Effect of nickel on catalytic behaviour of bimetallic Cu-Ni catalyst supported on mesoporous alumina for the hydrogenolysis of glycerol to 1,2-propanediol

Article abstract of DOI:10.1039/c4cy00320a

The catalytic conversion of glycerol to 1,2-propanediol by hydrogenolysis has potential use in the commercial biomass industry. However, the high hydrogen pressure required for the reaction is a major drawback. To overcome this limitation, in this study, we added nickel metal to a copper-based catalyst for both supplying hydrogen via aqueous-phase reforming (APR) of glycerol and improving selectivity for 1,2-propanediol in hydrogenolysis. The bimetallic Cu-Ni catalyst supported on mesoporous alumina (MA) was prepared by a sol-gel method. The prepared Cu-Ni catalyst contains ordered mesopores with high surface area and well-dispersed active sites, as confirmed by BET, TEM, XRD, and TPR. The 9Cu-1Ni/MA (molar ratio of copper to nickel: 9:1) catalyst showed the highest catalytic performance among the various xCu-yNi/MA catalysts in a low pressure of H₂. The XPS results showed that the surface ratio of Ni to (Cu + Ni) and Cu⁰/(Cu⁰ + Cu²⁺) is closely related to catalytic performance, selectivity and yield. The effect of nickel on the hydrogen production was experimentally proven by the time-on-stream tests over monometallic (Cu) and bimetallic (Cu-Ni) catalysts in the absence of hydrogen. The optimum value of the ratio of Ni to Cu is varied with the conditions in the presence of H₂. The reaction mechanism was proposed for the Cu-Ni bimetallic catalysts for hydrogenolysis with APR of glycerol. the Partner Organisations 2014.

Full text of DOI:10.1039/c4cy00320a

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Effect of nickel on catalytic behaviour of
bimetallic Cu–Ni catalyst supported on
mesoporous alumina for the hydrogenolysis
of glycerol to 1,2-propanediol†
Cite this: Catal. Sci. Technol., 2014,
4, 3191
*
Yang Sik Yun,‡ Dae Sung Park‡ and Jongheop Yi
The catalytic conversion of glycerol to 1,2-propanediol by hydrogenolysis has potential use in the
commercial biomass industry. However, the high hydrogen pressure required for the reaction is a major
drawback. To overcome this limitation, in this study, we added nickel metal to a copper-based catalyst
for both supplying hydrogen via aqueous-phase reforming (APR) of glycerol and improving selectivity for
1,2-propanediol in hydrogenolysis. The bimetallic Cu–Ni catalyst supported on mesoporous alumina (MA)
was prepared by a sol–gel method. The prepared Cu–Ni catalyst contains ordered mesopores with high
surface area and well-dispersed active sites, as confirmed by BET, TEM, XRD, and TPR. The 9Cu–1Ni/MA
(molar ratio of copper to nickel: 9 : 1) catalyst showed the highest catalytic performance among the
various xCu–yNi/MA catalysts in a low pressure of H2. The XPS results showed that the surface ratio of Ni
to (Cu + Ni) and Cu0/(Cu0 + Cu2+) is closely related to catalytic performance, selectivity and yield. The
effect of nickel on the hydrogen production was experimentally proven by the time-on-stream tests over
monometallic (Cu) and bimetallic (Cu–Ni) catalysts in the absence of hydrogen. The optimum value of
the ratio of Ni to Cu is varied with the conditions in the presence of H2. The reaction mechanism was
proposed for the Cu–Ni bimetallic catalysts for hydrogenolysis with APR of glycerol.
Received 13th March 2014,
Accepted 13th May 2014
DOI: 10.1039/c4cy00320a
www.rsc.org/catalysis
process of high O/C content. 1,2-PDO can be used as a
component of biodegradable functional fluids such as
1 Introduction
As a renewable and sustainable energy source, biodiesel is
an attractive biomass material replacing petroleum-derived
diesel fuel. Biodiesel is produced by the transesterification of
vegetable oil-based fatty acids, with a major byproduct, glyc-
erol (10 wt.% of production). In recent years, large amounts
of low-cost glycerol have been produced as a byproduct with
the expanding demand for biodiesel.1 Thus, the utilization of
glycerol for the production of value-added chemicals has
become a significant issue. A number of catalytic processes
for conversion of glycerol to value-added chemicals have been
reported in recently published papers.2–4 Among the several
methods for transformations of glycerol, the hydrogenolysis
to 1,2-propanediol (1,2-PDO) has been of interest as a route
to the chemical conversion of glycerol due to its reduction
de-icing reagents, antifreeze and coolants, and as precursors
in the synthesis of unsaturated polyester molecules and
pharmaceuticals.5 Therefore, the catalytic hydrogenolysis of
glycerol to 1,2-PDO has great potential for cost-effective processes.
Cu metal catalysts have shown superior performance in
terms of the selectivity of the hydrogenolysis reaction. They
have the intrinsic ability to catalyze the cleavage of C–O
bonds in the presence of H2 and have relatively poor activity
in catalyzing the cleavage of C–C bonds.6–8 In attempts to
enhance the activity, selectivity, and hydrogen utilization of
metallic Cu, various Cu-based bimetallic catalysts, such as
Cu–Cr, Cu–Zn, Cu–Al, Cu–Mg, and Cu–Ag, have been
studied.7,9–12 In particular, copper chromite shows a very
good performance in the hydrogenolysis of glycerol to
1,2-PDO and has a potential for a use in a commercial pro-
cess in industry. However, the major drawbacks including
the need for a high hydrogen pressure to obtain high yields
of 1,2-PDO and the low selectivity of Cu–X bimetallic catalysts
must be overcome for commercial use. Furthermore, a high
hydrogen pressure inevitably leads to an increase in operating
costs. Recently, in attempts to decrease hydrogen pressure,
several researchers have examined the possible use of
World Class University (WCU) Program of Chemical Convergence for Energy &
Environment (C2E2), School of Chemical and Biological Engineering, Institute of
Chemical Process, College of Engineering, Seoul National University (SNU),
Daehak-dong, Gwanak-gu, Seoul 151-741, Republic of Korea.
E-mail: jyi@snu.ac.kr; Tel: +82 2 880 7438
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4cy00320a
‡ These authors contributed equally.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 3191–3202 | 3191

