Methane catalytic combustion over CeO2-ZrO2-Sc2O3 mixed oxides

Article abstract of DOI:10.1016/j.apcata.2019.117235

In this work nanostructured CeO₂-ZrO₂-Sc₂O₃ mixed oxides prepared by citrate complexation method are studied. The effects of the co-substitution of ZrO₂ and Sc₂O₃ in the CeO₂ lattice are assessed in terms of the reducibility of the samples and their catalytic activity towards oxidation of methane. In situ X-ray absorption near edge structure (XANES) and X-ray powder diffraction (XPD) experiments with synchrotron light were performed to study the oxygen exchange capacity of these ternary oxides in reaction conditions and their structure stability. Catalytic experiments for complete methane combustion indicated a superior performance of samples containing Sc₂O₃ compared to the CeO₂-ZrO₂ one in terms of reaction onset temperature, conversion values and activation energy. Samples with 4 at% and 6 at% Sc doping exhibited a stable behavior at 800 °C during 7.5 h on-stream experiments. Structural stability was confirmed by in-situ XPD experiments with no secondary phase segregation or carbon deposition in reaction conditions.

Full text of DOI:10.1016/j.apcata.2019.117235

Applied Catalysis A, General 587 (2019) 117235
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Methane catalytic combustion over CeO -ZrO -Sc O mixed oxides
2
2
2 3
Lucía M. Toscania,b,⁎, Pablo A. Curyk , M. Genoveva Zimicz , Emilia B. Halac ,
c
d
e
f,g
h
a,b
Martín. E. Saleta , Diego G. Lamas , Susana A. Larrondo
a
UNIDEF-CONICET-MINDEF, Departamento de Investigaciones en Sólidos, J. B. De La Salle 4397, 1603, Villa Martelli, Pcia. de Buenos Aires, Argentina
Instituto de Investigación e Ingeniería Ambiental, UNSAM, Campus Miguelete, 25 de Mayo y Francia, 1650, San Martín, Pcia. de Buenos Aires, Argentina
Departamento de Ingeniería Química, Facultad de Ingeniería, Universidad de Buenos Aires, Pabellón de Industrias, Ciudad Universitaria, 1428, Buenos Aires, Argentina
Departamento de Física, Universidad Nacional del Sur, Instituto de Física del Sur, IFISUR-CONICET, Av. L.N. Alem 1253, 8000, Bahía Blanca, Pcia. de Buenos Aires,
b
c
d
Argentina
e
Departamento de Física de la Materia Condensada, Gerencia de Investigación y Aplicaciones, Centro Atómico Constituyentes, Comisión Nacional de Energía Atómica, Av.
General Paz 1499, 1650, San Martín, Pcia. de Buenos Aires, Argentina
f
Centro Atómico Bariloche (CNEA) and INN (CNEA-CONICET), Av. Ezequiel Bustillo 9500, 8400, S.C. de Bariloche, Argentina
Instituto Balseiro, Universidad Nacional de Cuyo, CNEA, Av. Ezequiel Bustillo 9500, 8400, S.C. de Bariloche, Argentina
CONICET and Laboratorio de Cristalografía Aplicada, Escuela de Ciencia y Tecnología, UNSAM, Campus Miguelete, 25 de Mayo y Francia, 1650, San Martín, Pcia. de
g
h
Buenos Aires, Argentina
A R T I C L E I N F O
A B S T R A C T
Keywords:
2 2 2 3
In this work nanostructured CeO -ZrO -Sc O mixed oxides prepared by citrate complexation method are stu-
CeO
2
-ZrO
2
2 3
died. The effects of the co-substitution of ZrO2 and Sc O in the CeO2 lattice are assessed in terms of the re-
Methane oxidation
In-situ XANES
In-situ XPD
ducibility of the samples and their catalytic activity towards oxidation of methane. In situ X-ray absorption near
edge structure (XANES) and X-ray powder diffraction (XPD) experiments with synchrotron light were performed
to study the oxygen exchange capacity of these ternary oxides in reaction conditions and their structure stability.
Catalytic experiments for complete methane combustion indicated a superior performance of samples containing
2 3 2 2
Sc O compared to the CeO -ZrO one in terms of reaction onset temperature, conversion values and activation
energy. Samples with 4 at% and 6 at% Sc doping exhibited a stable behavior at 800 °C during 7.5 h on-stream
experiments. Structural stability was confirmed by in-situ XPD experiments with no secondary phase segregation
or carbon deposition in reaction conditions.
1. Introduction
combustion as it is the hydrocarbon with the largest H:C ratio in its
chemical structure, becoming highly stable and hard to oxidize.
In the past few years, natural gas (NG) catalytic combustion became
Therefore, to allow for its combustion at T < 800 °C an active catalyst
a promising alternative to conventional fuel thermal combustion to feed
either gas turbines as for NG fed vehicles. NG catalytic combustion
allows the reduction of gaseous contaminants, especially NO due to the
x
is required to break-up the C–H bonds in its surface and, at the same
time, allow for the activation of the O
place [3].
2
molecule for the reaction to take
lower operating temperatures (T < 800 °C) required in contrast to
flame combustion [1,2]
Methane is, however, a strong greenhouse gas and great effort is
being made to reduce unburnt methane emissions from the exhaust of
combustion devices. In this sense, catalytic oxidation of methane has
emerged as a fundamental tool to reduce emissions.
Taking into account sustainable alternatives to fossil fuels, methane
can be obtained from the metabolism of anaerobic microorganisms
during organic matter decomposition in what is known as biogas.
Methane molecule requires high temperatures for its homogeneous
In spite of the fact that hydrocarbon oxidation mechanism over the
surface of the catalyst is complex, it is widely accepted that the acti-
vation process proceeds through hydrocarbon dissociative adsorption
followed by a rapid sequence of reactions between alkyl group and
uncoordinated oxygen species over the surface. In the particular case of
methane oxidation, two mechanisms are frequently called upon: one at
low temperatures involving chemisorbed oxygen (Langmuir-
Hinshelwood) and other at high temperatures involving oxygen in-
coming from the catalyst lattice (Mars-van-Krevelen) [3–5].
Traditionally, supported noble metals and transition metal oxides
⁎
Corresponding author. Present address: INN-CONICET-CNEA, Departamento de Caracterización de Materiales, Centro Atómico Bariloche (CAB-CNEA), Av.
