Silver-modulated SiO2-supported copper catalysts for selective hydrogenation of dimethyl oxalate to ethylene glycol

Article abstract of DOI:10.1016/j.jcat.2013.07.006

We present the application of a one-step urea-assisted gelation method to prepare a SiO₂-supported bimetallic catalyst composed of copper (Cu) and silver (Ag). Results show the remarkably enhanced performance of the catalyst for selective hydrogenation of dimethyl oxalate (DMO) to ethylene glycol (EG). Coupled with a series of characterization and kinetic studies, the improved activity is attributed to the formation of Cu nanoparticles containing Ag nanoclusters on the SiO₂ surface. The coherent interactions between the Cu and Ag species help create the active Cu⁺/Cu ⁰ species in a suitable proportion and prevent the transmigration of bimetallic nanoparticles during the hydrogenation process. The optimized CuAg/SiO₂ catalyst with an Ag/Cu atomic ratio of 0.05 has a balanced Cu⁺/Cu⁰ ratio and highly dispersed bimetal particles, which account for its high turnover frequency, EG selectivity of 97.0%, and excellent catalytic stability during the hydrogenation of DMO to EG for longer than 150 h.

Full text of DOI:10.1016/j.jcat.2013.07.006

Journal of Catalysis 307 (2013) 74–83
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Silver-modulated SiO -supported copper catalysts for selective
2
hydrogenation of dimethyl oxalate to ethylene glycol
a
b
a
a
b
b,
⇑
Ying Huang , Hiroko Ariga , Xinlei Zheng , Xinping Duan , Satoru Takakusagi , Kiyotaka Asakura ,
a,
⇑
Youzhu Yuan
a
State Key Laboratory of Physical Chemistry of Solid Surfaces and National Engineering Laboratory for Green Chemical Production of Alcohols-Ethers-Esters, College of Chemistry
and Chemical Engineering, Xiamen University, Xiamen 361005, China
Catalysis Research Center, Hokkaido University, Kita-ku N21W10, Sapporo, Hokkaido 001-0021, Japan
b
a r t i c l e i n f o
a b s t r a c t
Article history:
We present the application of a one-step urea-assisted gelation method to prepare a SiO -supported
bimetallic catalyst composed of copper (Cu) and silver (Ag). Results show the remarkably enhanced per-
formance of the catalyst for selective hydrogenation of dimethyl oxalate (DMO) to ethylene glycol (EG).
Coupled with a series of characterization and kinetic studies, the improved activity is attributed to the
2
Received 12 May 2013
Revised 3 July 2013
Accepted 8 July 2013
2
formation of Cu nanoparticles containing Ag nanoclusters on the SiO surface. The coherent interactions
+
0
between the Cu and Ag species help create the active Cu /Cu species in a suitable proportion and prevent
Keywords:
Copper
Silver
Hydrogenation
Dimethyl oxalate
Ethylene glycol
the transmigration of bimetallic nanoparticles during the hydrogenation process. The optimized CuAAg/
+
0
2
SiO catalyst with an Ag/Cu atomic ratio of 0.05 has a balanced Cu /Cu ratio and highly dispersed bimetal
particles, which account for its high turnover frequency, EG selectivity of 97.0%, and excellent catalytic
stability during the hydrogenation of DMO to EG for longer than 150 h.
Ó 2013 Elsevier Inc. All rights reserved.
1
. Introduction
Ethylene glycol (EG) is widely used as an antifreezing agent, a
The Cu-based catalysts have been studied extensively in the va-
por-phase hydrogenation of DMO. Cu-based catalysts allow for the
selective hydrogenation of carbon–oxygen bonds, while remaining
relatively inactive during carbon–carbon bond hydrogenolysis [6].
To date, CuACr catalysts remain the preferred industrial catalysts
for CTE because of their relatively high catalytic activity and dura-
bility [7–9]. Unfortunately, these catalysts require the use of toxic
chromium, which could endanger the safety of the environment
and the industrial workers. Therefore, different alternative carriers
component of polyester fiber, and a solvent, among others [1]. Gi-
ven the limited availability of crude oil resources and the increas-
ing demand for EG, novel approaches for the synthesis of EG are
being studied to substitute the production of EG derived from tra-
ditional petroleum [2,3]. The relatively green and economical coal
to EG (CTE) method is one possible alternative. The CTE process in-
volves the gasification of coal to syngas, followed by the coupling
of CO with nitrite esters to oxalates and the subsequent hydroge-
nation of oxalates to EG [1,4,5]. The selective hydrogenation of di-
methyl oxalate (DMO) to EG is one of the key reactions in the CTE
route that has drawn much attention. Several problems remain un-
solved despite considerable efforts in the study of DMO hydroge-
nation process. To date, EG selectivity and catalytic stability still
need improvement, the interaction with active sites remain con-
troversial, and the process of catalyst deactivation requires further
elucidation. A deeper understanding of the generic concepts is nec-
essary for more rational catalyst design.
(e.g., SiO
free Cu-based catalysts [10–12]. Among these carriers, the Cu/SiO
catalyst had the highest EG yield. Thus, the Cu/SiO catalyst was
considered a potential alternative for the conventional CuACr cat-
alyst. However, the Cu/SiO catalyst is relatively non-resistant to
2 2 3
, Al O , and ZnO) were evaluated for the synthesis of Cr-
2
2
2
sintering and is mechanically unstable for industrial operation
[6,13–16]. Thus, considerable effort has focused on modifying the
Cu/SiO catalyst using various methods, including the addition of
2
a second species [13–16], adoption of different preparation meth-
ods [3,17–21], and use of different silica structures for support
[22–26]. Cu, Ag, and Au are coinage metals belonging in the same
group in the periodic table. These metals possess a face-centered
cubic (fcc) structure, with close lattice parameters. In addition,
these metals have similar electronic structures and physical chem-
ical properties. These metals can also form intermetallic or alloy
phases with each other. The synergistic effects between the CuAAu
⇑
Corresponding authors. Fax: +81 11 706 9113 (K. Asakura), +86 592 2183047 (Y.
Yuan).
E-mail addresses: askr@cat.hokudai.ac.jp (K. Asakura), yzyuan@xmu.edu.cn
Y. Yuan).
[
27,28], CuAAg [29,30], and AuAAg [31,32] bimetallic catalysts
(
0
021-9517/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2013.07.006

Y. Huang et al. / Journal of Catalysis 307 (2013) 74–83
75
have been previously reported. Recent results showed that a bime-
tallic CuAAu catalyst with a large amount of Au could improve cat-
alyst performance during the hydrogenation of DMO to MG [33].
By contrast, a catalyst with a small amount of Au exhibited
remarkably enhanced activity and stability during the hydrogena-
tion of DMO to EG [34]. Spectroscopic studies revealed that the
CuAAu alloy nanoparticles (NPs) were formed on the catalyst sur-
faces, which were believed to retard the surface transmigration of
Cu species during hydrogenation. Furthermore, the ratio of surface
samples reached the preset temperatures for 30 min. Each diffrac-
tion pattern was identified by matching the results with reference
patterns included in the JCPDS data base. The full-width-at-half-
maximum of Cu (111) diffraction at a 2h of 43.2° was used to cal-
culate the Cu crystallite size using the Scherrer equation.
2
The hydrogen-temperature-programmed reduction (H -TPR) for
the as-calcined catalyst samples was carried out on a Micromeritics
Autochem II 2920 instrument connected to a Hiden Qic-20 mass
spectrometer (MS). A quartz U-tube was loaded with 100 mg of
the calcined catalyst and dried in an argon stream at 393 K for
+
0
Cu and Cu would vary, based on the amount of incorporated Au.
