Influence of preparation method on supported Cu-Ni alloys and their catalytic properties in high pressure CO hydrogenation

Article abstract of DOI:10.1039/c3cy00546a

Silica supported Cu-Ni (20 wt% Cu + Ni on silica, molar ratio of Cu/Ni = 2) alloys are prepared via impregnation, coprecipitation, and deposition- coprecipitation methods. The approach to co-precipitate the SiO₂ from Na₂SiO₃ together with metal precursors is found to be an efficient way to prepare high surface area silica supported catalysts (BET surface area up to 322 m² g^-1, and metal area calculated from X-ray diffraction particle size up to 29 m² g^-1). The formation of bimetallic Cu-Ni alloy nanoparticles has been studied during reduction using in situ X-ray diffraction. Compared to impregnation, the coprecipitation and deposition-coprecipitation methods are more efficient for preparation of small and homogeneous Cu-Ni alloy nanoparticles. In order to examine the stability of Cu-Ni alloys in high pressure synthesis gas conversion, they have been tested for high pressure CO hydrogenation (50 bar CO and 50 bar H₂). These alloy catalysts are highly selective (more than 99 mol%) and active for methanol synthesis; however, loss of Ni caused by nickel carbonyl formation is found to be a serious issue. The Ni carbonyl formation should be considered, if Ni-containing catalysts (even in alloyed form) are used under conditions with high partial pressure of CO. The Royal Society of Chemistry.