Ezequiel Bustillo 9500, 8400, S.C. de Bariloche, Argentina.
E-mail address: lucia.toscani@cab.cnea.gov.ar (L.M. Toscani).
https://doi.org/10.1016/j.apcata.2019.117235
Received 24 July 2019; Received in revised form 28 August 2019; Accepted 29 August 2019
Available online 16 September 2019
0926-860X/ © 2019 Elsevier B.V. All rights reserved.

L.M. Toscani, et al.
Applied Catalysis A, General 587 (2019) 117235
have been used as oxidation catalysts. Noble metals are the most ac-
tives, especially Pd and Pt, highly active to oxidize methane molecule
with synchrotron light were performed to study the oxygen exchange
capacity in reaction conditions and the structural stability of these
ternary oxides.
[
6–8]. However, their high cost, low thermal stability and low re-
sistance to sulphur poisoning, have motivated the development of new
catalysts. In this context, supported transition metal oxides, although
less active, are more resistant in demanding operating conditions and
are cost-effective compared to noble metals [3]. In particular, several
types of oxides have been used such as simple metal oxides like NiO,
2. Experimental
2.1. Catalyst preparation
Co
drotalcite-derived oxides such as Ni-Mg/Al or Co-Mg/Al oxides [9–14].
In this context, CeO has been widely studied due to its redox
properties and oxygen storage capacity. CeO has been used as a tex-
tural and structural promoter in supported noble and transition metal
catalysts and in mixed oxides doped with rare earth and transition
3
O
4
, Mn
2
O
3
; perovskite-type mixed oxides such as LaMnO
3
and hy-
x
Ce0.9Sc Zr0.1−xO2−δ (x = 0, 0.02, 0.04, and 0.06) mixed oxides
were synthesized from nitrate precursors following the citrate route.
Details about the synthesis procedure and sample characterization were
described in a previous work [23]. Finally, samples were fired at 800 °C
for 2 h in static air. Samples are referred hereinafter as CeZrScx, with x
2
2
the Sc atomic content (e.g. CeZrSc6 stands for Ce0.9Zr0.04Sc0.06O1.97
)
metals [6]. However, CeO
tures, being sintering the main cause of deactivation of the material due
to active site loss triggered by grain growth. ZrO addition to the
structure of CeO prevents sintering and improves redox properties of
the material [15,16]. In this sense, numerous catalysts based in CeO
and CeO -ZrO have been tested for methane catalytic combustion.
4,17–21].
Bozo et al. studied Ce/Zr mixed oxides prepared by coprecipitacion
method and obtained highest conversion rates for Ce:Zr ratio of 83:17
22]. Pengpanich et al. prepared CeO and CeO -ZrO (Ce:Zr = 75:25)
2
morphology is not stable at high tempera-
2.2. Catalyst characterization
2
2
The crystal structure of the samples was studied by means of con-
ventional X-ray powder diffraction (XPD) technique with a Philips
PW3710 diffractometer operated with Co-Kα (λ =1.7905 Å) radiation
and a graphite monochromator at 40 kV and 30 mA. Data was collected
in the angular region of 25–90 ° with a step size of 0.05 and a time per
step of 15 s.
2
2
2
[
[
2
2
2
2
Specific surface area was evaluated by means of N -physisorption
oxides by an hydrolysis route and concluded that catalytic activity for
methane oxidation is more dependent on structure and redox properties
with a Quantachrome Corporation Autosorb-1 equipment. Samples
were previously degassed with He at 90 °C during 12 h. Results were
obtained using the five point Brunauer − Emmett − Teller (BET)
method with the adsorption points of the isotherm in the following
relative pressure range: 0.05 − 0.35. Average particle size was esti-
rather than the BET surface area [18]. Larrondo et al. prepared CeO
ZrO mixed oxides by gel-combustion method and found and an op-
timum Ce:Zr ratio of 90:10 in terms of catalyst activity [19]. Zimicz
et al. studied the behavior of Ce0.9Zr0.1 for total oxidation of methane
2
-
2
−
1
−1
O
2
mated using the following expression DBET = 6ρ
SBET , where ρ is
using different amino acids in the gel-combustion synthesis and found a
marked influence of the final sample textural properties in catalytic
activity [15].
the theoretical density of the sample obtained from X-ray powder dif-
fraction (XPD) patterns after Rietveld refinements and SBET is the spe-
cific surface area.
All the aforementioned works correspond to stoichiometric oxides
in which ZrO was added to the CeO lattice as a textural and redox
2 2
The morphology of the samples was studied by means of scanning
electron microscopy (SEM) technique. SEM images were obtained with
a Zeiss Electron Beam SEM-Supra 40 placing the powdered samples
over an adhesive carbon-filled conductive tape to avoid charging pro-
blems.
promoter. In contrast, the incorporation of an aliovalent dopant can
alter the crystal lattice increasing the oxygen vacancy concentration,
thus improving oxygen mobility and catalytic activity. In line with this,
Palmqvist et at. studied the activity of CeO
2
oxides doped with Ca2+
,
Hydrogen temperature programmed reduction (TPR) experiments
were performed in a Micromeritics Chemisorb 2720 equipment to study
sample reducibility. The mass employed for each experiment was of
80 mg. Prior to TPR tests, samples were pretreated in He at 300 °C
during 30 min to remove any adsorbed species on the solid surface. TPR
3
+
3+
Nd
and Mn
[4]. In all cases, a promoting effect in methane con-
version was found upon doping, which the authors explained by cor-
relating oxide ion mobility in the lattice with catalytic activity. In
particular, they pointed out that one of the stages of the oxidation of
methane is oxygen transport from the lattice to allow the formation of
3
−1
was carried out with a 50 cm (STP) min
(5 vol.% H
heating ramp of
Hydrogen uptake was estimated using Thermal
Conductivity Detector (TCD) previously calibrated.
2
/Ar) flow from
CO
2
and H
2
O and, especially if this becomes the rate determining step,
room temperature up to 800 °C following a
−1
then a higher oxygen mobility in the solid can significantly favor cat-
alytic activity in the solid [4]. The addition of an aliovalent dopant to
the structure generates oxygen vacancies, exposing low coordination
cerium cations. In this case, the presence of defects enhances oxygen
mobility in the lattice, promoting its participation in oxidation reac-
tions.
