Abstract
The thermodynamic stability of the double perovskite YBaCo2O6−δ was studied using the coulometric titration technique and verified by measurements of the overall conductivity depending on oxygen partial pressure at a given temperature. As a result, the stability diagram of YBaCo2O6−δ was plotted. YBaCo2O6−δ was found to be thermodynamically stable in air at 850°C and higher temperatures, whereas its thermodynamic stability at 900°C is limited by the range of oxygen partial pressures −3.56 ≤ log(pO2/atm) ≤ −0.14. Oxygen content in YBaCo2O6−δ slightly decreases at 900°C from 5.035 at log(pO2/atm) = −0.14 to 4.989 in the atmosphere with log(pO2/atm) = −3.565 indicating a crucial role which variation of Co+3/Co+2 ratio plays in its stability. YBaCo2O6−δ decomposes into the mixture of YCoO3 and BaCoO3−z at the high pO2 stability limit, whereas YBaCo4O7, BaCo1−xYxO3−γ, and Y2O3 were identified as the products of its decomposition at the low pO2 one.
1. Introduction
Cobaltites REBaCo2O6−δ, where REis a rare earth metal, with the double perovskite structure have attracted great attention in the past decade due to their unique properties such as high oxide ion and electronic conductivity as well as very promising activity as cathodes in IT SOFCs [1–5]. Despite the fascinating properties of these materials, their successful commercial application as cathodes is restricted by large thermal expansion coefficient (CTE), which significantly exceeds CTE of the state-of-the-art electrolyte materials [4]. Nevertheless, there is a trend of lowering CTE with decreasing size of RE [4]. In this respect, YBaCo2O6−δ has advantage as compared to other double perovskites since it has the lowest CTE among all REBaCo2O6−δ oxides, which is close to that of doped ceria and zirconia, the state-of-the-art SOFC electrolytes [1–6]. This double perovskite was found to have the total conductivity high enough for using it as a cathode for IT SOFCs [6–15]. As a result, materials on the basis of YBaCo2O6−δ showed high performance as cathodes of IT SOFCs [6, 9, 10, 12, 15] and oxygen permeable membranes [16, 17]. It is generally recognized that knowledge of the thermodynamic stability limits of oxides materials is of key importance for understanding their properties and, therefore, their successful application in electrochemical devices for energy conversion and storage. However, the thermodynamic stability of YBaCo2O6−δ is poorly understood, and it remains a controversial topic so far. The authors [14, 15] by means of X-ray diffraction (XRD) found secondary phases in samples of YBaCo2O6−δ annealed at temperatures between 800°C and 850°C depending on ambient atmosphere and concluded that this oxide is unstable under aforementioned conditions. It is worth noting that, according to Kim et al. [15], YBaCo2O6−δ is unstable at 800°C in air while it was found [14] that this oxide is quite stable in air even at 850°C. On the contrary, YBaCo2O6−δ was found to be unstable at 850°C in atmosphere of nitrogen [14], whereas there is no evidence of this oxide decomposition at 800°C in the same ambient atmosphere [15]. Moreover, if YBaCo2O6−δ is unstable at certain temperature, then one could expect a singularity of temperature dependences of its oxygen content, electrical conductivity, and CTE at this temperature. However, there is no evidence of such singularity in literature except the work of Xue et al. [6] where the authors observed something like that for CTE of YBaCo2O6−δ at 800°C and 850°C. However, Xue et al. [6] did not comment this singularity, which was not confirmed in other works.
Thus, the priority purpose of the present work was to obtain reliable data on the thermodynamic stability of YBaCo2O6−δ by means of three independent methods such as coulometric titration, conductivity measurements, and homogenizing annealing of samples in atmosphere with controlled oxygen partial pressure. The novelty of the work consists in the constructing of the thermodynamic stability diagram for interesting and promising double perovskite YBaCo2O6−δ for the first time.
