important factor
to create PO is the moderate electrophilic oxygen species. There-
fore, besides neutralizing the surface Lewis acidity, we consider
that there may exist an interaction between K+ or Na+ and Cu spe-
cies (Cu0 or Cu+) [26,28]. The interaction can suppress the reactiv-
ity of lattice oxygen species, which are generally regarded to cause
the formation of acrolein (the allylic oxidation). Furthermore, the
modification by K+ or Na+ may stabilize the moderate electrophilic
oxygen species generated via the activation of oxygen on the near-
by copper sites (Cu0 or Cu+) [28]. All these factors are beneficial to
PO formation. With raising the reaction temperatures, the Cu0 or
Cu+ species is probably gradually oxidized to Cu2+ species (Supple-
mentary material, Fig. S5), which leads to produce abundant acro-
lein and COx, and the selectivity to PO is largely decreased.
However, when KCl or NaCl was used as a modifier, the situation
is totally varied. The role of K+ or Na+ is already presented above,
while the effect of ClÀ on Cu species is much severer. From Figs.
1–3, it is clear that ClÀ ions can result in the agglomeration of Cu
species to form agglomerate CuO. In this instance, the concentra-
tion of Cu0 or Cu+ is greatly diminished. Hence, the conversion of
C3H6 is rather lower, and the selectivity to PO is also decreased
by some extent compared to the catalyst modified by KAc or NaAc,
suggesting that larger CuO particle size (15 nm) is not suitable to
get PO. Previous investigation shows that NaCl is a good modifier
in Ag-based catalysts for C3H6 epoxidation reaction to obtain PO
[13,16]. They think that NaCl can make the adsorbed oxygen spe-
cies on the nearby silver sites relatively mildly electrophilic, which
is favorable for the epoxidation. However, in our present studies on
the Cu/SiO2 catalysts, the promotive effect of NaCl on producing PO
is not so remarkable, at least the promotive effect of NaCl is much
inferior to that of KAc.
Table 2
Catalytic performances of Cu/SiO2 catalysts with different CuO particle size for the
oxidation of C3H6 by O2.
CuO particle
size (nm)
T
(K)
PO formation rate
(mmol gÀ1 hÀ1
Acrolein formation rate
)
(mmol gÀ1 hÀ1
)
2.9
5.9
8.5
11.5
498 0.26
473 0.44
498 0.15
523 0.14
0.55
0.16
0.19
0.13
Reaction conditions: catalyst volume, 0.2 ml; C3H6:O2:N2 = 1:1:15; Space veloc-
ity = 30,000 hÀ1. Before reaction, all the catalysts were pretreated by H2 at 673 K.
We chose the reaction temperature at which the PO formation rate reaches the
maximum for each catalyst.
Table 3
Catalytic performances of (KAc)–Cu/SiO2 catalysts with different CuO particle size for
the oxidation of C3H6 by O2.
CuO particle
size (nm)
T
(K)
PO formation rate
(mmol gÀ1 hÀ1
Acrolein formation rate
)
(mmol gÀ1 hÀ1
)
2.9
5.9
8.5
11.5
523 2.50
523 0.43
523 0.24
573 0.25
1.13
0.20
0
0
Reaction conditions: catalyst volume, 0.2 ml; C3H6:O2:N2 = 1:1:15; Space veloc-
ity = 30,000 hÀ1. Before reaction, all the catalysts were pretreated by H2 at 673 K.
We chose the reaction temperature at which the PO formation rate reaches the
maximum for each catalyst.
(data not shown). The rate of acrolein formation also decreases
and the higher C3H6 conversion results from the combustion reac-
tion. These results suggest that the size of CuO particles is an abso-
lutely important factor in determining the catalytic performances
for C3H6 epoxidation reaction. The Cu/SiO2 catalyst with the size
of CuO particles at about 5.9 nm is the most beneficial to generate
PO. Once the CuO particle size exceeds 10 nm, the selectivity to PO
on this catalyst drops nearly to zero.
3.4. Size effect of CuO nanoparticles on the catalytic performances of
C3H6 epoxidation
Quantum size effect is a very common and important topic in
many fields including catalysis. However, the size effect of CuO
nanoparticles on the catalytic performances of C3H6 epoxidation
over Cu-based catalysts has hardly been investigated previously.
