Surface properties and physicochemical characterizations of a new type of anode material, La1-xSrxCr1-yRuy O3-δ, for a solid oxide fuel cell under methane at intermediate temperature

Article abstract of DOI:10.1006/jcat.2002.3588

Ru is an excellent catalyst for methane steam reforming but its use is limited by its cost and evaporation as RuO_x at high temperature. Lanthanum chromites doped with strontium and ruthenium were studied to develop a new anodic material for a solid oxide fuel cell under methane. To stabilize the ruthenum, it was inserted into the perovskite to form the anode material, La_1-xSr_xCr_1-yRu_yO ₃. Catalytic activity was determined with a test rig using GC as analyzing device. XPS showed that segregation of Sr occurred near the surface, in agreement with EDX analyses which indicated a concentration gradient of Sr within the subsurface. Sr doping allowed the stabilization of Ru in the tetravalent state in the perovskite even during reaction processes. The highest steam reforming efficiency was observed with 5% Ru at > 750°C for a steam/methane ratio of 0.175:1. No carbon deposit was detected and catalytic activity was stable for > 200 hr under reducing conditions.

Full text of DOI:10.1006/jcat.2002.3588

Journal of Catalysis 209, 25–34 (2002)
doi:10.1006/jcat.2002.3588
Surface Properties and Physicochemical Characterizations of a New
Type of Anode Material, La1−xSrxCr1−yRuyO3−δ, for a Solid Oxide
Fuel Cell under Methane at Intermediate Temperature
∗
A.-L. Sauvet,∗,1 J. Fouletier, F. Gaillard,† and M. Primet†
∗
Laboratoire d’Electrochimie et de Physico-Chimie des Mate´riaux et des Interfaces INPG, UJF et CNRS, 1130, rue de la piscine-Domaine
Universitaire, 38402 Saint Martin d’He`res Cedex, France; and †Laboratoire d’Application de la Chimie a` l’Environnement (LACE),
UMR, CNRS 5634, Universite´ Claude Bernard, 43 Bd du 11 novembre 1918, 69 622 Villeurbanne Cedex, France
Received August 13, 2001; revised January 21, 2002; accepted February 18, 2002
of this reaction, which produces local cooling in the cell.
An alternative process, called gradual internal steam re-
forming, has been proposed (2). This concept is based on
a coupling between the steam reforming reaction (Eq. 1)
and the electrochemical hydrogen oxidation (Eq. 2). This
allows the reaction to be delocalized over the entire elec-
trode surface and the steam content to be reduced since
water produced by hydrogen oxidation is used to convert
methane.
In order to develop a new anodic material for a solid oxide
fuel cell (SOFC) operated under methane, we studied lanthanum
chromites doped with strontium and ruthenium. Ruthenium is an
excellent catalyst for methane steam reforming but its use is lim-
ited by its cost and evaporation as RuOx at high temperature.
To stabilize the ruthenium, we inserted it into the perovskite to
form La1−xSrxCr1−yRuyO3. XRD, SEM, TPR, and EDX analyses
revealed the stability of ruthenium in the perovskite, in the tetrava-
lent state. The composition was homogeneous in the entire volume
of the electrode. Nevertheless, XPS analyses indicated a strontium
segregation for a thickness lower than 5 nm. The EDX analyses
also showed a strontium concentration gradient near the subsur-
face. The main result was that no loss of ruthenium was detected
even after treatments under air at 1100◦C. The strontium doping
allows the stabilization of the ruthenium in the perovskite in the
c
CH4 + H2O ꢀ 3H2 + CO
H2 + O2− ꢀ H2O + 2e−
[1]
[2]
Conversion of CO can proceed either via electrochemical
oxidation (Eq. 3),
tetravalent state.
ꢀ 2002 Elsevier Science (USA)
Key Words: perovskite; ruthenium; XPS; TPR; methane reform-
ing; catalytic activity.
CO + O2− ꢀ CO2 + 2e−,
[3]
or by the water–gas shift reaction,
1. INTRODUCTION
CO + H2O ꢀ CO2 + H2,
followed by reaction (Eq. 2).
ThecermetNi–YSZisthecommonlyusedanode. Indeed,
nickel is an excellent catalyst for methane reforming and for
electrochemical hydrogen oxidation. However, at this time,
employment of this anode is limited due to several prob-
lems, for example carbon deposition, which deactivates the
catalyst. Thus, dilution of natural gas is necessary.
Lanthanum chromites have been intensively studied as
promising interconnects for SOFC. The perovskite phase is
stable at high temperature (1000◦C) and across a wide PO2
range (1–10−20 bar) (3). Pure LaCrO3 is a p-type conductor
and shows a low electrical conductivity of ca. 0.6 S · cm−1
at 1000◦C in air. Alkaline earth ion (Mg, Sr, and Ca) dop-
ing enhances the electrical conductivity by two orders of
magnitude. Under hydrogen, Sr-doped LaCrO3 exhibits
the highest conductivity (4). The conduction mechanism of
[4]
One of the significant advantages of a high-temperature
solid oxide fuel cell (SOFC) is its ability to operate on fuels
such as natural gas or biogas. Conversion of methane can be
carried out along different process routes, i.e., external or
internal reforming (1). In the first case, the reforming reac-
tion takes place in a separate reforming reactor, a package
of outside fired tubes filled with nickel or noble metal cata-
lysts. For high methane conversion and to avoid carbon
deposition, steam has to be supplied by exceeding the stoi-
chiometry of the steam reforming reaction. The cost and the
volume of the reactor limit this process. Moreover the flex-
ibility of the SOFC’s response to load change is hindered
by the inertia of the reformer. Internal reforming should
be a more attractive solution. However, its development is
hindered by carbon deposition and by the endothermicity
1 To whom correspondence should be addressed. Fax: (33) 04 76 82 66
70. E-mail: alsauvet@hotmail.com.
25
0021-9517/02 $35.00
c
ꢀ 2002 Elsevier Science (USA)
All rights reserved.