View Article Online
Paper
Catalysis Science & Technology
the aqueous phase reforming of glycerol which can serve as
an in situ source of hydrogen in the hydrogenolysis of
glycerol to 1,2-PDO. Noble metal catalysts supported on
acidic supports, such as Pt/NaY, Pt/Al2O3, and Ru/Al2O3, were
examined in the past and showed a relatively good perfor-
mance for conversion and selectivity.13,14 As an alternate
strategy to supply in situ generated hydrogen, hydrogen
donor molecules were added to the reactant with a hydroge-
nation catalyst in a batch or in a continuous reactor.15 Using
these approaches, the developed catalyst and the system
showed good catalytic activity. However, other problems were
introduced, namely that when a noble metal is used in the
catalytic system, an additional separation process is neces-
sary. Therefore, the hydrogenolysis of glycerol at a lower
hydrogen pressure without a noble metal or additive hydro-
gen donor would be highly desirable in terms of developing a
commercially acceptable catalytic system.
Herein, to systematically explore the concept of the in situ
production of hydrogen and the hydrogenolysis reaction, we
developed a Cu–Ni bimetallic catalyst supported on meso-
porous alumina (Cu–Ni/MA). By a one-step sol–gel method
with a controlled hydrolysis rate, active metals were highly
dispersed on mesoporous alumina. Nano-sized metallic Cu
particles selectively cleaved C–O bonds, facilitating the
hydrogenolysis of glycerol. Ni metal supported on the alumina
was chosen to promote the aqueous phase reforming of
glycerol to generate hydrogen and enhance the utilization of
hydrogen. Moreover, we investigated the reducibility of
metallic Cu and the surface composition of the catalyst by
varying the Cu : Ni ratios. The findings indicated that the sur-
face ratios of Cu0/(Cu0 + Cu2+), which represents reducibility,
and Ni/(Ni + Cu) were the crucial factors in determining
the catalytic performance. Furthermore, a mechanistic study
revealed that the aqueous phase reforming of glycerol occurred
on the Ni sites in the bimetallic catalysts, leading to an enhanced
catalytic performance by supplying additional hydrogen.
the interaction between the surfactant and the metal precur-
sors. A small amount of water was slowly added dropwise to
the solutions at a rate of 0.2 ml min−1 for post-hydrolysis
until a blue or green precipitate was formed. The samples
were stirred for a further 12 h, and then dried at 353 K for
24 h. The solid product was finally calcined for 4 h at 823 K
in air. Various ratio-controlled xCu–yNi/MA catalysts were
denoted as 10Cu/MA, 9Cu–1Ni/MA, 7Cu–3Ni/MA, 5Cu–5–Ni/MA,
and 10Ni/MA.
2.2 Catalyst characterization
The catalysts were characterized by various methods, as
described below. Adsorption isotherms of nitrogen were
obtained with an ASAP2010 (Micromeritics) instrument. Pore
size distribution was determined by the BJH method applied
to the desorption branch of the nitrogen isotherm. The
surface morphology and particle size of the catalyst were
examined by high-resolution transmission electron micros-
copy (HR-TEM: 300 kV, JEOL, JEM-3010). X-ray diffraction
(XRD) data for the catalysts were collected by means of an
X-ray diffractometer (XRD, D-Max2500-PC, Rigaku Corp.)
using Cu Kα (λ = 1.5405 Å) radiation at 50 kV and 100 mA. To
examine the reducibility of the catalysts, TPR (temperature-
programmed reduction, AutoChem II, Micromeritics) was
performed using 10% hydrogen in argon with 1 gram of cata-
lyst in a conventional TPR set-up. The temperature was line-
arly ramped up to 1073 K at a rate of 10 K min−1. The
utilization of hydrogen by the catalysts was tested by H2-TPD
(temperature-programmed desorption, AutoChem II, Micro-
meritics) with a TCD (thermal conductivity detector). The
catalysts were reduced at 773 K for 3 h with 10% H2–N2 gas.
Under these conditions, the catalysts were partially saturated
with hydrogen. The temperature of the reduced catalyst was
decreased to ambient temperature and a stream of N2
(30 ml min−1) was then used to physically remove the adsorbed
H2. The measurements were conducted in flowing N2 from
ambient temperature to 1073 K at a ramping rate of
10 K min−1. The acid properties of the catalysts were measured
by NH3-TPD (temperature-programmed desorption, AutoChem
II, Micromeritics) with a thermal conductivity detector (TCD).
An NH3 adsorption step was carried out at ambient tempera-
ture. The NH3 that was physically adsorbed on the catalyst was
removed by heating at 373 K for 1 h in the presence of an inert
gas (He). The signals were measured in He flow from room
temperature to 1073 K at a ramping rate of 10 K min−1. The
copper metal areas were determined by the dissociative N2O
decomposition method using an AutoChem 2920 (Micro-
meritics, USA). In an each experiment, the catalyst was placed
in a U-shaped quartz reactor and in situ reduced in a flow of
H2–Ar at 773 K for 3h. Then the catalyst was exposed to N2O
gas at 333 K for 1 h, resulting in oxidation from Cu to Cu2O.
Finally, s-TPR was performed on the freshly oxidized Cu2O sur-
face to reduce the Cu2O to Cu, increasing the temperature up
to 723 K with a rate of 10 K min−1. The copper surface area was
calculated referring to the previously reported paper.18 After
the hydrogenolysis of glycerol, the amount of copper or nickel
2 Experimental
2.1 Preparation of Cu–Ni catalysts supported on mesoporous
alumina (Cu–Ni/MA)
The Cu–Ni/MA catalysts were synthesized by a one-step
sol–gel method16,17 using aluminum sec-butoxide (Al(sec-BuO)3,
Sigma Aldrich) as an aluminum source with control of the
hydrolysis rate via a post-hydrolysis method. Lauric acid was
used as a surfactant, and copper nitrate (Cu(NO3)2·2.5H2O,
Sigma Aldrich) and nickel nitrate (Ni(NO3)2·6H2O, Sigma
Aldrich) were mixed with the surfactant in an organic solu-
tion. Firstly, two solutions, the aluminum source and metal
precursors with the surfactant dissolved in sec-butyl alcohol
(Fluka) respectively, were prepared. The two solutions were
then mixed with vigorous stirring. A series of Cu–Ni/MA
bimetallic catalysts were prepared by controlling the molar
ratio of copper and nickel at 10 : 0, 9 : 1, 7 : 3, 5 : 5, and 0 : 10,
and the total loading amount of xCu–yNi was 33 wt.% in the
xCu–yNi/MA catalysts. After 30 min, micelles were formed by
3192 | Catal. Sci. Technol., 2014, 4, 3191–3202
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
that was leached from the system was analyzed by inductively
coupled plasma mass spectrometry (ICP-MS, PERKIN-ELMER
SCIEX, ELAN 6100).
shown in Fig. 1 and the corresponding textural properties are
summarized in Table 1. All of the tested Cu–Ni samples
exhibited a type IV isotherm with a hysteresis loop in the rel-
ative pressure range of 0.4–0.8, implying the presence of a
mesostructure. The samples had a uniform pore size within a
range of 3–6 nm, as confirmed by pore size distribution
curves in Fig. 1B. The prepared catalysts showed high BET
surface areas of over ca. 160 m2 g−1 due to their unique struc-
ture. However, as the ratio of Cu to Ni increases, the iso-
therms of the catalysts exhibit a slightly different hysteresis
loop with a decrease in both surface area and pore volume.
This result demonstrates that metallic Cu area is varied in
xCu–yNi/MA catalysts. As seen in Table 2, the Cu metal area
also increased upon increasing the amount of Ni. Concerning
pore size distribution, monometallic (10Cu/MA and 10Ni/MA)
catalysts showed symmetric pore size distribution curves as
shown in Fig. 1B. However, relatively broad and asymmetric
curves were observed after the addition of Ni. This is due to
the different interaction between metal ions (Cu and Ni) with
micelle complexes in the sol–gel process.19 The pore sizes of
the prepared catalysts decreased with the increase in Cu con-
tents because of the poor interaction of surfactant molecules
with Cu ions and partial pore blocking by metallic Cu, as
evidenced by the TGA and DTA curves.
2.3 Catalytic activity for the hydrogenolysis of glycerol
The reactions of the catalysts with glycerol were carried out
in a Teflon-lined stainless steel autoclave reactor (100 mL). In
each reaction, 50 g of an aqueous glycerol solution (80 wt.%
glycerol solution) and 0.02 mol of active metal (Cu, Ni) in the
catalysts were loaded in the reactor. Prior to the reaction, the
catalysts were reduced in a stream of 10% H2–N2 at 773 K
for 3 h. The reactor was then flushed 3 times with N2 to
remove the ambient air. The hydrogenolysis of glycerol was
performed with the reduced catalysts at a temperature of
493 K for 10 h with vigorous stirring (500 rpm). The reactant
and product in the liquid phase were analyzed using an
off-line gas chromatograph (GC, YL 6200, Younglin Instrument)
equipped with a flame ionization detector (FID) and a capil-
lary column (HP-INNOWAX, 30 m × 320 mm × 0.5 μm). The
gaseous products from the batch reactor were passed through
a condenser at 278 K resulting in separation of gas and liquid
phases after the reaction. The gas-phase products were
analysed by a gas chromatograph (GC, YL 6100, Younglin
Instrument) equipped with a TCD detector and a packed
column (ShinCarbon ST). The acquired experimental data
were analyzed using the following equations:
Fig. 2 shows the TG and DTA curves for the as-synthesized
10Cu/MA and 10Ni/MA catalysts. These data was collected in
order to gain insights into the nature of the interactions
between the metal ions and the surfactant. The weight loss at
temperatures of up to 500 K can be attributed to the removal
of the residual organic solvent and adsorbed water mole-
cules. The point where the slope becomes steep, at around
500 K, can be attributed to the loss of surfactant molecules. A
similar trend was observed in the weight loss curves of both
Glycerol conversion:
• Glycerol conversion:
moles of glycerol reacted
conversion % 
100


