Catalytic features of Ni supported on CeO2-ZrO2 solid solution in the steam reforming of glycerol for syngas production

Article abstract of DOI:10.1039/c4ra02886d

Catalysts containing 5 wt% Ni supported on Ce_xZr _1-xO₂ solid solutions prepared by the Pechini method were tested in steam reforming of glycerol (SRG) reactions. The catalysts were characterized by temperature programmed reduction with H₂ (H ₂-TPR), XANES with H₂-TPR, X-ray diffraction (XRD) for fresh and used catalysts, XRD in situ for reduced catalysts, N ₂-physisorption, oxygen storage capacity (OSC), and scanning electron microscopy (SEM). H₂-TPR analysis revealed various degrees of interaction between the NiO phase and the supports, which were confirmed by XANES analysis. XANES data also showed greater reduction of mixed supports, than of pure supports, showing that the presence of Zr⁴⁺ promotes CeO₂-ZrO₂ reduction. XRD confirmed the formation of CeO₂-ZrO₂ solid solutions, with cubic and tetragonal crystal structures, depending on the composition. N₂ physisorption analysis showed a greater surface area on the CeO₂-ZrO₂ solid solution supports than on the pure supports and this increase may be related to the insertion of ZrO₂ in the CeO₂ lattice. The SRG reaction was performed at 973 K with molar feed ratios of H ₂O-glycerol = 3 : 1. All catalysts exhibited catalytic activity for the reaction, but carbon formation was high on the catalysts supported on pure oxide. The 5% Ni50CZ catalyst was the most selective for syngas production in the SRG. In addition to the nickel activity, the experiments revealed the activity of the supports in oxidation reactions, which was assigned to the formation of oxygen vacancies by the interaction of CeO₂ with ZrO₂. This journal is the Partner Organisations 2014.

Full text of DOI:10.1039/c4ra02886d

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Catalytic features of Ni supported on CeO2–ZrO2
solid solution in the steam reforming of glycerol for
syngas production
Cite this: RSC Adv., 2014, 4, 31142
*
Thaisa Aparecida Maia and Elisabete Moreira Assaf
Catalysts containing 5 wt% Ni supported on CexZr1ꢀxO2 solid solutions prepared by the Pechini method
were tested in steam reforming of glycerol (SRG) reactions. The catalysts were characterized by
temperature programmed reduction with H2 (H2-TPR), XANES with H2-TPR, X-ray diffraction (XRD) for
fresh and used catalysts, XRD in situ for reduced catalysts, N2-physisorption, oxygen storage capacity
(OSC), and scanning electron microscopy (SEM). H2-TPR analysis revealed various degrees of interaction
between the NiO phase and the supports, which were confirmed by XANES analysis. XANES data also
showed greater reduction of mixed supports, than of pure supports, showing that the presence of Zr4+
promotes CeO2–ZrO2 reduction. XRD confirmed the formation of CeO2–ZrO2 solid solutions, with cubic
and tetragonal crystal structures, depending on the composition. N2 physisorption analysis showed a
greater surface area on the CeO2–ZrO2 solid solution supports than on the pure supports and this
increase may be related to the insertion of ZrO2 in the CeO2 lattice. The SRG reaction was performed at
973 K with molar feed ratios of H2O–glycerol ¼ 3 : 1. All catalysts exhibited catalytic activity for the
reaction, but carbon formation was high on the catalysts supported on pure oxide. The 5% Ni50CZ
catalyst was the most selective for syngas production in the SRG. In addition to the nickel activity, the
experiments revealed the activity of the supports in oxidation reactions, which was assigned to the
formation of oxygen vacancies by the interaction of CeO2 with ZrO2.
Received 1st April 2014
Accepted 20th June 2014
DOI: 10.1039/c4ra02886d
www.rsc.org/advances
C3H8O3 $ 3CO + 4H2
(1)
(2)
Introduction
3CO + 3H2O $ 3CO2 +3H2
The diminution of fossil fuel reserves and the pollution caused
by their combustion, combined with the continued rise in
energy demands, make hydrogen and syngas (CO + H2) attrac-
tive alternative sources of energy, when they are produced from
renewable resources such as alcohols derived from biomass,
because they pollute less than most fuels and contribute to
reducing global warming and our dependence on fossil fuels.1
Syngas production is economically important, this being the
highest cost stage in the production of liquid compounds, in
the gas-to-liquid (GTL) process. Focusing on this stage, a
potential way of producing syngas is the steam reforming of
glycerol (SRG). SRG has been studied recently and shown to be a
promising process for the production of syngas and/or H2;
furthermore, glycerol is generated in quantities exceeding
current needs, as a byproduct of the process of biodiesel
production, the transesterication of vegetable oils. According
to Reaction (1), up to four moles of hydrogen per mole of glyc-
erol could be formed by simple decomposition. Under steam-
reforming conditions, however, the water–gas shi reaction
(Reaction (2)) will occur simultaneously:2
Therefore, a maximum of seven moles of hydrogen could be
generated in the SRG and the global reaction is presented as:
C3H8O3 +3H2O / 3CO2 + 7H2
(3)
Some studies suggest that catalysts used for the steam
reforming of ethanol (SRE) reaction may also be active in the
SRG, among them stand out from Ni catalysts. Zhang et al.3
studied Co and Ni catalysts for the SRE and SRG processes,
noting that these catalysts are signicantly active and selective
in both reactions. However, they suffered from the disadvantage
of carbon formation.4 Therefore, it is necessary to develop
alternative catalysts that form less carbon formation; for
example, by using supports that favor the gasication of carbon
to CO or CO2, such as CeO2 and ZrO2.