10 °C min
.
a
Raman spectroscopy experiments were performed with a LabRam
HR Raman system (Horiba Jobin Yvon) equipped with two mono-
−
1
chromator gratings and a CCD detector. Spectral resolution of 1.5 cm
−1
was achieved by means of an 1800 g.mm grating and a 100 μm hole.
The spectrometer is coupled to a microscope equipped with 10x, 50x
and 100x magnification objectives. A He-Ne laser line at 632.8 nm was
employed as an excitation source and several measurements were car-
ried out to ensure minimal heating and no alteration of the sample.
In a previous work, we studied nanocrystalline CeO
oxides fired at 500 °C doped with Sc in terms of its structural
characterization and redox behavior. The main aim was to combine the
improved thermal stability provided by the ZrO with an increase in
2 2
-ZrO mixed
2 3
O
2
vacancy concentration upon Sc doping. Our results indicated that
homogeneous samples with superior redox behavior were achieved
with Sc incorporation [23]. However, no further studies were per-
formed in more demanding conditions such as higher firing tempera-
tures and methane oxidation reaction conditions.
In light of these results, the aim of this study is to analyze the effect
2 3 2 2
of adding Sc O to the CeO -ZrO lattice in catalytic activity toward
complete methane combustion. Structural, morphological and vacancy
concentration studies were performed to characterize the samples fired
at 800 °C. Furthermore, in-situ dispersive X-ray absorption near edge
structure (XANES) and X-ray powder diffraction (XPD) experiments
2.3. Catalytic activity assessment
Total oxidation of methane experiments were performed in a con-
ventional quartz fixed bed reactor operated isothermally at atmospheric
pressure. Temperature during experiments was monitored with a K-type
thermocouple axially placed in the center of the catalytic bed. The mass
of catalyst used in all tests was of 75 mg diluted with inert material in a
1:6 mass ratio to prevent the formation of hot spots. The reaction
mixture consisted of 1 vol.% CH
4
, 8 vol.% O
2
and 91 vol.% N
2
. Total
3
−1
feed Flow rate was of 300 cm min
set with mass flow controllers.
2

L.M. Toscani, et al.
Applied Catalysis A, General 587 (2019) 117235
The experiments were performed in the 400 °C–800 °C temperature
range, with steps of 50 °C. For each temperature, successive measure-
ments were performed until two consecutive analysis were identical.
Preliminary catalytic tests were performed to ensure that the con-
tribution of homogeneous phase reaction was negligible in the oper-
ating conditions and that there were no mass or heat transfer limita-
tions. All catalytic tests were performed with constant contact time
normalization was carried out. The fittings were performed using a
linear combination of the spectrum of Ce(NO ) , as a Ce(III) standard,
3 3
and the initial spectrum of each sample. It is important to note that by
following this procedure we obtain the degree of reduction achieved
and not the absolute Ce(III) content.
2.5. In situ XPD experiments
−
1
(
τ = 0.25 mg.min.ml ). Feed and exit compositions were determined
by on-line gas chromatography with a Perkin Elmer Clarus 500 Gas
chromatograph, equipped with an automatic injection valve, a CTR1
column and a Temperature Conductivity Detector (TCD), using He as a
carrier.
Time-resolved X-ray powder diffraction data were acquired at the
D10B-XPD beamline facility of the Brazilian Synchrotron Light
Laboratory (LNLS, Campinas, Brazil: Proposal XPD-20150256) using a
high-intensity and low-resolution configuration, without crystal ana-
lyzer and a Mythen 1 K detector [26]. The X-ray wavelength was set at
1.5498 Å. Samples were placed on the sample holder inside a furnace
Thermal stability was assessed in typical 8 h on-stream experiments
at 800 °C to evaluate carbon deposition and evolution of catalytic ac-
tivity. Methane conversion was calculated using the following expres-
with temperature control. In situ XPD patterns were collected upon
in
out
in
in
−1
sion: xCH4=(FCH4-FCH4)/(FCH4), where FCH4 represents the molar flow
heating following a ramp of 10 °C.min
in a mixture of 4 mol % CH
4
out
rate of methane in the inlet while FCH4 the methane molar flow rate in
and 8 mol % O (He balance) from room temperature to 800 °C with
2
the outlet. Complete selectivity to CO
balance matched ± 3% in all cases.
2
was always observed. Carbon
isothermal steps at 25, 650, 750 and 800 °C. Diffraction patterns were
recorded in the angular region 2θ = 20 − 80° during isothermal steps.
Rietveld refinements of these patterns were performed using FullProf
Suite open access software [27]. In addition, short diffraction patterns
were collected in the angular region 2θ = 23 − 37° in order to monitor
structural evolution upon heating. This 2θ range allowed us to calculate
Kinetic studies were performed using the same experimental setup
previously described, at 550 °C, 600 °C and 650 °C to estimate methane
partial order and the apparent activation energy of the reaction for all
samples. In these experiments, volumetric concentrations of methane in
the feed were 1%, 2%, 3% and 4%while oxygen concentration (8%) and
contact time were not modified. Specific rate was calculated following
the initial rate method for conversion values lower than 30%. Methane
partial order was studied assuming a power law kinetics. Finally,
Arrhenius plots were calculated following a pseudo first order kinetic
mechanism plotting ln[−ln(1−x)] vs 1/T, with x the methane con-
version. All samples exhibited a linear behavior, confirming the ex-
cellent agreement with the pseudo first order kinetic model, thus en-
abling to calculate the apparent activation energy for complete
methane combustion from the slope of the linear profiles.
2
the crystallite size upon heating, from the broadening of CeO (111)
Bragg peak, visible at ca. 28.4°, without exceeding a difference in 10 °C
during the data acquisition of a single scan. In addition, the main peak
2 3
of Sc O is apparent at 2θ = 31.4°. Therefore, any secondary phase
segregation which might take place at any temperature can be detected
in this region. Finally, any graphitic carbon that might be deposited as a
secondary reaction of methane oxidation would be visible at ca. 26°.
