Gas-phase epoxidation of 3,3,3-trifluoropropylene over Au/CuTiO2 catalysts with N2O as the oxidant

Article abstract of DOI:10.1016/j.jcat.2014.01.015

Gas-phase epoxidation of 3,3,3-trifluoropropylene (TFP) was conducted on a series of Au/CuTiO₂ catalysts with different Cu contents with N ₂O as the oxidant. These catalysts were effective for this reaction. The best catalytic performance was obtained on a catalyst containing 4.6 wt.% of Au and 0.9 wt.% of Cu (4.6Au/0.9CuTiO₂), with a steady-state 3,3,3-trifluoropropylene oxide (TFPO) formation rate of 72.4g _TFPOh^-1kgcat-1, which was much higher than that on a 2.1Au/TiO₂ catalyst (22.1g_TFPOh^-1kgcat-1). The enhancement was attributed to the higher Au content in the Cu-promoted catalyst and small Au particle size and more importantly to the complicated synergy between the AuCuTiO₂ interaction which might be the active sites for epoxidation. Also, high selectivity to TFPO up to 88% was obtained on the Cu-promoted catalyst, due to the proper electronic structure induced by the interaction between Au and low valent Cu species. Catalyst deactivation was due to the significant growth of Au particles and loss of AuCuTiO₂ interface because of the segregation of Cu species during the reaction.

Full text of DOI:10.1016/j.jcat.2014.01.015

Journal of Catalysis 312 (2014) 139–151
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Gas-phase epoxidation of 3,3,3-trifluoropropylene over Au/CuATiO2
catalysts with N2O as the oxidant
⇑
Chong-Xiang Sun, Yu Wang, Ai-Ping Jia, Shu-Xia Chen, Meng-Fei Luo, Ji-Qing Lu
Key Laboratory of the Ministry of Education for Advanced Catalysis Materials, Institute of Physical Chemistry, Zhejiang Normal University, Jinhua 321004, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Gas-phase epoxidation of 3,3,3-trifluoropropylene (TFP) was conducted on a series of Au/CuATiO2
catalysts with different Cu contents with N2O as the oxidant. These catalysts were effective for this reac-
tion. The best catalytic performance was obtained on a catalyst containing 4.6 wt.% of Au and 0.9 wt.% of
Cu (4.6Au/0.9CuATiO2), with a steady-state 3,3,3-trifluoropropylene oxide (TFPO) formation rate of
72:4 gTFPO hꢀ1 kgꢀ1, which was much higher than that on a 2.1Au/TiO2 catalyst (22:1 gTFPO hꢀ1 kgꢀ1).
Received 8 November 2013
Revised 12 January 2014
Accepted 17 January 2014
cat
cat
The enhancement was attributed to the higher Au content in the Cu-promoted catalyst and small Au par-
ticle size and more importantly to the complicated synergy between the AuACuATiO2 interaction which
might be the active sites for epoxidation. Also, high selectivity to TFPO up to 88% was obtained on the Cu-
promoted catalyst, due to the proper electronic structure induced by the interaction between Au and low
valent Cu species. Catalyst deactivation was due to the significant growth of Au particles and loss of
AuACuATiO2 interface because of the segregation of Cu species during the reaction.
Ó 2014 Elsevier Inc. All rights reserved.
Keywords:
3,3,3-Trifluoropropylene
Epoxidation
3,3,3-Trifluoropropylene oxide
Au catalyst
1. Introduction
In contrast, gas-phase propylene has been extensively investi-
gated in the literature. Catalysts for direct propylene epoxidation
3,3,3-Trifluropropylene oxide (TFPO) is a very important inter-
mediate in fluorochemical industry. Through ring-open reactions,
various valuable compounds can be synthesized using TFPO as
the starting material. For example, one of its derivatives, optically
mainly include Ag [4–6], Au [7–15], and Cu [16–19] as the active
components, using either molecular O2 (for Ag and Cu catalysts)
or a mixture of H2 and O2 (for Au catalysts). The use of molecular
oxygen usually leads to low selectivity to propylene oxide, because
active
a-trifluoromethyl alcohol has been frequently utilized in
the generated oxygen species are nucleophilic in nature (e.g., O2ꢀ
)
chiral moiety of liquid crystals, which could be synthesized
through a nucleophilic or acid catalyzed ring opening reaction of
optically active TFPO [1]. Also, TFPO is also a precursor for triflu-
oropropylene carbonate, an electrolyte for lithium batteries and
fuel cells [2]. Currently, the synthesis of TFPO exclusively goes
and they tend to attack the allylic hydrogen of CH3ꢀ group in the
propylene molecule, which could consequently suppress the selec-
tivity to the desired propylene oxide (PO) [20]. The mixture of H2
and O2 as the oxidant applied on Au catalysts supported on Ti-con-
taining materials (e.g., TiO2, TiO2ASiO2, Ti-MCM-41, TS-1, Ti-TUD)
leads to high selectivity to propylene oxide (PO), because the
in situ generated H2O2 on Ti site could form a peroxide species
(TiAOOH) as an important intermediate which could further react
with propylene [21]. Since the chemical properties of the oxygen
species are crucial in epoxidation reaction, nitrous oxide (N2O) is
also considered as a promising oxidant because it could be acti-
vated on metal sites in the catalysts and decomposes to N2 and
atomically adsorbed oxygen with a mild electrophilic character.
Thus, N2O has been used as a suitable oxidant for direct gas-phase
propylene epoxidation over iron oxides [20,22–29] and Au cata-
lysts [30,31]. For example, in an initial work by Duma and Hönicke
[20], Na-promoted iron oxide catalysts supported on silica were
tested for propylene epoxidation by N2O, and selectivities to PO
of 40–60% at propylene conversions in the range of 6–12% were ob-
served. Moreover, with the addition of proper promoters such as
through
a liquid-phase reaction with 3,3,3-trifluropropylene
(TFP) as the reactant. For example, TFPO could be synthesized by
mixing TFP with fuming sulfuric acid, followed by hydrolysis and
alkali treatment using 20% NaOH solution at 85 °C [3]. Such routes
have disadvantages such as difficulty in product separation and
pollution problem. In a heterogeneous catalysis point of view,
gas-phase epoxidation of TFP seems appealing because this process
may be more efficient and environmentally benign compared to
the liquid-phase reaction. But unfortunately, no work on this topic
has been reported in the literature yet.
⇑
Corresponding author. Fax: +86 579 2282595.
E-mail address: Jiqinglu@zjnu.cn (J.-Q. Lu).
http://dx.doi.org/10.1016/j.jcat.2014.01.015
0021-9517/Ó 2014 Elsevier Inc. All rights reserved.

