Vaporization of Sr- and Mg-Doped Lanthanum Gallate and Implications for Solid Oxide Fuel Cells

Article abstract of DOI:10.1149/1.1370967

Vaporization of the La_0.85Sr_0.15Ga_0.85Mg_0.15O _2.85, and La_0.90Sr_0.10Ga_0.80Mg_0.20O _2.85 perovskite phases was investigated by the use of Knudsen e

Full text of DOI:10.1149/1.1370967

E276
Journal of The Electrochemical Society, 148 ͑6͒ E276-E281 ͑2001͒
0013-4651/2001/148͑6͒/E276/6/$7.00 © The Electrochemical Society, Inc.
Vaporization of Sr- and Mg-Doped Lanthanum Gallate
and Implications for Solid Oxide Fuel Cells
Wioletta Kuncewicz-Kupczyk,a Dietmar Kobertz,b Miroslaw Miller,a
Lorenz Singheiser,b and Klaus Hilpertb, z
aWroclaw University of Technology, 50 370 Wroclaw, Poland
Research Center Julich, Institute for Materials and Processes in Energy Systems, 52425 Julich, Germany
b
¨
¨
Vaporization of the La0.85Sr0.15Ga0.85Mg0.15O2.85, and La0.90Sr0.10Ga0.80Mg0.20O2.85 perovskite phases was investigated by the use of
Knudsen effusion mass spectrometry in the temperature range of 1618-1886 K. The partial pressures of the gaseous species O2,
Mg, Sr, SrO, Ga, GaO, Ga2O, and LaO were determined over the samples investigated. The equilibrium partial pressures were
used for the calculation of thermodynamic activities of the components at 1800 K. The results are compared with thermodynamic
data of LaGaO3͑s͒ without additives. Implications of the present data for the potential use of the material in solid oxide fuel cell
technology are discussed as well.
© 2001 The Electrochemical Society. ͓DOI: 10.1149/1.1370967͔ All rights reserved.
Manuscript submitted October 13, 2000; revised manuscript received February 16, 2001.
chardt, Munich, pure anal.͒, and MgO ͑Merck, pure anal.͒ were
mixed with acetone and a binder ͓EtOH ϩ 5% Zusoplast ϩ 2%
polyvinylalcohol ͑PVA͔͒ in an agate mortar for 15 min and calcined
at 1273 K for 24 h in air. The calcined powder was ground again in
an agate mortar and pressed into pellets. After sintering the pellets at
1773 K for 24 h in air, the sample was characterized by X-ray
diffraction ͑XRD, X’Pert, Philips, Cu K␣͒. The XRD patterns ob-
tained are presented in Fig. 1. La͑NO3͒3•6H2O ͑Merck, min 99.0%͒,
Ga ͑Chempur, 99.999%͒, Mg͑NO3͒2•6H2O ͑Chempur, 99.9ϩ%͒,
Sr͑NO3͒2 ͑Riedel-de-Haean, min 99%͒, citric acid, and ethyl glycol
were used as starting materials in the preparation of sample B by the
Pechini method.7 Metallic gallium was solved in HNO3 ͑pure anal.͒
before the preparation of the sample. The sample was calcined at
1273 K for 20 h in air. The calcined powder was ground again in an
agate mortar and pressed into pellets. After sintering the pellets at
1673 K for 17 h in air the sample was characterized by XRD ͑see
Fig. 1͒. Sample B was, in addition, analyzed by the inductively
coupled plasma-optical emission spectra ͑ICP-OES͒ method to
check the deviation from the nominal chemical composition of the
prepared sample. Results are given in Table I.
A sample of the composition La0.90Sr0.10Ga0.80Mg0.20O2.85 sup-
plied by ECN, Petten, The Netherlands, was additionally investi-
gated and the results obtained were compared with those obtained
for a sample prepared by us. Figure 1 shows the XRD pattern of
samples used in this study. The SrLaGa3O7͑s͒ phase was detected in
both samples of the compositions La0.85Sr0.15Ga0.85Mg0.15O2.85 ͑A
and B͒ in addition to the main perovskite phase. The
La0.90Sr0.10Ga0.80Mg0.20O2.85 sample ͑C͒ showed XRD patterns origi-
nating from four phases as presented in Fig. 1.