Metallic catalysts supported on CeO2 and ZrO2 have been
receiving a lot of attention because of several interesting prop-
erties of these supports, such as thermal stability, electrical
conductivity and oxygen vacancies. In previous work, these
supports have been evaluated in steam processes for H2 or
syngas production. CeO2 has the ability to store and release
´
˜
˜
Instituto de Quımica de Sao Carlos – Universidade de Sao Paulo, Avenida Trabalhador
˜
˜
Saocarlense, 400, Sao Carlos, SP, Brazil. E-mail: eassaf@iqsc.usp.br
31142 | RSC Adv., 2014, 4, 31142–31154
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
oxygen, participating in the oxidation of CO and hydrocarbons.5
ZrO2 may be dissolved in CeO2 to increase the mobility of
oxygen and thermal stability.6
X-ray absorption near-edge structure spectroscopy (XANES) with
H2 TPR. XANES spectra were collected at the Ni K-edge and Ce
LIII-edge, at the D06A – XAS beamline of the Brazilian
Synchrotron Light Laboratory (LNLS) in Campinas, Brazil. A Si
(1 1 1) polychromator was used to select the X-ray beam from the
synchrotron light produced by the 1.37 GeV electron storage
ring, at a maximum current of 200 mA. The Ni K-edge and Ce
LIII-edge absorption spectra were recorded in the transmission
mode with a CCD camera, in a photon energy range from
8200 to 8800 eV for Ni K-edge and from 5500 to 6000 eV for Ce
LIII-edge. The oxidized samples were heated, in hydrogen
owing at 30 mL minꢀ1, from 298 to 923 K at 10 K minꢀ1 and
held at 923 K for 30 min. The XANES spectra at the Ni K-edge
and Ce LIII-edge were collected in situ during reduction,
throughout the temperature range.
Nickel catalysts supported on CeO2 were tested in SRG by
Manfro et al.7 The reaction was carried out in a batch reactor
with aqueous solutions of 1 and 10 wt% glycerol, at 523 and
543 K. A maximum glycerol conversion of 30% was achieved at
543 K with 1% glycerol solution. In the gas phase, the mole
fraction of H2 was always higher than 70% and formation of
CH4 was very low (<1%). Increasing the glycerol concentration
decreased the rate of conversion and the formation of H2.
Nichele et al.8 studied the Ni catalysts activity in H2
production from SRG, analyzing Ni addition and the effect of
different supports, such as TiO2, ZrO2 and SBA-15. NiZrO2
catalyst showed higher stability and H2 selectivity at 923 K,
concluding thus that this system may be promising for the
reaction of SRG and can get good results, according to the
reaction conditions used.
X-ray diffraction (XRD). XRD powder patterns were acquired in
a Rigaku Multiex diffractometer, to identify the crystalline
phases, with Cu Ka radiation (1.54178 nm).ꢀT1he Bragg angle, 2q,
was scanned between 5ꢁ and 80ꢁ at 2o min
.
Considering the advantages of using nickel as the catalyst for
steam reforming reactions and using CeO2 and ZrO2 as
supports, the objective of study was to investigate nickel cata-
lysts supported on CeO2–ZrO2 solid solutions used in SRG
reactions for the production of syngas.
XRD of activated catalysts: XRD patterns of activated cata-
lysts were recorded in situ at the XPD-10B beamline of the LNLS
synchrotron radiation facility. The measurements were made in
the reection mode, in a 2q interval from 25ꢁ to 55ꢁ, with
˚
radiation of l ¼ 1.54996 A, calibrated with a Si (1 1 1) mono-
chromator. The oxidized samples were heated, in a ow of H2,
from 298 K to the temperature of reduction (923 K), at 10 K
minꢀ1, and held at that temperature for 1 h and, aer that, a
diffraction pattern was collected in situ.
From the XRD patterns, it was possible to calculate the mean
crystal diameter (d) of the particles responsible for the most
intense peak, for the supports (dsupport), NiO phase (dNiO) and
metal phase Ni (dNi), by using Scherrer's formula (eqn (4)):
Experimental
Preparation of catalysts
The supports were produced by the polymeric precursor
method (Pechini method).9 First, zirconium carbonate (Alfa
Aesar) was stirred into a solution of citric acid in ethylene
glycol.9,10 Next, an aqueous solution of cerium ammonium
nitrate (Alfa Aesar) was added, to give the following mole frac-
tions of CeO2 in the nal oxides: 0; 0.25; 0.50; 0.75 and 1. The
mixture was heated to 383 K and the resin formed was cured at
this temperature for 24 h, dried at 393 K for 12 h and calcined in
two steps: at 773 K for 2 h (2 K minꢀ1) and 973 K for 3 h (2 K
minꢀ1).
Catalysts with 5 wt% Ni were prepared by impregnating the
supports with aqueous Ni(NO3)2$3H2O (Alfa Aesar) solution at
343 K for 3 h. The solid was then dried in an oven at 333 K for
12 h and nally calcined at 873 K for 3 h.
In this paper, the catalysts Ni/CexZr1ꢀxO2 (0 < x < 1) are
abbreviated to NiCZ, the supports being referred to as C; Z;
25CZ; 50CZ and 75CZ for ceria, zirconia, 25, 50 and 75% ceria in
zirconia, respectively. The catalysts with a nickel load of 5 wt%
were thus named 5% NiC, 5% NiZ, 5% Ni25CZ, 5% Ni50CZ and
5% Ni75CZ.
kl
d ¼
(4)
b cos q
where: d ¼ mean crystal size (nm), k ¼ proportionality constant,
which depends on the shape of particles, here assumed to be
spherical (k ꢂ 0.9), b ¼ width at half height of the most intense
peak (radians), l ¼ wavelength of incident radiation (nm) q ¼
half the Bragg angle (o).
N2-physisorption. The specic surface area was estimated with
a Quantachrome Nova 1000e analyzer, by applying B.E.T. theory
to the N2 adsorption isotherm.
Oxygen storage capacity (OSC). OSC measurements on the
ceria–zirconia oxides were made an apparatus similar to that
for TPR, attached to
a Quadrupole mass spectrometer
(Omnistar QMS200 of Balzers). The sample was reduced in
pure H2 owing (25 mL minꢀ1) at 923 K for 1 h, than the
surface was purged with He (50 mL minꢀ1) at the same
temperature for 1 h. Aer stabilization of the O2 signal in the
mass spectrometer, while passing an O2 stream (8% O2: He –
50 mL minꢀ1) in parallel to the reactor (by-pass mode), the O2
ow was passed through the sample to obtain the break-
through curve, from which the total OSC was determined.
These analyses were performed in the Catalysis and Chemical
Processes Laboratory, at the National Institute of Technology
(INT), Rio de Janeiro, Brazil.
