Critical Roles of Doping Cl on Cu2O Nanocrystals for Direct Epoxidation of Propylene by Molecular Oxygen

Article abstract of DOI:10.1021/jacs.0c03882

Direct epoxidation of propylene by molecular oxygen alone is one of the dream reactions in heterogeneous catalysis. Despite much effort, the yield of propylene epoxide is still too low to be commercially attractive due to the trade-off between conversion and selectivity. Here, we demonstrate that doping Cl into the lattice of Cu2O nanocrystals by the intergrowth method not only can enhance the catalytic selectivity and conversion of direct propylene epoxidation but also can solve the long-existing Cl loss problem. In particular, Cl-doped rhombic dodecahedral Cu2O with (110) exposing facets exhibited 63% PO selectivity with a 12.0 h-1 turnover frequency at 200 °C, outperforming any other coinage metal-based catalysts under mild conditions. Comprehensive characterization and theoretical calculations revealed that the Cl-decorated Cu(I) facilitated formation of electrophilic oxygen species, thus boosting the production of propylene oxide. This work provides a general strategy to develop catalysts and explore the promoter effect by creating uniform isolated anion doping to activate a nearby metal center by virtue of well-defined nanocrystals.

Full text of DOI:10.1021/jacs.0c03882

Subscriber access provided by FLORIDA INTL UNIV
Article
Critical roles of doping Cl on Cu2O nanocrystals for
direct epoxidation of propylene by molecular oxygen
Chao Zhan, Qiuxiang Wang, Lingyun Zhou, Xiao Han, Yongjin Wanyan, Jiayu
Chen, Yanping Zheng, Ye Wang, Gang Fu, Zhaoxiong Xie, and Zhong-qun Tian
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.0c03882 • Publication Date (Web): 23 Jul 2020
Downloaded from pubs.acs.org on July 23, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 7
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Critical roles of doping Cl on Cu O nanocrystals for direct
epoxidation of propylene by molecular oxygen
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†
†
†
Chao Zhan , Qiuxiang Wang , Lingyun Zhou , Xiao Han, Yongyin Wanyan, Jiayu Chen, Yanping
Zheng, Ye Wang, Gang Fu*, Zhaoxiong Xie*, Zhongqun Tian
State Key Laboratory of Physical Chemistry of Solid Surfaces, College of Chemistry and Chemical Engineering,
Collaborative Innovation Center of Chemistry for Energy Materials (iChEM), National Engineering Laboratory for Green
Chemical Productions of Alcohols, Ethers, and Esters, Fujian Provincial Key Laboratory of Theoretical and Computational
Chemistry and Department of Chemistry, Xiamen University, Xiamen 361005, China.
ABSTRACT: The direct epoxidation of propylene by molecular oxygen alone is one of the “dream reaction” in heterogeneous
catalysis. Despite much effort, the yield of propylene epoxide is still too low to be commercially attractive due to the trade-off between
conversion and selectivity. Here, we demonstrated that doping Cl into the lattice of Cu
2
O nanocrystals by the intergrowth method can
not only enhance the catalytic selectivity and conversion of direct propylene epoxidation but also solved the long-exist Cl loss
problem. Especially, Cl-doped rhombic dodecahedral Cu O with (110) exposing facets exhibited 63% PO selectivity with 12.0 h
2
turnover frequency at 200 °C, outperforming any other coinage metal based catalysts under mild condition. Comprehensive
characterizations and theoretical calculations revealed that the Cl decorated Cu(I) facilitated the formation of electrophilic oxygen
species, thus boosting the production of propylene oxide. This work provided a general strategy to develop catalysts and explore the
promoter effect by creating uniform isolated anion doping to activate nearby metal center in virtue of well-defined nanocrystals.
-
1
INTRODUCTION
hinder the mechanistic understanding of Cl effect at the atomic
level and prevent the mechanism-guided catalyst design.
Traditional surface-chemistry method has successfully applied
to many important catalytic processes,
adsorption of propylene under vacuum condition make it a great
Under growing pressures from environment, energy and
economy, it is highly desirable to upgrade amounts of existing
1
8,19
however, the weak
1
-3
catalytic processes.
Propylene oxide (PO), as a bulky
chemical (> 10 million tons per year), is mainly used in the
production of polyethers and propylene glycol. So far, the
13,20
challenge to initiate the DEP reaction.
