Promoting effect of boron oxide on Cu/SiO2 catalyst for glycerol hydrogenolysis to 1,2-propanediol

Article abstract of DOI:10.1016/j.jcat.2013.03.018

Cu/SiO₂ catalyst has been extensively applied in glycerol hydrogenolysis for its high selectivity to 1,2-propanediol, while suffering from severe deactivation easily. B₂O₃ is frequently used as an additive for stabilizing active species. Thus, a series of Cu/SiO₂ catalysts with various B₂O₃ loadings for glycerol hydrogenolysis were prepared via precipitation-gel method followed by impregnation with boric acid. These catalysts were fully characterized by ICP, BET, XRD (in situ XRD), N₂O chemisorption, H₂-TPR, NH ₃-TPD, IR, Raman, XPS, and TEM. Addition of B₂O ₃ to Cu/SiO₂ can greatly restrain the growth of copper particles and promote the dispersion of copper species upon calcination, reduction and reaction, which resulted in the enhanced catalytic activity and stability. The optimal 3CuB/SiO₂ reached complete conversion with 98.0% 1,2-propanediol selectivity. The strong correlation between 1,2-propanediol yield and Cu surface area gave direct evidence that the active Cu species were the primary active sites for glycerol hydrogenolysis.

Full text of DOI:10.1016/j.jcat.2013.03.018

Journal of Catalysis 303 (2013) 70–79
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Promoting effect of boron oxide on Cu/SiO2 catalyst for glycerol hydrogenolysis
to 1,2-propanediol
Shanhui Zhu a,b, Xiaoqing Gao c, Yulei Zhu a,c, , Yifeng Zhu a,b, Hongyan Zheng c, Yongwang Li a,c
⇑
a State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, PR China
b Graduate University of Chinese Academy of Sciences, Beijing 100039, PR China
c Synfuels China Co. Ltd., Taiyuan 030032, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Cu/SiO2 catalyst has been extensively applied in glycerol hydrogenolysis for its high selectivity to 1,2-pro-
panediol, while suffering from severe deactivation easily. B2O3 is frequently used as an additive for sta-
bilizing active species. Thus, a series of Cu/SiO2 catalysts with various B2O3 loadings for glycerol
hydrogenolysis were prepared via precipitation-gel method followed by impregnation with boric acid.
These catalysts were fully characterized by ICP, BET, XRD (in situ XRD), N2O chemisorption, H2-TPR,
NH3-TPD, IR, Raman, XPS, and TEM. Addition of B2O3 to Cu/SiO2 can greatly restrain the growth of copper
particles and promote the dispersion of copper species upon calcination, reduction and reaction, which
resulted in the enhanced catalytic activity and stability. The optimal 3CuB/SiO2 reached complete conver-
sion with 98.0% 1,2-propanediol selectivity. The strong correlation between 1,2-propanediol yield and Cu
surface area gave direct evidence that the active Cu species were the primary active sites for glycerol
hydrogenolysis.
Received 3 February 2013
Revised 19 March 2013
Accepted 21 March 2013
Keywords:
Glycerol
Hydrogenolysis
1,2-Propanediol
Boric oxide
Copper
Ó 2013 Elsevier Inc. All rights reserved.
1. Introduction
1,2-propanediol (1,2-PDO) which is an important commodity
chemical in the production of polyester resins, pharmaceuticals,
Declining fossil fuel reserves and increasing environmental is-
sues have stimulated intensive interest in developing renewable
biomass and biofuels [1,2]. In this context, biodiesel emerges as a
promising alternative to conventional fossil fuels for its non-toxic,
carbon natural, and biodegradable properties. Biodiesel is currently
produced by the transesterification of triglycerides with methanol
or ethanol, which concurrently manufactures large amounts of by-
product glycerol equivalent to ca. 10 wt.% of the overall biodiesel
production [3]. With the steady growth of biodiesel industry, huge
surplus of glycerol has been currently produced, which makes the
price of glycerol depreciate sharply. Moreover, the biomass-deri-
vate glycerol is a nontoxic, edible, and sustainable compound con-
taining highly multifunctional hydroxyl groups, making it as a
versatile platform chemical for producing valuable derivatives
[2,3]. Accordingly, significant work has been done to the transfor-
mation of glycerol by various catalytic processes, such as oxidation
[4], hydrogenolysis [5–8], dehydration [9], reforming [10], acetal-
ization [11], and esterification [12]. Among them, one of the most
promising approaches is the catalytic hydrogenolysis of glycerol to
tobacco humectants, paints, cosmetics, and antifreeze [13]. This
process offers a sustainable and economically competitive route
for the synthesis of 1,2-PDO from biomass-derived glycerol instead
of petroleum-based propylene oxide.
Supported Pt, Ru, Ir, Rh, Pd, Ni, and Cu catalysts have been com-
monly employed for glycerol hydrogenolysis to 1,2-PDO [14–20].
Despite the high activity of noble metal and Ni-based catalysts,
Cu-based catalysts generally afford excellent selectivity for pro-
ducing 1,2-PDO due to their intrinsic ability to cleave the C–O bond
in preference to the C–C bond in glycerol hydrogenolysis, which is
regarded as crucial for 1,2-PDO formation. Various supports
including SiO2 [21], MgO [22], ZnO [23], Al2O3 [24], Cr2O3 [13] as
well as zeolite [25] have been exhibited for Cu-based catalysts.
Bienholz et al. [23] obtained a highly active Cu/ZnO catalyst pre-
pared by oxalate gel method but with strong deactivation in water
which was an unavoidable by-product of this reaction. Cu/Cr2O3
[13] was identified as the most effective catalyst for relatively high
performance and stability, while the toxic chromium limited its
wide application for unresolved environmental issues. Among all
the Cu-based catalysts investigated, Cu/SiO2 [21,26–28] has been
extensively studied and attracted considerable attention owing to
its excellent selectivity, green benefit, and low cost in glycerol
hydrogenolysis to 1,2-PDO. Several techniques [21,26–28] involv-
ing sol–gel, impregnation, ion exchange, co-precipitation, precipi-
⇑
Corresponding author at: State Key Laboratory of Coal Conversion, Institute of
Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, PR China. Fax: +86
351 7560668.
E-mail address: zhuyulei@sxicc.ac.cn (Y. Zhu).
0021-9517/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2013.03.018

S. Zhu et al. / Journal of Catalysis 303 (2013) 70–79
71
tation-deposition, and precipitation gel have been utilized to pre-
pare Cu/SiO2 catalysts. Nevertheless, there are still numerous tech-
nical problems to be addressed or disadvantages that need to be
overcome. One of the most complicated issues is insufficient cata-
lytic activity and stability for glycerol hydrogenolysis. Vasiliadou
et al. [27] systematically investigated the deactivation behavior
of Cu/SiO2 catalyst in glycerol hydrogenolysis and suggested that
the deactivation was mainly attributed to the agglomeration of ac-
tive metallic phase.
ICP optical emission spectroscopy (Optima2100DV, PerkinEl-
mer) was exhibited to check the chemical compositions of as-pre-
pared samples.