Full text of DOI:10.1039/c3cy00546a

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Influence of preparation method on supported Cu–Ni
alloys and their catalytic properties in high pressure
CO hydrogenation
Cite this: Catal. Sci. Technol., 2014,
4, 378
a
a
b
a
Qiongxiao Wu, Winnie L. Eriksen, Linus D. L. Duchstein, Jakob M. Christensen,
bc
b
d
ae
Christian D. Damsgaard, Jakob B. Wagner, Burcin Temel, Jan-Dierk Grunwaldt
and Anker D. Jensen*
a
Silica supported Cu–Ni (20 wt% Cu + Ni on silica, molar ratio of Cu/Ni = 2) alloys are prepared via
impregnation, coprecipitation, and deposition–coprecipitation methods. The approach to co-precipitate
the SiO
2 2 3
from Na SiO together with metal precursors is found to be an efficient way to prepare high
2
−1
surface area silica supported catalysts (BET surface area up to 322 m g , and metal area calculated from
2
−1
X-ray diffraction particle size up to 29 m g ). The formation of bimetallic Cu–Ni alloy nanoparticles has
been studied during reduction using in situ X-ray diffraction. Compared to impregnation, the
coprecipitation and deposition–coprecipitation methods are more efficient for preparation of small and
homogeneous Cu–Ni alloy nanoparticles. In order to examine the stability of Cu–Ni alloys in high
pressure synthesis gas conversion, they have been tested for high pressure CO hydrogenation (50 bar
Received 25th July 2013,
Accepted 3rd November 2013
2
CO and 50 bar H ). These alloy catalysts are highly selective (more than 99 mol%) and active for
methanol synthesis; however, loss of Ni caused by nickel carbonyl formation is found to be a serious
issue. The Ni carbonyl formation should be considered, if Ni-containing catalysts (even in alloyed form)
are used under conditions with high partial pressure of CO.
DOI: 10.1039/c3cy00546a
www.rsc.org/catalysis
it was found that homogeneous alloys only formed when the
1
. Introduction
1
3,28
Cu–Ni alloys were rich in Ni.
Similarly, on the basis of
Cu–Ni alloys are widely used as catalysts in, for example, steam
in situ X-ray diffraction (XRD) analysis, we recently found that
silica supported Cu–Ni alloys prepared via impregnation
formed a homogeneous solid solution of Cu and Ni at high
1–6
7–9
reforming, dimethyl carbonate synthesis, hydrodeoxygena-
1
0–12
13–17
18,19
tion,
water gas shift,
CO
2
hydrogenation,
and CO
2
0–23
29
hydrogenation.
The chemical and physical properties of
nickel concentrations. At lower nickel content, there is a seg-
29
Cu–Ni alloys as well as the catalytic properties of the alloys
have been described extensively in the literature. However, the
structure of supported Cu–Ni catalysts is still unclear. With
respect to bulk Cu–Ni alloy formation, the phase diagram of
the binary Cu–Ni system shows phase segregation for a broad
regation into metallic Cu and Cu–Ni alloy. In that case, the
particle size was found to decrease significantly with increas-
ing Ni content, and the phase segregation seen for Cu-rich
compositions might be partly caused by the larger metal par-
ticle size in the Cu-rich range. The particle size has in some
cases been found to be an important parameter for alloy for-
2
4,25
range of mixing ratios at room temperature
and formation
2
9–31
of a solid solution only above a certain critical temperature,
mation.
For example, bimetallic, bulk Cu–Ag shows phase
typically found in the 200–300 °C range depending on the
segregation for a broad composition range, but when the clus-
2
4–27
alloy composition.
Regarding the supported Cu–Ni system,
ters are smaller than a critical size (ca. 1 nm), there is no
3
0
phase segregation at all. It has been reported that prepa-
ration of metal particles from nitrate precursors via impreg-
nation generally leads to significantly larger metal particles
a
Department of Chemical and Biochemical Engineering, Technical University of
Denmark, Søltofts Plads, 2800 Kgs. Lyngby, Denmark. E-mail: aj@kt.dtu.dk;
Fax: +45 4588 2285; Tel: +45 45252841
3
2–35
than preparation by, for example, co-precipitation.
b
Center for Electron Nanoscopy, Technical University of Denmark,
Cu-based catalysts are often prepared by a co-precipitation
approach since high metal loadings can be achieved in com-
bination with high metal dispersion (small particles) and
2
800 Kgs. Lyngby, Denmark
c
Center for Individual Nanoparticle Functionality, Technical University of
Denmark, 2800 Kgs. Lyngby, Denmark
36–38
d
improved stability.
In this work, we have investigated the
Haldor Topsøe A/S, Nymøllevej 55, 2800 Kgs. Lyngby, Denmark
e
effect of preparation methods on Cu-rich, Cu–Ni (Cu/Ni = 2/1,
Institute for Chemical Technology and Polymer Chemistry, Karlsruhe Institute of
Technology, Engesserstrasse 20, 76131 Karlsruhe, Germany
molar ratio) alloy formation on a silica support. We have prepared
378 | Catal. Sci. Technol., 2014, 4, 378–386
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
silica supported Cu–Ni catalysts via three different preparation
methods: incipient wetness impregnation, coprecipitation
and deposition–coprecipitation. The aim was to evaluate if pre-
cipitation methods would be able to yield homogeneous alloys
also in the Cu-rich composition range, where this is difficult
to achieve using impregnation methods. If this can be achieved,
it could potentially be of importance in many of the applica-
tions of Cu–Ni catalysts. Different characterization techniques
such as in situ XRD, BET surface area measurements, thermo-
gravimetric analysis (TGA) and transmission electron micros-
copy (TEM) have been used to evaluate the structure of the
supported Cu–Ni systems.
process, pH was controlled through the flows of the nitrate and
silicate solutions, and the temperature was maintained at 60 °C.
At the end of the precipitation process, the precipitate was
aged at 60 °C for 1 h (with continuous stirring). During this
hour, the pH value was maintained at 7 ± 0.02 by addition of
the remaining metal solution. Then, the precipitate was aged
at room temperature overnight. Afterwards, the precipitate was
filtered and washed three times with demineralized water,
followed by drying overnight at 110 °C. The sample was then
−1
calcined at 300 °C in a flow of N for 4 h (heating rate 2 °C min ).
2
This sample was denoted as Cu–Ni/SiO –Copr.
2
Deposition–coprecipitation. An aqueous solution of metal
nitrates [(Cu(NO ) ·3H O (0.0690 M), and Ni(NO ) ·6H O
Furthermore, Cu–Ni alloys have, as touched upon above,
attracted attention as catalysts for CO hydrogenation. Cu–Ni
catalysts have been reported to hydrogenate CO to different
3
2
2
3 2
2
(0.0345 M), 250 ml] was coprecipitated using a solution of
Na CO (Sigma-Aldrich, 500 ml, 0.103 M) in 1.0 L of preheated
(60 °C), demineralised water containing a suspension of
2
3
3
9,40
21
products such as higher alcohols,
methanol.
hydrocarbons, and
It is known that metallic Ni in the presence
of CO could form the volatile carbonyl Ni(CO) , which could
cause Ni particle sintering and loss of Ni under certain condi-
2
3,41–44
2
−1
SiO (207 m g , in a size range of 150–300 μm). During the
2
4
precipitation process, pH was maintained at 7 ± 0.05 by
controlling the flows of the nitrate and carbonate solutions,
and the temperature was maintained at 60 °C. At the end of
the precipitation process, the precipitate was aged at 60 °C
for 1 h (with continuous stirring). During this hour, the pH
value was maintained at 7 ± 0.02 by the addition of the
remaining metal solution. The precipitate was then aged at
room temperature overnight. Afterwards, the precipitate was
filtered and washed three times with demineralized water,
followed by drying overnight at 110 °C. The sample was then
4
5–47
tions. However, the influence of Ni carbonyl formation
has not been studied in our own previous work on this
system or in the general literature
4
3
20–23,44
on high pressure
CO hydrogenation over Cu–Ni catalysts. It remains unknown
whether Ni in the Cu–Ni alloys can be removed with the efflu-
ent syngas stream. It would thus be important to determine if
there is any Ni loss due to carbonyl formation. In order to
examine the stability of Cu–Ni, we have tested the Cu–Ni cata-
lysts in high pressure CO hydrogenation, and subsequently
analyzed the composition of both fresh and spent catalysts by
elemental analysis using inductively coupled plasma optical
emission spectroscopy (ICP-OES).
calcined at 300 °C in a flow of N for 4 h (heating rate 2 °C
2
−
1
2
min ). This sample was denoted as Cu–Ni/SiO –Dep–Copr.
2.2 Characterization
The chemical composition of both fresh and spent (tested in
high pressure CO hydrogenation) samples was determined by
an ICP-OES (Perkin-Elmer, model Optima 7300) using Ar as
2
. Experimental section
2
.1 Preparation of Cu–Ni precursor materials
Silica supported Cu–Ni catalyst precursors (20 wt% Cu + Ni,
Cu/Ni = 2 molar ratio) were prepared by incipient wetness
impregnation, coprecipitation, and deposition–coprecipitation.
The specific preparation procedures are described below.
Incipient wetness impregnation. An aqueous solution was
prepared by dissolving Cu(NO ) ·3H O and Ni(NO ) ·6H O
2 2 7
plasmogene. In a typical analysis, 5 g of K S O was melted
together with 0.25 g sample and dissolved in 10 ml HCl in
200 ml water and then analyzed.
2
The BET surface area was determined by N -adsorption at
77 K by means of a QuantaChrome Autosorb iQ gas sorption
2
analyzer. The specific surface area was analyzed using a 7-point,
3
2
2
3 2
2
(
(
Sigma-Aldrich) in demineralized water. Silica particles
600–1400 μm, with a BET surface area of 207 m g , supplied
linear BET plot in the range of p/p
0
= 0.05–0.3. Prior to the
2
−1
BET measurement, the sample was dried under nitrogen flow
−1
by Saint-Gobain Norpro) were impregnated with an amount of
the aqueous solution corresponding to the pore volume of
the support (1.2 ml g for SiO ). After 1 h aging at room tempera-
(200 Nml min ) at 150 °C for 2 h.
In order to understand the reduction and calcination
processes of the nitrate sample precursor (Cu–Ni/SiO –IWI),
−1
2
2
ture, the catalyst precursors were dried at 110 °C in air overnight.
TGA was carried out with a NETZSCH STA-449-F1 thermo-
analyzer. First, 15 mg of the sample was loaded into an alu-
mina sample holder. The sample was then heated to 400 °C
This sample was denoted as Cu–Ni/SiO
Coprecipitation. An aqueous solution of mixed metal nitrates
(Cu(NO ) ·3H O (0.0690 M) and Ni(NO ) ·6H O (0.0345 M), 250 ml]
2
–IWI.
−
1
[
in a 120 ml min flow of 4.2 mol% H /N (for reduction) or
2 2
3
2
2
3 2
2
−
1
with 25 ml of 37 wt% HNO
3
was coprecipitated using a
2
pure N (for calcination) with a heating rate of 2 °C min .
solution of Na Si O (Sigma-Aldrich, 500 ml, 0.213 M) in 1.0 L
In situ XRD was performed with a PANalytical X'Pert PRO
diffractometer equipped with an Anton Paar XRK 900 in situ
2
2 3
of preheated (60 °C), demineralized water. The addition of
HNO to the metal nitrate solution was required to keep the
3
cell and a gas flow control system for H
2
-TPR. Silica supported
pH constant at 7 ± 0.05 during the precipitation process due
to the basicity of the silicate solution. During the precipitation
Cu–Ni samples were heated in flowing 2 mol% H /He up to
300 °C for 1 h (heating rate 20 °C min below 160 °C and
2
−
1
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 378–386 | 379