26
SAUVET ET AL.
doped lanthanum chromites is relatively well established. the specific area by the Brunauer–Emmett–Teller (BET)
At high PO2, the substitution of a trivalent ion by a divalent method with nitrogen at 77 K. The powders were prepared
one is electronically compensated for by the formation of by the spray-pyrolysis technique (12) using an ultrasonic
chromium ions in the formal +IV oxidation state. The oxy- atomizer. The precursor solution consisted of lanthanum,
gen deviation from stoichiometry (δ) is close to zero. At low chromium, strontium, and ruthenium nitrates in stoichio-
PO2, δ increases and oxygen vacancies are formed. Charge metric ratio. The final solution concentration was 5 ×
neutrality is maintained by reduction of Cr(IV) to Cr(III) 10−2 mol · L−1. The aqueous solution was atomized by a
(5, 6). Moreover, La1−x Srx CrO3 is mechanically and chem- high-frequency (1.7 MHz) ultrasonic mist generator with
ically stable under reducing atmosphere during a period piezoelectric ceramic transducers. The gas (N2 + O2 mix-
of 1000 h (7). Vernoux (8) has shown that strontium-doped ture) carried the aerosol with a flow rate of 6 L · min−1
lanthanum chromites present interesting activity for hydro- through a tubular furnace heated at 800◦C.
gen electrochemical oxidation but do not catalyze methane
reforming (9).
After synthesis, the powder was sintered at 1100◦C for 1 h
in air to complete the solid-state reaction. The results of the
Some metals (Pt, Rh, Ru) are known to exhibit high ac- chemical analyses (realized in the CNRS Central Analysis
tivity for methane steam reforming. Suzuki et al. (10) have Service, Vernaison) are summarized in Table 1. To quantify
tested Ru/ZrO2 and Ru/Al2O3 cermets under SOFC anode the different elements, preliminary treatments were carried
operational conditions. Ru metal has high steam reforming out: for the lanthanum, the chromium, and the strontium an
reaction activity, carbon deposition resistance, and sinter- attack in perchloronitride was realized; for the ruthenium
ing resistance. Vernoux et al. (11) reached the same con- a fusion in soda at 450◦C was carried out. The analyses
clusion. However, cost and evaporation above 1200◦C to were made by inductive coupling plasma atomic emission
form RuO4 impede the use of ruthenium as a SOFC anode spectroscopy (IC-PAES).
material.
2.2. Characterization Techniques
The aim of this paper is to characterize anode materials
obtained by insertion of ruthenium into La1−x Srx CrO3 to
stabilize the ruthenium ion. We have synthesized the tested
powders by the pyrosol technique (12). The catalytic activ-
ity was determined with a test rig using gas chromatography
as analyzing device (13). The highest steam reforming ef-
ficiency was observed with 5% of ruthenium above 750◦C
for a steam/methane ratio equal to 0.175. No carbon depo-
sition was detected and the catalytic activity was stable for
more than 200 h under reducing conditions.
X-ray diffraction (XRD) patterns of the powders and
electrode deposits, sintered at 1100◦C for 4 h in air, were
recorded with a Siemens diffractometer using λKα(Cu) =
˚
1.5406 A. The step scans were taken over the range 2θ an-
gles from 20 to 80◦ with 0.04◦ steps.
TPR (temperature-programmed reduction) analyses
were carried out on LaCrO3 (LC), LS30, LS30R5, and
LS30R20. Each sample (100 mg) was placed on a fixed
quartz bed in a reactor and flushed with helium gas (20 ml
(STP) min−1) for 15 min at room temperature. Then, the
sample was pretreated in oxygen (20 ml (STP) min−1) from
The stabilization of the ruthenium ion in the perovskite
powder under reducing conditions at high temperature was
evidenced by various analysis techniques.
room temperature to 600◦C at a heating rate of 5◦C · min−1
.
Afterward, the sample was cooled to room temperature in
the same atmosphere and the reactor purged by pure He for
15 min. A 1% H2/He gas mixture (30 ml (STP) min−1) was
passed through the sample heated from room temperature
2. EXPERIMENTALS METHODS
2.1. Catalyst Preparation
Table 1 gives the list of the investigated materials. The to 900◦C at a constant rate of 20◦C · min−1, followed by a
particle sizes were measured by laser granulometry and plateau of ca. 30 min at 900◦C. The species followed were
TABLE 1
Investigated Materials and Results of Chemical Analyses
a
Sample
Symbols
%wt La
%wt Sr
%wt Cr
%wt Ru
Calculated formula by assuming O3
La0.8Sr0.2Cr0.95Ru0.05O3
La0.7Sr0.3Cr0.95Ru0.05O3
(1st synthesis)
LS20R5
LS30R5
49.15
43.90
7.10
10.69
20.74
21.07
2.49
2.44
La0.81Sr0.19Cr0.942Ru0.058O3
La0.72Sr0.28Cr0.945Ru0.055O3
La0.7Sr0.3Cr0.95Ru0.05O3
(2nd synthesis)
LS30R5
43.58
10.85
19.90
2.32
La0.72Sr0.28Cr0.944Ru0.056O3
La0.7Sr0.3Cr0.8Ru0.2O3
La0.6Sr0.4Cr0.95Ru0.05O3
LS30R20
LS40R5
42.51
38.46
10.38
14.70
17.08
21.62
7.20
2.51
La0.72Sr0.28Cr0.82Ru0.18O3
La0.62Sr0.38Cr0.944Ru0.056O3
a O was not analyzed.

NEW TYPE OF ANODE MATERIAL FOR SOFC UNDER METHANE
27
H2, CH4, H2O, N2 or CO, O2, and CO2 (m/e = 2, 15, 18, 28, tween 750 and 850◦C. The methane content and the steam
32, and 44 uma, respectively). For the TPD (temperature- content were, respectively, changed between 5 and 30% for
programmed desorption) experiment, the same procedure methane, 1.5 and 3.5% for steam. The experimental setup
was employed, but a pure He stream was used instead of was described in more detail in (13). Steam was the minor
1% H2/He.
reactive. To describe the activity, three parameters were
EDX analyses were carried out on the LS30R5 and used: the steam conversion (Eq. 5), the methane conver-
LS30R20 samples. The powders were milled and coated sion (Eq. 6), and the carbon monoxide selectivity (Eq. 7).
in organic resin. The resin was cut up in fine slices of 80 nm.