The La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 perovskite phase with differ-
ent Sr and Mg content has recently been proposed as a possible
candidate material for the electrolyte of solid oxide fuel cells
͑SOFC͒.1 The material shows a good oxygen ion conductivity at
about 800°C2 which is comparable to that of ZrO2 stabilized with 8
mol % Y2O3 ͑YSZ͒ at 1000°C. The latter material is commonly used
as the solid electrolyte in SOFCs. Electrolytes made of
La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 are useful for low SOFC operating
temperatures. Low operating temperatures between 750 and 800°C
allow serious technological and chemical problems to be avoided.
The study of the physicochemical properties of this compound is,
therefore, of great practical interest. For example, the vaporization
processes in the different atmospheres at the anode and cathode
sides of the SOFC have to be known under operating conditions. It
has been shown that lanthanum gallate vaporizes incongruently,3,4
which can lead to changes in the chemical composition of the elec-
trolyte and, as a consequence, to a change of its chemical and elec-
trochemical properties. Knowledge of the potential for the decrease
of this vaporization by the use of a doped LaGaO3 ceramic electro-
lyte is, therefore, of practical interest. There are no studies to date
dealing with the influence of the partial substitution of alkaline earth
metals for La and Ga in LaGaO3 on the volatility of the doped
material.
A further problem might be the alkaline earth carbonate forma-
tion in LaGaO3 base electrolytes if H2 /CO anode gases are used.5
Thermodynamic activities of the alkaline earth metal oxides SrO
and MgO in La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 have to be known in or-
der to make predictions about carbonate formation.6
This paper reports on experimental investigations of the vapor-
ization of the La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 perovskite phase car-
ried out by Knudsen effusion mass spectrometry. Thermodynamic
activities of oxide components in La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 are
determined. The results are used to predict the vaporization of
La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 under SOFC operating conditions at
the anode and cathode sides. Implications of the data for the poten-
tial use of La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 in SOFC technology are
discussed.
The vaporization studies of the samples were carried out by
¨
Knudsen effusion mass spectrometry at Research Center Julich ͑see,
e.g., Ref. 8, 9͒. The instrument of the MAT 271 type was supplied
by Finnigan MAT, Bremen, Germany, and is completely computer
controlled. The vapor species were ionized with an emission current
of 1 mA and an electron energy of 50 eV. Knudsen cells made of
tungsten and lined completely with iridium were employed in the
measurements. Temperatures were measured by an automatic py-
rometer of the ETSO-U type supplied by Dr. Georg Maurer GmbH,
Kohlberg, Germany, and calibrated using the melting points of
nickel, silver, gold, and platinum.
Six runs were carried out in order to elucidate the vaporization of
the different samples. The partial pressure of gallium oxide was
slightly decreasing at the beginning of each run and became nearly
constant after ca. 1 h of vaporization at about 1750 K. The runs
started at a comparatively high measurement temperature after the
ion intensities became time-independent. The measurement of ion
intensities was subsequently carried out at different decreasing tem-
peratures. Finally, several increasing temperatures were adjusted af-
Experimental
Two samples of the composition La0.85Sr0.15Ga0.85Mg0.15O2.85
were prepared by different methods: standard solid-state reaction
͑sample A͒ and Pechini method7 ͑sample B͒. In the standard solid-
state reaction method stoichiometric amounts of powders of La2O3
͑Fluka AG, 99.99%͒ and Ga2O3 ͑Chempur, 99.99%͒, SrO ͑Schu-
z E-mail: k.hilpert@fz-juelich.de
Downloaded on 2015-05-31 to IP 155.69.4.4 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 148 ͑6͒ E276-E281 ͑2001͒
E277
Figure 2. Fragmentation patterns of the Ga2O͑g͒ and LaO͑g͒ molecules on
electron impact ionization.
