Upload
ngothu
View
214
Download
2
Embed Size (px)
Citation preview
Post-print of Journal of Physical Chemistry C, 2015, 1119, 16708-16723
Ruthenium Effect on Formation Mechanism and
Structural Characteristics of LaCo1-xRuxO3
Perovskites and its Influence on Catalytic
Performance for Hydrocarbon Oxidative Reforming
Noelia Motaa, Laura Barrioa†*, Consuelo Alvarez-Galvána, François Fauthb, Rufino M.
Navarroa*, Jose Luis G. Fierroa
a Institute of Catalysis and Petrochemistry, CSIC, Marie Curie 2, Cantoblanco, 28049 Madrid,
Spain
b Experiments Division CELLS-ALBA, 08290 Cerdanyola del Vallès, Barcelona, Spain
DOI: 10.1021/acs.jpcc.5b04287
KEYWORDS LaCoO3, Ruthenium, Perovskites, X-ray diffraction, Raman, EXAFS.
1
ABSTRACT This work deals with the formation mechanism of LaCo1-xRuxO3 perovskites (x = 0,
0.05, 0.1, 0.2 and 0.4). In situ characterization of perovskite during formation were monitored
with X-ray diffraction and Raman spectroscopy techniques, revealing that perovskite formation
occurs via an oxo-lanthanum carbonate intermediate phase. Structural characterization of
perovskites showed structural changes in the perovskite as the Ru inserted in the structure
increases. It was observed that the insertion of Ru affects the bulk structure by creating rotational
and Jahn-Teller distortions in the perovskite structure. Raman spectroscopy completed the
description, proving the strong distortions of the lattice oxygen and the La-O coordination
induced by the presence of ruthenium. Such distorted configuration gave rise to a weakening of
metal-oxygen bonds, maximizing anionic mobility and reactants adsorption. Surface changes
were also observed with the insertion of Ru in the perovskite structure. XPS showed that there
are cobalt spinel species, unaltered by ruthenium, and lanthanum oxide species that become more
carbonated when Ru is present. The formation of carbonate-like structures is enhanced by
ruthenium, which must be interacting with lanthanum entities, loosening La-O bonds in order to
facilitate the adsorption of CO2. Relating these structural effects with catalytic performance in
hydrocarbons reforming, we can conclude that the structural distortion induced by ruthenium
favours catalytic stability, probably by stabilizing metallic Co and Co-Ru sites, increasing metal
dispersion and by making oxygen mobility easier in the disturbed La2O3 support.
1. INTRODUCTION
Wet impregnation of different supports is the commonly procedure employed to deposit metal
nanoparticles on a catalyst surface. This method is rather simple, but is not completely
reproducible as the distribution of the metal component across the surface is not homogeneous.
In this scenario, solid catalysts can be prepared using the metal ion precursors perfectly
2
distributed in a crystalline structure, which develops upon reduction highly dispersed and stable
metal particles on the substrate surface. Perovskite oxides (ABO3) have been extensively studied
in heterogeneous catalysis because they are a promising alternative to traditional supported
catalyst formulations, on account of their controllable physical and chemical properties due to
the wide range of ions and valences which this simple crystal structure can accommodate.1-3
Perovskites could act as precursors of catalysts containing perfectly distributed active metal (B)
in the perovskite structure (ABO3), which upon reduction develops highly dispersed and stable
active metal particles (B-site cations) on the surface of the oxide on element A.4-6 LaCoO3 is
particularly attractive as precursor of catalysts for hydrocarbon reforming because it is one of the
most reducible ABO3-type perovskites. After reduction, it forms highly dispersed Co particles in
close contact with La2O3, which has an important role in catalyst stability by favouring coke
gasification during hydrocarbon reforming. A way to achieve more active and stable catalysts is
to tailor catalytic properties by producing structural and electronic modifications through partial
substitution of Co sites with another cation in the perovskite lattice. Among the transition metals
for Co replacement in the perovskite lattice, ruthenium was particularly effective in catalytic
reforming of hydrocarbons for hydrogen production as shown in our previous works.7,8 Besides
the higher intrinsic activity of ruthenium, its better catalytic behaviour is attributed to the partial
distortion of the rhombohedral phase associated to high Co replacement, which manifests a
higher exposition of active phases formed during reaction. The synthetic route and formation
mechanism of perovskite-structured mixed oxides is of vital importance to determine the origin
of the reactivity obtained in catalysts derived from perovskites. The formation mechanism of
perovskites is scarcely studied in the literature. Perovskite formation has been studied by Ivanova
et al.,9,10 who performed a thermal analysis on a series of LaCoO3, LaCo1-xNixO3 and LaCo1-
xFexO3 samples prepared by a Pechini-like synthetic route. According to this study, the formation
of perovskite oxides starts from an amorphous hydrated lanthanum carbonate, goes through an
oxo-carbonate intermediate that finally decomposes to give rise to the perovskite oxide.
3
Taking into account the importance of the formation mechanism of perovskites in the origin of
the reactivity of the catalysts derived from them, we present herein an extensive characterization
analysis of the structural and chemical changes upon incorporating Ru to the LaCoO3 perovskite
lattice. Advanced in situ characterization by synchrotron-based wide-angle XRD and Raman
spectroscopy has been applied in order to unravel the formation mechanism of the perovskite
oxide. In line with this, we have studied the mechanism of Ru incorporation into the LaCo1-
xRuxO3 (x = 0.05-0.4) perovskite during annealing steps and how the resulting structures affect
catalyst performance in the hydrocarbon reforming reaction. Furthermore, characterization of the
formed perovskites has been performed by high resolution XRD, Raman, EXAFS and XPS in
order to relate structural changes with performance in the hydrocarbon oxidative reforming.
2. EXPERIMENTAL SECTION
2.1 Perovskite preparation
LaCo1−xRuxO3 perovskite oxides (x = 0, 0.05, 0.1, 0.2 and 0.4) have been prepared by a modified
citrate sol–gel method (Pechini method). 1 M aqueous nitrate solutions containing the precursor
cations La(NO3)3·6H2O (99.9% Alfa Aesar), Co(NO3)2·6H2O (97.7% Alfa Aesar) and RuCl3
(40.49% Ru Johnson Matthey) were added to a solution of citric acid (Alfa Aesar) and ethylene
glycol (99.5% Riedel-de Haën) (molar ratio ethylene glycol/citric acid = 1 and citric acid/(La +
(Co + Ru)) = 2.5). The mixture was stirred and heated at 70 ºC for 5 h in order to evaporate the
excess of solvent and promote polymerization. After some hours, a purple or black, highly
viscous gel was obtained. The resulting resin, which contains the metal cations inside a
polymeric network, was charred at 300 ºC for 2 h to remove the organic matter in order to obtain
the perovskite precursor. After that, the resin was milled to obtain a fine powder. For the
formation of the perovskite, the samples were calcined under air at 750 ºC for 4 h.
4
2.2 Physicochemical characterization
Formation of the perovskite structures was followed by Time Resolved X-ray diffraction (TR-
XRD), acquired at beamline X7B (λ = 0.3184 Å) of the National Synchrotron Light Source at
Brookhaven National Laboratory. Two-dimensional XRD patterns were collected with an image
plate detector (Perkin-Elmer). Each diffraction pattern was acquired in 3 min. The powder rings
were integrated using the FIT2D code. The sample (5-10 mg) was loaded into a quartz capillary
cell (1 mm diameter), which was attached to a flow system. A small resistance heater was
wrapped around the capillary, and the temperature was monitored with a 1.0 mm chromel-alumel
thermocouple that was placed straight into the capillary near the sample.11 Samples were heated
in a O2/He (5% vol. O2) flow up to 800 ºC. The relative product concentrations from the TR-
XRD experiments were measured with a 0–100 amu quadruple mass spectrometer (QMS,
Stanford Research Systems). A portion of the exit gas flow passed through a leak valve and into
the QMS vacuum chamber. QMS signals at mass-to-charge ratios of 2 (H2), 4 (He), 16 (O, CH4),
17 (OH), 18 (H2O), 28 (CO), 32 (O2) and 44 (CO2), were monitored and recorded during the
experiments.
