Gas-phase epoxidation of propylene by molecular oxygen over Ag-CuCl2/BaCO3 catalyst with low CuCl2 doping: Catalytic performance, deactivation and regeneration

Article abstract of DOI:10.1016/j.molcata.2016.08.019

Ag-MCl_x/BaCO₃ catalysts with different chloride promoters, prepared by reduction-deposition-impregnation method, were investigated for gas-phase epoxidation of propylene to propylene oxide (PO) by molecular oxygen. Ag-CuCl₂/BaCO₃ catalyst with 360?ppm of Cu and 400?ppm of Cl exhibits the best initial catalytic performance, in which PO selectivity of 71.2% and propylene conversion of 1.3% are achieved, but only PO selectivity of 13.9% is obtained at propylene conversion of 3.2% after reaction for 500?min. The catalytic reaction mechanism over Ag-CuCl₂/BaCO₃ catalyst follows Rideal-Eley mechanism, in which propylene in the gas phase reacts with molecular oxygen species adsorbed on the surface of Ag at the interface in close contact with CuCl₂ to produce PO, and with atomic oxygen species adsorbed on the surface of Ag nanoparticles to produce CO₂ and H₂O. One oxygen atom of molecular oxygen species reacts with propylene to form a PO molecule, and the left insufficient oxygen atoms react with propylene to produce oxygen-containing intermediates and then to form coke deposition which covers the active sites and thus results in the catalyst deactivation. The deactivated Ag-CuCl₂/BaCO₃ catalyst can be completely regenerated by combustion of coke deposition and then impregnation with appropriate amount of Cl.

Full text of DOI:10.1016/j.molcata.2016.08.019

Accepted Manuscript
Title: Gas-phase epoxidation of propylene by molecular
oxygen over Ag-CuCl2/BaCO3 catalyst with low CuCl2
doping: Catalytic performance, deactivation and regeneration
Author: Qing Zhang Guangtao Chai Yanglong Guo
Wangcheng Zhan Yun Guo Li Wang Yunsong Wang
Guanzhong Lu
PII:
S1381-1169(16)30343-0
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.molcata.2016.08.019
MOLCAA 10006
To appear in:
Journal of Molecular Catalysis A: Chemical
Received date:
Revised date:
Accepted date:
14-7-2016
13-8-2016
16-8-2016
Please cite this article as: Qing Zhang, Guangtao Chai, Yanglong Guo, Wangcheng
Zhan, Yun Guo, Li Wang, Yunsong Wang, Guanzhong Lu, Gas-phase epoxidation
of propylene by molecular oxygen over Ag-CuCl2/BaCO3 catalyst with low CuCl2
doping: Catalytic performance, deactivation and regeneration, Journal of Molecular
Catalysis A: Chemical http://dx.doi.org/10.1016/j.molcata.2016.08.019
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

<AT>Gas-phase epoxidation of propylene by molecular oxygen over Ag-CuCl2/BaCO3
catalyst with low CuCl2 doping: catalytic performance, deactivation and regeneration
<AU>Qing Zhang, Guangtao Chai, Yanglong Guo*
##Email##ylguo@ecust.edu.cn##/Email##, Wangcheng Zhan, Yun Guo, Li Wang, Yunsong
Wang, Guanzhong Lu* ##Email##gzhlu@ecust.edu.cn##/Email##
<AFF>Key Laboratory for Advanced Materials, Research Institute of Industrial Catalysis,
School of Chemistry and Molecular Engineering, East China University of Science and
Technology, Shanghai 200237, P R China
<PA>Fax: +86 21 64252923.
<ABS-Head><ABS-HEAD>Graphical abstract
<ABS-P>
<ABS-P><xps:span class="xps_Image">fx1</xps:span>
<ABS-HEAD>Highlights►
<remove picture pageno 1>Ag-CuCl2/BaCO3 catalyst was prepared by reduction-deposition-
impregnation method.
<remove picture pageno 1>Ag-CuCl2/BaCO3 catalyst with low CuCl2 doping exhibits better
catalytic performance.
<remove picture pageno 1>Epoxidation of propylene over Ag-CuCl2/BaCO3 catalyst follows
Rideal-Eley mechanism.
<remove picture pageno 1>Molecular oxygen species benefit epoxidation of propylene to
propylene oxide.► Ag-based catalyst is deactivated by coke deposition and can be entirely
regenerated.
<ABS-HEAD>ABSTRACT
<ABS-P>Ag-MClx/BaCO3 catalysts with different chloride promoters, prepared by
reduction-deposition-impregnation method, were investigated for gas-phase epoxidation of
propylene to propylene oxide (PO) by molecular oxygen. Ag-CuCl2/BaCO3 catalyst with 360
ppm of Cu and 400 ppm of Cl exhibits the best initial catalytic performance, in which PO
selectivity of 71.2% and propylene conversion of 1.3% are achieved, but only PO selectivity
of 13.9% is obtained at propylene conversion of 3.2% after reaction for 500 min. The catalytic
reaction mechanism over Ag-CuCl2/BaCO3 catalyst follows Rideal-Eley mechanism, in which
propylene in the gas phase reacts with molecular oxygen species adsorbed on the surface of
Ag at the interface in close contact with CuCl2 to produce PO, and with atomic oxygen
species adsorbed on the surface of Ag nanoparticles to produce CO2 and H2O. One oxygen
atom of molecular oxygen species reacts with propylene to form a PO molecule, and the left
insufficient oxygen atoms react with propylene to produce oxygen-containing intermediates
1

and then to form coke deposition which covers the active sites and thus results in the catalyst
deactivation. The deactivated Ag-CuCl2/BaCO3 catalyst can be completely regenerated by
combustion of coke deposition and then impregnation with appropriate amount of Cl.
<KWD>Keywords: Epoxidation of propylene; Propylene oxide; Molecular oxygen; Ag-based
catalyst; CuCl2 doping.
<H1>1. Introduction
Propylene oxide (PO) is a versatile synthetic intermediate widely used for production of
propylene glycols, polyether polyols, polyurethane, surfactants and other commercial
products. PO is produced mainly by the chlorohydrin process, the Halcon process and the
hydroperoxide process. However, the chlorohydrin process is environmentally unfriendly, the
Halcon process generates the equimolar coproducts which values depend on their demands in
the market [1] and the commercialization of the hydroperoxide process is hindered by the
expense and transport of H2O2 and the commercial installation [2-3]. Though Au-based
catalysts can produce PO with high selectivity by the in-situ production of H2O2 from H2-O2
mixture [4-11], H2 must be used as a sacrificial coreactant and its efficiency is usually low. As
oxygen is a cheap and green oxidant, gas-phase epoxidation of propylene by molecular
oxygen is the most ideal reaction process with high atomic economy, however, remains a
great challenge for researchers to attain higher PO selectivity.
Epoxidation of ethylene by molecular oxygen over Ag-based catalysts has been widely
used in the industrial production of ethylene oxide (EO) with the selectivity of 83-95 % for
many years [12-14], whereas, epoxidation of propylene by molecular oxygen over Ag-based
catalysts is still under investigation, in which unpromoted Ag catalysts are inefficient for PO
formation with PO selectivity of below 10 % [15-18]. Al2O3 is widely used as the support of
Ag-based catalysts for epoxidation of ethylene by molecular oxygen, however, is not an
efficient support for epoxidation of propylene by molecular oxygen, because high surface area
and the acidity of Al2O3 support easily lead to the complete combustion of propylene and PO
[17]. Various promoters, such as K, MoO3, Y2O3 and Cu, have been investigated to improve
2

PO selectivity [19-23]. In the industrial production process of EO, EO selectivity can be
improved by addition of chlorohydrocarbon in the feed gas, which indicates that Cl is benefit
for EO formation. Cl-promoted Ag catalysts were prepared to improve EO selectivity and Cl
effect was investigated [24-26]. In view of the similarity between ethylene and propylene,
both addition of chlorohydrocarbon in the feed gas [27] and modification of various chlorides,
such as NaCl, BaCl2, CsCl and CuCl, enhanced PO selectivity for epoxidation of propylene
by molecular oxygen [28-31]. The promotion of Cl was also investigated for epoxidation of
butadiene [32]. Cl was thought to work best among all halogen elements for its electronic
effect [33]. Cu catalyst was more efficient for epoxide formation than Ag catalyst, because
oxygen species adsorbed on Cu surface tended to attack C=C bond to form epoxide rather
than strip allylic hydrogen leading to combustion of propylene [34-36]. Cu-based catalysts
modified with VOx, RuOx and MnOx exhibited better catalytic performance for epoxidation of
propylene [37-39]. The low valance state of Cu was thought to be more effective for
epoxidation of propylene [40-42].
The reaction mechanism of epoxidation of propylene by molecular oxygen over Ag-based
catalysts has been discussed for many years. Because oxygen species adsorbed on the surface
of Ag catalyst are nucleophilic (i. e. high-basicity oxygen) rather than electrophilic (i. e. low-
basicity oxygen) [36], the markedly basic character of oxygen species leads to the stripping of
allylic hydrogen atoms and then the complete combustion of propylene to CO2, which results
in worse catalytic performance.
BaCO3 with low surface area and medium basicity is chosen as the support of Ag-based
catalysts for gas-phase epoxidation of propylene by molecular oxygen, in which 12.5 % of
propylene conversion and 36.9 % of PO selectivity are achieved [43]. In this work, Ag-
MClx/BaCO3 catalysts with different chloride promoters, prepared by the reduction-
deposition-impregnation method, were investigated for gas-phase epoxidation of propylene by
3