ter reaching the lowest temperature of the measurement range in
order to show the reproducibility of the vapor pressure measure-
ments. Runs 1-3 ͑sample A͒ and 4 and 5 ͑sample B͒ were carried out
for the same material. Between the runs, the samples were taken out
of the effusion cell, ground in an agate mortar to refresh the vapor-
ization surface, and studied again. The residual of sample B after
run 5 was analyzed by means of chemical analysis ͑Table I͒ and
XRD ͑Fig. 1͒ to check the possible change of sample due to deple-
tion of Ga2O3 from the sample during the vaporization measure-
ment. The relative accuracy of the chemical analysis performed by
ICP-OES method ͑instrument 34000-459 B, Fisons/ARL, Switzer-
land͒ was Ϯ3%.
Results
Ionic species and their assignment to neutral precursors.—The
ion species Oϩ2 , Gaϩ, Ga2ϩ , GaOϩ, Ga2Oϩ, Mgϩ, Srϩ, SrOϩ, Laϩ,
and LaOϩ were detected in the mass spectrum of the vapor over the
La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 samples. They were identified by
their mass, by the shutter effect, and by their isotope abundances.
The Oϩ signal originating from the samples was not detected be-
cause of interfering background ion intensities at m/e ϭ 16 amu.
The assignment of the Ga-containing ions to their neutral precursors
was carried out by the use of isothermal vaporization experiments of
La2O3-Ga2O3 samples and was reported in Ref. 4. It was assumed
that the other ion species are formed from the gaseous neutral spe-
cies determined by Chupka et al.10 on vaporizing pure La2O3 and by
Porter et al.11 on vaporizing pure MgO and SrO. The assignment of
the ions to their neutral precursors is given in Table II. Fragmenta-
tion patterns of the Ga2O and LaO molecules, as determined in Ref.
4 are presented in Fig. 2.
Figure 1. X-ray patterns of the samples A ͑a͒, and B ͑b͒, both of the
composition La0.85Sr0.15Ga0.85Mg0.15O2.85 and of the sample ͑c͒
(La0.90Sr0.10Ga0.80Mg0.20O2.85) after preparation and of sample B ͑d͒ after the
,
C
vaporization measurements.
Table I. Results of chemical analysis „by ICP-OES… of sample B
of the composition La0.85Sr0.15Ga0.85Mg0.15O2.85 after preparation
and after the vaporization study by KEMS „runs 4 and 5….
Figure 3 exemplifies the ion intensities corrected for isotopic
distribution, as measured in run 6 at different temperatures on va-
porizing the sample La0.90Sr0.10Ga0.80Mg0.20O2.85. The ion intensities
obtained from runs 1-6 at 1800 K by interpolation are listed in
Table II.
x ͑La͒
x ͑Sr͒
x ͑Ga͒
x ͑Mg͒
Nominal
After preparation
After vaporization study
0.425
0.428
0.438
0.0750
0.0720
0.0735
0.425
0.428
0.415
0.075
0.0725
0.0740
Partial pressures.—Partial pressures p(i) of the species i at the
temperature T were obtained from the equation
Downloaded on 2015-05-31 to IP 155.69.4.4 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E278
Journal of The Electrochemical Society, 148 ͑6͒ E276-E281 ͑2001͒
Figure 3. Ion intensities obtained in run 6 for different temperatures.
Figure 4. Vapor pressures over the sample LSGM 10-20 obtained in run 6.
MgO ϭ Mg͑g͒ ϩ 12O2͑g͒
SrO ϭ SrO͑g͒
͓3͔
͓4͔
p i͒ ϭ kT⌺I i͒ /␴ i͒
͓1͔
͑
͑
͑
͕
͖
where k and ⌺I(i) are the pressure calibration factor and the sum of
the intensities of the ions originating from the same neutral precur-
sor i. ␴(i) is the relative ionization cross section of the species i.