Raman spectra were acquired with a Renishaw inVia spectrophotometer, equipped with Leica
optics, a CCD detector cooled at -70 ºC and super-Notch holographic filters to get rid of the
elastic dispersion. A red laser (785 nm and maximum power of 300 mW) was chosen as an
excitation source. Photons dispersed by the sample were grated trough a 1200 lines/mm
monochromator before reaching the detector. The spectrometer was calibrated with a Si standard
using a Si band position at 520.3 cm-1. The ex situ Raman spectra of the annealed perovskites
along with the reference compounds La2O3, Co3O4 y RuO2, were recorded in a static mode
(centered at 700 cm-1) with a laser power of 0.3 mW, 10 s of exposure time, 20 accumulations, 1
cm-1 of spectral resolution and a 50x objective. For the in situ experiments during perovskite
formation, precursor samples were placed in a Linkam CCR 1000 cell. A 50 mL/min flow of
5
O2/N2 (21% vol. O2) was employed with a heating rate of 10 ºC/min up to 750 ºC holding this
temperature for 30 min. Temperature was raised in 100 ºC intervals and Raman spectra were
recorded at room temperature under the same conditions than the ex situ analysis.
High resolution XRD patterns were performed at beamline MSPD (λ = 0.4246 Å) at ALBA
Synchrotron Light Facility with the collaboration of ALBA staff.12 The powder samples were
loaded into thin (1/16’) kapton capillaries. One dimensional XRD patterns were collected by
continuous scanning of the so called MAD26 detector setup which is composed of 13 silicon
analyser crystals (Si 111 reflection) + scintillator/PMT detectors separated by ~1.5º angular
offsets.13 Each pattern was collected over 48 minutes in a 0-48 degree 2theta range. Rietveld
refinement was accomplished by the use of GSAS software.14 The instrument parameters
(Thompson-Cox-Hastings and asymmetry profile coefficients) were derived from the fit of a Si
reference pattern.15-18 The obtained patterns were compared with the Inorganic Crystal Structural
Database (ICSD) data for phase identification.
X-ray absorption measurements of annealed perovskites were carried out at beam line X18B of
the National Synchrotron Light Source at Brookhaven National Laboratory. A Si (111) double
crystal monochromator was used for energy selection. The monochromator was detuned by 20%
to suppress higher harmonic radiation. The intensities of the incident and transmitted X-rays
were monitored by ionization chambers. EXAFS spectra were acquired in transmission mode.
The energy resolution employed was 0.5 eV. Finely grounded powder samples were
homogeneously spread over kapton tape that was folded 3 to 4 times to achieve an optimal
energy jump. Sample data was acquired simultaneously with that of a 7 mm thin Co foil (for
energy calibration) at room temperature. Data analysis and background subtraction was
performed using Athena suite of programs.19
X-ray photoelectron spectra of the annealed perovskites were recorded on a VG Escalab 200R
spectrometer equipped with a hemispherical electron analyser and an Mg Kα (1253.6 eV), X-ray
6
source (12 kV and 10 mA). The powder samples were packed into small aluminium cylinders
and mounted on a sample rod placed in the pre-treatment chamber and degassed at 500 ºC for 1 h
before being moved into the analysis chamber. The base pressure of the ion-pumped analysis
chamber was maintained below 3.10-9 mbar during data acquisition. Charge effects on the
samples were corrected by fixing the binding energies of C1s peak at 284.8 eV due to
adventitious carbon. This reference gave binding energies values with accuracy of ±0.1 eV. Data
treatment was performed using “XPS peak” software. The spectra were decomposed with the
least square fitting routine using Gaussian/Lorentzian function and after subtracting a Shirley
background. Peak intensities were estimated by calculating the integral of each peak after
smoothing and subtracting a Shirley-type background.20 Atomic surface contents were estimated
from the areas of the peaks, corrected using the corresponding sensitivity factors.21
2.3 Activity tests
Catalytic tests for oxidative reforming of diesel were carried out in a fixed-bed continuous-flow
stainless steel reactor. The catalytic bed, 100 mg of catalysts, was placed in a tubular reactor (8
mm i.d.) with a coaxially centred thermocouple in contact with the catalytic bed. Prior to
reaction, perovskite precursors were flushed in H2/N2 (10% vol. H2, 50 mL/min) at 700 ºC for 1 h
before admission of feed mixture. The flow rates of diesel and water feeds were controlled by
liquid pumps and were preheated (200 ºC) in an evaporator before passing through the catalyst
bed in the reactor. Diesel fuel was provided by CEPSA (R&D Center, C14,4H27,4) and its sulphur
amount was 22 ppmw. Nitrogen gas was also fed to the evaporator to facilitate the evaporation
and passage of both the hydrocarbon and water. For the oxidative reforming of diesel, the
reactants were introduced into the reactor in a molar ratio of H2O/O2/C= 3/0.5/1. The total gas
flow rate was kept at 75 mL/min (GHSV= 20000 h-1). Activity was measured at atmospheric
pressure and 750ºC maintaining the reaction for 24 h at this temperature. The products were
analysed periodically by an on-line gas chromatograph (Varian 450-GC) equipped with a TC
7
detector and programmed to operate under high-sensitivity conditions. A 5A (CP7538) molecular
sieve column is used for H2, O2, N2, CO and CH4 separation and a PoraBOND Q (CP7354) for
CO2, C2H6, C2H4, C3H8 y C3H6.
The diesel conversion and hydrogen yield are defined as follows:
Diesel conversion (%):
mole C (CO2+CO+CH4+C2 H4+C2 H 6+C3 H6+C3 H 8) in reformatemole C (C14 . 4 H 27 .4 ) feed
×100
Hydrogen yield (%):
mole of H2 in reformatemaximum theoretical mole of H2
×100
3 RESULTS AND DISCUSSION
3.1 Characterization of the evolution of perovskites precursors during calcination
In order to follow the influence of Co substitution by Ru in the formation of LaCo1-xRuxO3
perovskite, we have studied the structural evolution of perovskite precursors during the
calcination process up to 800 ºC under oxidant atmosphere by in situ characterization using X-
ray diffraction and Raman spectroscopy. Figure 1 shows the time-resolved XRD patterns
obtained during annealing from room temperature to 800 ºC under a 5% O2/He flow of the
LaCo0.8Ru0.2O3 perovskite precursor. The diffraction pattern recorded at room temperature
showed a wide featured profile indicating the low crystallinity of the perovskite precursor. At
room temperature, two diffraction peaks at 7.5º and 9.0º matched the cobalt spinel Co3O4 phase
with a cubic structure (Fd3m). Also found in this sample is a wide diffraction feature, centred at
4.3º. The most probable composition associated to this signal is a lanthanum carbonate hydroxide
structure of the type La2(OH)6-2x(CO3)x.22 This structure could be generated during the synthesis
of the perovskite precursor. In the synthesis of the precursor, the organic citrate groups form an
8
organometallic complex, while the ethylene glycol gives rise to a resin-like polyester. When
treating this gel in an oxidizing atmosphere the organic ligands decompose, generating carbonate
species. These CO32- groups leave the sample during annealing, mainly in the form of CO2. Due
to the basic nature of La3+ cations, the CO2 is adsorbed on the surface, leading to the formation of
carbonate structures. At room temperature these carbonate phases are heavily hydroxylated, and
could therefore be responsible for the abovementioned XRD signal detected in the perovskite
precursor.22 Ruthenium oxides can also form amorphous hydroxylated species at room
temperature that can similarly contribute to this wide diffraction feature.
1817161514131211109876543
2
800
700
600
500
400
300
200
100
Temperature /ºC
800
600
400
200
Co3O4 (ICSD-27497) LaCoO3 (ICSD-201767) La2(OH)6-2x(CO3)x
La2O2(CO3) hexagonal
Figure 1. Time-resolved XRD patterns of LaCo0.8Ru0.2O3 sample during calcination in 5% O2/He
flow
XRD patterns evolve smoothly during calcination until temperature reaches 400 ºC. At this
temperature, the peak at 4.3º disappears, followed by the growth of a new broad peak at 6.0º,
while the signal of the cobalt spinel Co3O4 phase remains constant. It is difficult to make a
precise assignment with only one broad diffraction peak. However, the position of this peak
matches the most intense peak of a lanthanum oxo-carbonate (La2O2CO3) phase in a hexagonal
9
structure; therefore, it seems feasible to assign this new peak to such a lanthanum oxo-carbonate
compound. An analogous assignment to a lanthanum oxo-carbonate nanocrystallized compound
has already been made in bibliography by both thermogravimetric analysis and diffraction
experiments.23 Another way visualizing the structural changes at 400 ºC is to see the compression
of the hydroxocarbonate structure (peak at 2 = 4 º, d-spacing = 4.5 Å) by the loss of water and
CO2 molecules to a more compacted structure with an interplanar spacing of 3 Å.