Characterization of catalysts
Characterization for fresh catalysts
H2 temperature-programmed reduction (H2-TPR). The TPR
analyses were performed in a multipurpose analytical system
(SAMP3, Termolab Equipamentos Ltd, Brazil) with a thermal
conductivity detector (TCD), a reducing gas mixture of 2% H2–
98% Ar (v/v) and a heating rate of 10 K minꢀ1 from 298 to
1273 K.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 31142–31154 | 31143

View Article Online
RSC Advances
Paper
Characterization for spent catalysts
XRD aer SRG reaction. XRD powder pattern for 5% Ni50CZ
aer reaction were acquired under the same conditions of the
fresh catalysts.
SEM analyses. SEM photomicrographs were taken in a ZEISS
LEO 440 microscope (Cambridge, England), with an OXFORD
detector (7060), operating at 20 kV electron beam. The samples
were coated with 10 nm of gold in a BAL-TEC MED020 sputter
coating system (BAL-TEC, Liechtenstein) and kept in a desic-
cator until the time off analysis.
Catalytic tests
The catalytic tests were performed at atmospheric pressure, in a
xed-bed tubular quartz micro-reactor. Prior to the reaction,
150 mg of the catalyst was introduced into the reactor and
reduced in situ in owing H2 (30 mL minꢀ1) at 923 K (10 K
minꢀ1) for 1 h, to activate the catalyst. Aer that, the surface of
the catalyst was purged in a ow of pure N2 at the reaction
temperature (973 K). The catalytic bed temperature was
controlled with the aid of a thermocouple introduced into the
bed. The reaction was started in a hydrogen-free feed, when the
N2 was replaced with water–glycerol in a 3 : 1 molar ratio,
owing at 2.5 mL hꢀ1, supplied by a piston pump. This mixture
was vaporized in a preheating chamber (453 K) before reaching
the reactor. The gaseous products were analyzed with an in-line
gas chromatograph (VARIAN GC-3800) with two columns in
parallel, each with a TCD at the outlet. The columns were
packed with Porapak-N and 13X Molecular Sieve, operating
between 313 K and 353 K, with He and N2, respectively, as
carrier gas, owing at 10 mL minꢀ1. The liquid products
collected at the reaction outlet were analyzed with a gas chro-
matograph (Shimadzu), with H2 as carrier gas and HP5 capillary
column, operating between 308 K and 523 K. During prepara-
tion of the liquid products for GC analysis, the solution was kept
at temperature below 283 K, to avoid the evaporation of
products.
Fig. 1 Reduction profile of supports.
reduction of Ce4+ to Ce3+ by the removal of O2ꢀanions from the
lattice and formation of Ce2O3, as follows:12
2CeO2 + H2 / Ce2O3 + H2O
(5)
The reduction of Z occurs at 942 K. This small reduction peak
is attributed to Zr4+ / Zr3+, which may occur only on the
support surface.13 In the proles of mixed supports, 25CZ and
50CZ a single reduction peak is present, characteristic of a
single phase. According to the literature,14 the reduction of a
solid solution can be affected by oxygen mobility, which allows
the reduction of the bulk of the material at a lower temperature.
Indeed, the reduction in the bulk of this solid solution is
strongly promoted by the insertion of ZrO2 into the CeO2 lattice,
as can be observed in Table 1, which shows the H2 consumption
and OSC measurements for the supports. These analyses
showed a higher H2 consumption and higher oxygen storage
capacity for 25CZ and 50CZ. The 75CZ support shows an addi-
tional a high-temperature peak, unlike 50CZ and 25CZ. This
peak may correspond to the reduction of bulk ceria in
an inhomogeneous portion of the sample with a higher content
of Ce.
The glycerol conversions were calculated at the end of the
experiments by analyzing the amount of unreacted glycerol
contained in the cold trap, and were dened as mole percent of
glycerol consumed per mole of supplied reactant.
Fig. 2 shows the TPR proles of the catalysts. All the Ni
reduction peaks were attributed to the reduction of Ni2+ to Ni0.15
Previous H2-TPR investigations indicated that TPR peaks at
lower temperatures refer to the reduction of nickel oxide crys-
tallites and, those at higher temperatures can be attributed to
nickel species interacting more strongly with the support.16 All
catalysts show broadly similar proles, with three reduction
peaks. However, these peaks are shied to higher or lower
temperatures and vary in relative magnitude, showing more or
less interaction between nickel and support.17 For comparison,
the NiO reduction prole is also shown, with a single reduction
peak at Tmax ¼ 690 K. Since the supports were reduced over the
same temperature range as the nickel species, proles show co-
reduction of NiO and support oxides at these temperatures.
The peak at the lowest temperature (g) can be attributed to
reduction of NiO interacting weakly with the support. This peak
corresponds to the reduction of NiO species inuenced by
The catalysts 5% NiC, 5% Ni50CZ and 5% NiZ were also
subjected to temperature ramps, to assess the catalytic behav-
iour of these catalysts at various temperatures and to determine
the best temperature at which to work. The product composi-
tion was recorded for all catalysts over 6 h on stream at 973 K.
Results and discussion
H2 temperature-programmed reduction (H2-TPR)
Fig. 1 shows the CeO2–ZrO2 reduction proles of all supports.
The TPR prole of C shows two peaks, a small one at 833 K and a
stronger peak at 1193 K; this prole is similar to that reported in
the literature.11 The rst peak may be associated with oxygen
reduction on the surface of CeO2, which is covered with oxygen
anions (O2ꢀ or Oꢀ), and the second peak corresponds to the
reduction of CeO2 in the bulk of the material, through the
31144 | RSC Adv., 2014, 4, 31142–31154
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Table 1 Structural characteristics of the CZ supports
H2 consumption (mol H2) ꢃ 105
Experimental
OSC (mmol
Theoretical Reduction (%) O2 per g cat)
Support dsupport (nm) Lattice parametera (nm) SBET (m2 gꢀ1
)
C
25.9
8.4
7.3
7.8
13.2
0.5394
0.5377
0.5322
0.5268
—
18.3
57.6
48.3
47.9
29.3
1.30 (bulk), 0.80 (surface)
1.25 (bulk), 4.35 (solid solution) 21.2
8.50
7.00
0.45
29.0
7.3
25.8
59.0
96.8
1.1
265
355
419
380
67
75CZ
50CZ
25CZ
Z
14.5
7.3
40.4
a
based on the standard lattice parameter of CeO2: 0.5412 nm.
of vacancies in the support, which could also be related to the
higher OSC for this support.