molecular oxygen, in which effects of the chloride promoters, CuCl2 loading, the calcination
temperature, the reaction temperature and gas hourly space velocity were studied. The aim of
this work was to clarify oxygen species and active sites responsible for PO formation over
Ag-CuCl2/BaCO3 catalyst, and the catalytic reaction mechanism for epoxidation of propylene
by molecular oxygen was proposed. Moreover, the catalyst stability, the deactivation
mechanism and the regeneration way of Ag-CuCl2/BaCO3 catalyst were also investigated.
<H1>2. Experimental
<H2>2.1. Preparation of catalysts
Ag/BaCO3 catalyst was prepared by the reduction-deposition method as follows [43]: 8.0 g
of BaCO3 support was firstly immersed into the aqueous solution of ethylene diamine for 12 h
and then dried at 60 °C under vacuum for 4.5 h. After that, 2.0 g of formaldehyde (HCHO)
was dissolved in 40.0 mL of deionized water as the reductant, into which the dried amine-
pretreated BaCO3 support was added to form a slurry in the ice-water bath at 10 oC. After
stirring for 20 min, 0.6 g of AgNO3 dissolved in 40.0 mL of deionized water was added
dropwise into the above slurry, then the solid was filtered and washed with deionized water
and ethanol for several times. The solid was dried overnight at room temperature and then
calcined in N2 atmosphere at 250 oC for 4 h. Ag-CuCl2/BaCO3 catalysts were prepared by the
incipient impregnation of the dried Ag/BaCO3 catalyst without calcination with the aqueous
solution of CuCl2 for 12 h, dried at 60 °C and then calcined in N2 atmosphere at 250 oC for 4
h. The similar procedures were used to prepare Ag-MClx/BaCO3 catalysts modified with other
promoters (NH4Cl, NaCl, KCl, CaCl2 and FeCl3), except for Ag-CuCl/BaCO3 catalyst
prepared by the ball-milling treatment of the dried Ag/BaCO3 catalyst without calcination
with CuCl powder at 800 rpm for 10 h and then calcined in N2 atmosphere at 250 oC for 4 h.
CuCl2/BaCO3 catalyst was prepared by the incipient impregnation of BaCO3 support with the
aqueous solution of CuCl2 for 12 h, dried at 60 °C and then calcined in N2 atmosphere at 250
oC for 4 h.
4

<H2>2.2. Epoxidation of propylene
Epoxidation of propylene was carried out in a fixed-bed quartz reactor under the reaction
conditions of 0.6 g catalyst, 200 °C, 0.1 MPa and GHSV of 3000 h-1. The feed gas consisted
of 20 % of propylene, 10 % of O2 and balance N2. Reaction products were analyzed by two
on-line gas chromatographs equipped with a two packed columns (G.D.X-401 and Porapak
Q) with FID and TCD detectors. All lines between the reactor exit and the gas
chromatographs were heated to 120 °C in order to prevent condensation of products. PO
selectivity and propylene conversion were calculated on a carbon balance basis.
<H2>2.3. Characterization of catalysts
Elemental analysis was done by inductively coupled-plasma atomic emission spectroscopy
(ICP-AES) using a TJA IRIS ADVANTAG 1000 instrument.
XRD patterns were performed on a Bruker AXS D8 Focus diffractometer operated at 40
kV, 40 mA (Cu Kα radiation, λ = 0.15406 nm).
XPS spectra were recorded on a Thermo ESCALAB 250 spectrometer with a
monochromatized AlKα X-ray source (1486.6 eV) and a passing energy of 20 eV. C1s
(binding energy of 284.6 eV) of adventitious carbon was used as the reference.
H2-TPR experiments were carried out in a conventional flow system equipped with a TCD
detector. 500 mg samples of 40-60 mesh were loaded in a quartz reactor and then heated
programmedly from 25 oC to 350 oC at a heating rate of 10 oC·min-1 in a gas flow of H2-N2 (5
vol. % H2) of 45 mL·min-1.
TEM images were recorded on a JEM-2100 transmission electron microscope. The sample
was ultrasonically suspended in the ethanol solvent, and one drop of this slurry was deposited
on a molybdenum grid. The liquid phase was evaporated before the grid was loaded into the
microscope.
5

Temperature programmed desorption of O2 (O2-TPD) and C3H6 (C3H6-TPD), temperature
programmed oxidation of O2 (O2-TPO) and temperature programmed surface reaction of C3H6
(C3H6-TPSR) were carried out on a Micromeritics AutoChem 2920 Ⅱ chemisorption
analyzer with a Hiden HPR 20 mass spectrometer. For O2-TPD and C3H6-TPD, 500 mg
samples of 40-60 mesh were loaded in a quartz reactor and pretreated in a He flow of 30
mL·min-1 at 250 oC for 0.5 h. After it was cooled down to room temperature, a gas flow of
O2-He (3 vol. % O2) or a gas flow of C3H6-Ar (1 vol. % C3H6) of 20 mL·min-1 was introduced
into the reactor, the reactor temperature was raised to 200 oC at a rate of 10 oC·min-1 and was
kept at 200 oC for 0.5 h, then cooled down to room temperature. After that, a He flow of 20
mL·min-1 was switched. When the baseline was stable, the reactor temperature was raised to
600 oC at a heating rate of 10 oC·min-1. The MS signals of O2 (m/e = 32) or C3H6 (m/e = 41)
were recorded by the mass spectrometer. For C3H6-TPSR, a gas flow of C3H6-Ar (20 vol. %
C3H6) of 20 mL·min-1 was introduced into the reactor after the pretreatment with a gas flow of
O2-He (3 vol. % O2) of 20 mL·min-1 similar to O2-TPD. The MS signals of PO (m/e = 58) and
CO2 (m/e = 44) were recorded by the mass spectrometer. For O2-TPO experiment, a gas flow
of O2-He (3 vol. % O2) of 10 mL·min-1 was introduced into the quartz reactor loaded with 500
mg samples of 40-60 mesh. The reactor temperature was raised to 450 oC at a heating rate of
10 mL·min-1. The MS signals of O2 (m/e = 32), CO2 (m/e = 44) and H2O (m/e = 18) were
recorded by the mass spectrometer.
<H1>3. Results and discussion
<H2>3.1. Catalytic performance of Ag-MClx/BaCO3 catalysts
Table 1 shows effect of the chloride promoters on the catalytic performance of Ag-
MClx/BaCO3 catalysts for gas-phase epoxidation of propylene by molecular oxygen. As
shown in Table 1, Ag/BaCO3 catalyst without promoter exhibited high propylene conversion
of 12.5 % and medium PO selectivity of 36.9 %, in which CO2 was by-product with the
selectivity of 63.1 %. The modification of Ag/BaCO3 by the chloride promoters shown in
Table 1 all decreased propylene conversion probably because the introduction of chlorides
decreased the number of bare Ag sites [33]. The modification of NH4Cl, NaCl, CaCl2 or FeCl3
decreased propylene conversion significantly to < 1 %, the modification of KCl or CuCl2
6

exhibited propylene conversion of 1-2 %, and relatively high propylene conversion of 5.2 %
was achieved over Ag-CuCl/BaCO3 catalyst. The modification of the chlorides did not exert
positive effects on PO selectivity except for Ag-CuCl2/BaCO3 catalyst. PO selectivity reached
71.2 % at propylene conversion of 1.3 % over Ag-CuCl2/BaCO3 catalyst. Over all catalysts
modified with the chlorides, besides the main by-product of CO2, there was another by-
product of acrolein (PA) with the selectivity of below 1% which could almost be ignored.
Therefore Ag-CuCl2/BaCO3 catalyst exhibited better PO selectivity than Ag/BaCO3 catalyst
despite a decrease in propylene conversion for epoxidation of propylene by molecular oxygen,
which would be mainly investigated in detail.
<H2>3.2. Effect of CuCl2 loading
To gain further insights into the synergistic effect between CuCl2 and Ag, effect of CuCl2
loading on the catalytic performance of Ag-CuCl2/BaCO3 catalyst was investigated for
epoxidation of propylene by molecular oxygen, which is shown in Table 2. Ag loadings of all
Ag-containing catalysts were at the similar values of 4.3 wt. % measured by ICP-AES. As
shown in Table 2, both propylene conversion and PO selectivity were much lower and PA
selectivity was a little higher over CuCl2/BaCO3 catalysts than the corresponding Ag-
CuCl2/BaCO3 catalysts, which indicated that only the combination of Ag and CuCl2 was
effective for PO formation probably due to the significant synergistic effect between them.
Moreover, with an increase in CuCl2 loading from 180 ppm of Cu and 200 ppm of Cl up to
720 ppm of Cu and 800 ppm of Cl over Ag-CuCl2/BaCO3 catalyst, propylene conversion
significantly decreased from 2.9 % to 0.3 %, because Cl could cover and then poison the
active sites of Ag at high levels [33]. As for PO selectivity, Ag-CuCl2/BaCO3 catalyst with
the promoter of 180 ppm of Cu and 200 ppm of Cl obtained PO selectivity of only 15.5 %,
even lower than Ag/BaCO3 catalyst. However, it was interesting that Ag-CuCl2/BaCO3
catalyst with the promoter of 360 ppm of Cu and 400 ppm of Cl obtained PO selectivity of
71.2 %. With further increasing CuCl2 loading, PO selectivity decreased and CO2 selectivity
increased. Turnover frequency (TOF) values of PO formation based on Ag atoms on the
catalyst surface, i. e. moles of PO produced per mole of surface Ag per second, were
calculated [37, 44]. Compared with Ag/BaCO3 catalyst, TOF values of PO formation of Ag-
CuCl2/BaCO3 catalysts decreased largely due to much lower propylene conversion. Among
7