This nano-composite of lanthanum carbonate grows in intensity during calcination until the
temperature reaches 600 ºC. Finally, around 700 ºC the perovskite diffraction peaks start to
grow, as the Co3O4 peaks (at 7.5º and 9.0º) diminish in intensity. All LaCo1-xRuxO3 perovskite
precursors showed a similar behaviour during annealing to that described for the LaCo0.8Ru0.2O3
sample (Figure 1). The detailed temperatures for each transition for each of the LaCo1-xRuxO3
perovskite precursors are depicted in Table 1, while the evolution of the XRD peak intensities for
phases La2(OH)6-2x(CO3)x (4.0º), La2O2CO3 (6.0º) and perovskite (6.6º) during the annealing of
perovskite precursors can be followed in Figure 2A. From the results presented in this figure, the
degree of ruthenium replacement in the perovskite precursor has no effect on the formation
temperature of the perovskite phase. Analyses of gas evolved during annealing of the perovskite
precursors were followed by MS (signals of CO2 displayed in Figure 2B). An initial CO2
formation was observed at around 400 ºC, which accounts for the decarbonation and
dehydroxylation of the hydroxy-carbonates into oxo-carbonate structures, and a second, much
weaker, at high temperatures of 700 ºC that originates from the decomposition of the lanthanum
carbonate to form the final perovskite oxide.
SampleLa2O2CO3
Peak 2θ=6º
Perovskite
Peak 2θ=6.6º
CO2 decomposition (ºC)
Low Temp High Temp
x = 0 530 659 383 661
x = 0.05 527 679 377 689
10
x = 0.1 442 691 364 685
x = 0.2 421 686 353 698
x = 0.4 363 688 356 663
Table 1. Temperature of the maximum formation rate for lanthanum carbonate and perovskite
phases along with the temperature of CO2 decompositions during calcination of perovskite
precursors
100 200 300 400 500 600 700 8001
2
3
4
5
6
100 200 300 400 500 600 700 800
3.00E-009
6.00E-009
9.00E-009
1.20E-008
1.50E-008
1.80E-008
A)
x = 0.4
peak @ 2=4º peak @ 2=6º peak @ 2=6.6º
Nor
mal
ized
pea
k in
tens
ity
Temp /ºC
B)
x = 0.2
x = 0.1
x = 0.05
x = 0
Temp /ºC
x = 0.4
x = 0.2
x = 0.1
x = 0.05
CO
2 Par
tial p
ress
ure
(atm
)
x = 0
x10 MS
Figure 2. A) Evolution of normalized XRD peak intensities of the phases La2(OH)6-2x(CO3)x (4º),
La2O2CO3 (6º) and perovskite (6.6º) during the annealing of perovskite, B) MS CO2 signal versus
temperature during the calcination of perovskite precursors
In parallel with diffraction experiments, vibrational Raman spectroscopy was used for structural
determination and phase transitions. Figure 3 shows the evolution of the Raman spectra of the
perovskite precursors obtained during calcination from room temperature to 750 ºC under a 5%
11
O2/He flow. All samples showed a similar evolution of Raman spectra during calcination. The
Raman spectra obtained at room temperature is governed by the cobalt spinel signals at 690, 480
and 196 cm-1. The evolution of the spinel phase towards the perovskite phase is observed on
increasing the calcination temperature. This accounted for a diminishing intensity of the peak at
690 cm-1. Only after annealing at 750 ºC were the signals ascribed to the perovskite phase at 619
cm-1 observed. Both diffraction and the Raman data prove that annealing at high temperature (>
700 ºC) for a long time is needed to obtain a well-structured perovskite mixed oxide.
100 300 500 700 900 1100
689617480196
160
195689
520620
481750 ºC 30 min
600 ºC500 ºC
400 ºC
300 ºC
Cou
nts
/ a.u
.
Raman shift / cm-1
30 ºC
2000
x = 0.05
100 300 500 700 900 1100
689629
560196
689
522618
481196
750 ºC 30 min
600 ºC
500 ºC
400 ºC
300 ºC
Cou
nts
/ a.u
.
Raman / cm-1
30 ºC
2000
x = 0.1
100 300 500 700 900 1100
690619483196
692622481195
692481195
691483195
654691
537483196
686
542
479
195
750 ºC 30 min600 ºC500 ºC400 ºC
300 ºC
Cou
nts
/ a.u
.
Raman shift /cm-1
30 ºC
2000
x = 0.2
100 300 500 700 900 1100
688667611
480
392
195
620521
688478
195 750 ºC 30 min
600 ºC
500 ºC
400 ºC
300 ºC
Cou
nts
/ a.u
.
Raman shift / cm-1
30 ºC
500
x = 0.4
12
Figure 3. Evolution of Raman spectra during calcination of the LaCo1-xRuxO3 perovskite
precursors
3.2 Effect of ruthenium on the formation mechanism of LaCo1-xRuxO3 perovskite
Throughout the in situ measurements performed during calcination of perovskite precursors, only
crystalline phases corresponding to Co3O4 spinel and LaCo1-xRuxO3 perovskite oxides were
observed. For lanthanum entities, wide diffraction features assigned to lanthanum carbonate
species (hydrated and de-hydrated) appeared, and no contribution was detected from any
ruthenium phases (oxides, hydroxides or carbonates), neither in diffraction nor in Raman. The
absence of Ru phases is startling, particularly for the samples with high Ru content, in which a
segregated oxide should be easily distinguished by XRD or Raman. Compared to their Ru-free
LaCoO3 counterpart, those containing ruthenium show no peak shifts in the Co3O4 spinel phase
(neither in XRD nor in Raman signal), a fact that excludes the possibility of the Ru atoms
becoming incorporated into the spinel structure forming a Co2RuO4 oxide.24 At low temperature,
both Ru and La are known to form amorphous hydroxide and hydroxycarbonate phases.
Ru(OH)n decomposition to RuO2 occurs between 300-400 ºC,25 but even at much higher
temperatures no diffraction or Raman peaks from RuO2 are identified. Only the diffraction lines
from low crystallinity La2(OH)6-2x(CO3)x and La2O2CO3 phases are distinguished. The highly
disordered and quasi-amorphous nature of these structures makes them ideal candidates to
conceal ruthenium species in inter-laminar positions. Another less likely option is that ruthenium
is forming a segregated or independent amorphous oxide.