2. Materials and Methods
The powder samples of single phase YCoO3, BaCo2O3, and YBaCo2O6−δ were synthesized by means of glyserol-nitrate technique using Co, Y2O3, and BaCO3 as starting materials. Y2O3 and BaCO3 had a purity of 99.99%. Metallic Co was obtained by reduction of Co3O4 (purity 99.99%) in H2 atmosphere at 600°C. Y2O3 and BaCO3 were preliminary calcined in air at 1100°C and 600°C, respectively, for removal of adsorbed H2O and CO2.
Stoichiometric mixture of starting materials was dissolved in concentrated nitric acid (99.99% purity), and required volume of glycerol (99% purity) was added as a complexing agent and a fuel. Glycerol quantity was calculated according to full reduction of corresponding nitrates to molecular nitrogen N2. The as-prepared solutions were heated continuously at 100°C until complete water evaporation and pyrolysis of the dried precursor had occurred. The resulting ash was subsequently calcined for 10 hours at 1100°C for BaCoO3 as well as YBaCo2O6−δ and 900°C for YCoO3 to get the desired oxide powder.
The phase composition of the powder samples prepared accordingly was studied at room temperature by means of X-ray diffraction (XRD) with XRD-7000 diffractometer (Shimadzu, Japan) using Cu Kα radiation. XRD showed no indication of the presence of a second phase for the as-prepared oxides.
The chemical composition of all the oxides prepared was checked using an ICP spectrometer ICAP 6500 DUO and an atomic absorption spectrometer Solaar M6 (both supplied by Thermo Scientific, USA). All the as-prepared oxides were shown to have the stoichiometric composition with respect to metal cations within the accuracy of 2%. No impurities were found within the same accuracy range as well.
For the measurements of electrical conductivity, single phase powder of YBaCo2O6−δ was axially pressed into rectangular bars of 30 × 4 × 4 mm3 at 40 MPa and sintered at 1150°C for 24 h in air. The relative density of the sample bars used for measurements was found to be higher than 80%.
Thermodynamic stability limits of YBaCo2O6−δ were determined by three independent methods such as coulometric titration combined with EMF method, electrical conductivity measurements, and homogenizing annealing of samples in atmosphere with controlled oxygen partial pressure. The coulometric titration was also employed for measurements of the oxygen nonstoichiometry in YBaCo2O6−δ as a function of pO2 at a given temperature. The original coulometric titration setup and the measurement procedure are described in detail elsewhere [18, 19].
Total conductivity of YBaCo2O6−δ oxide as a function of pO2 at a given temperature was measured using 4-probe dc-method, and the original setup described in detail elsewhere [20].
Absolute value of δ in YBaCo2O6−δ sample was determined by direct reduction by hydrogen flux in the TG setup (TG/H2). Experimental details for this method are given elsewhere [19].
For homogenizing annealing experiments, samples of two different compositions were employed. The first one was equimolar mixture of YCoO3 and BaCoO3, whereas the second one was single phase powder of YBaCo2O6−δ. In both cases, a sample was heated first to 700°C in an atmosphere with a given pO2 and then was equilibrated at this temperature for 72 h followed by quenching to room temperature. After that, the quenched sample was grinded in mortar, and its phase composition was determined by XRD. Then, annealing temperature was increased on 100°C, and the rest of the experimental procedure was similar to that described above. After annealing at the highest temperature, the gas atmosphere surrounding the sample was changed, and measurement procedure was repeated as described above. Oxygen partial pressure in the gas atmosphere around the sample was adjusted using a YSZ-based electrochemical oxygen pump installed in the outer regulating unit and governed by the automatic controller (Zirconia 318, Russia). Three gas atmospheres with pO2 = 1, 0.21, and 10−3 adjusted accordingly with gas flow rate of about 50 ml/min (to avoid oxygen partial pressure gradients along the sample) were used.
3. Results and Discussion
3.1. Sample Characterization
X-ray diffraction patterns of the as-prepared single phase powder samples of YCoO3, YBaCo2O6−δ, and BaCoO3 are given in Figures 1–3, respectively. It should be mentioned that the pattern of BaCoO3 was interpreted as a mixture of two compounds BaCoO3 and BaCoO2.61.
