Sol-gel synthesis of Ni and Ni supported catalysts for hydrogen production by methane decomposition

Article abstract of DOI:10.1039/c4ra04379k

NiO, NiO-CuO, NiO-TiO₂and NiO-CeO₂were synthesized by the sol-gel method and further processed into catalysts for hydrogen production by the thermocatalytic decomposition of methane. The structures of the metal oxides and catalysts were extensively characterized by transmission electron microscopy, X-ray diffraction, energy dispersive X-ray spectroscopy, X-ray photoelectron spectroscopy, temperature-programmed reduction and N₂adsorption. The catalytic performance of the catalysts was evaluated in terms of methane conversion and carbon yield. NiO had a BET surface area of 66 m²g^-1. When compared with NiO, NiO-CuO had a lower BET surface area of 13 m²g^-1, while NiO-TiO₂and NiO-CeO₂had higher BET surface areas of 161 and 133 m²g^-1, respectively. The resulting supported catalysts required increased reduction temperatures and had higher Ni 2p_3/2binding energies, indicating strong metal-support interactions. For the Ni-Cu catalyst, the introduction of copper resulted in a lower Ni 2p_3/2binding energy due to the charge transfer from Cu to the adjacent Ni atoms. The Ni catalyst showed a stable methane conversion of 0.39 at 550 °C. The presence of copper in the Ni-Cu catalyst increased the working temperature but decreased methane conversion. The Ni-TiO₂and Ni-CeO₂catalysts exhibited stable methane conversions of 0.56 and 0.60, respectively, at 575 °C. The lifetime tests of the catalysts revealed that the Ni catalyst had a life span of 600 min at 550 °C, resulting in a carbon yield of 121 g C per g Ni. The Ni-Cu, Ni-CeO₂and Ni-TiO₂catalysts had life spans of 670, 790 and 840 min at 575 °C, resulting in carbon yields of 171, 204 and 256 g C per g Ni, respectively. The structures of the resulting carbon products were characterized by transmission electron microscopy and X-ray diffraction. It was found that the morphologies of the resulting carbon products were dependent on the type of catalyst used.

Full text of DOI:10.1039/c4ra04379k

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
RSC Advances
View Article Online
View Journal | View Issue
PAPER
Sol–gel synthesis of Ni and Ni supported catalysts
for hydrogen production by methane
decomposition
Cite this: RSC Adv., 2014, 4, 42159
ab
b
*
*
and Aik Chong Lua
Yi Shen
NiO, NiO–CuO, NiO–TiO2 and NiO–CeO2 were synthesized by the sol–gel method and further processed
into catalysts for hydrogen production by the thermocatalytic decomposition of methane. The structures of
the metal oxides and catalysts were extensively characterized by transmission electron microscopy, X-ray
diffraction, energy dispersive X-ray spectroscopy, X-ray photoelectron spectroscopy, temperature-
programmed reduction and N2 adsorption. The catalytic performance of the catalysts was evaluated in
terms of methane conversion and carbon yield. NiO had a BET surface area of 66 m2
compared with NiO, NiO–CuO had a lower BET surface area of 13 m2 gꢀ1, while NiO–TiO2 and NiO–CeO2
had higher BET surface areas of 161 and 133 m2 ꢀ1, respectively. The resulting supported catalysts
g
ꢀ1. When
g
required increased reduction temperatures and had higher Ni 2p3/2 binding energies, indicating strong
metal–support interactions. For the Ni–Cu catalyst, the introduction of copper resulted in a lower Ni
2p3/2 binding energy due to the charge transfer from Cu to the adjacent Ni atoms. The Ni catalyst
showed a stable methane conversion of 0.39 at 550 ꢁC. The presence of copper in the Ni–Cu catalyst
increased the working temperature but decreased methane conversion. The Ni–TiO2 and Ni–CeO2
catalysts exhibited stable methane conversions of 0.56 and 0.60, respectively, at 575 ꢁC. The lifetime
tests of the catalysts revealed that the Ni catalyst had a life span of 600 min at 550 ꢁC, resulting in a
carbon yield of 121 g C per g Ni. The Ni–Cu, Ni–CeO2 and Ni–TiO2 catalysts had life spans of 670, 790
and 840 min at 575 ꢁC, resulting in carbon yields of 171, 204 and 256 g C per g Ni, respectively. The
structures of the resulting carbon products were characterized by transmission electron microscopy and
X-ray diffraction. It was found that the morphologies of the resulting carbon products were dependent
on the type of catalyst used.