initial moles of glycerol
• Glycerol conversion into gas:
C atoms in gas products
XG % 
100


C atoms in feed
Selectivity of the liquid and gas-phase products:
• Yield of liquid products:
Nout
Nin
liquid products
Y % 
i 
100

Glycerol
• Selectivity to CO2, CO, and CH4
C atoms in species i
S % 
i 
100

C atoms in gas products
• Selectivity to hydrogen:
moles of H2 produced
3
SH % 

100


2
C atoms in gas products
7
3 Results and discussion
3.1 Characterization of the prepared xCu–yNi/MA catalysts
The pore structure of the reduced xCu–yNi/MA catalysts was
characterized by N2 adsorption–desorption isotherms as
Fig. 1 (A) Isotherm curve and (B) pore size distribution of the various
xCu–yNi/MA catalysts by N2 adsorption–desorption experiments.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 3191–3202 | 3193

View Article Online
Paper
Catalysis Science & Technology
Table 1 BET surface area and pore volume of the various xCu–yNi/MA catalysts by N2 adsorption–desorption experiments
Catalyst
BET surface area (m2 g−1
)
Pore size (nm)
Pore volume (cm3 g−1
)
10Cu/MA
161.6
179.1
186
219
306.9
3.3
3.5
4.3
3.6
4.6
0.17
0.26
0.35
0.38
0.47
9Cu–1Ni/MA
7Cu–3Ni/MA
5Cu–5Ni/MA
10Ni/MA
Table 2 Composition of the liquid- and gas-phase by-products over
xCu–yNi/MA catalysts in hydrogenolysis of glycerol. (Conditions: 493 K,
40 bar of H2, 12 h, 1 g of catalyst, 80 wt.% of glycerol solution)
Products
10Cu
9Cu–1Ni
7Cu–3Ni
5Cu–5Ni
10Ni
Liquid-phase (%)
Formaldehyde
Ethanol
1PO + 2 PO
Acetol
0.1
1
3.6
3.8
2.4
52.8
0.1
2.2
16
3
2.1
55.3
0.1
2.1
9.6
4.9
2.6
0.1
2.5
6.4
1.7
2.1
0.2
4.5
2.4
2.9
6.5
EG
1,2-PDO
26.3
27.3
24.3
Gas-phase (%)
Xglycerol
3
1.6
6.6
3.2
1.2
1.5
5.2
3.8
1.2
1.1
9.2
5.9
1.6
10.3
13.8
19.5
35.1
SCO
13.7
5.1
1.3
2
SCO
SCH
Fig. 3 TEM images of the xCu–yNi/MA catalysts, and 9Cu–1Ni/Al2O3
catalyst prepared by the impregnation method. Arrows indicate metal
particles.
4
9Cu–1Ni/Al2O3 (212.6 m2 g−1, 31.6%) were lower than that of
9Cu–1Ni/MA catalyst (292.6 m2 g−1, 43.4%). This indicates
that our preparation method is effective in terms of dispers-
ing active metal particles. Meanwhile, smaller metallic parti-
cles were confirmed to be present in other catalysts that
contain a lower content of Cu than that of 9Cu–1Ni/MA due
to the high dispersity of the active metal species.
Fig. 2 (A) TGA patterns and (B) heat flow curves of the 10Cu/MA and
10Ni/MA catalysts before the calcination process.
The crystalline phases of the xCu–yNi/MA catalysts were
investigated by X-ray diffraction measurements. Fig. 4 shows
XRD patterns of the calcined (A and B) and reduced (C and
D) xCu–yNi/MA catalysts, respectively. The diffraction pat-
terns of the calcined xCu–yNi/MA catalysts exhibit no charac-
teristic peaks for metals corresponding to Cu or Ni, except
for a set of peaks at 37.2°, 45.6°, and 66.9°, which are charac-
teristic of γ-Al2O3 even for catalysts with at high content of
Cu and Ni in the catalyst.16 In the region within 2θ = 30–50°
(Fig. 4B), featureless peaks at 2θ = 35.5°, 38.73° and 48.7° cor-
responding to CuO,20 and 37.0° and 43.3° corresponding to
NiO15 were observed in the diffraction patterns of the cal-
cined catalysts. This suggests that the method used to pre-
pare these catalysts results in small-sized metal species. The
weak signal at 2θ = 38.8°, corresponds to the presence of crys-
talline CuAl2O4 (ref. 21) in the 10Cu/MA catalyst, indicating
that a portion of the CuO species was transformed into a
CuAl2O4 phase during the calcination process. Similar to the
calcined catalysts, the characteristic peaks for γ-Al2O3 were
also detected in catalysts that were reduced at 773 K in a H2
atmosphere. However, in contrast to the calcined catalysts,
the XRD patterns of the reduced catalysts exhibited
catalysts for temperatures up to 500 K. However, the percent-
age of weight loss of the as-synthesized 10Ni/MA catalyst was
higher than that for the as-synthesized 10Cu/MA catalyst.
This result can be attributed to the stronger interactions of
surfactant molecules with Ni ions, compared to Cu ions, lead-
ing to the formation of well-developed micelle complexes.17
The DTA curve for the as-synthesized 10Ni/MA catalyst, which
showed a large exothermic peak at around 580 K, caused by
the removal of surfactant, provides further evidence for this
good interaction between surfactant molecules and Ni ions.
Fig. 3 shows TEM images of the reduced xCu–yNi/MA
and 9Cu–1Ni/Al2O3 catalysts. All of the catalysts that were
prepared by the sol–gel method exhibited a wormhole-like
structure with meso-size pores, consistent with the N2
adsorption–desorption data. Highly dispersed metal particles
with sizes of 20–30 nm were observed in the case of the
9Cu–1Ni/MA catalyst, which contains a large portion of Cu.
On the other hand, aggregated metal particles were found in
the 9Cu–1Ni/Al2O3 catalyst, which was prepared by the
impregnation method. The Cu metal area and dispersion of
3194 | Catal. Sci. Technol., 2014, 4, 3191–3202
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
catalysts appear to be highly dispersed on the mesoporous
alumina, confirming that active sites can be dispersed effec-
tively by this preparation method. The findings also show
that the higher the content of Cu species in the bimetallic
catalyst, the weaker the interaction between the Cu species
and the support. Hence the CuO can be easily reduced to
metallic Cu at a high content of Cu species.
The acid properties of the prepared samples were investi-
gated by an NH3-TPD technique, and the data are shown in
Fig. 6. All desorption peaks were observed below 550 K, indi-
cating that the catalysts have low acid strengths appropriate
for the conversion of glycerol to acetol rather than to acro-
lein.26 According to our observations, the area of the curve
corresponding to the amount of acid sites appeared to
change with varying Ni content, but the peak position
remained the same. These results can be attributed to varia-
tions in the surface area of the catalysts. However, the differ-
ences were not sufficient to cause a change in performance
due to their low effectiveness. It has been reported that
the ability of hydrogen utilization of catalyst affects the
hydrogenolysis of glycerol.27 The utilization of hydrogen
on the prepared catalysts was examined by H2-TPD as shown
in Fig. 7. No hydrogen desorption peak and a small peak
were observed for the 10Ni/MA and 10Cu/MA catalysts. How-
ever, the bimetallic catalysts (9Cu–1Ni/MA, 7Cu–3Ni/MA,
and 5Cu–5Ni/MA) show a remarkably increased hydrogen
Fig.
4 (A) XRD patterns of the calcined xCu–yNi/MA catalysts,
(B) expanded view of the data in (A), (C) XRD patterns of reduced
catalysts, and (D) expanded view of the data in (C).
characteristic peaks for metallic Cu (2θ = 43.3° and 50.4°)20
and Ni (2θ = 44.5° and 51.8°),22 indicating that the reduced
catalysts have a higher crystallinity compared to those of the
calcined catalysts (Fig. 4C and D). Therefore, it was con-
firmed that through a reduction process, the catalysts were in
a bimetallic phase with highly dispersed Cu and Ni sites,
thereby acting as active sites for C–O cleavage (hydrogenolysis)
and C–C cleavage (APR of glycerol) respectively.20,23
The reduction behavior of the calcined xCu–yNi/MA cata-
lysts was investigated by H2-TPR and the results are shown in
Fig. 5. The intense peaks in the range of 450–500 K corre-
spond to reduction of CuO species (denoted as α). Broad
peaks in the 700–950 K temperature range are due to the
reduction of various NiOx species (denoted as γ). It was
reported that the TPR signal attributable to the reduction of
CuO species can be shifted, depending on the dispersion of
the species on the support.15 In general, bulk CuO species
(denoted as β) start to be reduced at a temperature of around
550 K while highly dispersed CuO species can be reduced at
lower temperature (below 500 K) compared to bulk ones.24,25
Based on this explanation, the CuO species in the prepared
Fig. 6 NH3-TPD profiles of the xCu–yNi/MA catalysts. (Mass detector,
ion current of NH3 (m/z = 16)).
Fig. 5 H2-TPR profiles of the xCu–yNi/MA catalysts. (Reduced peaks
of α: highly dispersed CuO on the surface, β: CuO on bulk, and γ:
nickel aluminate).
Fig. 7 H2-TPD profiles of the xCu–yNi /MA catalysts. (Mass detector,
ion current of H2 (m/z = 2)).
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 3191–3202 | 3195