ꢀ
ꢁ
P
(H2O)f
The slices were deposited on a grid to analyze 200-nm zones.
Transmission electronic microscopy (TEM) and energy-
dispersive X-ray (EDX) analyses were carried out.
CH O = 100 × 1 −
,
[5]
2
P
(H2O)i
where P(H2O) is the initial steam pressure and P(H2O)f is the
i
Thermogravimetric analyses (ATG) of LS30R5 were per-
formed by placing 60 mg of a powder in a platinum crucible
and increasing the temperature in a 5% H2/Ar gas mix-
ture from−r1oom temperature to 950◦C at a constant rate of
final steam pressure.
ꢀ
ꢁ
P
(CH4)f
CCH = 100 × 1 −
,
[6]
4
P
(CH4)i
4◦C · min
.
Photoelectron spectra (XPS) were measured with a
XR3E2 model vacuum generator. The LS30R5 and LS40R5
powders were mixed with YSZ powder at a 75/25 volume
ratio to improve the adherence of the deposit on the
YSZ pellet. Anode slurries, made by adding ethanol,
polyvinyl butyral, polyvinyl pyrrolidone, and terpineol,
where P(CH4) is the initial methane pressure and P(CH4) is
i
f
the final methane pressure.
ꢀ
ꢁ
PCO
SCO = 100 ×
,
[7]
PCO + P
CO2
were pulverized by spray on an 8YSZ plate, 1.5 mm thick where PCO is the final carbon monoxide pressure and PCO2
and 18 mm in diameter. All the deposits were sintered at is the final carbon dioxide pressure.
1100◦C for 4 h in air. Before recording the spectra, the
pellet was sintered under different conditions, summarized
in Table 2. LS30R5-ref is a pellet only sintered at 1100◦C for
4 h in air, without treatment in reducing atmosphere. All
binding energies were referred to the C 1s line at 285 eV.
Microprobe analyses were carried out on LS30R5-ref,
LS30R5-H2, and LS30R5-CH4. The samples were covered
by organic resin and polished until mirror face. We mea-
sured the ruthenium content within the electrode thickness,
near the surface, in the middle of the deposit, and near the
electrolyte surface.
3. RESULTS AND DISCUSSION
3.1. Chemical Analysis
As seen in Table 1, the powders compositions matched
the desired stoichiometry. No loss of ruthenium was ob-
served even after sintering in air at 1100◦C and under cata-
lytic activity measurement condition, i.e., under more than
200 h in methane at 850◦C.
The specific area was 2 m2 · g−1 for all these powders.
From this result, we have calculated the grain size. The
density of these powders was measured by a picometer and
was around 3.8 g · cm−3. Assuming a spherical geometry
2.3. Catalytic Activity
3
(R = d×S , with R the grain radius, d the powder density,
The catalytic activities of these powders were evaluated
by analyzing the products by gas chromatography. The vec-
tor gas used was helium; the operating temperature was be-
and S the specific area), we obtained 0.3 µm as grain radius.
2
In the case of cylindrical geometry (R = d×S ), the grain ra-
dius was 0.25 µm. In summary, the grain size was around
0.5–0.6 µm.
TABLE 2
3.2. X-Ray Diffraction
Preliminary Treatments of Samples Studied by XPS
In Fig. 1 are compared the X-ray diffraction patterns of
LS30, LS20R5, LS30R5, and LS40R5 powders, sintered 1 h
at 1100◦C in air. No intermediate phase containing ruthe-
nium was detected. An intermediate phase, SrCrO4, was
detected around 25–30◦. However, as shown in Fig. 2, no
intermediate phase was observed on the X-ray diffraction
pattern of electrode deposits after sintering at 1100◦C for
4 h in air. The additional peaks observed on the LS30R5
deposit spectra correspond to YSZ added in the ink to im-
prove the adherence of the deposit on the YSZ pellet.
Sample
Preliminary treatments
Symbol
LS30R5
LS30R5
Sintered at 1100◦C 4 h in air
LS30R5-ref
LS30R5-H2
Sintered at 1100◦C 4 h in air and
treated at 850◦C, under H2 for 15 h
Sintered at 1100◦C 4 h in air and
treated at 850◦C, under CH4 for 15 h
Sintered at 1100◦C 4 h in air and
treated at 850◦C, under H2 for 15 h
Sintered at 1100◦C 4 h in air and
treated at 850◦C, under CH4 for 15 h
LS30R5
LS40R5
LS40R5
LS30R5-CH4
LS40R5-H2
LS40R5-CH4

28
SAUVET ET AL.
4500
4000
3500
3000
2500
2000
1500
1000
500
Si
Si
Si
Si
SrCrO
4
LS40R5
LS30R5
LS20R5
LS30
0
20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 80
2θ (scales)
FIG. 1. X-ray diffraction patterns obtained for LS30, LS20R5, LS30R5, and LS40R5 powders.
Crystallographic parameters are summarized in Table 3; activity of LaCr0.95Ru0.05 at 800◦C under a steam/methane
they are determined in a hexagonal unit cell with 0.025% ratioequalto0.75. Theinitialactivityofthispowderwaslow
√
error. The volume of a hexagonal system is V = ( 3/2)a2c. and increased until 100% of steam conversion after 40 h of
X-ray indexes verify the following relation: −h + k +l = 3n. analysis. This behavior was explained by a progressive and
The system has not a hexagonal unit cell but a rhomboedric reversible reduction of the ruthenium in the perovskite. In
one. Ruthenium substitution increases the volume. How- our studies, the steam conversion with LS20R5 or LS30R5
ever, with regard to X-ray spectra (few rays and sweeping was immediately equal to 100% for the whole ratio range
only to 80◦), complementary studies should be carried out studied, i.e., 0.07 and 0.7, and this reaction was limited by
to determine the limit of ruthenium insertion and the pos- steam content. We have not observed an increase in the
sible modification of the crystallographic parameters. The catalytic activity. So, the strontium doping seems to influ-
crystallite size was calculated by the half-width and was ca. ence the ruthenium stability in the perovskite and also the
30 nm.
catalytic activity. However, we have observed no difference
between LS20R5 activity and LS30R5 activity. We obtained
100% steam conversion down to 750◦C.