The values of the relative ionization cross sections ␴(i) given in
parentheses were used for the following gaseous species i: Ag
͑5.35͒, Au ͑6.82͒, Ni ͑5.08͒, O2 ͑1.25͒, Ga ͑8.62͒, GaO ͑6.12͒,
Ga2O ͑18.1͒, LaO ͑10.7͒, Mg ͑3.39͒, Sr ͑8.88͒, and SrO ͑6.30͒. They
were obtained on the basis of the cross sections of the elements
given in Ref. 12-15 and from the empirical ratios ␴(MO)/␴(M)
ϭ 0.71 and ␴(MO2)/␴(MO) ϭ 0.36 selected from the data re-
ported by Drowart16 for the transition metals M.
Pressure calibration was carried out by vaporizing pure Ag, Au,
and Ni at their melting temperatures. The pressure calibration con-
stant k amounted to (4.85 Ϯ 0.57)10Ϫ8 Pa s KϪ1 ͑runs 1-3͒, (2.72
Ϯ 0.33)10Ϫ8 Pa s KϪ1 ͑runs 4 and 5͒, and (2.82 Ϯ 0.34)10Ϫ8 Pa s
KϪ1 ͑run 6͒ for the runs given in parentheses. Partial pressures were
evaluated by the use of Eq. 1 for each measurement temperature. In
Fig. 4 we show as an example the partial pressures of vapor species
obtained in run 6 for the sample La0.90Sr0.10Ga0.80Mg0.20O2.85. Table
III shows the partial pressure equations of the different species ob-
tained for all measurements.
were evaluated from the measurements of the sample materials and
the pure components. Ga2O3, MgO, and SrO mean the respective
oxide components in the La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 samples or
these oxides as pure materials. The following equations result for
the computation of the thermodynamic activities by considering the
equilibrium constants Kpo(2), Kpo(3), and Kop(4) of the reactions Eq.
2 to 4
I͑Ga2Oϩ͒ • I͑Oϩ2 ͒
a Ga O ͒ ϭ
͓5͔
͓6͔
͑
2
3
Io Ga Oϩ͒ • Io Oϩ͒
͑
͑
2
2
ϩ
2
I Mgϩ͒ • I O ͒
1/2
͑
͑
a MgO͒ ϭ
͑
ϩ
Io Mgϩ͒ • Io O ͒
1/2
͑
͑
2
and
I SrOϩ͒
͑
a SrO͒ ϭ
͑
͓7͔
Io SrOϩ͒
͑
Io(i) denotes the ion intensities detected on vaporizing the pure
oxides. Thermodynamic activities resulted for each of the measure-
ment temperatures of the runs 1 to 6 by using Eq. 5-7 and the
measured ion intensities for the samples and for the pure compo-
nents. The intensities for pure Ga2O3͑s͒ ͑1479-1753 K͒, MgO͑s͒
͑1636-2029 K͒, and SrO͑s͒ ͑1651-2272 K͒ were measured in the
temperature ranges given in parentheses.
Thermodynamic activities.—The determination of the thermody-
namic activities of Ga2O3, MgO, and SrO in the samples was carried
out by vaporizing pure Ga2O3, MgO, and SrO before and after the
measurement of a sample. The equilibria
Ga2O3 ϭ Ga2O͑g͒ ϩ O2͑g͒
͓2͔
Downloaded on 2015-05-31 to IP 155.69.4.4 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 148 ͑6͒ E276-E281 ͑2001͒
E279
Table III. Partial pressure equations, ln p„i… ϭ ÀAÕT ϩ B, for the vapor species over LSGM samples.a
O2͑g͒
Ga͑g͒
GaO͑g͒
Ga2O͑g͒
Mg͑g͒
Sr͑g͒
SrO͑g͒
Run Sample
A
B
A
B
A
B
A
B
A
B
A
B
A
B
1
2
3
4
5
6
A
A
A
B
B
C
59898 30.837 56040 29.361 55172 22.137 68361 35.270 83039 40.264 34096 10.778 37811 11.705
66085 34.115 58882 30.945 57586 23.632 68251 35.084 79194 38.167 44795 16.555 35002 10.326
47995 23.974 59104 31.000 66096 28.591 63839 32.548
-
-
47890 18.470 50864 19.120
51907 27.057 52663 27.598 52791 22.568 62911 32.438 58307 26.758 28843
56782 29.752 51383 26.823 52478 22.488 62086 31.949 55631 25.334 24260
62490 32.396 59109 30.754 53860 22.424 66317 34.080 63316 29.634 30792
6.994 48248 16.716
4.548 46310 15.394
8.227 38762 11.312
a LaO͑g͒ was identified only at the highest measurement temperatures in runs 4-6.