XANES with H2-TPR
XANES in situ analyses during H2 TPR are very useful, as they
show the variation in electronic properties of the Ni atoms
during the reduction process. For this purpose, the most rele-
vant features of XANES spectra are the intensity of the white line
and pre-edge structures.21,22 XANES spectra of two of the cata-
lysts studied are shown in Fig. 3. Reference spectra of NiO and
metal Ni are presented in Fig. 4, representing the oxidation
states Ni2+ and Ni0 respectively. It can be veried that all the
samples gave proles similar to that of NiO at lower tempera-
tures, with the same white line at 8350 eV, and that this white
line is caused by Ni–O interaction.23–25 When the catalysts are
heated in a ow of H2, it is found that the intensity of the white
line falls, so that the spectrum takes on a prole similar to that
Fig. 2 Reduction profile of catalysts.
oxygen vacancies in the support. This occurs by an initial of metallic Ni, while the intensity of the pre-edge signal at 8330
weakening of the Ni–O bond or polarization of the H2 molecule, eV rises. These are results show the reduction of Ni(II) to Ni0,26
which becomes more reactive for NiO reduction in the neigh- conrming the TPR analyses. It is observed that, corroborating
borhood of oxygen vacancies.18 The second peak (b) indicates results obtained by TPR, the reduction of Ni species starts
reduction of NiO crystallites (observed by XRD). The third peak around 623 K and, by 823 K, is complete in all catalysts.
(s) can be attributed to the reduction of NiO species interacting
more strongly with the support.
Fig. 5 and 6 show XANES spectra recorded at the Ce LIII edge
for 5% NiCZ catalysts, during H2 TPR. The analyzes were per-
Previous H2-TPR investigations pointed out that the main formed from room temperature to 923 K. At room temperature,
features of CeO2–ZrO2 supports are the creation of oxygen the spectrum corresponds to Ce4+ species and a double white
vacancies as the solid solution is formed and the effect on the line was observed (named B and C). Due to the crystal eld
redox behavior of the interaction between these vacancies and splitting for the 5d level of Ce, these peaks were assigned to 2p3/
the supported catalytic oxides.19 Not only does the solid solution 2/(4f1L)5d1 and 2p3/2/(4f0)5d1 transitions, respectively. The “L”
inuence the NiO reduction, but also the NiO particles appar- denotes that an electron in the 2p orbital of oxygen is trans-
ently facilitate the reduction of CeO2–ZrO2 through the H2 ferred to the 4f orbital of a Ce atom and 5d1 refers to the excited
spillover effect during TPR or activation.20 This process can be electron in the unoccupied 5d state.27,28
explained in various ways: the H2 molecule may be dissociated
The increasing temperature of the TPR program causes
and adsorbed on NiO and hydrogen atoms can then be changes in peaks B and C in the 5% NiCZ spectra. There is a
distributed on to the CeO2–ZrO2. Another possibility is that the decrease in the intensity of C band and displacement of band B
H2 is activated at the contact point between NiO and solid to lower energies (indicated as A in Fig. 6).
solution particles. A third explanation is that the hydrogen
For the 5% NiC catalyst, the two peaks B and C, corre-
spillover effect could occur from a small number of Ni sites sponding to Ce4+, do not change, but, with the addition of Zr to
formed on the NiO surface, but not be detected by TPR.20 The the supports, the higher energy peak loses intensity and
amounts of H2 consumed in each TPR experiment are shown in reduction proles of catalysts supported on mixed oxides show
Table 1. It is seen that the support that led to the highest H2 a band A corresponding to the oxidation state Ce3+, indicating a
consumption was 50CZ. From this result, it is assumed that this mixture of two oxidation states, Ce3+ and Ce4+. For the catalysts
higher consumption may be associated with a greater number with higher Zr content, peak A, due to the dipole-allowed
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 31142–31154 | 31145

View Article Online
RSC Advances
Paper
Fig. 3 In situ XANES-H2 TPR spectra of the catalysts (a) 5% NiZ and (b) 5% Ni50CZ, at the NiK absorption edge.
Zr promotes CeO2–ZrO2 reduction at lower temperatures,
besides increasing the extent of reduction, which increases with
increasing Zr content, as conrmed by XANES-H2 TPR analysis
(Fig. 7).
X-ray diffraction (XRD)
Fig. 8 and 9 present the X-ray diffraction patterns of the
supports and catalysts. These were compared with published
patterns.32 The pattern of support C showed diffraction peaks at
Bragg angles 2q ¼ 28.6, 36.1, 47.4, 56.4, 59.2, 69.4, 76.7 and
79.4ꢁ, characterizing support C as a crystalline Ce oxide with a
cubic uorite structure (JCPDS 81-0792). The XRD pattern of the
support Z showed peaks at 2q ¼ 17.43, 24.17, 28.28, 31.65, 34.49,
35.33, 38.65, 40.90ꢁ; 45.60, 50.44, 54.16, 55.36, 62.82 and 65.79ꢁ
(JCPDS 83-0944), characteristic of the ZrO2 monoclinic struc-
ture, besides three main diffraction peaks of the tetragonal
structure of this oxide at 2q ¼ 30.38, 59.26 and 60.37ꢁ (JCPDS 79-
1768).33
Fig. 4 XANES spectra of reference compounds (metallic Ni and NiO).22
In the XRD patterns of the mixed oxides, the main peak at 2q
transition 2p3/2 / (4f1)5d,29,30 which characterizes the Ce3+ ¼ 28.6, ascribed to the ceria structure, shis to higher angles as
oxidation state, is more intense and thus the addition of Zr is ZrO2 is added, indicating the formation of solid solution. This is
correlated with a higher reducibility in mixed support.27,31 As due to the insertion of Zr4+ ions in the crystal lattice of CeO2,
observed in the TPR-H2 proles of the supports, the addition of which causes a change in the lattice parameter (Table 1). The
31146 | RSC Adv., 2014, 4, 31142–31154
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Fig. 5 In situ XANES-H2 TPR spectra of the 5% NiCZ catalysts at the Ce LIII absorption edge.
