CeO2-CoO phase diagram

Article abstract of DOI:10.1111/j.1151-2916.2003.tb03515.x

Phase equilibria in the CeO₂ - CoO system at temperatures above 1500°C were investigated. The microstructures and the phase compositions of the DTA (differential thermal analysis) samples and the quenched solid pellets were analyzed using SEM (scanning electron microscope), EDX (energy dispersive X-ray), and WDX (wavelength dispersive X-ray). A eutectic reaction was found at 1645 ± 5°C. The eutectic point was calculated to be at 82 ± 1.5 mol% CoO. The eutectic phases were the CeO₂-rich phase (containing <5 mol% CoO) and the CoO-rich phase (containing ～0.5 mol% CeO₂). At 1580°C, the solubility of CoO in CeO₂ was ～3 mol%.

Full text of DOI:10.1111/j.1151-2916.2003.tb03515.x

J. Am. Ceram. Soc., 86 [9] 1567–70 (2003)
journal
CeO ؊CoO Phase Diagram
2
Ming Chen, Bengt Hallstedt, A. Nicholas Grundy, and Ludwig J. Gauckler*
Institute of Nonmetallic Materials, Department of Materials, Swiss Federal Institute of Technology, ETH Z u¨ rich,
CH-8092 Z u¨ rich, Switzerland
8
9
Phase equilibria in the CeO ؊CoO system at temperatures
by Ranløv et al. In the response to Ranløv’s comment, Pound
2
above 1500°C were investigated. The microstructures and the
phase compositions of the DTA (differential thermal analysis)
samples and the quenched solid pellets were analyzed using
SEM (scanning electron microscope), EDX (energy dispersive
X-ray), and WDX (wavelength dispersive X-ray). A eutectic
reaction was found at 1645 ؎ 5°C. The eutectic point was
calculated to be at 82 ؎ 1.5 mol% CoO. The eutectic phases
also noticed that “[t]he data for the doped CeO in this work was
2
probably affected (unfortunately) by the presence of excess NiO
1
0
and CoO in the specimens.” Recently Hrovat et al. found that at
1200°C, the solid solubility of cobalt oxides in CeO was at the
2
limit of energy dispersive spectroscopy (EDS) analysis, i.e.,
around 0.5 at.%, far less than Dontsov’s and Pound’s results.
Hrovat et al. attributed the discrepancy to different sintering
temperatures, i.e., 1600°C in Dontsov’s, 1500°C in Pound’s, and
1200°C in Hrovat’s experiments.
were the CeO -rich phase (containing <5 mol% CoO) and the
2
CoO-rich phase (containing ϳ0.5 mol% CeO ). At 1580°C, the
2
1
1
solubility of CoO in CeO was ϳ3 mol%.
Sirman et al. measured the lattice parameters of solid pellets
sintered at 1400°C for 50 h. They concluded that the solubility
limit of CoO in CGO is less than 0.5 mol% at 1400°C. They also
measured the diffusion profile of cobalt into CGO at different
2
I. Introduction
2
ϩ
temperatures. Co ions were shown to diffuse very slowly into
CGO.
N THE last few decades there has been much interest in the use
I
of ceria-based ceramics for applications in solid oxide fuel cells
The purpose of the present work was to investigate the phase
(
SOFCs) and in oxygen sensors. Pure stoichiometric CeO has the
2
equilibria in the CeO ϪCoO system at temperatures above
2
fluorite type of structure with space group Fm3m over the
1
500°C, in particular the solubility of CoO in CeO . The compo-
1
2
temperature range from room temperature to the melting point.
Replacing Ce ions with divalent or trivalent cations can intro-
sitions of the samples are hence presented in mole percent of CoO.
A tentative experimental phase diagram is also presented.
4
ϩ
duce a high concentration of oxygen vacancies into the CeO
crystal lattice. It has been shown that the ionic conductivity of
2
2
doped ceria can be 4 to 5 times higher than that of yttria-fully-
II. Experimental Procedure
stabilized zirconia (YSZ) at intermediate temperatures
3
(
600Ϫ700°C). This makes ceria-based ceramics very attractive as
Powders of the pure oxides were used: CeO (Alfa, 99.9%),
Co O (Alfa, 99.7%). One composition (corresponding to 20
mol% CeO Ϫ80 mol% CoO) was investigated. The oxides were
mixed with ethanol in an agate mortar for 2 h, then dried, pressed
into pellets, and prefired at 1000°C for 50 h. The pellets were
ground, pressed into pellets again, and sintered at 1580°C for 40 h
and quenched in air.