3.3. Catalytic Activity
In all the cases, no carbon deposition was detected and
We observed that LS20R5 and LS30R5 have very simi- no degradation of the catalytic activity was observed after
lar catalytic activities. Vernoux (9) has studied the catalytic more than 200 h under operating conditions.
4000
3500
3000
2500
2000
1500
SrCrO4
YSZ
1000
LS30R5
YSZ
500
0
YSZ
YSZ
LS30R5 deposit
20
26
32
38
44
50
56
62
68
74 80
2θ (scales)
FIG. 2. X-ray diffraction patterns obtained for LS30R5 powder and LS30R5 deposit.

NEW TYPE OF ANODE MATERIAL FOR SOFC UNDER METHANE
TABLE 3
29
Calculated Crystallographic Parameters of the Studied Samples
in the Hexagonal System
❛
❛
❛
Volume (A3)
˚
c (A)
Sample
a (A) b (A)
La0.75Sr0.25CrO3, data JCPDS 5.403
5.403 13.301
5.507 13.562
5.494 13.438
5.500 13.425
5.496 13.330
336.27
356.19
351.27
351.70
348.70
LS20R5
LS30R5
LS30R20
LS40R5
5.507
5.494
5.500
5.496
3.4. Scanning Electronic Microscopy Analysis (SEM)
FIG. 4. TPR spectra of LC, LS30, LS30R5, and LS30R20. Profile of
signal at m/e = 2 amu.
Nosignificantdifferencewasobservedasafunctionofthe
strontium content. The grain size was around 0.2–0.5 µm,
with a few 1-µm agglomerates. These results agree with the
grain size measured from the specific area.
As shown in Fig. 3 (EDX cartography), all the elements
are homogeneously distributed. No segregation of ruthe-
nium was observed.
Compared to the results obtained for LC, it seems
that the hydrogen quantity consumed for the reduction of
strontium-doped samples cannot be explained by the re-
duction of Cr(III) to Cr(II). Moreover, ruthenium has no
influence on the hydrogen quantity necessary for the re-
duction. The quantity of hydrogen consumed was close to
the strontium content (Tables 4 and 5). The difference ob-
served between the hydrogen quantity consumed and the
strontium micromole quantity results probably from the
secondary-phase SrCrO4, observed on RX spectra real-
ized with these powders sintered at 1100◦C during 1 h in
air.
As the reduction peak was not observed with LC pow-
der, it could be related to strontium. This peak could be
ascribed to the reduction of Cr(IV), which is produced to
compensate electronically for the substitution of a triva-
lent ion La(III) with a divalent Sr(II) according to the
3.5. Temperature-Programmed Reduction (TPR)
Figure 4 shows the TPR spectra run on for samples LC,
LS30, LS30R5, and LS30R20 under 1% H2/He. In the range
25–400◦C, no chemical reduction was detected. Undoped
sample (LC) revealed a weak reduction peak at 530◦C.
Strontium-doped samples had a large reduction peak which
started at 550◦C and which was incomplete at 900◦C. With
ruthenium-doped samples, a shift of the peak toward lower
temperatures was observed. The hydrogen quantities con-
sumed at ambient temperature and during the reduction
treatment at 900◦C are given in Table 4.
FIG. 3. EDX cartography of LS30R20.

30
SAUVET ET AL.
TABLE 4
1.E-10
4.E-10
4.E-10
3.E-10
3.E-10
2.E-10
TPR
TPR Quantitative Analyses
8.E-11
6.E-11
4.E-11
2.E-11
0.E+00
H2 consumed
during the
H2 consumed
at 25◦C
reduction peak
Sample
(µmol H2 · g−1
)
(µmol H2 · g−1
)
LC
LS30
LS30R5
LS30R20
3
25
1749
1619
1873
Not detectable
1
97
0
200
400
600
800
Temperature / °C
Kro¨ger–Vink notation (14):
1.E-10
8.E-11
6.E-11
4.E-11
2.E-11
0.E+00
3.E-09
2.E-09
1.E-09
0.E+00
TPD
(1 − x)LaO1.5 + xSrO + CrO1.5
= (1 − x)LaLx a + xSrLꢁ a + (1 − x)CrCx r + xCr•Cr
+ (3 − 0.5x)OOx + 0.5VOx .
[8]
Similar results have previously been described with
strontium-substituted lanthanum manganites (15) and with
calcium-substituted lanthanum chromites (16).
The ruthenium, Ru(IV), was substituted for Cr(IV). The
reduction peak does not decrease as a function of ruthe-
nium content.
0
200
400
600
800
Temperature / °C
(1 − x)LaO1.5 + xSrO + (1 − y)CrO1.5 + yRuO2
= (1 − x)LaLx a + xSrLꢁ a + [1 − (x + y)]CrCx r + xCr•Cr
+ yRu•Cr + [3 − 0.5(x − y)]OOx + 0.5(x − y)VOx . [9]
FIG. 5. TPD and TPR spectra of LS30R20. Profiles of signals at
m/e = 44 and 32 amu.
Moreover, in comparing the chemical potential of
Cr2O3/CrO2 (µ◦ = +10.575 kJ) and Ru2O3/RuO2 (µ◦ =
−56.304 kJ) at 25◦C, Cr(IV) would be reduced before
Ru(IV).
The co-doping with Sr(II) is therefore compensated for
by the formation of Cr(IV). The ruthenium(IV) is stabilized
because the Cr(IV) is more easily reduced. Thisexplanation
only holds if the amount of hydrogen consumed is totally
used to reduce Cr(IV) to Cr(III). More particularly, one can
consider that some carbonate species are present in the bulk
of the perovskite (or more likely on its surface) and may
be thermally decomposed, leading to the evolution of CO2.