Even worse, the Cl
4
loss is another big issue existed in Cl-modified catalysts, which
not only corrodes the apparatus but also seriously deteriorates
catalysts.
majority of PO is produced by the chlorohydrin or peroxidation
processes, which faces with security and environment
5
,6
problems. To solve this, direct epoxidation of propylene by
molecular oxygen alone (hereafter denoted as DEP) has been
In present work, we showed that when using NH
reductant, a small amount of Cl could dope into the crystal
lattice of Cu O nanocrystals, which significantly improved both
2
OH·HCl as
6-10
proposed as an ideal alternative. Although the oxygen atoms
on the Cu surface were claimed to have high selectivity for
2
1
1-13
the activity and selectivity for DEP and solved the long-exist Cl
loss problem. Combined with appropriate crystal facet (110),
DEP,
unsatisfactory.
could only be achieved at very low propylene conversion
the unmodified Cu catalysts were still
6
,10,14,15
The high PO selectivity (40% ~ 50%)
the as-prepared Cl-contained rhombic dodecahedral Cu
nanocrystals (Cl-RD-Cu O) displayed one order of magnitude
higher PO formation rate than that of the Cl-free RD-Cu O and
2
O
(
usually less than 0.25%), while when the conversion increased,
the PO selectivity dramatically decreased.
2
6
,9,14
Such trade-off
2
NH
4
Cl solvent post-treatment cases (NH
O achieved near 100% PO selectivity with 0.45
h turnover frequency (TOF), and the 63% PO selectivity was
4 2
Cl-RD-Cu O). At 125
between conversion and selectivity was suggested to be
originated from the series-parallel network of propylene
oxidation in which complete combustion pathways would
branch off from the selective one at the reactant and any
°
C, Cl-RD-Cu
2
-1
-1
retained with ~12 h TOF at 200 °C. Comprehensive
5,6,11
characterizations revealed that Cl penetrated into the Cu O
lattice as deep as tens of nm and Cu-Cl bonds were formed
during the synthesis process for Cl-RD-Cu
intermediates.
Thus, it is highly desirable to develop
2
efficient strategies to increase conversion of propylene without
compromising the PO selectivity.
2
O, distinguishing
from that of NH Cl-RD-Cu O. Density functional theory (DFT)
4
2
Industrially, it has been widely accepted that Cl is an effective
promoter for ethylene epoxidation.
no consensus on the Cl effect for the DEP reaction.
Conventionally, Cl is introduced by post-treatment method, i.e.
impregnating the as-synthesized catalysts into the chlorine salt
solvents. However, such a post-treatment method usually
causes the chemical and structural heterogeneity of Cl.
Moreover, the particle size, shape, and surface coordination of
the Cu based catalysts leads to additional complexity. All these
calculations showed that the substitutional Cl rendered the
nearby Cu(I) more reactive for the activation of molecular
5,6,16,17
However, there is still
1
4,16
-
-
oxygen, producing the electrophilic O
2
and O species, which
interacted preferentially with the C=C double bond rather than

-C-H bond, thus preventing the by-reactions.
RESULTS AND DICUSSION
Characterization and catalytic performance of the as-
prepared Cl-RD-Cu
2
O
and RD-Cu
2
O
nanocrystals.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 7
Compared with the single-crystal model systems, nanocrystals
(SAED) further confirmed the exposure of (110) facets (Figure
1b, 1d). X-ray diffraction (XRD) identified that both of them
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
with well-defined structures and relatively large surface area are
able to work under catalytic relevant condition, thus bridging
the material gap and pressure gap when investigating the
are Cu
2Theta of the XRD peak of Cl-RD-Cu
lattice expansion (Figure S5) and broadening diffraction peaks
of Cl-RD-Cu O can be attributed to its smaller crystallite size
and Cl doping compared to RD-Cu
band gap of Cl-RD-Cu O was wider than that of RD-Cu
eV vs. 2.0 eV), accounting for the color change from brick red
2
O nanocrystals (Figure 1e). The slight shift to lower
2
O reflected a slight
2
1-23
structure-function relationship.
prepared by wet chemical methods. The RD-Cu
exposing (110) facets are chosen because they have been
reported as the most selective catalyst for PO among Cu
RD-Cu
2
O nanocrystals were
2
O nanocrystals
2
1
5,24,25
2
O.
Experimentally, the
O (2.3
2
O
2
2
2
2
nanocrystals. Here, NH
reductant, respectively, resulting in Cl-RD-Cu
Details in Supporting Information, Figures S1-S4). As shown
2
OH·HCl and glucose were used as
2
6
2
O and RD-Cu
2
O
(RD-Cu
450 nm illumination, their emission spectra were similar, but
the relative intensity of Cl-RD-Cu O at longer wavelength
(~560 nm) is slightly higher than that of RD-Cu O (Figure S8).