2
solid electrolytes for developing low-temperature SOFCs and
low-temperature oxygen sensors.
3
4
2
One of the drawbacks of ceria-based ceramics is the high
3
sintering temperature. It varies from 1300° to 1700°C, depending
on the dopant, the starting powder, and the sintering method. Such
high sintering temperatures lead to considerable grain coarsening
and therefore poor mechanical properties. By doping with a small
4
Differential thermal analysis (DTA) was done on the prefired
powder, using simultaneous thermal analysis (STA 501, B a¨ hr
Thermoanalyse GmbH). The powder was heated up to 1680°C at
a heating rate of 10°C/min and then cooled down to room
temperature at Ϫ10°C/min. The experiments were conducted in an
atmosphere of flowing air. Platinum crucibles were found to be
more inert toward the oxide melt than Al O crucibles, and were
amount of Co O (less than 5 cat.%), Kleinlogel and Gauckler
3
4
attained dense Ce0.8Gd0.2O2Ϫx (CGO, Ͼ99% theoretical density)
at temperatures below 900°C, with an average grain size of less
5
than 150 nm. This was confirmed by Lewis et al. They suggested
that the transformation from Co O to CoO around 900°C could be
3
4
important for enhancing densification.
Until now, no phase diagram has been available for the
2
3
thus used.
CeO ϪCoO system. Some work has been done on the solubilities
2
The DTA samples (molten and resolidified) and the quenched
solid pellets were polished with diamond paste and coated with
carbon. Their microstructures were analyzed using a field emission
scanning electron microscope (LEO 1530, LEO Electron Micros-
copy, Ltd.). The phase compositions were measured using energy
dispersive X-ray (EDX: Z-MAX 30, Thermo NORAN) and using
wavelength dispersive X-ray (WDX: SX50, CAMECA). The
accelerating voltage was set at 15 kV for both EDX and WDX
measurements.
of CoO in pure CeO and in CGO at temperatures above 1000°C.
2
6
Based on lattice parameter measurements, Dontsov et al. con-
cluded that 35 mol% CoO can be dissolved into CeO at 1600°C.
2
7
Later, Pound suggested that a 10 mol% substitutional solid
solution of CoO in CeO can be formed by sintering mixed oxide
2
disks at 1500°C for 5 h. His evaluation of the formation of the solid
solution was based on the comparison of the d-spacings between
pure ceria and the mixed oxide disks. His results were questioned
The overall compositions of the sintered pellets were analyzed
using EDX, as visible cobalt contamination was found in the
furnace after sintering. The CoO concentration in the quenched
pellets was around 78 Ϯ 2 mol%. In the case of the DTA
measurements, no visible cobalt contamination was found. The
distinct weight change shown in the TG (thermogravimetry) curve
was caused by the decomposition of Co O into CoO. The effect
R. S. Roth—contributing editor
Manuscript No. 186607. Received October 27, 2002; approved May 20, 2003.
Supported by the Swiss National Science Foundation, under Contract No. NFP
3
4
of cobalt evaporation and CeO nonstoichiometry on the compo-
2
0-53542.98.
Member, American Ceramic Society.
2
*
sitions of the DTA samples can be neglected. In our paper, the
1
567

1
568
Journal of the American Ceramic Society—Chen et al.
Vol. 86, No. 9
nominal composition of 20 mol% CeO Ϫ80 mol% CoO was used
for both the sintered pellets and the DTA samples.
probe diameter) could be as fine as 1Ϫ2 nm. However, the
generated X-ray signals were from a much larger volume centered
on the beam spot. In the present work, the Co K lines and the Ce
L lines were used to quantify the phase compositions. The
estimated depth of the X-ray generation volume was around 1 ␮m,
and the diameter of the volume was probably in the same range. In
2
III. Results and Discussion
(1) Eutectic Temperature and Compositions
the fine eutectic structure, the average grain size of the CeO -rich
2
The prefired powder underwent an endothermic reaction at
phase was around 1.5 ␮m. It is clear that the composition of the
neighboring CoO-rich phase will influence the measured compo-
8
82 Ϯ 5°C, corresponding to the transformation from Co O to
3
4
CoO. Before DTA measurements, the powder was prefired at
000°C and Co O should have transformed to CoO. However, the
sition of the CeO -rich phase, which makes the detected concen-
2
1
tration of CoO in CeO too high. In the coarse eutectic structure,
3
4
2
newly formed CoO still could transform back to Co O during
the grains of the CeO -rich phase were a little bit larger, but the
3
4
2
cooling and intermediate grinding, as Co O is more stable than
measured composition of the CeO -rich phase may have also been
3
4
2
CoO below 800°C. An endothermic melting peak during heating
and an exothermic solidification peak during cooling were found at
high temperature. The melting temperature was 1644 Ϯ 2°C and
the solidification temperature was 1646 Ϯ 2°C. As shown later,
typical eutectic structures were found in the DTA sample. This
influenced by the composition of the neighboring CoO-rich phase.