Fig. 6 Selected XANES in situ spectra at the Ce LIII absorption edge for 5% NiCZ catalysts undergoing H2 TPR.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 31142–31154 | 31147

View Article Online
RSC Advances
Paper
insertion of increasing amounts of Zr4+ into the cubic uorite
structure decreases the cell volume because of the smaller ionic
4+
radius of Zr4+ (0.84 A), with relative to that of Ce (0.97 A),
leading to progressive shrinkage of the lattice parameter.34 The
crystal structure of the solid solution depends on the compo-
sition. For the supports with higher Ce content, the diffracto-
grams show a prole closer to the cubic uorite structure of
CeO2, with a shi in 2q consistent with the insertion of Zr4+
mentioned above. Higher contents of zirconia transform the
structure to tetragonal.33
˚
˚
Besides the diffraction peaks of the supports, the XRD
patterns of the catalysts (Fig. 9) also show weak diffraction
peaks at Bragg angles 2q ¼ 37.29; 43.29 and 62.96ꢁ, which are
characteristic of the NiO phase (JCPDS 44-1159).
Structural characteristics of the supports and catalysts
Fig. 7 Comparison of the XANES-H2 TPR in situ spectra reduction of
catalysts at 923 K – influence of Zr addition in the support reduction.
Table 1 summarizes some structural characteristics of the
supports, such as mean crystallite diameter, BET surface areas,
H2 consumption in TPR and OSC, determined by the techniques
of X-ray diffraction, N2 physisorption, temperature pro-
grammed reduction and oxygen storage capacity. It was
observed that Zr4+ addition led to a reduction in the crystallite
size (dsupport). These smaller values can be interpreted as a result
of loss of crystallinity with solid solution formation.35 It is
observed that the reduced crystallite size in mixed oxide
supports has resulted in an increase in BET surface area, sug-
gesting that the addition of zirconium to the solid solution
leads to textural changes.
H2 consumption values were calculated from TPR proles,
considering the experimental and theoretical H2 consumption.
It is observed that these values are correlated with the surface
area of the supports. For the C support, it is apparent that the H2
consumption corresponding to surface reduction is less than
the H2 consumption calculated for the bulk reduction, is typical
behavior of low-surface-area materials, also indicating that the
reduction of support C begins on the surface and progressively
affects the bulk region.12 The Z support showed a low level of
reduction, also reported by other authors.13,36,37
Fig. 8 X-ray diffractograms of supports.
The H2 consumption by the mixed oxides was higher than
that by the pure oxide supports. This higher consumption could
be related to the greater BET surface area and to the solid
solution composition. A higher BET surface area favors the
availability of cerium at the surface and consequently a higher
H2 consumption. Moreover, the insertion of zirconium can
facilitate oxygen transport from bulk to the surface, facilitating
the solid solution reduction.6,12
The H2 consumption and oxygen storage capacity (OSC) data
also agree. A higher consumption of H2 was accompanied by a
greater OSC. According with Gutierrez-Ortis et al.,37 reduction in
the mixed oxide of cerium is as follows:
CexZr1ꢀxO2 + dH $ CexZr1ꢀxO2ꢀd + dH O + Vo
2
2
Through the reduction of cerium, oxygen ions are removed
from the surface (oxygen being eliminated by the formation of
H2O molecules), leading to the formation of oxygen vacancies
Fig. 9 X-ray diffractograms of catalysts.
31148 | RSC Adv., 2014, 4, 31142–31154
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
(Vo). The presence of these vacancies greatly boosts the oxygen reforming of other alcohols and hydrocarbons, the following
mobility from the bulk to the surface. Thus, it is possible to reactions may occur in the steam reforming of glycerol.2,8,38
relate the reduction process, vacancy formation, the mobility of
C3H8O3 $ 3CO + 4H2
C3H8O3 + H2O / CO2 + 2CO + 5H2
CH4 + H2O $ CO + 3H2
CH4 + CO2 $ 2CO + 2H2
2CO / CO2 + C
(1)
oxygen and thus the OSC of the supports. Observing the H2
consumption and OSC data, it is apparent that the 50CZ and.
25CZ supports have greater mobility of oxygen than the
others.
Table 2 summarizes some structural properties of 5% Ni/
CeO2–ZrO2 catalysts, such as NiO crystallite size, BET surface
area and H2 consumption. The crystallite sizes (dNiO) were
calculated by the Scherrer equation from the main peak of the
nickel phase present in the diffraction pattern (2q ¼ 43.3ꢁ).
Comparing the surface areas analysis in Tables 1 and 2, it
appears that the catalysts have less adsorption capacity than the
supports. This is due to the nickel oxide coating on the support
surface and obstruction of the pores.
Through the values of NiO crystal size (dNiO) for all catalysts,
it can be seen that they decrease with increasing amounts of
Zr4+ in the support, indicating that the NiO is nely dispersed
on the support, perhaps due to greater NiO interaction with the
solid solution, promoted by zirconia. The XRD in situ analyses
were performed for 5% NiC and 5% Ni50CZ catalysts. Aer
reduction, the nickel species present on both supports was in
the form of metallic Ni, and its calculated crystal size (dNi),
based on the peak at 2q ꢂ 44.0ꢁ, was 27.0 nm for 5% NiC and
20.7 nm for 5% Ni50CZ, from which it can be conrmed that the
ceria–zirconia solid solution enhanced dispersion of the nickel
species on the support.38
The H2 consumption showed that all catalysts had a nickel
phase reduction over 80%. The literature reports that NiO
reduction is inuenced by the Ni–Ce interaction. Hydrogen
might be dissociatively adsorbed on NiO and hydrogen atoms
spilled over to CeO2–ZrO2 particles, or hydrogen could be acti-
vated at the contact point of between the NiO and solid solution
particles. However, from the H2 consumption data, a correla-
tion cannot be established between the composition of the solid
solution and nickel reducibility.17
(6)
(7)
(8)
(9)
C2H4 / polymers / C
(10)
Reactions at 673 K showed low conversion rates to gaseous
products, when compared to higher temperatures. Condensed
reaction products were yellowish in color and gas chromatog-
raphy analysis identied unsaturated compounds (not quanti-
ed). The GC-MS analyses at this temperature identied large
amounts of the following liquid products: acetic acid, 1,2,3-
butanetriol, 1,2-propanediol, propanal and butanoic acid.