View Article Online
Paper
Catalysis Science & Technology
−
1
1
1
°C min above 160 °C). Each XRD measurement took
ni y
i
0 minutes, and the temperature was maintained during the
Si 
100%
ni y
i
XRD measurement. After reduction, the sample was cooled
to room temperature, and then a long XRD scan (640 minutes)
was performed in order to estimate the particle size of the
formed metallic phase. An estimate of the average particle
diameter (dp,XRD) was obtained from diffraction peak broaden-

Here n (n ≥ 1) is the number of carbon atoms in component
i
i
i
i, and y is the mole fraction of component i in the product.
The space time yield (STYMeOH) of methanol is the production
rate of methanol per mass of catalyst:
4
8
ing by means of the Scherrer equation:
MMeOH FMeOH
1
1
0
.9

h

STYMeOH

kg kgcat
dp,XRD



mcat
1/2
cos



Here, MMeOH is the molar mass of methanol and FMeOH is the
molar flow rate of methanol out of the reactor. We report the
STY in the units of mass of methanol produced per hour and
per mass of catalyst (mcat), which includes both active metals
and support.
Here λ is the X-ray wavelength, θ is the Bragg angle, and β1/2
is the full width at half maximum of the diffraction peak
corrected for instrumental broadening (using a highly crystal-
line Si standard).
TEM images were acquired using an FEI Titan 80-300 aber-
ration corrected microscope operated at 300 kV and an FEI
Tecnai T20G2S microscope operated at 200 kV. The samples
were dispersed in dry form on the TEM grids, which were
Au-grids coated with a holey carbon film.
3
. Results and discussion
3.1. Catalyst structure
3.1.1 ICP-OES elemental analyses and BET surface area
studies. The ICP-OES elemental analysis results for all fresh
2
.3 CO hydrogenation
samples are listed in Table 1. The desired composition was
CO hydrogenation experiments were performed in a high pres-
sure, fixed-bed flow reactor setup. Detailed descriptions of the
experimental setup and the experimental procedure have been
provided in ref. 49–52. The catalyst is located in a quartz tube
within a stainless steel pressure shell. As the inside of the
quartz tube is pressurized, the pressure shell is also pressur-
ized with nitrogen to ensure that no pressure gradient exists
across the quartz tube wall. High pressure CO hydrogenation
20 wt% Cu + Ni metal loading on silica, and the samples are
close to this value. In addition, the Cu/Ni atomic ratios are
also close to the expected atomic ratio of Cu/Ni = 2. Com-
2
−1
pared to the original SiO support (207 m g ) the specific
2
surface area of the impregnated sample (Cu–Ni/SiO –IWI)
2
2
−1
has decreased about 44% to 116 m g . This is presumably
due to blocking of the pore structure by the loaded metals.
Using the same silica, the surface area of the deposition–
−
1
was carried out at P = 100 bar, T = 275 °C, GHSV = 4000 h
based on total bed volume and gas flow referring to 298.15 K
and 1 bar), feed: H /CO = 2 (vol/vol). The temperature was mea-
coprecipitation sample, Cu–Ni/SiO
2
2
–Dep–Copr, increases to
53 m g . The increase in the surface area can presumably
be ascribed to the presence of small precipitated nanoparticles
in the sample. The coprecipitated sample, Cu–Ni/SiO –Copr,
(
2
−1
2
sured in a steel thermo pocket touching the external surface of
the quartz tube. The product was analyzed by using an online
gas chromatograph equipped with flame ionization and ther-
mal conductivity detectors (6890N GC-FID/TCD from Agilent
Technologies). The characterized oxygenates were: DME, meth-
anol, ethanol, 1-propanol, 1-butanol, 2-butanol, 2-methyl-1-
propanol and ethyl acetate. Additionally, methane, ethane,
ethene, propane, CO, CO and H were analyzed. A bubble flow
2
−1
2
shows the highest specific surface area of 322 m g . This
approach to co-precipitate the SiO2 from Na SiO3 together
2
with metal precursors seems to be an efficient way to prepare
high surface area silica supported catalysts. The specific sur-
face areas of the three samples prepared by different
methods are in increasing order: impregnation < deposition–
coprecipitation < coprecipitation.
2
2
meter was used for determining the volumetric flow rate of the
reactor effluent. Prior to the reaction, all catalyst precursors
were reduced in situ at atmospheric pressure by a flow of
3.1.2 In situ XRD and TGA studies. The diffraction patterns
marked “im” in Fig. 1 are the XRD patterns at selected
temperatures during in situ reduction of the Cu–Ni/SiO –IWI
2
1
2 2
.4 mol% H in N for 12–14 h at 280 °C (which was reached
precursor in a flow of 2 mol% H /He. The catalyst precursor
−
1
2
using a heating rate of 1 °C min ). The carbon mass balance
contains Cu/Ni hydroxynitrates that are formed during the
drying of impregnated metal nitrate precursors at elevated
temperature (100 °C), which results in agglomeration during
was generally fulfilled to within 5 mol%. The conversion of CO
(X
CO) is calculated from the molar flow rates of CO (FCO) into
and out of the reactor:
53,54
hydrolysis of metal nitrate hydrate to metal hydroxynitrate.
in
CO
out
There are no observed reflections of either CuO or NiO in
the XRD patterns during the reduction process. The main
reduction occurs between 230 °C and 240 °C. Metallic Cu and
Ni have two main reflections, belonging to the [111] and the [002]
crystal planes, at 43.3° and 50.4° for Cu and at 44.5° and 51.8°
F
 F
CO
XCO