1073 K. Figure 5 shows the results of the computations carried out
The activity of La2O3 could only be estimated due to the very
small intensities of the ions Laϩ and LaOϩ observed on vaporizing
the La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 samples. The estimation was car-
ried out for runs 4 and 5 considering the equilibrium
by using the thermodynamic activity of Ga2O3 at 1073 K and the
17
* * *
F A C T 2.2 program together with the thermodynamic data of
* * *
the different gaseous species given in the F A C T 2.2 database.
The thermodynamic activity of Ga2O3 in
La2O3͑s͒ ϭ 2LaO͑g͒ ϩ 1/2O2͑g͒
͓8͔
La0.90Sr0.10Ga0.80Mg0.20O2.85 was estimated from the value at 1800 K
in Table IV by assuming that the chemical potential of Ga2O3 in
La0.90Sr0.10Ga0.80Mg0.20O2.85 is independent of temperature. This as-
sumption is based on the general experience that the Gibbs energy
of formation of the perovskites from their component oxides is prac-
tically temperature independent. A value of a͑G2O3͒ ϭ 2.3•10Ϫ5
thus resulted for 1073 K.
in the sample material and in the sample of the composition
42 mol % Ga O ϩ 58 mol % La O ͖. The latter sample consists
͕
2
3
2
3
of the two phases LaGaO3͑s͒ and La4Ga2O9͑s͒; the activity of La2O3
in this sample at 1800 K was previously determined by us as 0.47.4
The thermodynamic activity in the La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2
sample was evaluated from the equation
Figure
5
shows that the volatility of Ga2O3 in
ϩ
1/2
I LaOϩ͒2I O ͒
La0.90Sr0.10Ga0.80Mg0.20O2.85 at the anode side of SOFC is high. The
high volatility is due to the formation of GaOH͑g͒ and decreases if
the H2O/H2 ratio is raised. The latter means that the volatility is the
highest at the entrance of the fuel gas. The partial pressure of the
most abundant vapor species GaOH at the anode side changes by
less than a factor of 1.8 if the temperature is reduced from 1073 to
973 K. That means a decrease of the operating SOFC temperature
by 100 K reduces the Ga2O3 vaporization only slightly. The results
indicate that the use of electrolytes made of LaGaO3 base perovs-
kites without protective measures at the anode side of SOFC can be
a problem.
͑
͑
2
͕
͖
a La O ͒ ϭ
LSGM 0.47
͓9͔
͑
2
3
ϩ
I LaOϩ͒2I O ͒
1/2
͑
͑
2
͕
͖
LG
where LSGM and LG means that the ion intensities were measured
over the sample La0.85Sr0.15Ga0.85Mg0.15O2.85 ͑sample B͒ and over
the two-phase sample LaGaO ϩLa Ga O , respectively.
͕
͖
9
3
4
2
Table IV summarizes the thermodynamic activities of the com-
ponents in the La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 samples at 1800 K ob-
tained by all runs.
Discussion
The values for the thermodynamic activity of MgO and SrO in
Implications
for
SOFC.—Vaporization
of
La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 are high ͑cf. Table IV͒ and can be
La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 is an important property for the tech-
nical application of this material and for its production process ͑e.g.,
sintering͒. It is shown that this material vaporizes incongruently
mainly in the form of the different Ga-containing species and oxy-
gen. The volatility of the La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 materials
depends on the oxygen partial pressure of the surrounding atmo-
sphere and the temperature. The volatility of the material under
SOFC operating conditions is important for its practical use as an
electrolyte in SOFCs. The composition La0.90Sr0.10Ga0.80Mg0.20O2.85
is most important for practical application due its high ionic conduc-
tivity. The estimation of the volatility was, therefore, carried out for
this composition. In addition, different H2 /H2O ratios were selected
covering the range present at the anode side and a temperature of
Table IV. Thermodynamic activities of the LSGM components
obtained at 1800 K.