Carbon dioxide could then react with hydrogen present
in the stream to form CO + H2O (reverse water–gas shift
reaction) or CH4 +H2O (methanation). The observation of
profiles at m/e = 15 and 28 amu indicates that neither CH4
nor CO evolution is detected within the temperature range
where reduction occurs. The profiles reported in Fig. 5 are
relative to signals at m/e = 32 and 44 amu (related to O2
and CO2, respectively), under TPR and TPD conditions,
for a LS30R20 sample. TPD is realized under a stream of
pure He. Under both TPR and TPD conditions, a peak re-
lated to CO2 evolution is observed within the temperature
range where reduction occurs. The CO2 profiles are quite
similar in H2/He and in pure He. The quantities evolved
are 5.6 and 6.5 µmol of CO2/g, respectively. These amounts
are very small and roughly correspond to the amount
of Sr expected at the sample surface. The absence of
methane production from CO2 during TPR is consistent
with the presence of nonmetallic ruthenium species. Like
nickel-based ones (17), ruthenium-based catalysts may
be used successfully for CO2 methanation (18) but they
necessitate Ru to be in the reduced state (19). Profiles
at m/e = 32 amu are very different under TPR and TPD
conditions. In the TPR experiment, no peak is detected.
Conversely, under TPD conditions, a strong peak related
to O2 evolution is observed, for a temperature higher than
700◦C or so. The amount of oxygen detected is 280 µmol
TABLE 5
Compositions of the Samples Calculated from the Chemi-
cal Analysis, Except for LS30, Which Was the Theoretical Com-
position
La
Sr
Cr
Ru
M
Sample (µmol · g−1) (µmol · g−1) (µmol · g−1) (µmol · g−1) (g · mol−1
)
LS30
LS30R5
LS30R20
3132
3168
3085
1342
1232
1200
4474
4153
3513
0
246
771
223.52
227.29
233.38

NEW TYPE OF ANODE MATERIAL FOR SOFC UNDER METHANE
TABLE 6
31
3.7. XPS Experiments
Compositions (at%) of the Different Elements Detected
by EDX for LS30R5
In spite of the deposit characteristics, i.e., porous and
not smooth, microprobe analysis revealed no ruthenium
or strontium concentration gradient with LS30R5-H2 and
LS30R5-CH4. The composition was homogeneous over the
whole thickness of the deposit, around 40 µm. Complemen-
taryresultshavebeenobtainedbyXPS, whichprovidemore
surface-sensitive data (around 0.5–5 nm).
In Fig. 6 are represented the ratio between the atomic
percentage determined for each element and the La + Sr
+ Cr + Ru sum. This representation allows one to elimi-
nate the O 1s and C 1s participation and to highlight pos-
sible surface segregation. Several observations could be
made.
Theor at%
(zone)
O
(60)
Cr
(19)
Sr
(6)
Ru
(1)
La
(14)
Area
1
8
9
54.9
59.9
53.7
60.2
56.0
59.9
72.0
24.1
21.7
23.9
21.8
22.7
20.4
13.4
3.2
2.8
3.4
2.7
3.3
3.3
1.9
1.8
1.4
1.6
1.3
2.1
1.5
1.6
16
14.2
13.9
16.0
14.9
11.0
10
11
12
13
14
2
15
16
17
18
19
58.7
56.1
59.6
56.4
60.1
22.1
22.6
20.5
23.3
20.6
3.2
3.6
2.9
3.1
2.4
1.2
1.4
1.6
1.4
1.2
14.7
16.3
15.4
15.8
15.7
• For LS30R5-ref, the ratios are similar to those obtained
for the bulk.
• The treatments in reducing atmosphere (hydrogen or
methane) modify the surface composition with a large de-
crease in lanthanum content and an increase in strontium
content for LS30R5 and for LS40R5.
• The ratio (La + Sr)/sum is constant for all the samples,
near the theoretical value of 0.5. The perovskite has the
structure ABO3. The increase in strontium content occurs
to the detriment of lanthanum content, and the balance
between elements on site A (La and Sr) and the elements
on site B (Cr or Ru) is maintained.
of O2/g and therefore would necessitate 560 µmol of H2/g
to be fully converted to H2O, as suggested by profiles in
TPR conditions. This quantity, not employed for Cr(IV)
reduction into Cr(III), corresponds to the excess of
hydrogen consumption observed when comparing data
from Tables 4 and 5 (H2 consumed during reduction peak
with respect to Sr concentration).
3.6. TEM and EDX Experiments
• For elements on site B (Cr and Ru), treatment un-
der hydrogen or under methane seems to favor a slight in-
crease in ruthenium content and the concomitant decrease
in chromium content.
The results obtained by EDX are summarized in Table 6
for LS30R5 and in Table 7 for LS30R20. The measured
atomic compositions are similar to those expected, except
for strontium. A deficit of strontium compared to the
theoretical atomic percentage was obtained. The main
point concerns the ruthenium content: no loss of ruthenium
was observed even for LS30R20 even after calcination in
air at 1100◦C.
The lanthanum spectra are similar to that measured
for La2O3. Energy position of La 3d5/2 (834.8 eV) and
spin-orbit splitting (16.8 eV) correspond exactly to that
0.6
0.5
0.4
0.3
0.2
0.1
0
TABLE 7
Compositions (at%) of the Different Elements Detected
by EDX for LS30R20
Theor at%
(zone)
O
Cr
Sr
(6)
Ru
(4)
La
(14)
Area
1
(60)
(16)
7
8
9
10
11
12
13
33.9
46.9
46.5
53.3
61.3
31.3
45.2
28.5
25.1
25.6
25.6
15.7
35.2
27.3
3.9
3.1
3.7
4.3
2.2
4.8
4.4
8.0
4.7
3.8
1.9
3.8
3.9
2.7
25.6
20.2
20.3
16.9
17.0
24.7
20.4
LS30R5- LS30R5-ref LS30R5-H2 LS30R5-
bulk CH4
LS40R5- LS40R5-H2 LS40R5-
bulk CH4
2
14
15
16
17
18
19
46.0
55.4
61.6
53.4
51.4
53.8
25.5
24.8
19.9
18.0
17.3
19.1
3.8
2.7
2.6
5.3
6.2
4.3
4.4
1.3
1.3
7.1
7.8
6.5
20.2
15.9
14.6
16.2
17.3
16.3
%Sr / (%La + %Sr + %Cr + %Ru)
%La / (%La + %Sr + %Cr + %Ru)
FIG. 6. Ratio between the atomic percentages of Sr and La and the
sum (La + Sr + Cr + Ru) determined by XPS analyses.