3. Results and discussion
3.1. Catalyst characterization
2.4. In situ XANES experiments
XPD patterns are presented in Fig. 1. Results showed that all sam-
In-situ X-ray absorption near-edge structure (XANES) experiments
ples exhibited a cubic fluorite-type crystal structure characteristic of
CeO -rich mixed oxides. No secondary phases of Sc or ZrO were
detected either after synthesis or after catalytic tests. Sample crystallite
sizes (DXPD) calculated from (111) Bragg peak are presented in Table 1,
displaying a value of ca. 10 nm and showing no differences with in-
were performed in the Ce LIII-edge at the D06A-DXAS dispersive
beamline of the Brazilian Synchrotron Light Laboratory (LNLS,
Campinas, Brazil, Proposal: 20,160,090) [24]. A Si (111) mono-
chromator was used altogether with a CCD detector to collect the ab-
sorption spectra in transmission mode. Self-supporting discs were pre-
pared by mixing the sample powder with boron nitride, which has no
significant absorption in the measured energy ranges. The discs were
located in a sample-holder, with a thermocouple attached to it, inside a
furnace with temperature control. The sample holder was placed in a
quartz reactor with inlet and outlet gas lines. Inlet gas composition was
set with a gas-mixing station provided with mass flow controllers and
exit composition was assessed with a Pfeiffer Omnistar mass spectro-
meter.
2
2
O
3
2
creasing Sc
Results from N
Surface area values were in the range of 35-48 m g
2
O
3
content.
2
-physisorption experiments are indicated in Table 1.
2
−1
and estimated
particle sizes were of ca. 20 nm for all samples. Although no significant
differences were evidenced among samples, the degree of agglomera-
tion (DBET/DXRD) resulted slightly larger for sample CeZrSc4.
In situ catalytic tests were performed to assess sample activity to-
wards total oxidation of methane (TOM) using a reactor inlet flow
3
−1
3
−1
consisting of a mixture of: 1.6 cm . min
1
O
2
, 0.8 cm . min CH
:CH = 2) corresponds to the
stoichiometric one for TOM. Measurements were performed at the
4
and
3
−1
7.6 cm . min He. Molar feed ratio (O
2
4
−
1
temperature range: 25–800 °C with a heating of 10 °C.min
and a
dwell time of 30 min at 600 °C, 700 °C and 800 °C. Although samples
were diluted with boron nitride to limit sample absorbance, care was
taken to maintain spatial time used during conventional catalytic ex-
−1
periments (τ = 0.25 mgcatalyst.min. ml ).
Data analysis was performed by means of the software Athena in-
cluded in the IFEFFIT software package (Version 1.2.11) [25]. A linear
combination fitting of standard spectra was carried out in order to
evaluate the fractions of reduced and oxidized species in the samples
during the course of the experiments. Before proceeding with the linear
combination fitting, background subtraction and spectrum
Fig. 1. XPD patterns of samples fired at 800 °C.
3

L.M. Toscani, et al.
Applied Catalysis A, General 587 (2019) 117235
Table 1
Results from XPD, N
2 2
-physisorption and H -TPR experiments.
1
(m2 g−1
2
uptake (mmol g−1
LT/IHT3
Sample
D
XPD (nm)
S
g
)
D
BET (nm)
DBET/DXPD
Total H
2
)
I
Reduction (%)
CeZr
10.2
10.7
10.4
10.8
48
43
35
46
18
20
24
19
1.7
1.8
2.3
1.7
0.9
1.0
1.2
1.2
1.2
1.1
1.0
0.9
37
38
47
45
CeZrSc2
CeZrSc4
CeZrSc6
1
Calculated using Scherrer equation with peak (111).
Calculated from specific surface area.
Ratio of intensities of low temperature (LT) and high temperature (HT) peaks from TPR profiles.
2
3
Fig. 2. SEM micrographs of sample CeZrSc2 fired at 800 °C.
−
1
SEM image analysis indicated no significant differences in sample
On the other hand, the signal evidenced at 600 cm is attributed to a
second order dispersion mode caused by the presence of oxygen va-
cancies in the lattice [30]. These so called intrinsic oxygen vacancies
are present in oxides with crystallite sizes in the nanometric domain
due to the large area/volume ratio of the samples.
morphology with Sc incorporation. SEM micrographs of sample
CeZrSc2 are exhibited in Fig. 2, allowing a clear identification of small
particles of mean diameter DSEM = 13 nm on the walls of platy struc-
tures.
−1
TPR profiles are depicted in Fig. 3, evidencing in all cases the two
In Fig. 4b a detail of the region 500-700 cm is displayed enlarged.
−
1
3+
characteristic peaks from CeO
2
reduction. The first peak is usually as-
2
Besides the peak at 600 cm , upon Sc incorporation into the CeO -
ZrO
2
structure, an additional mode becomes apparent at 550 cm 1.
−
cribed to the reduction of surface Ce whilst the high temperature peak
is generally associated to bulk Ce reduction [28]. Results show that,
upon incorporation of 2% Sc, reduction onset temperature increases,
being this tendency reverted with higher Sc contents. In fact, CeZrSc4
and CeZrSc6 exhibited the lowest onset temperatures and highest total
reduction values as shown in Table 1. Therefore, as sample surface area
is virtually the same in all cases and morphology studied by SEM ex-
hibited no variations whatsoever, the changes in the redox behavior can
This mode is usually ascribed to the presence of extrinsic oxygen va-
cancies generated by aliovalent doping to compensate charge un-
balance in the lattice. Therefore, it is clear that there is an increase in
sample defects with Sc addition. These results, along with the fact that
2 3
the modes corresponding to Sc O are absent in the Raman profile,
confirm scandium incorporation into the lattice of the mixed oxide.
The main F2g mode frequencies are displayed in Table 2 together
with the (I550+I600)/IF2g ratio for each sample. This last value allows a
straightforward comparison of the amount of both intrinsic and ex-
trinsic oxygen vacancies in the lattice in the different samples [32].
oxides doped with Sc
larger amount of oxygen vacancies in contrast to undoped CeO
being this value maximum for CeZrSc6 sample.
To conclude this characterization section, it is worth pointing out
that Raman and TPR results indicate that Sc3+ incorporation to the Ce-
Zr lattice allows the formation of additional defects and an increase in
oxygen mobility, promoting sample reducibility, especially in samples
CeZrSc4 and CeZrSc6. These properties are of great interest for good
oxidation catalysts.
2 3 2 2
be assigned to the incorporation of Sc O to the CeO -ZrO lattice. It is
worth noticing that the intensity of the low temperature peak exhibits a
slight increase in Sc containing samples compared to the CeZr sample,
whereas the intensity of the high temperature peak significantly in-
creases upon Sc incorporation. In fact, the ratio of the low temperature
to high temperature peak intensity (ILT/IHT) decreases continuously
with Sc incorporation to the point of being this ratio lower than 1 in the
case of CeZrSc6 sample as portrayed in Table 1 and Figure S1. These
Results indicate that CeO
2
-ZrO
2
2
O
3
present a
-ZrO
2
2
,
2 3 2 2
results could indicate that the addition of Sc O to the CeO -ZrO
system effectively increases oxygen mobility in the bulk of the lattice
thus increasing the contribution of bulk Ce reduction in the ternary
samples.