Powder X-ray diffraction (XRD) patterns were recorded with a
D2/max-RA X-ray diffractometer (Bruker, Germany) operating
with Cu K
a radiation at 30 kV and 10 mA with a scanning angle
(2h) ranging from 10° to 90° at the scanning rate of 5°/min. The
mean crystallite size of Cu was calculated by Scherrer equation
according to the full width at half maximum (FWHM) of Cu
(111) diffraction at 2h of 43.2°. For in situ XRD measurement, the
sample remained in pure H2 at a flow rate of 30 cm3 minꢁ1. Tem-
perature-ramping programs were exhibited from 25 °C to 150,
To obtain highly dispersed and stable Cu nanoparticles has been
supposed to be a dominant factor in achieving excellent catalytic
performance for glycerol hydrogenolysis to 1,2-PDO. Modification
of Cu-based catalyst by an effective promoter is needed to improve
the catalytic activity and stability owing to the enhanced dispersion
of Cu, specific surface area, and interaction between Cu and dopant.
Zheng et al. [29] claimed that B2O3-modified Ni/SiO2 catalyst pos-
sessed high metal dispersion together with suitable acidity and pre-
sented outstanding catalytic performance for hydrogenation of
ethylbenzene because of the strong interaction between surface
nickel and boron species. B2O3 has emerged as a promising candi-
date for its sintering-resistant property and suitable acidity
[29,30]. Appropriate acid components are essential for glycerol
hydrogenolysis according to the dehydration–hydrogenation
bifunctional mechanism, as elucidated by our previous reports
[6,7]. Nevertheless, no literature has been published regarding
structure-behavior correlation of B2O3-modified Cu/SiO2 catalysts
for glycerol hydrogenolysis. Thus, a series of B2O3-doped Cu/SiO2
catalysts prepared by precipitation-gel procedure followed by
impregnation with boric acid were first reported for selective
hydrogenolysis of glycerol to 1,2-PDO. These catalysts were deeply
characterized by various techniques including ICP, BET, XRD (in situ
XRD), N2O chemisorption, H2-TPR, NH3-TPD, IR, Raman, XPS, and
TEM. The structure–activity relationship of these B2O3-doped Cu/
SiO2 catalysts in glycerol hydrogenolysis was also discussed in
detail.
200, 250, 300, 350, and 400 °C at a heating rate of 5 °C minꢁ1
.
The XRD patterns were obtained after the sample reached the pre-
set temperature for 30 min.
H2-TPR measurements were conducted in Auto Chem. II 2920
equipment (Mircromeritics, USA) using 0.10 g sample for each run.
Typically, the catalyst was pretreated in Ar at 150 °C for 30 min
and then cooled to 30 °C. Subsequently, the Ar flow was switched
to 10%H2–Armixedgas anda coldtrapof 2-propanol-liquid nitrogen
slurry was added to condense the water gas. The H2-TPR was started
from 30 °C to 500 °C at a heating rate of 5 °C/min and simultaneously
monitored by a thermal conductivity detector (TCD).
N2O chemisorption was carried out in the same apparatus as
above. The catalyst sample was placed into a U-shaped quartz tube
and reduced according to the same H2-TPR procedure described
above. The specific surface area of Cu was performed by dissocia-
tive N2O adsorption on the surface of copper with the pulse titra-
tion method based on the following equations: 2Cu
(s) + N2O ? N2 + Cu2O (s). Pure nitrogen was employed to detect
the consumption of N2O. The specific surface area of Cu was esti-
mated from the total amount of N2O consumption with
1.46 ꢂ 1019 copper atoms per square meter.
NH3-TPD was performed in the same apparatus as H2-TPR. Prior
to the adsorption of NH3, the catalyst (0.30 g) was pretreated in He
at 400 °C to clean the surface from moisture and other adsorbed
gases for 1 h. After cooling to 100 °C, the catalyst was saturated
with pure NH3 for 30 min and then purged with He to remove
the physisorbed NH3 for 30 min. Subsequently, the sample was
heated to 700 °C at a ramp rate of 5 °C/min and the NH3 desorption
was detected by a mass spectrometer (MS, Agilent).
The IR spectra were carried out on a Vertex 70 (Bruker) FT-IR
spectrophotometer in the range of 400–4000 cmꢁ1. The powder
samples were mixed with KBr (2 wt.%) and pressed into translu-
cent disks at room temperature.
Raman spectroscopy was performed on a Renishaw-UV–vis Ra-
man System 1000 using a CCD detector at room temperature. The
532 nm of the air-cooled frequency-doubled Nd–Yag laser was em-
ployed as the exciting source with a power of 30 MW.
X-ray photoelectron spectroscopy (XPS) and Auger electron
spectroscopy (XAES) were measured on a VG MiltiLab 2000 spec-
2. Experimental
2.1. Catalyst preparation
The Cu/SiO2 catalyst with 20 wt.% CuO loading was prepared by
precipitation-gel method, according to previous reports [21,26]. A
28 wt.% ammonia (Sinopharm Chemical Reagent Co., Ltd., China
(SCRC)) aqueous solution was dripped into Cu(NO3)2ꢀ3H2O
(0.5 mol/L, SCRC) aqueous solution under vigorous stirring at
80 °C until the pH of mixed solution reached 6.5. Subsequently, a
calculated amount of colloidal silica aqueous solution (SiO2,
30 wt.%, Qingdao Ocean Chemical CO., Ltd., China) was added into
the remaining solution of the precipitate and then a gel was
formed. After that, the gel was aged at 80 °C for 4 h, filtrated,
washed with hot deionized water for 5 times, and dried overnight
at 80 °C to form the powder of Cu/SiO2 precursor. The B2O3-modi-
fied Cu/SiO2 catalysts were prepared by incipient wetness impreg-
nation of Cu/SiO2 precursor with aqueous solutions containing the
desired amount of H3BO3 (SCRC). After impregnation, these sam-
ples were dried overnight at 80 °C and then calcined at 400 °C in
static air for 4 h. The obtained catalysts are designated as xCuB/
SiO2, where the x represents the nominal mass loading of boron.
trometer with Mg Ka radiation and a multichannel detector. Prior
to each test, the calcined sample was reduced in H2 at 250 °C for
2 h. The obtained binding energies were calibrated using the C1s
peak at 284.6 eV as the reference. The experiment error was given
within 0.1 eV.
TEM measurement was performed with a JEM-1011 electron
microscope operating at an 80 kV voltage. The samples were sus-
pended in ethanol with an ultrasonic dispersion for 20 min and
deposited on copper grids coated with amorphous carbon films.
2.2. Catalyst characterization
N2 adsorption–desorption isotherms were measured at –196 °C
using a Micromeritics ASAP 2420 instrument after degassing at
300 °C to remove physically adsorbed impurities for 8 h in vacuum.
BET surface area and BJH pore size distribution were calculated
according to the desorption branch of the isotherms.
2.3. Catalytic test
Hydrogenolysis of glycerol was performed in a vertical fixed-
bed reactor (i.d. 12 mm, length 600 mm) made of a stainless steel
tube. Typically, 4.0 g catalyst (20–40 mesh) was loaded into the

72
S. Zhu et al. / Journal of Catalysis 303 (2013) 70–79
constant temperature section of the reactor between two layers of
quartz sand. Prior to each test, the catalyst was in situ reduced in
flowing H2 (100 cm3 minꢁ1) at 250 °C for 2 h at atmospheric pres-
sure. After reduction, the system was cooled to 200 °C and pres-
sured to 5.0 MPa. A 10 wt.% glycerol aqueous (or 2-propanol)
solution was continuously pumped into the reactor with a high-
pressure pump along with co-feeding H2 of gas flowing at
150 cm3 minꢁ1. The liquid and gas products were condensed and
collected in a gas–liquid separator immersed in an ice-water trap.