The space groups used for the XRD patterns indexing along with the refined cell parameters are given in Table 1 in comparison with those reported in literature.
As seen, the cell parameters found in the present study and those reported earlier are in good agreement with each other.
3.2. Oxygen Nonstoichiometry and Thermodynamic Stability of YBaCo2O6−δ
Oxygen nonstoichiometry in YBaCo2O6−δ as a function of T and pO2 was measured by means of coulometric titration technique in the ranges 800 ≤ (T, °C) ≤ 1050 and −5 ≤ log(pO2/atm) ≤ 0, respectively. Taking into account the absolute oxygen content determined by TG/H2 (5.016 ± 0.002 at 900°C in air), the coulometric titration curves measured for YBaCo2O6−δ at different temperatures were recalculated in pO2 dependences of its oxygen content, 6−δ, given in Figure 4.
As an example, such dependence obtained accordingly at 900°C is shown in Figure 5.
Vertical segments of the titration curve (Figure 5) correspond to the sample decomposition and clearly indicate its thermodynamic stability limits with respect to reduction (at low pO2) and oxidation (at high pO2). The curve plot enclosed between vertical segments corresponds to the oxygen content change in YBaCo2O6−δ within the thermodynamic stability region. It is worth noting that the pO2 dependence of oxygen content measured at this temperature and shown in Figure 5 exhibits inflection when oxygen content of the double perovskite reaches the value of 5. The same behavior was found earlier [18] for another double perovskite GdBaCo2O6−δ. However, unlike GdBaCo2O6−δ the double perovskite with yttrium possesses really narrow homogeneity range with respect to oxygen since oxygen content changes in the vicinity of 5 and its overall variation is less than 0.83% within the thermodynamic stability region at 900°C. In other words, such narrow homogeneity region corresponds to a small change of the average oxidation state of cobalt; that is, even slight reduction of Co3+ or oxidation of Co2+ may result in YBaCo2O6−δ oxide decomposition. Coulometric measurements carried out at 900°C were stopped after reaching the highest pO2 value, 0.72 atm, and the coulometric cell was fast cooled down to room temperature. XRD of the YBaCo2O6−δ sample cooled accordingly showed the presence of the yttrium cobaltite YCoO3 and the barium cobaltites BaCoO3 and BaCoO2.63 as products of YBaCo2O6−δ decomposition. Taking into account possible oxygen nonstoichiometry of the complex oxides, a decomposition reaction can be written aswhere a value of the oxygen content, 6−δ, depends on temperature and varies from 5.012 at 800°C up to 5.035 at 900°C. It is worth noting that reaction (1) is completely in line with the finding of the authors [15] that the perovskites YCoO3 and BaCoO3 are products of YBaCo2O6−δ decomposition in air (pO2 = 0.21 atm) at temperatures lower than 850°C.
In order to find products of YBaCo2O6−δ decomposition at low pO2 stability limit, its single phase sample was annealed at 1000°C in gas atmosphere with log(pO2/atm) = −4 for 12 hours and then quenched to ice at −18°C. The XRD pattern of the sample prepared accordingly showed the presence of YBaCo4O7, BaCo1−xYxO3−z, and Y2O3. Therefore, the decomposition of YBaCo2O6−δ at low pO2 stability limit occurs according to the following reaction:where , , , and , since the oxygen content in the hexagonal YBaCo4O7 is very close to 7 in the vicinity of its low pO2 stability limit; for example, it comes to 6.99 at 900°C and log(pO2/atm) = −3.5 [27].
It is worth noting that decomposition of YBaCo2O6−δ is quite different from that of another double perovskite, for instance GdBaCo2O6−δ, which decomposes at low pO2 according to completely different reaction [28]:
Thermodynamic stability limits of YBaCo2O6−δ determined accordingly are summarized in the stability diagram shown in Figure 6 as log(pO2) = f(1/T).