View Article Online
Paper
Catalysis Science & Technology
desorption peak. This suggests that the hydrogen utilization
of the catalyst was enhanced significantly when Ni is added
to the 10Cu/MA catalyst. The nickel seems to play a role not
only in a higher production of hydrogen, but also in an
enhancement of the hydrogen utilization of the catalyst.
This enhancement may have a positive influence on the
hydrogenolysis of glycerol.
When Ni was added to 10Cu/MA (9Cu–1Ni/MA), the conver-
sion increased up to 76.6% but selectivity was essentially
unchanged compared to 10Cu/MA, resulting in the highest
yield (42.4%) of 1,2-PDO. It suggests that H2 would be pro-
duced from the APR of glycerol on metallic Ni and would
have a positive effect on the yield of 1,2-PDO. In addition,
metallic Ni can facilitate the utilization of active hydrogen in
hydrogenation, which would promote the second step, and
selectivities were decreased significantly over 7Cu–3Ni/MA
and 5Cu–5Ni/MA even lower than monometallic 10Cu/MA.
This may be because a decline in the amount of Cu and an
increase in that of Ni on the surface layer induced a prefer-
ential C–C cleavage rather than C–O cleavage. This result
suggests that 9Cu–1Ni/MA has the optimal Cu : Ni ratio for a
high activity and selectivity in the hydrogenolysis of glycerol
due to the ensemble effect. When Ni was added into the Cu
monometallic catalyst, the catalytic activity was significantly
decreased. This is due to the presence of H2 and the charac-
teristics of bimetallic catalysts. Gandarias and coworkers also
reported that the activity was not increased with increasing
in the nickel contents caused by the hydrogen. Under the
hydrogen environment, the glycerol was favorably adsorbed
on Cu metal, and the adsorption sites on Ni were mostly
occupied by the hydrogen.29,30 Thus, the hydrogenolysis of
glycerol is more dominant than aqueous phase reforming in
the presence of H2. Therefore, there is an optimum amount
of nickel content in xCu–yNi bimetallic catalysts for the
highest conversion of glycerol. It is well known that in bime-
tallic catalysts, the addition of a second metal induces signifi-
cant changes in both activity and selectivity in catalytic
reactions.22,31–33 For further investigation of the activity of
xCu–yNi bimetallic catalysts, the selectivity of various liquid-
and gas-phase products was tabulated in Table 2. The selec-
tivities of C < 3 products (ethylene glycol, ethanol, methane)
on the xCu–yNi bimetallic catalysts were slightly increased
with increasing Ni content. The amount of produced
1-propanol (1-PO) and 2-propanol (2-PO) rapidly increased
after the addition of Ni to the Cu monometallic catalyst. 1-PO
and 2-PO can be produced by additional dehydration and
hydrogenation of 1,2-PDO or 1,3-PDO.30 Therefore, the high
selectivity of 1-PO and 2-PO indicates that our catalytic
system follows the dehydration–hydrogenation mechanism
rather than the glyceraldehyde-based mechanism. It should
be noted that glyceraldehyde was not detected. The conver-
sion of glycerol to the gas-phase products also decreased with
increasing Ni content in xCu–yNi/MA catalysts. However, for
the 10Ni/MA catalyst, much more glycerol was converted to
the gas-phase products than for the other catalysts because
of the conversion of glycerol by aqueous phase reforming.
These results indicate that the nickel plays a role in promot-
ing the additional dehydration–hydrogenation of produced
propylene glycols on xCu–yNi/MA catalysts in the presence of
H2. The C–C cleavage of glycerol does not largely occur over
the xCu–yNi catalysts in a pressurized H2 environment.
3.2 Catalytic activity of the prepared xCu–yNi/MA catalysts in
glycerol hydrogenolysis
A catalytic activity test for the hydrgenolysis of glycerol was
carried out using reduced xCu–yNi/MA catalysts with differ-
ent Cu to Ni ratios. Glycerol conversion, yield and selectivity
to 1,2-PDO are shown in Fig. 8A. 10Cu/MA showed quite high
activity (62.4%) and selectivity (52.8%). These results are con-
sistent with previously reported data, showing that metallic
Cu is selective for the cleavage of C–O bonds in this reac-
tion.28 The 10Ni/MA catalyst exhibited the highest conversion
of glycerol among the prepared catalysts but had the lowest
selectivity for 1,2-PDO due to the fact that it promotes C–C
cleavage.23 A remarkable change in activity and selectivity
was observed in bimetallic catalysts. The bimetallic catalysts were
less active than monometallic ones, except for 9Cu–1Ni/MA.
Fig. 8 Catalytic activity test for hydrogenolysis of glycerol in a batch
reactor of (A) the xCu–yNi/MA catalysts and (B) CuCr2O4 and 9Cu–1Ni/Al2O3
catalysts prepared by the sol–gel and impregnation methods.
(Conditions: 493 K, 40 bar of H2, 12 h (24 h), 1 g of catalyst, 80 wt.% of
glycerol solution).
In order to confirm the superior catalytic performance of
the xCu–yNi/MA catalyst, catalytic tests were performed using
3196 | Catal. Sci. Technol., 2014, 4, 3191–3202
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
copper chromite (CuCr) and Cu–Ni/Al2O3 catalysts prepared
by the sol–gel and impregnation methods. It is well known
that copper chromite is one of the most effective catalysts for
the hydrogenolysis of glycerol to 1,2-PDO after a reduction
process when a 1 : 2 ratio of Cu : Cr is used.34 In order to con-
firm the difference resulting from the preparation methods,
9Cu–1Ni/Al2O3 was prepared with the same weight percent of
metal on a commercial gamma alumina support and the
same ratio of Cu : Ni as was used for 9Cu–1Ni/MA, followed
by a reduction process at 773 K for 3 h. Fig. 8B shows the
results of a catalytic activity test for the hydrogenolysis of
glycerol over CuCr, 9Cu–1Ni/Al2O3 (imp.), and 9Cu–1Ni/MA
for 12 h and 24 h in a batch reactor. The CuCr exhibited a
72.5% conversion and a 48.1% selectivity, which are slightly
lower values than those for 9Cu–1Ni/MA. This confirms
that 9Cu–1Ni/MA is a highly appropriate catalyst for the
hydrogenolysis of glycerol. In the case of 9Cu–1Ni/Al2O3 pre-
pared by the impregnation method, a lower conversion of
glycerol with a significant decline in selectivity from 55.3% to
41.9% was observed. This observation can be attributed to
the poor dispersion of metallic Cu on the alumina support,
as confirmed in TEM images (Fig. 3). This result suggests
that our preparation method is effective in terms of dispers-
ing the active metal, thereby resulting in the generation of a
high performance catalyst. In addition, a high performance
was found for 9Cu–1Ni/MA for 24 h, with a yield approaching
nearly 60%.
Fig. 9 XPS spectra and their deconvolution spectra of the reduced
samples in Cu 2p3/2 region; (A) 10Cu/MA, (B) 9Cu–1Ni/MA, (C) 7Cu–3Ni/MA,
and (D) 5Cu–5Ni/MA.
9Cu–1Ni/MA. This result means that a small amount of nickel
leads to an enhancement in the reducibility of surface copper
oxide on mesoporous alumina-supported Cu. On the basis of
the calculated data in Table 3, we developed a correlation plot
for the surface atomic % of Ni with Cu0/(Cu0 + Cu2+) ratios on
the surface layers of bimetallic catalysts versus their catalytic
performance. The data points were from the results of XPS
analyses and activity tests involving the 10Cu/MA, 9Cu–1Ni/MA,
7Cu–3Ni/MA and 5Cu–5Ni/MA catalysts (Fig. 10). The
9Cu–1Ni/MA catalyst had the highest Cu0/(Cu0 + Cu2+) ratio, as
shown in Fig. 10A. It is found that the ratio and reducibility of
the Cu species in the catalysts were closely related to the yield
of 1,2-PDO. Furthermore, this result indicates that the role
of Ni is not only to supply hydrogen by the aqueous phase
reforming of glycerol, but also to increase the reducibility of
Cu metal.
Partial pressure of hydrogen is an important factor to
determine the catalytic performance.39 Thus, it can be
expected that the in situ supply of hydrogen has a large effect
on selectivity for 1,2-PDO. In order to investigate the effect of
Ni metal on hydrogen production (which can affect 1,2-PDO
selectivity), we developed a correlation plot of the ratio of
Ni/Cu + Ni on the surface layers of the samples versus the
selectivity. To improve the reliability of the correlation plot,
two additional samples (8Cu–2Ni/MA, 6Cu–4Ni/MA) were pre-
pared and analyzed by XPS and subjected to reaction testing.
In addition, the conversion of glycerol for each sample was
adjusted to a similar value by controlling reaction time. As
shown in Fig. 10B, selectivity for 1,2-PDO increases with an
increase in the Ni ratio on the surface layers. This can sup-
port the large influence of hydrogen supply on selectivity for
1,2-PDO in liquid-phase glycerol hydrogenolysis. Based on
the results from Fig. 10A and B, it appears that reduced Cu
contributes to the selective conversion of glycerol to acetol
with selective cleavage of C–O bonds, and Ni contributes to
the increase in selectivity for 1,2-PDO by supplying additional
3.3 Effect of the surface ratio of Cu–Ni bimetallic catalysts on
glycerol hydrogenolysis
It is well known that the oxidation state of the active
site influences its catalytic performance, especially in metallic
catalysts.35,36 Hence, we performed XPS analyses of the
reduced catalysts containing Cu species. From this analysis,
it is possible to obtain information on how the oxidation
state of Cu and its proportion affects catalytic activity. Finally,
information can be obtained related to the reason why the
Cu–Ni/MA catalysts with various Cu : Ni ratios exhibit differ-
ent catalytic performance. Fig. 9 shows XPS spectra in the
2p3/2 region of Cu for samples with various Cu/Ni ratios. The
area of the XPS spectrum increases with increasing the fraction
of Cu in the catalysts, implying that Cu species are abundant
on the surface layers. These Cu 2p3/2 spectra can be divided
into two peaks at 932.7 and 933.6 eV, which are assigned to
metallic Cu and Cu2+ in CuO respectively.37,38 Deconvoluted
Cu 2p3/2 spectra of the 10Cu/MA, 9Cu–1Ni/MA, 7Cu–3Ni/MA,
5Cu–5Ni/MA samples are shown in Fig. 9A, B, C, and D, respec-
tively. The spectra of the bimetallic catalysts were deconvoluted
into two main peaks, metallic Cu and Cu2+ from CuO. The rela-
tive surface ratios of the Cu0 to Cu2+ in the prepared catalysts
measured by XPS are listed in Table 2. A higher Cu0/(Cu0 +
Cu2+) ratio was obtained in cases of bimetallic catalysts with
higher Cu contents, implying that reducibility increases with
increasing the fraction of Cu, consistent with the TPR data.
The 10Cu/MA catalyst showed a lower reducibility than that for
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 3191–3202 | 3197