Raman spectra obtained in ambient conditions for all samples are
displayed in Fig. 4a. The main Raman F2g mode observed at ca.
3
.2. Catalytic activity assessment
−1
4
68 cm is characteristic of fluorite type oxides and is attributed to the
symmetric stretching of oxygen around cerium cations in the CeO vi-
brational unit [29,30]. Additional features can be observed clearly in
8
Catalytic activity results for TOM are presented in Fig. 5. The light-
off curves indicate that all samples were active for methane combustion
above 550 °C. Samples containing Sc exhibited superior perfor-
mance than Ce-Zr sample in the whole temperature range. Temperature
values for 50% conversion of methane (T50) are exhibited in Table 2,
being the conversion profiles of CeZrScx samples shifted towards lower
−1
-1
−1
the inset of Fig. 4a at 260 cm , 600 cm and 1170 cm . The peaks
2 3
O
-
1
-1
observed at 260 cm and 1170 cm are related to second order trans-
verse modes which indicate the presence of defects in the crystal
structure and are characteristic of nanostructured fluorite oxides [31].
4

L.M. Toscani, et al.
Applied Catalysis A, General 587 (2019) 117235
Table 2
Results from Raman spectra analysis.
Sample
F2g band position (cm−1)
(I550+I600)/IF2g (%)
CeZr
470
470
468
468
2.9
3.8
4.6
4.8
CeZrSc2
CeZrSc4
CeZrSc6
4 4
Fig. 5. CH conversion as a function of reaction temperature (1 vol.% CH in
−
1
8
2
vol.%O , τ = 0.25 mg.min.ml ).
toward oxidation of methane.
The effect of methane partial pressure on the reaction rate is pre-
sented in Fig. 6. In all cases a linear dependency in the ln rate vs ln pCH4
plot was evidenced, indicating that the experimental power law type
kinetic model is valid. Based on this model all samples exhibited a mean
reaction order with respect to methane comprised between (0.8–1).
Complete methane catalytic combustion has been reported to follow
a pseudo first order kinetic law [3,4,6,18]. Reaction mechanism is
complex and involves the competition for active sites and several ad-
2
Fig. 3. H -TPR profiles plotted as (a) TCD signal vs temperature and (b) Degree
of Ce reduction vs temperature.
sorption/desorption equilibria of CO, CO
2 2 4 2
, H O, CH and O molecules
temperatures when compared to the one of CeZr sample. In all ex-
periments, the only product detected was CO and the carbon balance
2
[
4]. However, reaction rate equation can be assumed to follow a pseudo
first order kinetic model assuming fast oxygen supply from gas phase to
between feed and exit flows matched 100 ± 3%. In addition, no evi-
dence of carbon formation was detected by XPD experiments or SEM
images of spent catalysts.
The activity rates assessed using the initial rate method (CH con-
4
version < 30%) are presented in Table 3. All Sc-containing samples
exhibited higher rates than the CeZr one, indicating a superior capacity
replace the oxygen incoming from the lattice and low site occupation
with CO
2 2 4
and H O. These conditions can be fulfilled if CH con-
centration in the feed is low and high temperatures are used [4].
Arrhenius plots for TOM are presented in Fig. 7. In all cases a linear
dependency was evidenced, confirming the pseudo first order kinetic
−
1
and 1170 cm−1 are
Fig. 4. (a) Raman spectra of all samples fired at 800 °C acquired at room temperature. In the inset the smaller contributions at 260 cm
−
1
−1
highlighted (b) region close to peaks at 550 cm
and 600 cm
enlarged.
5

L.M. Toscani, et al.
Applied Catalysis A, General 587 (2019) 117235
Table 3
2 2 2 3
Catalytic activity for total oxidation of methane of CeO -ZrO -Sc O samples studied.
Sample
T50(°C)
Specific rate (107 mol g−1 s-1
)
Intrinsic rate (107 mol m−2 s-1
)
Ea
At 550 °C
6.2
8.5
10.8
9.4
At 600 °C
19.2
24.8
29.4
27.3
At 650 °C
48.0
58.4
72.1
67.2
At 650 °C
1.0
1.4
2.1
1.5
(Kj mol−1)
141.2
129.3
124.3
127.3
CeZr
718
703
700
700
CeZrSc2
CeZrSc4
CeZrSc6
Fig. 8. Methane conversion vs. time-on stream during stability tests at 800 °C
−
1
(
4 2
4 vol.% CH in 8 vol.% O , τ = 0.25 mg.min.ml ).
Fig. 6. Reaction rate (r) vs pCH4at 600 °C and 650 °C for samples (a) CeZr,
b)CeZrSc2, (c) CeZrSc4 and (d) CeZrSc6 at constant PO2 = 0.08 atm.
(
especially in samples CeZrSc4 and CeZrSc6 that stand out as a pro-
mising alternative for their application in oxidation reactions.
Fig. 8 exhibits the effect of time on stream on catalytic activity in a
more concentrated methane feed, consistent with the stoichiometric
2 4
O :CH ratio of 2 for complete methane oxidation. After 2 h stabiliza-
tion period, stable activity was observed for samples CeZrSc4 and
CeZrSc6 in contrast to CeZrSc2 and CeZr. Sample CeZrSc6 exhibited no
deactivation whatsoever in the whole period assessed, indicating that
the addition of Sc to the system not only improves methane conversion
values but contributes to attain stable catalytic activity.
SEM micrographs of the spent sample CeZrSc4 are exhibited in
Fig. 9. These micrographs indicate that after stability tests during 8 h at
800 °C in reaction conditions some morphological changes took place.
These images are representative of all samples studied and reveal that
although large particles preserve its shape and size, crystallites forming
part of the walls of these particles do exhibit some size growth. This
growth is consistent with the results observed in XPD results obtained in
situ (Section 3.5) where a growth from 9 nm from room temperature to
13 nm at 800 °C is found to take place.