140
C.-X. Sun et al. / Journal of Catalysis 312 (2014) 139–151
Rb2SO4 [28] and K+ [24,29], the catalytic performance of the iron
oxide catalysts could be even enhanced. As for the nature of the ac-
tive Fe sites, Zhang et al. [24] prepared KCl-modified Fe-MFI and
Fe-MCM-41 catalysts with iron species in different locations and
tested for the epoxidation of propylene with nitrous oxide, and
the authors concluded that the highly dispersed extra-framework
iron species were responsible for the epoxidation of propylene
with nitrous oxide. Au catalysts have been also applied in propyl-
ene epoxidation with N2O. Chimentão et al. [30] prepared AuACu
alloy catalysts supported on TiO2 and tested them for propylene
epoxidation, and the authors found that the sample with a Cu/Au
ratio of 3/1 showed the best catalytic performance, with a propyl-
ene conversion of 4.3% and a PO selectivity of 26.3% at 573 K. The
PO selectivity could be enhanced by the modified preparation of
AuACu/TiO2 catalysts. By using thiol-capped AuACu alloy nano-
particles, the PO selectivity could be increased up to 52.2% [31].
As an analog of propylene, the gas-phase epoxidation of TFP
could be inspired by the experiences of propylene epoxidation.
Moreover, the absence of CH3ꢀ group in TFP molecule avoids the
attack of allylic hydrogen by oxygen species, and thus, a high selec-
tivity to TFPO could be possibly obtained. Thus, in this work, a ser-
ies of Au/CuATiO2 catalysts were prepared and tested for TFP
epoxidation with N2O as the oxidant. The catalytic behaviors of
these catalysts were investigated, which were correlated with their
chemical natures.
X-ray diffraction (XRD) patterns were recorded using a PANalyt-
ic X’Pert PW3040 diffractometer with Cu K _radiation operated at
40 kV and 40 mA. The patterns were collected in a 2 theta ranged
a
from 10° to 90°, with a scanning step of 0.15°/s.
High-resolution transmission electron microscopy (HRTEM)
was performed on a JEM-2100F microscopy with a field emissive
gun, operated at 200 kV and with a point resolution of 0.24 nm.
Particle distributions were analyzed by measuring at least 100 par-
ticles in representative regions.
The reduction properties of the samples were measured by
hydrogen temperature-programmed reduction (H2-TPR) experi-
ments. Fifty milligram of the catalyst was placed in a quartz reactor
and pretreated in a He flow (30 ml minꢀ1) at 100 °C for 1 h to re-
move the water. Then, it was heated from r.t. to 800 °C at a rate
of 10 °C minꢀ1 in a H2AN2 flow (5 vol% H2, 30 ml minꢀ1). The
hydrogen consumption during the reduction was determined by
a gas chromatograph with a thermal conductivity detector (TCD).
The water produced in TPR was trapped on a 5 Å molecular sieve.
X-ray photoelectron spectra (XPS) were recorded using a VGES-
CALAB 250Xi spectrometer with Al K
a radiation (1486.6 eV). The
voltage and power for the measurements were 12.5 kV and
250 W, respectively. The vacuum in the test chamber during the
collection of spectra was kept at 2 ꢂ 10ꢀ8 Pa. In order to obtain
true information of the sample under reaction conditions, before
XPS measurement, the catalyst was pretreated in a reaction gas
mixture (TFP/N2O/N2 = 3.5/3.5/28.0 cm3 minꢀ1) from r.t. to reac-
tion temperature in 4 h (same conditions as in the reaction pro-
cess). But for the spent sample, it was pretreated in He flow
(30 ml minꢀ1) at 100 °C for 1 h to remove the water. The spectra
obtained, once the background was removed, were fitted to
Lorentzian and Gaussian lines to obtain the number of compo-
nents, peak position, and their areas. The adventitious C1s line at
284.6 eV was used as an internal standard. It is worth noting that
to avoid reduction in oxidized metal species (Aun+ and Cun+) under
high vacuum environment by X-ray during the measurement, 30
scans were used for each sample. Also, in order to distinguish
Cu0 and Cu+ species, Auger lines of Cu species in the samples were
investigated. To avoid the interference of Ti 2S with Cu LMM, XR4
2. Experimental
2.1. Catalyst preparation
The Cu/TiO2 support was prepared using an impregnation
method. In a typical process, certain amount of Cu(NO3)2 was dis-
solved in water, then 2 g of TiO2 (P25, Degussa, SBET = 55 m2 gꢀ1
)
was added. The suspension was stirred at r.t. for 1 h and kept for
6 h. The solid was dried and heated at 200 °C overnight. Other me-
tal modified TiO2 support was prepared in a similar manner.
The supported Au catalysts were prepared using the deposi-
tion–precipitation method (DP). A 100 cm3 solution of HAuCl4ꢁ4H2-
O (2 mg cmꢀ3, Jiuyue Chem. 99.8 wt.%) was heated to 70 °C under
vigorous stirring. After adjusting the pH of the solution to 7 using
a 0.2 M NaOH solution, 1 g of support was added, and the suspen-
sion was aged at 70 °C for 1 h. After cooling to RT, the solid was col-
lected via centrifugation, washed with 10 cm3 of deionized water,
and centrifuged again. The process was repeated for five times. Fi-
nally, the solid was vacuum dried at r.t. overnight and the resulting
solid was not further calcined. The catalysts were designated as
xAu/yCuATiO2, with the x and y referring to the contents of Au
and Cu (wt.%) in the catalyst, respectively. A reference Au catalyst
twin anode with Mg Ka radiation (1253.6 eV) was used.
Temperature-programmed desorptions of N2O and TFP were
performed on a homemade apparatus. 0.1 g of the catalyst was
loaded in the middle of a quartz tubular reactor attached with a
thermal couple. The catalyst was pretreated in
a He flow
(30 ml minꢀ1) at 100 °C for 1 h to remove the water. Then, the cat-
alyst was cooled down to r. t. Then, a N2O or TFP gas with a flow
rate of 20 ml minꢀ1 was introduced to the sample for 30 min, fol-
lowed by purging of He for another 30 min. The sample was heated
from r.t. to 800 °C with a ramp of 10 °C minꢀ1 in He flow. The outlet
gas was monitored by a mass spectrometer (MS, Qic-20 Benchtop,
HidenAnalytical). For the N2O desorption, m/e signals of 28, 32, and
44 were monitored, corresponding to N2, O2, and N2O, respectively.
For the TFP desorption, m/e signals of 18, 44, 20, 96, 29, 43, 98, cor-
responding to H2O, CO2, TFP, HF, trifluoropropylene oxide (TFPO)
and trifluoropropionaldehyde (TFPA), respectively.
supported on Cu-modified c-Al2O3 (denoted as Au/CuAAl2O3) was
prepared in a similar manner.
2.2. Characterizations
Fourier transform infrared (FTIR) spectra of the samples were
recorded on a NEXUS670 spectrometer equipped with a MCT
detector. Self-supported sample wafers (diameter = 1 cm) were
prepared from 15 mg of sample by pressing at about 3 MPa. It
was pretreated in the IR cell with a He flow (30 ml minꢀ1) at
100 °C for 1 h to remove the water. Spectra were recorded at
4 cmꢀ1 resolution and averaged over 100 scans.
Actual contents of Au and Cu (or other promoters) in the cata-
lysts were determined by inductively coupled plasma atomic emis-
sion spectrometry measurements (ICP-AES, Optima 7300DV,
Perkin Elmer). The content of Cl in the catalyst was measured by
an X-ray fluorescence spectrometry (ARL ADVANT’X Intelli Power
4200).
Specific surface areas (SBET) of the catalysts were measured from
a multipoint Brunauer–Emmett–Teller (BET) analysis of the nitro-
gen adsorption and desorption isotherms at 77 K, recorded on a
Quantachrome Autosorb-1 apparatus. Before measurement, the
sample was dried at 100 °C for 4 h under vacuum.
2.3. Catalytic testing
Epoxidation of propylene was carried out in a quartz tubular
microreactor (i.d. = 8 mm, length = 180 mm) using 0.3 g catalyst

C.-X. Sun et al. / Journal of Catalysis 312 (2014) 139–151
141
of 100–140 mesh size without dilution. The catalyst was heated in
I
(a) Fresh catalysts
I
Anatase
a reaction gas mixture (TFP/N2O/N2 = 3.5/3.5/28.0 cm3 minꢀ1) from
r.t. to reaction temperature in 4 h. Products were analyzed online
using a gas chromatograph (Shimazu GC-2014) equipped with a
flame ionization detector and a thermal conductivity detector, at-
tached, respectively, to a FFAP capillary column and a Unibeads C
compact column. The FFAP capillary column was used to detect
oxygenates (i.e., trifluoroacetaldehyde (TFAA), trifluoropropylene
oxide (TFPO), trifluoroacetone (TFAC), trifluoropropionaldehyde
(TFPA)), while the Unibeads C column was used to detect N2O,
O2, CO2, and H2O. To avoid the possibility of production of HF dur-
ing the reaction which may be harmful to the analysis system, the
outlet gas passed through a plastic trap containing H2O.
I
I'
I' Rutile
I
I
I
I
I
I
I'
I'
I
I'
I
1.0Cu/TiO2
5.0Au/3.9Cu-TiO2
4.9Au/2.1Cu-TiO2
4.6Au/1.6Cu-TiO2
4.6Au/0.9Cu-TiO2
3.4Au/0.5Cu-TiO2
2.4Au/0.1Cu-TiO2
2.1Au/TiO2
The following definitions were used:
20
30
40
50
60
70
80
90
100
TFP conversion ¼ moles ofðoxygenates
þ CO2=3Þ=moles of TFP in feed:
2 theta /degree
Anatase
Rutile
I"
I
I'
(b) spent catalysts
N2Oconversion ¼ ðmoles of N2O in feed
ꢀmoles of N2O in outletÞ=moles of N2O in feed:
I
I"
TiOF2
I
TFPO selectivity ¼ moles of TFPO=moles of ðoxygenates
þ CO2=3Þ:
I'
I
I
I
I
I'
I
I'
I
I
I'
2.1Au/TiO2
3.4Au/0.5Cu-TiO2
4.6Au/0.9Cu-TiO2
4.9Au/2.1Cu-TiO2
N2Oselectivity ¼ Conversion of TFP
I"
I"
I"
ꢂ Selectivity to TFPO=Conversion of N2O:
I"
I"
3. Results and discussion
3.1. Physical properties
20
30
40
50
60
70
80
90
100
2 theta /degree
Table 1 lists the surface areas, the actual Cu and Au and Cl con-
tents in the samples. The surface areas of the prepared catalysts are
39–53 m2 gꢀ1, with the high Cu loading samples having low sur-
face areas. Also, it is found that the Cu-free Au/TiO2 catalyst has
an Au loading of 2.1%, while with the addition of Cu, the Au content
gradually increases, with the highest Au content of 5.05% obtained
on a sample with Cu content of 3.92%. The enhanced Au capture
efficiency is because the isoelectric point (IEP) of CuO is 8.5–9.5,
which is much higher than that of the TiO2 (4.0–5.0). During the
DP process, the [Au(OH)3Cl]ꢀ species would be repulsed by the neg-
atively charged TiO2 surface at a pH value of 7. But with the addi-
tion of Cu species on the TiO2, the support surface could be
positively charged at pH = 7, which would attract [Au(OH)3Cl]ꢀ
species and consequently result in high Au content in the final
sample. Cl contents in the samples are very low, ranging from trace
amount to 0.09 wt.%.
Fig. 1. XRD patterns of (a) fresh and (b) spent Au catalysts.
diffraction peaks in forms of anatase (PDF 21-1272) and rutile (PDF
21-1276). In addition, no diffraction peaks due to Au or Cu species
are detected, suggesting that the these species are highly dispersed
on the support. For the spent catalysts (after reaction), additional
diffraction peaks at 2h of 23.6°, 33.4°, 48.1°, and 54.2° are detected,
which could be assigned to TiOF2 (PDF 08-0060). The formation of
the TiOF2 phase is accompanied by the decrease in the intensities
of the diffraction peaks of anatase TiO2. The formation of TiOF2
could be due to the reaction between anatase TiO2 and HF formed
during the oxidation of TFP (TiO2 + 2HF ? TiOF2 + H2O). The re-
verse reaction has been reported in a recent work [32].
Fig. 2 shows the HRTEM and Fourier transformation (FT) images
of various catalysts. For the 1.0CuATiO2 sample, no distinct CuO
particles are observed, and the clear lattice fringe of 0.352 nm is as-
signed to TiO2(101) plane. For the 2.1Au/TiO2 catalyst, lattice
fringes of 0.236 and 0.254 nm correspond to Au(111) and
Au2O3(131) planes, respectively. Also, the presence of Au2O3 in
the catalyst is verified by the FT identification, with the spots at
0.274, 0.276 and 0.254 nm corresponding to (4, 2, 0), (3, ꢀ1, ꢀ1),
and (ꢀ1, ꢀ3, 1) planes of Au2O3, respectively. For the
3.4Au/0.5CuATiO2 catalyst, Au and CuO particles are observed.
The presence of Au and Cu2O is also detected in the 4.6Au/
0.9CuATiO2. And the Cu2O is verified by the FT spots at 0.243,
0.243, and 0.212 nm, which correspond to (1, 1, 1), (ꢀ1, 1, 1),
and (2, 0, 0) planes of Cu2O, respectively. For the 5.0Au/3.9CuATiO2
catalyst, in addition to Au2O3 and CuO, AuCu alloy is also detected,
with the FT spots at 0.226, 0.239, and 0.164 nm corresponding to
(0, 2, 0), (2, 0, 1), and (2, ꢀ2, 1) planes of AuCu alloy. Note that
the lattice fringe of Au(111) plane (0.236 nm) is very close to that
Fig. 1 shows the XRD patterns of the catalysts. For the fresh cat-
alysts (Fig. 1a), it is found that all these catalysts show typical TiO2
Table 1
Physical properties of various Au catalysts.
Catalyst
Content (wt.%)
SBET
Average Au
particle size (nm)
(m2 gꢀ1
)
Au
Cu
Cl
2.1Au/TiO2
2.10
2.41
3.43
4.64
4.60
4.93
5.05
–
–
Trace
Trace
Trace
0.02
0.03
0.06
0.09
0
48
47
45
42
42
40
39
53
165
2.5
2.7
3.0
3.3
3.3
3.8
4.2
–
2.4Au/0.1CuATiO2
3.4Au/0.5CuATiO2
4.6Au/0.9CuATiO2
4.6Au/1.6CuATiO2
4.9Au/2.1CuATiO2
5.0Au/3.9CuATiO2
1.0CuATiO2
0.14
0.53
0.88
1.62
2.09
3.92
1.00
0.88
4.9Au/0.8CuAAl2O3
4.95
0.06
n.d.