Ag-CuCl2/BaCO3 catalysts with different CuCl2 loadings, Ag-CuCl2/BaCO3 catalyst with 360
ppm of Cu and 400 ppm of Cl achieved the highest TOF value of 2.0*10-4 s-1 of PO
formation. Therefore only the appropriate modification amount of CuCl2 promoter of Ag-
CuCl2/BaCO3 catalyst can improve PO selectivity for epoxidation of propylene by molecular
oxygen.
<H2>3.3. Effect of reaction conditions
Fig. 1 shows effect of the reaction temperature on the catalytic performance of Ag/BaCO3
catalyst and Ag-CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400 ppm of Cl. Over
Ag/BaCO3 catalyst, with increasing the reaction temperature, PO selectivity decreased
gradually from 40.3 % at the reaction temperature of 80 oC to 36.9 % at the reaction
temperature of 200 oC, but propylene conversion increased obviously from 0.4 % at the
reaction temperature of 80 oC to 12.5 % at the reaction temperature of 200 oC. However, over
Ag-CuCl2/BaCO3 catalyst, with an increase in the reaction temperature, PO selectivity
decreased significantly from nearly 100 % at the reaction temperature of 80 oC to 71.2 % at
the reaction temperature of 200 oC and only 47.5 % at the reaction temperature of 240 oC, but
propylene conversion increased gradually, in which propylene conversions were 0.2 %, 1.3 %
and 5.5 %, respectively, at the reaction temperatures of 80 oC, 200 oC and 240 oC. As shown
in Fig. 2, good linear relations between the logarithm of the reaction rate of epoxidation of
propylene and the reciprocal of the reaction temperature were obtained over Ag/BaCO3 and
Ag-CuCl2/BaCO3 catalysts, in which the apparent activation energies were 40.7 KJ·mol-1 and
23.0 KJ·mol-1, respectively. The decrease in the apparent activation energy after modification
of CuCl2 may imply the change of the reaction pathway and the product distribution, which
agrees well with the catalytic performance of Ag/BaCO3 and Ag-CuCl2/BaCO3 catalysts
shown in Table 2.
8

Fig. 3 shows effect of gas hourly space velocity (GHSV) on the catalytic performance of
Ag-CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400 ppm of Cl. A decrease in GHSV from
6000 h-1 to 1500 h-1 increased propylene conversion from 0.4% to 2.0% and PO selectivity
from 70.1% to 75.4 %. Over Ag-CuCl2/BaCO3 catalyst, PO selectivity of 75.4 % and
propylene conversion of 2.0 % at GHSV of 1500 h-1 were better than those at GHSV of 3000
h-1 in which PO selectivity of 71.2 % and propylene conversion of 1.3 % were obtained.
<H2>3.4. Characterization of Ag-CuCl2/BaCO3 catalyst
To gain information about the valence states of Ag, Cu and Cl, XPS spectra of Ag-
CuCl2/BaCO3 catalysts with different CuCl2 loadings were performed. As shown in Fig. 4A,
the satellite peak at 942 eV was not detected over all Ag-CuCl2/BaCO3 catalysts, which
implied the absence of Cu2+ species. The binding energies of Cu 2p3/2 and Cu 2p1/2 of all
catalysts were 933.1-934.8 eV and 952.8-954.5 eV respectively, which indicated the valence
state of Cu was between +1 and +2, and there existed strong interaction between CuCl2 and
Ag. The binding energies of Cu 2p3/2 and Cu 2p1/2 of Ag-CuCl2/BaCO3 catalyst with 360 ppm
of Cu and 400 ppm of Cl were the lowest, 933.1 eV and 952.8 eV respectively, which
indicated that there existed the strongest interaction between Cu and Ag and thus was
beneficial for PO formation. Fig. 4B shows Cl 2p XPS spectra of Ag-CuCl2/BaCO3 catalyst,
in which the binding energy was 198 eV assignable to Cl- species. 45 The peak of Cl 2p of Ag-
CuCl2/BaCO3 catalyst with higher CuCl2 loading was not obvious because of the surface
enrichment of Cu species. As shown in Fig. 4C, the binding energies of Ag 3d3/2 and Ag 3d5/2
of all Ag-CuCl2/BaCO3 catalysts were at the same values of 374.1 eV and 368.1 eV,
respectively, corresponding to Ag0 [46, 47], which indicated that the valence state of Ag
didn’t change with modification of CuCl2 probably because of the low CuCl2 doping. As
reported in the literature, the oxidation state of Ag of Ag-Pt/Al2O3 catalyst didn’t change due
to the very low Pt loading of 500 ppm [48]. Therefore there exists the strongest interactions
9

between Ag and Cu favorable for PO formation over Ag-CuCl2/BaCO3 catalyst with 360 ppm
of Cu and 400 ppm of Cl.
Table 3 shows the surface composition of Ag-CuCl2/BaCO3 catalysts with different CuCl2
loadings determined by XPS. The surface concentration of Ag of Ag/BaCO3 catalyst was 12.4
atom %. Interestingly, the surface concentration of Ag of Ag-CuCl2/BaCO3 catalyst with 180
ppm of Cu and 200 ppm of Cl was 12.5 atom %, which was almost the same as that of
Ag/BaCO3 catalyst. This indicated that 180 ppm of Cu and 200 ppm of Cl were mainly
located on the surface of BaCO3 support rather than the surface of Ag nanoparticles. With
further increasing CuCl2 loading, the surface concentration of Ag decreased gradually, which
indicated that the extra part of Cu and Cl were located on the surface of Ag nanoparticles
besides a part of Cu and Cl located on the surface of BaCO3 support, which agreed well with
H2-TPR results shown in Fig. S2. With increasing CuCl2 loading, the ratio of Ag to Cu
decreased obviously, and the surface concentration of Cu and Cl increased gradually except
for the situation of Cl over Ag-CuCl2/BaCO3 catalysts with higher CuCl2 loadings, in which
Cl was hardly detected because of the surface enrichment of Cu and thus maybe be the reason
for the surface nonstoichiometric ratio of Cu to Cl.
<H2>3.5. Catalytic reaction mechanism
<H3>3.5.1. Reaction kinetics
The reaction mechanism of epoxidation of propylene by molecular oxygen over Ag-
CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400 ppm of Cl was investigated for its best
catalytic performance among all Ag-CuCl2/BaCO3 catalysts with different CuCl2 loadings.
Fig. 5 shows dependence of the reaction rates of epoxidation of C3H6 on the partial pressure
of O2 or the partial pressure of C3H6 at 200 oC. As shown in Fig. 5, the reaction orders of O2
or C3H6 were 0.13 or 0.94, approximately 0 and 1, respectively, that was to say, 푟푐 퐻
=
3
6
푘퐶 퐻 ∙ 푃퐶 퐻 0.94 ∙ 푃푂
≈ 푘퐶 퐻 ∙ 푃퐶 퐻6, which indicated that the reaction process followed
3 6 3
0.13
3
6
3
6
2
Rideal-Eley mechanism, similar to Ag/BaCO3 catalyst [43]. This also implied that the
modification of CuCl2 didn’t change the reaction mechanism.
<H3>3.5.2. C3H6-TPD
A better understanding of the adsorption behavior of the reactants on the catalyst surface is
crucial for gaining insight into the reaction mechanism. Fig. 6 shows C3H6-TPD profiles of
CuCl2/BaCO3, Ag/BaCO3 and Ag-CuCl2/BaCO3 catalysts. As shown in Fig. 6, only a
desorption peak of C3H6 was observed at about -30 oC over all the three catalysts, which
10