2 2 2 3 4
Fig. 7. TOM Arrhenius plots for CeO -ZrO -Sc O samples (1 vol.% CH in
−
1
8
2
vol.%O , τ = 0.25 mg.min.ml ).
model for methane oxidation. Apparent activation energies calculated
from the slopes of linear fittings are displayed in Table 3, indicating a
decreasing tendency upon incorporation of Sc into the Ce-Zr lattice.
Palmqvist et al. obtained values of apparent activation energy for TOM
3.3. In situ XANES experiments
In situ XANES experiments at the Ce LIII-edge were performed in
order to study the oxidation state of the Ce present in the samples in
reaction conditions. Two spectra corresponding to sample CeZrSc6 ac-
quired at 25 °C and 800 °C in reaction conditions are exhibited in
Fig. 10. It is clear that ceria is only faintly reduced at the end of the
experiments, as seen in the shift of the energy edge.
−
1
in the range of 110–130 kJ.mol
Mn, Pb y Nd fired at 600 °C [4]. On the other hand, Pengpanich et al.
calculated the activation energy for Ce0.75Zr0.25 sample fired at
00 °C in 100.8 kJ.mol . In this sense, the values of activation energy
for Ce mixed oxides doped with Ca,
O
2
−
1
5
−
1
obtained in the present work (127–141 kJ mol ) fall in the range ex-
pected for moderately active oxides [4, HYPERLINK \l "Ref18″ \o "[18]
M.G. Zimicz, I.O. Fabregas, D.G. Lamas, S.A. Larrondo, Materials Re-
search Bulletin" \h 18]. It should be pointed out that reported values
from the other authors correspond to samples fired at lower tempera-
tures, 500 and 600 °C in contrast to the 800 °C from samples studied in
this work. Activity results clearly indicate a promoting effect of Sc,
The results from XANES profile fittings are plotted against tem-
perature in Fig. 11a, while Fig. 11b displays mass spectrometry data
collected during experiments. These results indicate a very low degree
of Ce reduction (αCe < 3%) in reaction conditions. However, in all cases
a stepwise increase in the Ce reduction value was observed upon tem-
perature raise. The fact that methane conversion boost is accompanied
by the reduction of the Ce present in the support can be explained by
the Mars-van Krevelen reaction oxidation mechanism involving in the
6

L.M. Toscani, et al.
Applied Catalysis A, General 587 (2019) 117235
4 2
-8%mol O
(τ = 025 mg min cm−3) and (c) fresh
Fig. 9. SEM micrographs of sample CeZrSc4 (a), (b) after stability tests at a T =800 °C during 8 h in 4 mol % CH
sample fired at 800 °C in air.
first stage an O2 uptake from the oxide lattice followed by a replen-
−
predominant, it is generally assumed that active sites for methane
dissociation over CeO -based catalysts are the exposed oxygen atoms
with low coordination (3 fh sites) on the (111) surface [33]. Methane
dissociation is followed by the formation of a CH radical and an ad-
sorbed H* atom. These two species bind to two surface oxygen atoms,
leading to the formation of a methyl radical (CH -) and a hydroxyl
radical (HO-). Hydrogen adsorption leads to the reduction of one of the
neighboring Ce atoms. Afterwards, a series of intermediate steps take
3 2
place in which H atoms are abstracted from CH * to form CH * and CH*
ishment stage from the gas phase [20].
2
Results reported in this work confirm Ce participation in the redox
mechanism for TOM. In particular, sample CeZrSc6 exhibits a rapid
stabilization of the degree of reduction indicating that the dynamic
equilibrium between reduction and oxidation of Ce cations is rapidly
achieved [32]. In contrast, sample CeZrSc4 exhibits a continuous oxi-
dation with increasing temperature after the initial boost in the degree
of reduction. This behavior can be explained by the different capacity of
the samples to deliver oxygen form its lattice. It should be noted that
3
3
until CO is formed by further reducing the Ce with the formation of an
oxygen vacancy [8,33]. Finally, adsorbed CO reduces Ce by binding
with another oxygen allowing the formation of CO . Water is formed by
2
among different aliovalent dopants incorporated to the CeO
2
lattice in
3
+
the literature, Sc
has a large association enthalpy to oxygen va-
cancies, being therefore an optimum in the concentration of oxygen
vacancies that can be generated in order to enhance ionic conductivity,
the binding of adjacent OH groups. The breaking of the C–H bond is
generally the rate-limiting step in all methane activation processes
[33].
The results reported here coincide in that the dynamic equilibrium
in oxygen vacancy concentration involved in the reaction mechanism is
key to understanding the activity and stability of the catalyst towards
methane oxidation.
3+
which is significantly smaller than with other +3 cations such as Gd
3
+
or Sm
[23]. In fact, as we discussed in Raman spectroscopy results,
3+
there is an increase in the amount of sample defects with Sc
in-
−1
corporation due to the presence of an additional mode at 550 cm in
the Raman profile, clearly resolved in the case of sample CeZrSc6. This
mode is usually ascribed to the presence of extrinsic oxygen vacancies
generated by aliovalent doping to compensate charge unbalance in the
lattice and indicates a large dopant-defect interaction that can be det-
rimental to the oxygen mobility in the lattice. Results from XANES
experiments indicate a faster re-oxidation rate in sample CerZrSc4 in
contrast to samples CeZr and CeZrSc6, behavior that can be ascribed to
the different oxygen mobility of the samples.
Mass spectrometry results (Fig. 11b) indicate significant CH
uptake above 600 °C with the concomitant production of CO and H
evidencing catalytic activity for TOM. CH and O
are more pronounced in the case of CeZrSc4 and CeZrSc6 samples
compared to CeZr sample. This tendency is also exhibited for CO and
O production signals. Results obtained from experiments in the
conventional laboratory fixed bed reactor in a feed composition of
4 mol % CH and 8 mol % O (N balance) (Figure S2) showed at 600 °C
values of CH conversion of 5–10% and of 75–90% at 800 °C, detecting
only CO as reaction product. Samples CeZrSc4 and CeZrSc6 exhibited a
4
and O
2
2
2
O,
4
2
consumption signals
2
H
2
Methane complete oxidation over CeO
2
has been reported to follow
4
2
2
the MvK mechanism, in which CH reduces de surface of the doped Ce
4
4
oxide followed by re-oxidation from gas phase oxygen [7,8,33].
Results reported in literature indicate that the first step is hydrogen
abstraction from the methane molecule over the Ce surface [7,8,33]. As
2
superior performance in the whole temperature range and carbon bal-
ance matched 100% ± 3% with no evidence of carbon deposits in the
SEM micrographs of spent catalysts.