For all the samples, the transition from the perovskite precursor to the perovskite structure occurs
through an oxo-lanthanum-carbonate intermediate phase. This intermediate phase decomposes
mostly at around 400 ºC into an oxo-carbonate structure and, finally, this oxycarbonate forms the
final perovskite at around 700 ºC. Table 1 summarizes the temperatures at which the transitions
13
between the different phases are observed during the annealing of perovskite precursors. From
the evolution of the peak at 6.6º due to the formation of the perovskite, two main features are
derived. One is that its intensity is growing continuously until quenching the experiment, which
suggests that the kinetics of the formation is slow. Secondly, the position of the perovskite
diffraction peak does not change sharply with time, pointing to single-step perovskite oxides
formation. If the Ru were incorporated at a later stage than Co on the structure, there would
appear peak shifts in the diffraction patterns of the ruthenium-containing samples during
annealing (see Figure 2). There is no correlation between Ru content and the temperatures of
formation of either carbonate or perovskite phases (data in Table 1). The temperature for the
transition from La2(OH)6-2x(CO3)x to La2O2CO3 is between 420 ºC to 530 ºC, whereas the
temperature range for perovskite formation is 660-690 ºC. Ruthenium presence does not seem to
alter the formation mechanism of the perovskite oxides. The sample with highest Ru content (x =
0.4) has a different behaviour than the rest of the samples. First of all, the position of the most
intense peak of the perovskite phase is shifted toward smaller values. As shown in previous
sections, this sample crystallizes in a monoclinic phase. The position shift causes the perovskite
peak at 6.6º to overlap with the signal of the La2O2CO3 phase at 6.0º, owing to which, unlike in
the other samples in the series, we cannot observe the disappearance of the lanthanum oxo-
carbonate phase. Besides, the formation temperature of the perovskite oxide is the highest of the
series. Such a high temperature is needed to obtain Ru3+ species.26
Based on these analyses we can propose two alternative mechanisms for the formation of LaCo1-
xRuxO3 perovskites (the * marks the crystalline phases that are clearly distinguished in the
diffraction patterns and in the Raman spectra):
T > 350 ºC: 2La2(OH)6-2x(CO3)x ∆→ La2O2CO3 + xCO2 + 3H2O [1a]
T > 700 ºC La2O2CO3 + 2/3Co3O4* + xRuO2 + nO2∆→ 2LaCo1-xRuxO3* + CO2 [2a]
14
T > 350 ºC 2La(OH)6-2y(CO3)y-Rux ∆→ La2O2CO3-Rux + yCO2 + 3H2O [1b]
T > 700 ºC 3La2O2CO3-Rux + 2Co3O4* + ½O2 ∆→ 6LaCo1-xRuxO3* + 3CO2 [2b]
The first mechanism involves the segregation of Ru species in the form of amorphous ruthenium
oxide phase (mechanism a), while the second requires the incorporation of Ru species in the
lanthanum oxycarbonate phase (mechanism b). In the first, the formation of the perovskite phase
can only take place if a solid-state reaction between three different phases, La 2O2CO3, Co3O4 and
RuO2 takes place simultaneously. In the second, a very disordered carbonate phase, containing
La3+ ions and Ru4+ species, reacts with the cobalt spinel to form the perovskite oxide as shown
schematically in Figure 4. The highly unlikely event of a solid-state reaction between three
different phases makes mechanism b statistically and kinetically preferential over mechanism a.
3 La2(CO3)3(H2O)8 Rux
O2, D
H2O + CO2
3 La2O2(CO3)2Rux
O2, D
H2O + CO2
Ste
p1
T
350º
C
3LaCo1-xRuxO3Co3O4 3 La2O2(CO3)2Rux
O2, D
CO2
+
O2, D
CO2
+
Ste
p2
T >
700
ºC
Figure 4. Proposed scheme for the formation of LaCo1-xRuxO3 perovskite oxide
3.3 Characterizacion of perovskites after calcination
The high-resolution X-ray diffraction patterns of the LaCo1−xRuxO3 calcined samples, along with
the refined data, are displayed in Figure 5 A. The diffraction pattern of LaCoO3 sample showed
15
strong reflexions at 8,97º and 9,05º corresponding to the rhombohedral (R3c) structure 8 of
perovskite with a minor contribution at 7.5º and 9.0º of cobalt spinel (Co3O4) indicative of the
high degree of incorporation of the La and Co oxides into the perovskite structure. The
diffraction patterns of the LaCo1−xRuxO3 samples show profiles corresponding to single
perovskite structures without peaks attributable to ruthenium oxides (Figure 5A). The partial
substitution of Co by Ru evidences changes in the rhombohedral structure of the perovkite as the
changes the diffraction lines characteristic of the of LaCoO3 sample indicated in Figure 5B. In
this figures it is observed that the diffraction lines of the Ru-substituted perovskites shifted to
lower angles respect to the diffraction lines characteristic of the rhombohedral LaCoO3
perovskite phase that result in a modification of the structure of pure LaCoO3. As the degree of
Ru replacement increased, the rhombohedral perovskite structure became increasingly more
distorted, which can be accounted for by a shift in the peaks toward smaller angles and by a
lesser splitting of the peaks at 9º. The maximum distortion in rhombohedral structures is
achieved for the sample with 20% of the atomic substitution of Co by Ru. For a degree of Ru
atomic substitution greater than 20%, the rhombohedral phase is no longer stable and the
LaCo0.6Ru0.4O3 sample crystallizes in a double perovskite structure with monoclinic
symmetry.27,28 . No diffraction peaks were observed from any crystalline phase associated to
ruthenium oxides Table 2 summarizes the results of X-ray Rietveld refinement: lattice
dimension, crystallite size as obtained by the Scherrer equation, occupancy of B site and weight
fraction of each crystallographic phase. Rietveld refinement shows that as the Ru substitutes Co
positions in the lattice, a cell expansion is observed. This expansion could be caused by the
higher ionic radii of the Ru3+ (0.68 Å) as compared to the Co3+ ions (0.55 Å). The cell expansion
may also be caused by charge redistribution between Ru and Co ions. Ru ions are very stable as
+4 cations and they may incorporate as such into the perovskite structure. If that is the case,
some Co3+ ions must get reduced ion order to compensate charges. The substitution of cobalt by
ruthenium is confirmed by an increased occupancy of B sites as obtained in the Rietveld
16
refinement. Another important parameter, also affected by the presence of Ru, is the crystallite
size determined by the Scherrer equation. As Ru is incorporated in the perovskite lattice, the
particle size diminishes reaching its minimum for the LaCo0.8Ru0.2O3. Results from the Rietveld
refinement show that the presence of ruthenium hinders phase segregation since cobalt spinel
phase fraction diminishes when ruthenium is added. Additionally, the distortion induced by the
presence of ruthenium also allows the formation of smaller crystallite sizes.
120x103
100
80
60
40
20
0
Cou
nts
/a.u
.
161514131211109876
2
Observed Rietveld Refinement Co3O4 (ICSD-27497) LaCoO3 (ICSD-201767)
x = 0
x = 0.05
x = 0.1
x = 0.2
x = 0.4
A)
120x103
100
80
60
40
20
0
Cou
nts
9.29.08.88.6
2
x = 0
x = 0.05
x = 0.1
x = 0.2
x = 0.4
B)
Figure 5. High resolution XRD patterns and Rietveld refinement of the calcined LaCo1-xRuxO3
perovskites (x=0, 0,05, 0,1, 0,2 and 0,4) (A), Inset of the doublet peak at 9º showing the peak
shift and symmetry loss with increasing Ru substitution in the perovskite (B)
Sample Space group a (Å) b (Å) c (Å)size
(nm)
Occ
. B
site
%weight
LaCoO3 La2CoRuO6 Co3O4
x = 0 R3c
rhombohedral
5.442 5.442 13.102 42.5 1.05 97.77 0.00 2.23
x = 0.05 5.457 5.457 13.142 32.7 1.09 98.13 0.00 1.87
x = 0.1 5.478 5.478 13.193 31.4 1.10 98.56 0.00 1.44
17
x = 0.2 5.505 5.505 13.393 28.3 1.20 98.35 0.00 1.65
x = 0.4P21/n
monoclinic5.571 5.608 7.870 30.8 1.18 11.78 87.69 0.53
Table 2. Structural parameters from XRD Rietveld refinement of the calcined LaCo1−xRuxO3
perovskites
Figure 6 shows the XANES spectra of the calcined LaCo1−xRuxO3 perovskites analysed in
comparison with the CoO and Co3O4 reference oxides. The XANES spectra of the parent
LaCoO3 and the Ru-substituted samples are similar to those previously published for LaCoO3 by
Thornton et al.29 The XANES region of the Co K-edge deals with the electronic transition
between the core 1s electrons and the empty states of the 4s and 4p shell, where the 1s4p
transition is the one allowed by selection rules while the pre-edge features are governed by the
symmetry-forbidden Co3+ 1s3d transitions. In order to better analyse the subtle variations in the
XANES signal, we will also study the 1st and 2nd derivatives of the absorption spectra (Figure 6).
The maximum for the first derivative provides the position of the absorption edge (E0) of each
sample as summarized in Table 3. The absorption edge for the pure LaCoO3 sample is 7724.4 eV
in agreement with the reported value.30 Upon substituting the samples with increasing Ru
amounts, a shift towards smaller energies of the adsorption peak is observed. This effect
accounts for a partial reduction of the Co3+ ions to Co2+. This is followed by all the samples in the
series except for the one with the highest Ru content (x = 0.4), for which cobalt species are fully
oxidized to +3. The partial reduction of the Co3+ ions to Co2+ when ruthenium is incorporated is
in full agreement with the Rietveld refinement data showed above which point to the reduction
of Co to 2+ oxidation state to compensate the introduction of Ru4+ ions in the lattice of
perovskite. Ru K-edge XANES data corroborates this face showing a Ru4+ oxidation state in all
the samples.