Fig. 11 indicates the TEM images of Cu/SiO2 samples calcined at
different temperatures with respective particle size distributions.
The sizes of CuO particles are adjusted by raising the calcination
temperatures from 723 to 1073 K. The CuO particle size in each
catalyst is presented in Fig. 11 and is increased from 2.9 to
11.5 nm with the elevated calcination temperatures. According to
Scherrer formula D = Kk/(b cos h), the size of CuO particles in Cu/
SiO2 calcined at 983 K and 1073 K is about 9.1 nm and 15.1 nm,
respectively (Supplementary material, Fig. S6). These results are
in good agreement with those of TEM images. Tables 2 and 3 show
the catalytic performances of Cu/SiO2 and (KAc)–Cu/SiO2 catalysts
with different CuO particle size for the oxidation of C3H6 by O2,
respectively. According to the above results, the Cu/SiO2 catalyst
modified by KAc and pretreated in H2 shows the highest PO selec-
tivity. Thus, in this series of experiments, we chose KAc as the
modifier, and all the Cu/SiO2 catalysts were pretreated by H2 before
reaction. In Table 2, the rate of PO formation reaches a maximum
on the Cu/SiO2 catalyst with about 5.9 nm CuO particle size
(0.44 mmol gÀ1 hÀ1), and then gradually decreases when the CuO
particle size is ranged from about 2.9 to 8.5 nm. Moreover, the
combustion product of COx is not observed, and the only side prod-
uct is acrolein. The variety of the acrolein formation rate is just
opposite to that of PO. Once the size of CuO particles reaches
11.5 nm, PO is hardly produced at the same reaction temperature.
Although the PO formation rate remains 0.14 mmol gÀ1 hÀ1 due to
the higher C3H6 conversion at higher reaction temperature, PO
selectivity is not more than 10% and COx selectivity exceeds 80%
We also studied the size effect of CuO particles on C3H6 epoxi-
dation reaction over (KAc)–Cu/SiO2 catalysts. In Table 3, the high-
est rate of PO formation can reach 2.5 mmol gÀ1 hÀ1 on the (KAc)–
Cu/SiO2 catalyst with the size of CuO particles at about 2.9 nm. This
result is the highest in all Cu-based catalysts reported for C3H6
epoxidation reaction by O2 [17,18,25–28]. However, the rates of
both PO formation and acrolein formation decrease with the in-
crease of CuO particle size from 2.9 to 11.5 nm. In addition, the de-
crease of the PO formation rate is very sharp especially when the
CuO particle size is increased from about 2.9 to 5.9 nm. Once the
size of CuO particles exceeds 8 nm, acrolein is hardly produced.
The combustion reaction is predominant and the selectivity to
COx exceeds 75%. In fact, COx is detected on all the (KAc)–Cu/SiO2
catalysts and the selectivity to COx increases fast with the increase
of CuO particle size. These phenomena are very different from
those observed on the Cu/SiO2 catalysts. In the case of the (KAc)–
Cu/SiO2 catalyst series which was prepared with the CuO particle
size ranging from 2.9 to 11.5 nm, it is found that the smaller the
size of the CuO particles, the higher the activity for PO formation.
On the other hand, the catalyst with large CuO particle size tends
to produce COx instead of PO. This may be due to the higher con-
centration of Cu+ and Cu0 species in smaller CuO particles.
Comparing the catalytic performances on Cu/SiO2 catalysts with
and without KAc modification, we find that the rate of PO forma-
tion is only slightly enhanced by KAc modification when the CuO
particle size is in the ranges of 5.9–11.5 nm. Namely, the effect of
KAc modification on PO formation is very small. On the other hand,
the rate of acrolein formation is suppressed after the modification
by KAc, and COx is the main product. These results are quite differ-
ent from those of the Cu/SiO2 catalyst with 2.9 nm CuO particle

174
W. Su et al. / Journal of Catalysis 268 (2009) 165–174
size, on which KAc modification largely increases the rates of PO
formation and acrolein formation especially for PO formation.