142
C.-X. Sun et al. / Journal of Catalysis 312 (2014) 139–151
Fig. 2. HRTEM and FT images of various catalysts.
of the AuCu(111) plane; thus, the possibility of the presence of
AuCu alloy in other catalysts could not be excluded. Concerning
the Au particle size, it seems that the particle size of Au slightly in-
creases with increasing Cu content in the catalysts, ranging from
2.5 nm for the 2.1Au/TiO2 to 4.2 nm for the 5.0Au/3.9CuATiO2
catalyst (Table 1).
species (i.e., Cu2+), then the proportion of the oxidized Cu species
in contact with Au species in each catalyst could be roughly
estimated. Based on the peak areas of the reduction peaks at
3.2. Reduction properties
Fig. 3 demonstrates H2-TPR profiles of the catalysts. For the
1.0CuATiO2, two reduction peaks at 250 and 280 °C are observed,
which could be assigned to typical reduction of CuO with different
sizes [33]. For the 2.1Au/TiO2 and 2.4Au/0.1CuATiO2 catalysts, no
obvious reduction peaks are detected, implying that the Au species
in the catalysts are in forms of metallic Au0 or the oxidized Au spe-
cies could be readily reduced at very low temperature before the
profiles were recorded [34]. With increasing Cu content in the cat-
alyst, weak reduction peaks at 120 and 300 °C are observed, and
both peaks progressively increase in their intensities. The low-tem-
perature reduction peak (at 100 °C) could be attributed to the
reduction in oxidized Cu species in close contact with Au, which
is probably due to a typical spillover effect [35]; while the peak
at 300 °C could be assigned to the reduction isolated Cu oxides,
but it may also contain contribution of the partial reduction of
the TiO2 support due to spillover effect. These results suggest that
part of the oxidized Cu species is in intimate contact with the Au
species, and other oxidized Cu species are isolated on the TiO2 sup-
port. In addition, the progressive shift to higher temperature of the
reduction peak at 100 °C with increasing Cu loading implies that
the interaction between Au and Cu species is less intense in the
high Cu loading sample (i.e., 5.0Au/3.9CuATiO2) than that in the
low Cu loading one (i.e., 4.6Au/1.6CuATiO2). Moreover, if we
assume that the peaks are all due to the reduction in oxidized Cu
1.0Cu-TiO2
5.0Au/3.9Cu-TiO2
4.9Au/2.1Cu-TiO2
4.6Au/1.6Cu-TiO2
x 6
4.6Au/0.9Cu-TiO2
x 6
x 3
x 3
3.4Au/0.5Cu-TiO2
2.4Au/0.1Cu-TiO2
2.1Au/TiO2
100
200
300
400
500
Temperature / oC
Fig. 3. H2-TPR profiles of various catalysts.

C.-X. Sun et al. / Journal of Catalysis 312 (2014) 139–151
143
100–140 and about 300 °C, the proportions of the Cu species in
contact with Au in the 3.4Au/0.5CuATiO2, 4.6Au/0.9CuATiO2,
4.6Au/1.6CuATiO2, 4.9Au/2.1CuATiO2 and 5.0Au/3.9CuATiO2 cata-
lysts are 28.3%, 34.4%, 81.0%, 81.4%, and 68.9%, respectively. It can
be seen that this value first increases and then slightly decreases
with Cu content in the catalyst. This is understandable because
with high coverage of Cu species on the support (high Cu content),
Au cations could be selectively deposited on these Cu species
during the DP process due to its higher IEP value (8.5–9.5 for
CuO) compared to that of TiO2 (4.0–5.0). But excessive Cu content
(i.e., 5.0Au/3.9CuATiO2) would certainly result in isolated Cu
species on the support as the catalyst contains more Cu than
level could be deconvoluted into three components centered at
84.0, 84.6, and 86.3 eV, which could be assigned to Au0, Au+ and
Au3+, respectively [36]. Note that the spectrum of the 2.1Au/TiO2
catalyst contains only two components (Au0 and Au+), while the
others contain three components (Au0, Au+, and Au3+). Also, it is
found that the binding energy (BE) of Au0 in the 2.1Au/TiO2 is
83.6 eV, while those of the Cu-containing catalysts (3.4Au/
0.5CuATiO2, 4.6Au/0.9CuATiO2, and 4.9Au/2.1CuATiO2) gradually
shift to higher values (Table 2). Similar trend is also found in the
BEs of Au3+, namely the BE of Au3+ in the 3.4Au/0.5CuATiO2 is
85.9 eV, while that in the 4.9Au/2.1CuATiO2 is 86.4 eV. Moreover,
the analysis of surface composition of Au species reveals that the
2.1Au/TiO2 catalyst contains almost equal contents of Au0 and
Au+, but the Cu-containing catalysts contain less Au0 and more
Au3+. For example, the 2.1Au/TiO2 catalyst contains 49.6% surface
Au0 and 50.4% Au+, while the concentrations of Au0, Au+, and
Au3+ on the 4.9Au/2.1CuATiO2 catalyst are 30.0%, 37.9%, and
33.1%, respectively. Cu 2p3/2 spectra of the catalysts are shown in
Fig. 4b. The main peak could be deconvoluted into two peaks
centered at 932.8 and 933.8 eV, which could be assigned to Cu+
ꢀ1
ꢀ1
Au (0:254 mmolAu
3.9CuATiO2).
g
and 0:609 mmolCu
g
in the 5.0Au/
cat
cat
3.3. Oxidation states analysis
Oxidation states of the surface species are analyzed by XPS, as
shown in Fig. 4 and the detailed results are summarized in Table 2.
For the Au 4f spectra (Fig. 4a), the main peaks of the Au 4f7/2 core
(a) Au 4f
(b) Cu 2p 3/2
(c) Cu LMM
4.9Au/2.1Cu-TiO2
4.9Au/2.1Cu-TiO2
4.9Au/2.1Cu-TiO2
4.6Au/0.9Cu-TiO2
3.4Au/0.5Cu-TiO2
1.0Cu-TiO2
4.6Au/0.9Cu-TiO2
4.6Au/0.9Cu-TiO2
3.4Au/0.5Cu-TiO2
2.1Au/TiO2
3.4Au/0.5Cu-TiO2
1.0Cu-TiO2
80
82
84
86
88
90
92
94 928 930 932 934 936 938 940 942 944 946 948 906 908 910 912 914 916 918 920 922 924
Kinetic energy / eV
Binfing energy / eV
Binding energy / eV
Fig. 4. (a) Au 4f, (b) Cu 2p core level and (c) Cu LMM Auger spectra of Au catalysts.
Table 2
Binding energies and surface compositions of Au and Cu species in catalysts.
Catalyst
Au 4f7/2 (molar content (%))
Cu 2p3/2 (molar content (%))
Surface composition
Au0
Au+
Au3+
Cu0+Cu+
Cu2+
Au/Ti
Au/Cu
2.1Au/TiO2
1.0Cu/TiO2
83.6 (49.6)
–
84.8 (50.4)
–
84.6 (30.0)
84.6 (37.0)
84.8 (37.9)
–
–
– (–)
–
–
–
0.0088
–
–
–
932.6 (32.1)
932.8 (72.2)
932.8 (76.8)
932.7 (25.4)
932.3 (73.3)
932.3 (91.0)
934.9 (67.9)
934.1 (27.8)
934.3 (23.2)
933.8 (74.6)
934.5 (26.7)
934.4 (9.0)
3.4Au/0.5CuATiO2
4.6Au/0.9CuATiO2
4.9Au/2.1CuATiO2
4.6Au/0.9CuATiO2 spent
4.6Au/0.9CuATiO2-300
83.8 (50.7)
84.1 (30.1)
84.3 (30.0)
84.1 (93.0)
84.0 (100.0)
85.9 (19.3)
86.3 (32.9)
86.4 (33.1)
86.1 (7.0)
–
0.055
0.077
0.067
0.139
0.041
1.29
1.04
0.41
1.47
0.55