2 2
O) to orange (Cl-RD-Cu O). (Figure S6-S7). Under
(
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
in Figure1a-d, the scanning electron microscopy (SEM) and
transmission electron microscope (TEM) images revealed that
2
2
both Cl-RD-Cu
2
O and RD-Cu
2
O have dodecahedron crystal
In addition, X-ray photoelectron spectroscopy (XPS) of Cu 2p
showed that the surface copper atoms on both nanocrystals were
almost in +1 valence (Figure 1f).
shape with the size in between 200 nm to 400 nm and 300 nm
to 600 nm, respectively. Selected area electron diffraction
Figure 1. Structure characterizations and catalytic performance of the as-prepared RD-Cu
nanocrystals. (a) SEM image of Cl-RD-Cu O. (b) TEM image of Cl-RD-Cu O and corresponding SAED pattern (the inset). (c) SEM
image of RD-Cu O. (d) TEM image of RD-Cu O and corresponding SAED pattern (the inset). (e) XRD patterns of Cl-RD-Cu O and
RD-Cu O. (f) XPS patterns of Cu of Cl-RD-Cu O and RD-Cu O. (g) PO selectivity and formation rate of Cl-RD-Cu O and RD-Cu
at 125 °C. (h) Catalytic performance of Cl-RD-Cu O at different temperatures. (i) The PO selectivity as a function of TOF for typical
coinage metal based catalysts for DEP under similar condition (the details were shown in Table S1). The gray dots are Ag-based
catalysts; the green dot is Au-based catalyst; the white dots are Cu-based catalysts, the blue dots are RD-Cu O, and the orange dots
are Cl-RD-Cu O.
Although having the similar morphology and structure, Cl-
RD-Cu O and RD-Cu O exhibited dramatically differerent
2 2
O and Cl-RD-Cu O
2
2
2
2
2
2
2
2
2
2
O
2
2
2
temperature were illustrated in Figure 1h. At 200 °C, 1.0 %
3 6
C H conversion with 63% PO selectivity was achieved, and the
2
2
-1
-1
catalytic performances towards DEP (Figure 1g-i). Figure 1g
PO formation rate attained 500 umol·g ·h . During the
o
o
showed that >95% PO selectivity was obtained at 125 °C with
warming/cooling processes between 125 C to 200 C, there was
no significant decay in the activity and selectivity (Figure S9);
meanwhile, there was no significant change in the phase and the
morphology of the catalysts (Figures S10-S12). Under the same
-1 -1
2
0 umol·g ·h PO formation rate for Cl-RD-Cu
2
O, while
O was employed. The
detailed catalytic performance of Cl-RD-Cu O at different
nearly no reaction occurred when RD-Cu
2
2
ACS Paragon Plus Environment

Page 3 of 7
Journal of the American Chemical Society
experiment condition, the RD-Cu
conversion with less than 30% PO selectivity (Figures S13-S15),
2
O exhibited only 0.3%
with Cu atoms arranged in face-centered cubic sites and anions
tetrahedrally coordinated by Cu atoms. In fact, it has been
proved that partial lattice oxygen atoms in Cu O could be
substituted by Cl atoms. Thus, we referred such a method as
“intergrowth method”. To confirm the existence of Cl,
comprehensive characterizations were employed, such as
energy dispersive spectroscopy (EDS), scanning transmission
electron microscopy (STEM), XPS, high-sensitivity low-
energy ion scattering spectroscopy (HS-LEISS) and Raman
spectroscopy. STEM elemental mappings confirmed the
1
2
3
4
5
6
7
8
9
1
5
consistent with that of RD-Cu
the in-situ FTIR of RD-Cu O and Cl-RD-Cu
there exists a signal of CH bending-vibration at 1441 cm ,
lending to support the production of propylene oxide through
2
O previously reported. From
2
2
7-29
2
2
O, we can see that
-1
2
1
5
propylene oxidation (Figure S16-17). Meanwhile, no obvious
IR signal attributed to oleic acid could be observed on RD-Cu
Figure S18). Figure 1i plotted the selectivity of coinage-metal
2
O
(
based catalysts versus the TOF under the similar reaction
condition (details in Figure S19 and Table S2). It was clear that
most catalysts distributed in the region with TOF below 8 h
and PO selectivity less than 60%. However, Cl-RD-Cu
appeared on the upper right corner with 12.0 h TOF and 63%
PO selectivity, outperforming any other coinage-metal catalysts.
2
existence of Cl in Cl-RD-Cu O (Figure 2a). The EDS results
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
-1
showed that Cl-RD-Cu O contained ~2.0 wt% residual Cl, the
2
2
O
corresponding Cu/Cl atomic ratio was 22.8:1 (Figure S20 and
Table S3). The XPS results also demonstrated the existence of
Cl (Figure 2b), and the concentration of Cl in the near surface
region could reach 6.7 wt%, the corresponding Cu/Cl atomic
ratio was 6.9:1(Figure S21 and Table S4). The Cl 2p spectrum
of Cl-RD-Cu O shows a Cl 2p3/2 peak at 198.3 eV and a Cl 2p1/2
2
peak at 199.9 eV, together with a spacing of 1.6 eV and a peak
area ratio of about 2:1 (Figure 2b), which indicates that chlorine
-1
atoms enters into the lattice of Cu
2
O through substituting the
30
lattice oxygen atoms and bonding with copper atoms. Indeed,
-1
-1
the Raman peaks at 68 cm and 196 cm verified the existence
31,32
of Cu-Cl bond in Cl-RD-Cu
2
O (Figure 2c).