indicated that C3H6 couldn’t adsorb on the surface of all the three catalysts at the reaction
temperature of 200 oC and the modification of CuCl2 didn’t change the weak interaction
between C3H6 and Ag surface [49].
<H3>3.5.3. O2-TPD
Fig. 7 shows O2-TPD profiles of Ag-CuCl2/BaCO3 catalysts with different CuCl2 loadings.
As shown in Fig. 7, the desorption peaks of O2 were observed from 100 oC to 320 oC, which
indicated that O2 could adsorb on the catalyst surface at the reaction temperature of 200 oC
and there was strong interaction between O2 and the catalyst surface. This further confirms
that the reaction process is that propylene in the gas phase reacts with oxygen species
adsorbed on the catalyst surface to produce products and thus follows Rideal-Eley
mechanism, which is consistent with the reaction orders shown in Fig. 5.
Oxygen chemisorption on silver surfaces has been studied in great detail over the past two
decades [50-56]. Oxygen species adsorbed on silver surfaces are divided into three categories:
surface chemisorbed molecular oxygen species, surface chemisorbed atomic oxygen species
and subsurface or bulk-dissolved atomic oxygen species. The atomic oxygen species were
considered as the active oxygen species for epoxidation of both ethylene and propylene in
some literatures [13, 46, 57-58]. However, there is a different viewpoint in this work. As
shown in Fig. 7, there were three types of the overlapping peaks of O2-TPD from 50 oC to 600
oC over Ag-CuCl2/BaCO3 catalysts with CuCl2 loadings not more than 360 ppm of Cu and
400 ppm of Cl. Based on O2-TPD profiles of an electrolytic silver catalyst reported in the
literature [52], these three peaks of O2-TPD were attributed to desorption of three types of
oxygen species from the catalyst surface. The peak of O2-TPD at 138-150 oC was attributed to
desorption of chemisorbed molecular oxygen species. The peak of O2-TPD at 222-241 oC was
attributed to desorption of chemisorbed atomic oxygen species. The peak of O2-TPD at 359-
424 oC arose from segregation and desorption of bulk-dissolved (subsurface) oxygen species.
11

Because the reaction temperature was 200 oC, the former two oxygen species were discussed
as follows.
Over Ag/BaCO3 catalyst, the desorption temperatures of molecular oxygen species and
atomic oxygen species were 150 oC and 222 oC, respectively. When Ag/BaCO3 catalyst was
modified with 180 ppm of Cu and 200 ppm of Cl, the desorption peak of molecular oxygen
species shifted to low temperature at 140 oC, and the desorption peak of the atomic oxygen
species shifted to high temperature at 225 oC. When Ag/BaCO3 catalyst was modified with
360 ppm of Cu and 400 ppm of Cl, the desorption temperature of molecular oxygen species
also shifted to lower temperature at 138 oC, and the desorption peak of atomic oxygen species
also shifted to higher temperature at 241 oC with the significantly decreased intensity. With a
further increase in CuCl2 loading not less than 540 ppm of Cu and 600 ppm of Cl, the O2
uptakes decreased obviously and O2 almost couldn’t adsorb on the catalyst surface. As shown
in Fig. 7 and Table 2, the appropriate CuCl2 loading of Ag-CuCl2/BaCO3 catalyst can
decrease the desorption temperature of molecular oxygen species and increase the desorption
temperature of atomic oxygen species, which is in favor of epoxidation of propylene by
molecular oxygen. As shown in Table 2, PO selectivity was still low (15.5 %) over Ag-
CuCl2/BaCO3 catalyst with 180 ppm of Cu and 200 ppm of Cl, although there were molecular
oxygen species adsorbed on the catalyst surface shown in Fig. 7. As shown in Fig. 7, a strong
desorption peak of atomic oxygen species was observed over this catalyst with the same
intensity as Ag/BaCO3 catalyst, which indicated that propylene preferred to react with atomic
oxygen species to produce a large amount of CO2 and PO formation was low leading to low
PO selectivity. When Ag/BaCO3 catalyst was modified with 360 ppm of Cu and 400 ppm of
Cl, the intensity of the desorption peak of atomic oxygen species decreased significantly, thus
propylene tended to react with molecular oxygen species to produce a large quantity of PO
leading to high PO selectivity. An increase in CuCl2 loading greatly suppresses the
12

dissociative adsorption of oxygen on the catalyst surface. Herein, molecular oxygen species
are necessary but not sufficient for high PO selectivity. Less dissociative adsorption of
oxygen is also required. High PO selectivity needs the balance between electrophilic and
nucleophilic oxygen species, that is to say, molecular oxygen species and atomic oxygen
species respectively in this work, which is also thought to be related to an increase in EO
selectivity [26].
Fig. 8 shows O2-TPD profiles of Ag-MClx/BaCO3 catalysts. A very interesting
phenomenon was observed that molecular oxygen species disappeared only with the
coexistence of atomic oxygen species and subsurface oxygen species and desorption of
atomic oxygen species shifted to lower temperature over all Ag-MClx/BaCO3 catalysts except
for Ag-CuCl2/BaCO3 catalyst. Combined with the catalytic performance of epoxidation of
propylene by molecular oxygen over Ag-MClx/BaCO3 catalysts shown in Table 1, only
modification of CuCl2 enhanced PO selectivity compared with Ag/BaCO3 catalyst, which
indicated that PO selectivity was higher over Ag-based catalysts with molecular oxygen
species adsorbed on the catalyst surface and molecular oxygen species were the active oxygen
species reacting with propylene to produce PO and atomic oxygen species were the active
oxygen species reacting with propylene to produce CO2.
Table 4 shows the catalytic performance of epoxidation of propylene by molecular oxygen
over Ag/BaCO3 catalysts modified with Cu, Cl or CuCl2, Ag/BaCO3 catalyst and
CuCl2/BaCO3 catalyst. Compared with Ag/BaCO3 catalyst, over both Ag-Cu/BaCO3 and Ag-
Cl/BaCO3 catalysts, propylene conversion and PO selectivity decreased significantly, which
indicated that the modification of only Cu or Cl couldn’t improve PO selectivity except for
the cooperative modification of Cu and Cl. Fig. 9 shows O2-TPD profiles of Ag/BaCO3
catalysts modified with Cu, Cl or CuCl2, Ag/BaCO3 catalyst and CuCl2/BaCO3 catalyst. The
O2 uptakes over CuCl2/BaCO3 catalyst were small, which indicated oxygen was difficult to
adsorb on the catalyst surface and thus accounted for low propylene conversion. Molecular
oxygen species were difficult to adsorb on the surface of both Ag-Cu/BaCO3 and Ag-
Cl/BaCO3 catalysts, and molecular oxygen species could adsorb on the surface of Ag-
CuCl2/BaCO3 catalyst with the cooperative modification of Cu and Cl. Therefore the
adsorption of molecular oxygen species on the catalyst surface is necessary for achieving high
PO selectivity and molecular oxygen species are the active oxygen species for PO formation.
<H3>3.5.4. C3H6-TPSR
13

To investigate the reactivity of molecular oxygen species and atomic oxygen species with
propylene, C3H6-TPSR experiments were performed. Fig. 10 shows the products formed
during C3H6-TPSR experiments over Ag-CuCl2/BaCO3 catalysts with different CuCl2
loadings. The main products are PO (m/e = 58) and CO2 (m/e = 44) detected by mass
spectrometer. As shown in Fig. 10A, there were a weak peak of PO at ~80 oC over Ag-
CuCl2/BaCO3 catalyst with 180 ppm of Cu and 200 ppm of Cl and a weak peak of PO at ~100
oC over Ag-CuCl2/BaCO3 catalyst with 540 ppm of Cu and 600 ppm of Cl, but a sharp peak
of PO at ~90 oC over Ag-CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400 ppm of Cl. The
temperature of PO formation was below the desorption temperature of molecular oxygen
species, which confirmed that PO was produced by propylene reacting with molecular oxygen
species. No PO peaks were observed over Ag/BaCO3 catalyst and Ag-CuCl2/BaCO3 catalyst
with 720 ppm of Cu and 800 ppm of Cl. As for the formation of CO2 shown in Fig. 10B, a
sharp peak of CO2 at ~70 oC was observed over Ag/BaCO3 catalyst and this peak became
weaker with increasing CuCl2 loading, which indicated that the modification of CuCl2
suppressed the formation of CO2 from the deep oxidation of PO. Over Ag/BaCO3 and Ag-
CuCl2/BaCO3 catalysts, there was a broad peak of CO2 at ~200 oC and the intensity of this
peak was much higher than that at ~70 oC, which indicated that CO2 was mainly produced by
propylene reacting with atomic oxygen species. As shown in Fig. 7 and Fig. 10, the
appropriate CuCl2 doping of Ag-CuCl2/BaCO3 catalyst can decrease the desorption
temperature of molecular oxygen species in favor of PO formation, and increase the
desorption temperature of atomic oxygen species in favor of depressing combustion of
propylene and deep oxidation of PO, which benefits epoxidation of propylene by molecular
oxygen.
<H3>3.5.5. Catalytic reaction mechanism
14