32
SAUVET ET AL.
of La2O3 (20). The different treatments do not modify 3.8. Thermogravimetric Analyses (TGA)
form and position of peaks. The 3d5/2 and 3d3/2 stron-
Two plateaus were observed; the first weight loss began at
tium peaks obviously result from a mixture of at least
two chemical species. As a function of the sample, peak
maximum is between 134.5 and 135.2 eV, with no ap-
parent relation to preliminary treatment. In the litera-
ture, we find 134.4 eV for Sr and 135.3 eV for SrO (21)
but we have also to consider species such as strontium
carbonates. For chromium, we observed essentially a Cr
2p3/2 peak maximum at 576.5 eV, corresponding to Cr(III)
(22, 23). The spectrum of LS30R5-ref exhibits two weak
peaks, at 580.1 and 589.7 eV. They can be attributed to
the presence of Cr(VI), as reported by some authors for
La1−x Srx CrO3 compounds obtained by the sol–gel route
(24). Chromium(VI) compounds are characterized by Cr
2p3/2 peaks ranging from 578.3 to 580.1 eV, according to
the chemical form, i.e., CrO3, CrO24−, or Cr2O72− (22, 25).
No trace of Cr(VI) can be detected after treatment under
hydrogen or methane. Moreover, on LS30R5-ref, it seems
that the peak shape at the low binding energy side indi-
cates the presence of Cr(IV) (26). The determination of
the ruthenium chemical state by XPS is not easy, as the Ru
3d5/2 peaks for Ru(0) and Ru(IV) are very close and the
spin-orbit splitting is the same (27, 28). Moreover, C 1s and
Sr 3p1/2 peaks interfere with Ru 3d3/2 and 3d5/2, respec-
tively.
Figure 7 shows the oxygen spectra for the different sam-
ples. For LS30R5-ref, oxygen in structure (529.9 eV) is es-
sentially observed. This result is in agreement with values
obtained for LaMn1−x Cux O3 (29), 529.4–529.8 eV, and for
La1−x Srx MnO3 (20), 529.3–529.9 eV. After treatment un-
der hydrogen or methane, other structures at higher bind-
ing energy appear. This phenomenon is independent of the
gas nature (hydrogen or methane) and increases with stron-
tium content. It seems to be correlated with oxygen species
adsorption on oxygen vacancies, during transfer in air, after
reducing treatments. As an example, Tejuca et al. (30) ob-
served 533.2 eV for adsorbed water molecules and 531.1 eV
for hydroxyl groups.
250◦C and the second one at 800◦C. This last temperature is
200◦C higher than the beginning of hydrogen consumption
observed by TPR. However, the temperature of the second
TGA plateau corresponds to the beginning of water evolu-
tion, at about 750◦C, observed during the TPR experiments.
The temperature difference between the beginning of re-
duction and the beginning of water evolution (evidenced by
both TGA and MS) precludes the possibility of a direct for-
mation/desorptionofwatermolecules. Itisplausiblethatin-
termediate hydroxyl species (resulting from water dissocia-
tive adsorption) are formed after reduction of metallic ions,
which leads to water desorption at higher temperatures, ac-
cording to condensation of adjacent hydroxyl groups. The
total mass loss was around 1.8 mg (3 wt%), corresponding
to a deviation of the stoichiometry of 0.14. This mass loss is
in good agreement with the mass of water, determined from
the quantity of hydrogen consumed, evolved above 750◦C
during TPR experiment. Larsen et al. (31) have studied the
conduction mechanism in La1−x Srx Cr1−yVyO3−δ as a func-
tion of the oxygen pressure. At low PO2, the oxygen loss is
electronically compensated for by Cr(IV) reduction. These
authors show that in considering the oxide chemical po-
tentials and assuming that Cr(IV) and V(IV) are stabilized
similarly in the perovskite lattice, the reduction of Cr(IV)
is expected to occur at much higher PO2 than that of V(IV).
In this case, δ increases with decreasing PO2 and reaches a
saturation value equal to (x − y)/2 (see Eq. 9). If we calcu-
late the saturation value of δ in our case, we obtain 0.125.
This value is close to the experimental result (δ = 0.14). This
observation confirms that the ruthenium is in a tetravalent
state.
3.9. Interpretation
The main use of ruthenium metal, in common with the
other group VIII metals, is as a catalyst. In solution, ruthe-
nium is employed in the tetravalent oxidation state (32–34),
the most stable form of ionic ruthenium.
In reactions between a gas and a solid, for example in
methane oxidation or methane reforming by carbon diox-
ide or steam, ruthenium is metallic and supported on SiO2
or Al2O3. Matsui et al. (35) studied the reaction mechanism
of carbon dioxide reforming of methane with Ru supported
on La2O3, Y2O3, ZrO2, and Al2O3. During the reaction, a
part of metallic ruthenium reacted with CH4 to give Ru–
CHx ; simultaneously, ruthenium metal could be oxidized
by CO2 to give Ru–Ox and CO; then, oxygen transfer from
Ru–Ox and Ru–CHx took place to give CO and metal-
lic ruthenium. The results of XRD analyses of Ru/La2O3
catalyst after CO2 reforming exhibited diffraction peaks
assignable to RuO2 and Ru metal. Guerrero-Ruiz et al. (36)
carried out isotopic experiments in syngas production from
25000
LS40R5-CH4
LS40R5-H2
20000
LS30R5-CH4
LS30R5-H2
15000
10000
LS30R5-Ref
5000
540
535
530
525
Binding energy / eV
FIG. 7. Comparison of the oxygen XPS spectra obtained for the dif-
ferent samples.