144
C.-X. Sun et al. / Journal of Catalysis 312 (2014) 139–151
and Cu2+, respectively [37]. Also, the oxidized Cu species could be
identified by the presence of satellite peaks at 941.7 and
944.1 eV [38]. Moreover, surface composition analyses reveal that
the addition of Au on the CuATiO2 support significantly reduces
the amount of Cu2+ species, as the 1.0CuATiO2 contains 32.1%
Cu+ and 67.9% Cu2+ while the 4.6Au/0.9CuATiO2 catalyst contains
76.8% Cu+ and 23.2% Cu2+. But further increasing Cu content in
the catalyst (e.g., 4.9Au/2.1CuATiO2) results in the lower Cu+ and
higher Cu2+ surface contents. Since the Cu 2p3/2 peaks of Cu0
and Cu+ species are very close, Auger lines of these Cu-containing
samples were also investigated (Fig. 4c). Peaks at kinetic energies
of 916.8, 917.7, and 918.6 eV could be assigned to Cu+, Cu2+, and
Cu0, respectively. Note that the peaks in Fig. 4c are broad and a
quantitative analysis is difficult; however, it can be seen that the
1.0CuATiO2 contains mostly oxidized Cu species (Cu+ and Cu2+),
while the 3.4Au/0.5CuATiO2 and 4.6Au/0.9CuATiO2 catalysts pos-
sibly contain both metallic and oxidized Cu species. For the high
Cu loading sample (4.9Au/2.1CuATiO2), the peak maximum
slightly shifts to lower energy compared to those of the
3.4Au/0.5CuATiO2 and 4.6Au/0.9CuATiO2, indicating that the pro-
portion of oxidized Cu species in this sample could be higher than
those in the latter ones. The presence of Au and Cu species in var-
ious oxidation states is consistent with the results reported by Duh
et al. [34] on a series of Au/CuOxATiO2 catalysts prepared in a sim-
ilar manner. In addition, surface compositions of the catalysts were
also measured and the results are listed in Table 2. The surface
molar ratio of Au/Ti generally increases with Au content in the cat-
alyst. The surface molar ratios of Au/Cu in the 3.4Au/0.5CuATiO2,
4.6Au/0.9CuATiO2 and 4.9Au/2.1CuATiO2 catalysts are 1.29, 1.04,
and 0.41, respectively, which are lower than the corresponding
bulk Au/Cu ratios (2.21 for the 3.4Au/0.5CuATiO2, 1.66 for the
4.6Au/0.9CuATiO2, and 0.76 for the 4.9Au/2.1CuATiO2). The combi-
nation of the XPS results clearly suggests that the Au supported on
CuATiO2 has a strong interaction between Au and Cu species,
which leads to the changes in their electronic properties. With
the presence of cationic Cu species in the catalyst, charges on Au
species transfer to the oxidized Cun+ (n = 1 or 2) species because
the electronic negativity of Cu+ (5.04) or Cu2+ (10.28) is higher than
that of Au0 (2.54). Similar results were also observed in the litera-
ture. Duh et al. [34] found that the BE of Cu 2p3/2 shifted from 934.0
to 933.1 eV after Au deposition on a Au/1%CuOxATiO2 catalyst,
while the BE of Au 4f7/2 shifted from 84.2 eV on a Au/TiO2 catalyst
to 84.6 eV on the Au/1%CuOxATiO2.
100
80
60
40
20
0
(b)
4.6Au/0.9Cu-TiO2
4.3Au/0.8Ag-TiO2
4.5Au/0.7Fe-TiO2
4.4Au/0.7Ni-TiO2
4.7Au/0.9Zn-TiO2
2.1Au/TiO2
0
100
200
300
400
500
600
6
5
4
3
2
1
0
(a)
4.6Au/0.9Cu-TiO2
4.3Au/0.8Ag-TiO2
4.5Au/0.7Fe-TiO2
4.4Au/0.7Ni-TiO2
4.7Au/0.9Zn-TiO2
2.1Au/TiO2
0
100
200
300
400
500
600
Reaction time / min
Fig. 5. TFP epoxidation over Au catalysts modified with different metal oxides.
Reaction conditions: T = 300 °C; P = 0.1 MPa; N2/N2O/TFP = 28:3.5:3.5 ml minꢀ1
.
100
(b)
80
2.4Au/0.1Cu-TiO2
3.4Au/0.5Cu-TiO2
4.6Au/0.9Cu-TiO2
4.6Au/1.6Cu-TiO2
4.9Au/2.1Cu-TiO2
5.0Au/3.9Cu-TiO2
2.1Au/TiO2
60
40
20
1.0Cu-TiO2
3.4. Epoxidation of TFP
0
7
6
5
4
3
2
1
0
0
100
200
300
400
500
600
First of all, direct epoxidation of TFP was carried out on several
Au/TiO2 catalysts modified with different metal oxides and the re-
sults are shown in Fig. 5. The detected main carbon-containing
products are TFPO, trifluoropropionaldehyde (TFPA), and CO2. The
2.1Au/TiO2 catalyst has a steady-state TFP conversion of 0.8% at
the reaction temperature of 300 °C. The modification of catalyst
with Ag, Fe, or Ni does not have positive effect on the activity,
while the addition of Zn and Cu in the Au/TiO2 catalyst (4.7Au/
0.9ZnATiO2 and 4.6Au/0.9CuATiO2) promotes the catalytic activ-
ity. Particularly, the 4.6Au/0.9CuATiO2 catalyst has a steady-state
TFP conversion of 2.4%, which is three times as high as that ob-
tained on the un-promoted 2.1Au/TiO2. Therefore, hereafter, we fo-
cused on the Cu-promoted catalysts.
2.4Au/0.1Cu-TiO2
3.4Au/0.5Cu-TiO2
4.6Au/0.9Cu-TiO2
4.6Au/1.6Cu-TiO2
4.9Au/2.1Cu-TiO2
5.0Au/3.9Cu-TiO2
2.1Au/TiO2
(a)
1.0Cu-TiO2
0
100
200
300
400
500
600
Fig. 6 shows the TFP epoxidation over the Au/CuATiO2 catalysts
with various Cu contents and detailed results are listed in Table 3.
From Fig. 6, it is found that the initial TFP conversions on the cat-
alysts are up to 4.1%, but they gradually declines, implying the
deactivation of the catalysts. Among the catalysts, the Au-free
1.0CuATiO2 catalyst is almost inactive for this reaction. With the
addition of Cu in the Au/TiO2 catalyst, the TFP conversion increases,
Reaction time / min
Fig. 6. TFP epoxidation over Cu-modified Au catalysts. Reaction conditions:
T = 300 °C; P = 0.1 MPa; N2/N2O/TFP = 28:3.5:3.5 ml minꢀ1
.
but high content of Cu leads to a decline in the activity. For exam-
ple, the 2.4Au/0.1CuATiO2 catalyst has a steady-state TFP conver-