A schematic diagram of epoxidation of propylene by molecular oxygen over Ag-
CuCl2/BaCO3 catalysts with different CuCl2 loadings was proposed in Fig. 11. As shown in
Table 3, compared with Ag/BaCO3 catalyst, when 180 ppm of Cu and 200 ppm of Cl were
doped into Ag-CuCl2/BaCO3 catalyst, the surface concentration of Ag didn’t change, which
implied that Cu and Cl didn’t locate on the surface of Ag nanoparticles but on the surface of
BaCO3 support. Taking into account O2-TPD profiles shown in Fig. 7, molecular oxygen
species became more active and atomic oxygen species almost didn’t change compared with
Ag/BaCO3 catalyst, we could infer that there was an interaction at the interface between
CuCl2 and Ag located on the surface of BaCO3 support. Meanwhile, oxygen was adsorbed
directly on the surface of Ag at the interface in close contact with CuCl2 to form more active
molecular oxygen species, and oxygen was adsorbed on the surface of Ag nanoparticles and
then dissociated to form atomic oxygen species. Because no Cu or Cl was located on the
surface of Ag nanoparticles, the adsorption of atomic oxygen species didn’t reduce leading to
low PO selectivity over Ag-CuCl2/BaCO3 catalyst with the modification of 180 ppm of Cu
and 200 ppm of Cl. When 360 ppm of Cu and 400 ppm of Cl were doped, the surface
concentration of Ag decreased, which indicated that some CuCl2 were located on the surface
of Ag nanoparticles and depressed the dissociation adsorption of oxygen, and thus molecular
oxygen species almost maintained leading to higher PO selectivity. When more CuCl2 were
doped, more and more surface of Ag nanoparticles and BaCO3 support was covered with
CuCl2, then was difficult to adsorb oxygen to produce molecular oxygen species and atomic
oxygen species, which resulted in lower propylene conversion and PO selectivity. Molecular
oxygen species adsorbed at the interface between Ag nanoparticles and BaCO3 support are
responsible for PO formation, which probably can explain the size effect of Ag nanoparticles
over Ag/BaCO3 catalyst for epoxidation of propylene by molecular oxygen [43]. Therefore
the positive effects of the appropriate modification of CuCl2 are twofold to improve PO
15

selectivity: some parts of CuCl2 make adsorbed molecular oxygen species more active to
produce PO easily, and other parts of CuCl2 depress the dissociation adsorption of oxygen to
atomic oxygen species leading to combustion of propylene to CO2.
<H2>3.6. Stability of Ag-CuCl2/BaCO3 catalysts
Fig. 12 shows the stability of Ag-CuCl2/BaCO3 catalysts and CuCl2/BaCO3 catalysts with
different CuCl2 loadings. Over the four CuCl2/BaCO3 catalysts, PO selectivity kept nearly
constant, although propylene conversion increased to a maximum at time on stream of 50 h
and then decreased gradually with an increase in time on stream. Among the four Ag-
CuCl2/BaCO3 catalysts, Ag-CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400 ppm of Cl
exhibited the best initial catalytic performance with propylene conversion of 1.3 % and PO
selectivity of 71.2 %, however, with increasing time on stream, PO selectivity decreased
rapidly from 71.2 % to 17.4 % at the first 200 min and then retained almost stable from 200
min to 500 min, meanwhile, propylene conversion increased slowly. PO selectivity was 13.9
% at propylene conversion of 3.2 % after reaction for 500 min over Ag-CuCl2/BaCO3 catalyst
with 360 ppm of Cu and 400 ppm of Cl. Over Ag-CuCl2/BaCO3 catalysts with higher CuCl2
loadings, the changing trend of PO selectivity was similar to that of Ag-CuCl2/BaCO3 catalyst
with 360 ppm of Cu and 400 ppm of Cl, in which PO selectivity decreased significantly to
nearly 15 % and propylene conversion decreased to almost zero. Interestingly, Ag-
CuCl2/BaCO3 catalyst with 180 ppm of Cu and 200 ppm of Cl maintained both propylene
conversion and PO selectivity, although the initial catalytic performance was low, in which
PO selectivity was about 14 % at propylene conversion of about 4 %. This catalyst showed a
good stability but was not efficient for epoxidation of propylene by molecular oxygen.
<H2>3.7. Deactivation of Ag-CuCl2/BaCO3 catalyst
<H3>3.7.1. Deactivation of the catalyst
16

Considering that Ag-CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400 ppm of Cl had the
best initial catalytic performance and the deactivation trends of Ag-CuCl2/BaCO3 catalyst
with higher CuCl2 loadings were similar, the deactivation of Ag-CuCl2/BaCO3 catalyst with
360 ppm of Cu and 400 ppm of Cl after reaction for 500 min was mainly studied. The main
reasons of deactivation are poisoning, fouling, thermal degradation (sintering or evaporation)
initiated by the high temperature, mechanical damage and leaching [59]. Table 5 compares the
catalytic performance and physico-chemical properties of the fresh and the deactivated Ag-
CuCl2/BaCO3 catalysts. As shown in Table 5, Ag and Cu contents of the fresh and deactivated
catalysts determined by ICP-AES didn’t change and were 4.3 wt. % and 350 ppm,
respectively, which indicated that the catalyst didn’t lose the active components in the course
of the deactivation. Epoxidation of propylene by molecular oxygen over Ag-based catalyst is
a size-dependent reaction and the particle size of Ag nanoparticles has an effect on the
catalytic performance [29, 60-61]. The results show that Ag nanoparticles didn’t grow up
during the deactivation procedure, namely, 23.6 nm after reaction for 500 min compared with
23.1 nm of the fresh catalyst, which indicated that Ag nanoparticles existed stable on BaCO3
support and the aggregation didn’t happen. As shown in Fig. S3, no obvious changes were
observed from TEM images of the fresh and the deactivated catalysts.
To investigate whether the reactants and the products had a negative effect on Ag-
CuCl2/BaCO3 catalyst, the catalytic performance of the fresh catalyst pretreated with reactants
and products independently at 200 oC was evaluated and shown in Table 6. As shown in
Table 6, the catalytic performance of Ag-CuCl2/BaCO3 catalyst after the pretreatments almost
didn’t become worse, which indicated that the reactants and the products had little effect on
the catalyst except for the pretreatment with PO-N2. Both propylene conversion and PO
selectivity decreased a little after the pretreatment with PO-N2 for 0.5 h, more obviously after
the pretreatment for 2 h. PO was thought to adsorb easily and irreversibly on the surface of
17

Au-based catalysts which led to the worse catalytic performance [62-63]. However, PO
selectivity of 62.1 % after the pretreatment with PO-N2 for 2 h was much higher than PO
selectivity of 22.0 % after reaction for 2 h shown in Fig. 12B, which indicated that PO
adsorption might not be the main reason for the catalyst deactivation and there should be
another important factor responsible for the catalyst deactivation.
<H3>3.7.2. Surface composition of the catalyst
The heterogeneous catalysis is a reaction which occurs on the catalyst surface and the
surface properties of the catalyst are responsible for the catalytic performance. The changes of
the catalyst surface between the fresh and deactivated Ag-CuCl2/BaCO3 catalysts were
investigated by XPS shown in Fig. S4. The binding energies of Ag, Cl and Cu elements
almost didn’t change after reaction for 500 min, which indicated that both the valence states
of the three elements and the interaction between the three elements didn’t change. However,
the peak intensities of Cu 2p (Fig. S4A) and Cl 2p (Fig. S4B) of the deactivated catalyst
declined remarkably compared with the fresh catalyst. The surface contents of Cu and Cl of
the deactivated catalyst shown in Table 7 were 0.8 atom % and 0.5 atom % respectively,
much less than those of the fresh catalyst with 2.1 atom % and 0.9 atom %. The surface
content of Ag decreased a little from 10.2 atom % to 9.2 atom %. The ratio of Ag to Cu
increased remarkably from 4.8 to 11.5. All these results indicate that parts of the catalyst
surface are coated with coverings and coke deposition is formed during the epoxidation
reaction, which probably accounts for the catalyst deactivation.
<H3>3.7.3. Coke deposition
O2-TPO experiment is performed to demonstrate whether coke deposition is formed during
the reaction. Fig. 13 shows O2-TPO profiles of the deactivated Ag-CuCl2/BaCO3 catalysts
with different CuCl2 loadings. As shown in Fig. 13, CO2 and H2O were produced and O2 was
consumed over all the four catalysts, which indicated that coke deposition was formed on the
18