100%
in
F
CO
The selectivity to a given product (S ) is based on the total
number of carbon atoms in the characterized products:
i
380 | Catal. Sci. Technol., 2014, 4, 378–386
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
Table 1 ICP-OES elemental analysis and BET surface area measurement results of fresh samples
a
Analyzed loading (wt%)
Metal loading in reduced catalyst
Cu + Ni (wt%)
Cu/Ni (atomic ratio)
BET surface area
2
−1
Catalysts
Cu
Ni
Na
Nominal
Measured
m g
Cu–Ni/SiO
Cu–Ni/SiO
Cu–Ni/SiO
2
2
2
–IWI
–Copr
–Dep–Copr
9.17
13.0
12.8
4.1
5.30
4.87
—
0.027
0.205
23.6
22.0
20.8
2.00
2.00
2.00
2.07
2.26
2.12
116
322
253
WCu WNi
a
The metal loading in the reduced sample is calculated from the equation WCu+Ni

100% on the basis of the analyzed
WCu WNi WSiO2
contents of Cu, Ni and Si.
Fig. 2 TGA analysis of the Cu–Ni/SiO
2
–IWI sample in a flow (120 ml
−1
−1
min ) of N
2
and 4.2 mol% H
2
/N
2
respectively (heating rate 2 °C min ).
Fig. 1 XRD patterns at selected temperatures during in situ reduction
of the impregnated (im), deposition–coprecipitated (dc) and coprecipitated
−1
(
2 2
cp) Cu–Ni/SiO samples in a flow (100 ml min ) of 2 mol% H /He. Phase
peak which begins at 210 °C and is centered at 250 °C. The
designations: ▲ – Cu/Ni hydroxynitrate, ● – metallic Cu, and ■ – Cu–Ni alloy.
calcination temperatures for Cu and Ni nitrates are reported to
55–57
be between 210 and 280 °C.
The first peak, centered at
for Ni (ICSD file 52265), respectively. In Fig. 1, the first
reflection is located at 43.3° with a small shoulder at 43.7°
and the second reflection is located at 50.4° with a shoulder
at 50.7°. The reflections at 43.3° and 50.4° are assigned to a
metallic Cu phase, and the shoulders located in between the
metallic phases of Cu and Ni are assigned to Cu–Ni alloys.
Based on ICP-OES analysis, the Cu/Ni molar ratio is 2.07. The
Cu–Ni alloy phase composition calculated using Vegard's law
is Cu/Ni = 2.0. From these analyses, about 4 mol% of isolated
metallic Cu is estimated to exist in the reduced catalysts. It
seems that the majority of Cu is alloyed with Ni, while a frac-
tion of Cu is segregated from the alloy phase. It was reported
that the standard reduction potential for Cu is 0.337 eV, which
250 °C, can therefore be ascribed to mass loss resulting from
nitrate decomposition during calcination. The second peak cen-
tered around 260 °C in reducing gas flow is probably related to
the beginning reduction of CuO, which is more easily reduced,
and the reduced Cu then acts as a catalyst and enhances the
reduction of Ni, whereby a double peak structure is seen in the
mass loss curve. This might further indicate that Cu and Ni
are not reduced simultaneously and that may be a part of the
reason for the observed phase segregation.
The diffraction patterns marked “cp” in Fig. 1 show the
in situ XRD scans on the Cu–Ni/SiO –Copr sample during the
2
reduction process. No reflections of either CuO or NiO are
observed in the XRD patterns in the as prepared oxide.
Hence, these oxides are finely dispersed and in an X-ray invis-
ible, amorphous form. The Cu–Ni alloy starts to emerge at
220 °C. Two symmetric reflections located at ~43.4° and at
~50.5° are observed after reduction. Interestingly, it seems
that only the Cu–Ni alloy is formed in this case. According to
15,28
is higher than that of Ni (−0.25 eV).
Cu is therefore more
easily reduced than Ni, and reduced Cu can subsequently act
as a catalyst to shift the reduction of Ni to lower temperature
whereby the Cu–Ni alloys are formed. According to the Scherrer
48
equation, the average particle sizes are about 11 nm for Cu–Ni
alloys and about 54 nm for Cu. In order to understand the
reduction and calcination processes of nitrate precursors, TGA
4
8
the Scherrer equation, the particle size is about 4 nm for
the Cu–Ni alloy phase. A very similar reduction process is
analyses were performed in inert gas (N
gas (4.2 mol% H /N ) flow. Fig. 2 shows the mass loss rate as a
function of temperature in the two different gas atmospheres.
In the flow of 4.2 mol% H /N , there are three major mass loss
peaks. The first peak starts at 210 °C and is centered at 250 °C,
while the second and third peaks are centered at around 260
2
) flow and reducing
observed for the Cu–Ni/SiO –Dep–Copr sample in the in situ
2
2
2
XRD studies (patterns marked “dc” in Fig. 1). Two symmetric
reflections located at ~43.5° and at ~50.6° are observed after
reduction, indicating the formation of a Cu–Ni alloy without
the formation of a separate Cu phase. The particle size of the
Cu–Ni alloy particles is estimated from XRD peak broadening
to be 5 nm.
2
2
2
and 265 °C, respectively. In the flow of N , there is one major
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 378–386 | 381