Received 11th May 2014
Accepted 18th August 2014
DOI: 10.1039/c4ra04379k
www.rsc.org/advances
reforming of natural gas using nickel as a catalyst has been the
most widely used process for the industrial production of
1. Introduction
hydrogen. However, the production of hydrogen by this multi-
step process results in high emissions of carbon dioxide, and
the end-product always contains some quantities of carbon
monoxide, preventing the resulting hydrogen from being
directly used in fuel cells. Alternatively, the thermocatalytic
decomposition of methane (TDM) has been of great interest
because only pure hydrogen and high quality carbon are
obtained in this process.3–6 COx-free hydrogen produced by
TDM can be used in fuel cell applications without further
separation and purication processes. In addition, carbon is
sequestered in the form of valuable solid carbon nanotubes or
nanobers rather than as a notorious greenhouse gas in the
form of carbon dioxide. The feasibility of the TDM process is
highly dependent on the efficiency and durability of the catalyst.
Nickel has proven to be the most efficient catalyst and has been
extensively studied for the TDM process.7–10 Considerable work
has focused on catalyst preparation and the selection of a
catalyst support. Impregnation,11–15 co-precipitation16–22 and the
fusion method23–26 have been widely used to prepare active
It is established that global warming is taking place due to
human activities, in particular the combustion of fossil fuels.
Due to its immense negative impacts on human communities
and ecosystems, global warming has been the most dangerous
environmental problem of the contemporary world and has
attracted great attention from the scientic and political
communities. To address this urgent problem, it is critical to
control the emissions of greenhouse gases and develop renew-
able energy sources. Hydrogen is considered to be one of the
most efficient and environmentally friendly energy carriers, and
considerable research has been carried out into eco-friendly
methods of hydrogen production.1,2 For decades, the steam
aCollege of Light Industry and Food Sciences, South China University of Technology,
Guangzhou, 510640, P. R. China. E-mail: feyshen@scut.edu.cn; Fax: +86
2087113843; Tel: +86 2087113843
bSchool of Mechanical and Aerospace Engineering, Nanyang Technological University,
50 Nanyang Avenue, 639798, Republic of Singapore. E-mail: maclua@ntu.edu.sg; Tel:
+65 67905535
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 42159–42167 | 42159

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
RSC Advances
Paper
catalysts. The sol–gel technique is a well-established method to loaded into a quartz tube reactor. A mixture of hydrogen and
prepare various catalysts with high porosities and uniform nitrogen with a volume ratio of 1 : 3 was introduced as the
dispersions of active components on supports.27,28 However, reducing agent at a ow rate of 20 mL minꢀ1. The temperature
limited work has been reported on the sol–gel syntheses of of the reactor was increased at a rate of 2.5 ꢁC minꢀ1. The
catalysts for the TDM process.29 The supported catalysts composition of the gas stream was continuously analysed. NiO,
obtained by the sol–gel technique have been reported to exhibit NiO–CuO, NiO–TiO2 and NiO–CeO2 were reduced to Ni, NiCu,
strong metal–support interactions, which could stabilize the Ni–TiO2 and Ni–CeO2, respectively.
active component at high reaction temperatures.30 In this work,
nickel oxide and nickel oxide supported on copper oxide, tita-
nium oxide and cerium oxide were synthesized by the sol–gel
method using a nontoxic and biodegradable ethoxylated sor-
bitan ester as the structure-directing agent. Aer calcination
and reduction, the resulting active nickel particles and sup-
ported nickel particles were investigated as catalysts for the
TDM process. The structures and catalytic performance of the
resulting catalysts were fully characterized. The effects of copper
promoter and supports on the catalytic activities of the catalysts
were also investigated.