C.-X. Sun et al. / Journal of Catalysis 312 (2014) 139–151
145
Table 3
Summary of reaction results over various catalysts.
c
Catalyst
Conv.a (%)
TFP
Selectivity (%)
TFPO formation rate
TOFinit. (ꢂ10ꢀ3 sꢀ1
)
N2O
TFPO
TFPA
CO2
N2Ob
g
ꢁ hꢀ1 ꢁ kgꢀ1
g
ꢁ hꢀ1 ꢁ gꢀ1
ꢀ1
ꢀ1
cat
TFPO
TFPO
Au
2.1Au/TiO2
0.8
1.2
1.9
2.4
1.4
0.8
0.5
0.3
0.2
4.1
5.2
6.8
7.1
5.4
3.2
2.8
1.8
2.0
80.0
76.0
81.0
88.0
80.0
78.0
65.0
40.0
23.1
3.0
10.0
8.0
4.0
8.0
10.0
20.0
45.0
49.0
17.0
14.0
11.0
8.0
12.0
12.0
15.0
15.0
27.9
15.6
17.5
22.6
29.7
20.7
19.5
11.6
6.7
22.1
39.9
53.8
72.4
37.0
21.8
10.2
4.3
1.05
1.33
1.58
1.56
0.80
0.44
0.20
–
3.25
1.72
5.21
5.72
2.15
0.99
1.22
–
2.4Au/0.1CuATiO2
3.4Au/0.5CuATiO2
4.6Au/0.9CuATiO2
4.6Au/1.6CuATiO2
4.9Au/2.1CuATiO2
5.0Au/3.9CuATiO2
1.0CuATiO2
4.9Au/0.8CuAAl2O3
2.3
1.6
0.033
Reaction conditions: T = 300 °C; P = 0.1 MPa; N2/N2O/TFP = 28:3.5:3.5 ml minꢀ1; TFP: trifluoropropylene; TFPO: trifluoropropylene oxide; TFPA: trifluoropropionaldehyde.
a
Data obtained at steady state (reaction time = 300 min).
b
N2O selectivity (%) = 100 ꢂ TFP conv. ꢂ Selectivity to TFPO/N2O conv.
c
Calculated based on initial TFP conversion in Fig. 6 (reaction time = 30 min) and Au dispersion (Au particle size obtained by TEM).
sion of 1.2%, while that on the 5.0Au/3.9CuATiO2 is 0.5%. In addi-
tion, the optimal Cu content seems to be 0.9% as the highest stea-
dy-state TFP conversion (2.4%) is obtained on the 4.6Au/
0.9CuATiO2 catalyst. Also, it is found that the conversion of N2O
for each catalyst is much higher than the conversion of TFP
(Table 3), probably due to the fact that the combustion of one
TFP molecule to CO2 and H2O will consume about seven N2O mol-
ecules (CF3CHCH2 + 7N2O ? 3CO2 + H2O + F2 + HF) and N2O could
decompose to N2 and O2 (2N2O ? 2N2 + O2). Consequently, the
selectivity to N2O (defined as moles of produced TFPO/moles of
consumed N2O) is less than 30% (Table 3). Concerning the selectiv-
ity, the 1.0CuATiO2 sample has a low selectivity to TFPO (40%), but
the addition of Cu significantly enhances the selectivity to TFPO. In
general, the selectivity to TFPO over the Cu-containing catalysts is
65–88%, depending on the Cu content in the catalyst. The highest
TFPO selectivity (88%) is also obtained on the 4.6Au/0.9CuATiO2
catalyst. The calculation of TFPO formation rate (Table 3) shows
that the maximum yield of TFPO is achꢀie1ved on the 4.6Au/
much lower than those on the 4.6Au/0.9CuATiO2, which is due to
the higher Au content (4.6%) in the 4.6Au/0.9CuATiO2 catalyst
compared to the Cu-free 2.1Au/TiO2 (2.1%). Also, the intensities
of the signals are less intense over the 5.0Au/3.9CuATiO2 catalyst.
Peak areas of the N2 and N2O signals in the spectra of the
1.0CuATiO2, 2.1Au/TiO2, 4.6Au/0.9CuATiO2 and 5.0Au/3.9CuATiO2
catalysts are 1.52 ꢂ 10ꢀ8
,
1.53 ꢂ 10ꢀ8
,
6.37 ꢂ 10ꢀ8
,
and
5.04 ꢂ 10ꢀ8, respectively, indicating that the amounts of adsorbed
N2O differ on the catalysts, with the 4.6Au/0.9CuATiO2 possessing
the highest value. Also, TFP–TPD experiments (Fig. 7b) show that
TFP decomposes to CO2 and H2O at two temperature regions of
350–550 °C and 600–750 °C, probably due to the sequential oxida-
tion of different intermediates. Meanwhile, the intensities of the
product signals are higher on the 4.6Au/0.9CuATiO2 catalyst than
on the 1.0CuATiO2, 2.1Au/TiO2 or 5.0Au/3.9CuATiO2 samples.
Interestingly, the products observed during the TFP–TPD process
are mainly CO2 and H2O, and no distinct organic products are de-
tected. Since there is no N2O present in the TFP–TPD process, the
results indicate that the lattice oxygen in the metal oxides (AuOx,
CuOx, and TiO2) could non-selectively react with the TFP molecules
and produce CO2 and H2O, due to the fact that these oxygen species
are nucleophilic (O2ꢀ). In contrast, the reaction between TFP and
N2O (Figs. 5 and 6) clearly produces TFPO, revealing that the
importance of the presence of N2O in the feedstock because it
can decomposes on metal sites to produce N2 and oxygen species
with a mild electrophilic character [20] which could attack C@C
bond to produce epoxide. The TPD results suggest that the adsorp-
tions of N2O and TFP are more pronounced on the Cu-promoted
4.6Au/0.9CuATiO2 than on the 1.0CuATiO2 or 2.1Au/TiO2 counter-
part, most probably due to the more active sites in the former sam-
ple because of its higher Au content in this sample (but with
slightly larger Au particle size). Therefore, the enhanced TFPO for-
mation on the Cu-promoted catalysts could be contributed to the
higher Au contents in these catalysts compared to the bare Au/
TiO2 due to a higher Au capture efficiency during the preparation
process (DP method) and to small Au particle sizes in these sample
(2–3 nm). However, the samples with high Cu contents contain lar-
ger Au particles compared to those with low Cu contents (Table 1
and Fig. 2), which leads to the reduction in surface Au atoms and
consequently the activity. This point is also confirmed by the TFPO
formation rates calculated based on the Au content in the catalyst
0.9CuATiO2, with a rate of 72:4 gTFPO hꢀ1 kg
.
cat
The above results indicate that the prepared Au/TiO2 catalysts
are very selective for TFP epoxidation when N2O is used as the oxi-
dant, although the activities are not high (with TFP conversions
less than 3%). The highest steady-state TFPO formation rate ob-
tained on the 4.6Au/0.9CuATiO2 is 72:4 gTFPO hꢀ1 kgꢀ1, corre-
cat
hꢀ1. This rate is comparable to
ꢀ1
sponding to 0:65 mmolTFPO
g
cat
the results reported in the literature for propylene epoxidation
over AuACu/TiO2 alloy catalysts (0:36—1:96 mmol gcat hꢀ1) [30],
although the catalysts in the current work and in the literature
were prepared in different manners. It is clear that the Au-free
sample 1.0CuATiO2 is not active for TFP epoxidation (Table 3)
while the 2.1Au/TiO2 catalyst is more active than the 1.0CuATiO2,
which implies that the presence of Au in the catalyst plays very
important role in the reaction. Also, the addition of small amount
of Cu promotes the activity. The promoting effect of Cu in propyl-
ene epoxidation has been discussed in the literature. For example,
Llorca et al. [31] compared activities of propylene epoxidation with
N2O over Au1Cu3/TiO2 and Au/TiO2 catalysts with similar Au con-
tents (1.14–1.23 wt.%) and Au particle sizes (5.1–5.8 nm), and the
authors found that the former was much more active than the lat-
ter, due to the site isolation effects by alloying.
In order to investigate the catalytic behaviors of the catalysts in
the current work, temperature-programmed desorption of N2O and
TFP was conducted on several catalysts and the results are shown
in Fig. 7. In Fig. 7a, it can be seen that N2O molecules could adsorb
on all samples, and N2O decomposition starts at about 300 °C to
form N2 and O2 following the reaction 2N2O ? 2N2 + O2. However,
the signal intensities on the 1.0CuATiO2 or 2.1Au/TiO2 catalyst are
ꢀ1
ꢀ1
(Table 3). For example, the formation rate in unit of g
hꢀ1
g
TFPO
Au
on the 2.1Au/TiO2 is 1.05, which is comparable to those obtained
on the Au/CuATiO2 catalysts with small amount of Cu (1.10–
ꢀ1
ꢀ1
1.56 g
hꢀ1
g
). However, heavy loading of Cu (Cu content
TFPO
Au
>0.9%) in the catalyst causes a decline in activity, as the 4.9Au/
2.1CuATiO2 and 5.0Au/3.9CuATiO2 have formation rates of 0.44