a(La2O3)a
Run
Sample
a(Ga2O3)
ϫ 103
a(MgO)
a(SrO)
1
2
3
4
5
6
A
A
A
B
B
C
2.9
2.0
1.6
2.9
2.4
1.7
0.50
0.59
—
0.61
0.64
0.70
0.76
0.90
0.89
0.36
0.28
0.30
0.3
0.3
Figure 5. Partial pressures over the sample of the composition
La0.90Sr0.10Ga0.80Mg0.20O2.85 in an H2 /H2O atmosphere at 1073 K for differ-
ent H2O mole fractions.
a Estimated values.
Downloaded on 2015-05-31 to IP 155.69.4.4 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E280
Journal of The Electrochemical Society, 148 ͑6͒ E276-E281 ͑2001͒
used to estimate the conditions for the formation of MgCO3 and
SrCO3 under an atmosphere present at the anode according to the
reactions
La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 samples do not depend on the oxygen
partial pressure since the oxygen vacancies, ␦, in the perovskite
structure are independent from the oxygen partial pressure. This is
in contrast to other doped perovskites containing transition metals
͑e.g., Ref. 20͒.
MgO ϩ CO2͑g͒ ϭ MgCO3͑s͒
SrO ϩ CO2͑g͒ ϭ SrCO3͑s͒
͓10͔
͓11͔
The thermodynamic activity of Ga2O3 could be also evaluated by
use of the equilibrium
where MgO and SrO means MgO and SrO bound in the perovskite.
By using a͑MgO͒ ϭ 0.40, a͑SrO͒ ϭ 0.13, a CO2 partial pressure of
0.1 bar, and the data from the IVTANTHERMO database18 the fol-
lowing relations are obtained for the change of the Gibbs energy of
reactions Eq. 10 and 11
Ga2O3͑s͒ ϭ 2Ga͑g͒ ϩ 3/2 O2͑g͒
from the equation
͓14͔
ϩ
2
I Gaϩ from Ga͒2I O ͒
3/2
͑
͑
͕
͖
o
⌬ G T͒/J molϪ1 ϭ Ϫ99743 ϩ 199.63T for MgCO3 ͓12͔
2
3
a Ga O ͒ ϭ
͓15͔
͑
͑
ϩ
I Gaϩ from Ga͒2I O ͒
3/2
f
͑
͑
͕
͖
2
and
Values of 1.9 • 10Ϫ3 and 1.7 • 10Ϫ3 were, for example, obtained in
this way for the runs 3 and 4, respectively. Taking into account that
͑i͒ Reaction 14, involves 3.5 gaseous species in comparison to only
two species in Eq. 2 and ͑ii͒ there is the necessity of subtraction of
part of the total Gaϩ intensity formed by fragmentation from
Ga2O͑g͒, the use of Eq. 15 in the calculations must be considered as
substantially less accurate than the method used in this work. From
this point of view, a͑Ga2O3͒ values obtained by both methods are in
agreement.
⌬ G T͒/J molϪ1 ϭ Ϫ229890 ϩ 193.29T for SrCO3 ͓13͔
͑
f
This means that the formation of MgCO3 and SrCO3 is possible at
temperatures below 500 and 1189 K, respectively; the compounds
are not stable at temperatures above this value. The temperature
dependence of the thermodynamic activity of MgO and SrO is small
for high activity values, which is the case if the chemical potential is
assumed to be temperature independent.
We also tried to estimate the SrO activity by use of the equilib-
rium
Condensed phase equilibria.—Samples A and B of the chemical
composition La0.85Sr0.15Ga0.85Mg0.15O2.85 contained the SrLaGa3O7
phase in addition to the perovskite phase after preparation. This was
independent from the preparation method as can be seen in Fig. 1.
According to the phase diagram studies carried out by us, the sample
should, however, be a single phase.19 A possible reason for this
observation might be a small overstoichiometry in the Ga2O3 com-
ponent. This was, however, not confirmed by the chemical analysis
of sample B conducted with a fresh-prepared sample ͑see Table I͒.