2.4 Catalyst characterization
Field emission scanning electron microscopy (FESEM; JSM-
7600F, JEOL) and transmission electron microscopy (TEM;
JSM2010, JEOL) were used to observe the morphology of the
catalysts and the resulting carbon materials. To analyse the
crystalline structures of the catalysts and the resulting carbon
materials, X-ray diffraction (XRD) patterns were obtained by a
diffractometer (PW1830, Philips) equipped with Cu-Ka radia-
˚
tion of 1.54 A. The N2 adsorption–desorption isotherms were
obtained using the accelerated surface area and porosimetry
system (ASAP 2010, Micromeritics). The pore surface area was
calculated using the Brunauer–Emmett–Teller (BET) method;
the pore volume and pore size were calculated by the Barret–
Joyner–Halenda (BJH) method. Energy dispersive X-ray spec-
troscopy (EDX) and X-ray photoelectron spectroscopy (XPS) were
used to analyse the elemental composition and binding energy
of the samples, respectively. To characterize the reactive cata-
lysts aer hydrogen reduction, a passivation treatment was
conducted. Aer the nickel oxide was reduced in the quartz
tube, it was subsequently cooled to ambient temperature in a
nitrogen atmosphere. An O2/N2 (0.5%/99.5%) mixture stream
was then introduced to the catalysts for 4 h at room temperature
before the sample was removed from the reactor.
2. Experimental sections
2.1 Chemicals
For the synthesis of the catalysts, the following chemicals were
used: precursors – nickel(II) nitrate hexahydrate, copper(II)
nitrate trihydrate, cerium(III) nitrate hexahydrate, and titaniu-
m(IV) butoxide (Ti(OBu)4); acids – hydrochloric acid (HCl) and
acetic acid; surfactant – ethoxylated sorbitan ester (Tween 20);
and solvent – ethanol. These chemicals were used as received.
2.2 Synthesis of NiO and NiO on metal oxide supports
Nickel oxide and nickel oxide on metal oxide supports were
synthesized using a similar sol–gel method reported by Fan
et al.31 In a typical nickel oxide synthetic process, 4 mL of Tween
20, 40 mmol of acetic acid, 24 mmol of HCl and 30 mmol of
nickel(II) nitrate hexahydrate were dissolved in 60 mL of
ethanol. The mixture was stirred vigorously for 2 h at room
temperature and then aged at 60 ꢁC in an oven for 2 days. Aer
the complete evaporation of ethanol, the as-synthesized
composite was calcined at 500 ꢁC in air for 5 h. For the synthesis
of supported nickel oxides, nickel(II) nitrate hexahydrate and
the other metal salt with a mole ratio of 1 : 1 were used as
precursors. For instance, in the synthesis of nickel oxide sup-
ported on titanium oxide, 4 mL of Tween 20, 40 mmol of acetic
acid, 24 mmol of HCl, 15 mmol of nickel(II) nitrate hexahydrate
and 15 mmol of Ti(OBu)4 were dissolved in 60 mL of ethanol.
The synthetic procedures were identical to those of the nickel
oxide. For convenience, the nickel oxide, nickel oxide supported
on copper oxide, nickel oxide supported on titanium oxide and
nickel oxide supported on cerium oxide were denoted as NiO,
NiO–CuO, NiO–TiO2 and NiO–CeO2, respectively.
2.5 Characterization of catalytic activity
The catalytic activities of the catalysts were tested experimen-
tally.33,34 In a typical process, a predetermined amount of cata-
lyst was loaded into the mid-section of a quartz tube reactor
with an internal diameter of 10 mm and a length of 800 mm.
The reactor was surrounded and heated by an electric furnace
(Lenton) with a temperature-programmed controller (Carbonite
2416), and the temperature of the reactor was accurately
monitored using a K-type thermocouple. The ow rates of the
gases were measured and controlled by mass ow meters (Alicat
Scientic). Prior to the catalytic reaction, the system was ushed
with nitrogen to remove the air. The concentrations of the gases
at the reactor outlet were analysed by a calibrated gas chro-
matograph (Agilent 6890-5973 GC-MS network systems) equip-
ped with a thermal conductivity detector.