Consequently, different catalytic behaviors would be observed un-
der similar conditions. The catalytic performance of the CuAAg/
1 h. After cooling the catalysts to room temperature under an argon
À1
atmosphere, a flow of 5% H
2
–95% N
2
(50 mL min ) was fed into the
SiO
2
catalysts prepared by a deposition–precipitation method has
catalyst bed. The temperature was then ramped linearly from the
ambient temperature to 1073 K, at a rate of 10 K min . A 5A zeolite
À1
been evaluated for the DMO hydrogenation reaction, with MG as
the main product [35]. However, the structure and structure–activ-
ity relationships of these CuAAg catalysts remain ambiguous.
In this work, we incorporate a small amount of Ag into Cu/SiO
using a urea-assisted gelation method. Additionally, we found that
the bimetallic CuAAg/SiO catalysts could efficiently promote the
hydrogenation of DMO to EG. Kinetic and spectroscopic studies
show that an intimate interaction between Cu and Ag NPs, as well
as a cooperative effect between these, might be essentially respon-
sible for the observed enhanced catalysis.
trap was connected to the reactor outlet to remove any moisture.
The hydrogen consumption was simultaneously monitored by a
thermal conductivity detector (TCD) and MS.
Transmission electron microscopy (TEM) images were obtained
using a Tecnai F30 apparatus operated at 300 kV. The composition
analysis of each metal particle was performed using energy-disper-
sive X-ray spectroscopy (EDS) at scanning TEM (STEM) mode. The
powdered catalyst was dispersed in ethanol using ultrasound at
room temperature. The as-obtained solution was then dropped
into the carbon-coated molybdenum grids.
2
2
Ultraviolet–visible light (UV–vis) diffuse reflectance spectros-
copy (DRS) of the as-reduced catalysts was collected using a UV–
vis–NIR Cary 5000 scanning spectrophotometer. All catalyst pre-
2
. Experimental
2.1. Catalyst preparation
cursors were freshly reduced in a 5% H
2 2
–95% N atmosphere at
6
23 K for 4 h. The as-reduced samples were carefully collected un-
2
The bimetallic CuAAg/SiO catalysts with a preset metal loading
der an argon atmosphere at room temperature and sealed in glass
bottles before the UV–vis DRS measurements.
of 10 wt% were prepared using a urea-assisted gelation method
13]. Briefly, 4.5 g of 40 wt% LudoxAS-40 colloidal silica was dis-
persed in a 100 mL aqueous solution containing 6.0 g of urea, a
known amount of AgNO , Cu(NO O, and aqueous ammonia
Á3H
28 wt%) in a round-bottomed flask. The suspension was vigor-
[
The Cu and Ag K-edge XAFS measurements were performed
using the transmission mode at the BL12C and NW10 beam line
of the PF-AR and Photon Factory, Institute of Materials Structure
Science, High Energy Accelerator Research Organization (IMSS-
KEK), Japan (Proposal Nos. 2012G680 and 2012G644). The storage
ring was operated at 2.5 GeV and 450 mA using the top-up mode. A
Si (111) double crystal monochromator was used in the quick scan
3
)
3 2
2
(
ously stirred at 353 K in an oil bath for 4 h. The obtained precipi-
tate was separated by hot filtration, washed thrice with
deionized water, dried at 393 K overnight, and calcinated at
6
23 K for 4 h. The catalyst precursor was denoted as Cu
SiO , where x represents the atomic ratio of Ag and Cu.
A Cu/SiO catalyst with 10 wt% Cu and an Ag/SiO catalyst with
0 wt% Ag were similarly prepared using a urea-assisted gelation
method.
1 x
AAg /
2
mode. After reduction under 5% H –95% Ar at 623 K for 2 h, the
2
samples were sealed in glass cells under He (Cu) or Ar (Ag). The
XAFS spectra were then obtained at room temperature. The EXAFS
analysis was performed using the REX (version 2.5) program (Rig-
2
2
1
3
aku). The Fourier transformation of the k -weighted EXAFS oscilla-
tion from the k space to the r space was performed over the range
À1
2.2. Catalyst characterizations
of 30–140 nm to obtain a radial distribution function. The inver-
sely filtered Fourier data were analyzed using a non-linear least
À1
The N adsorption–desorption isotherms for the catalysts
2
squares curve fitting method in the k range of 30–140 nm . For
were measured at 77 K using a Micromeritics TriStar II 3020
porosimetry analyzer. The samples were degassed at 573 K for
the curve fitting analysis, the phase shift and amplitude functions
of the AgACu, AgAAg, and AgAO shells were calculated using the
FEFF program.
X-ray photoelectron spectroscopy (XPS) and Auger electron
spectroscopy (XAES) were performed using a JPS-9010MC photo-
3
h prior to the measurements. The specific surface area (SBET
was calculated using the Brunauer–Emmett–Teller (BET) method,
which adopted the isotherm data in a relative pressure (P/P
)
0
)
range of 0.05–0.2. The mesopore size distributions were evalu-
ated from the desorption branch of the isotherm using the Bar-
rett–Joyner–Halenda (BJH) method. The total pore volume
electron spectrometer equipped with an Al K
source (h = 1486.6 eV). Prior to the measurements, each sample
was pressed into a thin disk and pretreated in an atmosphere of
a X-ray radiation
m
4
depended on the absorbed N
approximately 0.99.
2
volume at a relative pressure of
5% H
2
–95% Ar (5 Â 10 Pa) at 623 K for 1 h in an auxiliary pretreat-
ment chamber. After the pretreatment, the sample was introduced
into the XPS chamber to avoid exposure to air. The XPS spectra of
as-calcined and pretreated samples were recorded at room tem-
perature. The binding energy (BE) was calibrated using C 1s peak
at 284.6 eV as reference with an uncertainty of ±0.2 eV.
X-ray diffraction (XRD) patterns for the catalyst samples were
obtained on a PANalytical X’pert Pro Super X-ray diffractometer
using Cu Ka radiation (k = 0.15418 nm) with a scanning angle
(
2h) ranging from 10° to 90°, tube voltage of 40 kV, and a current
of 30 mA. For the in situ XRD measurements, a 5% H
ture was introduced to the system at a flow rate of 50 mL min
2
–95% N
2
mix-
The Cu dispersion in each catalyst was determined by dissocia-
À1
.
tive N
2
O chemisorption and hydrogen pulse reduction using a
Temperature ramping programs were performed at room temper-
ature as well as at 523, 573, 623, 673, 723, 773, 873, and 973 K,
with a rate of 2 K min . The XRD patterns were collected after
Micromeritics Autochem II 2920 apparatus with a TCD. Typically,
100 mg of Cu/SiO
2
was calcined at 623 K, reduced in 5% H
2
–95%
À1
À1
N
2
(50 mL min ) at 623 K for 4 h, and cooled to 333 K. Pure N
2
O

7
6
Y. Huang et al. / Journal of Catalysis 307 (2013) 74–83
À1
(
30 mL min ) was then introduced for 30 min to completely oxi-
dize the surface Cu atoms into Cu O. The hydrogen pulse reduction
of the surface Cu O to metallic Cu was conducted at 573 K to en-
3. Results and discussion
2
2
3.1. Catalytic activity and stability
sure that the chemisorbed oxygen could immediately react with
the high-purity hydrogen supplied from a 0.4 mL loop. Excess
moisture was removed using a 5A zeolite dehydration trap. Hydro-
gen pulse-dosing was repeated until the pulse area remained con-
stant. The consumed amount of hydrogen was obtained by
subtracting the area of the first few pulses from the area of the
other pulses. The Cu loading of all the reduced catalysts was ana-
lyzed by inductively coupled plasma optical emission spectrome-
try (ICP-OES) using a Thermo Electron IRIS Intrepid II XSP. The Cu
dispersion was calculated by dividing the amount of chemisorption
sites by the total number of supported Cu atoms per gram of the
catalyst.