The rate of PO formation is enhanced by about one order after
KAc modification. These suggest that the (KAc)–Cu/SiO2 catalyst
with 2.9 nm CuO particle size is a very special sample and is differ-
ent from other Cu/SiO2 catalysts. From the above discussion, there
is an interaction between Cu species and K+, which is favorable for
PO formation [26]. This interaction may be weaker and weaker
with increasing CuO particle size, especially when the size of
CuO particles exceeds 6 nm. Thus, in the case of the (KAc)–Cu/
SiO2 catalyst (2.9 nm), the strong interaction between K+ and Cu
species may further stabilize the electrophilic oxygen species
[28], which switches the reaction pathway from the formation of
acrolein and COx to that of PO and leads to high PO selectivity.
References
[1] T.A. Nijhuis, M. Makkee, J.A. Moulijn, B.M. Weckhuysen, Ind. Eng. Chem. Res. 45
(2006) 3447.
[2] M.G. Clerici, G. Bellussi, U. Romano, J. Catal. 129 (1991) 159.
[3] Z.W. Xi, N. Zhou, Y. Sun, K.L. Li, Science 292 (2001) 1139.
[4] T. Hayashi, K. Tanaka, M. Haruta, J. Catal. 178 (1998) 566.
[5] A.K. Sinha, S. Seelan, S. Tsubota, M. Haruta, Angew. Chem. Int. Ed. 43 (2004)
1546.
[6] E. Ananieva, A. Reitzmann, Chem. Eng. Sci. 59 (2004) 5509.
[7] Y. Wang, W. Yang, L.J. Yang, X.X. Wang, Q.H. Zhang, Catal. Today 117 (2006)
156.
[8] X.X. Wang, Q.H. Zhang, Q. Guo, Y.C. Lou, L.J. Yang, Y. Wang, Chem. Commun.
(2004) 1396.
[9] J.R. Monnier, Appl. Catal. A 221 (2001) 73.
[10] R.P. Wang, X.W. Guo, X.S. Wang, J.Q. Hao, G. Li, J.H. Xiu, Appl. Catal. A 261
(2004) 7.
[11] M.F. Luo, J.Q. Lu, C. Li, Catal. Lett. 86 (2003) 43.
[12] G.J. Jin, G.Z. Lu, Y.L. Guo, Y. Guo, J.S. Wang, X.H. Liu, Catal. Lett. 87 (2003)
249.
4. Conclusions
[13] G.J. Jin, G.Z. Lu, Y.L. Guo, Y. Guo, J.S. Wang, X.H. Liu, Catal. Today 93 (2004)
173.
Both Cu0 and Cu+ species in Cu/SiO2 catalyst show C3H6 epoxi-
dation activity. Cu2+ species in Cu/SiO2 hardly produces PO, but
mainly produces acrolein and COx. IR studies indicate that C3H6
is strongly bonded to Cu0 and Cu+ sites since most adsorbed CO
[14] R.P. Wang, X.W. Guo, X.S. Wang, J.Q. Hao, Catal. Today 93 (2004) 217.
[15] J.Q. Lu, J.J. Bravo-Suarez, M. Haruta, S.T. Oyama, Appl. Catal. A 302 (2006)
283.
[16] J.Q. Lu, M.F. Luo, H. Lei, C. Li, Appl. Catal. A 237 (2002) 11.
[17] J.Q. Lu, M.F. Luo, H. Lei, X.H. Bao, C. Li, J. Catal. 211 (2002) 552.
[18] J.Q. Lu, M.F. Luo, C. Li, Chin. J. Catal. 25 (2004) 5.
[19] M. Akimoto, K. Ichikawa, E. Echigoya, J. Catal. 76 (1982) 333.
[20] R.L. Cropley, F.J. Williams, O.P.H. Vaughan, A.J. Urquhart, M.S. Tikhov, R.M.
Lambert, Surf. Sci. 578 (2005) L85.
can be replaced by adsorbed C3H6.
Cu0 results in a red shift of mCO with respect to CO adsorbed on
Cu+ or Cu0 ‘‘solo”. Moreover, donation of C3H6 to Cu0 is stronger
than that to Cu+. -Back donation of d electrons of Cu+ ions to
p
donation of C3H6 to Cu+ or
p
[21] R.L. Cropley, F.J. Williams, A.J. Urquhart, O.P.H. Vaughan, M.S. Tikhov, R.M.