146
C.-X. Sun et al. / Journal of Catalysis 312 (2014) 139–151
(b) TFP
5.0Au/3.9Cu-TiO2
(a) N2O
N2
5.0Au/3.9Cu-TiO2
CO2
H2O
HF
TFPA
TFPO
TFP
O2
N2O
4.6Au/0.9Cu-TiO2
2.1Au/TiO2
4.6Au/0.9Cu-TiO2
2.1Au/TiO2
CO2
H2O
N2
HF
O
TFPA
TFPO
TFP
N22O
CO2
H O
N2
H2F
TFPA
O2
N2O
TFPO
TFP
1.0Cu-TiO2
1.0Cu-TiO2
CO2
N2
O2
N2O
H2O
HF
TFPA
TFPO
TFP
100
200
300
400
500
600
700
800
100
200
300
400
500
600
700
800
Temperature / oC
Temperature / oC
Fig. 7. Mass signals of (a) N2Oꢀ and (b) TFP–TPD on various catalysts.
hꢀ1
g
, respectively, which is most likely due to the
using N2O as the oxidant. Despite the complexity of the catalyst sys-
tem, Llorca et al. [31] concluded that the active sites of the AuACu/
TiO2 catalysts could be located at the AuACu alloy and TiO2 interface
and the catalytic performance of the catalyst was influenced by var-
ious parameters such as metal–support interaction, Au particle size,
and decoration of Cu on Au surface. In the current work, AuCu alloy
is observed in some catalysts (Fig. 2), and the intimate contact be-
tween Au and Cu is evidenced by the H2-TPR (Fig. 3) and XPS
(Fig. 4) results. Although a solid conclusion on the active sites for
TFP epoxidation cannot be reached at the current stage, a few
remarks can still be made. The first one is the importance of the
presence of Au in the catalyst. As the Au-containing catalysts are
more active and selective than the Au-free 1.0CuATiO2 (Table 3),
it suggests that the Au species are crucial for this reaction. The sec-
ond one is the influence of the support. A reference Au catalyst con-
ꢀ1
ꢀ1
and 0.20 g
TFPO
Au
growth of Au particles in these samples. Similar conclusion has
been also reported in the case of propylene epoxidation over
AuACu/TiO2 with N2O. Llorca et al. [31] calcined a Au1Cu3/TiO2 cat-
alyst at different temperatures (573–873 K) and tested for propyl-
ene epoxidation with N2O. The authors found that the catalyst
calcined at 673 K had the best performance (propylene conver-
sion = 4.2%), while the catalysts calcined at 873 K had a inferior
performance with a propylene conversion of 2.5%, which was due
to the progressive growth of Au particles upon calcination, chang-
ing from 5.7 nm at 673 K to 9.8 nm at 873 K.
Note that the Au contents in the catalysts are different, and it
is better to compare the activities of the catalysts in terms of
turnover frequencies (TOFs). The TOF values are calculated and
summarized in Table 3. Note that the TOFs are initial (taken at
reaction time of 30 min) to avoid possible aggregation of Au par-
ticles during the reaction (especially at high temperature i.e.,
300 °C). Also, the TOFs values (0.99–5.72 ꢂ 10ꢀ3 sꢀ1) are compa-
rable to the results reported by Llorca et al. [31] for propylene
epoxidation over various alloyed AuCu/TiO2 catalysts using
N2O as the oxidant. Under similar reaction conditions
(propylene/N2O molar ratio = 1/1, reaction temperature = 300 °C),
the authors reported TOFs of 0.8–4.5 ꢂ 10ꢀ3 sꢀ1 on the catalysts
calcined at 300 °C, although the particle sizes of those AuACu
alloys were slightly larger (2.0–5.3 nm) than those in the current
work. In addition, it is found that the TOF values based on Au
dispersions in the catalysts first increase and then decline
dramatically (Table 3), with the highest value of 5.72 ꢂ 10ꢀ3 sꢀ1
obtained on the 4.6Au/0.9CuATiO2 catalyst. These results suggest
that the reaction is structure-sensitive, and the optimal Au parti-
cle size for this reaction is about 3 nm, which is consistent with
the conclusion obtained on the Au/TiO2 catalysts for propylene
epoxidation [7].
taining 4.9 wt.% of Au and 0.8 wt.% of Cu supported on
c-Al2O3
(denoted as 4.9Au/0.8CuAAl2O3) was prepared and tested for TFP
epoxidation under the same conditions (Table 3). It is obvious that
the 4.9Au/0.8CuAAl2O3 catalyst shows much lower TFP conversion
(0.2%) compared to the 4.6Au/0.9CuATiO2. Moreover, the product
distributions on these two catalysts are completely different. For
the 4.9Au/0.8CuAAl2O3 catalyst, the selectivity to the desired TFPO
is about 23.1%, and the main products are TFPA (selectivity = 49.0%)
and CO2 (selectivity = 27.9%). The different behaviors on these two
catalysts clearly indicate the importance of the presence of TiO2 in
the catalyst. The third one is the presence of Cu in the Au catalyst.
Note that the initial TOF values of the 3.4Au/0.5CuATiO2 and
4.6Au/0.9CuATiO2 catalysts are higher than that of the Cu-free
2.1Au/TiO2, suggesting that proper interaction between Au and Cu
might lead to geometry and/or electronic properties of the metal
species which is helpful to the epoxidation. Based on these findings,
therefore, it is reasonable to believe that the complex synergy
created via the interaction of AuACuATiO2 is important for the
epoxidation capability and the different initial TOF values on the
catalysts might be due to the subtle changes in such interaction,
although deep investigations of structural properties of these cata-
lysts are required to obtain a concrete conclusion.
The different initial TOF values of the catalysts may reflect the
differences in the active sites of the catalysts for TFP epoxidation.
The nature of the active sites/phases of the AuACu/TiO2 systems
has been discussed in the literature for direct propylene epoxidation

C.-X. Sun et al. / Journal of Catalysis 312 (2014) 139–151
147
Another worth noting issue is that the high selectivity to TFPO
obtained on the catalysts in the current work. The TFPO selectivi-
ties are 65–88% for the Au-containing catalysts in the present
work, which are much higher than those values for propylene
epoxidation over AuACu/TiO2 catalysts (10–50%) [30,31]. The main
by-products in those works [30,31] were acrolein and COx (CO and
CO2). The formation of acrolein in propylene oxidation is a result of
nucleophilic oxygen reacting with the allylic hydrogen of CH3
group in propylene molecule, and the formation of COx is due to
the strong electrophilic oxygen reacting with C@C bond or further
combustion of organic intermediates [20]. Other than propylene,
the CH3 group is replaced by a CF3 group in TFP molecule, and
the pathway of reaction between oxygen species and allylic hydro-
gen is blocked; thus, high selectivity to TFPO could be achieved.
Although high selectivity to TFPO is generally achieved on the
Cu-promoted Au/TiO2 catalysts, heavy loading of Cu obviously sup-
presses the TFPO selectivity, as the 3.4Au/0.5CuATiO2 has a TFPO
selectivity of 81% while the 5.0Au/3.9CuATiO2 has a selectivity of
only 65%. Also, the 1.0CuATiO2 catalyst has an even lower TFPO
selectivity (40%). It has been discussed that Cu could effectively en-
hance the PO selectivity in propylene epoxidation. Generally, it was
reported that only the Cu species in low oxidation states (Cu0 or
Cu+) were active for PO formation using molecular O2 as the oxi-
dant [16,18,39,40], while high-valent Cu species (e.g., Cu2+) usually
lead to the formation of acrolein and COx [16,41]. In the present
work, oxidation states of Cu and Au in the catalysts were analyzed
by XPS (Fig. 4 and Table 2). It is found that the 1.0CuATiO2 and
5.0Au/3.9CuATiO2 samples contain 67.9% and 74.6% of Cu2+ spe-
cies, respectively, while those with low Cu contents have less than
30% of Cu2+. The main by-products obtained on the 1.0CuATiO2
and 5.0Au/3.9CuATiO2 samples are TFPA (due the reaction of oxy-
gen species with hydrogen in the @CH2 group) and CO2 (due to
combustion of organic intermediates). For the 2.1Au/TiO2 and the
low Cu content Au/CuATiO2 samples, the TFPO becomes dominant.
Thus, it appears that that the low valent Cu species (Cu+) are ben-
eficial to the enhancement of TFPO selectivity, but high-valent Cu
species such as Cu2+ are detrimental to the TFPO selectivity, which
is in good agreement with the conclusions in propylene
epoxidation.
as a reason for the alternation of product selectivity because the
catalytic performance is extremely sensitive to the Au particle size.
In propylene epoxidation with H2 and O2 over Au catalysts, it has
been recognized that Au particles with a size of 2 nm were the
most effective for the PO formation [7]. Propylene epoxidation
with N2O also has a strong particle size dependency. For example,
Llorca et al. [31] compared Au1Cu3/TiO2 catalyst calcined at
573–873 K and found the optimal product selectivities were
obtained on the catalyst calcined at 673 K with a mean Au particle
size of about 5.7 nm (PO selectivity = 52.2%; acrolein selectiv-
ity = 18.0%; COx selectivity = 19.6%). When the catalyst was
calcined at 873 K, the Au particles became larger (9.8 nm), and con-
sequently the selectivity to PO significantly decreased to 39.9%,
100
(b)
Sel. to CO2
80
2.1Au/TiO2
4.6Au/0.9Cu-TiO2
60
40
20
0
Sel. to CO2
Sel. to TFPO
Sel. to TFPO
0
100
200
300
400
500
600
25
(a)
20
15
10
5
2.1Au/TiO2
4.6Au/0.9Cu/TiO2
Moreover, as the N2O decomposes on the metal sites, and the
electronic property of the formed oxygen species by N2O decompo-
sition is crucial to the resulting products, which may also alter due
to the interaction between Au and Cu. In propylene molecule, the
CH3-group donates electron to C@C bond and the charge density
of C@C could be increased, which needs mild electrophilic oxygen
species in order to form PO while the oxygen species with strong
electrophilic feature would react with C@C bond to produce other
oxygenates and COx. However, as the CF3-group in TFP molecule
could strongly attract charge from C@C bond, therefore oxygen
species with strong electrophilic nature are required to react with
the charge-deficient C@C bond to form epoxide. The electronic nat-
ure of the oxygen species could be tuned by the catalyst through
the interaction between the oxygen species and the metal sites
in the catalyst. In the current work, the XPS results (Fig. 4 and
Table 2) clearly shows shifts in the BEs of Au and Cu species,
depending on the Cu contents in the catalysts. BEs of Au species
shift to higher values with the addition of Cu (Table 2), suggesting
the charge transferring from Au to Cu. Consequently, the oxygen
species adsorbed on the charge-deficient Au species could be more
electrophilic, which could preferably attack the C@C bond to form
epoxide. As the highest TFPO selectivity (88%) is obtained on the
4.6Au/0.9CuATiO2 catalyst, it seems that the oxygen species with
finely adjusted charge density are suitable for TFP epoxidation,
although more careful surface analyses of the catalysts are re-
quired in order to reach a concrete conclusion. On the other hand,
changes in Au particle sizes in the catalysts may also be concerned
0
0
100
200
300
400
500
600
Reaction time / min
Fig. 8. Effect of oxidant on catalytic performance of 4.6Au/0.9CuATiO2 catalyst.
Reaction conditions: T = 300 °C; P = 0.1 MPa; N2/O2/TFP = 28:3.5:3.5 ml minꢀ1
.
7
100
6
80
5
4.6Au/0.9Cu-TiO2
4
3
2
1
0
60
40
20
0
4.6Au/0.9Cu-TiO2-300
0
100
200
300
400
500
600
Reaction time/min
Fig. 9. Effect of calcination on catalytic performance of 4.6Au/0.9CuATiO2 catalyst.
Reaction conditions: T = 300 °C; P = 0.1 MPa; N2/N2O/TFP = 28:3.5:3.5 ml minꢀ1
.