The liquid products were determined by a gas chromatography
(GC, Ruihong chromatogram analysis Co., Ltd., China) using a cap-
illary column (DB-WAX, 30 m ꢂ 0.32 mm) and a flame ionization
detector. The tail gas was off-line analyzed by a GC (Huaai chro-
matogram analysis Co., Ltd., China) using a capillary column (OV-
101, 60 m ꢂ 0.25 mm) and a TCD. The obtained products were also
identified by GC–MS (6890N, Agilent, USA). The conversion of glyc-
erol and selectivity of products were calculated based on the fol-
lowing equations:
Cu dispersion (23.9%) and lowest particle size (4.2 nm). However,
further increase in B2O3 loading (5CuB/SiO2) declined the Cu dis-
persion mildly, which was most probably ascribed to partial cover-
ing effect of surplus B2O3 on the Cu surface (see below) [30].
3.1.2. Structural properties of the catalysts
The XRD patterns of Cu/SiO2 catalysts with different B2O3 load-
ings after the calcination at 400 °C are displayed in Fig. 1A. The dif-
fraction peaks at around 35.6° and 38.8° are assigned to CuO phase
[31]. With the increasing B2O3 loading, the intensity of diffraction
peaks for CuO declined notably, and the peak width broadened
gradually. These results meant that the crystallization degree of
CuO and particle size calculated by Scherrer equation monotoni-
cally decreased with an increase in B2O3 content. Moreover, the
introduction of B2O3 resulted in a slight shift of CuO diffraction
peaks toward higher 2h values. A detailed study of the predomi-
nant 32.0–42.0° reflections confirmed this effect, showing the con-
tinuous shift toward lower lattice parameters with an increase in
boron content. This behavior was indicative of structure defects
on the surface which might be originated from the incorporation
of partial B3+ into Cu2+ cation [31]. Almost no peaks of B2O3 were
detected, most probably due to the homogeneous distribution of
B2O3 on the surface. Interestingly, only 5CuB/SiO2 catalyst pre-
sented weak characteristic diffraction peak at 27.8° corresponding
to B2O3 [30], reflecting that surplus B2O3 formed crystalline parti-
cle on the surface and covered partial cupreous species, as deduced
by N2O chemisorption. In addition, a broad and diffuse diffraction
peak centered at about 21.7° was observed for all the catalysts,
which was assigned to amorphous SiO2 [21].
moles of glycerol ðinÞ ꢁ moles of glycerol ðoutÞ
Conversion ð%Þ ¼
moles of glycerol ðinÞ
ꢂ 100
moles of one product
moles of all products
Selectivity ð%Þ ¼
ꢂ 100
3. Results
3.1. Characterization of the catalysts
The XRD patterns of Cu/SiO2 with various B2O3 contents after ex
situ reduction at 250 °C were also checked. To prevent and dimin-
ish phase transformation that likely proceeded during exposure to
air, the reduced samples were carefully collected and preserved in
liquid paraffin at room temperature and sealed in glass bottles un-
til XRD analysis. As compiled in Fig. 1B, for all the catalysts the
broad and diffuse overlapping peak at around 19.3° was attributed
to the collective contribution of amorphous silica and paraffin [21].
After reduction, the XRD diffraction peaks of CuO disappeared,
while the peaks at around 43.2°, 50.4°, and 74.1° attributed to Cu
emerged in these catalysts [32]. The Cu characteristic peaks en-
larged and their intensities diminished apparently as a result of
B2O3 doping. The average Cu particle size calculated by Scherrer
equation declined gradually from 17.6 nm for Cu/SiO2 to 4.2 nm
for 5CuB/SiO2. Due to the weak interaction between copper species
and silica support, the copper particles tended to mobilize and
aggregate to larger particles during the calcination and reduction
processes. In contrast, the incorporation of B2O3 was capable of
controlling the growth of Cu particle size upon heat treatment,
which led to the improved dispersion of Cu species. On the other
hand, generally the average Cu particle size based on XRD was
higher than that of N2O chemisorption. The nanoparticles below
4 nm cannot be detected by XRD, which can affect the accuracy
3.1.1. Physicochemical properties of the catalysts
The textural properties derived from N2 adsorption–desorption
isotherms of the Cu/SiO2 and B2O3-doped catalysts are listed in Ta-
ble 1. It can be observed that the BET surface area and average pore
volume decreased obviously with increasing B2O3 loading, while
the average pore size enhanced. This may be attributed to pore
blockage caused by B2O3 molecules which are small enough to en-
ter micropores of Cu/SiO2. As shown in Table 1, the bulk weight
loadings of Cu and B elements determined by ICP test were close
to the nominal values, despite small discrepancy owing to the loss
in washing procedure. The dissociative N2O chemisorption (Ta-
ble 1) on the reduced catalyst was to determine the Cu dispersion,
specific surface area, and particle size. Doping B2O3 to Cu/SiO2 cat-
alysts decreased Cu particle size significantly and simultaneously
increased Cu dispersion and surface area remarkably. The role of
B2O3 in enhancing the dispersion was consistent well with the case
of Ni-B/SiO2 [29]. This tendency is explained by the fact that B2O3
can be instrumental in preventing thermal transmigration and
aggregation of Cu nanoparticles during the calcination and reduc-
tion, which may relate to the strong interaction between copper
and boron species. Among them, 3CuB/SiO2 afforded the highest
Table 1
The physicochemical properties and chemical compositions of B2O3-doped Cu/SiO2 catalysts.
SBET (m2 gꢁ1
)
Dpore (nm)
Vpore (cm3 gꢁ1
)
Cu content (wt.%)a
B content (wt.%)a
dCu (nm)b
Cu dispersion (%)c
SCu
dCu (nm)c
c
Catalyst
Cu/SiO2
146.6
117.4
104.7
98.8
8.8
9.2
9.3
10.9
11.1
0.392
0.380
0.348
0.337
0.294
14.65
14.31
13.88
13.95
13.52
0
17.6
14.2
12.3
6.8
7.7
13.7
15.8
23.9
22.1
7.6
13.0
7.3
6.3
4.2
4.5
0.5CuB/SiO2
1CuB/SiO2
3CuB/SiO2
5CuB/SiO2
0.50
1.06
2.73
4.59
12.9
15.7
23.6
20.2
92.5
4.2
a
b
c
Determined by ICP.
Average Cu particle size calculated by XRD.
Determined by N2O chemisorption.

S. Zhu et al. / Journal of Catalysis 303 (2013) 70–79
73
(A)
CuO
B2 O3
Cu/SiO2
0.5CuB/SiO2
1CuB/SiO2
3CuB/SiO2
5CuB/SiO2
40
10
20
30
40
50
60
70
80
90
°
2θ/
°
2θ/
#
(B)
# Cu
#
#
Cu/SiO2
0.5CuB/SiO2
1CuB/SiO2
3CuB/SiO2
5CuB/SiO2
10
20
30
40
50
60
70
80
90
°
2θ/
Fig. 1. XRD patterns of B2O3-doped Cu/SiO2 catalysts upon (A) calcination; (B) reduction.
[33]. This may explain the discrepancy between XRD and N2O
chemisorption results.
due to B2O3 doping. To gain further insight into the structure infor-
mation, the FTIR spectra of these calcined samples were conducted.
As shown in Fig. 2, the Cu/SiO2 sample displayed three characteris-
tic bands which centered at 1110, 800, and 470 cmꢁ1 correspond-
ing to the various vibration modes of Si-O bonds in the amorphous
SiO2 [21]. Compared to bulk SiO2, these bands still preserved and
actually unchanged even addition of B2O3 to Cu/SiO2, revealing
the weak interaction between CuO and SiO2, consistent well with
To further explore the phase evolution of representative Cu/SiO2
and 3CuB/SiO2, the in situ XRD technique was exhibited in flowing
H2 under different reduction temperatures. As illustrated in Fig. S1
(in the Supplementary material), the Cu/SiO2 sample displayed two
obvious peaks at ca. 35.6° and 38.8° characteristic of CuO at 25 °C.