It is probable that the detected 5 mol% CoO in CeO was still
2
higher than the real concentration. For the measured composition
of the CoO-rich phase, the influence from the neighboring CeO2-
rich phase should have been much less, since the main component
of the eutectic was the CoO-rich phase. The measured concentra-
indicates that in the CeO ϪCoO system, there is a eutectic reaction
2
at 1645 Ϯ 5°C.
tion of 0.5 mol% CeO in CoO should be close to the real value.
2
The microstructure of the DTA sample is shown in Fig. 1. The
Therefore, we tend to conclude that, in the eutectic, the CeO -rich
2
light phase is the CeO -rich phase, and the dark phase is the
phase may have contained Ͻ5 mol% CoO. The concentration of
2
CoO-rich phase. As Fig. 1 shows, the DTA sample is mainly
composed of eutectics; only a few large ceria particles can be
found in the whole sample. It is clear that the composition of the
eutectic point is close to the overall composition of the DTA
CeO in the CoO-rich phase was ϳ0.5 mol%.
2
SEM image analyses show that the dark phase covered 70 Ϯ 2%
area fraction (i.e., volume fraction). We conclude that the eutectic
in the CeO ϪCoO system consists of 30 vol% of the CeO -rich
2
2
sample (20 mol% CeO Ϫ80 mol% CoO), but somewhat richer in
phase and 70 vol% of the CoO-rich phase. Therefore, the eutectic
point is calculated to be at 82 Ϯ 1.5 mol% CoO. This is consistent
with the previous conclusion from Fig. 1.
2
CoO. Both fine and coarse eutectic structures were found, which
may have resulted from changing growth kinetics.
The compositions of the eutectic phases were determined using
EDX. The measurements were done both on the fine and on the
coarse eutectic structures. The probe (the electron beam spot) was
selected to be in the center of the analyzed grain. For the fine
(
2) Subsolidus Phase Equilibria at 1580°C
To study the subsolidus equilibria in the CeO ϪCoO system,
2
eutectic structure, the measurements show that the CeO -rich
quenched solid pellets of 20 mol% CeO Ϫ80 mol% CoO (sintered
2
2
phase contained 10Ϫ11 mol% CoO and the CoO-rich phase
at 1580°C for 40 h) were investigated. The microstructure is shown
in Fig. 2.
contained 0.5Ϫ1 mol% CeO . For the measurements on the coarse
2
eutectic structure, large grains both for the CeO -rich phase and for
The phase compositions were first determined using EDX.
Special line scan measurements were made across the phase
2
the CoO-rich phase were selected, to make sure that the distance
between the beam spot and the phase boundary was Ն2 ␮m; 5
mol% CoO in the CeO -rich phase and around 0.5 mol% CeO in
boundary. Figure 3 shows the CeO and the CoO concentration
2
profiles along one of these lines. The CeO concentration profile
2
2
2
the CoO-rich phase were detected. The phase compositions mea-
sured on the coarse eutectic structure differed from the ones
measured on the fine eutectic structure. The main discrepancy lay
shows that the measured CeO2 concentration in the CoO-rich
phase is around 0.5 mol%, at the limit of EDX analysis. Also as
shown in Fig. 3, the measured CoO concentration in the CeO -rich
2
in the CoO concentration in the CeO -rich phase.
For the field emission scanning electron microscope, the diam-
eter of the electron beam spot impacted on the sample surface (the
phase is around 2 mol%. This is quite different from the reported
solubility data. To verify this, WDX measurements were also done
on the same sample. The results are shown in Fig. 4. No major
2
Fig. 1. Microstructure of the DTA sample. The DTA measurements were
conducted on the prefired powder of 20 mol% CeO Ϫ80 mol% CoO in an
2
atmosphere of flowing air. The powder was heated up to 1680°C at a
heating rate of 10°C/min and then cooled down to room temperature at
Ϫ10°C/min.