148
C.-X. Sun et al. / Journal of Catalysis 312 (2014) 139–151
while the selectivities of acrolein and COx increased to 22.4% and
30.1%, respectively. In our case, the TFPO selectivities over the Au
catalysts are virtually same (about 80%) except for the 5.0Au/
3.9CuATiO2 catalyst (65%). Note that the former catalysts have
smaller Au particles than the 5.0Au/3.9CuATiO2 catalyst, therefore,
it seems that the large Au particles in the 5.0Au/3.9CuATiO2 cata-
lyst might be another reason for the loss of selectivity to TFPO.
possible. To investigate the effect of oxidant, TFP epoxidation
was carried out using molecular O2 as the oxidant under the same
reaction conditions. As shown in Fig. 8, under the reaction condi-
tions (T = 300 °C; P = 0.1 MPa; N2/O2/TFP = 28/3.5/3.5 ml minꢀ1),
the tested catalysts (2.1Au/TiO2 and 4.6Au/0.9CuATiO2) exhibit
completely different behaviors compared to those with N2O as
the oxidant (Fig. 6). Using molecular O2 as the oxidant, the TFP con-
versions (Fig. 8a) on the two catalysts are much higher than those
using N2O as the oxidant, particularly for the Cu-modified 4.6Au/
0.9CuATiO2 catalyst (15.1% using O2 vs 2.4% using N2O). However,
the selectivities to TFPO are much lower when molecular O2 is used
as the oxidant (Fig. 8b), with a TFPO selectivity of about 10%
3.5. Effects of oxidant and calcination on catalytic performance
Since the N2O oxidant could decompose to N2 and O2 following
the reaction 2N2O ? 2N2 + O2, a reaction between TFP and O2 is
4.6Au/0.9Cu-TiO2 fresh
Au 4f
Cu 2p 3/2
Cu LMM
4.6Au/0.9Cu-TiO2-300 fresh
Au 4f
Cu 2p 3/2
Cu LMM
x 2
80
82
84
86
88
90
92
94 928 930 932 934 936 938 940 942 944 946 948 906 908 910 912 914 916 918 920 922 924
Binding energy / eV
Kinetic energy / eV
Binding energy / eV
Fig. 10. Comparison of TEM images and XPS spectra of 4.6Au/0.9CuATiO2 and 4.6Au/0.9CuATiO2-300 catalysts.

C.-X. Sun et al. / Journal of Catalysis 312 (2014) 139–151
149
obtained on the 4.6Au/0.9CuATiO2 catalyst and a CO2 selectivity of
about 80% (the other by-product is TFPA). The O2 conversions on
the 2.1Au/TiO2 and 4.6Au/09CuATiO2 catalysts are estimated to
be 13.5 and 74.8%, respectively, which are much higher than those
of N2O when N2O is used as the oxidant (Table 3). These results im-
ply that the reaction between TFP and O2 is much easier compared
to the reaction between TFP and N2O (high TFP conversion), but the
use of molecular O2 as the oxidant leads to combustion of TFP to
produce CO2 (low TFPO selectivity). This is because the oxygen spe-
cies produced by molecular O2 activation on transition metal oxi-
des (such as CuO) are nucleophilic (such as O2ꢀ), and could react
with the olefin molecules to produce organic by-products and fur-
ther oxidize them to CO2; while the oxygen species formed by N2O
decomposition have electrophilic nature which could attack C@C
bond of the olefin molecule to produce epoxide [20]. Moreover,
comparison of the behaviors of the 2.1Au/TiO2 and 4.6Au/
0.9CuATiO2 catalysts shows that the TFP conversion is lower on
the 2.1Au/TiO2 (5.2%) than that on the 4.6Au/0.9CuATiO2 (15.1%)
but the TFPO selectivity is higher on the former (37%) than that
on the latter (10%), which suggests that molecular O2 activated
on Au is relatively selective for epoxide formation but O2 molecules
activated on Cu species are non-selective.
In a previous work by Llorca et al. [31], the AuACu/TiO2 cata-
lysts were calcined at different temperatures (573–873 K) for pro-
pylene epoxidation using N2O as the oxidant, and it was found that
the calcination treatment exerted significant impact on the cata-
lytic performance (i.e., propylene conversion and PO selectivity).
In the current work, the Au/CuATiO2 catalysts were not calcined,
therefore, it is worthwhile to investigate the effect of calcination
on the catalytic behaviors of the catalysts. We calcined the best-
performed catalyst (4.6Au/0.9CuATiO2) at 300 °C for 4 h in air (de-
noted as 4.6Au/0.9CuATiO2-300) and tested it for TFP epoxidation
under the same conditions. As can be seen in Fig. 9, it is found that
the calcined sample (4.6Au/0.9CuATiO2-300) has slightly higher
TFPO selectivity (about 92%) than the 4.6Au/0.9CuATiO2 (88%),
but the former has a lower steady-state TFP conversion (1.9%) com-
pared to the latter (2.4%). To investigate the different structural
properties of these two catalysts, TEM images and XPS spectra of
the catalysts were compared, as shown in Fig. 10. The presence
of AuCu alloy and Cu oxide in the 4.6Au/0.9CuATiO2-300 is verified
by the lattice fringes measurements and FT images. For example,
the FT spots at 0.198, 0.223 and 0.223 nm correspond to (ꢀ2, 0,
0), (1, 1, 1) and (ꢀ1, 1, 1) planes of AuCu alloy, respectively. Also,
it seems that the calcination treatment leads to the growth of me-
tal particles. The XPS spectra of these two catalysts (Fig. 10 and
Table 2) also clearly show different features. It is found that after
calcination at 300 °C, the 4.6Au/0.9CuATiO2-300 contains mostly
Au0 and low valent Cu species (Cu0 and Cu+), indicating that the
calcination treatment of the catalyst leads to reduction in oxidized
metal species in the catalyst. The presence of low valent Cu species
are beneficial to the epoxide selectivity [16,18,39,40], which could
explain the higher TFPO selectivity on the 4.6Au/0.9CuATiO2-300
catalyst. However, the surface Au/Cu ratio on the 4.6Au/
0.9CuATiO2-300 is 0.55, which is much lower than that on the
4.6Au/0.9CuATiO2 (1.04). The lowered Au/Cu ratio in the 4.6Au/
0.9CuATiO2-300 catalyst is caused by either the growth of Au par-
ticles or the covering of some Au entities by Cu species during the
calcination process. The decoration of Cu on the Au is also con-
firmed by the H2-TPR results (Fig. 11). The reduction peak at
280 °C in the 4.6Au/0.9CuATiO2 obviously shifts to lower temper-
ature (175 °C) in the 4.6Au/0.9CuATiO2-300, implying a strong
interaction between Au and Cu is generated after calcination, prob-
ably through a decorating effect or the formation of AuCu alloy.
The coverage of Au entities by Cu enrichment was reported in a
previous work. Llorca et al. [31] in the alloyed AuACu/TiO2 system
that surface enrichment of Cu progressively occurred during the
334
280
475
240
175
spent 4.6Au/0.9Cu-TiO2
4.6Au/0.9Cu-TiO2
131
126
254
4.6Au/0.9Cu-TiO2-300
100
200
300
400
500
600
700
Temperature / oC
Fig. 11. H2-TPR profiles of fresh and spent 4.6Au/0.9CuATiO2 and 4.6Au/
0.9CuATiO2-300 catalysts.
catalyst calcination, which accounted for the decrease in catalyst
activity. Therefore, in this work, the inferior performance of the
4.6Au/0.9CuATiO2-300 could be attributed to the growth of Au
particles, or more importantly to the structural changes in the cat-
alyst, i.e., decoration of Cu on the Au surface.
3.6. Catalyst deactivation
The catalysts employed in the current suffer deactivation during
the reaction, which is commonly observed in the Au catalysts for
epoxidation [7]. In direct propylene epoxidation, one main reason
for the deactivation of Au catalysts is the irreversible adsorption
of PO on the catalyst surface. Mul et al. [42] studied the propylene
epoxidation reaction using C3H6/H2/O2/He gas mixture and the
adsorption of PO on Au/TiO2 and Au/TiO2 (1.6 wt.%)/SiO2 catalysts.
These authors observed that PO adsorbed irreversibly on both cat-
alysts resulting in the formation of bidentate propoxy moieties.
The strong adsorption of the propylene oxide product on adjacent
Ti sites to form a bidentate species results in the eventual occupa-
tion of the sites responsible for epoxidation. In order to investigate
the surface species on the catalyst, FTIR spectra of the fresh and
spent 4.6Au/0.9CuATiO2 catalysts are compared, as well as TFPO
adsorbed on the fresh sample as shown in Fig. 12. The TFPO ad-
sorbed 4.6Au/0.9CuATiO2 catalyst contains feature bands at 1177
and 1272 cmꢀ1, which could be assigned to asymmetric and
fresh 4.6Au/0.9Cu-TiO2
5
spent 4.6Au/0.9Cu-TiO2
fresh 4.6Au/0.9Cu-TiO2 + TFPO
1000 1200 1400 1600
2800 3000 3200 3400
Wavenumber / cm-1
Fig. 12. FTIR spectra of (a) fresh and (b) spent 4.6Au/0.9CuATiO2 catalysts.