These two peaks became remarkably weak while new peaks at ca.
43.2°, 50.4°, and 74.1° characteristic of Cu appeared with the
increasing temperature to 150 °C. With continuous increasing
reduction temperature, these metallic Cu characteristic peaks
sharpened gradually, due to the inevitable sintering and aggrega-
tion of copper particles upon heat treatment. Based on Scherrer
equation, the mean copper particle size significantly increased
from 15.7 to 25.5 nm, when the reduction temperature enhanced
from 200 to 400 °C. Similar trend of phase evolution has been ob-
served in the case of 3CuB/SiO2 (Fig. S2). Nevertheless, the mean
copper particle size grew moderately from 4.2 to 7.8 nm with the
increasing temperature from 200 to 400 °C, indicating that the
B2O3 played a role in immobilizing copper species and suppressing
sintering of Cu nanoparticles upon heat treatment.
the Raman results (see below).
A weak band centered at
670 cmꢁ1 was ascribed to cupric phyllosilicate over Cu/SiO2 sample
[34]. The intensity ratio for IR bands at 670 and 800 cmꢁ1 enhanced
continuously with the increasing B2O3 loading, suggesting that the
relative content of cupric phyllosilicate to amorphous silica im-
proved gradually. The cupric phyllosilicate which is a kind of cop-
per silicate with lamellar structure may originate from the unique
precipitation-gel method. The structure is formed through the
reaction of Cu2+ complex with the silanol groups of silica surface
via hydrolytic adsorption to produce „SiOCuII monomer and con-
secutive polymerization [34]. Addition of B2O3 into Cu/SiO2 may
facilitate to stabilize this complex on the surface that would
decompose to form cupric phyllosilicate upon calcination at
400 °C. Therefore, it can be inferred that the black color was attrib-
uted to the CuO species while the blue color was due to the forma-
tion of partial cupric phyllosilicate on the surface. Meanwhile, the
band at approximately 1420 cmꢁ1 appeared and the intensity in-
Interestingly, it has been observed that the calcined Cu/SiO2
sample presented mainly dark color while the color of B2O3-doped
samples gradually became blue with the increasing B2O3 content
(Fig. S3), indicating the appearance of at least two copper species

74
S. Zhu et al. / Journal of Catalysis 303 (2013) 70–79
1100
a
ß
470
5CuB/SiO2
3CuB/SiO2
1CuB/SiO2
1420
f
800
670
e
d
c
0.5CuB/SiO2
Cu/SiO2
b
a
100
200
300
400
500
Temperature (oC)
400
600
800
1000
1200
1400
1600
1800
2000
Fig. 4. H2-TPR profiles of B2O3-doped Cu/SiO2 catalysts.
Wavenumber (cm-1)
Fig. 2. FTIR spectra of calcined (a) SiO2, (b) Cu/SiO2, (c) 0.5CuB/SiO2, (d) 1CuB/SiO2,
(e) 3CuB/SiO2, and (f) 5CuB/SiO2.
species on the support surface, reflecting the strong CuO and
B2O3 interaction. Moreover, it can be observed that only the
5CuB/SiO2 emerged two bands at 143 and 247 cmꢁ1 corresponding
to B2O3 [35], suggesting that the B2O3 species were highly dis-
persed on the surface in the other modified catalysts, in agreement
with the XRD results.
creased with the increasing boron content, which was assigned to
the vibration of B2O3 [35].
The Raman spectra of Cu/SiO2 catalysts with and without B2O3
dopants are illustrated in Fig. 3. For comparison, the Raman spectra
of bulk CuO were also examined, which presented three character-
istic bands at 293, 340 and 627 cmꢁ1, respectively [36]. For B2O3-
free Cu/SiO2 catalyst, these characteristic bands remained and
practically unchanged, suggesting that CuO and SiO2 did not inter-
act substantially. However, these bands underwent obvious shift
toward lower wavenumbers and widened gradually with the
increasing boron content. Specifically, the vibration band in Ag
mode of CuO [36] ranged from 293 cmꢁ1 for Cu/SiO2 to 282 cmꢁ1
for 5CuB/SiO2. These features could be related to the structure de-
fects generated by incorporation of heterovalent atoms [36]. The
incorporation of smaller ions (B3+) into CuO structure and conse-
quent lattice contraction of CuO as well as Cu–O vibration fre-
quency may induce shift in the CuO characteristic band position,
as evidenced by XRD measurements. The generation of surface de-
fects in the CuO structure due to the incorporation of B3+ was indic-
ative of the existence of close contact between copper and boron
3.1.3. Reducibility and surface acidic properties
In order to explore the reduction behavior of these catalysts, the
H2-TPR technique was measured. As displayed in Fig. 4, the Cu/SiO2
sample presented an asymmetric reduction peak, implying at least
two different Cu species, which can be deconvoluted into two
peaks centered at about 165 and 236 °C, respectively. All the TPR
patterns of these samples can be roughly divided in two reductive
peaks which were designated as low-temperature peak
high-temperature peak b, respectively. Irrespective of peak
a
and
or
a
b, these peaks shifted toward higher temperature with the increas-
ing B2O3 content, most presumably owing to the strong interaction
between CuO and B2O3. According to previous reports [26,31,36],
the low temperature peak can be assigned to the reduction of
highly dispersed CuO with small nanoparticles while the high-tem-
perature peak can be related to the reduction of bulk CuO with
large size. Thus, the copper species in Cu/SiO2 sample was mainly
present in the form of bulk CuO, indicating the weak meal-support
interaction, which was probably stemmed from the poor interac-
tion between the precipitated species copper hydroxide and the
surface of silica support [21]. With the promotion of B2O3, the peak
5CuB/SiO2
3CuB/SiO2
area of low-temperature peak
a was elevated considerately,
reflecting that addition of B2O3 was beneficial for the formation
of large amounts of highly dispersed CuO nanoparticles. This may
be associated with the close contact between CuO and B2O3 on
the SiO2 surface, which facilitates to inhibit the aggregation of cop-
per species and form small nanoparticles upon heat treatment. The
nanoparticles possessing a lower average CuO size presented
kinetically faster reduction than that of bulk CuO. Indeed, the H2-
TPR results demonstrate that there is strong interaction between
copper and boron species, consistent with the Raman spectra
results.
1CuB/SiO2
0.5CuB/SiO2
Cu/SiO2
It is well established that the acidity plays an important role in
determining the catalytic performance of glycerol hydrogenolysis
on the basis of bi-functional reaction mechanism [6,14,37]. Thus,
the accessible surface acidic sites of these catalysts were probed
by NH3-TPD. As shown in Fig. 5, the NH3-TPD profile of Cu/SiO2 cat-
alyst presented a broad peak centered at 375 °C, indicating that
1/2
CuO
100
200
300
400
500
600
700
800
Raman (cm-1)
Fig. 3. Raman spectra of bulk CuO and calcined B2O3-doped Cu/SiO2 catalysts.