View Article Online
Paper
Catalysis Science & Technology
Long scan (640 minutes) XRD patterns at room temperature
do form bimetallic alloys when the particles are very small
31
after reduction of silica supported Cu–Ni samples are shown
in Fig. 3. The XRD particle size estimates of Cu–Ni alloys
(on the order of 1.5 nm or smaller). It has been concluded
that the behavior of small bimetallic clusters may be very dif-
4
8
31
(
based on the Scherrer equation ), the fraction of metallic Cu
ferent from that of substantially larger clusters. Using Monte
30
in the reduced catalysts (for the impregnated catalyst) and the
metal surface areas of all reduced samples are calculated and
summarized in Table 2. The particle sizes of Cu–Ni alloys are
nm and 5 nm for catalysts prepared by coprecipitation and
deposition–coprecipitation methods, respectively – significantly
smaller than the particle size of the catalyst prepared via the
Carlo simulations, Christensen et al. suggested that this
would also be the case for the Cu–Ni system, and the simula-
tions indicated that particles up to a size of 1.9 nm would be
alloyed even at a temperature of absolute zero. This is because
the energy cost in interface free energy upon formation of
segregated particles counteracts the gain in free energy upon
4
3
0
impregnation method on SiO (11 nm).
phase segregation. Depending on the particle size and the
temperature, the supported particles may therefore be alloyed,
even at temperatures where the bulk system is immiscible,
and this can help to explain why the smaller particles in the
precipitated samples appear to be fully alloyed, while the larger
particles in the impregnated sample exhibit a segregation into
Cu and Cu–Ni alloy.
2
With respect to bulk Cu–Ni alloy formation, the phase dia-
gram for the binary Cu–Ni bulk system shows phase segrega-
2
4,25
tion for a broad range of mixing ratios at room temperature
and formation of a solid solution only above a certain critical
temperature typically found in the 200–300 °C range depending
2
4–27
on the alloy composition.
In the case of Cu rich composi-
tions, we have found in this study, as well as in a previous
3.1.3 TEM studies of the supported Cu–Ni catalyst precursors.
Fig. 4 shows TEM images of the catalyst precursors prepared
by the three different preparation methods as well as the
corresponding particle size distributions. A clear difference in
structure among these three catalyst precursors can be observed.
2
9
one, that impregnated Cu–Ni samples form segregated phases
of metallic Cu and Cu–Ni alloy. However, the present results
for precipitated samples show that supported Cu–Ni in fact
can form homogeneous alloys even when being rich in Cu.
We have previously found that when the Cu–Ni alloys are rich
in Cu, the particle size is larger than that for Ni-rich Cu–Ni
alloys that show little or no segregation, and the particle size
2
For the impregnated catalyst (Cu–Ni/SiO –IWI), the particle size
distribution is broad as shown in Fig. 4a and d. The observed
particles span the range of 3 to 15 nm, and the average particle
size is 8.5 nm. It was previously reported for silica-supported
29
may thus also have an impact on the miscibility. Some
bimetallic systems such as Cu–Ru and Cu–Re are completely
metal nitrates that the drying step is of vital importance for
5
8,59
53,54
immiscible in the bulk.
However, Cu–Ru and Cu–Re
the particle size distribution.
Drying at elevated temperature
(
90 °C) resulted in agglomeration of metal hydroxynitrate
53,54
particles.
This could help to explain the rather broad par-
ticle size distribution observed for the impregnated catalyst in
the present study. It has previously been reported that drying
at ambient temperatures could prevent the formation of large
metal hydroxynitrate crystals, but the subsequent calcination
60
step still resulted in broad particle size distributions.
For both the deposition–coprecipitation prepared catalyst
(
Cu–Ni/SiO –Dep–Copr) and the catalyst prepared using
2
2
coprecipitation (Cu–Ni/SiO –Copr), the particle sizes are dis-
tributed in a narrow range, and most of the particles are in
the size range between 3 and 5 nm. The average particle sizes
are 4.3 nm and 4.1 nm for the Cu–Ni/SiO
2
–Dep–Copr and
Cu–Ni/SiO –Copr catalysts, respectively. A comparison of the
2
size of the metal precursor particles from TEM (Fig. 4) to
the metal particle size determined by in situ XRD (Table 2)
shows that the particle sizes of the metal precursors in the
Fig. 3 Long scan (640 minutes) XRD patterns at room temperature
after reduction of silica supported Cu–Ni samples. ● – metallic Cu and
■
– Cu–Ni alloys.
Table 2 XRD particle size estimates for Cu–Ni alloys, the fraction of metallic Cu in the reduced catalysts and estimated metal surface areas of all
reduced samples
a
b
c
cat−1
d
2
cat−1
Catalysts
d
p,XRD [nm]
Fraction of metallic Cu [mol%]
Mass of metal [g g
]
S
metal [m g ]
CuNi/SiO
CuNi/SiO
CuNi/SiO
2
2
2
–IWI
–C
–DC
11
4
5
4
0
0
17.1
17.7
18.3
10.4
29.4
24.4
a
b
c
XRD particle size estimate for Cu–Ni alloys. Calculated from Vegard's law and ICP-OES results. Active metal is metallic Cu + Ni, the weight
d
of metal material per gram of catalyst precursor, which is determined using ICP-OES analysis. Metal surface area calculated from the alloy
particle size and the metal content.
382 | Catal. Sci. Technol., 2014, 4, 378–386
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
is 2 to 3 times higher for the coprecipitated and deposition–
−
1
−1
coprecipitated catalysts (0.52 and 0.66 kg kgcat h , respec-
−
1
−1
tively) than for the impregnated catalyst (0.19 kg kgcat h ).
This is most likely due to the differences in the metal particle
sizes (see Table 2). The high selectivity towards methanol over
Cu–Ni catalysts is in agreement with our previous results on
43
20,23,41,42,44,61
CuNi (1/1 molar ratio) alloys and with other studies.
However, a significant amount (53–82 wt%) of Ni has been
lost after the CO hydrogenation test (Table 4), which is
ascribed to loss by volatile nickel carbonyl formation similar
4
5–47
to what is known from pure Ni.
The loss of Ni means that
there is a variable Ni-content in the catalyst during the test. It
is interesting to note that the selectivity and activity in MeOH
synthesis have been very stable during even prolonged tests.
Fig. 5 shows the development of space time yield (STY) and
selectivity to methanol as functions of time on stream for the
Cu–Ni/SiO –Dep–Copr catalyst. The methanol STY increases
2
−
1
−1
from 0.46 to 0.66 kg kgcat
h
during the first 10 h and then
stabilizes. In our previous paper, we have found a similar acti-
vation process over the silica supported CuNi catalysts as
4
3
well. The activation process could be related to introduction
of syngas (flushing the reactor volume), surface segregation
effects and Ni loss from Cu–Ni alloys. The catalyst precursor
is first reduced in H
2 2
/N gas flow, and then exposed to the
more reducing CO and H mixture. After reduction in H /N ,
2
2
2
Cu presumably occupies the first surface layer due to its lower
surface energy, whereas under reaction conditions, Ni may
2
Fig. 4 TEM images of (a) Cu–Ni/SiO –IWI (Cu and Ni present as the
nitrates), (b) Cu–Ni/SiO
and (c) Cu–Ni/SiO –Copr (Cu and Ni present as the oxides) and particle
size distributions of (d) Cu–Ni/SiO –IWI, (e) Cu–Ni/SiO –Dep–Copr, and
f) Cu–Ni/SiO –Copr precursor samples.
2
–Dep–Copr (Cu and Ni present as the oxides),
move to the surface due to the stronger bonding of CO to Ni
43,62,63
2
than to Cu.
In our previous work, we proposed that Ni
2
2
is pulled out to the surface of the alloy particles and forms
(
2
4
3
a bimetallic Cu–Ni surface. However, the present results
suggest that Ni not only segregates to the surface but is
even being leached from the surface of the particles very
quickly due to the high CO partial pressure. After a short initial
phase, diffusion of Ni to the surface presumably becomes rate
controlling for the Ni removal. We therefore in the end effec-
tively operate with an essentially pure Cu surface for CO hydro-
genation. It is of interest to note that other researchers have
tested Cu–Ni alloys in CO hydrogenation at high CO partial
as-prepared samples are very similar to the corresponding
reduced alloy particles. Thus, there does not appear to be dis-
cernible sintering during the reduction process, except for the
impregnated sample where the reduction leads to large isolated
Cu particles (54 nm) in addition to the alloy particles.
3
.2 CO hydrogenation
44
pressure conditions. For example, Fraga et al. tested CO
The steady state performance in CO hydrogenation of silica
supported Cu–Ni catalysts prepared by different methods is
shown in Table 3. The major product is methanol with a
selectivity above 99 mol% in all cases, while the rest of the
product consists of a small amount of oxygenates (ethanol
and/or DME) and methane. The space time yield of methanol
2 3
hydrogenation over Cu–Ni/Al O catalysts and noted a change
in product yields from hydrocarbons and higher alcohols
towards MeOH over about 10 hours of operation. The reason
was not explained by the authors, but may be rationalized in
the light of our observations of changes in the alloy composition.
When the catalyst contains Ni, it produces hydrocarbons and
Table 3 Results for CO hydrogenation at steady state over silica supported Cu–Ni catalysts prepared by different methods. Experimental
−1
conditions: P = 100 bar, T = 275 °C, GHSV = 4000 h and H
2
/CO = 2.0 vol/vol
a
b
Sample
TOS
h]
42
28
26
X
CO
Carbon based selectivity (CO
2
-free) [mol%]
STYMeOH
cat−
1
−1
[
[%]
8.5
14.3
16.6
MeOH
99.7
99.1
Ethanol
0.2
0.2
Methane
0.1
0.6
DME
0
0.1
0.1
[kg kg
h ]
Cu–Ni/SiO
Cu–Ni/SiO
Cu–Ni/SiO
2
2
2
–IWI
–Copr
–Dep–Copr
0.19
0.52
0.66
99.2
0.3
0.2
a
b
TOS: time on stream. CO conversion.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 378–386 | 383