Vapor-phase DMO hydrogenation is known to comprise sev-
eral continuous reactions (Scheme 1), including DMO hydrogena-
tion to methyl glycolate (MG), MG hydrogenation to EG, and deep
hydrogenation of EG to ethanol (EtOH). Moreover, by-products of
1,2-butanediol (1,2-BDO) and 1,2-propanediol (1,2-PDO) are also
produced as a result of the dehydration reaction between EG
and EtOH or methanol [1,16]. To rigorously compare the catalytic
activity between Cu/SiO2 and Cu AAg /SiO , the DMO conversion
1
x
2
was limited to less than 35% by adjusting the DMO weight liquid
hourly space velocity (WLHSVDMO). The DMO conversion data
were then used to calculate the TOF according to the Cu disper-
M
sion (TOFCu) and total metal dispersion (TOF ), as listed in Table 1.
2.3. Catalytic testing
The TOFCu exhibited volcano-type improvement with the increas-
ing Ag content in Cu
lue (18.3 h ) with the Ag/Cu atomic ratio of 0.05:1. The trends in
1
AAg
x
/SiO
2
, which reached its maximum va-
À1
The catalytic performance for DMO hydrogenation was evalu-
ated using a fixed-bed microreactor equipped with a computer-
controlled auto-sampling system. Typically, 200 mg of the catalyst
precursor (40–60 meshes) was loaded into the center of the reac-
tor, with both sides of the catalyst bed packed with quartz powder
the TOF
M
of the catalysts were similar to those exhibited by
was slightly higher, as com-
TOFCu. However, the value of TOF
M
pared with TOFCu when the Ag/Cu atomic ratio was below 0.1.
Therefore, several interactions must have occurred between Ag
and Cu. Further increasing the Ag loading restrained the catalytic
activity. Additionally, this restriction was more evident when the
monometallic Ag catalyst was taken into consideration. The re-
sults demonstrated the synergistic effect between Cu and Ag dur-
ing the hydrogenation of DMO to EG.
DMO hydrogenation was conducted at 463 K under
WLHSVDMO of 1.05 h , a H
molar ratio of 80. The resulting bimetallic catalysts had more than
97% DMO conversion (Table S1 in the Supporting information).
However, the Cu/SiO
sion and 30.7% selectivity to EG, whereas the Ag/SiO
was almost inactive. That is, Ag incorporated with Cu/SiO
dramatically improve the EG yield, as compared with the monome-
tallic Cu/SiO or Ag/SiO . The Ag/Cu atomic ratio had a notably sig-
nificant effect on tuning the catalytic capability of DMO
hydrogenation. Introducing a very small amount of Ag (Ag/Cu
atomic ratio, 0.02; Ag loading, 0.26 wt%) in the Cu/SiO
(
40–60 meshes). The catalyst was activated under a 5% H
2
–95% N
2
À1
flow (50 mL min ) at 623 K for 4 h, with a ramping rate of
2
À1
K min . The catalyst was cooled to the desired reaction temper-
ature, and pure H
2
(99.999%) was allowed to pass through the cat-
À1
alyst bed. A DMO methanol solution (0.01 g mL ) was then
pumped into the reactor using a Series III digital HPLC pump (Sci-
entific Systems, Inc.) with a system pressure of 3.0 MPa. The outlet
stream was sampled using an automatic Valco 6-ports valve sys-
tem and analyzed using an online gas chromatograph (GC9790
Gas Chromatograph, Zhe Jiang Fu Li Analytical Instrument Co.,
Ltd.) with a flame ionization detector and a KB-Wax capillary col-
a
À1
2
pressure of 3.0 MPa, and a H
2
/DMO
2
catalyst demonstrated 87.2% DMO conver-
catalyst
could
2
umn (30 m  0.45 mm  0.85
lm) at intervals of 30 min.
2
The initial turn over frequency (TOF) of the reaction was mea-
sured when the DMO conversion was lower than 35%. The TOF va-
lue was based on the Cu dispersion or the number of surface metal
atoms estimated by the metal dispersion, according to an equation
in the literature [36]. This value indicated the moles of DMO con-
verted per hour by each mole of the metal on the catalyst surface
2
2
2
catalyst in-
creased the EG yield to 58.4%, which was twice as that of the
monometallic Cu/SiO catalyst. By increasing the Ag/Cu atomic
2
À1 À1
molDMO molmetal (surf)
À1
(
h
or simply h ).
O
O
O
O
HO
O
HO
+
2 H2
+ 2 H2
CH OH
+ H2
H O
CH OH
OH
OH
3
O
3
2
Scheme 1. Reaction pathway for the hydrogenation of DMO to MG, EG, and EtOH.
Table 1
TOF of Cu
1
AAg
x
/SiO
2
catalysts for DMO hydrogenation.a
Conversion (%)
b
À1
c
À1
Catalyst
Selectivity (%)
EG
TOFCu (h
)
M
TOF (h )
MG
Cu/SiO
2
16.0
22.3
33.8
24.5
16.4
3.7
4.8
8.2
8.8
8.1
5.1
1.6
95.2
91.8
91.2
91.9
94.9
98.4
10.1
13.5
18.3
16.1
13.3
–
11.3
15.1
20.6
16.0
12.1
5.5
Cu
1
Cu
1
Cu
1
Cu
1
AAg0.02/SiO
AAg0.05/SiO
AAg0.1/SiO
2
2
2
AAg0.2/SiO
2
d
Ag/SiO
2
a
b
c
À1
Reaction conditions: T = 463 K, P(H
TOFCu was calculated by Cu dispersion.
2
) = 3.0 MPa, H
2
/DMO molar ratio = 80, WLHSVDMO = 3.6 h
.
TOF
M
was calculated by metal dispersion.
d
À1
WLHSVDMO = 1.05 h
.

Y. Huang et al. / Journal of Catalysis 307 (2013) 74–83
77
ratio to 0.05, the catalyst produced the maximum EG yield of
5.9%. However, when the Ag/Cu atomic ratio was further in-
creased to 0.1, the catalyst had a lower EG yield.
The catalytic performance of the Cu/SiO and optimized Cu1-
AAg0.05/SiO catalysts were further investigated under different
WLHSVDMO values (Fig. 1). The conversion of DMO over the Cu1-
AAg0.05/SiO catalyst remained higher than 99% when the
WLHSVDMO varied from 0.3 h to 1.05 h . Subsequently, the con-
The long-term catalytic behaviors of the Cu/SiO
2
and optimized
8
Cu
1
AAg0.05/SiO catalysts were further evaluated at 463 K and
2
3.0 MPa under a H
0.6 h . The DMO conversion and EG selectivity of the Cu/SiO
2
/DMO molar ratio of 80 and a WLHSVDMO of
cat-
À1
2
2
2
alyst dramatically decreased after 100 h to approximately 60% and
10%, respectively (Fig. 2), whereas the Cu AAg0.05/SiO catalyst re-
tained its initial activity even after more than 150 h. These results
suggested that adding the proper amount of Ag to the Cu/SiO cat-
1
2
2
À1
À1
2
version over Cu/SiO
2
began to decline when WLHSVDMO was set at
alyst can remarkably enhance DMO hydrogenation and effectively
improve catalyst stability.
À1
À1
À1
0
.6 h . At a set of WLHSVDMO ranged from 0.3 h to 0.75 h , the
selectivity to EG was approximately 97% over the Cu
catalyst while the selectivity to MG and other by-products (EtOH,
,2-PDO, and 1,2-BDO) were lower than 1% and 2%, respectively.