S. Zhu et al. / Journal of Catalysis 303 (2013) 70–79
75
(A)
5CuB/SiO2
Cu 2p
5CuB/SiO2
3CuB/SiO2
1CuB/SiO2
3CuB/SiO2
1CuB/SiO2
0.5CuB/SiO2
Cu/SiO2
0.5CuB/SiO2
Cu/SiO2
930
935
940
945
950
955
960
200
300
400
500
600
Temperature (oC)
Binding energy (eV)
Fig. 5. NH3-TPD profiles of B2O3-doped Cu/SiO2 catalysts.
medium acid sites existed on the catalyst surface [30]. The NH3
desorption peak area improved gradually with the enhancement
of boron content, maximized on 3CuB/SiO2, and then decreased
slightly. Meanwhile, the NH3 desorption temperature shifted to-
ward higher temperature with an increase in boron loading. Thus,
it can be inferred that addition of boron species increased acidic
amount and also enhanced acidic strength, in well agreement with
the previous reports [29,38].
(B)
B 1s
5CuB/SiO2
3CuB/SiO2
1CuB/SiO2
3.1.4. Surface chemical states
XPS analysis was used to identify surface chemical states after
reduction and the compositions of these catalysts; the results are
summarized in Table 2. According to previous reports [32,39],
the Cu+ was almost not detected for all the catalysts, as indicated
by XAES spectra. Therefore, the two peaks (Fig. 6A) centered at
ca. 932.5 and 952.5 eV were mainly ascribed to Cu 2p3/2 and Cu
2p1/2 peaks of Cu0, respectively [30,39]. Meanwhile, at all cases
the absence of satellite peaks at 944–945 eV related to Cu2+ were
observed [39], suggesting that the Cu2+ species (from copper oxide
and cupric phyllosilicate) were reduced completely, consistent
with the XRD results. As shown in Fig. 6B, the B 1s peak at about
193.5 eV was indicative of B3+. The lack of a peak at 188.7 eV re-
vealed that no B0 was present, which was different from the report
by Zhao et al. [32]. The disparity may stem from the different
reduction procedure as Zhao et al. used 350 °C as reduction tem-
perature at which partial B3+ can be reduced to B0.
As listed in Table 2, there was almost no variation in the binding
energy of Si 2p. The Cu 2p3/2 peak shifted to higher binding energy
when adding B2O3 to Cu/SiO2, while the B 1s moved to lower bind-
ing energy. The diminishing electron density in the Cu0 nucleus ar-
ose from the electron scavenging property of B2O3 [30]. Thus, an
electron donor–acceptor interaction between Cu and B2O3 may oc-
cur on the silica surface. These results demonstrated the existence
of strong interaction between copper and boron species wherein
the electron likely transferred from Cu0 to B3+. Additionally, despite
0.5CuB/SiO2
189
192
195
198
Binding energy (eV)
Fig. 6. Cu 2p (A) and B 1s (B) XPS photoemission peaks of reduced B2O3-doped Cu/
SiO2 catalysts.
almost the same bulk Cu loading, the surface concentration of Cu
over Cu/SiO2 increased greatly due to B2O3 doping, indicating that
the incorporation of B2O3 apparently enhanced Cu dispersion and
thus generated more active sites, in well line with N2O chemisorp-
tion results. The Cu/B atomic molar ratio estimated from XPS anal-
ysis was much lower than the values determined by ICP test. The
low detection of Cu on the surface was possibly ascribed to the fact
that partial Cu species were covered by B2O3 as a consequence of
the unique preparation method, which was also in accordance with
the results deduced by N2O chemisorption.
3.2. Catalytic performance of the catalysts
The catalytic performance of B2O3-doped Cu/SiO2 catalysts was
tested in a continuous fixed-bed reactor at 200 °C and 5.0 MPa
Table 2
XPS results for B2O3-doped Cu/SiO2 catalysts.
Catalyst
Binding energy (eV)
Cu 2p3/2
Surface elemental concentration (at.%)
Cu/B molar ratio
B 1s
Si 2p
Cu
B
–
Cu/SiO2
932.5
932.7
932.8
932.9
933.0
–
103.4
103.4
103.4
103.4
103.4
2.9
3.5
4.4
5.1
4.5
–
0.5CuB/SiO2
1CuB/SiO2
3CuB/SiO2
5CuB/SiO2
193.5
193.4
193.2
193.1
3.0
4.2
6.4
1.2 (4.9)a
1.0 (2.2)
0.80 (0.87)
0.41 (0.50)
11.0
a
Values in parenthesis are determined by ICP test.

76
S. Zhu et al. / Journal of Catalysis 303 (2013) 70–79
Table 3
Catalytic performance of glycerol hydrogenolysis over different catalysts.a
# Cu
#
Catalyst
TOFb
Conversion
(%)
Selectivity (%)
(hꢁ1
)
#
Acetol EG Othersc
3CuB/SiO2 -2-propanol
#
1,2-
PDO
Cu/SiO2
0.5CuB/
SiO2
0.88
1.24
62.1
83.3
89.5
92.1
6.0
3.6
2.6 1.9
1.2 3.1
O
3CuB/SiO2-H2
1CuB/SiO2
3CuB/SiO2
5CuB/SiO2
1.38
1.53
1.47
91.3
100
98.5
94.7
98.0
96.7
2.1
0.5
1.3
1.0 2.2
0.3 1.2
0.6 1.4
a
Reaction conditions: 200 °C, 5.0 MPa, 10 wt.% glycerol aqueous solution, H2/
glycerol = 123:1 (molar ratio), WHSV = 0.075 hꢁ1
.
b
Cu/SiO2-H2O
60
TOF is moles of glycerol converted per mole of surface Cu sites and per hour.
Reaction conditions: 200 °C, 5.0 MPa, 10 wt.% glycerol aqueous solution, H2/glyc-
erol = 123:1 (molar ratio), WHSV = 0.3 hꢁ1
.
c
Others include ethanol, methanol, propane, etc.
30
40
50
70
80
θ
°
/
2
using water as solvent; the results are summarized in Table 3. For
B2O3-free Cu/SiO2 catalyst, the conversion of glycerol and selectiv-
ity to 1,2-PDO were 62.1% and 89.5%, respectively. As B2O3 was
introduced into Cu/SiO2 catalysts, the catalytic activity and 1,2-
PDO selectivity were improved significantly. Among the catalysts
employed, the 3CuB/SiO2 catalyst achieved the superior perfor-
mance, up to complete glycerol conversion with 98.0% 1,2-PDO
selectivity, which is among the best reported results for the previ-
ous Cu-based catalysts [21–25,28]. However, further increase in
B2O3 amount resulted in minor decline of conversion and 1,2-
PDO selectivity. Turnover frequency (TOF) is calculated to reflect
the intrinsic activity; the results are also listed in Table 3. Com-
pared to Cu/SiO2, the addition of B2O3 increased TOF dramatically.
Fig. 8. XRD patterns of spent Cu/SiO2 and 3CuB/SiO2 catalysts.
24 h, when using water as solvent. Contrarily, the 3CuB/SiO2 cata-
lyst can stabilize to 56 h under the identical reaction conditions.
However, further prolonging reaction time led to distinct decrease
in glycerol conversion. Bienholz et al. [23] asserted that water can
cause serious aggregation of Cu nanoparticles and result in a tre-
mendous loss of copper surface area. Similar results have been ob-
tained in our B2O3-modified Cu/SiO2 catalysts, as evidenced by XRD
and TEM techniques. As illustrated in Fig. 8, the Cu particle size of
spent Cu/SiO2 after 33 h in water increased remarkably from 17.6
to 34.2 nm owing to the significant agglomeration of copper spe-
cies. In contrast, the Cu particle size of spent 3CuB/SiO2 catalyst
grew moderately from 6.8 to 12.1 nm despite with longer reaction
time (72 h). The TEM results of spent catalysts are displayed in
Fig. 9, which can give direct information on the distribution of
Cu particles and morphology. Big copper particles with an average
diameter up to 30 nm were present over spent Cu/SiO2, consistent
well with the XRD results. Furthermore, the majority of them
underwent extremely serious aggregation and formed irregularly
shaped large clusters. Compared with spent Cu/SiO2, the large
amounts of copper nanoparticles were homogenously distributed
on 3CuB/SiO2, while some of them moderately aggregated. The
above results indicated that B2O3 can be instrumental in control-
ling the Cu particle size and suppressing the aggregation of copper
species during the reaction.