surface of all the four catalysts during the epoxidation reaction. Coke deposition covers the
active sites, which results in the worse catalytic performance and then the catalyst
deactivation. Coke deposition is usually considered as a common reason for the catalyst
deactivation for hydrocarbon reactions [64-66]. Interestingly, Ag-CuCl2/BaCO3 catalyst with
180 ppm of Cu and 200 ppm of Cl had a good stability shown in Fig. 12A, although there was
also coke deposition formed during the epoxidation reaction. Maybe these active sites covered
with coke deposition didn’t participate in the epoxidation reaction, so the catalytic
performance didn’t become worse even if the active sites were covered with coke deposition.
Over other three catalysts, coke deposition covered the active sites responsible for PO
formation, which resulted in a significant decrease in PO selectivity and the deactivation of
catalysts.
<H3>3.7.4. Mechanism of coke deposition
Fig. 14 shows the stability of Ag-CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400 ppm
of Cl after increasing the oxygen concentration in the feed gas. As shown in Fig. 14,
unfortunately, the catalyst stability was not improved, and PO selectivity still decreased
rapidly and propylene conversion increased slightly with an increase in time on stream, which
indicated that the catalyst deactivation still occurred and more oxygen in the feed gas couldn’t
make the complete combustion of coke deposition even with increasing the molar ratio of
C3H6 to O2 to 1:4.5 (the molar ratio of the complete combustion of C3H6 to CO2 and H2O).
Therefore changing reaction conditions can’t improve the catalyst stability and coke
deposition is formed intrinsically during the epoxidation reaction, that is to say, the oxygen-
containing intermediates cover the active sites, which leads to the formation of coke
deposition and then the catalyst deactivation.
As shown in Fig. 12B, Ag-CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400 ppm of Cl
deactivated rapidly with an increase in time on stream, whereas both Ag-Cu/BaCO3 with 360
19

ppm of Cu and Ag-Cl/BaCO3 with 400 ppm of Cl catalysts exhibited a good stability in spite
of the bad catalytic performance shown in Fig. S5. As shown in Fig. S6, compared with the
deactivated Ag-CuCl2/BaCO3 catalyst, the peak intensities of CO2 and H2O of Ag-Cu/BaCO3
and Ag-Cl/BaCO3 catalysts were so weak that little coke deposition was formed on the
surface of Ag-Cu/BaCO3 and Ag-Cl/BaCO3 catalysts during the epoxidation reaction. As
shown in Fig. 9, molecular oxygen species were difficult to adsorb on the surface of both Ag-
Cu/BaCO3 and Ag-Cl/BaCO3 catalysts, and molecular oxygen species could adsorb on the
surface of Ag-CuCl2/BaCO3 catalyst. An oxygen atom of molecular oxygen species reacts
with propylene to form a PO molecule and the left oxygen atom would like to react with
propylene to form CO2 and H2O with the molar ratio of O2 to C3H6 of 4.5. Insufficient oxygen
atoms react with propylene to form oxygen-containing intermediates, which can’t desorb
easily and cover the active sites, and thus result in the catalyst deactivation. Over Ag-
Cu/BaCO3 and Ag-Cl/BaCO3 catalysts without molecular oxygen species adsorbed, there
were no coke deposition and catalyst deactivation. As shown in Fig. 13A, the coke deposition
was observed on the surface of Ag-CuCl2/BaCO3 catalyst with 180 ppm of Cu and 200 ppm
of Cl, however, this catalyst exhibited a good stability shown in Fig. 12A. As shown in Fig. 7,
over Ag-CuCl2/BaCO3 catalyst with 180 ppm of Cu and 200 ppm of Cl, a large quantity of
adsorbed atomic oxygen species reacted with propylene to produce CO2 and molecular
oxygen species in comparison were so few that the coverage of the active sites of molecular
oxygen species adsorbed by coke deposition didn’t affect the catalytic performance. However,
over Ag-CuCl2/BaCO3 catalysts with higher CuCl2 loadings, when the amount of atomic
oxygen species adsorbed decreased and molecular oxygen species played an important role
for PO formation, the coverage of the active sites of molecular oxygen species adsorbed by
coke deposition would result in the catalyst deactivation.
20

Herein, the existence of molecular oxygen species and the balance between molecular
oxygen species and atomic oxygen species are necessary for achieving high PO selectivity.
However, these conditions will result in the coke deposition and then the catalyst deactivation.
Therefore, over Ag-CuCl2/BaCO3 catalyst, high PO selectivity and good catalytic stability are
conflicting and the deactivation of Ag-CuCl2/BaCO3 catalyst with high PO selectivity seems
to be inevitable.
<H2>3.8. Regeneration of Ag-CuCl2/BaCO3 catalyst
The catalyst deactivation was caused by the coverage of active sites by oxygen-containing
intermediates produced during the epoxidation reaction, which was probably reversible. Table
8 shows the catalytic performance of the regenerated Ag-CuCl2/BaCO3 catalyst with 360 ppm
of Cu and 400 ppm of Cl under different regeneration conditions. When the deactivated
catalyst was directly impregnated with Cl or Cu or Cl and Cu (Entry 3, 4, 5, 6), the catalytic
performance didn’t recover because propylene conversion became lower than that of the fresh
catalyst although PO selectivity increased greatly. The removal of coke deposition by the
combustion method is thought to be an easier method to recover the catalytic performance,
which is usually used for a deactivated catalyst with coke deposition to restore the activity
[59]. So the deactivated Ag-CuCl2/BaCO3 catalyst was regenerated in the quartz reactor under
a gas flow of O2-N2 (10 vol. %) of 20 mL·min-1 at 200 oC for 30 min, after which the surface
contents of Ag, Cu and Cl of the regenerated catalyst were 11.4 atom %, 2.1 atom % and 0.4
atom %, respectively, determined by XPS. Compared with the fresh catalyst, the surface
contents of Ag and Cu of the regenerated catalyst was almost the same except for lower
surface content of Cl. Unfortunately, PO selectivity of the regenerated catalyst was still low of
15.5 % (Entry 7) probably due to loss of Cl during the combustion of coke deposition. So the
deactivated catalyst was impregnated with Cl after the combustion of coke deposition, and
when 200 ppm of Cl was supplemented, both propylene conversion and PO selectivity of the
21

regenerated catalyst were the same as the catalytic performance of the fresh catalyst (Entry 9).
Therefore the deactivated Ag-CuCl2/BaCO3 catalyst can be completely regenerated by
combustion of coke deposition and then impregnation with appropriate amount of Cl.
<H1>4. Conclusions
Among Ag-MClx/BaCO3 catalysts with different chloride promoters, Ag-CuCl2/BaCO3
catalyst with 360 ppm of Cu and 400 ppm of Cl, prepared by the reduction-deposition-
impregnation method, exhibits the best catalytic performance for gas-phase epoxidation of
propylene by molecular oxygen due to the significant synergistic effect between Ag and
CuCl2, in which PO selectivity of 71.2 % and propylene conversion of 1.3 % are achieved
under the reaction conditions of 20 % C3H6 -10 % O2-70 % N2, 200 oC, 0.1 MPa, 3000 h-1.
The catalytic reaction mechanism of epoxidation of propylene by molecular oxygen over Ag-
CuCl2/BaCO3 catalyst follows Rideal-Eley mechanism, in which propylene in the gas phase
reacts with molecular oxygen species adsorbed on the surface of Ag at the interface in close
contact with CuCl2 to produce PO and with atomic oxygen species adsorbed on the surface of
Ag nanoparticles to produce CO2 and H2O. The appropriate CuCl2 doping of Ag-
CuCl2/BaCO3 catalyst can make adsorbed molecular oxygen species more active to produce
PO easily, meanwhile, depress the dissociation adsorption of oxygen to atomic oxygen
species leading to combustion of propylene to CO2, which benefits epoxidation of propylene
by molecular oxygen. Over Ag-CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400 ppm of
Cl, only PO selectivity of 13.9 % is obtained at propylene conversion of 3.2 % after reaction
for 500 min. One oxygen atom of molecular oxygen species reacts with propylene to form a
PO molecule and the left insufficient oxygen atoms react with propylene to produce oxygen-
containing intermediates and then to form the coke deposition which covers the active sites
and thus results in the catalyst deactivation. The deactivation of Ag-CuCl2/BaCO3 catalyst
with high PO selectivity seems to be inevitable. The deactivated Ag-CuCl2/BaCO3 catalyst
22