150
C.-X. Sun et al. / Journal of Catalysis 312 (2014) 139–151
symmetric vibration of CAF band in the TFPO molecule, respec-
tively [43]. However, no similar bands are observed on the spent
sample, and its spectrum is almost identical to that of the fresh
4.6Au/0.9CuATiO2 catalyst, suggesting that no organic compounds
adsorb on the catalyst surface during the reaction. Another key fact
for catalyst deactivation is the growth of Au particles during the
reaction [11]. In the current work, comparison of the TEM images
of the fresh and spent 4.6Au/0.9CuATiO2 catalysts was made and
shown in Fig. 13. For the fresh catalyst, the mean Au particle size
is about 3 nm, while obvious growth of Au particles is observed
in the spent sample, with a mean Au particle size of 4–6 nm. The
small Au particles are believed to be active for propylene epoxida-
tion [7], while the growth of Au particles certainly causes the
reduction in the amount of active sites and consequently suppress
the activity. Also, a comparison of the XPS spectra of the fresh and
spent 4.6Au/0.9CuATiO2 was made (Fig. 13). It is found that the
spent catalyst contains mostly metallic Au0, suggesting that the
oxidized Au species decompose to metallic Au during the reaction
at high temperature, which is consistent with the results of the cal-
cined sample (4.6Au/0.9CuATiO2-300, Fig. 10). Also, the spent cat-
alyst contains slightly lower content of low valent Cu species
(Cu0 + Cu+ = 73.3%) compared to the fresh one (Cu0 + Cu+ = 76.8%),
4.6Au/0.9Cu-TiO2
Au 4f
Cu 2p3/2
Cu LMM
4.6Au/0.9Cu-TiO2 spent
Au 4f
Cu LMM
Cu 2p 3/2
80
82
84
86
88
90
92
94 928 930 932 934 936 938 940 942 944 946 948 906 908 910 912 914 916 918 920 922 924
Binding energy / eV
Binding energy / eV
Kinetic energy / eV
Fig. 13. TEM images and XPS spectra of fresh and spent 4.6Au/0.9CuATiO2 catalysts.

C.-X. Sun et al. / Journal of Catalysis 312 (2014) 139–151
151
indicating that some low valent Cu species are oxidized during the
References
reaction. Moreover, one distinct difference is found, that is, the sur-
face Au/Cu ratio in the spent sample (Au/Cu = 1.47, Table 2) is
much higher than that in the fresh one (Au/Cu = 1.04, Table 2). This
difference suggests that segregation of Au and Cu occurs during the
reaction. Such phenomenon is also observed on the 4.6Au/
0.9CuATiO2-300 catalyst. The spent 4.6Au/0.9CuATiO2-300 has a
surface Au/Cu ratio of 1.15, which is much higher than that on
the fresh 4.6Au/0.9CuATiO2-300 (0.55). The segregation of Au
and Cu is also confirmed by a comparison of the H2-TPR results
of the fresh and spent 4.6Au/0.9CuATiO2 catalyst (Fig. 11). The
spent 4.6Au/0.9CuATiO2 catalyst shows two weak reduction peaks
at 240 and 334 °C, and a strong reduction peak at 475 °C. The peaks
at 240 and 334 °C could be assigned to the reduction in oxidized Cu
species and the peak at 475 °C could be assigned to the reduction
of TiOF2. Compared to the fresh 4.6Au/0.9CuATiO2 catalyst, the ab-
sence of the reduction peak at 131 °C in the spent sample implies
that the intimate contact between Au and Cu disappears after reac-
tion, which could lead to the reduction in the amount of the active
sites (probably at AuACuATiO2 interface) and thus the activity. It is
worth noting that the calcination treatment leads to strong inter-
action between Au and Cu as in the 4.6Au/0.9CuATiO2-300 cata-
lyst, while phase segregation occurs during the reaction,
probably due to the different gas environments (calcination in air
and reaction in N2O + TFP). Therefore, we believe that the catalyst
deactivation is possibly caused by the structural change in the cat-
alyst, i.e., the growth of Au particles at high temperature and loss of
AuACuATiO2 interface due to the segregation of Cu during the
reaction.
[1] T. Katagiri, K. Uneyama, J. Fluor. Chem. 105 (2000) 285.
[2] K. Xu, Chem. Rev. 104 (2004) 4303.
[3] C.G. Krespan, US Patent 5101058, 1992, assigned to E. I. Du Pont de Nemours
and Company.
[4] J. Lu, M. Luo, H. Lei, C. Li, Appl. Catal. A: Gen. 237 (2002) 11.
[5] A. Palermo, A. Husain, M. Tikhov, R. Lambert, J. Catal. 207 (2002) 331.
[6] J. Lu, J.J. Bravo-Suárez, A. Takahashi, M. Haruta, S.T. Oyama, J. Catal. 232 (2005)
85.
[7] T. Hayashi, K. Tanaka, M. Haruta, J. Catal. 178 (1998) 566.
[8] T.A. Nijhuis, T. Visser, B.M. Weckhuysen, J. Phys. Chem. B 109 (2005) 19309.
[9] A.K. Sinha, S. Seelan, S. Tsubota, M. Haruta, Angew. Chem. Int. Ed. 43 (2004)
1546.
[10] B. Chowdhury, J.J. Bravo-Suárez, M. Daté, S. Tsubota, M. Haruta, Angew. Chem.
Int. Ed. 45 (2006) 412.
[11] N. Yap, R.P. Andres, W.N. Delgass, J. Catal. 226 (2004) 156.
[12] L. Cumaranatunge, W.N. Delgass, J. Catal. 232 (2005) 38.
[13] J. Lu, X. Zhang, J.J. Bravo-Suárez, K.K. Bando, T. Fujitani, S.T. Oyama, J. Catal. 250
(2007) 350.
[14] J. Lu, X. Zhang, J.J. Bravo- Suárez, T. Fujitani, S.T. Oyama, Catal. Today 147
(2009) 186.
[15] T. Liu, P. Hacarlioglu, S.T. Oyama, M.F. Luo, X.R. Pan, J.Q. Lu, J. Catal. 267 (2009)
202.
[16] J. Lu, M. Luo, H. Lei, X. Bao, C. Li, J. Catal. 211 (2002) 552.
[17] O.P.H. Vaughan, G. Kyriakou, N. Macleod, M. Yikhov, R.M. Lambert, J. Catal. 236
(2005) 401.
[18] H. Chu, L. Yang, Q. Zhang, Y. Wang, J. Catal. 241 (2006) 225.
[19] Y. Wang, H. Chu, W. Zhu, Q. Zhang, Catal. Today 131 (2008) 496.
[20] V. Duma, D. Hönicke, J. Catal. 191 (2000) 93.
[21] J.J. Bravo-Suárez, K.K. Bando, J. Lu, M. Haruta, T. Fujitani, S.T. Oyama, J. Phys.
Chem. C 112 (2008) 1115.
[22] E. Ananieva, A. Reitzmann, Chem. Eng. Sci. 59 (2004) 5509.
[23] X. Wang, Q. Zhang, S. Yang, Y. Wang, J. Phys. Chem. B 109 (2005) 23500.
[24] Q. Zhang, Q. Guo, X. Wang, T. Shishido, Y. Wang, J. Catal. 239 (2006) 105.
[25] A. Costine, T. O’Sullivan, B.K. Hodnett, Catal. Today 112 (2006) 103.
[26] Y. Wang, W. Yang, L. Yang, X. Wang, Q. Zhang, Catal. Today 117 (2006) 156.
[27] T. Thömmes, S. Zürcher, A. Wix, A. Reitzmann, B. Kraushaar-Czarnetzki, Appl.
Catal. A Gen. 318 (2007) 160.
[28] B. Moens, H.D. Winne, S. Corthals, H. Poelman, R. De Gryse, V. Meynen, P. Cool,
B.F. Sels, P.A. Jacobs, J. Catal. 247 (2007) 86.
[29] S. Yang, W. Zhu, Q. Zhang, Y. Wang, J. Catal. 254 (2008) 251.
[30] R.J. Chimentão, F. Medina, J.L.G. Fierro, J. Llorca, J. Mol. Catal. A: Chem. 274
(2007) 159.
[31] J. Llorca, M. Domínguez, C. Ledesma, R.J. Chimentão, F. Medina, J. Sueiras, I.
Angurell, M. Seco, O. Rossell, J. Catal. 258 (2008) 187.
[32] Z. Huang, Z. Wang, K. Lv, Y. Zheng, K. Deng, ACS Appl. Mater. Interfaces 5
(2013) 8663.
[33] M.F. Luo, J.M. Ma, J.Q. Lu, Y.P. Song, Y.J. Wang, J. Catal. 246 (2007) 52.
[34] F.C. Duh, D.S. Lee, Y.W. Chen, Mod. Res. Catal. 2 (2013) 1.
[35] R. Kramer, M. Andre, J. Catal. 58 (1979) 287–295.
[36] Q. Fu, H. Saltsburg, M. Flytzani-Stephanopoulos, Science 301 (2003) 935.
[37] M.P. Casaletto, A. Longo, A. Martorana, A. Prestianni, A.M. Venezia, Surf.
Interface Anal. 38 (2006) 215–218.
4. Conclusion
This work demonstrates that Au/TiO2 catalysts are effective for
direct gas-phase epoxidation of TFP with N2O as the oxidant. With
the addition of Cu, the catalytic performance could be enhanced.
The promoted activity on these catalysts could be ascribed to the
higher Au contents in the catalyst by Cu addition and more impor-
tantly the complicated synergy between AuACuATiO2. Also, it is
found that the addition of Cu improves the selectivity to TFPO,
mainly due to the presence of the low valent Cu+ species, while
the presence of high-valent Cu2+ species results in decline of
selectivity.
[38] K.S. Kim, J. Electron. Spectrosc. Relat. Phenom. 3 (1974) 217.
[39] D. Torres, N. Lopez, F. Illas, R.M. Lambert, Angew. Chem. Int. Ed. 46 (2007)
2055.
[40] W. Su, S. Wang, P. Ying, Z. Feng, C. Li, J. Catal. 268 (2009) 165.
[41] J.R. Monnier, G.W. Hartley, J. Catal. 203 (2001) 253.
[42] G. Mul, A. Zwijnenburg, B. van der Linden, M. Makkee, J.A. Moulijn, J. Catal. 201
(2001) 128.
Acknowledgment
[43] Y. Peng, B. Sun, P. Wu, Appl. Spectrosc. 62 (2008) 295.
This work was financially supported by Natural Science Foun-
dation of China (Grant No. 20703039).