The 3CuB/SiO2 catalyst achieved the highest TOF, up to 1.53 hꢁ1
,
which was superior or at least compared to the previous Cu/SiO2
catalysts [21,28]. The combined results clearly confirmed the high
efficiency of B2O3-modified Cu/SiO2 catalyst. Concurrently, the
selectivity to acetol diminished with the increasing B2O3 loading,
minimized on 3CuB/SiO2, and then increased slightly. The intro-
duction of B2O3 into Cu/SiO2 impaired the selectivity of ethylene
glycol (EG), indicating that the C–C cleavage reaction was sup-
pressed to some extent, which was related to the increased acidity
[40].
As stability was critical for the practical application for glycerol
hydrogenolysis, the long-term performance of representative Cu/
SiO2 and 3CuB/SiO2 catalysts was conducted. As shown in Fig. 7,
the Cu/SiO2 catalyst presented serious deactivation after only
To address this issue, the glycerol hydrogenolysis was per-
formed over 3CuB/SiO2 using 2-propanol as solvent. As shown in
Fig. 7, no obvious decline in the activity was observed even after
150 h time-on-stream over 3CuB/SiO2 catalyst. Concurrently, the
product distribution did not show any appreciable fluctuation dur-
ing the complete test. These results demonstrated that the 3CuB/
SiO2 catalyst was rather robust under 2-propanol solvent for glyc-
erol hydrogenolysis. As shown in Fig. 8, the mean Cu particle size of
spent 3CuB/SiO2 in 2-propanol solvent only grew mildly up to
9.6 nm. In addition, the Cu nanoparticles in TEM image (Fig. 9C)
were spherical and distributed uniformly on the surface without
obvious aggregation. It can be concluded that the solvent had
important impact on the stability of Cu-based catalysts, in well line
with the report by Bienholz et al. [23].
100
80
60
Conv. over Cu/SiO2
Sel. over Cu/SiO2
40
20
Conv. over 3CuB/SiO2
Sel. over 3CuB/SiO2
Conv. over 3CuB/SiO2-2-propanol
Sel. over 3CuB/SiO2-2-propanol
4. Discussion
0
20
40
60
80
100
120
140
4.1. The effect of B2O3 doping on the structural and surface features of
Cu/SiO2 catalysts
Time on stream (h)
Fig. 7. The long-term performance of Cu/SiO2 and 3CuB/SiO2 catalysts. Reaction
conditions: 200 °C, 5.0 MPa, 10 wt.% glycerol aqueous solution (or 2-propanol), H2/
It is generally accepted that B2O3 is considered as an additive of
thermal stability and applied widely for various catalysts
glycerol = 123:1 (molar ratio), WHSV = 0.075 hꢁ1
.

S. Zhu et al. / Journal of Catalysis 303 (2013) 70–79
77
100 nm
100 nm
(B)
(A)
100 nm
(C)
Fig. 9. TEM images of spent (A) Cu/SiO2 using water as solvent, (B) 3CuB/SiO2 using water as solvent, and (C) 3CuB/SiO2 using 2-propanol as solvent.
[29,41,42]. The quantum mechanics calculations reveal that boron
atoms are thermodynamically stable at step and subsurface sites
over CoB/Al2O3 catalyst under Fischer–Tropsch synthesis condi-
tions which are even harsher than ours [41]. However, the under-
lying mechanism by which the stabilizing effect occurs is still
subject to some debate. In the presence of Ni-B/SiO2 catalyst pre-
pared by sol–gel approach, Zheng et al. [29] ascribed the enhance-
ment in the dispersion of active species to the strong interaction
between surface Ni and B2O3. On the other hand, Tupabut et al.
[42] claimed that the strong interaction between Co3O4 and sup-
port MCM-41 induced by B2O3 doping can suppress sintering of
Co3O4 particles and improve the dispersion of active species. In a
series of research on CuB/HMS catalysts by Yin et al. [38], boron
loading, boron source, and the preparation method were found to
affect Cu particle size and catalytic performance of dimethyl oxa-
late hydrogenation. They argued that B2O3 can make Cu-based cat-
alysts expose more active sites or a synergetic effect may lie in
between the copper and boron species. It seems that all of them
provide plausible explanations which are highly dependent on
the catalytic systems and preparation methods.
In our case, the stabilizing effect of B2O3 on copper species can
be explained as follows. Firstly, the strong interaction between
copper and boron species is capable of retarding the aggregation
of copper species as well as the growth of crystalline particles dur-
ing the processes of calcination, reduction and reaction, which re-
sults in remaining high dispersion of copper species. Based on FTIR,
Raman and TPR results, the interaction between CuO and support
SiO2 is rather weak, which readily leads to the transmigration of
copper species and further coagulation during thermal treatment.
Contrarily, as revealed by Raman, XPS and TPR results, the strong
interaction between copper and boron species can effectively con-
fine the lateral growth of Cu nanoparticles and improve thermal
stability, even in hydrothermal reaction. He et al. [30] have pro-
posed that the acidity and electron affinity of B2O3 are higher than
that of SiO2, which is favorable to the formation of strong interac-
tion between copper and boron species. Secondly, addition of B2O3
to Cu/SiO2 catalysts facilitates to induce the generation of cupric
phyllosilicate which can provide more stable copper species during
the course of reduction [32,34]. Note that all the spent catalysts
presented red and the blue color disappeared, confirming that
the blue cupric phyllosilicate can be reduced to red copper. This
was also verified by undetectable Cu2+ from XPS. The cupric phyl-
losilicate species cannot only enhance the dispersion of copper
species, but also stabilize the active species during heat treatment.
Finally, the doping of B2O3 positively generates new surface defects
which are typically more reactive than fully coordinated species,
consistent well with the report by Zhao et al. [32]. The XRD and Ra-
man characterization results implied that partial Cu2+ was likely
substituted by B3+, which can affect planes, corner, and edge atoms,
correspondingly change both surface structure and electronic
property and eventually form new surface defects. Natesakhawat
et al. [39] suggested that surface defects contained coordinately
unsaturated sites and were more reactive in CO2 hydrogenation
to methanol. Guo et al. [31] indicated that the new formation of
surface defects can hinder the crystallization of catalyst compo-
nents and growth of crystal grain over La-doped Cu/ZrO2 catalyst.
On the other hand, the partial surface Cu might be occupied by
boron species on the surface, as evidenced by XPS results. Addi-
tionally, B2O3 can also be utilized to adjust acidity. Likewise, the
improvement of boron oxide in acidic property has been demon-
strated by others [29,38]. Moreover, the introduction of B2O3 also
affected BET surface area and pore size distribution to some extent,
although the nature of mesoporous material was not changed.