3. Results and discussion
3.1 Structures of the catalysts
2.3 Temperature-programmed reduction of nickel oxide
The elemental compositions of the as-synthesized metal oxides
The temperature-programmed reduction (TPR) of nickel oxide were determined by EDX; the results are shown in Table 1. All
was carried out using an experimental set-up equipped with a the fresh metal oxides contained a small quantity of chlorine,
temperature-programmed controller similar to that described e.g., the fresh NiO–CeO2 sample consisted of 9 mmol% chlorine
by Liotta et al.32 In a typical process, 0.25 g of metal oxide was atom (Table 1 and Fig. 1). The existence of chlorine atoms in the
42160 | RSC Adv., 2014, 4, 42159–42167
This journal is © The Royal Society of Chemistry 2014

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
Paper
RSC Advances
Table 1 Atomic composition, BET surface area and pore size of the metal oxidesc
b
˚
Sample
NiO
Elemental compositiona (mmol%)
BET surface area (m2 gꢀ1
)
Pore size (A)
Ni (47.8) : O (48.5) : Cl (3.7)*
Ni (47.2) : O (52.8)**
66
161
133
13
36
90
98
37
NiO–TiO2
NiO–CeO2
NiO–CuO
Ni (15.4) : Ti (16.3) : O (64.4) : Cl (3.9)*
Ni (13.7) : Ti (18.9) : O (67.4)**
Ni (14.5) : Ce (15.7) : O (60.8) : Cl (9)*
Ni (8.7) : Ce (20.9) : O (70.4)**
Ni (20.3) : Cu (24.4) : O (52.7) : Cl (2.6)*
Ni (21.4) : Cu (25.6) : O (53)**
a
Determined by EDX. For each sample, ve tests were carried out, and the average values are given. b Determined by 4V/A, where V and A are the
pore volume and BET surface area, respectively. c *Results for the samples before washing. ** Results for the samples aer thorough washing.
Fig. 2 N2 adsorption–desorption isotherms of the metal oxides.
Fig. 1 EDX curves of NiO–CeO2 before and after washing.
fresh metal oxides in the form of NiCl2 (as conrmed by XRD
but not shown here) was due to the incomplete hydrolysis of
Ni2+ during the sol–gel process. This phenomenon was also
reported by Fan et al.31 The presence of NiCl2 as an impurity in
the fresh metal oxides was detrimental to the subsequent
catalytic activity in the TDM process. Hence, NiCl2 was removed
by thoroughly washing the as-synthesized metal oxides with DI
water; the results are evidenced by the EDX results in Table 1
and Fig. 1. Aer washing, the NiO content was determined to be
40.5, 15.3 and 43.8% in the NiO–TiO2, NiO–CeO2 and NiO–CuO
catalysts, respectively. The nitrogen adsorption–desorption
isotherms of the metal oxides are shown in Fig. 2. The BET
surface areas and pore sizes are shown in Table 1. NiO had a BET
surface area of 66 m2 gꢀ1. When compared with NiO, NiO–CuO
had a lower BET surface area of 13 m2 gꢀ1, while NiO–TiO2 and
NiO–CeO2 had higher BET surface areas of 161 and 133 m2 gꢀ1
,
respectively.
Fig. 3 shows the TEM micrographs of the metal oxides. The
NiO particles showed cuboid-like morphologies with lengths of
50–120 nm, as shown in Fig. 3A. Closer examination revealed
that the surface of the NiO particle was rough, and some worm-
like pores of 1–2 nm were observed (inset in Fig. 3A). The
Fig. 3 TEM micrographs of (A) NiO, (B) NiO–CuO, (C) NiO–CeO2 and
(D) NiO–TiO2.
formation of these worm-like pores was attributed to the super- diameters from 40 to 100 nm. Some worm-like pores were also
molecular templating of Tween 20 surfactant. The NiO–CuO observed in the particles. In contrast, NiO–CeO2 and NiO–TiO2
particles exhibited irregular-spherical structures with varying had average particle sizes of 11 and 9 nm, respectively, which
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 42159–42167 | 42161

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
View Article Online
RSC Advances
Paper
were much smaller than those of NiO and NiO–CuO. This in the XRD patterns, conrming that the metal oxides were free
indicates that the presence of Ce4+ and Ti4+ had signicant of impurities aer the thorough washing.