However, the selectivity to EG over the Cu/SiO was lower than
0% and MG as the partial hydrogenation product became domi-
1 2
AAg0.05/SiO
3
3
.2. Catalyst characterizations
1
2
.2.1. Physicochemical properties of the catalysts
The metal loadings of the catalysts were determined using ICP-
5
À1
nant when the WLHSVDMO reached 0.75 h . In this case, the selec-
tivity to MG was 43.2% and almost no other by-products were
OES, as listed in Table 2. The actual loadings of Cu and Ag were
slightly lower, as compared with the theoretical value. This differ-
ence occurred because the weakly absorbed metallic ions on the
silica gel were eluted during the filtration process. However, the
actual atomic ratio of Ag/Cu in the catalysts was close to the preset
value. The physicochemical properties of the catalysts are summa-
rized in Table 2. The introduction of Ag had no obvious effect on
the surface area of the catalysts, whereas the pore volume and
average pore diameter were slightly increased when the Ag/Cu
atomic ratios were lower than 0.1. The values then decreased after
further introduction of Ag. The concurrent oxidation of metallic Ag
yielded over the Cu/SiO
WLHSVDMO, the selectivity to EG over the Cu
2
catalyst. Further increasing the
AAg0.05/SiO catalyst
1
2
was also decreased, accompanying with the increase in selectivity
to MG and suppression of excess hydrogenation or dehydration
products. The EG yield over Cu AAg /SiO catalysts was clearly a
1 x 2
function of WLHSVDMO (Fig. S1 in the Supporting information)
and easy to obtain, thereby indicating the significant influence of
the WLHSVDMO. Better catalytic performance with higher tolerance
2
to WLHSVDMO was observed with the bimetallic CuAAg/SiO cata-
lysts for the hydrogenation of DMO, as compared with the Cu/SiO
catalyst.
2
species while determining the Cu dispersion by N
ignored because the reaction of N O with the metallic Ag species
2
O at 333 K was
2
occurs at a relatively high temperature (658 K) [37] or requires
long reaction times (5 h) when the temperature is below 373 K
[38]. In addition, this phenomenon occurs when the Ag content
was very low in the bimetallic catalysts. The Cu dispersion of the
1
00
9
9
8
8
7
7
5
0
5
0
5
0
bimetallic CuAAg/SiO
persion initially increased and reached its maximum at an Ag/Cu
atomic ratio of 0.05, with a value of 43.1 m g . The Cu dispersion
then gradually decreased with the further increase in Ag loading.
The results suggested that the introduction of the proper amount
of Ag facilitated the higher dispersion of Cu particles. The Cu1-
AAg0.05/SiO had the highest Cu dispersion and notably presented
2
the optimal catalytic performance under identical reaction
conditions.
2
catalysts is summarized in Table 2. This dis-
2
À1
Cu/SiO2
Cu -Ag0.05/SiO2
1
3.2.2. TEM, EDS and XRD
0
.3
0.6
0.9
1.2
The morphologies and structural details of the bimetallic
catalysts were examined using TEM (Fig. 3). The metal nanoparti-
cles were distributed uniformly on the surfaces of the silica. The
−1
WLHSVDMO / h
1
00
1
00
8
6
4
2
0
0
0
0
0
8
6
4
2
0
0
0
0
0
Cu/SiO2
Conv. over Cu -Ag0.05/SiO2
1
Cu -Ag0.05/SiO
1
2
Selec. over Cu -Ag0.05/SiO2
1
Conv. over Cu/SiO2
Selec. over Cu/SiO2
0
.3
0.6
WLHSVDMO / h
0.9
1.2
0
25
50
75
100
125
150
−1
Time on stream / h
Fig. 1. DMO hydrogenation over Cu/SiO
2
and Cu
1
AAg0.05/SiO
2
catalysts as
a
Fig. 2. DMO hydrogenation over Cu/SiO
2
and Cu
1
AAg0.05/SiO
2
catalysts as a
function of WLHSVDMO. Reaction conditions: T = 463 K, P(H
molar ratio = 80.
2
2
) = 3.0 MPa, H /DMO
function of time on stream. Reaction conditions: T = 463 K, P(H
2 2
) = 3.0 MPa, H /
À1
DMO molar ratio = 80, WLHSVDMO = 0.6 h
.

7
8
Y. Huang et al. / Journal of Catalysis 307 (2013) 74–83
Table 2
1 x 2
Physicochemical properties of Cu AAg /SiO catalysts.
Contenta (wt%)
Ag/Cu atomic ratio
S
BET (m2
g
À1
)
V
pore (cm3
g
À1
)
Dpore (nm)
Metal dispersion (%)
Catalyst
Cu
Ag
By N
2
O
By TEM
Cu/SiO
2
8.7
8.3
7.9
7.3
6.6
–
–
–
225
217
225
217
188
109
108
0.82
0.90
0.90
0.96
0.63
0.45
0.34
12.5
13.8
13.7
14.9
12.1
15.2
9.9
35.3
38.6
43.1
40.5
36.3
–
31.6
33.8
38.4
37.2
33.8
7.1
Cu
1
Cu
1
Cu
1
Cu
1
AAg0.02/SiO
AAg0.05/SiO
AAg0.1/SiO
AAg0.2/SiO
2
2
0.26
0.65
1.2
2.0
9.1
–
0.018
0.048
0.097
0.18
–
2
2
Ag/SiO
SiO
2
2
–
–
–
–
a
Determined by ICP-OES analysis.
Fig. 3. TEM images of as-reduced catalysts. (a) Cu/SiO
2
, (b) Cu
1
AAg0.02/SiO
2
, (c) Cu
1
AAg0.05/SiO
2
, (d) Cu
1
2 1 2
AAg0.1/SiO , and (e) Cu AAg0.2/SiO .
particle diameter distribution showed that mean particle size of all
the samples was approximately 3 nm (Fig. S2 in the Supporting
information). High-resolution TEM (HRTEM) of a typical metal par-
ticle in the Cu AAg0.05/SiO catalyst (Fig. S3 in the Supporting
1 2
information) indicated that the 0.188 nm interval of lattice fringes
was smaller, as compared with those monometallic Ag (111)
on the selected fifteen metal particles. The STEM-EDS results re-
vealed that both Cu and Ag elements were detectable, but the
Ag/Cu atomic ratio was higher as compared with the present value
in most of the selected metallic particles, particularly when the ac-
tual Ag/Cu ratio was higher than 0.05. In other words, segregation
of the Ag species could occur in bulk particles.
(
0.236 nm), Ag (200) (0.204 nm), and Cu (111) (0.209 nm), but lar-
The XRD patterns showed that the as-calcined catalysts were
ger than Cu (200) (0.181 nm). This result may indicate the forma-
tion of the CuAAg alloy after reduction. The total metal dispersion
of the catalysts was calculated in accordance with the TEM, as
listed in Table 2. The total metal dispersion of the catalysts im-
proved with the increasing Ag/Cu atomic ratio, which agrees with
the observed trend of Cu dispersion. The dispersion of the Cu1-
amorphous, except for the Ag/SiO
mation). The introduction of Ag to the Cu/SiO
nificantly change the XRD patterns of the catalyst. Neither Cu nor
Ag species could be detected in these precursors, thereby implying
that Cu and Ag were highly dispersed on the porous silica support
by the urea-assisted gelation method. However, well-dispersed
2
(Fig. S5 in the Supporting infor-
catalyst did not sig-
2
2
À1
AAg0.05/SiO
2
catalyst reached its maximum at 38.4 m g
.
pure Ag/SiO
We then conducted in situ XRD characterization to monitor the
phase evolution of the Cu/SiO and optimized Cu AAg0.05/SiO
catalysts as the reduction temperature was increased under a 5%
–95% N atmosphere. As shown in Fig. 4, the characteristic peak
of Cu (111) (JCPDS 04-0836) at 2h of 43.2° was detected because
2
catalysts could not be prepared using this method.