Furthermore,
HS-LEISS was performed to explore the distribution of Cl. As
shown in Figure 2d, with increasing the etching depth, the
relative intensity of Cl decreased, but the signal of Cl still
maintained as deep as 18 nm.
The catalytic performance and durability of different Cl
doping Cu
was prepared using the conventional post-treatment method by
impregnating the RD-Cu O nanocrystals into the NH Cl solvent.
Again, from TEM and XRD, no significant difference could be
observed for NH Cl-RD-Cu O when comparing with pristine
RD-Cu O (Figures S22 and S23). In terms of catalytic
performance, NH Cl-RD-Cu O was better than RD-Cu O, but
still far inferior to Cl-RD-Cu O, as shown in Figure 2e and
Figure S24-S26. Without special description, NH Cl-RD-Cu
in the manuscript referred to RD-Cu O with optimal
concentration of NH Cl (0.1 wt%, Figure S27). Moreover,
octahedral Cu O (OCT-Cu O), and Cl-OCT-Cu O through
intergrowth method and NH Cl-OCT-Cu O (0.1 wt%) through
2 4 2
O catalysts. For comparison, the NH Cl-RD-Cu O
2
4
4
2
2
4
2
2
2
4
2
O
2
4
2
2
2
4
2
post-treatment method were also prepared. Once again, the
similar trend in catalytic performance could be observed: Cl-
OCT-Cu O> NH Cl-OCT-Cu O> OCT-Cu O, as shown in
2 4 2 2
Figure 2e and Figures S28-S31. From above discussion, it was
inferred that the intergrowth method would lead to formation
more uniform Cl, which decorated Cu(I) to generate Cl-Cu
bonds. On the contrary, post-treatment method would produce
Figure 2. The characterizations of the doping Cl and
improved catalytic performance. (a) High-angle annular
dark-field image of a single Cl-RD-Cu
elemental mapping images of Cu, O and Cl. (b) XPS of Cl for
Cl-RD-Cu O and RD-Cu O. (c) Raman spectroscopy for Cu-
based catalysts. (d) HS-LEIS for the Cu and Cl distribution of
Cl-RD-Cu O. (e) The PO selectivity as a function of the PO
formation rate for different Cu O nanocrystals at different
temperatures. All data in Figure 2e are from catalytic tests of
different Cu O nanocrystals at different temperatures.
From the viewpoint of synthesize method, the main
difference between Cl-RD-Cu O and RD-Cu O was the
reductants they used. We suggested that when employing
NH OH·HCl as reductant, a small amount of Cl would be
residual on/in Cl-RD-Cu O during the process of crystal growth,
as CuCl also has a low solubility product constant of only
2
O nanocrystal and
4
various kinds of Cl, such as NH Cl salt, surface adsorbed Cl ion,
substitutional Cl, and/or CuCl. Additionally, in virtue of
nanocrystal models, it was revealed that there existed synergy
between facet effects and Cl doping effects, accounting for
2
2
2
higher catalytic performance of Cl-RD-Cu
OCT-Cu O. Considering that CuCl having Cl-Cu-Cl group only
showed very poor DEP activity (Figure S32), the excellent
performance of Cl-RD-Cu O and Cl-OCT-Cu O would be
ascribed to the formation of Cl-Cu-O moieties. Thus, the
substitutional chlorine combining with the Cu O(110) facet
2
O than that of Cl-
2
2
2
2
2
2
2
2
accounts for the excellent performance of propylene
epoxidation by molecular oxygen.