These results of GC-MS analyses corroborate with Buffoni et al.4
Metallic Ni favors C–C bond cleavage in the glycerol mole-
cule, leading to CO and H2 formation. However, on thermody-
namic grounds, the steam reforming reaction, being
endothermic, is limited at lower temperatures, favoring the
formation of liquid products. Besides this, it was observed that
a high pressure-drop developed in the reactor, due to the
deposition of carbon in the catalytic bed and consequent cata-
lyst deactivation. The precursor of this carbon may be ethylene,
formed by dehydration of glycerol.3,4 The ethylene may be
formed aer an initial C–C bond cleavage, forming the radicals
CH2(OH)–CH(OH) and CH2–OH, which subsequently evolve
towards the formation of C2H4, H2 and CH4.8,39,40
As expected, increasing temperature led to a rise in gas
product yields, mainly favouring the formation of H2 and CO,
which was highest at 973 K. A high reaction temperature
enhanced both the reforming reaction and the decomposition
of glycerol, producing more H2, CO and CO2, as seen in Fig. 10.38
Methane is stable in the system at the lower temperatures, but
its content decreases strongly as the temperature rises, owing to
the steam and dry reforming of CH4 (Reactions (7) and (8)).
Another product that is not favoured in the highest temperature
is CO2. Above 873 K, it is probable that the increasing inuence
Catalytic tests
Catalytic activity was assessed for 5% NiZ, 5% NiC and 5%
Ni50CZ catalysts, at temperatures varying between 673 K and
973 K (Fig. 10). The gas products obtained were typically a
reformed gas mixture obtained by reforming oxygenated
hydrocarbons. It has been reported that, similarly to the
Table 2 Structural properties of 5%NiCZ catalysts
H2 consumption (mol H2) ꢃ 105
Catalyst
dNiO(nm)
SBET (m2 gꢀ1
)
Experimental
Theoretical
Reduction (%)
5% NiC
28.8
26.8
21.8
18.2
13.3
16
51
36
42
25
7.5
7.8
7.0
6.9
8.3
8.7
8.5
8.5
8.8
8.7
86.2
91.8
82.3
78.4
95.4
5% Ni75CZ
5% Ni50CZ
5% Ni25CZ
5% NiZ
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 31142–31154 | 31149

View Article Online
RSC Advances
Paper
Fig. 10 Influence of temperature on the gaseous products of SGR catalyzed by 5NiC; 5Ni50CZ and 5NiZ.
of the reverse water–gas shi reaction (Reaction (2)) and/or the (reverse of Reaction (7)) is exothermic reaction, and thus, at
dry reforming of methane reduces the amount of CO2.2 The higher temperature, methanation is unfavoured, resulting to a
increase in CO and H2 yields with rising temperature may also decrease in CH4 yield at equilibrium.41
be due to glycerol decomposition (Reaction (1)).
Among the catalysts analyzed, it was found that 5% Ni50CZ
In view of the better performance of the catalysts at higher was the most stable during the time of reaction, besides
temperatures, all catalysts were tested over 6 hours at 973 K. achieving the best H2, CO and CO2 formation (favoring Reac-
Fig. 11 and Table 3 present the results of the catalytic tests. tions (1) and (6)) and a low carbon deposit (Table 3). This higher
Under these conditions, complete glycerol conversion occurred activity and stability may be related to its structural properties.
on all catalysts. In the most of catalysts, the H2 formation This catalyst showed a mixture of crystalline phases (cubic and
decreased during the reaction, accompanied by an increase in tetragonal) in the support, according to the XRD analysis. The
ethylene formation and possible catalytic deactivation (Reac- incorporation of Zr4+ in the CeO2 lattice may lead to a structural
tion (10)).3,4
distortion, which promotes easier reduction of the support and
Fig. 11 shows the time proles of the GSR reactions for all the thus a higher activity for the 50CZ solution than for pure C and
5% NiCZ catalysts and Table 3 shows the mean selectivity of Z supports. In the characterization analyses of this catalyst, it
gaseous products formed on these catalysts during the time on was observed that the addition of Zr4+ favored a high support
stream. For all catalysts, the CO formation was greater than CO2 reduction in H2 TPR and oxygen storage capacity. The high
formation.41 Thus, it is likely that the inuence of the reverse activity and selectivity for H2, CO and CO2 on 5% Ni50CZ might
water–gas shi reaction is strongest at this temperature, which be due to the OSC. OSC is a important factor to SRG reaction
is consistent with the thermodynamics of the process. This because a higher OSC indicates a better oxygen mobility, which
higher CO formation was accompanied by low CH4 formation, helps in gasifying the coke deposited on the catalyst during the
indicating the parallel occurrence of the steam and dry methane reaction and thus keeps the catalyst active.41,44
reforming reactions (Reaction (7) and (8)) and the activity of
Besides the structural change, another factor that may
these catalysts for these reactions at 973 K, in accordance with promote the higher activity of this catalyst is a strong interac-
studies of Ni catalysts supported on various supports, which tion between Ni and support. The addition of Zr4+ addition also
showed methane reforming activity.42,43 Another reason for the promotes a better nickel dispersion on the support and conse-
lower CH4 formation is to the fact that the methanation reaction quently a reduced nickel mean crystal size for 5% Ni50CZ,
31150 | RSC Adv., 2014, 4, 31142–31154
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Fig. 11 Gaseous product yields over 6 h of the steam reforming of glycerol at 973 K.
compared to 5% NiC, as was observed in the in situ XRD anal-
The analysis of carbon formation presented in (Table 3)
yses. This better dispersion implies stronger interaction and shows a considerably heavier deposit in the 5NiZ catalyst and
more intimate contact between Ni and support, improving the the gaseous product results show that this catalyst formed less
probable oxygen transfer from support to Ni.41,45 According to H2, CO and CO2 and more ethylene and CH4, than the other
Takeguchi et al.,20 the interaction between nickel species and catalysts. Thus, this catalyst undergoes greater deactivation by
oxygen vacancies may create active Ni–Ce3+ centers (synergistic carbon formation, probably through Reactions (9) and (10).
effect). This interaction may increase catalytic activity, since the Carbon formation was also great on 5% NiC, and this catalyst
vacancies activate the oxygen species.
also formed a large amount of ethylene. On the mixed supports,
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 31142–31154 | 31151

View Article Online
RSC Advances
Table Gaseous products and carbon formation of the steam
Paper
3
reforming glycerol on 5% NiCZ catalysts at 973 K for 6 h
Gaseous products (mol prod./mol
glycerol converted)
Carbon
Catalyst
H2
CH4
CO
CO2
C2H4
(mmol hꢀ1
)
5NiZ
0.96
1.29
3.05
1.97
2.51
0.25
0.05
0.21
0.17
0.14
0.96
0.65
1.33
1.03
1.00
0.11
0.37
0.77
0.58
0.63
1.22
0.37
—
1.19
1.37
3.11
0.42
0.78
1.51
1.64
5Ni25CZ
5Ni50CZ
5Ni75CZ
5NiC
carbon formation decreased with increasing Zr4+ in the solid
solution.