The high-angle annular dark-field scanning transmission elec-
tron microscopy (HAADF-STEM) images of the as-reduced Cu1-
2
1
2
AAg
Supporting information). The tiny bright spots in the images corre-
spond to metal nanoparticles supported on the SiO . In combina-
x 2
/SiO bimetallic catalysts were shown in Fig. S4 (in the
H
2
2
2
2
+
tion with X-ray energy dispersive spectroscopy (X-EDS) system
equipped with a sub-nanometer probe, we can locate the conver-
gent electron beam (0.5 nm) at any position of the sample to obtain
more detail features of metal nanoparticles such as element distri-
bution and composition under HAADF-STEM mode. The composi-
tions of individual metallic particles on the several as-reduced
2
several Cu species on the Cu/SiO catalyst precursor were re-
duced to Cu0 when the reduction temperature reached 523 K.
According to the Scherrer equation, the Cu crystallite size of the
Cu/SiO
perature was lower than 773 K, whereas the Cu crystallite size of
the Cu AAg0.05/SiO catalyst was smaller than 3.2 nm at the same
reduction temperature range. The diffraction peaks of the
2
catalyst was smaller than 4 nm when the reduction tem-
1
2
1 x 2
Cu AAg /SiO catalysts were measured by the point EDS analysis

Y. Huang et al. / Journal of Catalysis 307 (2013) 74–83
79
0
0
were performed to investigate the reducibility of the as-calcined
samples with different Ag/Cu atomic ratios. A reduction peak for
♦
Cu
Cu size
1000
(a)
♦
♦
♦
9
73 K
the as-calcined Ag/SiO
porting information). The as-calcined Cu/SiO
2
catalyst was not obvious (Fig. S8 in the Sup-
presented a reduction
4
4
4
4
.5 nm
.1 nm
.0 nm
.0 nm
9
8
23 K
73 K
2
peak at 518 K, which was assigned to the reduction of highly dis-
persed CuO to Cu0 and copper phyllosilicate to Cu+ [3,39]. The
823 K
773 K
723 K
reduction peaks of the bimetallic samples gradually shifted to a
lower value with the increasing Ag/Cu atomic ratio. The introduc-
tion of Ag had the tendency to decrease the reduction temperature,
which was consistent with the results of Zhou et al. [30]. These re-
sults suggest that several significant interactions occurred between
the Cu and Ag species.
4
3
3
.0 nm
.9 nm
.6 nm
6
6
73 K
23 K
3
3
3
.6 nm
.6 nm
.5 nm
5
73 K
23 K
r.t.
5
3.2.4. UV–vis DRS
2
UV–vis DRS of the as-reduced Ag/SiO catalyst had a strong
2
0
40
60
80
absorption peak at 395 nm, which is typical for the well-known
surface plasmon resonance (SPR) band of Ag NPs (Fig. S9 in the
2θ / °
2
Supporting information). Meanwhile, Cu/SiO showed a very weak
0
0
1000
♦ Cu
(
b)
Cu size
and broad SPR band at approximately 600 nm. The UV–vis DRS
spectra of the bimetallic catalysts produced a single SPR band be-
tween 400 nm and 600 nm, thereby revealing that it was not a
physical mixture of the individual metals [40]. The increasing Ag/
Cu atomic ratio gradually shifted the SPR band into the blue range.
Therefore, the electronic structure of the bimetallic samples chan-
ged after the introduction of Ag. This change occurred because the
SPR bands of the metal particles are produced by the collective
oscillations of the conduction electrons. The d-band energy level
continuously changed because of the increasing amount of Ag
[40,41].
♦
♦
9
73 K
4
3
3
3
.1 nm
.5 nm
.5 nm
.4 nm
9
8
23 K
73 K
8
7
7
23 K
73 K
23 K
3
.2 nm
.1 nm
3
6
6
73 K
23 K
3
3
3
.0 nm
.0 nm
.0 nm
5
73 K
23 K
r.t.
60
5
<
3.0 nm
3.2.5. XPS and XAES
XPS was employed to investigate the surface valence state of
the as-reduced CuAAg/SiO catalysts. For the Cu/SiO catalyst,
2 2
the Cu 2p XPS peaks corresponding to Cu 2p3/2 and Cu 2p1/2 ap-
peared at 932.5 eV and 952.4 eV, respectively (Fig. S10 in the Sup-
porting information). Nevertheless, the BE of Cu 2p3/2 was shifted
to higher values with the increasing Ag content, thereby indicating
2
0
40
80
2θ / °
Fig. 4. In situ XRD patterns of as-calcined catalysts as a function of the reduction
temperature under a 5% H –95% N atmosphere. (a) Cu/SiO and (b) Cu
SiO
2
2
2
1
AAg0.05/
2
.
that the electronic effect was more obvious. The absence of the
Cu
of the pure Cu catalyst. Therefore, adding the proper amount of Ag
to the Cu/SiO catalyst could improve the Cu dispersion. This
observation was in accordance with the results of the dissociative
O chemisorption and TEM. When the reduction temperature un-
der the 5% H –95% N flow is further increased, the Cu diffraction
peaks of Cu/SiO and Cu AAg0.05/SiO become sharper. Thus, simi-
1
AAg0.05/SiO
2
catalyst were, to some extent, broader than those
2+
2
d ? 3d satellite peaks at 934.9 or 933.5 eV suggested that Cu
0
+
was successfully reduced to Cu and/or Cu after reduction at
2
6
23 K [13].
2
The Ag 3d XPS spectra of the as-reduced Ag/SiO showed two
N
2
peaks centered at 367.9 and 373.9 eV, which is typical for the
metallic Ag 3d5/2 and Ag 3d3/2 (Fig. S11 in the Supporting informa-
tion), as reported in the literature (367.9–368.1 eV and 373.9–
2
2
2
1
2
lar phase evolution behavior is observed in both catalysts.
After a long-term catalytic test, the catalysts were carefully col-
lected under a hydrogen atmosphere at room temperature and
sealed in glass bottles to protect the samples from oxidation. The
3
74.1 eV, respectively) [42]. In addition, the Ag 3d5/2 BE values of
Ag O and AgO range from 367.6 eV to 367.7 eV and 367.2 eV to
67.4 eV, respectively [42]. These reports are contrary to the typi-
2
3
cal positive core level BE shifts of metal cations in ionic materials.
As shown in Fig. S10, the Ag 3d5/2 peak of the as-reduced bimetallic
catalysts shifted to relatively higher BE values, as compared with
the monometallic Ag/SiO catalyst. These results implied that the
2
Ag species of the bimetallic catalysts had a greater tendency to-
wards electronic richness as compared with those of the monome-
tallic Ag/SiO
Ag O system. However, the positive shift in Ag may also occur due
to the smaller size cluster [43].
Given that the Cu and Cu species have almost similar BE val-
ues, the XAES spectra was conducted to further discriminate the
1 x 2
surface Cu and Cu species of the as-reduced Cu AAg /SiO cata-
lysts. Asymmetry and overlapping Auger peaks were observed in
the XAES spectra of Cu LMM (Fig. 5). These observations strongly
XRD diffraction peaks of the metallic copper in the Cu/SiO
AAg0.05/SiO catalysts were significantly different from each other
Fig. S6 in the Supporting information). The Cu diffraction peak of
the spent Cu/SiO catalyst at 2h of 43.2° was sharper, as compared
with that of the Cu AAg0.05/SiO catalyst. The changes in the Cu1-
AAg0.05/SiO catalyst before and after the long-term reaction were
2
and Cu1-
2
(
2
1
2
2
, according to the anomalous spectral shift of the Ag/
2
not significant. Furthermore, the TEM images of these catalysts re-
vealed consistent results (Fig. S7 in the Supporting information).