18
7700 7710 7720 7730 77407700 7720 7740 7760 7780
1st d
eriv
ativ
e of
abs
orpt
ion
data
Energy /eV
Co2+ LaCoO3
x = 0.2
x = 0.1
x = 0.05
x = 0.4
x = 0.2
x = 0.1
x = 0.05
x = 0
x = 0.05
x = 0.1
x = 0.2
CoO
Co3O
4
x = 0
x = 0.4
Co3O
4
Nor
m. a
bs
Energy /eV
CoO
x = 0
Co3O4
CoO
7700 7710 7720 7730 7740
x = 0.4
2nd d
eriv
ativ
e of
abs
orpt
ion
data
Energy /eV
Co(1s)Co(4p)
LMCT
Figure 6. Normalized Co K-edge XANES spectra of calcined LaCo1-xRuxO3 perovskites and
their 1st and 2nd derivatives
Sample E0 Oxidation state
CoO7720.
5+2
Co3O4 7721. +2, +3
19
7
x = 07724.
4+3
x = 0.057723.
4+2, +3
x = 0.17722.
7+2, +3
x = 0.27722.
2+2, +3
x = 0.47724.
0+3
Table 3. Co K-edge energies and main oxidation states of calcined LaCo1−xRuxO3 perovskites
along with CoO and Co3O4 reference samples
Figure 7A shows the k2-weighted Co K-edge EXAFS spectra of the analysed samples. As Ru is
incorporated in the structure, oscillations are broadened and signal is flattened at high k values.
This behaviour can easily be ascribed to an increased disorder due to the distortion induced by
the presence of Ru in the coordination of Co atoms. The sample with the highest Ru content, x =
0.4, shows a different behaviour as it regains order in its structure. Moreover, the oscillations in
k-space of this sample are very similar to those of the LaCoO3 parent structure, suggesting
similar chemical first coordination shells in both samples. This similar short-range ordering
(octahedral oxygen coordination around Co atoms) for the LaCo0.6Ru0.4O3 sample as compared to
the LaCoO3 is consistent with evolution to a different long-range structure, as observed in the
distinct phase obtained by XRD. Figure 7b shows phase uncorrected interatomic distance
obtained by Fourier transformation of (k) over the k-space range between 2-12.5 Å-1. For
the parent LaCoO3 perovskite, the first peak in the radial distribution at 1.5 Å is assigned to the
20
first shell Co-O distance. Weak peaks at 2.3 and 2.6 Å correspond to Co-O-O and Co-O-La
distances. The strong peak at 3.1 Å is attributed to Co-O-Co(Ru) distance in adjacent octahedra.
It was noted that the absolute values of the interatomic distances obtained by EXAFS are too
short, as compared to the real distances, due to the inherent phase shift of the uncorrected
EXAFS dispersion data. For Ru-containing samples, the appearance of a shoulder is observed at
short distances (0.95 Å) indicating a Co-O bond shortening. Such distortion in the Co-O bond
distance may be due to a Jahn-Teller distortion in the CoO6 octahedra, arising from the different
ionic radii of Co3+, Co2+ and Ru4+ cations occupying B sites. The Jahn-Teller distortion causes the
splitting of the first shell Co-O bond distance, elongating some bonds and shortening others. At
high R-values the disorder created by the incorporation of Ru leads to a flattening of the signal.
The different bond distances at R > 4 Å due to the variable second coordination shell of Co-O-
Ru paths are caused by the combined Jahn-Teller and rotational distortions of MO6 octahedra.
The EXAFS signal shows that the disorder induced by Ru incorporation affects the local
environment of Co in addition to the long-range effects observed by XRD.
21
Figure 7. The k2-weighted Co K-edge EXAFS spectra of the calcined LaCo1-xRuxO3 perovskites
(x=0, 0,05, 0,10, 0,20, 0,40) (A) and the derived radial distribution function for each sample (B)
The Raman spectra of the LaCo1−xRuxO3 annealed samples, along with the Co3O4 and RuO2
reference oxides, are displayed in Figure 8. The spinel Co3O4 (cubic structure)31 shows Raman
modes at 196, 481 (Eg), 521 (F2g), 619 (F2g) and 690 (A1g) cm-1, whereas the Raman spectrum of
RuO2 (tetragonal phase) possesses three vibrational modes at 523 (Eg), 644 (A1g) and 710 cm-1
(B2g).32 The LaCoO3 with perovskite structure displays bands at 157, 196, 416, 480, 521, 619,
647 and 689 cm-1.2,33 The region from 400 to 700 cm-1 is assigned to Co-O bending and stretching
modes. At low wavenumbers we can also observe a rotational Raman band of the Co-O
octahedral at 196 cm-1 and the mode of La-O bending at 157 cm-1. The most intense Raman
modes of the LaCoO3 perovskite overlap with the Co3O4 signal, making it difficult to quantify
phase segregation from the Raman spectra. This is due to a similar local environment around the
CoO6 octahedra in both phases. In the region from 400 to 700 cm-1, characteristic of the Co-O
stretching and bending local vibrations, the LaCoO3 sample presented a Raman peak at around
690 cm-1, followed by a wide band centred at 647 cm-1 and a small peak at 635 cm-1. In this
region, the Ru-substituted samples presented the characteristic peak at 690 cm-1 slightly shifted
towards a higher wavenumber while the signal at 600-650 cm-1, arising from oxygen mobility in
the structure, became broader and more intense. The general trend is for the oxygen mobility to
improve with Ru incorporation. The shift in the position of the 690 cm -1 is ascribed to different
bond distances in the vibration modes, caused by distortion in the M-O octahedral as Ru is
incorporated. The sample with highest Ru content, x = 0.4, crystallized in the double perovskite
structure La2CoRuO6 with a monoclinic symmetry and originated an intense sharp Raman mode
at 665 cm-1 with a weaker peak at 689 cm-1 and a wide band at 597 cm-1. As observed from XRD
results, there were no signals arising from ruthenium oxides in segregated phases, confirming the
absence of segregation of ruthenium phases from the perovskite structure. The presence of
ruthenium not only alters the coordination of cobalt atoms, but it also has strong effect on the
22
lanthanum environment. For the low wavenumber region (below 200 cm-1), assigned to La-O
roto-vibrations there also are variations dependent on Ru content. For the Ru-substituted
perovskites, this Raman mode is broadened and split into two components, suggesting a
distortion on La-O bonds.
100 125 150 175 200 400 500 600 700 800 900
x = 0.4
x = 0.2
x = 0.1
x = 0.05
Co-Ostretching
Co-ObendingRotationalLa bending
RuO2
x = 0
Inte
nsity
Raman shift /cm-1
x 0,03 Co3O
4
23
Figure 8. Raman spectra of the calcined LaCo1-xRuxO3 perovskites along with Co3O4 and RuO2 as
reference
The surface composition and oxidation state of the calcined LaCo1−xRuxO3 perovskites was
determined by XPS. The Co 2p level of all LaCo1−xRuxO3 perovskites (Figure 9) shows the BE of
the most intense Co 2p3/2 peak of the Co 2p doublet centred at 780.1 eV with a satellite line at
790.2 eV. These values are consistent with the presence of Co3O4 species on the perovskite
surface.34,35 Cobalt signal was unaltered in the Ru-substituted samples and no peak shift or
broadening was observed upon the incorporation of ruthenium, except for the sample with the
highest Ru content (x = 0.4). For this sample the Co 2p doublet appeared somewhat broadened,
the satellite peak being less intense. This behaviour points to a surface enrichment in Co3+
species.34
765 775 785 795 805 815
790.2
780.1
c.p.
s. /
a.u.
B.E. / eV
500 x = 0
765 775 785 795 805 815
790.1
779.7
c.p.
s. /
a.u.
B.E. / eV
1000 x = 0.05
765 775 785 795 805 815
789.6
780.1
c.p.
s. /
a.u.
B.E. / eV
1000 x = 0.1
765 775 785 795 805 815
790,1
780,4
c.p.
s. /
a.u.
B.E. / eV
1000 x = 0.2
24
Figure 9: Co 2p XP spectra of the calcined LaCo1-xRuxO3 (x=0, 0.05, 0.1, 0.2 and 0.4) perovskite
oxides.