(
2
111) surface of CeO is the most stable and, therefore, the most
Fig. 10. (a) XANES spectra acquired at 25 °C and 800 °C in reaction conditions for CeZrSc6 sample (b) The region near the absorption edge (E0) is zoomed.
7

L.M. Toscani, et al.
Applied Catalysis A, General 587 (2019) 117235
Fig. 11. Results from in-situ XANES experiments in LIII-Ce edge: (a) Ce re-
duction degree vs temperature in complete combustion reaction conditions
Fig. 13. Results from in-situ synchrotron XPD experiments in reaction condi-
tions (pO = 0.08 atm, pCH = 0.04 atm, τ = 0.25 mg.min.cm ): (a)
2 4
−
3
−3
(pO
2
= 0.08 atm, pCH
4
= 0.04 atm, τ = 025 mg.min.cm ), (b) Mass spectro-
metry signals during the in-situ catalytic tests.
Crystallite size evolution with temperature and (b) Lattice parameter evolution
with temperature for all samples.
3
.4. In situ XPD experiments
exhibited a crystallite size at room temperature of ca. 9 nm. This value
remained between 9–10 nm in the temperature range (25–650) °C.
Beyond 650 °C all samples exhibited crystallite growth, however, sam-
ples CeZr and CeZrSc6 displayed a steeper growth compared to sample
CeZrSc4. This sample proved to experience less sintering than the other
samples in reaction conditions.
In situ XPD experiments were performed in order to study the
structural stability of the samples in reaction conditions. Results are
plotted in Fig. 12 and indicate that all samples exhibited a cubic fluorite
type crystal structure and retained this structure throughout the whole
2 3 2
experiment up to 800 °C. No secondary phases of Sc O or ZrO were
detected, indicating structure stability in catalytic conditions. In addi-
tion, no signals of graphitic carbon were observed in agreement with
carbon balance results in catalytic experiments and SEM images of
spent catalysts, which showed no formation of carbonaceous species
during complete combustion experiments.
The broadening of the (111) Bragg peak upon heating, shown in
Fig. 12, evidences the increase of sample crystallites at temperatures
close-to or exceeding firing temperature. Crystallite size evolution with
temperature is presented in Fig. 13. Results indicate that all samples
Lattice parameters obtained from Rietveld refinement of patterns
recorded in reaction conditions in several temperatures are presented in
Fig. 13 and Table 4. It is noteworthy that sample CeZrSc4 exhibited a
slightly larger lattice parameter at room temperature; however, at
650 °C, 750 °C and 800 °C this sample displayed the lowest values of
lattice parameter in contrast to samples CeZr and CeZrSc4. It should be
noted that in this case there are three competing causes for lattice
parameter differences: firstly, the addition of a smaller cation such as
3+
Sc
2 2
to the CeO -ZrO lattice might account for a lattice contraction,
2 4
Fig. 12. In-situ synchrotron XPD patterns acquired in reaction conditions (pO = 0.08 atm, pCH
= 0.04 atm, τ = 0.25 mg min cm−3).
8

L.M. Toscani, et al.
Applied Catalysis A, General 587 (2019) 117235
Table 4
DXAS-20160090 and XPD-20150256 and Agencia Nacional de
Promoción Científica y Tecnológica (PICT 2016 N°1921, PICT 2015
N°3411 and PICT 2013 N° 1587). The authors would like to recognize
the staff of the XPD and DXAS beamlines and LQU from the LNLS for
providing assistance during the experiments. No competing interests to
declare.
In-situ XPD results in reaction conditions at different temperatures: Lattice
parameter (a) and thermal expansion coefficient (α).
α (10 K−1
1
6
Sample
a
25°C (Å)
a
650°C (Å)
a
750°C (Å)
a
800°C (Å)
)
CeZr
CeZrSc4
CeZrSc6
5.3892(1)
5.3894(6)
5.3910(7)
5.4380(0)
5.4367(5)
5.4405(0)
5.4466(8)
5.4461(9)
5.4500(6)
5.4514(2)
5.4506(7)
5.4555(2)
14.7
14.4
15.1
Appendix A. Supplementary data
1
Calculated from lattice parameters obtained from Rietveld refinement.
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcata.2019.117235.
secondly, the addition of an aliovalent dopant can enhance oxygen
vacancy generation, responsible for lattice expansion and, thirdly, re-
action conditions with the presence of CH
4
might trigger Ce4+ reduc-
References
tion with the concomitant oxygen vacancy generation and subsequent
lattice expansion. Results presented in the previous sections indicated
that sample CeZrSc4 exhibited a faster redox rate, displaying larger Ce
reduction values in TPR experiments and a faster oxidation rate in re-
action conditions during in situ XANES experiments. The fact that this
[1] P. Gelin, M. Primet, Appl. Catal. B 39 (2002) 1–37.
[
2] M. Hoffmann, S. Kreft, G. Georgi, G. Fulda, M. Pohl, D. Seeburg, C. Berger-Karin,
E.V. Kondratenko, S. Wohlrab, Appl. Catal. B 179 (2015) 313–320.
3] A. Urda, I. Popescu, T. Cacciaguerra, N. Tanchoux, D. Tirchit, I. Marcu, Appl. Catal.
A Gen. 464 (2013) 20–27.
[
3
+
4+
[4] A.E.C. Palmqvist, E.M. Johansson, S.G. Jaras, M. Muhammed, Catal. Letters 56
1998) 69–75.
sample showed the lowest Ce /Ce
ratio in reaction conditions
(
might account for the lowest values in the lattice parameter and total
expansion coefficient (Table 4).
[
5] P. Ciambelli, S. Cimini, S. De Rossi, M. Faticanti, L. Lisi, G. Minelli, I. Pettiti,
P. Porta, G. Russo, M. Turco, Appl. Catal. B 24 (2000) 243–253.
6] S. Thaicharoensutcharittham, V. Meeyoo, B. Kitiyanan, P. Rangsunvigit,
T. Rirksomboon, Cat. Comm. 10 (2009) 673–677.
[
4. Conclusions
[
7] J.B. Miller, M. Malatpure, Appl. Catal. A Gen. 495 (2015) 54–62.
[
8] W. Tang, Z. Hu, M. Wang, G.D. Stucky, H. Metiu, E.W. McFarland, J. Catal. 273
(
2010) 125–137.