NEW TYPE OF ANODE MATERIAL FOR SOFC UNDER METHANE
33
methane on Ru/Al2O3 and Ru/SiO2. They observed that catalytic mechanism in which metallic ruthenium activates
an initial oxidation of the Ru surface would probably de- reactants.
lay the formation of H2 and CO. They also indicated that
some authors propose that the most active form under dry solid-state aspects of oxidation catalysis, showed the influ-
reforming conditions was oxidized noble metal forms. ence of the solid-state effects on the catalytic activity. The
Gellings and Bouwmeester (40), in a recent study of the
Several studies have shown the influence of the catalyst main parameters of the catalytic activity are the following:
supports on the catalytic activity. As an example, Ferreira-
• nature of conduction (ionic, electronic or mixed)
Aparicio et al. (37) studied some mechanistic aspects of the
• nature of the point defects
dry methane reforming over Ru/SiO2 and Ru/Al2O3. The
• interaction between defects
support has great influence on the activity for dehydrogena-
• oxygen vacancies formation in reducing atmosphere
tion of methane and on the kind of carbonaceous species.
In this connection, Driscoll et al. (41), who studied the
methane oxidation on Li/MgO, showed that O− ions at the
surface activated methane reaction and reacted with Li+ to
form bidefects as [Li+O−]. In the same way, Steghuis et al.
(42) observed that active sites for partial methane oxidation
with TiO2 are O− at the surface of the catalyst. Saracco et al.
(43) studied the methane decomposition with magnesium-
doped lanthanum chromites. The magnesium present in the
perovskite structure leads to a significant increase in the
catalytic activity. They suggested that magnesium is located
on the chromium sites of the perovskite according to MgCꢁ r
which is accompanied by charge compensation, i.e., an in-
crease in the charge of chromium, which are then present as
Cr•Cr. It is proposed that the increase in the catalytic activity
can be explained by higher oxygen reactivity near the CrC• r
defects.
In our stud•y, strontium doping is compensated for by for-
mation of CrCr. Under reducing conditions, oxygen vacan-
cies are produced and Cr(IV) is reduced to Cr(III). On the
surface of the catalyst, oxygen vacancy formation could be
neutralized by strontium segregation within the subsurface.
The oxygen vacancies could assist the steam dissociation in
hydroxyl groups. Moreover, after reaction under methane,
we have observed an increase in hydroxyl group concen-
tration at the surface and a decrease in oxygen vacancies
compared to the values obtained under hydrogen.
With Ru/SiO2, activation of both reactants takes place on
the ruthenium surface. The accumulation of carbon ad-
species formed from methane decomposition on the metal-
lic particles finally impedes carbon dioxide dissociation and
induces rapid deactivation of the catalyst. In contrast, with
Ru/Al2O3, the alumina support participates in the activa-
tion of carbon dioxide by supplying hydroxyl groups (38,
39) according to
OHAl O + Ru → Ru–OH,
2
3
1
2
Ru–OH + Ru–C → CO + H2 + Ru.
These authors proposed that the dry reforming of methane
occurs through a Langmuir–Hinshelwood-type mechanism
for Ru/SiO2 in which both reactants are activated on the
metallic phase. In contrast, a bifunctional mechanism, in
which surface hydroxyl groups play a major role in the re-
sistance to deactivation, is proposed for Ru/Al2O3.
Under our operating conditions of methane steam re-
forming, ruthenium is not metallic and not supported but
is inserted as an ionic species into strontium-doped lan-
thanum chromites. The powder and deposit characteriza-
tions and the catalytic experiments have indicated the fol-
lowing:
• Metallic ruthenium is not detected even after reaction;
TPR experiments allow detecting the presence of Cr(IV);
ruthenium in tetravalent state would be substituted for
Cr(IV); the quantity of hydrogen consumed during TPR
was not related to the ruthenium content; Cr(IV) was more
easily reduced than Ru(IV).
• Strontium is the key factor in explaining the stabi-
lization of ionic ruthenium in the perovskite structure;
Vernoux (9) observed a progressive and reversible reduc-
tion of ruthenium when LaCr0.95Ru0.05O3 was used as a
catalyst; the behavior difference can be explained only by
the strontium doping.
4. CONCLUSION
All the experimental results obtained by the different
techniques indicate that the ruthenium ion is stabilized in
the tetravalent state in the matrix. Moreover, the XPS ana-
lyses reveal that segregation of strontium occurs near the
surface. This result is in agreement with the EDX analyses
that indicate a concentration gradient of strontium within
the subsurface. We cannot describe the different steps of
the mechanism but some essential results were observed
and they will allow optimization of new catalysts.
• Strontium segregation was observed by XPS experi-
ments; we observed also that, after treatment under reduc-
ing condition, other oxygen structures appear.
• Oxygen vacancies play a part in the catalytic mecha-
nism.
• Strontium doping is compensated for by the formation
The catalytic activity of strontium- and ruthenium-doped of Cr(IV), and under reducing conditions, Cr(IV) is more
lanthanum chromites could not be explained by the usual easily reduced than Ru(IV); thus, the strontium doping

34
SAUVET ET AL.
allows one to stabilize the ruthenium in the tetravalent state 17. Twigg, M. V., “Catalyst Handbook,” 2nd ed., Manson, London, 1996.
18. Fujita, S. I., and Takezawa, N., Chem. Eng. J. 68, 63 (1997).
19. Gupta, N. M., and Kamble, V. S., Indian J. Chem. Sect. A 35, 557 (1996).
20. Yong, W. Y., and Thorn, R. J., J. Chem. Phys. Solids 41, 75 (1980).
21. Van Dovere, H., and Verhoeven, J. A., J. Electron Spectrosc. Rel. Phe-
in the perovskite even during reaction processes.