View Article Online
Paper
Catalysis Science & Technology
Table 4 ICP-OES analysis of both fresh and spent Cu–Ni catalysts
Fresh catalyst [wt%]
Spent catalyst [wt%]
Ni
loss
[%]
a
c
TOS
[h]
b
b
Sample
Cu
9.17
Ni
Si
Ni loading
Cu
Ni
Si
Ni loading
Cu–Ni/SiO
Cu–Ni/SiO
Cu–Ni/SiO
2
2
2
–IWI
–Copr
–Dep–Copr
166
28
26
4.1
5.30
4.87
20
30.4
34.4
7.3
6.4
5.3
11.0
12.4
13.0
1.26
1.95
2.20
35.5
33.8
33.5
1.4
2.2
2.5
82
65
53
13.0
12.8
fresh
WNi
a
b
100% . c Ni loss is
TOS: time on stream. The Ni loading in the reduced catalyst is calculated from the equation WNi

spent WCu WNi WSiO2
Ni
Si
Ni
(
fresh)  (spent)
Si
calculated from the equation Ni loss 
100% .
Ni
(
fresh)
Si
not only often used as an alloying component with Cu for high
pressure CO hydrogenation, but also used as a promoter for high
pressure CO hydrogenation to higher alcohols in other catalyst
6
4–69
systems.
The present results show that loss of Ni by carbonyl
formation is not only a problem for pure Ni catalysts, but also
something that must be considered when Ni-containing alloy
catalysts are used under conditions with elevated partial pres-
sure of CO.
4
Conclusion
Fig. 5 Space time yield and selectivity of methanol as functions of
time on stream for the Cu–Ni/SiO –Dep–Copr catalyst. The operating
conditions are: P = 100 bar, T = 275 °C, GHSV = 4000 h (based on
2
Silica supported, bimetallic Cu–Ni (20 wt% Cu + Ni on silica,
Cu/Ni = 2, molar ratio) alloy catalysts have been prepared via
incipient wetness impregnation, coprecipitation, and deposition–
coprecipitation methods and tested for CO hydrogenation.
−1
total bed volume and gas flow referring to 298.15 K and 1 bar), feed:
2
H /CO = 2 (vol/vol).
2 2 3
The approach to co-precipitate the SiO from Na SiO together
higher alcohols, but as the composition shifts towards pure
Cu, the catalyst changes to produce mainly methanol. A very
with metal precursors seems to be an efficient way to prepare
high surface area silica supported catalysts with small Cu–Ni
23
2
−1
recent study by Xiao et al. made similar observations without
mentioning that Ni was lost from the catalyst. Their XRD data
on CuNi3 catalysts (Cu/Ni = 1/3, molar ratio) clearly showed
that the freshly reduced catalyst consisted of a Cu–Ni alloy,
but after the CO hydrogenation test, the XRD reflections had
particles (BET surface area up to 322 m g , and metal area
2
−1
calculated from XRD particle size up to 29 m g ). TEM studies
on as-prepared catalyst precursors show that both the deposi-
tion–coprecipitation and coprecipitation methods yield small,
homogeneous particles in the size range between 3 and 5 nm,
while catalyst preparation by impregnation yields larger particles
(8.5 nm). The particle sizes of the metal precursors in the as-
prepared samples are similar to the corresponding reduced alloy
particles, and there does thus not appear to be discernible
sintering during the reduction process. An exception is the
impregnated sample where the reduction leads to large isolated
Cu particles (54 nm) in addition to Cu–Ni alloy particles. Moni-
toring of the crystal structure using XRD during in situ reduction
shows that coprecipitation and deposition–coprecipitation
methods are more efficient than the impregnation method for
preparation of small and especially homogeneous Cu–Ni alloy
nanoparticles. The small size of the metal particles in the precip-
itated samples is most likely the reason why homogeneous alloys
are formed.
2
3
shifted towards the position for metallic Cu.
For pure metallic Ni, it is well established that the thermo-
dynamics may favor formation of volatile Ni(CO) under some
4
conditions. For example, when the partial pressure of CO is
higher than 0.2 bar and the temperature is lower than 425 °C,
carbonyl formation will take place and removal of nickel from
4
6
the catalytic reactor occurs. According to thermodynamic
4
6
calculations, Shen et al. reported that conditions for which
−
11
the equilibrium Ni(CO) pressure is less than ca. 1 × 10 bar
4
result in stable methanation activity. Conversely, much higher
partial pressures of Ni(CO)
4
will lead to the transport of nickel
through the catalyst bed and ultimately to the removal of
4
6
nickel from the reactor. It is therefore not surprising that
Ni was lost in our high CO pressure tests on Cu–Ni catalysts
as well as in the mentioned examples from the literature. For
example, at a starting CO pressure of 50 bar, the equilibrium
The supported Cu–Ni catalysts were tested in high pres-
sure CO hydrogenation. The selectivity towards methanol was
always higher than 99 mol%; however, analyses of spent cata-
lysts show that serious Ni loss by carbonyl formation occurs
for the Cu–Ni alloy catalysts during high pressure CO
4
Ni(CO) pressure is very high (about 1 to 5 bar in the 250–300 °C
range) – significantly higher than the upper limit for stable
46
operation given by Shen et al. It should be noticed that Ni is
384 | Catal. Sci. Technol., 2014, 4, 378–386
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
hydrogenation. Due to the strong bonding of CO to Ni, it
seems quite possible that loss of Ni by carbonyl formation
will also take place for other Ni containing alloys if these are
used under conditions with high partial pressures of CO. Our