View Article Online
Paper
Catalysis Science & Technology
Table 3 Summary of Cu metal area, Cu/Ni atomic ratio, and surface state of Cu from XPS deconvolution data of the xCu–yNi/MA catalysts
Cu metal areaa
Catalyst
(m2 g−1
)
Cu/Nib
Ni on surf.c (%) Cu on surf.c (%) Cu0 (932.7 eV) Cu2+ from CuO (933.6 eV) Cu0/(Cu0 + Cu2+
)
10Cu/MA
165.2
0
100
91.93
72.61
60.16
2859.2
2328.5
1594.3
951.5
696.1
236.1
516.9
742.3
0.820
0.908
0.755
0.562
9Cu–1Ni/MA 292.6
7Cu–3Ni/MA 376.1
5Cu–5Ni/MA 497.8
9.03
2.34
1
8.07
27.38
39.83
a
c
Calculated from H2 uptake by s-TPR.18 b Atomic composition obtained from theoretical calculation. Atomic surface composition obtained
from XPS results.
hydrogen. Meanwhile, it is thought that 10Ni/MA, a solely Ni
catalyst, is mainly active in the formation of gas-phase prod-
ucts rather than the formation of acetol. It can be concluded
that the high performance of 9Cu–1Ni/MA may be due to the
optimum Cu and Ni ratio on the surface layer for selective
C–O bond cleavage and hydrogen production of glycerol APR.
reported that the deactivation of active sites is due to the
leaching of the metal species, the sintering of active material
and the change in the surface state of reduced active
metal.14,40,41 When the reaction is conducted under H2 pres-
sure, the reaction proceeds via the acetol pathway (fast hydro-
genation of the acetol after the rate-determining step;
glycerol dehydration), and coke deposition does not have a
significant effect on the decrease of the active sites.42 We first
checked the leaching of Cu or Ni after the reaction by ICP-MS
analysis, which confirmed that the leaching of Cu and Ni
ions rarely occurred during the reaction (0.25 ppm of Cu and
0.013 ppm of Ni in aqueous phase products). The results of
the recycling test of the 9Cu–1Ni/MA catalyst are shown in
Fig. 11. After the first recycle test, the catalyst showed a glyc-
erol conversion of 69.8% (7% less than a fresh sample) and a
39.7% yield of 1,2-PDO (2.7% less than a fresh sample). As
seen in Fig. 12A and B, the particle size of the metallic Cu
was slightly larger than a fresh one, and the structure and
morphology of the catalyst was not changed after the reac-
tion. However, after the 2nd and 3rd recycling tests, the
metallic Cu was rapidly sintered, resulting in the growth of
the particle size from 10–20 nm to 40–50 nm. From the XRD
patterns, the peak for CuO (Cu2+) was sharply increased and
an additional peak of Cu2O (Cu+) appeared at 2θ = 42.6°
(JCPDS 34-1354). This suggests that the reduced metallic
copper was oxidized and sintered during the hydrogenolysis
of glycerol in the aqueous phase, resulting in the rapid
deactivation of the catalyst and a corresponding decrease in
activity (57.2% conversion and 33.9% yield of 1,2-PDO). The
3.4 Durability and reusability of the prepared catalyst
In order to investigate the durability and reusability of the
9Cu–1Ni/MA catalyst, the catalyst was recycled three times
after the reaction. For the recycling test, the product obtained
using the used catalyst was transferred to a centrifugation
tube, and the used catalyst was separated from the liquid-
phase product by centrifugation. The used catalyst was then
washed with distilled water three times to remove the
adsorbed organic species, and then dried at 80 °C for 12 h.
In liquid-phase glycerol hydrogenolysis, it was previously
Fig. 10 Correlations of catalytic activity for the glycerol
hydrogenolysis (A) with copper surface state; Cu0/(Cu0 + Cu2+) and
(B) with surface atomic ratio of Ni/(Cu + Ni) of the xCu–yNi/MA
catalysts (surface atomic % of Ni and Cu: 17% and 83% in 8Cu–2Ni/MA,
33% and 67% in 6Cu–4Ni/MA by XPS analysis).
Fig. 11 Catalytic recycling tests of the 9Cu–1Ni/MA catalyst in glycerol
hydrogenolysis.
3198 | Catal. Sci. Technol., 2014, 4, 3191–3202
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
hydrogenolysis of glycerol to 1,2-PDO. From the previous lit-
erature reports, most studies suggested that the conversion
of glycerol into 1,2-PDO involves dehydration to acetol and
subsequent hydrogenation to 1,2-PDO.43–45 All of the pro-
posed mechanisms are based on the use of a bi-functional
catalyst with additional hydrogen gas for the hydrogenation
reaction. Based on this mechanism, Cu in the Cu–Ni/MA
selectively cleaves the C–O bonds, leading to the conversion
of acetol. In addition, the hydrogen generated from the aque-
ous phase reforming of glycerol on Ni then leads to the con-
version of acetol to 1,2-PDO. Therefore, the acetol to 1,2-PDO
step can be controlled by the hydrogen pressure produced
from the aqueous phase reaction of glycerol by the Ni metal
catalyst. An activity test as time-on-stream without hydrogen
pressure was carried out in order to confirm the proposed
mechanistic pathway. Monometallic (10Cu/MA) and bimetal-
lic (7Cu–3Ni/MA) catalysts were used as model catalysts to
observe the role of metallic Ni on the hydrogenolysis of glyc-
erol. The conversion of glycerol and product selectivity for
10Cu/MA and 7Cu–3Ni/MA are shown as a function of time
in Fig. 13. It was observed that the 1,2-PDO can be produced
over only a Cu monometallic catalyst in the absence of hydro-
gen. This is due to the APR of glycerol on the distinct role of
Cu and Al in aqueous phase. Mane et al.21 revealed that
in situ generated hydrogen from the APR of glycerol was
dissociatively adsorbed on copper metal, causing hydroxyl-
ation of the Al surface. The formaldehyde as an intermediate
was detected via dehydrogenation and C–C cleavage of glyc-
erol on Cu. In our case, it seemed that the Cu/MA catalyst
also plays a role in C–C cleavage of glycerol via hydroxylation
of the mesoporous alumina evidenced by detected formalde-
hyde, although the catalytic efficiency of copper metal is not
remarkable. In 7Cu–3Ni/MA, acetol was rapidly produced and
then diminished after 3 h with increasing yield of 1,2-PDO,
as shown in Fig. 13B. At the same time, a high conversion
was recorded. This indicates that the hydrogenation of acetol
involved the use of hydrogen produced by in situ generation
on the metallic Ni. However, this reaction pathway does
not proceed in the case of a catalyst only containing Cu
(Fig. 13A). The conversion of glycerol reached a plateau with
a low value after 5 h of reaction time. Selectivity for 1,2,-PDO
and acetol was almost maintained. This can be attributed to
the fact that there was no hydrogen to be used in the hydro-
genation reaction.
Fig. 12 (A) XRD patterns and (B) HR-TEM images of the 9Cu–1Ni/MA
catalyst after recycle tests.
morphology of the catalyst was also significantly altered as
seen in the HR-TEM image after the 3rd recycle test. In order
to clarify the effect of Ni addition on catalytic stability,
recycling tests using monometallic (10Cu/MA) and bimetallic
(7Cu–3Ni/MA) catalysts were carried out (Fig. S1†). The
glycerol conversion of the 10Cu/MA catalyst was higher than
that of 7Cu–3Ni/MA. However, the trend was inversed after
the 3rd recycling tests. It was also observed that the 1,2-PDO
selectivity of the Cu/MA catalyst decreased while the selectiv-
ity of the 7Cu–3Ni/MA catalyst for 1,2-PDO was maintained
during the recycling test. These results could arise from the
lower deactivation of bimetallic catalysts compared to mono-
metallic catalysts in terms of sintering and oxidation of
metallic Cu. These results were confirmed by TEM and
XRD analysis (Fig. S2 and S3†). It can be concluded that the
addition of Ni enhances the catalytic stability in glycerol
hydrogenolysis. Although a severe decrease in the activity was
observed, the selectivity for 1,2-PDO was not significantly
reduced during the recycle test. This is because the active
sites that converts glycerol into acetol could be decreased by
the sintering process. In addition, the particle size of the
metallic nickel was unchanged and its metallic phase was
retained during the reaction (as evidenced by XRD data),
resulting in maintaining a continuous supply of hydrogen
which can affect the selectivity of 1,2-PDO.
The time-on-stream tests of the other xCu–yNi/MA
catalysts were carried out under the same conditions. On
increasing the amount of Ni in xCu–yNi/MA catalysts, the
conversion of glycerol increased due to the aqueous phase
reforming (APR) on Ni metal (Fig. S4†). This result was quite
different from the result (Fig. 8) in the presence of H2. As
described before, the relatively low hydrogen pressure
enhances the adsorption of glycerol on Ni metal, resulting
in increasing the conversion of APR of glycerol. After 10 h,
9Cu–1Ni/MA showed the highest value of selectivity of
1,2-PDO, however, the yield of 1,2-PDO on 7Cu–3Ni/MA was
higher than the other catalysts (Table S1†). A little amount of
3.5 Mechanistic discussion with time-on-stream behavior
To develop a technology that produces 1,2-PDO from glycerol
without the need for an external hydrogen source, it is neces-
sary to understand the mechanism responsible for the
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 3191–3202 | 3199