• The activity of this catalyst for methane steam reform-
inginamethane-richatmosphere(steam/methaneratioless
than 1) is similar to that of ruthenium metal; however, we
nom. 21, 265 (1980).
avoid ruthenium loss during preliminary treatment and the 22. Halada, G. P., and Clayton, C. R., J. Electrochem. Soc. 138, 2921
(1991).
agglomeration of ruthenium particles during reaction.
23. Battistoni, C., Dormann, J. L., Fiorani, D., Paparazzo, E., and
Viticoli, S., Solid State Commun. 39, 581 (1981).
REFERENCES
24. Liu, X., Su, W., Lu, J., Pei, L., and He, L., J. Alloys Compounds 305,
21 (2000).
1. Ledjeff, K., Rohrbach, T., and Schaumberg, G., in “Proceedings of
the 2nd International Symposium of SOFC’s” (F. Grosz, P. Zegers,
S. C. Singhal, and O. Yamamoto, Eds.), p. 323; Electrochemical Society,
Pennington, NJ, 1991.
2. Vernoux, P., Guindet, J., and Kleitz, M., J. Electrochem. Soc. 145, 3487
(1998).
25. Shuttleworth, D., J. Phys. Chem. 84, 1629 (1980).
26. Cimino, A., De Angelis, A., Luchetti, A., and Minelli, G., J. Catal. 45,
316 (1976).
27. Nyholm, R., and Martensson, N., Solid State Commun. 40, 311 (1981).
28. Kim, Y. J., Gao, Y., and Chambers, S. A., Appl. Surf. Sci. 120, 250
(1997).
3. Nakamura, T., Petzow, G., and Gauckler, L. J., Mater. Res. Bull. 14, 29. Tabata, K., Hirano, Y., and Suzuki, E., Appl. Catal. A 170, 245
649 (1979).
(1998).
4. Mori, M., Yamamoto, T., Itoh, H., and Abe, T., in “Proceedings of
the 2nd European SOFC Forum” (B. Thorstensen, Ed.), p. 465. Oslo,
1996.
5. Larsen, P. H., Hendriksen, P. V., and Mogensen, M., J. Thermal Anal.
49, 1263 (1997).
30. Tejuca, L. G., Bell, A. T., Fierro, J. L. G., and Pena˜, M. A., Appl. Surf.
Sci. 31, 301 (1998).
31. Larsen, P. H., Hendriksen, P. V., and Mogensen, M., J. Therm. Anal.
49, 1263 (1997).
32. Caron, V., Thesis. INPG, Grenoble, 2000.
6. Mizusaki, J., Yamauchi, S., Fueki, K., and Ishikawa, A., Solid State
Ionics 12, 119 (1984).
7. Milliken, C., Emangovan, S., and Khandler, A., in “Proceedings
SOFC’s III” (S. C. Singhal and H. Iwahara, Eds.), p. 335. Electro-
chemical Society, Pennington, NJ, 1993.
33. Ungermann, C., Landis, V., Moya, S. A., Cohen, H., Walker, H.,
Pearson, R. G., Rinker, R. G., and Ford, P. C., J. Am. Chem. Soc.
101, 5922 (1979).
34. Atwood, D. K., and De Vries, T., J. Am. Chem. Soc. 84, 2659 (1962).
35. Matsui, N., Anzai, K., Akamatsu, N., Nakagawa, K., Ikenaga, N., and
Suzuki, T., Appl. Catal. A 179, 247 (1999).
8. Vernoux, P., Ionics 3, 270 (1997).
9. Vernoux, P., Thesis. INPG, Grenoble, 1998.
36. Guerrero-Ruiz, A., Ferreira-Aparicio, P., Bachiller-Baeza, M. B., and
Rodriguez-Ramos, I., Catal. Today 46, 99 (1998).
37. Ferreira-Aparicio, P., Rodriguez-Ramos, I., Anderson, J. A., and
Guerrero-Ruiz, A., Appl. Catal. A 202, 183 (2000).
38. Schuurman, Y., Marquez-Alvarez, C., Kroll, V. C. H., and
Mirodatos, C., Catal. Today 46, 185 (1998).
10. Suzuki, M., Sasaki, H., and Otoshi, S., in “Proceedings of the 2nd In-
ternational Symposium of SOFC’s” (F. Grosz, P. Zegers, S. C. Singhal,
and O. Yamamoto, Eds.), p. 323; Electrochemical Society, Pennington,
NJ, 1991.
11. Vernoux, P., Guindet, J., Gehain, E., and Kleitz, M., in “Proceedings of
SOFC’s V” (U. Stimming, S. C. Singhal, H. Tagawa, and W. Lehnert,
Eds.), p. 219. Electrochemical Society, Pennington, NJ, 1997.
39. Ferreira-Aparicio, P., Rodriguez-Ramos, I., and Guerrero-Ruiz, A.,
Appl. Catal. A 148, 343 (1997).
12. Djurado, E., Bouvier, P., and Lucazeau, G., J. Solid State Chem. 149, 40. Gellings, P. J., and Bouwmeester, H. J. M., Catal. Today 58, 1 (2000).
399 (2000).
41. Driscoll, J., Martir, W., Wang, J.-X., and Lunsford, J. H., J. Am. Chem.
Soc. 107, 58 (1985).
42. Steghuis, G., Van Ommen, J. G., and Lercher, J. A., Stud. Surf. Sci.
Catal. 119, 831 (1998).
43. Saracco, G., Scibilia, G., Iannibello, A., and Baldi, G., Appl. Catal. B
8, 229 (1996).
13. Sauvet, A. L., and Fouletier, J., J. Power Sources 101, 259 (2001).
14. Kro¨ger, F. A., and Vink, H., J. Solid State Phys. 3, 307 (1956).
15. Ponce, S., Pen˜a, M. A., and Fierro, J. L. G., Appl. Catal. B 24, 193
(2000).
16. Baker, R. T., and Metcalfe, I. S., Appl. Catal. A 126, 297 (1995).