2
The transmigration and aggregation of Cu occurred on the Cu/SiO
2
0
+
catalyst after the long-term reaction. Thus, we speculated that Cu
agglomeration was one of the key factors that caused the deactiva-
tion of the catalyst. The addition of Ag was clearly important for
retarding the aggregation of the metallic Cu crystallites.
0
+
0
+
indicated that the Cu and Cu species coexisted in the reduced cat-
alysts. The deconvolution results are listed in Table 3. The surface
Cu and Cu distributions were significantly affected by the Ag/Cu
3
.2.3. H
In several cases, the presence of a noble metal affected the reduc-
ibility of the surface metal oxides. The H -TPR characterizations
2
-TPR
+
0
+
0
+
2
atomic ratio. The Cu /(Cu + Cu ) intensity ratio was initially

8
0
Y. Huang et al. / Journal of Catalysis 307 (2013) 74–83
+
Cu
0
Cu
A
(
e)
2
1
0
0
0
(
(
d)
c)
(c)
(
b)
(
b)
(
a)
(
a)
0
1
2
3
4
5
6
9
00
905
910
915
920
925
R / 0.1 nm
Kinetic energy / eV
Fig. 5. Cu LMM XAES spectra of the as-reduced Cu
different Ag/Cu atomic ratios. (a) Cu/SiO , (b) Cu AAg0.02/SiO
d) Cu AAg0.1/SiO , and (e) Cu AAg0.2/SiO
1
AAg
x
/SiO
, (c) Cu
2
catalysts with
2
1
2
1
AAg0.05/SiO
2
,
40
B
(
1
2
1
2
.
(
c)
enhanced until it reached its maximum value of 53.4% with the Ag/
Cu atomic ratio of 0.05, before it was gradually decreased with the
further increase in the Ag loading. The Auger parameters (APs) of
Cu and Cu+ were close to the previously reported values of
2
0
0
χ
(
b)
0
1
851.0 and 1847.0 eV, respectively [44].
Based on the XPS data, the surface Ag/Cu ratio was slightly high-
(
a)
er than the observed value in the samples with an extremely small
amount of Ag, such as Cu AAg0.02/SiO and Cu AAg0.05/SiO . In
1
2
1
2
addition, the surface Ag/Cu ratio was clearly lower than the actual
value when that was higher than 0.05. The results verified the pos-
sible segregation of the Ag species in bulk particles by STEM-EDS
0
2
4
6
8
10
12
14
−
1
k / 10 nm
(
Fig. S4 in the Supporting information). Nonetheless, the results
3
Fig. 6. (A) Fourier transforms and (B) k -weighted (k) of Ag K-edge EXAFS for (a) Ag
foil, (b) Ag/SiO , and (c) Cu AAg0.05/SiO
also demonstrated that Ag species influence the surface Cu NPs,
by facilitating contact between Cu and Ag atoms and providing
possible synergistic effects.
2
1
2
.
The structural parameters derived from the EXAFS curve fitting
analysis are listed in Table 4 and Table S2. For the Ag K-edge EXAFS
3.2.6. EXAFS
1 2
of the Cu AAg0.05/SiO , a CuAAg shell at a bond distance (R) of
Extended X-ray absorption fine structure (EXAFS) measure-
ments were conducted to further investigate the near-neighbor
atomic environment of the Cu and Ag atoms. The Fourier transfor-
0.264 nm with a coordination number (N) of 11 was observed
without considering the AgAAg distance by neglecting the AgAAg
contribution. When the AgAAg coordination was taken into con-
sideration, a AgACu shell at R = 0.263 nm with N = 10 and an
AgAAg shell at R = 0.289 nm with N = 2.7 were detected. Here,
the AgACu coordination number was corrected using the AgAAg
coordination number because of the absence of a good AgACu ref-
mations of the Ag K-edge EXAFS spectra of Cu
SiO , and Ag foil are shown in Fig. 6. The first nearest neighbor dis-
tance of the Ag/SiO was clearly close to that of the Ag foil. The Cu1-
AAg0.05/SiO catalyst showed a relatively broad peak at a position
lower than the peak for the Ag foil; this peak of at approximately
.230 nm was longer than that of Ag O (0.204 nm) [29], thereby
1 2
AAg0.05/SiO , Ag/
2
2
2
1 2
erence compound in this study. For the Cu AAg0.05/SiO catalyst,
0
2
the N of AgACu was 78% of the total N of Ag, which suggested that
most of the Ag had bonded with Cu to form the CuAAg alloy. Addi-
tionally, a small part of the Ag (ca. 22%) formed Ag clusters. The
implying the existence of an AgACu bond. For the Fourier transfor-
mations of the Cu K-edge EXAFS, the first nearest neighbor distance
of Cu/SiO
2
and the Cu
1
AAg0.05/SiO
2
catalyst showed no obvious dif-
much smaller AgAAg coordination number of the Cu
SiO
1
AAg0.05/
ferences from that of the Cu foil (Fig. 7).
2
(N = 2.7), as compared with the Ag foil (N = 12), indicated
Table 3
Cu LMM deconvolution results of Cu/SiO
2
and Cu
1
AAg
x
/SiO
2
catalysts.
KEa (eV)
Cu+
APb (eV)
Cu+
Cu 2p3/2 BE (eV)
X
Cu+ (%)
c
Catalyst
Ag/Cu atomic ratio
Cu0
Cu0
Cu/SiO
2
–
914.0
914.2
914.0
913.9
913.9
918.1
918.1
917.9
918.0
917.9
1846.5
1486.9
1486.8
1846.8
1846.8
1850.6
1850.8
1850.7
1850.9
1850.8
932.5
932.7
932.8
932.9
932.9
38.3
40.7
53.4
46.4
40.9
Cu
1
Cu
1
Cu
1
Cu
1
AAg0.02/SiO
AAg0.05/SiO
AAg0.1/SiO
AAg0.2/SiO
2
2
0.038
0.059
0.060
0.079
2
2
a
b
c
Kinetic energy.
Auger parameter.
+
+
0
Intensity ratio between Cu and (Cu + Cu ) by deconvolution of Cu LMM XAES spectra.

Y. Huang et al. / Journal of Catalysis 307 (2013) 74–83
81
5
4
3
2
1
0
0
0
0
0
0
Cu. For the Cu K-edge, the Cu
1
AAg0.05/SiO
2
sample had a CuACu
shell at R = 0.252 nm with N = 9.2. For the Ag content of 0.05 with
an AgACu shell of 10, as determined from the Ag K-edge and based
on the relation: NAgACu  CAg = NCuAAg  CCu, the calculated NCuAAg
A
was ca. 0.5. This value was too small to be detected. The Cu/SiO
2
(
(
c)
b)
sample consisted of a CuACu shell (N = 6.2 at R = 0.253 nm) and a
relatively weak CuAO shell (N = 0.8 at R = 0.183 nm), which im-
plied that the Cu species were somewhat oxidized.