observations can help to explain observations of transient
behavior in the selectivity during CO hydrogenation previ-
ously left unexplained in the literature.
19 E. Vesselli, E. Monachino, M. Rizzi, S. Furlan, X. Duan,
C. Dri, A. Peronio, C. Africh, P. Lacovig and A. Baldereschi,
ACS Catal., 2013, 3, 1555–1559.
20 J. Nerlov and I. Chorkendorff, J. Catal., 1999, 181, 271–279.
21 Y. Wang, S. Qi and Y. Hou, React. Kinet. Catal. Lett., 2000,
70, 213–217.
22 P. Sermon and Y. Wang, Stud. Surf. Sci. Catal., 1993,
7
5, 2773–2776.
2
3 K. Xiao, X. Qi, Z. Bao, X. Wang, L. Zhong, K. Fang, M. Lin
Acknowledgements
and Y. Sun, Catal. Sci. Technol., 2013, 3, 1591–1602.
The experimental work was financed by the Danish Ministry for
Science, Technology and Development under the “Catalysis
for Sustainable Energy” initiative, the Technical University of
Denmark, and Haldor Topsøe A/S. We thank Mr. Helge
Kildahl Rasmussen from DTU Physics for aid in the XRD mea-
surement and Haldor Topsøe A/S for the ICP-OES analyses.
24 P. Van der Plank and W. Sachtler, J. Catal., 1968, 12, 35–44.
25 W. H. Sachtler and P. Van der Plank, Surf. Sci., 1969, 18, 62–79.
26 Binary alloy phase diagrams, ed. T. B. Massalski, H. Okamoto,
P. R. Subramanian and L. Kacprzak, ASM International, 1990.
27 Desk handbook: phase diagrams for binary alloys, ed. H. Okamoto,
ASM International, 2000.
2
2
8 A. Naghash, T. Etsell and S. Xu, Chem. Mater., 2006, 18, 2480–2488.
9 Q. Wu, L. D. L. Duchstein, G. L. Chiarello, J. M. Christensen,
C. D. Damsgaard, C. F. Elkjær, J. B. Wagner, B. Temel,
J. D. Grunwaldt and A. D. Jensen, ChemCatChem, 2013, DOI:
10.1002/cctc.201300628.
References
1
2
L. C. Chen and S. D. Lin, Appl. Catal., B, 2011, 106, 639–649.
J. Calles, A. Carrero and A. Vizcaíno, Microporous Mesoporous
Mater., 2009, 119, 200–207.
30 A. Christensen, P. Stoltze and J. Nørskov, J. Phys.: Condens.
Matter, 1995, 7, 1047.
31 J. H. Sinfelt, Acc. Chem. Res., 1987, 20, 134–139.
32 E. Guerreiro, O. Gorriz, G. Larsen and L. Arrúa, Appl. Catal., A,
2000, 204, 33–48.
33 J. Van de Loosdrecht, S. Barradas, E. Caricato, N. Ngwenya,
P. Nkwanyana, M. Rawat, B. Sigwebela, P. Van Berge and
J. Visagie, Top. Catal., 2003, 26, 121–127.
3
4
5
F. Wang, Y. Li, W. Cai, E. Zhan, X. Mu and W. Shen, Catal.
Today, 2009, 146, 31–36.
V. Bergamaschi and F. Carvalho, Mater. Sci. Forum, 2008,
5
91, 734–739.
R. Pérez-Hernández, G. Mondragón Galicia, D. Mendoza
Anaya, J. Palacios, C. Angeles-Chavez and J. Arenas-Alatorre,
Int. J. Hydrogen Energy, 2008, 33, 4569–4576.
6
7
8
9
A. Vizcaíno, A. Carrero and J. Calles, Int. J. Hydrogen Energy,
34 A. J. Van Dillen, R. J. A. M. Terörde, D. J. Lensveld,
J. W. Geus and K. P. De Jong, J. Catal., 2003, 216, 257–264.
35 J. R. A. Sietsma, J. D. Meeldijk, M. Versluijs-Helder,
A. Broersma, A. J. Dillen, P. E. de Jongh and K. P. de Jong,
Chem. Mater., 2008, 20, 2921–2931.
36 R. Herman, K. Klier, G. Simmons, B. Finn, J. B. Bulko and
T. Kobylinski, J. Catal., 1979, 56, 407–429.
37 P. Porta, S. De Rossi, G. Ferraris, M. Lo Jacono, G. Minelli
and G. Moretti, J. Catal., 1988, 109, 367–377.
38 Z. Huang, F. Cui, H. Kang, J. Chen, X. Zhang and C. Xia,
Chem. Mater., 2008, 20, 5090–5099.
39 S. Uchiyama, Y. Ohbayashi, T. Hayasaka and N. Kawata,
Appl. Catal., 1988, 42, 143–152.
2
007, 32, 1450–1461.
J. Bian, M. Xiao, S. Wang, Y. Lu and Y. Meng, Catal. Commun.,
009, 10, 1142–1145.
X. Wu, Y. Meng, M. Xiao and Y. Lu, J. Mol. Catal. A: Chem.,
006, 249, 93–97.
2
2
J. Bian, M. Xiao, S. J. Wang, Y. X. Lu and Y. Z. Meng, Appl.
Surf. Sci., 2009, 255, 7188–7196.
1
1
0 A. Ardiyanti, S. Khromova, R. Venderbosch, V. Yakovlev and
H. Heeres, Appl. Catal., B, 2012, 117, 105–117.
1 V. Yakovlev, S. Khromova, O. Sherstyuk, V. Dundich,
D. Y. Ermakov, V. Novopashina, M. Y. Lebedev, O. Bulavchenko
and V. Parmon, Catal. Today, 2009, 144, 362–366.
1
2 P. M. Mortensen, J. D. Grunwaldt, P. A. Jensen, K. Knudsen
and A. D. Jensen, Appl. Catal., A, 2011, 407, 1–19.
40 S. Uchiyama, Y. Obayashi, M. Shibata, T. Uchiyama, N. Kawata
and T. Konishi, J. Chem. Soc., Chem. Commun., 1985, 1071–1072.
41 E. B. Pereira and G. A. Martin, Appl. Catal., A, 1993,
103, 291–309.
42 J. Llorca, P. Ramirez de la Piscina, N. Homs, E. Brum Pereira,
P. Moral and G. Martin, Stud. Surf. Sci. Catal., 1997,
107, 9–14.
43 F. Studt, F. Abild-Pedersen, Q. Wu, A. D. Jensen, B. Temel,
J. D. Grunwaldt and J. K. Nørskov, J. Catal., 2012, 293, 51–60.
44 M. A. Fraga and E. Jordăo, React. Kinet. Catal. Lett., 1998,
64, 331–336.
45 J. Engbæk, O. Lytken, J. H. Nielsen and I. Chorkendorff,
Surf. Sci., 2008, 602, 733–743.
1
1
3 J. H. Lin and V. V. Guliants, ChemCatChem, 2012, 4, 1611–1621.
4 N. Schumacher, A. Boisen, S. Dahl, A. A. Gokhale, S. Kandoi,
L. C. Grabow, J. A. Dumesic, M. Mavrikakis and I. Chorkendorff,
J. Catal., 2005, 229, 265–275.
1
1
1
1
5 J. H. Lin, P. Biswas, V. V. Guliants and S. Misture, Appl.
Catal., A, 2010, 387, 87–94.
6 T. J. Huang, T. C. Yu and S. Y. Jhao, Ind. Eng. Chem. Res.,
2
006, 45, 150–156.
7 L. Y. Gan, R. Tian, X. B. Yang, D. H. Lu and Y. Zhao, J. Phys.
Chem. C, 2012, 116, 745–752.
8 L. E. Cratty Jr and W. W. Russell, J. Am. Chem. Soc., 1958,
8
0, 767–773.
46 W. Shen, J. Dumesic and C. Hill, J. Catal., 1981, 68, 152–165.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 378–386 | 385