XRD analysis of the sample B residual from the Knudsen cell
shows the rearrangement of solid phases during the vaporization
measurement. The SrLaGa3O7 phase disappeared and the SrLaGaO4
phase was identified in this residual besides the perovskite phase.
According to phase diagram of the system ͑Ref. 19͒, the SrLaGaO4
phase becomes stable in this case.
SrO͑s͒ ϭ Sr͑g͒ ϩ 1/2O2͑g͒
͓16͔
from the equation
ϩ
1/2
I Srϩ͒I O ͒
͑
͑
2
͕
͖
o
LSGM
1/2
͖
a SrO͒ ϭ
͑
͓17͔
ϩ
2
I Srϩ͒oI O ͒
͑
͑
͕
The values obtained in this way were by 25-100% different from
those evaluated by using of Eq. 7. Values of Ͼ1 ͑run 3͒, 0.21 ͑run
5͒, and 0.79 ͑run 6͒ resulted, for example, for samples A, B, and C,
respectively. Possible reasons of these differences are ͑i͒ relative
large scatter of Srϩ ion intensities and ͑ii͒ possible fragmentation of
SrO͑g͒ forming Srϩ.
The thermodynamic activities of Ga2O3 and MgO in the samples
A and B, which are of the same chemical composition, show very
good agreement. This is not the case for the SrO activities, which
differ by a factor of two to three between both samples as follows
from Table IV. Due to the low SrOϩ ion intensity, the activity of
SrO was obtained with less accuracy in comparison to the activities
of MgO and Ga2O3. However, the difference of SrO activity as
measured for samples A and B cannot be explained by this uncer-
tainty. The possible reason for this inconsistency might be small
differences of sample composition caused by different preparation
methods ͑see Experimental section͒.
Probable overall uncertainties were not evaluated for the partial
pressures given in Table III since it is difficult to estimate uncertain-
ties for the ionization cross sections of gaseous oxide species. In
contrast to the partial pressures, the thermodynamic activities do not
depend on ionization cross sections but only on ion intensity ratios.
This leads to comparatively small uncertainties for the thermody-
namic activities. These uncertainties were estimated as 25, 25, and
35%, respectively, for the activities of Ga2O3, MgO, and SrO, con-
sidering the error of the calibration constant, the statistical scatter,
and the error in the extrapolation of the SrOϩ ion intensities up to
1800 K carried out for the calibration measurements using pure SrO.
The absolute error of the temperature measurement estimated as
Ϯ10 K was additionally used in the error computation for the ther-
modynamic activities. The estimated uncertainty of the values of the
La2O3 activity in sample B is a factor of two.
Sample
C
of
the
chemical
composition
La0.90Sr0.10Ga0.80Mg0.20O2.85 contained the four phases LaGaO3,
La4Ga2O9, SrLaGa3O7, and SrLaGaO4 before the vaporization study
͑see Fig. 1͒. This might be caused by the incomplete reaction of the
starting components during the sample preparation. Phase relations
in the quasi-quaternary Ga2O3-La2O3-MgO-SrO system show that
these phases cannot exist together in thermodynamic equilibrium.
Equilibration is assumed at the beginning of the vaporization experi-
ments.
The partial pressures of the gallium-containing species were
slightly decreasing at the beginning of the vaporization experiments.
This might be explained by the continuous Ga depletion from the
samples resulting in the formation of additional phases in the
samples. Therefore, thermodynamic data obtained in this study al-
ways refer to the B site poor perovskite phase. This phase remains in
equilibrium with other phases depending on the quasi-quaternary
phase diagram, which is still little known. However, as follows from
chemical analysis of the initial sample and of the residuals after the
vaporization study carried out for sample B, the total composition of
samples changes only slightly during the experiment. This means,
that the amount of other phases formed in the Knudsen cell is often
too small to identify all of them by means of XRD analysis.