3.2.7. Structure–performance relationship
The monometallic Cu/SiO catalyst with relatively low Cu dis-
2
persion showed poor catalytic activity and stability. After incorpo-
rating a small amount of Ag (x 6 0.05), the bimetallic catalyst
exhibited higher Cu dispersion and enhanced catalytic activity
and stability. The XRD patterns and TEM images of the catalysts
after long-term catalytic test showed that the addition of Ag could
effectively suppressed the aggregation of the Cu NPs during the
(
a)
0
1
2
3
4
5
6
R / 0.1 nm
2
long-term reaction process. Characterizations of EDS, H -TPR, UV–
B
6
4
2
0
0
0
0
vis DRS, XPS, and XAES revealed that the electronic interactions oc-
curred between Cu and Ag species, as reflected by the fact that the
(
c)
+
0
surface Cu and Cu distributions were significantly affected by the
+
0
+
(
b)
Ag/Cu atomic ratio. The Cu /(Cu + Cu ) intensity ratio was in-
creased from 38.3% to 53.4% after incorporating 0.65 wt% Ag
χ
(x = 0.05) to Cu/SiO
the Cu AAg0.05/SiO
catalyst. Moreover, the Ag K-edge EXAFS of
showed that a AgACu shell at R = 0.263 nm
2
1
2
(
a)
with N = 10 and a AgAAg shell at R = 0.289 nm with N = 2.7 were
detected when the AgAAg coordination was taken into account.
In other words, the Cu NPs were most probably embedded with
Ag nanoclusters. The coherent interactions between Cu NPs and
Ag nanoclusters could induce the formation of CuAAg alloys during
-
20
+
0
0
2
4
6
8
−
10
12
14
the pretreatment, with important roles for modulating the Cu /Cu
1
ratio and retarding the aggregation of Cu NPs during the reaction
process. With the further increase in the Ag/Cu atomic ratio
k / 10 nm
3
Fig. 7. (A) Fourier transforms and (B) k -weighted (k) of Cu K-edge EXAFS for (a) Cu
foil, (b) Cu/SiO , and (c) Cu AAg0.05/SiO .
2 1 2
(x P 0.1), excess Ag nanoclusters would cover or block the surface
of Cu NPs, which made the Cu dispersion decreased and the active
Cu species inaccessible. Based on the analysis of the structural char-
Table 4
Curve fitting analysis of Ag K-edge EXAFS of Cu
1
AAg0.05/SiO
2
, Ag/SiO
2
, and Ag foil.
1 x
acterization results, a possible schematic diagram of the Cu AAg /
Na
Rb (nm)
Ec (eV)
d
e
2
SiO catalysts with varied Ag/Cu atomic ratio is proposed in Fig. 8.
Sample
Shell
r
(nm)
R
f
Simultaneous existence of Cu and Cu0 species on silica-sup-
+
Cu
Cu
1
AAg0.05/SiO
AAg0.05/SiO
2
2
AgACu
AgACu
AgAAg
AgAAg
AgAAg
11
10
2.7
8.9
12
0.264
0.263
0.289
0.285
0.285
2.2
1.5
4.1
3.6
2.8
0.0082
0.0076
0.0083
0.0081
0.0082
0.029
0.012
ported Cu-based catalysts has been extensively discussed
1
[
3,13,25,45–47]. It is generally accepted that the synergistic effect
0
+
Ag/SiO
Ag foil
2
0.052
0.044
of Cu and Cu determines the catalytic performance in DMO
hydrogenation and a balanced distribution of Cu+ and Cu0 is one
a
b
c
of the key issues for obtaining outstanding hydrogenation perfor-
Coordination number.
Bond distance between absorber and backscatter atoms.
Inner potential correction to account for the difference in the inner potential
0
mance [3,13,25,45]. Poels and Brands [46] reported that the Cu
+
species dissociatively adsorbs H
2
, while the Cu species may stabi-
between the sample and the reference compound.
lize the methoxy and the acyl species, which are also important
d
Debye–Waller factor.
Residual factor.
+
e
intermediates in the hydrogenation of DMO. Moreover, Cu species
may function as electrophilic or Lewis acid sites to polarize the
C@O bond via the electron lone pair on oxygen [3,25], thus improv-
ing the reactivity of the ester group in DMO.
the presence of tiny metallic Ag clusters with sizes less than a few
nanometers [29]. Moreover, the total N of Ag was almost 12, there-
by indicating that the small Ag cluster was mainly surrounded by
The cooperative effect of Cu0 and Cu species is inevitably
+
1 x 2
believed to influence the catalytic performance of Cu AAg /SiO
Fig. 8. Schematic diagram of Cu
1 x
AAg /SiO
2
catalysts with different Ag/Cu atomic ratios.

8
2
Y. Huang et al. / Journal of Catalysis 307 (2013) 74–83
the excellent activity and stability of the bimetallic CuAAg/SiO
catalyst.
2
+
Cu
5
0
Acknowledgments
4
2
0
0
We gratefully acknowledge the financial support from the Na-
tional Basic Research Program of China (2011CBA00508), the Nat-
ural Science Foundation of China (20923004 and 21173175), the
Research Fund for the Doctoral Program of Higher Education
TOF byDM
(
20110121130002), the Program for Changjiang Scholars and Inno-
TOF byDCu
vative Research Team in University (IRT1036), and Cooperative Re-
search Program of Catalysis Research Center, Hokkaido University
1
0
(
Grant No. 10B0043).
0
.00
0.05
0.10
0.15
0.20
Ag/Cu atomic ratio
Appendix A. Supplementary material
+
0
+
Fig. 9. TOF and Cu /(Cu + Cu ) intensity ratio as a function of the Ag/Cu atomic
ratio.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2013.07.006.
catalysts. By combining the catalytic results with the Cu and metal
+
0
References
dispersion, as well as the Cu /Cu ratio, the effects of Ag/Cu atomic
+
0
ratio and the Cu /Cu ratio on the TOF can be discussed with re-
[
1] H.L. Yue, Y.J. Zhao, X.B. Ma, J.L. Gong, Chem. Soc. Rev. 41 (2012) 4218.
+
0
+
spect to the number of surface metal atoms. The Cu /(Cu + Cu )
intensity ratio and TOF both showed a volcano-type dependence
[2] R.A. Kerr, R.F. Service, Science 309 (2005) 101.
[3] L.F. Chen, P.J. Guo, M.H. Qiao, S.R. Yan, H.X. Li, W. Shen, H.L. Xu, K.N. Fan, J.
Catal. 257 (2008) 172.
+
0
+
on the Ag/Cu atomic ratio (Fig. 9). The Cu /(Cu + Cu ) ratio and
TOF value were proportionally increased with Ag content when
[
[
4] G.H. Xu, Y.C. Li, Z.H. Li, H.J. Wang, Ind. Eng. Chem. Res. 34 (1995) 2371.
5] Z.N. Xu, J. Sun, C.S. Lin, X.M. Jiang, Q.S. Chen, S.Y. Peng, M.S. Wang, G.C. Guo,
ACS Catal. 3 (2013) 118.
+
0
+
the Ag/Cu atomic ratio was less than 0.05. The Cu /(Cu + Cu ) ratio
and TOF value simultaneously reached their summits in the case of
Ag/Cu atomic ratio at 0.05. After that, both were decreased gradu-
ally. Thus, besides the coverage or blockage of Cu NPs by excess Ag,
[
[
[
6] D.S. Brands, E.K. Poels, A. Bliek, Appl. Catal. A: Gen. 184 (1999) 279.
7] L.R. Zehner, R.W. Lenton, US Patent 4 112 245, 1978.
8] F. Poppelsdorf, Eur. Patent 0 060 787, 1982.
[9] S. Tahara, K. Fujii, K. Nishihira, M. Matsuda, K. Mizutare, US Patent 4 453 026,
+
0
+
1984.
the balanced Cu /(Cu + Cu ) ratio was disturbed with relative high
content of surface Ag species, resulting in the decrease in catalytic
[
[
10] K. Hirai, T. Uda, Y. Nakamura, US Patent 4 614 728, 1986.
11] H. Miyazaki, K. Hirai, T. Uda, Y. Nakamura, S. Ikezawa, T. Tsuchie, US Patent 57
122 946, 1982.