CeO -ZrO -Sc O mixed oxides were prepared by citrate com-
2
2
2 3
[
9] Z. Jiang, J. Yu, J. Cheng, T. Xiao, M. Jones, Z. Hao, P. Edwards, Fuel Process.
plexation route. Ceria reduction values from TPR experiments and va-
cancy concentration from Raman tests both increase with increasing Sc
content, indicating superior redox behavior of doped samples. Structure
and morphology of samples fired at 800 °C remained stable in catalytic
conditions without phase segregation or significant particle agglom-
eration. In-situ XPD experiments with synchrotron radiation demon-
strated structural stability in operating conditions.
Catalytic activity was enhanced with increasing Sc content with an
excellent carbon balance match ( ± 3%), indicating that the catalysts
exhibit a marked tolerance to carbon formation. Regarding stability
tests, samples CeZrSc4 and CeZrSc6 exhibited a remarkably stable be-
havior during 8 h time-on-stream experiments.
Technol. 91 (2010) 97–102.
[10] G. Saracco, F. Geobaldo, G. Baldi, Appl. Catal. B 20 (1999) 277–288.
[
11] S. Cimino, S. Colonna, S. De Rossi, M. Faticanti, L. Lisi, I. Pettiti, P. Porta, J. Catal.
05 (2002) 309–317.
2
[
12] T. Shishido, M. Sukenobu, H. Morioka, M. Kondo, Y. Wang, K. Takaki, K. Takehira,
Appl. Catal. A Gen. 223 (2002) 35–42.
[13] F.F. Tao, J.J. Shan, L. Nguyen, Z. Wang, S. Zhang, L. Zhang, Z. Wu, W. Huang,
S. Zeng, P. Hu, Nat. Commun. 6 (2015) 7798.
[
14] X. Wang, Y. Liu, Y. Zhang, T. Zhang, H. Chang, Y. Zhang, Appl. Catal. B 229 (2018)
2–62.
[15] M.G. Zimicz, D.G. Lamas, S.A. Larrondo, Cat. Comm. 15 (2011) 68.
5
[
[
[
[
[
16] M.G. Zimicz, I.O. Fabregas, D.G. Lamas, S.A. Larrondo, Mater. Res. Bull. 850
(2011).
17] D. Terribile, A. Trovarelli, J. Llorca, C. De Leitenburg, G. Dolcetti, Catal. Today 43
(1998) 79–88.
18] S. Pengpanich, V. Meeyoo, T. Rirksomboon, K. Bunyakiat, Appl. Catal. 234 (2002)
Preliminary kinetic studies indicated higher reaction rates and
2
21.
lower apparent activation energies for methane oxidation for CeO
ZrO -Sc samples in contrast to the binary CeO -ZrO sample.
Apparent activation energy was minimum for sample CeZrSc4 with a
2
-
19] S.A. Larrondo, M.A. Vidal, B. Irigoyen, A.F. Craievich, D.G. Lamas, I.O. Fabregas,
G.E. Lascalea, N.E. Walsoe de Reca, N. Amadeo, Catal. Today 107 (2005) 53–59.
20] M.M. Pakulska, C.M. Grgicak, J.B. Giorgi, Appl. Catal. A Gen. 332 (2007) 124–129.
[21] A.J. Abreu, A.F. Lucredio, E.M. Assaf, Fuel Process. Technol. 102 (2012) 140–145.
22] C. Bozo, N. Guilhaume, E. Garbowski, M. Primet, Catal. Today 59 (2000) 33–45.
2
2
O
3
2
2
−
1
value of Ea = 1243 kJ.mol
.
[
In-situ XANES experiments allowed to confirm the participation of
[23] L.M. Toscani, A.F. Craievich, M.C.A. Fantini, D.G. Lamas, S.A. Larrondo, J.
Phys.Chem.C 120 (2016) 24165–24175.
the lattice in the redox mechanism for TOM as an increase in the Ce
2–
[24] J.C. Cezar, N.M. Souza-Neto, C. Piamonteze, E. Tamura, F. Garcia, E.J. Carvalho,
R.T. Neueschwander, A.Y. Ramos, H.C.N. Tolentino, A. Caneiro, N.E. Massa,
Mj. Martinez-Lope, J.A. Alonso, J.P. Itief, J. Synchrotron Rad. 17 (2010) 93–102.
[25] B. Ravel, M. Newville, J. Synchrotron Rad. 12 (2005) 537–541.
26] F. Furlan Ferreira, E. Granado, W. Carvalho, S.W. Kycia, D. Bruno, R. Droppa, J.
Synchrotron Rad. 13 (2006) 46–53.
reduction value (O delivered from the lattice) upon temperature rise,
followed by Ce re-oxidation (oxygen incorporated from the gas phase)
at constant temperature.
[
These results indicate a promoting effect of Sc
2 3
O incorporation in
the CeO -ZrO in terms of redox properties and catalytic activity.
2
2
[
27] J. Rodríguez-Carvajal, FULLPROF: a Program for Rietveld Refinement and Pattern
Matching Analysis, Satellite Meeting on Powder Diffraction of the XV IUCr Congress
(
1990) 127.
Declaration of Competing Interest
The authors declare no competing financial interest
Acknowledgements
[
28] P. Fornasiero, G. Balducci, R. Di Monte, J. Kaspar, V. Sergo, G. Gubitosa, A. Ferrero,
M. Graziani, J. Catal. 164 (1996) 173.
[29] S. Askrabic, Z.D. Dohcevic-Mitrovic, M. Radovic, M. Scepanovic, Z.V. Popovic, J.
Raman Spectrosc. 40 (2009) 650–655.
[
[
[
30] M. Guo, J. Lu, Y. Wu, Y. Wang, M. Luo, Langmuir 27 (2011) 3872–3877.
31] I. Iglesias, G. Baronetti, F. Mariño, Solid State Ion. 309 (2017) 123–129.
32] M.G. Zimicz, F.D. Prado, A.L. Soldati, D.G. Lamas, S.A. Larrondo, J. Phys. Chem. C
1
19 (2015) 19210–19217.
The doctoral scholarship granted by CONICET to Dr. Lucía M.
Toscani is gratefully acknowledged. This work was supported by the
Brazilian Synchrotron Light Laboratory (LNLS) under proposals D06-
[
33] D. Knapp, T. Ziegler, J. Phys. Chem. C 112 (2008) 17311–17318.
9