2
2
Cl-modified catalysts usually encountered the Cl loss
problems which would deteriorate the catalytic performance
-7
1
2
.72×10 , and CuCl and Cu O share similar structural features
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 7
and contaminate the products.5,6,17,33 As shown in Figure 3a, the
PO formation rate of NH Cl-RD-Cu O was slightly higher than
RD-Cu O at the beginning; however, with increasing the
reaction time, the former quickly deactivated, even worse than
the latter after 300 min. This indicated that the active sites on
RD-Cu
demonstrated that more than 45% Cl in NH
be lost with the temperature increasing to 250 °C, while
negligible Cl loss in Cl-RD-Cu O was observed (Figures S35-
S38, Table S4). In addition, Figure 3d displayed the three-
dimensional distribution of Cl in NH Cl-RD-Cu O and Cl-RD-
O by using Cu peak area as reference. For NH Cl-RD-
O, nearly no Cl signal could be detected deeper than 2.7 nm
O. As shown in Figure 3c, quasi in-situ XPS
2
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Cl-RD-Cu O would
4
2
4
2
2
2
NH
PO formation rate of Cl-RD-Cu
magnitude higher than that of RD-Cu
but also kept nearly unchangeable even after 420 min. Figure
b compared the conversion and PO selectivity of RD-Cu O,
Cl-RD-Cu O and NH Cl-RD-Cu O. The Cl modification by
post-treatment method would slightly decrease the C
conversion. At the initial stage, both Cl-RD-Cu O and NH
RD-Cu O exhibited ~60 % PO selectivity, nearly twice higher
than that of RD-Cu O (28 %). With increasing the reaction time,
slight decrease in the PO selectivity could be observed for Cl-
RD-Cu and RD-Cu O, while, for NH Cl-RD-Cu O,
4
Cl-RD-Cu
2
O was not stable (Figure S33). By contrast, the
O was not only one order of
O and NH Cl-RD-Cu O,
4
2
3
4
2
Cu
Cu
2
4
2
4
2
2
(Figure S39), while Cl would penetrate into the lattice as deep
as 18.8 nm for Cl-RD-Cu O. All these indicated that the post-
treatment method only decorated the Cu O with Cl from outside,
3
2
2
2
4
2
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
H
6
thus suffering from serious Cl loss, while the intergrowth
method gave rise to a lattice substitutional Cl, which was
stabilized by the Madelung potential. Indeed, DFT calculations
showed that the substitutional Cl over Cu O(110) is more stable
2
than the surface adsorbed Cl in terms of binding energy (-2.71
eV vs -0.63 eV).
2
4
Cl-
2
2
2
O
2
4
2
significant decay in PO selectivity was obtained at 420 min.
With elongating the reaction time, the PO formation rate of
Reaction mechanisms for DEP over Cu
Cu O(110).
2
O(110) and Cl-
2
NH
former had relatively low conversion. Interestingly, we found
that the PO selectivity of NH Cl-RD-Cu O would be recovered
when putting the used NH Cl-RD-Cu O again into the NH Cl
4 2 2
Cl-RD-Cu O became worse than that of RD-Cu O as the
To better understand the promotional effect of substitutional
Cl, DFT calculations were carried out (see Computational
details in Supporting Information). As reported previously,
DEP involved a complicate reaction network in which complete
combustion pathways would be always competitive with the
4
2
4
2
4
solvent, although the conversion could not be completely
recovered (Figure S34). This indicated that the host structure of
Cu O would not be destroyed during the catalysis.
2
5
,6,11,35
selective one.
Herein we would compare the energy
barriers of different pathways to evaluate the PO selectivity.
15
Similar to previous theoretical works, seven-layer slab with O
rich planes was used to model Cu
possible types of Cl species on/in the Cu
considered (Figure S40). DFT calculations showed that the
surface adsorbed Cl would block the adsorption of O , while the
2
O(110) surface. Herein, four
2
O(110) were
2
interstitial Cl was unstable since the space was too limit to
accommodate the Cl. Thus, we concentrated on the roles of the
substitutional Cl, including surface substitutional Cl (Cl-
2 2
Cu O(110)) and subsurface substitutional Cl (Clsub-Cu O(110))
(Figure S40-41). Bader charge analysis showed that replacing
the surface lattice oxygen by Cl increased the electron density
on Cu sites. The ratio of substituting Cl to Cu on the Cl-
2
Cu O(110) was calculated to be 1:8, close to 1:6.9 estimated by
XPS. Previously, lattice oxygen atoms were claimed to be
1
5
active, and the presence of Cl was proposed to reduce barrier
3
6
for the formation of epoxide product. Thus, we first
considered the epoxidation mechanism with lattice oxygen
atom. The reaction network, the optimized structures of
intermediates and key transition states (TSs) on Cl-Cu
2
O (110)
Figure 3. The durability of RD-Cu
NH Cl-RD-Cu O. (a) PO formation rate as a function of time
at 200 °C. (b) PO selectivity (Histogram) and conversion
Scatter plot) at 200 °C. (c) Cl loss of Cl-RD-Cu O and NH Cl-
RD-Cu O during the catalysis process. (d) The area ratio of
Cl/Cu peak with different etching depth of Cl-RD-Cu O and
NH Cl-RD-Cu O from HS-LEIS.
Quasi in-situ XPS and HS-LEISS were employed to further
illustrate the difference of Cl nature in NH Cl-RD-Cu O and Cl-
2 2
O, Cl-RD-Cu O and
and Cu O (110) surfaces were demonstrated in Figures S42-45.
2
4
2
According to our DFT calculations, the surface substitutional
Cl would reduce the energy barrier from oxametallopropylene
(
2
4
(OMP) intermediate to PO by 0.79 eV as compared with
2
unmodified Cu(110) (TS3ʹ shown in Figures S42-45), in line
36
2
with previous work. However, Cl doping would not change
4
2
the situation that -H bond oxidation was still competitive with
the C=C bond oxidation when the lattice oxygen was employed
(Figure S42).