Lambert, J. Am. Chem. Soc. 127 (2005) 6069.
p
p*
antibonding orbitals of C3H6 also takes place, which results in a dis-
tinct C@C bond weakening. C3H6 is weakly adsorbed on Cu2+, how-
ever, the charge transfer between C3H6 and Cu2+ is not obvious.
KAc is found to be an excellent modifier for PO formation, while
ClÀ modification only slightly enhances PO selectivity. The modifi-
cation by K+ can switch the main reaction route from allylic oxida-
tion to epoxidation. The small CuO particle size is most favorable
for PO formation especially when Cu/SiO2 sample is modified by
KAc, and large CuO particle size tends to produce acrolein and
COx. The concentration of Cu0 and Cu+ species is much higher in
small CuO particles. Thus, the isolated Cu species and small CuOx
clusters modified by KAc may be the active sites for C3H6 epoxida-
tion reaction using molecular oxygen.
[22] R.M. Lambert, F.J. Williams, R.L. Cropley, A. Palermo, J. Mol. Catal. A 228 (2005)
27.
[23] D. Torres, N. Lopez, F. Illas, R.M. Lambert, Angew. Chem. Int. Ed. 46 (2007)
2055.
[24] J.R. Monnier, G.W. Hartley, J. Catal. 203 (2001) 253.
[25] O.P.H. Vaughan, G. Kyriakou, N. Macleod, M. Tikhov, R.M. Lambert, J. Catal. 236
(2005) 401.
[26] H. Chu, L. Yang, Q.H. Zhang, Y. Wang, J. Catal. 241 (2006) 225.
[27] W.M. Zhu, Q.H. Zhang, Y. Wang, J. Phys. Chem. C 112 (2008) 7731.
[28] Y. Wang, H. Chu, W.M. Zhu, Q.H. Zhang, Catal. Today 131 (2008) 496.
[29] C.J.G. Vandergrift, P.A. Elberse, A. Mulder, J.W. Geus, Appl. Catal. 59 (1990)
275.
[30] A.R. Balkenende, C.J.G. Vandergrift, E.A. Meulenkamp, J.W. Geus, Appl. Surf. Sci.
68 (1993) 161.
[31] J. Datka, E. Kukulska-Zajac, J. Phys. Chem. B 108 (2004) 17760.
[32] S. Bordiga, R. Buzzoni, F. Geobaldo, C. Lamberti, E. Giamello, A. Zecchina, G.
Leofanti, G. Petrini, G. Tozzola, G. Vlaic, J. Catal. 158 (1996) 486.
[33] G. Cordoba, R. Arroyo, J.L.G. Fierro, M. Viniegra, J. Solid State Chem. 123 (1996)
93.
[34] F. Boccuzzi, S. Coluccia, G. Martra, N. Ravasio, J. Catal. 184 (1999) 316.
[35] S. Velu, K. Suzuki, M. Okazaki, M.P. Kapoor, T. Osaki, F. Ohashi, J. Catal. 194
(2000) 373.
[36] M. Haruta, Cattech 6 (2002) 102.
[37] J.M. Coxon, R. Maclagan, A. Rauk, A.J. Thorpe, D. Whalen, J. Am. Chem. Soc. 119
(1997) 4712.
[38] E. Kukulska-Zajac, N. Kumar, T. Salmi, D.Y. Murzin, J. Datka, Catal. Today 100
(2005) 407.
Acknowledgments
This work was financially supported by the National Basic Re-
search Program of China (Grants 2009CB623507), the National
Natural Science Foundation of China (NSFC, Grants 20590363),
and Program Strategic Scientific Alliances between China and the
Netherlands (Grants 2008DFB50130).
[39] R. Burch, J.P. Breen, F.C. Meunier, Appl. Catal. B 39 (2002) 283.
[40] R. Burch, Phys. Chem. Chem. Phys. 8 (2006) 5483.
[41] K.I. Hadjiivanov, G.N. Vayssilov, Adv. Catal. 47 (2002) 307.
[42] K. Hadjiivanov, H. Knozinger, Phys. Chem. Chem. Phys. 3 (2001) 1132.
[43] J. Datka, E. Kukulska-Zajac, W. Kobyzewa, Catal. Today 114 (2006) 169.
[44] E. Broclawik, P. Rejmak, P. Kozyrac, J. Datka, Catal. Today 114 (2006) 162.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2009.09.017.