View Article Online
Paper
Catalysis Science & Technology
the hydrogenolysis of glycerol. It should be noted that the
reaction with CTH needs both the additional hydrogen donor
compound and a lot of water with glycerol in the feed. In this
case, the efficiency of the unit weight of catalyst is very low.
The optimum amount of Ni in the Cu–Ni bimetallic catalyst
and the reaction pathway are also different in our system
because of the high concentration of glycerol. The reaction
pathway of glycerol hydrogenolysis is still controversial between
dehydrogenation–dehydration–hydrogenation (Montassier et al.46)
and dehydration–hydrogenation (Dasari et al.47) mechanisms.
In our system, the hydrogenolysis of glycerol is followed by
the dehydration–hydrogenation route. It is known that the
dehydration–hydrogenation route is suitable for the condi-
tions of acidic sites in the catalyst, hydrogen pressure and
small amount of water in feed.48 Therefore, at first, the
oxygen on terminal OH groups of glycerol are adsorbed on
Cu metal, then the dehydration into acetol occurs at the
acidic sites of the alumina support (Scheme 1). In the absence
of Ni, a small amount of 1,2-PDO is produced because of the
absence of hydrogen. After the addition of Ni, there are two
possible pathways to produce hydrogen by APR of glycerol on
Ni metal.49 Based on the adsorption energy calculation of
glycerol,50,51 both C–C and C–O bonds are formed on the Ni
surface. The binding energies and preferred adsorption sites
for atomic C and O on various metals are as follows: Pt
(C–metal: −7.42 eV, O–metal: −4.59 eV), Pd (C–metal: −6.93 eV,
O–metal: −4.49 eV) and Ni (C–metal: −6.81 eV, O–metal:
−5.60 eV). Therefore, the adsorbed species could be metal–carbon
or metal–oxygen bonds. For the formation of metal–carbon
(route A), the glycerol first undergoes reversible dehydrogena-
tion steps to become adsorbed on Ni metal.49 Thus, it is
thought that the adsorption of metal–oxygen (route B) favor-
ably occurs in a pressurized H2 environment. Adsorbed
Fig. 13 Time-on-stream activity for (A) 10Cu/MA and (B) 7Cu–3Ni/MA
catalysts without a hydrogen supply in a batch reactor. (Conditions:
220 °C, no pressure, 1 g of catalyst, 80 wt.% of glycerol solution).
1-PO and 2-PO was produced. This implies that the addi-
tional dehydration–hydrogenation of propylene glycol does
not occur in the absence of hydrogen gas due to the domi-
nant reaction of the aqueous phase reforming on the Ni
metal. On the 5Cu–5Ni/MA catalyst, the acetol selectivity and
the yield of 1,2-PDO rapidly decreased, although hydrogen
was produced much more than the other catalysts. From the
activity test in the absence of H2, the ratio of Ni to Cu was
found to be an important factor for the high yield of
1,2-PDO. The optimum value of the ratio of Ni to Cu is differ-
ent with respect to the pressurized atmosphere; the presence
of hydrogen gas or the absence of hydrogen gas. From these
results, it can be confirmed that metallic Cu and Ni catalyses
C–O cleavage and glycerol APR to supply hydrogen gas for
the reaction.
Previously, Gandarias et al.15 reported the hydrogenolysis
of glycerol over Cu–Ni bimetallic catalysts using formic acid
as a hydrogen donor (CTH, catalytic transfer hydrogen). In
their report, the 20Ni–15Cu catalyst showed the highest yield
of 1,2-PDO among the various xNi–yCu catalysts in a feed of
0.06 molglycerol g−1 at 493 K for 16 h. The value of 1,2-PDO
cat
yield was larger than that of our prepared catalyst, 7Cu–3Ni/MA;
about 10% (reaction conditions: 0.43 molglycerol g−1cat, 493 K,
10 h in Fig. 13B). However, the production of 1,2-PDO per
gram of catalyst in our system is much larger than that of the
previously reported system when considering the amount of
reacted glycerol. This implies that aqueous phase reforming
(APR) is more effective than CTH to supply the hydrogen to
Scheme 1 Proposed reaction pathway for hydrogenolysis of glycerol
to 1,2-propanediol with aqueous phase reforming over Cu–Ni/MA
catalyst.
3200 | Catal. Sci. Technol., 2014, 4, 3191–3202
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
hydrogen from the APR of glycerol on Ni is delivered to the
primary or terminal carbon of the glycerol which is adsorbed
onto Cu. After the protonation to terminal OH by the acidic
sites, the Cu metal cleaves the C–O bond, then the hydrogen
is transferred from Ni metal to the terminal carbon at an
adjacent Cu metal, simultaneously.42 When the APR is used
for the production of hydrogen in glycerol hydrogenolysis,
both ratio control of Cu to Ni and the reaction conditions are
significant factors in the production of 1,2-PDO with high
yield. In addition, CTH requires an additional donor
compound and semi-continuous reactor for high yield of