View Article Online
Paper
Catalysis Science & Technology
4
7 C. Mirodatos, H. Praliaud and M. Primet, J. Catal., 1987,
07, 275–287.
58 J. Sinfelt, J. Catal., 1973, 29, 308–315.
59 G. Meitzner, G. Via, F. Lytle and J. Sinfelt, J. Chem. Phys.,
1985, 83, 353.
60 P. Munnik, M. Wolters, A. Gabrielsson, S. Pollington,
G. Headdock, J. Bitter, P. de Jongh and K. de Jong, J. Phys.
Chem. C, 2011, 115, 14698–14706.
61 J. Nerlov and I. Chorkendorff, Catal. Lett., 1998, 54, 171–176.
62 G. Tréglia, B. Legrand and P. Maugain, Surf. Sci., 1990,
225, 319–330.
1
4
4
8 P. Scherrer, Nachr. Ges. Wiss. Göttingen, 1918, 26, 98–100.
9 J. M. Christensen, P. M. Mortensen, R. Trane, P. A. Jensen
and A. D. Jensen, Appl. Catal., A, 2009, 366, 29–43.
5
0 J. M. Christensen, L. D. L. Duchstein, J. B. Wagner,
P. A. Jensen, B. Temel and A. D. Jensen, Ind. Eng. Chem. Res.,
2
012, 51, 4161–4172.
5
1 Q. Wu, J. M. Christensen, G. L. Chiarello, B. Temel,
J. D. Grunwaldt and A. D. Jensen, Catal. Today, 2013,
63 K. Wandelt and C. R. Brundle, Phys. Rev. Lett., 1981, 46, 1529–1532.
64 H. Qi, D. Li, C. Yang, Y. Ma, W. Li, Y. Sun and B. Zhong,
Catal. Commun., 2003, 4, 339–342.
65 T. Tatsumi, A. Muramatsu, T. Fukunaga and H. Tominaga,
in Proc. 9th Int. Cong. Catal., ed. M. J. Phillips and
M. Ternan, 1988, vol. 2, pp. 618–622.
2
15, 162–168.
5
5
2 J. M. Christensen, P. A. Jensen and A. Jensen, Ind. Eng.
Chem. Res., 2011, 50, 7949–7963.
3 S. Catillon-Mucherie, F. Ammari, J. M. Krafft, H. Lauron-Pernot,
R. Touroude and C. Louis, J. Phys. Chem. C, 2007, 111,
1
1619–11626.
66 M. Xiang, D. Li, H. Xiao, J. Zhang, W. Li, B. Zhong and
Y. Sun, Catal. Today, 2008, 131, 489–495.
5
5
5
5
4 S. Catillon-Mucherie, H. Lauron-Pernot and C. Louis, J. Phys.
Chem. C, 2010, 114, 11140–11147.
5 S. Yuvaraj, L. Fan-Yuan, C. Tsong-Huei and Y. Chuin-Tih,
J. Phys. Chem. B, 2003, 107, 1044–1047.
6 J. G. Jackson, R. W. Fonseca and J. A. Holcombe, Spectrochim.
Acta, Part B, 1995, 50, 1449–1457.
7 J. Mu and D. Perlmutter, Thermochim. Acta, 1982, 56, 253–260.
67 N. Zhao, R. Xu, W. Wei and Y. Sun, React. Kinet. Catal. Lett.,
2002, 75, 297–304.
68 E. C. Alyea, D. He and J. Wang, Appl. Catal., A, 1993,
104, 77–85.
69 C. Ma, H. Li, G. Lin and H. Zhang, Appl. Catal., B, 2010,
100, 245–253.
386 | Catal. Sci. Technol., 2014, 4, 378–386
This journal is © The Royal Society of Chemistry 2014