can be completely regenerated by combustion of coke deposition and then impregnation with
appropriate amount of Cl due to loss of Cl during combustion of coke deposition.
Appendix A. Supplementary Material
Supplementary material associated with this article can be found in the online version.
Acknowledgments
This work was supported by National Basic Research Program of China (2013CB933200)
and Commission of Science and Technology of Shanghai Municipality (15DZ1205305).
<REF>References
<BIBL>
[1] T.A. Nijhuis, M. Makkee, J.A. Moulijn, B.M. Weckhuysen,;1; Ind. Eng. Chem. Res. 45
(2006) 3447-3459.
[2] G. Blanco-Brieva, M.C. Capel-Sánchez, M.P. Frutos, A. Padilla-Polo, J.M. Campos-
Martín, J.L.G. Fierro,;1; Ind. Eng. Chem. Res. 47 (2008) 8011-8015.
[3] M. Okumura, T. Fujitani, J. Huang, T. Ishida,;1; ACS Catal. 5 (2015) 4699-4707.
[4] T. Hayashi, K. Tanaka, M. Haruta,;1; J. Catal. 178 (1998) 566-575.
[5] J. Chou, E.W. McFarland,;1; Chem. Commun. 14 (2004) 1648-1649.
[6] J.Q. Lu, X.M. Zhang, J.J. Bravo-Suárez, K.K. Bando, T. Fujitani, S.T. Oyama,;1; J. Catal.
250 (2007) 350-359.
[7] W.S. Lee, R. Zhang, M.C. Akatay, C.D. Baertsch, E.A. Stach, F.H. Ribeiro, W.N.
Delgass,;1; ACS Catal. 1 (2011) 1327-1330.
[8] C.H. Liu, Y.J. Guan, E.J.M. Hensen, J.F. Lee, C.M. Yang,;1; J. Catal. 282 (2011) 94-102.
[9] B. Chowdhury, K.K. Bando, J.J. Bravo-Suárez, S. Tsubota, M. Haruta,;1; J. Mol. Catal. A:
Chem. 359 (2012) 21-27.
[10] W.S. Lee, M.C. Akatay, E.A. Stach, F.H. Ribeiro, W.N. Delgass,;1; J. Catal. 313 (2014)
104-112.
[11] X. Feng, X.Z. Duan, H.Y. Cheng, G. Qian, D. Chen, W.K. Yuan, X.G. Zhou,;1; J. Catal.
325 (2015) 128-135.
[12] A. Chongterdtoonskul, J.W. Schwank, S. Chavadej,;1; J. Mol. Catal. A: Chem. 372
(2013) 175-182.
[13] X.L. Jing, H.X. Wang, H.M. Chen, J.L. Huang, D.H. Sun, Q.B. Li,;1; RSC Adv. 4 (2014)
27597-27603.
[14] W. Diao, C.D. DiGiulio, M.T. Schaal, S. Ma, J.R. Monnier,;1; J. Catal. 322 (2015) 14-
23.
[15] A. Palermo, A. Husain, M.S. Tikhov, R.M. Lambert,;1; J. Catal. 207 (2002) 331-340.
[16] J.Q. Lu, J.J. Bravo-Suárez, A. Takahashi, M. Haruta, S.T. Oysma,;1; J. Catal. 232 (2005)
85-95.
[17] J.Q. Lu, J.J. Bravo-Suárez, M. Haruta, S.T. Oyama,;1; Appl. Catal. A: Gen. 302 (2006)
283-295.
[18] S. Vajda, S. Lee, K. Sell, I. Barke, A. Kleibert, V.V. Oeynhausen, K. Meiwes-Broer,
A.F. Rodríguez, J.W. Elam, M.M. Pellin, B. Lee, S. Seifert, R.E. Winans,;1; J. Chem. Phys.
131 (2009) 121104.
[19] F.W. Zemichael, A. Palermo, M.S. Tikhov, R.M. Lambert,;1; Catal. Lett. 80 (2002) 93-
98.
[20] G.J. Jin, G.Z. Lu, Y.L. Guo, Y. Guo, J.S. Wang, W.Y. Kong, X.H. Liu,;1; J. Mol. Catal.
A: Chem. 93-95 (2004) 173-182.
23

[21] W. Yao, G.Z. Lu, Y.L. Guo, Y. Guo, Y.Q. Wang, Z.G. Zhang,;1; J. Mol. Catal. A:
Chem. 276 (2007) 162-167.
[22] X. Zheng, Q. Zhang, Y.L. Guo, W.C. Zhan, Y. Guo, Y.S. Wang, G.Z. Lu,;1; J. Mol.
Catal. A: Chem. 357 (2012) 106-111.
[23] X. Zheng, Y.L. Guo, Y. Guo, Q. Zhang, X.H. Liu, L. Wang, W.C. Zhan, G.Z. Lu,;1;
Rare Metals 34 (2015) 477-490.
[24] J.T. Jankowiak, M.A. Barteau,;1; J. Catal. 236 (2005) 379-386.
[25] M.O. Ozbek, I. Onal, R.A.V. Santen,;1; ChemCatChem 5 (2013) 443-451.
[26] T.C.R. Rocha, M. Hävecker, A. Knop-Gericke, R. Schlögl,;1; J. Catal. 312 (2014) 12-16.
[27] W. Yao, X. Zheng, Y.L. Guo, W.C. Zhan, Y. Guo, G.Z. Lu,;1; J. Mol. Catal. A: Chem.
342-343 (2011) 30-34.
[28] G.Z. Lu, X.B. Zuo,;1; Catal. Lett. 58 (1999) 67-70.
[29] J.Q. Lu, M.F. Luo, H. Lei, C. Li,;1; Appl. Catal. A: Gen. 237 (2002) 11-19.
[30] J.T. Jankowiak, M.A. Barteau,;1; J. Catal. 236 (2005) 379-386.
[31] M.F. Luo, J.Q. Lu, C. Li,;1; Catal. Lett. 86 (2003) 43-49.
[32] J.R. Monnier, J.L. Stavinoha, G.W. Hartley,;1; J. Catal. 226 (2004) 321-333.
[33] R.M. Lambert, R.L. Cropley, A. Husain, M.S. Tikhov,;1; Chem. Commun. 10 (2003)
1184-1185.
[34] R.L. Cropley, F.J. Williams, O.P.H. Vaughan, A.J. Urquhart, M.S. Tikhov, R.M.
Lambert,;1; Surf. Sci. 578 (2005) 85-88.
[35] D. Torres, N. Lopez, F. Illas, R.M. Lambert,;1; J. Am. Chem. Soc. 127 (2005) 10774-
10775.
[36] D. Torres, N. Lopez, F. Illas, R.M. Lambert,;1; Angew. Chem. 119 (2007) 2101-2104.
[37] L.J. Yang, J.L. He, Q.H. Zhang, Y. Wang,;1; J. Catal. 276 (2010) 76-84.
[38] A. Seubsai, B. Zohour, D. Noon, S. Senkan,;1; ChemCatChem 6 (2014) 1215-1219.
[39] A. Seubsai, M. Kahn, B. Zohour, D. Noon, M. Charoenpanich, S. Senkan,;1; Ind. Eng.
Chem. Res. 54 (2015) 2638-2645.
[40] W.G. Su, S.G. Wang, P.L. Ying, Z.C. Feng, C. Li,;1; J. Catal. 268 (2009) 165-174.
[41] J.L. He, Q.G. Zhai, Q.H. Zhang, W.P. Deng, Y. Wang,;1; J. Catal. 299 (2013) 53-66.
24

[42] A. Marimuthu, J.W. Zhang, S. Linic,;1; Science 339 (2013) 1590-1593.
[43] Q. Zhang, Y.L. Guo, W.C. Zhan, Y. Guo, L. Wang, Y.S. Wang, G.Z. Lu,;1; Catal. Today
http://dx.doi.org/10.1016/j.cattod.2016.01.015.
[44] J.K. Plischke, M.A. Vannice,;1; Appli. Catal. 42 (1988) 255-283.
[45] J.Q. Lu, M.F. Luo, C. Li,;1; React. Kinet. Catal. Lett. 86 (2005) 219-224.
[46] S. Ghosh, S.S. Acharyya, R. Tiwari, B. Sarkar, R.K. Singha, C. Pendem, T. Sasaki, R.
Bal,;1; ACS Catal. 4 (2014) 2169-2174.
[47] L. Ma, D.S. Wang, J.H. Li, B.Y. Bai, L.X. Fu, Y.D. Li,;1; Appl. Catal. B: Environ. 148-
149 (2014) 36-43.
[48] F. Gunnarsson, H. Kannisto, M. Skoglundh, H. Härelind,;1; Appl. Catal. B: Environ.
152-153 (2014) 218-225.
[49] K. Kershen, H. Celio, I. Lee, J.M. White,;1; Langmuir 17 (2001) 323-328.
[50] C. T. Campbell,;1; Surf. Sci. 157 (1985) 43-60.
[51] M. Dean, M. Bowker,;1; J. Catal. 115 (1989) 138-147.
[52] G.W. Busser, O. Hinrichsen, M. Muhler,;1; Catal. Lett. 79 (2002) 49-54.
[53] G.I.N. Waterhouse, G.A. Bowmaker, J.B. Metson,;1; Appl. Surf. Sci. 214 (2003) 36-51.
[54] K. Yamazaki, T. Kayama, F. Dong, H. Shinjoh,;1; J. Catal. 282 (2011) 289-298.
[55] T.E. Jones, T.C.R. Rocha, A. Knop-Gericke, C. Stampfl, R. Schlögl, S. Piccinin,;1; ACS
Catal. 5 (2015) 5846-5850.
[56] J. Derouin, R.G. Farber, S.L. Heslop, D.R. Killelea,;1; Surf. Sci. 641 (2015) 1-4.
[57] V.I. Bukhtiyarov, V.V. Kaichev, E.A. Podgornov, I.P. Prosvirin,;1; Catal. Lett. 57 (1999)
233-239.
[58] L. Cheng, C. Yin, F. Mehmood, B. Liu, J. Greeley, S. Lee, B. Lee, S. Seifert, R.E.
Winans, D. Teschner, R. Schlogl, S. Vajda, L.A. Curtiss,;1; ACS Catal. 4 (2014) 32-39.
[59] J.A. Moulijn, A.E. Diepen, F. Kapteijn,;1; Appl. Catal. A: Gen. 212 (2001) 3-16.
[60] Y. Lei, F. Mehmood, S. Lee, J. Greeley, B. Lee, S. Seifert, R.E. Winans, J.W. Elam, R.J.
Meyer, P.C. Redfern, D. Teschner, R. Schlögl, M.J. Pellin, L.A. Curtiss, S. Vajda,;1; Science
328 (2010) 224-228.
[61] L.M. Molina, S. Lee, K. Sell, G. Barcaro, A. Fortunelli, B. Lee, S. Seifert, R.E. Winans,
J.W. Elam, M.J. Pellin, I. Barke, V.V. Oeynhausen, Y. Lei, R.J. Meyer, J.A. Alonso, A.F.
25