Consistency and uncertainties of data.—The thermodynamic ac-
tivities of Ga2O3 in the La1ϪxSrxGa1ϪyMgyO3Ϫ(xϩy)/2 samples are
similar to the value for undoped LaGaO3, 2.17 • 10Ϫ3 at 1700 K and
3.35 • 10Ϫ3 at 1800 K.4 It is interesting to note that the thermody-
namic activities of the oxide components in the
Downloaded on 2015-05-31 to IP 155.69.4.4 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 148 ͑6͒ E276-E281 ͑2001͒
E281
7. M. P. Pechini, U.S. Pat. 3,330,697 ͑1967͒.
8. V. L. Stolyarova and G. A. Semenov, Mass Spectrometric Study of the Vaporiza-
tion of Oxide Systems, Wiley, New York ͑1994͒.
9. K. Hilpert, Rapid Commun. Mass Spectrom., 5, 175 ͑1991͒.
10. W. A. Chupka, M. G. Inghram, and R. F. Porter, J. Chem. Phys., 24, 792 ͑1956͒.
11. R. F. Porter, W. A. Chupka, and M. G. Inghram, J. Chem. Phys., 23, 1347 ͑1955͒.
12. R. S. Freund, R. C. Wetzel, R. J. Shul, and T. R. Hayes, Phys. Rev. A, 41, 3575
͑1990͒.
13. J. L. Cooper, Jr., G. A. Pressley, Jr., and F. E. Stafford, J. Chem. Phys., 44, 3946
͑1965͒.
14. F. A. Stevie and M. J. Vasile, J. Chem. Phys., 74, 5106 ͑1981͒.
15. J. B. Mann, J. Chem. Phys., 46, 1646 ͑1967͒.
16. J. Drowart, in Advances in Mass Spectrometry 1985, J. F. J. Todd, Editor, p. 195,
Wiley, New York ͑1986͒.
Acknowledgments
The authors thank the Deutsche Forschungsgemeinschaft, DFG,
and the International Bureau of the German Ministry of Science and
Technology for financial support.
¨
Research Center Julich assisted in meeting the publication costs of this
article.
References
1. I. Ishihara, M. Honda, T. Shibayama, H. Minami, H. Nishiguchi, and Y. Takita, J.
Electrochem. Soc., 145, 3177 ͑1998͒.
2. P. Huang and A. Petric, J. Electrochem. Soc., 143, 1644 ͑1996͒.
3. K. Yamaji, T. Horita, M. Ishikawa, N. Sakai, and H. Yokokawa, Solid State Ionics,
121, 217 ͑1999͒.
4. W. Kuncewicz-Kupczyk, D. Kobertz, M. Miller, C. Chatillon, and K. Hilpert, In
preparation.
5. B. C. H. Steele, in Proceedings of 1st European Solid Oxide Fuel Cell Forum, U.
Bossel, Editor, p. 375, Baden, Switzerland ͑1994͒.
*
* *
17. F A C T ͑Facility for the Analysis of Chemical Thermodynamics͒, A. D. Pelton,
W. T. Thompson, and C. W. Bale, McGill University/Ecole Polytechnique, Mon-
treal.
18. IVTANTHERMO—A Thermodynamic Database and Software System for the Per-
sonal Computer, V. S. Yungman, V. A. Medvedev, I. V. Veits, and G. A. Bergman,
CRC Press and Begell House, Boca Raton, FL ͑1993͒.
6. D. H. Peck, M. Miller, H. Nickel, D. Das, and K. Hilpert, in Proceedings of the 4th
International Symposium on Solid Oxide Fuel Cells (SOFC IV), M. Dokiya, O.
Yamamoto, H. Tagawa, and S. C. Singhal, Editors, PV 95-1, p. 858, The Electro-
chemical Society Proceedings Series, Pennington, NJ ͑1995͒.
19. A. Matraszek, D. Kobertz, L. Singheiser, K. Hilpert, W. Kuncewicz-Kupczyk, and
M. Miller, in Proceedings of the 7th International Symposium on Solid Oxide Fuel
Cells, June 3-8, 2001, Tsukuba, Japan, In press.
20. N. Trofimenko and H. Ullmann, Solid State Ionics, 118, 215 ͑1999͒.
Downloaded on 2015-05-31 to IP 155.69.4.4 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).