1,2-PDO.48 Therefore, the APR is a more effective source of
hydrogen for the conversion of glycerol to various polyols.
4 N. D. Kim, J. R. Park, D. S. Park, B. K. Kwak and J. Yi, Green
Chem., 2012, 14, 2638.
5 Y. Nakagawa and K. Tomishige, Catal. Sci. Technol., 2011,
1, 179.
6 L. Huang, Y.-L. Zhu, H.-Y. Zheng, Y.-W. Li and Z.-Y. Zeng,
J. Chem. Technol. Biotechnol., 2008, 83, 1670.
7 M. Balaraju, V. Rekha, P. S. Sai Prasad, R. B. N. Prasad and
N. Lingaiah, Catal. Lett., 2008, 126, 119.
8 S. Sato, M. Akiyama, K. Inui and M. Yokota, Chem. Lett.,
2009, 38, 560.
9 Z. Xiao, J. Xiu, X. Wang, B. Zhang, C. T. Williams, D. Su and
C. Liang, Catal. Sci. Technol., 2013, 3, 1108.
10 N. D. Kim, S. Oh, J. B. Joo, K. S. Jung and J. Yi, Korean J.
Chem. Eng., 2010, 27, 431.
11 M. Balaraju, K. Jagadeeswaraiah, P. S. S. Prasad and
N. Lingaiah, Catal. Sci. Technol., 2012, 2, 1967.
12 J. Zhou, L. Guo, X. Guo, J. Mao and S. Zhang, Green Chem.,
2010, 12, 1835.
13 E. D'Hondt, S. Van de Vyver, B. F. Sels and P. A. Jacobs,
Chem. Commun., 2008, 6011.
14 D. Roy, B. Subramaniam and R. V. Chaudhari, Catal. Today,
2010, 156, 31.
15 I. Gandarias, J. Requies, P. L. Arias, U. Armbruster and
A. Martin, J. Catal., 2012, 290, 79.
16 Y. Kim, B. Lee and J. Yi, Korean J. Chem. Eng., 2002, 19, 908.
17 P. Kim, Y. Kim, H. Kim, I. K. Song and J. Yi, J. Mol. Catal. A:
Chem., 2005, 231, 247.
18 G. V. Sagar, P. V. R. Rao, C. S. Srikanth and K. V. R. Chary,
J. Phys. Chem. B, 2006, 110, 13881.
4 Conclusions
Cu–Ni bimetallic catalysts supported on mesoporous alumina
(MA) with different metal compositions were successfully syn-
thesized using a post-hydrolysis sol–gel method. The Cu–Ni/MA
catalysts exhibited mesoporosities and had a relatively high
metal dispersion (ca. 33 wt.%), resulting in the formation of
nano-sized metal particles. Among the prepared samples, the
9Cu–1Ni/MA catalyst showed the highest catalytic performance,
because of the increased hydrogen utilization and the high
surface ratio of Cu0/(Cu0 + Cu2+). On the Cu–Ni bimetallic cata-
lysts, the surface ratio of Ni/(Cu + Ni) and the Cu0/(Cu0 + Cu2+)
were related to the selectivity and yield in the hydrogenolysis of
glycerol to 1,2-PDO. Through the addition of Ni on Cu-based
catalysts, the reducibility of Cu was increased and hydrogen
sources were generated in situ by the aqueous phase reforming
of glycerol. The generated hydrogen was subsequently used for
the hydrogenation of acetol in the hydrogenolysis of glycerol to
1,2-PDO. By the activity test in the absence of H2, the ratio of
Ni to Cu was an important factor to the high yield of 1,2-PDO.
The optimum value of the ratio of Ni to Cu is different with
respect to the pressurized atmosphere; the presence of hydro-
gen gas or the absence of hydrogen gas. Their different reac-
tion pathways are proposed over Cu–Ni bimetallic catalysts for
the hydrogenolysis of glycerol with aqueous phase reforming.
19 P. Kim, Y. Kim, H. Kim, I. K. Song and J. Yi, Appl. Catal., A,
2004, 272, 157.
20 R. B. Mane, S. E. Kondawar, P. S. Niphadkar, P. N. Joshi,
K. R. Patil and C. V. Rode, Catal. Today, 2012, 198, 321.
21 R. B. Mane and C. V. Rode, Green Chem., 2012, 14, 2780.
22 I. Jiménez-Morales, F. Vila, R. Mariscal and A. Jiménez-López,
Appl. Catal., B, 2012, 117–118, 254.
23 G. Wen, Y. Xu, H. Ma, Z. Xu and Z. Tian, Int. J. Hydrogen
Energy, 2008, 33, 6657.
24 X. Wang, X. Pan, R. Lin, S. Kou, W. Zou and J.-X. Ma, Int. J.
Hydrogen Energy, 2010, 35, 4060.
25 F. E. López-Suárez, A. Bueno-López and M. J. Illán-Gómez,
Appl. Catal., B, 2008, 84, 651.
26 A. Alhanash, E. F. Kozhevnikova and I. V. Kozhevnikov, Appl.
Catal., A, 2010, 378, 11.
27 B. K. Kwak, D. S. Park, Y. S. Yun and J. Yi, Catal. Commun.,
2012, 24, 90.
28 A. Torres, D. Roy, B. Subramaniam and R. V. Chaudhari, Ind.
Eng. Chem. Res., 2010, 49, 10826.
Acknowledgements
This work was supported by the National Research Founda-
tion of Korea (NRF) grant funded by the Korea government
(MEST) (no. 2013R1A2A2A01067164).
29 I. Gandarias, S. G. Fernández, M. E. Doukkali, J. Requies
and P. L. Arias, Top. Catal., 2013, 56, 995.
Notes and references
1 M. G. Musolino, L. A. Scarpino, F. Mauriello and
R. Pietropaolo, ChemSusChem, 2011, 4, 1143.
30 I. Gandarias, P. L. Arias, J. Requies, M. B. Güemez and
J. L. G. Fierro, Appl. Catal., B, 2010, 97, 248.
2 B. Katryniok, S. Paul, M. Capron and F. Dumeignil,
ChemSusChem, 2009, 2, 719.
3 M. D. Soriano, P. Concepción, J. M. L. Nieto, F. Cavani,
S. Guidetti and C. Trevisanut, Green Chem., 2011, 13, 2954.
31 R. L. Manfro, T. P. M. D. Pires, N. F. Ribeiro and
M. M. V. M. Souza, Catal. Sci. Technol., 2013, 3, 1278.
32 V. Ponec, Appl. Catal., A, 2001, 222, 31.
33 B. Coq and F. Figueras, Coord. Chem. Rev., 1998, 178–180, 1753.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 3191–3202 | 3201