4
2
ACS Paragon Plus Environment

Page 5 of 7
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. The proposed reaction network of propylene oxidation by adsorbed O
a) the schematic diagram of the reaction mechanism. X represented Cl atom or O atom for Cl-Cu
potential energy profiles of the reaction of propylene with O on Cl-Cu O (110) surface; (c) the comparison of energy barriers of DEP
on Cl-Cu O (110) (orange) and Cu O (110) (blue) surfaces.
2
on Cl-Cu
2
O (110) and Cu
2
O (110) surfaces.
(
2
O (110) or Cu
2
O (110); (b) the
2
2
2
2
Alternatively, we shifted our attention on the adsorbed
oxygen species. Figure 4 displayed the reaction mechanisms of
the catalytic cycle that involved the epoxidation of two
low temperature. In contrast, -H abstractions via TS2 needed
to conquer barriers of 0.89 eV and 1.01 eV for Cl-Cu O(110)
2
and Cu O(110), respectively (in Figure 4c). By comparing the
2
propylene molecules with one adsorbed O
2
(the optimized
energy differences between TS2 and TS1, it was clear that the
structures of intermediates and transition states were shown in
Figures S46-48). Experimentally, PO was not able to form in
substitutional Cl did enhance the selectivity.
2
The first PO formation consumed one O atom of adsorbed O ,
the absence of O
oxygen (Figure S49). Indeed, DFT calculations showed that the
adsorption energy of O on Cu of Cl-Cu O(110) (I) was
calculated as -0.66 eV, 0.09 eV more stable than that on
Cu O(110). We found that the adsorbed O species on both
surfaces were negatively charged and O-O bonds were
2
, highlighting the significant of the molecular
and the other O atom (IV) remained on the surface. In this case,
when the second propylene approached, OMP (V) would be
readily formed without any barrier, while forming allyl
intermediate (VI) via -H abstraction (TS4) had to pass through
a barrier of 0.2 eV ~ 0.4 eV. Thus, OMP would be the
predominant product under mild condition on both surfaces.
Starting from OMP, there existed three competitive pathways,
namely ring-closing (V to VII), -H abstraction (V to VIII) and
2
2
2
2
-
elongated, indicating the formation of superoxide species (O
2
).
-
Due to its electrophilic nature, O
2
preferred to attack the C=C
bond rather than the C-H bond. Thus, two possible routes
should be considered, namely oxygen insertion (from I to II in
1
,2 H shift (V to IX). It was worth noting that ring-closing
directly produced PO (VII) and the latter two routes gave rise
to allylic oxygen (VIII) and propanal (IX) intermediate,
respectively. The DFT calculations showed that ring-closing
(TS5) and -H abstraction (TS6) were competitive on Cu O
2
(110) with barriers of 0.55 eV and 0.50 eV, respectively, while
Figure 4a, b) and -H abstraction (from I to III). In the former
-
case, O
2
would directly interact with the C=C bond to produce
peroxirane intermediate (II) via TS1 by overcoming barriers of
.39 eV and 0.69 eV for Cl-Cu O(110) and Cu O(110),
respectively. Such a low barrier of TS1 for Cl-Cu O(110) nicely
explained the high TOF observed on Cl-RD-Cu O, especially at
0
2
2
2
1,2 H shift through TS7 was energetically unfavorable (1.16
eV). By contrast, for Cl-Cu O (110), the most favorable
2
2
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 7
pathway starting from OMP was ring-closing (TS5, 0.51 eV),
followed by -H abstraction (TS6, 0.89 eV) and 1, 2 H shift
(TS7, 1.52 eV) (Figure 4c).
(6) Huang, J.; Haruta, M. Gas-phase propene epoxidation over
coinage metal catalysts. Res. Chem. Intermed. 2012, 38, 1-24.
7) Lei, Y.; Mehmood, F.; Lee, S.; Greeley, J.; Lee, B.; Seifert, S.;
Winans, R. E.; Elam, J. W.; Meyer, R. J.; Redfern, P. C.; Teschner, D.;
Schlögl, R.; Pellin, M. J.; Curtiss, L. A.; Vajda, S. Increased silver
activity for direct propylene epoxidation via subnanometer size effects.
Science 2010, 328, 224-228.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
In addition, our DFT calculations demonstrated that the
subsurface substitutional Cl catalysts also steered the reaction
on the way to producing PO. However, the promotional effect
was less significant than that of the surface substitutional Cl
(8) Su, W.; Wang, S.; Ying, P.; Feng, Z.; Li, C. A molecular insight
2 2
into propylene epoxidation on Cu/SiO catalysts using O as oxidant. J.
Catal. 2009, 268, 165-174.
(Details are shown in Figure S50).
CONCLUSIONS
Herein we fully demonstrated the advantages of forming Cl-
Cu-O moiety by Cl substituting into Cu O nanocrystals, which
2
significantly promote the catalytic performance of DEP and
solve the long-exist Cl loss problem. Especially, the Cl-doping
effect and the crystal facet effect could work in synergy such
(9) Chu, H.; Yang, L.; Zhang, Q.; Wang, Y. Copper-catalyzed
propylene epoxidation by molecular oxygen: Superior catalytic
+
performances of halogen-free K -modified CuO
006, 241, 225-228.