Fig. 2. SEM image of the quenched solid pellet sintered at 1580°C for
40 h. The nominal composition of the pellet is 20 mol% CeO Ϫ80 mol%
2
CoO.

September 2003
CeO ϪCoO Phase Diagram
1569
2
Table I. Calculated Vegard’s
Slopes and Reported Solubility
Limits in CeO at 1600°C
2
8
from Ranløv et al.
Estimated
Vegard’s slope
Solubility limit
Dopant
(pm/mol%)
(mol%)
CaO
SrO
MgO
CoO
NiO
0.0
ϩ0.3
Ϫ0.5
Ϫ0.5
Ϫ0.6
16
9
2
?
ϳ0
linear change in lattice parameter with concentration of solute) for
each solute. It is hence possible to make a rough estimation of the
solubility limit for different dopants, by empirical comparison of
8
calculated Vegard’s slopes. This has been done by Ranløv et al.
The Vegard slopes that they calculated are listed in Table I,
together with the solubility limits in CeO at 1600°C reported from
2
the literature. It is seen that the solid solution of CoO in CeO has
2
a relatively steep Vegard’s slope. It is thus unlikely that at 1600°C,
Fig. 3. EDX line scan measurement on the quenched solid pellet sintered
at 1580°C for 40 h. The nominal composition of the pellet is 20 mol%
1
0 mol% CoO or more can be dissolved into CeO . The actual
2
result of a solubility of 3 mol% CoO in CeO at 1580°C seems
2
CeO Ϫ80 mol% CoO. The line scan was made across the phase boundary.
2
more reasonable.
difference can be found between Fig. 3 and Fig. 4. For WDX
measurements, the experimental error (1␴) for the CoO concen-
(3) Phase Diagram
tration in the CeO -rich phase is Ϯ0.09 mol%, very small
Our proposed phase diagram is shown in Fig. 5, together with
2
compared with the measured value. For EDX measurements, the
experimental error (1␴) is Ϯ0.7 mol%. It is obvious that in the
present work, the phase compositions measured using EDX are
almost as reliable as the ones measured using WDX, although
EDX analyses are less precise than WDX analyses. Thus, we
solubility data from other sources. No compound has been found
in the CeO ϪCoO system. The melting temperature of pure
stoichiometric CeO
2
1
is taken as 2477°C, from Mogensen et al.
2
1
3
Chen et al. made a thermodynamic assessment of the CoϪO
system. Their assessed congruent melting temperature of CoO is
1
4
conclude that at 1580°C, the CeO solubility in CoO is very small,
1824°C, very close to the one adopted by the JANAF tables
2
Յ0.5 mol%. The CoO solubility in CeO2 is around 3 mol%.
This is different from Dontsov’s data (30 mol% CoO in CeO2
at 1600°C) and Pound’s data (10 mol% CoO in CeO2 at
(1830 Ϯ 10°C). In the present work, we take the melting
temperature of CoO as 1824°C.
As mentioned before, our measured eutectic temperature is
around 1645 Ϯ 5°C and the composition of the calculated eutectic
point is at 82 Ϯ 1.5 mol% CoO. The eutectic phases are the
1
500°C). The discrepancy cannot be easily attributed to tem-
perature differences.
For solid solutions with fluorite-type structure, there is a
relation between substitutional solubility and radius and valence
CeO
phase (containing ϳ0.5 mol% CeO
solubility in CeO is around 3 mol%. The CeO
is very small, Յ0.5 mol%.
2
-rich phase (containing Ͻ5 mol% CoO) and the CoO-rich
). At 1580°C, the CoO
solubility in CoO
2
8
12
differences between host and dopant cations. Kim studied the
effect of ionic radii and valence of the dopant on the ionic
conductivity of ceria solid solutions. He found that the solubility
limit of the same periodic group elements in CeO2 decreases
linearly with the square of Vegard’s slope (which describes the
2
2
Fig. 4. WDX line scan measurement on the quenched solid pellet sintered
at 1580°C for 40 h. The nominal composition of the pellet is 20 mol%
CeO Ϫ80 mol% CoO. The line scan was made across the phase boundary.
Fig. 5. Tentative phase diagram of the CeO ϪCoO system.
2
2

1
570
Journal of the American Ceramic Society—Chen et al.
Vol. 86, No. 9
5
G. S. Lewis, A. Atkinson, and B. C. H. Steele, “Sintering of Gadolinia-Doped
Ceria at Reduced Temperature”; pp. 773–82 in Proceedings of the Fourth European
Solid Oxide Fuel Cell Forum (Lucerne, Switzerland, 2000). Edited by A. J. McEvoy.