A nonlinear character of log(pO2) = f(1/T) dependence corresponding to low pO2 stability limit of YBaCo2O6−δ is related probably to a real composition of BaCo1−xYxO3−γ as a product of the decomposition reaction (2). Composition of both cation and anion sublattice of this compound is expected to depend significantly on temperature and oxygen partial pressure in ambient gas atmosphere.
Overall electrical conductivity of YBaCo2O6−δ was measured as a function of pO2 at a given temperature in the range 900 ≤ T, °C ≤ 1050 with step of 50°C. The overall conductivity measured as a function of pO2 at different temperatures is shown in Figure 7. As seen in Figure 7, the overall conductivity first slightly decreases with pO2 descent in ambient atmosphere down to certain threshold value (arround pO2 = 10−2, 10−3, 10−3.5, and 10−4 atm at 1050, 1000, 950, and 900°C, resp.), indicating in favor of electron holes as dominant charge carriers, and then abruptly drops down on about an order of magnitude; afterwards it remains unchanged in practical term upon further pO2 decrease. Such conductivity drop observed at low pO2 at all temperatures studied in the present work is obviously related to the YBaCo2O6−δ decomposition upon reaching the low pO2 stability limit.
For the sake of comparison, the values of pO2 at which the conductivity drop was observed are also given in Figure 6 depending on reciprocal temperature. As seen, the datasets on stability obtained by means of different techniques are in agreement with each other pretty well.
In order to confirm the high pO2 stability limit, a single phase sample of YBaCo2O6−δ was annealed at 800°C for 10 hours in air followed by quenching to room temperature. XRD of the so-prepared sample showed the presence of YCoO3 and BaCoO3−z. On the other hand, annealing of equimolar mixture of YCoO3 and BaCoO3−z for 72 hours under the same conditions was found not to lead to the formation of the double perovskite whilst it is formed in result of this mixture annealing at 900°C in air for the same time. These findings are in full coincidence with the stability diagram plotted (Figure 6) as well as with the data of the authors [15]. It follows from the stability diagram shown in Figure 5 that the double perovskite YBaCo2O6−δ is thermodynamically stable in air only at 850°C and higher temperatures while it can be formed only as a metastable phase below this temperature. The last conclusion is in full agreement with the results of the paper by Zhang et al. [14] where YBaCo2O6−δ was found to be stable at 850°C in air. The results of the thermodynamic stability investigation were further used to optimize the synthesis routine for YBaCo2O6−δ as described elsewhere [29].
4. Conclusion
The thermodynamic stability and oxygen nonstoichiometry of the double perovskite YBaCo2O6−δ was studied using coulometric titration technique. The stability diagram of YBaCo2O6−δ was plotted. The found limits of its thermodynamic stability were successfully verified by measurements of the overall conductivity as a function of oxygen partial pressure at given temperatures. YBaCo2O6−δ was shown to be thermodynamically stable in the range of pO2 between certain threshold values at a given temperature likewise YBaCo4O7±δ studied earlier [27]. For instance, YBaCo2O6−δ was found to be thermodynamically stable in air at 850°C and higher temperatures, whereas its thermodynamic stability at 900°C is limited by the range of oxygen partial pressures −3.56 ≤ log(pO2/atm) ≤ −0.14. Oxygen content in YBaCo2O6−δ was found to decrease slightly at 900°C from 5.035 at log(pO2/atm) = −0.14 to 4.989 in the atmosphere with log(pO2/atm) = −3.565. Such narrow homogeneity region corresponds to a small change of the average oxidation state of cobalt; that is, even slight reduction of Co3+ or oxidation of Co2+ may result in YBaCo2O6−δ oxide decomposition. YBaCo2O6−δ decomposes into the mixture of YCoO3 and BaCoO3−z at the high pO2 stability limit, whereas YBaCo4O7, BaCo1−xYxO3−γ, and Y2O3 were identified as the products of its decomposition at the low pO2 one.
Conflicts of Interest
The authors declare that there are no conflicts of interest regarding the publication of this paper.
Acknowledgments
This work was supported by the Ural Federal University within the framework of Act 211 of Government of the Russian Federation, Agreement no. 02.A03.21.0006.