4.2. The effect of B2O3 doping on the catalytic behavior of Cu/SiO2
catalysts
Bienholz et al. [28] proposed that there was a linear relationship
between the copper specific surface area and catalytic performance
of glycerol hydrogenolysis, irrespective of the preparation proce-

78
S. Zhu et al. / Journal of Catalysis 303 (2013) 70–79
1.6
100
80
3CuB/SiO2
35
28
21
14
7
5CuB/SiO2
1.4
1.2
1.0
0.8
1CuB/SiO2
0.5CuB/SiO2
60
40
0
Cu/SiO2
6
9
12
15
18
21
24
Cu surface area (m2 g-1)
4
6
8
10
12
14
Fig. 10. 1,2-PDO yield and copper dispersion as a function of copper surface area for
the B2O3-doped Cu/SiO2 catalysts. Reaction conditions: 200 °C, 5.0 MPa, 10 wt.%
Particle size (nm)
glycerol aqueous solution, H2/glycerol = 123:1 (molar ratio), WHSV = 0.075 hꢁ1
.
Fig. 11. TOF value as a function of copper particle size for the B2O3-doped Cu/SiO2
catalysts. Reaction conditions: 200 °C, 5.0 MPa, 10 wt.% glycerol aqueous solution,
H2/glycerol = 123:1 (molar ratio), WHSV = 0.3 hꢁ1
.
dure, additive, and support. For Cu-based catalysts, the similar lin-
ear correlation between activity and copper surface area has also
been widely reported, including dimethyl oxalate hydrogenation
[43], 5-methylfurfuryl alcohol hydrogenolysis [44], and carbon
dioxide hydrogenation to methanol [39]. Thereby, the metallic cop-
per surface area was supposed to be the predominant factor in the
structure–activity relationship for Cu-based catalysts during glyc-
erol hydrogenolysis. Accordingly, the best catalytic performance
of optimal 3CuB/SiO2 (Table 3) was tentatively ascribed to the most
Cu surface area and highest Cu dispersion which could be origi-
nated from the formation of abundant highly dispersed Cu nano-
particles due to B2O3 modification.
To gain further insight into the nature of active species in glyc-
erol hydrogenolysis over B2O3-doped Cu/SiO2 catalysts, we at-
tempted to correlate 1,2-PDO yield at 200 °C with the Cu surface
area determined by N2O chemisorption. As compiled in Fig. 10, a
closely linear relationship between 1,2-PDO yield and Cu surface
area was obtained, giving clear evidence that metallic Cu involved
as the active site in glycerol hydrogenolysis. The surface metallic
Cu site density was mainly responsible for the catalytic activity,
which can offer a rational explanation for the boron content effect
on the catalytic performance. Moreover, it is expected that the Cu
dispersion and specific surface area are strongly correlated because
the dispersion is inversely proportional to the radius of spherical
particle [39], as confirmed by TEM.
Based on the above discussion, it is reasonably to infer that the
TOF value may have obvious correlation with Cu particle size. As
illustrated in Fig. 11, the TOF value linearly declined with the
increasing Cu particle size over these catalysts, providing direct
evidence that glycerol hydrogenolysis was a structure-sensitive
reaction [27]. However, these results cannot rule out the role of
other factors, such as the textural properties, chemical environ-
mental of copper species and surface acid sites, which may also af-
fect the catalytic behavior of glycerol hydrogenolysis.
It is well-known that glycerol hydrogenolysis to 1,2-PDO pro-
ceeds dehydration of glycerol primary hydroxyl group to acetol
and subsequent hydrogenation [6]. Our previous report [45] re-
vealed that an approximately stoichiometric relationship was ex-
isted between acetol and 1,2-PDO. Similar tendency was also
obtained over B2O3-modified Cu/SiO2 catalysts. As shown in Ta-
ble 3, addition of B2O3 to Cu/SiO2 catalyst improved 1,2-PDO selec-
tivity moderately at the expense of acetol, implying that the
surplus acetol over Cu/SiO2 can be converted into 1,2-PDO. This
can be explained by the enhanced surface Cu species due to B2O3
doping. Thereby, it can be inferred that the B2O3-free Cu/SiO2 cat-
alyst could provide enough acidic sites to proceed dehydration of
glycerol, while lacking adequately available surface copper sites
to undergo hydrogenation, which led to the formation of much
acetol. Furthermore, the linear correlation between 1,2-PDO yield
with Cu surface area also indicates that acetol hydrogenation over
active copper sites may be the rate-determining step [46]. How-
ever, the surface acidic sites play indispensable role in glycerol
hydrogenolysis as they can protonate the hydroxyl group of glyc-
erol and promote glycerol dehydration. To sum up, the optimal cat-
alytic activity of 3CuB/SiO2 is supposed to be the combined
contribution of the most surface Cu sites and acidic sites.
4.3. The effect of B2O3 doping on the stability of Cu/SiO2 catalysts
It is well established that Cu-based catalysts are susceptible to
agglomeration of active metallic phase during glycerol hydrogenol-
ysis in aqueous solution, which is the primary factor for strong
deactivation [23,27,47,48]. Accordingly, a number of additives
including ZnO [47], Ru [27], Ga2O3 [48], and Cr2O3 [13] have been
employed to immobilize copper particle size. The stability of cop-
per catalysts can be distinctly improved by B2O3 doping, such as
in the hydrogenation of dimethyl oxalate [30]. As mentioned
above, the strong interaction between copper and boron species
was supposed to stabilize active copper species and suppress sur-
face transmigration of small nanoparticles during glycerol hydrog-
enolysis, resulting in good catalytic stability (Fig. 7). Moreover, the
formation of cupric phyllosilicate also favored to provide stable
copper nanoparticles upon reduction, which may play a role in
improving the stability.
However, the frequent transition of copper species induced by
water presumably impaired the strong interaction between copper
and boron species and led to the accelerated sintering of copper
nanoparticles [30], which finally caused the deactivation of 3CuB/
SiO2. Contrarily, the 3CuB/SiO2 catalyst presented extremely excel-
lent stability in 2-propanol solvent. The combined results indicated
that the reaction solvent played a key role in preventing the aggre-
gation of copper species. Recently, Zhou et al. [49] concluded that
sintering due to the combined contribution of water, hydrogen,
glycerol, temperature, and duration was responsible for the deacti-
vation of Ag/Al2O3 catalyst in glycerol hydrogenolysis. Based on
our results, it can be speculated that the deactivation was mainly
attributed to water and reaction duration for Cu-based catalysts.

S. Zhu et al. / Journal of Catalysis 303 (2013) 70–79
[7] S. Zhu, Y. Zhu, S. Hao, L. Chen, B. Zhang, Y. Li, Catal. Lett. 142 (2012) 267.
79
5. Conclusions
[8] Y. Amada, Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige, Appl. Catal.
B: Environ. 105 (2011) 117.
It has been demonstrated that the B2O3 dopant can substan-
tially interact with copper species, induce the formation of cupric
phyllosilicate, and generate surface defects, which can be instru-
mental in preventing the aggregation of Cu nanoparticles and pro-
moting the dispersion of copper species upon heat treatment and
reaction.
Addition of suitable B2O3 to Cu/SiO2 catalyst pronouncedly en-
hanced the activity, 1,2-PDO selectivity, and stability for glycerol
hydrogenolysis. Among the catalysts tested, 3CuB/SiO2 afforded
the best catalytic performance, up to complete conversion with
98.0% 1,2-PDO selectivity. The close linear relationship between
1,2-PDO yield and copper surface area revealed that the active Cu
surface area was predominantly responsible for the catalytic
behavior of glycerol hydrogenolysis. Moreover, glycerol hydrogen-
olysis was structure-sensitive, as corroborated by the strong corre-
lation between TOF value with Cu particle size.
[9] F. Wang, J.L. Dubois, W. Ueda, J. Catal. 268 (2009) 260.