For the LaCoO3 sample, the La 3d spectra (Figure 10) shows a characteristic doublet for each La
3d5/2 and La 3d3/2 component at 833.3 and 835.3 eV. Taking into consideration the position and
shape of the peaks, the first contribution at 833.3 eV is assigned to La3+ in a perovskite
environment35,36 where the second contribution at 835.3 is assigned to La3+ combined with
hydroxyl37 or carbonate groups (La2(CO3)3, La2O2CO3).37,38 The Ru-containing samples show a
poorly resolved La 3d doublet profile as well as a shift towards higher BE values, which suggest
the increasing content in lanthanum hydroxide or carbonate species induced by the presence of
ruthenium. The O 1s spectra of the annealed perovskites show differences depending upon the
degree of Ru substitution (Figure 11). The O 1s spectra of all perovskites show three different
contributions due to lattice oxygen (529 eV), hydroxide/carbonate species (531 eV) and a
characteristic tail around 533.5 eV related to oxygen from molecular water strongly adsorbed on
the surface. The Ru-substituted perovskites presented higher percentages of surface oxygen in
765 775 785 795 805 815
c.p.
s. /
a.u.
B.E. / eV
1000779.9
789.5
x = 0.4
765 775 785 795 805 815
789.6
780.1
c.p.
s. /
a.u.
B.E. / eV
1000 x = 0.1
765 775 785 795 805 815
790,1
780,4
c.p.
s. /
a.u.
B.E. / eV
1000 x = 0.2
25
the form of hydroxide/carbonate than in the form of lattice oxygen. This fact could be related to
the higher concentration of lanthanum hydroxide or carbonate surface species previously
observed in the analysis of La 3d levels that could be associated to the Co substitution by Ru in
the perovskite lattice.
The surface composition of the calcined LaCo1−xRuxO3 perovskites (Table 4) was determined
from XPS data. The comparison of nominal and surface concentration of Co and Ru calculated
from XPS intensities are presented in Figure 12. It was observed that the surface concentration of
Co was lower than the nominal value in the case of the samples with higher Ru substitution
(LaCo0.8 Ru0.2O3 and LaCo0.8 Ru0.4O3), indicating a loss of cobalt at surface level. The Ru surface
exposure varied with the Ru loading in the perovskite precursor. As it is shown in Fig. 12, the
relative surface concentration of ruthenium proportionally increased with the Ru loading in the
perovskite precursor except for the sample with higher Ru substitution for which a strong surface
concentration was detected.
825 830 835 840 845 850 855 860 865
835.3833.3
c.p.
s. /
u.a.
B.E. / eV
2000 x = 0
825 830 835 840 845 850 855 860 865
835.4833.4
c.p.
s. /
a.u.
B.E. / eV
2500 x = 0.05
825 830 835 840 845 850 855 860 865
835.4833.6
c.p.
s. /
a.u.
B.E. / eV
2500 x = 0.1
825 830 835 840 845 850 855 860 865
835.7833.8
c.p.
s. /
u.a.
B.E. / eV
2500 x = 0.2
26
Figure 10: La3d XP spectra of the calcined LaCo1-xRuxO3 (x=0, 0.05, 0.1, 0.2 and 0.4) perovskite
oxides.
524 526 528 530 532 534 536 538
532.3
530.6
528.7
c.p.
s. /
a.u.
B.E. / eV
1000 x = 0
524 526 528 530 532 534 536 538
532.1
530.6
528.6
c.p.
s. /
a.u.
B.E. / eV
1000 x = 0.05
825 830 835 840 845 850 855 860 865
835.4833.6
c.p.
s. /
a.u.
B.E. / eV
2500 x = 0.1
825 830 835 840 845 850 855 860 865
c.p.
s. /
a.u.
B.E. / eV
5000
837.7
834.2
x = 0.4
825 830 835 840 845 850 855 860 865
835.7833.8
c.p.
s. /
u.a.
B.E. / eV
2500 x = 0.2
524 526 528 530 532 534 536 538
532.8
531.1529.1
c.p.
s. /
a.u.
B.E. / eV
1000 x = 0.2
524 526 528 530 532 534 536 538
533.5
530.8528.8
c.p.
s. /
u.a.
B.E. / eV
1000 x = 0.1
27
Figure 11: O1s XP spectra of the annealed LaCo1-xRuxO3 (x=0, 0.05, 0.1, 0.2 and 0.4) perovskite
oxides.
Figure 12 XPS surface Co/La and Ru/La atomic ratio of calcined LaCo1−xRuxO3 (x=0, 0.05, 0.1,
0.2 and 0.4) perovskites
Sample Co/La* Ru/La* (Co+Ru)/La* O(OH-/CO32-)/Ored O/(Co+Ru+La)*
524 526 528 530 532 534 536 538
532.8
531.1529.1
c.p.
s. /
a.u.
B.E. / eV
1000 x = 0.2
524 526 528 530 532 534 536 538
533.5
530.8528.8
c.p.
s. /
u.a.
B.E. / eV
1000 x = 0.1
524 526 528 530 532 534 536 538
533.8
530.7
528.5
c.p.
s. /
a.u.
B.E. / eV
1000 x = 0.4
28
x = 01.03
(1.00)-
1.03
(1.00)1.03
2.53
(1.50)
x =
0.05
0.95
(0.95)
0.08
(0.05)
1.03
(1.00)1.53
2.58
(1.50)
x = 0.11.09
(0.90)
0.11
(0.10)
1.20
(1.00)1.92
2.21
(1.50)
x = 0.20.65
(0.80)
0.22
(0.20)
0.87
(1.00)1.21
2.77
(1.50)
x = 0.40.35
(0.60)
0.20
(0.40)
0.56
(1.00)3.31
2.54
(1.50)
*In brackets nominal atomic ratio
Table 4. XPS surface atomic ratio of calcined LaCo1−xRuxO3 (x=0, 0.05, 0.1, 0.2 and 0.4)
perovskites
3.4 Influence of Ru incorporation in the catalytic performance of perovskites for hydrocarbon
reforming
Characterization of perovskites showed that the partial substitution of Co by Ru led to
differences in their structure, crystallite size and surface characteristics. The structural effects
observed on the LaCo1-xRuxO3 perovskites have strong influence on their catalytic behaviour in
the oxidative reforming of diesel. The activity of the catalysts derived from LaCo1-
xRuxO3perovskite precursors for the oxidative reforming of diesel was measured in terms of
diesel conversion and hydrogen yield. Figure 13 shows the evolution of diesel conversion with
the reaction time for each catalyst. It is observed that each catalyst evolved in a different way
with time-on-stream, depending on the catalyst precursor. Figure 14 shows the hydrogen yield
29
obtained in the oxidative reforming of diesel on the catalysts derived from LaCo1-xRuxO3
perovskites. The catalysts displayed differences in the initial activity (0-8 h) and stability with
time-on-stream. These differences are indicative of differences in the initial concentration and
stability of the Co, Ru and La phases present in catalysts that are affected by the ruthenium
incorporation in the perovskites precursors.
Figure 13. Evolution of diesel conversion with time on stream during the oxidative reforming of
diesel over catalyst derived from LaCo1-xRuxO3 perovskite precursors (x = 0, 0.05, 0.1, 0.2 and
0.4)
30
60
55
50
45
40
35
30
% H
2 y
ield
4035302520151050
% at. Ru
0-8 h 16-24 h
Figure 14. Initial and steady-state hydrogen yield from oxidative reforming of diesel over
catalyst derived from LaCo1-xRuxO3 perovskite precursors as a function of ruthenium content on
perovskite (x = 0, 0.05, 0.1, 0.2 and 0.4)
It is known that the initial activity of the reforming catalysts is related with the surface exposure
of metal active Co and Ru sites.8 In this sense, the observed changes in the initial reforming of
samples should be related with differences in the dispersion of Co and Ru developed after the
reduction of the perovskite precursors. XRD, EXAFS and Raman analysis on LaCo1-xRuxO3
perovskite precursors indicated that the partial substitution of Co by Ru into LaCoO3 perovskite
led to structural changes associated with the ruthenium incorporation into the perovskite lattice.