10) Marimuthu, A.; Zhang, J.; Linic, S. Tuning selectivity in
x
/SBA-15. J. Catal.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
(
propylene epoxidation by plasmon mediated photo-switching of Cu
oxidation state. Science 2013, 339, 1590-1593.
(11) Torres, D.; Lopez, N.; Illas, F.; Lambert, R. M. Low-basicity
oxygen atoms: a key in the search for propylene epoxidation catalysts.
Angew. Chem. Int. Ed. 2007, 46, 2055-2058.
that Cl-decorated RD-Cu
2
O achieved the excellent DEP
performance. By means of comprehensive characterizations
and DFT calculations, we found that substitutional Cl facilitated
the formation of electrophilic oxygen species and the
production of PO, highlighting the importance to modulate an
active site through anionic doping. Finally, more attention
should be paid to the effects of surface/lattice impurities as they
would be inevitably introduced during the synthesis.
(12) Kulkarni, A.; Bedolla-Pantoja, M.; Singh, S.; Lobo, R. F.;
Mavrikakis, M.; Barteau, M. A. Reactions of propylene oxide on
supported silver catalysts: insights into pathways limiting epoxidation
selectivity. Top. Catal. 2012, 55, 3-12.
(13) Cowell, J. J.; Santra, A. K.; Lambert, R. M. Ultraselective
epoxidation of butadiene on Cu(111) and the effects of Cs promotion.
J. Am. Chem. Soc. 2000, 122, 2381-2382.
(14) Vaughan, O. P. H.; Kyriakou, G.; Macleod, N.; Tikhov, M.;
Lambert, R. M. Copper as a selective catalyst for the epoxidation of
propene. J. Catal. 2005, 236, 401-404.
ASSOCIATED CONTENT
Supporting Information.
The Supporting Information is available free of charge on the ACS
Publications website.
Experimental section, SEM images, TEM images, XPS, XRD, HS-
LEIS, and theoretical calculation (PDF)
(15) Hua, Q.; Cao, T.; Gu, X.-K.; Lu, J.; Jiang, Z.; Pan, X.; Luo, L.;
Li, W.-X.; Huang, W. Crystal-plane-controlled selectivity of Cu
2
O
catalysts in propylene oxidation with molecular oxygen. Angew. Chem.
Int. Ed. 2014, 53, 4856-4861.
(16) Lu, J. Q.; Luo, M. F.; Lei, H.; Bao, X. H.; Li, C. Epoxidation of
AUTHOR INFORMATION
Corresponding Author
propylene on NaCl-modified VCe1-xCu oxide catalysts with direct
x
molecular oxygen as the oxidant. J. Catal. 2002, 211, 552-555.
(17) Rocha, T. C. R.; Hävecker, M.; Knop-Gericke, A.; Schlögl, R.
*
zxxie@xmu.edu.cn; gfu@xmu.edu.cn
Author Contributions
These authors contributed equally.
Promoters in heterogeneous catalysis: the role of cl on ethylene
epoxidation over Ag. J. Catal. 2014, 312, 12-16.
(18) Ertl, G. Reactions at solid surfaces; John Wiley & Sons, 2010;
Vol. 14.
†
(19) Somorjai, G. A.; Li, Y. Introduction to surface chemistry and
Notes
catalysis; John Wiley & Sons, 2010.
(20) Williams, F. J.; Cropley, R. L.; Vaughan, O. P. H.; Urquhart, A.
J.; Tikhov, M. S.; Kolczewski, C.; Hermann, K.; Lambert, R. M.
Critical Influence of Adsorption Geometry in the Heterogeneous
Epoxidation of “Allylic” Alkenes:ꢀ Structure and Reactivity of Three
Phenylpropene Isomers on Cu(111). J. Am. Chem. Soc. 2005, 127,
17007-17011.
The authors declare no competing financial interests.
ACKNOWLEDGMENT
The authors acknowledge support from the National Key Research
and Development Program of China (2017YFA0206801,
2017YFA0207303), and the National Natural Science Foundation
of China (No. 21931009, 91845102 and 21773190). The authors
also thank Q. Kuang, Q. H. Zhang, Q. H. Xie, Z. B. Shen, K. Cheng,
S. J. Xie, L. Chen, S. Hu, Z. C. Lei for helpful discussions, and H.
H. Fang for help with in-situ FTIR measurements.
(
21) Bell, A. T. The impact of nanoscience on heterogeneous
catalysis. Science 2003, 299, 1688-1691.
22) Huang, W. Oxide Nanocrystal Model Catalysts. Acc. Chem. Res.
(
2
016, 49, 520-527.