[10] L. Zhang, A.M. Karim, M.H. Engelhard, Z. Wei, D.L. King, Y. Wang, J. Catal. 287
(2012) 37.
[11] J. Deutsch, A. Martin, H. Lieske, J. Catal. 245 (2007) 428.
[12] S. Zhu, Y. Zhu, X. Gao, T. Mo, Y. Zhu, Y. Li, Bioresour. Technol. 130 (2013) 45.
[13] M.A. Dasari, P.P. Kiatsimkul, W.R. Sutterlin, G.J. Suppes, Appl. Catal. A: Gen. 281
(2005) 225.
[14] E.S. Vasiliadou, E. Heracleous, I.A. Vasalos, A.A. Lemonidou, Appl. Catal. B:
Environ. 92 (2009) 90.
[15] Y. Nakagawa, Y. Shinmi, S. Koso, K. Tomishige, J. Catal. 272 (2010) 191.
[16] I. Gandarias, P.L. Arias, J. Requies, M.E. Doukkali, M.B. Güemez, J. Catal. 282
(2011) 237.
[17] M.G. Musolino, L.A. Scarpino, F. Mauriello, R. Pietropaolo, ChemSusChem 4
(2011) 1143.
[18] N.D. Kim, J.R. Park, D.S. Park, B.K. Kwak, J. Yi, Green Chem. 14 (2012) 2638.
[19] S. Xia, R. Nie, X. Lu, L. Wang, P. Chen, Z. Hou, J. Catal. 296 (2012) 1.
[20] T. Miyazawa, Y. Kusunoki, K. Kunimori, K. Tomishige, J. Catal. 240 (2006) 213.
[21] Z.W. Huang, F. Cui, H.X. Kang, J. Chen, X.Z. Zhang, C.G. Xia, Chem. Mater. 20
(2008) 5090.
[22] Z.L. Yuan, J.H. Wang, L.N. Wang, W.H. Xie, P. Chen, Z.Y. Hou, X.M. Zheng,
Bioresour. Technol. 101 (2010) 7088.
Compared to Cu/SiO2, the 3CuB/SiO2 catalyst exhibited superior
long-term performance when using water as solvent, which was
mainly related to the stabilizing effect of B2O3 on Cu nanoparticles
and strong interaction between copper and boron species. Never-
theless, the 3CuB/SiO2 catalyst still suffered from deactivation after
56 h time-on-stream, presumably due to the sintering and aggre-
gation of active metallic species. The lifespan of 3CuB/SiO2 can be
greatly extended to 150 h, when 2-propanol was employed as
the reaction solvent, instead of water. The important role of solvent
on the stability for Cu-based catalysts will be further investigated
in our future work.
[23] A. Bienholz, F. Schwab, P. Claus, Green Chem. 12 (2010) 290.
[24] I. Gandarias, J. Requies, P.L. Arias, U. Armbruster, A. Martin, J. Catal. 290 (2012)
79.
[25] L.Y. Guo, J.X. Zhou, J.B. Mao, X.W. Guo, S.G. Zhang, Appl. Catal. A: Gen. 367
(2009) 93.
[26] Z.W. Huang, F. Cui, H.X. Kang, J. Chen, C.G. Xia, Appl. Catal. A: Gen. 366 (2009)
288.
[27] E.S. Vasiliadou, A.A. Lemonidou, Appl. Catal. A: Gen. 396 (2011) 177.
[28] A. Bienholz, H. Hofmann, P. Claus, Appl. Catal. A: Gen. 391 (2010) 153.
[29] J. Zheng, Z. Xia, J. Li, W. Lai, X. Yi, B. Chen, W. Fang, H. Wan, Catal. Commun. 21
(2012) 18.
[30] Z. He, H. Lin, P. He, Y. Yuan, J. Catal. 277 (2011) 54.
[31] X. Guo, D. Mao, G. Lu, S. Wang, G. Wu, J. Mol. Catal. A: Chem. 345 (2011) 60.
[32] S. Zhao, H. Yue, Y. Zhao, B. Wang, Y. Geng, J. Lv, S. Wang, J. Gong, X. Ma, J. Catal.
297 (2013) 142.
[33] L. Ma, D.H. He, Z.P. Li, Catal. Commun. 9 (2008) 2489.
[34] J. Gong, H. Yue, Y. Zhao, S. Zhao, L. Zhao, J. Lv, S. Wang, X. Ma, J. Am. Chem. Soc.
134 (2012) 13922.
Acknowledgment
[35] S.C. Neumair, R. Kaindl, R.D. Hoffmann, H. Huppertz, Solid State Sci. 14 (2012)
229.
[36] G. Águila, F. Gracia, P. Araya, Appl. Catal. A: Gen. 343 (2008) 16.
[37] A. Wawrzetz, B. Peng, A. Hrabar, A. Jentys, A.A. Lemonidou, J.A. Lercher, J. Catal.
269 (2010) 411.
This work was financially supported by the Major State Basic
Research Development Program of China (973 Program) (No.
2012CB215305).
[38] A. Yin, J. Qu, X. Guo, W.L. Dai, K. Fan, Appl. Catal. A: Gen. 400 (2011) 39.
[39] S. Natesakhawat, J.W. Lekse, J.P. Baltrus, P.R. Ohodnicki, B.H. Howard, X. Deng,
C. Matranga, ACS Catal. 2 (2012) 1667.
Appendix A. Supplementary material
[40] I. Jiménez-Morales, F. Vila, R. Mariscal, A. Jiménez-López, Appl. Catal. B:
Environ. 117–118 (2012) 253.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2013.03.018.
[41] K.F. Tan, J. Chang, A. Borgna, M. Saeys, J. Catal. 280 (2011) 50.
[42] P. Tupabut, B. Jongsomjit, P. Praserthdam, Catal. Lett. 118 (2007) 195.
[43] A. Yin, X. Guo, W.L. Dai, K. Fan, J. Phy. Chem. C 113 (2009) 11003.
[44] K.L. Deutsch, B.H. Shanks, J. Catal. 285 (2012) 235.
[45] L. Huang, Y.L. Zhu, H.Y. Zheng, Y.W. Li, Z.Y. Zeng, J. Chem. Technol. Biotechnol.
83 (2008) 1670.
[46] Y. Nakagawa, X. Ning, Y. Amada, K. Tomishige, Appl. Catal. A: Gen. 433–434
(2012) 128.
[47] S. Panyad, S. Jongpatiwut, T. Sreethawong, T. Rirksomboon, S. Osuwan, Catal.
Today 174 (2011) 59.
Reference
[1] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411.
[2] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Green Chem. 10 (2008)
13.
[3] C.H.C. Zhou, J.N. Beltramini, Y.X. Fan, G.Q.M. Lu, Chem. Soc. Rev. 37 (2008) 527.
[4] D. Liang, J. Gao, H. Sun, P. Chen, Z. Hou, X. Zheng, Appl. Catal. B: Environ. 106
(2011) 423.
[5] S. Zhu, Y. Zhu, S. Hao, H. Zheng, T. Mo, Y. Li, Green Chem. 14 (2012) 2607.
[6] S. Zhu, Y. Qiu, Y. Zhu, S. Hao, H. Zheng, Y. Li, Catal. Today (2012), http://
dx.doi.org/10.1016/j.cattod.2012.09.011.
[48] A. Bienholz, R. Blume, A. Knop-Gericke, F. Girgsdies, M. Behrens, P. Claus, J. Phy.
Chem. C 115 (2010) 999.
[49] J. Zhou, J. Zhang, X. Guo, J. Mao, S. Zhang, Green Chem. 14 (2012) 156.