Figure 15 shows the change in the crystallite size of perovskite and the reduction of Co3+ to Co2+
sites as function of the Ru content in the perovskite. The figure shows that as ruthenium is
incorporated in the perovskite, the Co reduction degree increases and the perovskite crystallite
31
size diminishes. The LaCo0.6Ru0.4O3 sample does not follow the trend and partially recuperates
both crystallite size and Co oxidation state. The abovementioned changes in size and reduction
degree of Co associated with the ruthenium incorporation in the perovskite coincide with the
trend in the initial reforming activity of the samples as observed in Figure 14. The lower size and
higher reduction degree of Co sites in perovskites may facilitate both the reduction of the
perovskite and the surface exposure of active Co and Ru sites that could be in the origin of the
differences in the initial activity observed among the samples. In addition, the partial substitution
of Co by Ru in LaCoO3 perovskite led to the distortion of the rhombohedral structure as a
consequence of the insertion of ruthenium cations into the Co position. This distortion makes
easier the oxygen mobility that facilitates the reduction of perovskites to form the catalysts.
7724.5
7724.0
7723.5
7723.0
7722.5
7722.0
7721.5
7721.0
7720.5
Abs
orpt
ion
Edg
e /e
V
4035302520151050
% at. Ru
42
40
38
36
34
32
30
Crystallite size /nm
E0 LaCo1-xRuxO3 Crystallite size
Co3O4 (Co+2
+ Co+3
)
CoO (Co+2
)
Figure 15. Co reduction degree (from absorption edge position) along with crystallite size of the
LaCo1-xRuxO3 perovskite as a function of ruthenium content
All catalysts show a decrease in activity after the first hours on stream indicating some
deactivation (Figure 14). However, the deactivation is lower as the Ru content in perovskite
increases (Figure 14). Catalyst deactivation by carbonaceous deposits is the main factor to
analyse in order to justify the evolution of catalysts under reforming conditions. Higher
carbonaceous deposits on reforming catalysts imply higher deactivation rates. Dispersion of
32
metal particles and the nature of supports strongly affect the formation of carbonaceous deposits
on reforming catalysts. Carbonaceous deposits are favoured on metal particles of larger size
while supports as La2O3 assist in coke removal from catalyst surfaces. In literature, positive
effect of lanthanum as support of noble metals applied to steam reforming of hydrocarbons are
explained by the ability of lanthanum to adsorb CO2 forming lanthanum oxycarbonates which
participate in coke gasification.39,40 Taking these facts into account, the evolution of catalysts
deactivation under reforming conditions could be related with the differences in cobalt and
ruthenium dispersion as well as their contact with lanthanum entities associated with the
ruthenium incorporation in the perovskite precursor. As commented previously, the partial
substitution of Co by Ru in the perovskite precursor probably generated a better Co and Ru
dispersion that could contribute to lower the carbon deposits on catalysts. In addition, the
deactivation differences observed on catalysts may also be related to the differences in
lanthanum coordination and chemistry derived from the structural distortion induced by
ruthenium substitution in the perovskite. As it was observed in the XPS and Raman spectra, the
cobalt substitution by ruthenium in the perovskite modifies the lanthanum characteristics
favouring the formation of carbonated species at the surface of the catalytic precursors. The
formation of carbonate-like structures is enhanced by ruthenium, which must be interacting with
lanthanum entities, loosening La-O bonds in order to facilitate the adsorption of CO2. Therefore,
the lower deactivation observed on catalysts with higher ruthenium substitution could be also
related with the enhancement of the capacity of the catalysts to form lanthanum oxycarbonates
which participate in coke gasification.
CONCLUSIONS
The in situ studies during the calcination process by time-resolved synchrotron X-ray diffraction
and Raman spectroscopy demonstrated that ruthenium content in LaCo1-xRuxO3 perovskites did
33
not affect the kinetics or formation temperature of the perovskite oxide. The formation
mechanism of LaCo1-xRuxO3 likely involves a solid-state reaction between cobalt spinel and
lanthanum oxide-carbonate in close contact with ruthenium species. Structural characterization
of perovskites showed structural changes in the perovskite with the insertion of Ru in the
structure. It was observed that the insertion of Ru affects the bulk structure by creating rotational
and Jahn-Teller distortions in the perovskite structure. In this way, the formation of a single
perovskite phase was observed, either in a rhombohedral (x = 0), distorted rhombohedral (x =
0.05, 0.1 and 0.2) or monoclinic (x = 0.4) structure, depending on the amount of Ru incorporated
into the perovskite structure. EXAFS analysis of Co K-edge showed strong Jahn-Teller
distortions around the CoO6 octahedra, in addition to the rotational distortions observed in the
diffraction patterns. Raman spectroscopy completed the description, proving the strong
distortions of the lattice oxygen and the La-O coordination induced by the presence of
ruthenium. Such distorted configuration gave rise to a weakening of metal-oxygen bonds,
maximizing anionic mobility and reactants adsorption. Surface changes were also observed with
the insertion of Ru in the perovskite structure. XPS showed that there are cobalt spinel species,
unaltered by ruthenium, and lanthanum oxide species that become more carbonated when Ru is
present. The formation of carbonate-like structures is enhanced by ruthenium, which must be
interacting with lanthanum entities, loosening La-O bonds in order to facilitate the adsorption of
CO2. Relating these structural effects with catalytic performance in hydrocarbons reforming, we
can conclude that the structural distortion induced by ruthenium favours catalytic stability,
probably by stabilizing metallic Co and Co-Ru sites, increasing metal dispersion and by making
oxygen mobility easier in the disturbed La2O3 support.
AUTHOR INFORMATION
Corresponding Author
34
*e-mail: r. [email protected]; phone +34915854774
Present Addresses
† Velocys, Milton Park, OX14 4SA United Kingdom.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the support provided by the Spanish Ministry of Economy
and Competitiveness (MINECO) under grant CTQ2013-48669-P and by CAM under grant
P2013/MAE-2882. Our appreciation goes to I. Peral and C. Popescu for their help in the
acquisition of HR-XRD at MSPD ALBA beamline. We thank J. Hanson and W. Xu for their
help in performing the TR-XRD experiment at X7B beamline (NSLS). We also wish to thank N.
Marinkovic for his help in the XAFS measurements at X18B (NSLS). Use of the National
Synchrotron Light Source, BNL, was supported by the US DoE, Office of Basic Energy
Sciences, under Contract No. DE-AC02-98CH10886.
REFERENCES
35
1. Tejuca, L. G.; Fierro, J. L. G.; Tascón, J. M. D. Structure and reactivity of perovskite-type
oxides. Adv. Catal. 1989, 36, 237-328.
2. Iliev, M. N.; Abrashev, M. V. Raman phonons and Raman Jahn-Teller bands in perovskite-
like manganites. J. Raman Spectrosc. 2001, 32, 805-811.
3. Peña, M. A.; Fierro, J. L. G. Chemical structures and performance of perovskite oxides. Chem.
Rev. 2001, 101, 1981-2017.
4. Pietri, E.; Barrios, A.; Gonzalez, O.; Goldwasser, M. R.; Pérez-Zurita, M. J.; Cubeiro, M. L.;
Goldwasser, J.; Leclercq, L.; Leclercq, G.; Gingembre, L. Perovskites as catalysts precursors for
methane reforming: Ru based catalysts. Stud. Surf. Sci. Catal. 2001, 136, 381-386.
5. Tomishige, K.; Kanazawa, S.; Suzuki, K.; Asadullah, M.; Sato, M.; Ikushima, K.; Kunimori,
K. Effective heat supply from combustion to reforming in methane reforming with CO2 and O2:
comparison between Ni and Pt catalysts. Appl. Catal., A 2002, 233, 35-44.
6. Goldwasser, M. R.; Rivas, M. E.; Pietri, E.; Pérez-Zurita, M. J.; Cubeiro, M. L.; Gingembre,
L.; Leclercq, L.; Leclercq, G. Perovskites as catalysts precursors: CO2 reforming of CH4 on Ln1-
xCaxRu0.8Ni0.2O3 (Ln = La, Sm, Nd). Appl. Catal., A 2003, 255, 45-57.
7. Navarro, R. M.; Alvarez-Galvan, M. C.; Villoria, J. A.; González-Jiménez, I. D.; Rosa, F.;
Fierro, J. L. G. Effect of Ru on LaCoO3 perovskite-derived catalyst properties tested in oxidative
reforming of diesel. Appl. Catal., B 2007, 73, 247-258.