According to the literature,46–48 CexZr1ꢀxO2-supported cata-
lysts show more resistance to carbon formation than those on
pure CeO2 and ZrO2, because incorporation of Zr4+ in the CeO2
lattice can promote oxygen species reduction, both at the
surface and in the bulk of the solid solution, which would favor
oxidation reactions, and result in carbon gasication.
Characterization of spent catalyst aer the SRG reaction
On the basis of the results of the catalytic tests, further char-
acterization of the catalyst with the best performance (5%
Ni50CZ) was carried out aer the reaction. The carbon formed
on the used catalysts was subjected to XRD and SEM analysis.
Fig. 12 shows the XRD patterns for fresh and used catalyst. In
the second pattern, the metallic Ni0 phase is seen at 2q ꢂ 44ꢁ
and 52ꢁ, indicating that the nickel remained totally or partially
reduced in the catalytic test. Besides the diffraction peaks for
phase, the XRD pattern of the 5% Ni50CZ used catalyst also
shows more dened peak and intense diffraction peaks at Bragg
angles 2q ꢂ 25.8, 36.6, 40.3, 42.5 and 68.1ꢁ, which are charac-
teristic of carbon graphite (JCPDS 898487).
Fig. 13 SEM image of (A) fresh and (B) used 5% Ni50CZcatalyst.
photomicrograph of the fresh catalyst and, it can be seen that
the surface of the catalyst consists of large particles with a semi-
spherical tendency. This catalyst, aer the SRG reaction
(Fig. 13B), shows the carbon laments formed on the catalyst
surface. The diffractogram for this catalyst aer reaction
(Fig. 12) also indicated carbon graphite formation, with high
crystallinity. Carbon graphite is not easily detected by SEM, but
lamentous carbon nanotube structures can be seen in the SEM
images.
In addition to the XRD, SEM photomicrograph of fresh and
used catalyst are presented in Fig. 13. Fig. 13A shows the
As already reported,49,50 the presence of species such as NiO
crystallites that interact weakly with the support (free NiO), is
responsible for the carbon formation on nickel catalysts. Thus,
this kind of free NiO, identied in the TPR prole of the catalyst
5% Ni50CZ in Fig. 2 (peak g) may favour the formation of
carbon laments on this catalyst. However, the formation of
carbon lament does not affect the catalytic activity, but results
in the variation of the catalytic bed by obstruction.
The catalytic test for SRG on 5% Ni50CZ showed high CO
formation, which may be related to the carbon formation. Small
NiO crystallites may promote the formation of lamentous
carbon at a low growth rate, as may the more dispersed Ni
particles, due to a low driving force for diffusion of carbon
through the crystallites.49 This continuous formation of carbon
Fig. 12 XRD of used and fresh 5% Ni50CZ.
31152 | RSC Adv., 2014, 4, 31142–31154
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
may then be accompanied in part by carbon gasication, 13 W. P. Dow, Y. P. Wang and T. J. Huang, J. Catal., 1996, 160,
leading to the higher formation of carbon monoxide.
155–170.
14 E. Beryariva, P. Fornasiero, J. Kaspar and M. Graziani, Catal.
Today, 1998, 45, 179–183.
15 J. D. A. Bellido and E. M. Assaf, J. Power Sources, 2008, 177,
24–32.
Conclusions
The present results show that increasing the reaction temper-
ature in the steam reforming of glycerol favors the formation of 16 M. Diskin, R. H. Cunningham and R. M. Ormerod, Catal.
CO and H2, while lower temperatures lead to catalyst deactiva- Today, 1998, 46, 147–154.
tion, possibly by the formation of carbon. The incorporation of 17 S. Xu and X. Wang, Fuel, 2005, 84, 563–567.
Zr4+ into the CeO2 lattice leads to a structural change in the 18 J. B. Wang, Y. L. Tai, W. P. Dow and T. J. Huang, Appl. Catal.,
mixed oxides and favors reduction of oxide with increasing of Zr
content. This structural change, together with a stronger nickel– 19 M. H. Youn, J. G. Seo, K. M. Cho, S. Park, D. R. Park, J. C. Jung
support interaction, causes the 5Ni50CZ catalyst to be the most and I. K. Song, Int. J. Hydrogen Energy, 2008, 33, 5052–5059.
stable during the SRG reaction, favoring increased formation of 20 T. Takeguchi, S. Furukawa and M. Inoue, J. Catal., 2001, 202,
H2 and CO and reduced carbon formation at 973 K. For this 14–24.
catalyst, it was also observed that the CeO2–ZrO2 combination 21 L. P. R. Profeti, E. A. Ticianelli and E. M. AssaF, Int. J.
favors the removal of carbon deposits, owing to the increased Hydrogen Energy, 2009, 34, 5049–5060.
mobility of oxygen in the solid solution. Considering these 22 L. P. R. Profeti, J. A. C. Dias, J. M. Assaf and E. M. Assaf, J.
results, it was concluded that the association between nickel Power Sources, 2009, 190, 525–533.
and CeO2–ZrO2 solid solution presents a promising catalytic 23 J. Juan-Juan, M. C. Roman-Martınez and M. J. Illan-Gomez,
A, 2001, 218, 69–79.
´
´
´
´
system for SRG.
Appl. Catal., A, 2004, 264, 169–174.
24 S. Takenaka, E. Kato, Y. Tomikubo and K. Otsuka, J. Catal.,
2003, 219, 176–185.
25 W. Wen, J. E. Calderon, J. L. Brito, N. Marinkovic,
J. C. Hanson and J. A. Rodriguez, J. Phys. Chem. C, 2008,
112, 2121–2128.