View Article Online
Paper
Catalysis Science & Technology
34 N. D. Kim, S. Oh, J. B. Joo, K. S. Jung and J. Yi, Top. Catal.,
2010, 53, 517.
43 J. Chaminand, L. Djakovitch, P. Gallezot, P. Marion, C. Pinel
and C. Rosier, Green Chem., 2004, 6, 359.
35 Y. Zhu, W. Lu, H. Li and H. Wan, J. Catal., 2007, 246, 382.
36 M. S. Kumar, N. Hammer, M. Rønning, A. Holmen, D. Chen,
J. C. Walmsley and G. Øye, J. Catal., 2009, 261, 116.
37 A. Crucq, L. Degols, G. Lienard and A. Frennet, Stud. Surf.
Sci. Catal., 1983, 17, 137.
44 T. Miyazawa, Y. Kusunoki, K. Kunimori and K. Tomishige,
J. Catal., 2006, 240, 213.
45 A. Perosa and P. Tundo, Ind. Eng. Chem. Res., 2005, 44, 8535.
46 C. Montassier, D. Giraud and J. Barbier, Fine Chem. Heterog.
Catal., 1998, 165.
38 F. Vila, M. Lopez Granados, M. Ojeda, J. L. G. Fierro and
R. Mariscal, Catal. Today, 2012, 187, 122.
47 C.-W. Chiu, M. A. Dasari and G. J. Suppes, AIChE J.,
2006, 52, 3543.
39 M. A. Dasari, P. Kiatsimkul, W. R. Sutterlin and G. J. Suppes,
Appl. Catal., A, 2005, 281, 225.
48 A. Martin, U. Armbruster, I. Gandarias and P. L. Arias, Eur. J.
Lipid Sci. Technol., 2013, 115, 9.
40 Z. Huang, F. Cui, H. Kang, J. Chen and C. Xia, Appl. Catal., A,
2009, 366, 288.
41 Z. Huang, F. Cui, J. Xue, J. Zuo, J. Chen and C. Xia, Catal.
Today, 2012, 183, 42.
49 R. R. Davda, J. W. Shabaker, G. W. Huber, R. D. Cortright
and J. A. Dumesic, Appl. Catal., B, 2005, 56, 171.
50 R. Alcalá, M. Mavrikakis and J. A. Dumesic, J. Catal.,
2003, 218, 178.
42 I. Gandarias, P. L. Arias, J. Requies, M. E. Doukkali and
M. B. Güemez, J. Catal., 2011, 282, 237.
51 B. Liu and J. Greeley, Phys. Chem. Chem. Phys., 2013,
15, 6475.
3202 | Catal. Sci. Technol., 2014, 4, 3191–3202
This journal is © The Royal Society of Chemistry 2014