8. Mota, N.; Álvarez-Galván, M. C.; Al-Zahrani, S. M.; Navarro, R. M.; Fierro, J. L. G. Diesel
fuel reforming over catalysts derived from LaCo1-xRuxO3 perovskites with high Ru loading. Int.
J. Hydrogen Energy 2012, 37, 7056-7066.
9. Ivanova, S.; Senyshyn, A.; Zhecheva, E.; Tenchev, K.; Nikolov, V.; Stoyanova, R.; Fuess, H.
Effect of the synthesis route on the microstructure and the reducibility of LaCoO 3. J. Alloys
Compd. 2009, 480, 279-285.
36
10. Ivanova, S.; Senyshyn, A.; Zhecheva, E.; Tenchev, K.; Stoyanova, R.; Fuess, H. Crystal
structure, microstructure and reducibility of LaNixCo1-xO3 and LaFexCo1-xO3 perovskites (0 < x <
0.5). J. Solid State Chem. 2010, 183, 940-950.
11. Chupas, P. J.; Qiu, X.; Hanson, J. C.; Lee, P. L.; Grey, C. P.; Billinge, S. J. L. Rapid-
acquisition pair distribution function (RA-PDF) analysis. J. Appl. Crystallogr. 2003, 36, 1342-
1347.
12. Fauth, F.; Peral, I.; Popescu, C.; Knapp, M. The new material science powder diffraction
beamline at ALBA synchrotron. Powder Diffr. 2013, 28, S360-S370.
13. Peral, I.; McKinlay, J.; Knapp, M.; Ferrer, S. Design and construction of multicrystal
analyser detectors using Rowland circles: application to MAD26 at ALBA. J. Synchrotron Rad.
2011, 18, 842-850.
14. Toby, B. H. EXPGUI, a graphical user interface for GSAS. J. Appl. Crystallogr. 2001, 34,
210-213.
15. Howard, C. J. The approximation of asymmetric neutron powder diffraction peaks by sums
of Gaussians. J. Appl. Crystallogr. 1982, 15, 615-620.
16. Thompson, P.; Cox, D. E.; Hastings, J. B. Rietveld refinement of Debye–Scherrer
synchrotron X-ray data from Al2O3. J. Appl. Crystallogr. 1987, 20, 79-83.
17. Finger, L. W.; Cox, D. E.; Jephcoat, A. P. A correction for powder diffraction peak
asymmetry due to axial divergence. J. Appl. Crystallogr. 1994, 27, 892-900.
18. Stephens, P. W. Phenomenological model of anisotropic peak broadening in powder
diffraction. J. Appl. Crystallogr. 1999, 32, 281-289.
19. Ravel, B.; Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: Data analysis for X-ray
absorption spectroscopy using IFEFFIT. J. Synchrotron Rad. 2005, 12, 537-541.
20. Shirley, D. A. High-resolution X-ray photoemission spectrum of the valence bands of gold.
Phys. Rev. B: Condens. Matter Mater. Phys. 1972, 5, 4709-4714.
37
21. Wagner, C. D.; Davis, L. E.; Zeller, M. V.; Taylor, J. A.; Raymond, R. H.; Gale, L. H.
Empirical atomic sensitivity factors for quantitative analysis by electron spectroscopy for
chemical analysis. Surf. Interface Anal. 1981, 3, 211-225.
22. Bernal, S.; Díaz, J. A.; García, R.; Rodríguez-Izquierdo, J. M. Study of some aspects of the
reactivity of La2O3 with CO2 and H2O. J. Mater. Sci. 1985, 20, 537-541.
23. Laberty-Robert, C.; Fontaine, M. L.; Mounis, T.; Mierzwa, B.; Lisovytskiy, D.; Pielaszek, J.
X-ray diffraction studies of perovskite or derived perovskite phase formation. Solid State Ionics
2005, 176, 1213-1223.
24. Mandrus, D.; Keppens, V.; Chakoumakos, B. C. Spin-glass formation in Co2RuO4. Mater.
Res. Bull. 1999, 34, 1013-1022.
25. Balek, V.; Labhsetwar, N. K.; Mitsuhashi, T.; Haneda, H.; Šubrt, J.; Zeleňák, V. Study of the
preparation of ruthenia based catalytic materials by heating their precursors. J. Mater. Sci. 2004,
39, 3095-3103.
26. Labhsetwar, N. K.; Watanabe, A.; Mitsuhashi, T. New improved syntheses of LaRuO3
perovskites and their applications in environmental catalysis. Appl. Catal., B 2003, 40, 21-30.
27. Thornton, G.; Tofield, B. C.; Hewat, A. W. A neutron diffraction study of LaCoO3 in the
temperature range 4.2 < T < 1248 K. J. Solid State Chem. 1986, 61, 301-307.
28. Bos, J. W. G.; Attfield, J. P. Crystal and magnetic structures of the double perovskite
La2CoRuO6. J. Mater. Chem. 2005, 15, 715-720.
29. Thornton, G.; Orchard, A. F.; Rao, C. N. R. A study of LaCoO3 and related materials by X-
ray photoelectron spectroscopy. J. Phys. C: Solid State Phys. 1976, 9, 1991-1998.
30. Haas, O.; Struis, R. P. W. J.; McBreen, J. M. Synchrotron X-ray absorption of LaCoO3
perovskite. J. Solid State Chem. 2004, 177, 1000-1010.
31. Batista, M. S.; Santos, R. K. S.; Assaf, E. M.; Assaf, J. M.; Ticianelli, E. A. High efficiency
steam reforming of ethanol by cobalt-based catalysts. J. Power Sources 2004, 134, 27-32.
38
32. Devadas, A.; Baranton, S.; Napporn, T. W.; Coutanceau, C. Tailoring of RuO2 nanoparticles
by microwave assisted "Instant method" for energy storage applications. J. Power Sources 2011,
196, 4044-4053.
33. Motin Seikh, M.; Sudheendra, L.; Narayana, C.; Rao, C. N. R. A Raman study of the
temperature-induced low-to-intermediate-spin state transition in LaCoO3. J. Mol. Struct. 2004,
706, 121-126.
34. Goldwasser, M. R.; Rivas, M. E.; Lugo, M. L.; Pietri, E.; Pérez-Zurita, J.; Cubeiro, M. L.;
Griboval-Constant, A.; Leclercq, G. Combined methane reforming in presence of CO2 and O2
over LaFe1-xCoxO3 mixed-oxide perovskites as catalysts precursors. Catal. Today 2005, 107-108,
106-113.
35. Giraudon, J.-M.; Elhachimi, A.; Wyrwalski, F.; Siffert, S.; Aboukaïs, A.; Lamonier, J.-F.;
Leclercq, G. Studies of the activation process over Pd perovskite-type oxides used for catalytic
oxidation of toluene. Appl. Catal., B 2007, 75, 157-166.
36. Armelao, L.; Bandoli, G.; Barreca, D.; Bettinelli, M.; Bottaro, G.; Caneschi, A. Synthesis and
characterization of nanophasic LaCoO3 powders. Surf. Interface Anal. 2002, 34, 112-115.
37. Glisenti, A.; Galenda, A.; Natile, M. M. Steam reforming and oxidative steam reforming of
methanol and ethanol: The behaviour of LaCo0.7Cu0.3O3. Appl. Catal., A 2013, 453, 102-112.
38. Lima, S. M.; Assaf, J. M.; Peña, M. A.; Fierro, J. L. G. Structural features of La 1-xCexNiO3
mixed oxides and performance for the dry reforming of methane. Appl. Catal., A 2006, 311, 94-
104.
39. Zhang, Z. L.; Verykios, X. E. Carbon dioxide reforming of methane to synthesis gas over
Ni/La2O3 catalysts. Appl. Catal., A 1996, 138, 109-133.
40. Slagtern, A.; Schuurman, Y.; Leclercq, C.; Verykios, X.; Mirodatos, C. Specific features
concerning the mechanism of methane reforming by carbon dioxide over Ni/La2O3 catalyst. J.
Catal. 1997, 172, 118-126.
39
t Table of Contents Graphic
3 La2(CO3)3(H2O)8 Rux
O2, D
H2O + CO2
3 La2O2(CO3)2Rux
O2, D
H2O + CO2
Ste
p1
T
350º
C
3LaCo1-xRuxO3Co3O4 3 La2O2(CO3)2Rux
O2, D
CO2
+
O2, D
CO2
+
Ste
p2
T >
700
ºC
40