(23) Kuang, Q.; Wang, X.; Jiang, Z.; Xie, Z.; Zheng, L. High-energy-
surface engineered metal oxide micro- and nanocrystallites and their
applications. Acc. Chem. Res. 2014, 47, 308-318.
REFERENCES
(24) Huang, W.-C.; Lyu, L.-M.; Yang, Y.-C.; Huang, M. H. Synthesis
(
1) Poliakoff, M.; Licence, P. Green chemistry. Nature 2007, 450,
810.
(2) Poliakoff, M.; Fitzpatrick, J. M.; Farren, T. R.; Anastas, P. T.
Green chemistry: science and politics of change. Science 2002, 297,
07-810.
3) Anastas, P.; Eghbali, N. Green chemistry: principles and practice.
of Cu O nanocrystals from cubic to rhombic dodecahedral structures
2
and their comparative photocatalytic activity. J. Am. Chem. Soc. 2012,
134, 1261-1267.
(25) Kuo, C.-H.; Huang, M. H. Morphologically controlled synthesis
8
of Cu O nanocrystals and their properties. Nano Today 2010, 5, 106-
2
(
116.
Chem. Soc. Rev. 2010, 39, 301-312.
(4) Myers, R. L. The 100 Most Important Chemical Compounds;
Greenwood Press: London, 2007.
(26) Huang, J.-Y.; Madasu, M.; Huang, M. H. Modified
Semiconductor Band Diagrams Constructed from Optical
Characterization of Size-Tunable Cu O Cubes, Octahedra, and
2
(
5) Khatib, S. J.; Oyama, S. T. Direct oxidation of propylene to
propylene oxide with molecular oxygen: a review. Catal. Rev. 2015,
7, 306-344.
Rhombic Dodecahedra. J. Phys. Chem. C 2018, 122, 13027-13033.
(27) Wu, S.; Yin, Z.; He, Q.; Lu, G.; Zhou, X.; Zhang, H.
5
Electrochemical deposition of Cl-doped n-type Cu O on reduced
2
graphene oxide electrodes. J. Mater. Chem. 2011, 21, 3467-3470.
ACS Paragon Plus Environment

Page 7 of 7
Journal of the American Chemical Society
(
28) Bai, Q.; Wang, W.; Zhang, Q.; Tao, M. n-type doping in Cu
with F, Cl, and Br: A first-principles study. J. Appl. Phys. 2012, 111,
23709.
29) Yu, L.; Ba, X.; Qiu, M.; Li, Y.; Shuai, L.; Zhang, W.; Ren, Z.;
Yu, Y. Visible-light driven CO reduction coupled with water oxidation
on Cl-doped Cu O nanorods. Nano Energy 2019, 60, 576-582.
2
O
ruthenium–copper-based catalysts for epoxidation of propylene. Catal.
Lett. 2018, 148, 586-600.
1
2
3
4
5
6
7
8
9
0
(34) Chen, G.; Zhao, Y.; Fu, G.; Duchesne, P. N.; Gu, L.; Zheng, Y.;
Weng, X.; Chen, M.; Zhang, P.; Pao, C.-W.; Lee, J.-F.; Zheng, N.
Interfacial effects in iron-nickel hydroxide–platinum nanoparticles
enhance catalytic oxidation. Science 2014, 344, 495-499.
(35) Düzenli, D.; Atmaca, D. O.; Gezer, M. G.; Onal, I. A density
(
2
2
(30) Ci, J.-W.; Tu, W.-C.; Uen, W.-Y.; Lan, S.-M.; Zeng, J.-X.; Yang,
T.-N.; Shen, C.-C.; Jhao, J.-C. Chlorine-Doped n-Type Cuprous Oxide
Films Fabricated by Chemical Bath Deposition. J. Electrochem. Soc.
2
functional theory study of partial oxidation of propylene on Cu O(001)
and CuO(001) surfaces. Appl. Surf. Sci. 2015, 355, 660-666.
2
014, 161, D321-D326.
31) Hochheimer, H. D.; Shand, M. L.; Potts, J. E.; Hanson, R. C.;
(36) Torres, D.; Illas, F.; Lambert, R. M. Towards an understanding
of promoter action in heterogeneously catalyzed ethene epoxidation:
why chlorine is the best halogen. J. Catal. 2008, 260, 380-383.
(
Walker, C. T. Pressure dependence of the Raman scattering by copper
halides. Phys. Rev. B 1976, 14, 4630-4636.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
32) Kaminow, I. P.; Turner, E. H. Temperature dependence of raman
scattering and the electro-optic properties of CuCl. Phys. Rev. B 1972,
, 1564-1578.
33) Seubsai, A.; Uppala, C.; Tiencharoenwong, P.; Chukeaw, T.;
Chareonpanich, M.; Zohour, B.; Noon, D.; Senkan, S. High stability of
5
(
Table of Contents artwork here
ACS Paragon Plus Environment