effects on the hydrolysis of Ni2+ during the sol–gel process. The
Fig. 5 shows the hydrogen uptake of the metal oxides as a
hydrolysis constants (Kas) of metal ions are highly dependent on function of temperature. NiO showed a reduction peak at
the ionic charge and radius. The pKa values follow the order of approximately 377 ꢁC as began to be reduced at about 350–
Ni2+ (9.9) > Cu2+ (7.9) > Ce4+ (ꢀ1.1) > Ti 4+ (ꢀ4.0).35 Therefore, the 450 ꢁC. NiO–CuO metal oxides exhibited two distinguishable
hydrolysis of Ni2+ could be overwhelmed by those of Ce4+ and reduction peaks centred at about 235 and 345 C, correspond-
ꢁ
Ti4+ during the sol–gel process. In this study, the inhibition of ing to the reduction of Cu2+ and Ni2+, respectively. The Cu2+ can
the hydrolysis of Ni2+ could be further intensied due to the be reduced at lower temperatures because it has a higher
scarcity of water, which originated exclusively from the reaction standard reduction potential (Cu2+ / Cu0 +0.34 V) than Ni2+
of ethanol and acetic acid during the sol–gel process. No (Ni2+ / Ni0 ꢀ0.26 V). The presence of metallic Cu has been
distinct pores were observed on the surfaces of NiO–CeO2 and reported to enhance the reducibility of the Ni species due to the
NiO–TiO2 particles, indicating the failure of the templating fact that metallic copper could dissociate the hydrogen mole-
effect of the surfactant. This could be related to the branched cule into hydrogen atoms, which could spill over to the surface
structure of Tween 20 and the kinetics of the sol–gel process.
of the neighbouring NiO. This led to a lower NiO reduction
Fig. 4 shows the XRD patterns of the metal oxides. The XRD temperature in the NiO–CuO metal oxides when compared with
pattern of NiO showed diffraction peaks at 37.3, 43.3, 62.9, 75.4 that of NiO.34 The NiO–TiO2 metal oxides showed a broad
ꢁ
and 79.4ꢁ, which could be assigned to the rhombohedral phase reduction peak centred at 409 C, which was attributed to the
of NiO (JCPDS #44-1159). For NiO–CuO, separate NiO and CuO reduction of Ni species. A closer inspection of the TPR curve
diffractions were observed, indicating that NiO and CuO did not showed that this broad peak had a shoulder at lower tempera-
form a solid solution during the sol–gel process. The peaks in tures and a long tail at higher temperatures. This suggested that
the XRD patterns for NiO–CeO2 and NiO–TiO2 could be ascribed supported catalysts such as Ni/Al2O3 and Ni/MgO had higher
to the diffractions of NiO and the supports (CeO2 JCPDS: #1- reduction temperatures due to the interactions between NiO
0800; TiO2 JCPDS: #1-0562). The crystal sizes of NiO in the metal and the support.36 Correspondingly the shoulder at lower
oxides were calculated from the strongest (012) diffraction temperatures in the TPR curve of NiO–TiO2 could be ascribed to
peaks and are given in Table 2. No other phases were identied the reduction of the surface NiO particles that were weakly
bound to TiO2, while the long tail at higher temperatures was
Fig.
5 Temperature-programmed reduction curves of the metal
Fig. 4 XRD patterns of the metal oxides before hydrogen reduction.
oxides.
Table 2 Binding energies of Ni 2p3/2 obtained from curve-fitted values of experimental XPS spectra, crystal sizes obtained from XRD patterns and
reduction temperatures determined by TPR
Ni 2p3/2 (eV)
Ni(I)
Crystal sizeb (nm)
Before reduction
a
Sample
Ni(II)
Ni(III)
Satellite
Aer reduction
RTc (ꢁC)
Ni
852.1
851.6
852.7
852.8
854.9
854.3
855.4
855.6
856.5
857.8
855.9
857.1
861.3
861.9
861.5
862.1
36
26
20
21
24
30
14
18
377
235, 340
409
Ni–Cu
Ni–TiO2
Ni–CeO2
418
a
b
c
Estimated experimental error: ꢂ0.1 eV. Determined from XRD patterns based on the Scherrer equation. RT ¼ reduction temperature as
determined from TPR tests.
42162 | RSC Adv., 2014, 4, 42159–42167
This journal is © The Royal Society of Chemistry 2014