Acknowledgements
The authors acknowledge Brazilian federal and state funding
agencies CNPQ and FAPESP for nancial assistance, the LNLS
˜´
(Brazilian Synchrotron Light Source Laboratory) for the XANES 26 V. M. G. DelaCruz, J. P. Holgado, R. Perenıguez and
´
facilities (DXAS beamlines) and the researcher Marco Andre
A. Caballero, J. Catal., 2008, 257, 307–314.
Fraga of National Institute of Technology (INT-Rio de Janeiro, 27 M. G. Zimicz, S. A. Larrondo, R. J. Prado and D. G. Lamas, Int.
Brazil) for Oxygen Capacity Storage (OSC) analyses.
J. Hydrogen Energy, 2012, 37, 14881–14886.
28 K. Kaneko and Y. Tamaura, J. Phys. Chem. Solids, 2009, 70,
1008–1014.
29 K. Chayakul, T. Srithanratana and S. Hengrasme, Catal.
Today, 2011, 175, 420–429.
Notes and references
1 L. C. Lima and T. N. Veziroglu, Int. J. Hydrogen Energy, 2001,
26, 39–45.
2 Y. Cui, V. Galvita, L. Rihko-Struckmann, H. Lorenz and
K. Sundmacher, Appl. Catal., B, 2009, 90, 29–37.
3 B. Zhang, X. Tang, Y. Li, Y. Xu and W. Shen, Int. J. Hydrogen
Energy, 2007, 32, 2367–2373.
30 J. Zhang, Z. Wu, T. Liu, T. Hu, Z. Wu and X. Ju, J. Synchrotron
Radiat., 2001, 8, 531–532.
31 S. Ricote, G. Jacobs, M. Milling, Y. Ji, P. M. Patterson and
B. H. Davis, Appl. Catal., A, 2006, 303, 35–47.
32 Powder Diffraction File (PDF), Joint Committee on Powder
Diffraction Standards (JCPDS), International Center of
Diffraction Data (ICDD), Pennsylvania, USA, 1994 (CD ROM).
4 I. N. Buffoni, F. Pompeo, G. F. Santori and N. N. Nichio,
Catal. Commun., 2009, 10, 1656–1660.
5 R. G. Rao and B. G. Mishra, Bull. Catal. Soc. India, 2003, 2, 33 C. Diagne, H. Idriss and A. Kiennemann, Catal. Commun.,
122–134.
2002, 3, 565–571.
6 S. Letichevsky, C. A. Tellez, R. R. Avillez, M. I. P. Silva, 34 L. M. Martinez, M. Araque, J. C. Vargas and A. C. Rogera,
M. A. Fraga and L. G. Appel, Appl. Catal., B, 2005, 58, 203–
210.
7 R. L. Manfro, A. F. Costa, N. F. P. Ribeiro and
Appl. Catal., B, 2013, 132–133, 499–510.
35 J. Kaspar, P. Fornasiero, G. Balducci, R. Di Monte, H. Hickey
and V. Sergo, Inorg. Chim. Acta, 2003, 349, 217–226.
M. M. V. M. Souza, Fuel Process. Technol., 2011, 92, 330–333. 36 J. D. A. Bellido, E. Y. Tanabe and E. M. Assaf, Appl. Catal., B,
8 V. Nichele, M. Signoretto, F. Menegazzo, A. Gallo, V. Dal 2009, 90, 485–488.
Santo, J. Cruciani and G. Cerrato, Appl. Catal., B, 2012, 37 J. I. Gutierrez-Ortis, B. Rivas, R. L. Fonseca and
111–112, 225–232.
9 M. P. Pechini, Sprague Electric Co., US Pat. 3330697, 26 Aug 38 K.
J. R. G. Velasco, Appl. Catal., A, 2004, 269, 147–155.
Kamonsuangkasem, S. Therdthianwong
and
1963.
A. Therdthianwong, Fuel Process. Technol., 2013, 106, 695–
703.
ˆ
10 P. J. B. Marcos and D. Gouvea, Ceramica, 2004, 50, 38–42.
11 A. Holmgren and B. Andersson, J. Catal., 1998, 178, 14–25.
12 S. Damyanova, K. Pawelec, K. Arishtirova, M. V. Martinez-
Huerta and J. L. G. Fierro, Appl. Catal., A, 2008, 337, 86–96.
39 M. R. Nimlos, S. J. Blanksby, X. Qian, M. E. Himmel and
D. K. Johnson, J. Phys. Chem. A, 2006, 110, 6145–6156.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 31142–31154 | 31153

View Article Online
RSC Advances
Paper
40 T. A. Semelsberger and R. L. Borup, J. Power Sources, 2006, 45 J. Chen, Q. Wu, J. Zhang and J. Zhang, Fuel, 2008, 87, 2901–
155, 340–352. 2907.
41 C. D. Dave and K. K. Pant, Renewable Energy, 2011, 36, 3195– 46 H. Song and U. S. Ozkan, J. Catal., 2009, 261, 66–74.
3202.
47 T. A. Maia, J. M. Assaf and E. M. Assaf, Mater. Chem. Phys.,
2012, 132, 1029–1034.
48 S. Shao, A. W. Shi, C. L. Liu, R. Z. Yang and W. S. Dong, Fuel
Process. Technol, 2014, 125, 1–7.
42 X. Wang, M. Li, M. Wang, H. Wang, S. Li, S. Wang and X. Ma,
Fuel, 2009, 88, 2148–2153.
43 S. Adhikari, S. D. Fernando, S. D. F. To, R. M. Bricka,
P. H. Steele and A. Haryanto, Energy Fuels, 2008, 22, 1220– 49 K. Y. Koo, H. S. Roh, Y. T. Seo, D. J. Seo, W. L. Yoon and
1226. S. B. Park, Appl. Catal., A, 2008, 340, 183–190.
´
44 M. Araque, L. M. Martınez, T. , J. C. Vargas and A. C. Roger, 50 J. W. Lee and C. Z. Li, Carbon, 2008, 46, 1208–1217.
Catal. Today, 2011, 176, 352–356.
31154 | RSC Adv., 2014, 4, 31142–31154
This journal is © The Royal Society of Chemistry 2014