European Fuel Cell Forum, Oberrohrdorf, Switzerland, 2000.
IV. Conclusions
In the CeO ϪCoO system, a eutectic reaction was found at a
2
temperature of 1645 Ϯ 5°C and the evaluated eutectic point is at
2 Ϯ 1.5 mol% CoO. The eutectic phases are the CeO -rich phase
6
G. Dontsov, G. Vitter, and C. D e´ portes, “Study of the Fluorite Phase in the
8
CeO –ZrO –CoO System” (in French), Rev. Int. Hautes Temp. Refract., 9 [1] 147–52
2
2
2
(containing Ͻ5 mol% CoO) and the CoO-rich phase (containing
(1972).
7
2
B. G. Pound, “The Characterization of Doped CeO Electrodes in Solid Oxide
ϳ0.5 mol% CeO ). At 1580°C, ϳ3 mol% CoO can be dissolved
2
Fuel Cells,” Solid State Ionics, 52 [1–3] 183–88 (1992).
into CeO2.
8
J. Ranløv, F. W. Poulsen, and M. Mogensen, “Comment on ‘The Characterization
of Doped CeO2 Electrodes in Solid Oxide Fuel Cells’ by B. G. Pound, Solid State
Ionics, 52, 183–88 (1992),” Solid State Ionics, 61 [4] 277–79 (1993).
Acknowledgments
9
B. G. Pound, “Response to Comment by J. Ranløv, F. W. Poulsen, and M.
Mogensen on ‘The Characterization of Doped CeO2 Electrodes in Solid Oxide Fuel
Cells’,” Solid State Ionics, 61 [4] 281–84 (1993).
We are gratefully indebted to Anne-Chantal Risold, Dainius Perednis, and David
Sager for their invaluable help with the EDX and WDX.
10
M. Hrovat, J. Holc, and S. Bernik, “Subsolidus Phase Equilibria in the
La O –CeO –Co O System,” J. Mater. Res., 14 [5] 1692–94 (1999).
2
3
2
2 3
1
1
J. D. Sirman, D. Waller, and J. A. Kilner, “Cation Diffusion into a CGO
References
Electrolyte”; pp. 1159–67 in Proceedings of The Fifth International Symposium on
Solid Oxide Fuel Cells (SOFC-V) (Aachen, Germany, 1997). Edited by U. Stimming,
S. C. Singhal, H. Tagawa, and W. Lehnert. The Electrochemical Society, Pennington,
NJ, 1997.
1
M. Mogensen, N. M. Sammes, and G. A. Tompsett, “Physical, Chemical and
Electrochemical Properties of Pure and Doped Ceria,” Solid State Ionics, 129 [1–4]
6
3–94 (2000).
1
2
2
D.-J. Kim, “Lattice Parameters, Ionic Conductivities, and Solubility Limits in
R. T. Dirstine, R. N. Blumenthal, and T. F. Kuech, “Ionic Conductivity of Calcia,
Yttria, and Rare Earth-Doped Cerium Dioxide,” J. Electrochem. Soc., 126 [2] 264–69
1979).
C. Kleinlogel, “Cathode Supported Thin Electrolytes and Nano Sized Ceria Solid
Solutions for Solid Oxide Fuel Cells”; Ph.D. Thesis. Swiss Federal Institute of
Technology, Z u¨ rich, Switzerland, 1999.
4ϩ 4ϩ 4ϩ 4ϩ 4ϩ
2
Fluorite-Structure MO Oxide (M ϭ Hf , Zr , Ce , Th , U ) Solid Solutions,”
(
J. Am. Ceram. Soc., 72 [8] 1415–21 (1989).
3
13
M. Chen, B. Hallstedt, and L. J. Gauckler, “Thermodynamic Assessment of the
Co–O System,” J. Phase Equilib., 24 [3] 212–27 (2003).
1
4
M. W. Chase, C. A. Davies, J. R. Downey, D. J. Frurip, R. A. McDonald, and
A. N. Syverud, “JANAF Thermochemical Tables, 3rd ed.,” J. Phys. Chem. Ref. Data,
Suppl. 1, 14, 923–26 (1985).
4
C. Kleinlogel and L. J. Gauckler, “Sintering and Properties of Nanosized Ceria
Solid Solutions,” Solid State Ionics, 135 [1–4] 567–73 (2000).
Ⅺ