Rodríguez, A. Kleibert, S. Giorgio, C.R. Henry, K.H. Meiwes-Broer, S. Vajda,;1; Catal.
Today 160 (2011) 116-130.
[62] G. Mul, A. Zwijnenbyrg, B.V.D. Linden, M. Makkee, J.A. Moulijn,;1; J. Catal. 201
(2001) 128-137.
[63] T.A. Nijhuis, B.M. Weckhuysen,;1; Catal. Today 117 (2006) 84-89.
[64] N. Madaan, S. Gatla, V.N. Kalevaru, J. Radnik, B. Lücke, A. Brückner, A. Martin,;1; J.
Catal. 282 (2011) 103-111.
[65] S.I. Sanchez, M.D. Moser, S.A. Bradley,;1; ACS Catal. 4 (2014) 220-228.
[66] C.Z. Wang, N.N. Sun, N. Zhao, W. Wei, Y.H. Sun, C.G. Sun, H. Liu, C.E. Snape,;1;
Fuel 143 (2015) 527-535.
</BIBL>
<Figure>Fig. 1. Effect of the reaction temperature on the catalytic performance of Ag/BaCO3
catalyst (A) and Ag-CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400 ppm of Cl (B) for
epoxidation of propylene by molecular oxygen.
Reaction conditions: 0.6 g catalyst, 20 % C3H6 -10 % O2-70 % N2, 0.1 MPa, 3000 h-1.
<Figure>Fig. 2. Logarithm of the reaction rate of epoxidation of propylene versus the reciprocal of
reaction temperature over Ag/BaCO3 catalyst and Ag-CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400
ppm of Cl.
<Figure>Fig. 3. Effect of GHSV on the catalytic performance of Ag-CuCl2/BaCO3 catalyst with 360 ppm
of Cu and 400 ppm of Cl for epoxidation of propylene by molecular oxygen.
Reaction conditions: 0.6 g catalyst, 20 % C3H6 -10 % O2-70 % N2, 200 oC, 0.1 MPa.
<Figure>Fig. 4. Cu 2p XPS (A), Cl 2p XPS (B) and Ag 3d XPS (C) spectra of Ag-CuCl2/BaCO3 catalysts
with different CuCl2 loadings.
<Figure>Fig. 5. Dependence of the reaction rates of epoxidation of C3H6 on the partial pressure of O2 (A)
or the partial pressure of C3H6 (B) over Ag-CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400 ppm of Cl.
<Figure>Fig. 6. C3H6-TPD profiles of CuCl2/BaCO3 catalyst, Ag/BaCO3 catalyst and Ag-CuCl2/BaCO3
catalyst with 360 ppm of Cu and 400 ppm of Cl.
<Figure>Fig. 7. O2-TPD profiles of Ag-CuCl2/BaCO3 catalysts with different CuCl2 loadings.
<Figure>Fig. 8. O2-TPD profiles of Ag-MClx/BaCO3 catalysts.
26

<Figure>Fig. 9. O2-TPD profiles of Ag/BaCO3 catalysts modified with different promoters and
CuCl2/BaCO3 catalyst.
<Figure>Fig. 10. C3H6-TPSR profiles of Ag-CuCl2/BaCO3 catalysts with different CuCl2 loadings. The
MS signals of PO (m/e = 58) (A) and CO2 (m/e = 44) (B).
<Figure>Fig. 11. Proposed schematic diagram of epoxidation of propylene by molecular
oxygen over Ag-CuCl2/BaCO3 catalysts with different CuCl2 loadings.
<Figure>Fig. 12. Stabilities of Ag-CuCl2/BaCO3 catalysts and CuCl2/BaCO3 catalysts with
different CuCl2 loadings for epoxidation of propylene by molecular oxygen. Propylene
conversion (A) and PO selectivity (B) of Ag-CuCl2/BaCO3 catalysts with 180 ppm of Cu and
200 ppm of Cl (■), 360 ppm of Cu and 400 ppm of Cl (●), 540 ppm of Cu and 600 ppm of Cl
(▲), 720 ppm of Cu and 800 ppm of Cl (▼). Propylene conversion (A) and PO selectivity (B)
of CuCl2/BaCO3 catalysts with 180 ppm of Cu and 200 ppm of Cl (▓▓▓□), 360 ppm of Cu
and 400 ppm of Cl (◇), 540 ppm of Cu and 600 ppm of Cl (☆), 720 ppm of Cu and 800 ppm
of Cl (○).
Reaction conditions: 0.6 g catalyst, 20 % C3H6 -10 % O2-70 % N2, 200 oC, 0.1 MPa, 3000 h-1.
<Figure>Fig. 13. O2-TPO profiles of the deactivated Ag-CuCl2/BaCO3 catalysts with different
CuCl2 loadings. 180 ppm Cu of and 200 ppm of Cl (A), 360 ppm of Cu and 400 ppm of Cl
(B), 540 ppm of Cu and 600 ppm of Cl (C), 720 ppm of Cu and 800 ppm of Cl (D).
<Figure>Fig. 14. Stability of Ag-CuCl2/BaCO3 catalyst with 360 ppm of Cu and 400 ppm of Cl under the
feed gas with different O2 concentration.
Tables
<Table>Table 1 Effects of the chloride promoters on the catalytic performance of Ag-MClx/BaCO3
catalysts with 400 ppm of Cl.
Propylene
conversion (%)
Selectivity (%)
Chlorides
PO
PA
--
CO2
63.1
77.7
75.4
83.0
75.2
87.7
28.6
73.6
--
12.5
0.3
0.5
2.0
0.4
0.6
1.3
5.2
36.9
21.5
24.3
16.8
24.3
12.0
71.2
26.1
NH4Cl
NaCl
KCl
CaCl2
FeCl3
CuCl2
CuCl
0.8
0.3
0.2
0.5
0.3
0.2
0.3
Reaction conditions: 0.6 g catalyst, 20 % C3H6 -10 % O2-70 % N2, 200 oC, 0.1 MPa, 3000 h-1.
<Table>Table 2 Effect of CuCl2 loading on the catalytic performance of CuCl2/BaCO3 and
Ag-CuCl2/BaCO3 catalysts.
27

28

Reaction conditions: 0.6 g catalyst, 20 % C3H6 -10 % O2-70 % N2, 200 oC, 0.1 MPa, 3000 h-1.
a Determined from XRD patterns shown in Fig. S1 by Scherrer’s equation.
<Table>Table 3 The surface composition of Ag-CuCl2/BaCO3 catalysts with different CuCl2 loadings.
Surface composition
Catalysts
Ag (atom %) Cu (atom %) Cl (atom %)
Ag/Cu
--
7.8
4.8
1.4
Ag/BaCO3
12.4
12.5
10.2
3.6
--
--
0.7
0.9
--
Ag-180ppmCu-200ppmCl/BaCO3
Ag-360ppmCu-400ppmCl/BaCO3
Ag-540ppmCu-600ppmCl/BaCO3
Ag-720ppmCu-800ppmCl/BaCO3
1.6
2.1
2.6
2.9
2.9
--
1.0
<Table>Table 4 The catalytic performance of Ag/BaCO3 catalysts modified with different promoters and
CuCl2/BaCO3 catalyst.
Propylene
conversion (%)
Selectivity (%)
Catalysts
PO
36.9
6.2
21.5
71.2
28.1
PA
--
--
0.8
0.2
1.9
CO2
63.1
93.8
77.7
28.6
70.0
Ag/BaCO3
Ag-Cu/BaCO3
Ag-Cl/BaCO3
Ag-CuCl2/BaCO3
CuCl2/BaCO3
12.5
6.9
0.3
1.3
0.1
a
b
Reaction conditions: 0.6 g catalyst, 20 % C3H6 -10 % O2-70 % N2, 200 oC, 0.1 MPa, 3000 h-1.
a Cu(NO3)2 was used as the precursor of Cu with 360 ppm of Cu.
b NH4Cl was used as the precursor of Cl with 400 ppm of Cl.
<Table>Table 5 The comparison between the fresh and deactivated Ag-CuCl2/BaCO3
catalysts with 360 ppm of Cu and 400 ppm of Cl after reaction for 500 min.
Propylene
conversion
(%)
Ag
content
(wt. %) a
4.3
Cu
Ag
Selectivity (%)
content
(ppm) a
350
crystallite
size (nm) b
23.1
Catalysts
Fresh
PO
71.2
PA
0.2
CO2
1.3
28.6
29