+
activity. This dependence suggests that a balanced surface Cu and
0
[12] H. Miyazaki, T. Uda, K. Hirai, Y. Nakamura, H. Ikezawa, T. Tsuchie, US Patent 4
85 890, 1986.
13] Z. He, H.Q. Lin, P. He, Y.Z. Yuan, J. Catal. 277 (2011) 54.
[14] H.Q. Lin, X.L. Zheng, Z. He, J.W. Zheng, X.P. Duan, Y.Z. Yuan, Appl. Catal. A: Gen.
45–446 (2012) 287.
[15] A.Y. Yin, C. Wen, X.Y. Guo, W.L. Dai, K.N. Fan, J. Catal. 280 (2011) 77.
16] C. Wen, Y.Y. Cui, A.Y. Yin, K.N. Fan, W.L. Dai, ChemCatChem 5 (2013) 138.
[17] A.Y. Yin, X.Y. Guo, K.N. Fan, W.L. Dai, ChemCatChem 2 (2010) 206.
Cu concentration is important to achieve remarkable catalytic
5
performance. The results likewise demonstrate that the hydroge-
nation of DMO is a structure-sensitive reaction. Consequently, add-
ing the appropriate amount of Ag can be an effective way to
[
4
+
0
modulate the proportion of surface Cu and Cu to obtain excellent
activity and stability for the hydrogenation of DMO.
[
[
18] Y.Y. Zhu, S.R. Wang, L.J. Zhu, X.L. Ge, X.B. Li, Z.Y. Luo, Catal. Lett. 135 (2010)
75.
2
[
19] L. Lin, P.B. Pan, Z.F. Zhou, Z.J. Li, J.X. Yang, M.L. Sun, Y.G. Yao, Chin. J. Catal. 32
(2011) 957.
4
. Conclusions
[
[
[
20] S.R. Wang, X.B. Li, Q.Q. Yin, L.J. Zhu, Z.Y. Luo, Catal. Commun. 12 (2011) 1246.
21] S.R. Wang, L.J. Zhu, Y.Y. Zhu, X.L. Ge, X.B. Li, Chin. Chem. Lett. 22 (2011) 362.
22] C. Wen, A.Y. Yin, Y.Y. Cui, X.L. Yang, W.L. Dai, K.N. Fan, Appl. Catal. A: Gen. 543
(2013) 82.
23] X.Y. Guo, A.Y. Yin, W.L. Dai, K.N. Fan, Catal. Lett. 132 (2009) 22.
24] A.Y. Yin, X.Y. Guo, W.L. Dai, H.X. Li, K.N. Fan, Appl. Catal. A: Gen. 349 (2008) 91.
25] A.Y. Yin, X.Y. Guo, W.L. Dai, H.X. Li, K.N. Fan, J. Phys. Chem. C 113 (2009) 11003.
The incorporation of an appropriate amount of Ag into Cu/SiO
2
was significantly demonstrated to enhance the catalytic activity
and stability for the hydrogenation of DMO to EG. The EG yield
over the optimized catalyst Cu
as compared with the monometallic Cu/SiO
with a H
Cu AAg0.05/SiO
[
[
[
1
AAg0.05/SiO
2
was 2.2 times higher,
at 463 K and 3.0 MPa,
2
[26] A.Y. Yin, X.Y. Guo, W.L. Dai, K.N. Fan, J. Phys. Chem. C 114 (2010) 8523.
[27] C.L. Bracey, P.R. Ellis, G.J. Hutchings, Chem. Soc. Rev. 38 (2009) 2231.
[
[
À1
2
/DMO molar ratio of 80, and a WLHSVDMO of 1.05 h . The
28] X.Y. Liu, A.Q. Wang, L. Li, T. Zhang, C.Y. Mou, J.F. Lee, J. Catal. 278 (2011) 288.
29] K. Shimizu, K. Shimura, M. Nishimura, A. Satsuma, RSC Adv. 1 (2011) 1310.
1
2
catalyst can maintain almost 100% DMO conver-
sion and more than 97% EG selectivity for over 150 h under the
optimized conditions.
[30] J.X. Zhou, L.Y. Guo, X.W. Guo, J.B. Mao, S.G. Zhang, Green Chem. 12 (2010)
835.
31] J.W. Zheng, H.Q. Lin, Y.N. Wang, X.L. Zheng, X.P. Duan, Y.Z. Yuan, J. Catal. 297
2012) 110.
[32] A. Sandoval, A. Aguilar, C. Louis, A. Traverse, R. Zanella, J. Catal. 281 (2011) 40.
1
[
A series of spectroscopic studies were used to explore the struc-
(
2
ture of the catalysts, including N O chemisorption, TEM, in situ
[
[
33] A.Y. Yin, C. Wen, W.L. Dai, K.N. Fan, J. Mater. Chem. 21 (2011) 8997.
34] Y.N. Wang, X.P. Duan, J.W. Zheng, H.Q. Lin, Y.Z. Yuan, H. Ariga, S. Takakusagi, K.
Asakura, Catal. Sci. Technol. 2 (2012) 1637.
XRD, EDS, H -TPR, UV–vis DRS, XPS, XAES, and EXAFS. Introducing
2
the appropriate amount of Ag facilitates the higher dispersion of Cu
NPs. This higher dispersion is caused by the formation of Cu NPs
containing Ag nanoclusters. The coherent interactions between
the Cu and Ag species induce the formation of CuAAg alloys. The
distribution of surface Cu and Cu0 species can be modulated by
varying the Ag/Cu atomic ratio. Therefore, the interactions be-
tween Cu and an appropriate amount of Ag species can help sus-
[35] B.W. Wang, Q. Xu, H. Song, G.H. Xu, J. Nat. Gas Chem. 16 (2007) 78.
[36] J.R. Anderson, Structure of Metallic Catalysts, Academic Press, New York, 1975.
[
37] T. Furusawa, K. Seshan, J.A. Lercher, L. Lefferts, K. Aika, Appl. Catal. B: Environ.
7 (2002) 205.
38] J.J.F. Scholten, J.A. Konvalinka, F.W. Beekman, J. Catal. 28 (1973) 209.
3
+
[
[39] C.J.G. Van Der Grift, A. Mulder, J.W. Geus, Appl. Catal. 60 (1990) 181.
[
40] H.T. Beyene, V.S.K. Chakravadhanula, C. Hanisch, T. Strunskus, V.
Zaporojtchenko, M. Elbahri, F. Faupel, Plasmonics 7 (2012) 107.
41] U. Kreibig, M. Vollmer, Optical Properties of Metal Clusters, Springer, Berlin,
1995.
+
0
tain
a
suitable Cu /Cu
proportion and restrain the
[
transmigration of copper NPs, which are largely responsible for

Y. Huang et al. / Journal of Catalysis 307 (2013) 74–83
83
[
42] C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, in: G.E. Muilenberg (Ed.),
Handbook of X-ray Photoelectron Spectroscopy, Perkin–Elmer, Physical
Electronics Division, Eden Prairie, Minnesota, 1979.
43] M.G. Mason, Phys. Rev. B 27 (1983) 748.
44] K.P. Sun, W.W. Lu, F.Y. Qiu, S.W. Liu, X.L. Xu, Appl. Catal. A: Gen. 252 (2003)
[45] J.L. Gong, H.R. Yue, Y.J. Zhao, S. Zhao, L. Zhao, J. Lv, S.P. Wang, X.B. Ma, J. Am.
Chem. Soc. 134 (2012) 13922.
[46] E.K. Poels, D.S. Brands, Appl. Catal. A: Gen. 191 (2000) 83.
[47] V. Gutierreza, M. Dennehyb, A. Diezb, M.A. Volpea, Appl. Catal. A: Gen. 437
(2012) 72.
[
[
2
43.