Synthesis and thermal expansion property of (Ba_0.5Sr_0.5)_0.9Bi_0.1Co_0.8Fe_0.2O_3−δ cathode materials for IT-SOFCs

Abstract

Perovskite cathode material, (Ba_0.5Sr_0.5)_0.9Bi_0.1Co_0.8Fe_0.2O_3−δ (BSBCF-0.1), was synthesized by a combined citrate–EDTA complexing method for intermediate temperature solid oxide fuel cells. Bi was successfully added to the perovskite structure. A small amount of Bi could enter into the A-site of the perovskite structure, and ionic Bi would enter the B-site of the BSCF structure when the concentration was larger than 0.1. Thermal expansion coefficient (TEC) of BSBCF-0.1 is lower than that of undoped BSCF in the working temperature range. The decreased TEC of BSBCF may be more suitable for the thermal compatibility of cell components. The maximum power density is 220 and 270 mW cm⁻² at 600 and 650 °C, respectively. The result suggested that the perovskite BSBCF material could be considered as the cathode for IT-SOFCs.

---

1. Introduction

Solid oxide fuel cells (SOFCs), a new kind of generating equipment, have attracted much attention due to their high-efficiency, low-pollution, modular design and fuel flexibility. Recently, many efforts have been made to reduce the working temperature of SOFCs in order to avoid a series of problems such as the high-cost metallic interconnectors, chemical and thermal compatibility of cell components and performance degradation.^1–4 However, SOFCs operating at a relatively lower temperature need a special cathode material with relatively high conductivity and proper thermal expansion coefficient.^5,6 The perovskite structure (ABO₃) Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3−δ (BSCF), with the mixed ionic and electronic conductivity, was selected as a candidate cathode material owing to its high catalytic activity for oxygen reduction and good oxygen ion permeability.^7,8 It is well known that the electrochemical activity is deteriorated abruptly when the temperature is decreased for SOFCs, and the cell life has a close relationship with the compatibility between cathode and electrolyte materials. However, a larger thermal expansion coefficient (TEC) of BSCF will cause thermally incompatibility with electrolyte. There has been an attempt to resolve these drawbacks by the substitution of A- or B-site cations in the ABO₃ structure. Meng et al. reported that the Zr-substituted BSCF perovskite cathode material, Ba_0.5Sr_0.5(Co_0.6Zr_0.2)Fe_0.2O_3−δ, has excellent thermal stability, although it has an increased area specific resistance (ASR) compared to BSCF.⁹ Chemical stability for BSCF has also been improved by adding Ti in B-site of the ABO₃ perovskite oxide.¹⁰ The rare earth ions such as Sm³⁺,¹¹ Gd³⁺ (ref. 12) and Nd³⁺ (ref. 13) were also added into the A-site cations of BSCF to elevate the electrical conductivity, but this increased the TEC of BSCF and degenerated the thermally incompatibility between the cell's components. It is reported that Bi could be doped into SrFeO_3−δ perovskite structure, and the material is compatible with the Sm doped ceria electrolyte for low TEC.¹⁴ Considering that Bi could be introduced to the perovskite structure, it was selected to reduce the TEC and increase the conductivity for BSCF.

In this study, Bi was introduced to the perovskite oxide of BSCF to form a new (Ba_0.5Sr_0.5)_0.95Bi_0.05Co_0.8Fe_0.2O_3−δ cathode material for intermediate temperature SOFCs. The major aim is to develop a special cathode material with low TEC and enhanced electrical performance. The phase structure, TECs and electrochemical properties are also discussed.

2. Experiment

(Ba_0.5Sr_0.5)_1−xBi_xCo_0.8Fe_0.2O_3−δ (x = 0, 0.1, 0.125, 0.15) (noted as BSCF (x = 0) and BSBCF-x) powders were synthesized by a glycine–nitrate combustion process.¹⁵ The analytical reagents Ba(NO₃)₂, Sr(NO₃)₂, Bi(NO₃)₃·6H₂O, Co(NO₃)₂·6H₂O and Fe(NO₃)₃·9H₂O were used as starting materials. All the nitrate salts were dissolved in EDTA (ethylenediaminetetraacetic acid) and NH₄OH solution under stirring at 80 °C. Then, the citric acid solution was added, which was used as a complexing reagent. The molar ratio of metal ions : EDTA : citric acid was fixed at 1 : 1 : 1.5. The mixed solution was stirred and heated at 120 °C for 48 h. The BSBCF precursor was sintered at 800 °C for 12 h, and then at 1050 °C for 12 h to obtain the final powder. The powders were compacted into a cylinder with a diameter of 13 mm under 400 MPa uniaxial pressing. The compacted cylinders were sintered at 1100 °C for 6 h with a heating and cooling rate of 5 °C min⁻¹. Two sides of dense samples were polished to parallel by sand paper for TEC measurement. The approximate dimensions of BSCF, BSBCF-0.1, BSBCF-0.125, BSBCF-0.15 samples were about Φ11 × 7, Φ11 × 5, Φ11 × 3, Φ11 × 6 mm³, respectively. Densities exceeded 94% of the theoretical value as measured by the Archimedes method.

In cell fabrication, the dense electrolyte Ce_0.82La_0.18O_2−δ (LDC) was prepared by a solid-state reaction similar to our previous work.¹⁶ Anode materials were made of 80 wt% LDC–NiO (with a ratio of 60 : 40 wt%) and 20 wt% starch, and then mixed with appropriated terpineol to form a mixed slurry. Anode slurry was painted on one side of the dense LDC electrolyte, and then the bilayer was sintered at 1400 °C for 6 h. Four kinds of ground cathode powders (60%) were dispersed homogeneously into terpineol (40%) to obtain the cathode slurry. The mixed cathode slurry was painted on the other side of LDC supporter and sintered at 1100 °C for 4 h to form a single cell of BSBCF-x|LDC|NiO–LDC. The fuel cell has an electrolyte thickness of ∼1 mm, cathode thickness of 80 μm and anode thickness of 0.7 mm. Single fuel cells with an effective cathode area of 0.45 cm² were prepared. In the cell performance test, the cathode was exposed to air, and the anode side was exposed to 3% H₂O + humidified H₂ at a flow rate of 50 ml min⁻¹.

The crystal structures of the powders were characterized by X-ray diffraction (XRD) using an X’pert PRO diffractometer and Cu Kα radiation in the 2θ range of 10–80°. Scanning electron microscopy (SEM) and energy dispersive spectrometry (EDS) were employed to examine the morphology and element distribution of the sample. The linear thermal expansion coefficient was measured on the thermal expansion apparatus (LINSEIS DIL L76) using quartz as the reference material with a heating rate of 5 °C min from room temperature to 800 °C. The performances of the fuel cells were determined by a PARSTAT 2273 electrochemical workstation (Princeton, USA) at 600 °C and 650 °C.

3. Results and discussion

3.1 Crystal structure

Fig. 1 shows the X-ray diffraction patterns of the BSBCF-x (x = 0, 0.1, 0.125, 0.15) samples. It can be observed that all the samples possess the cubic perovskite-type structure similar to that reported by Niedrig.¹⁷ The lattice parameters of BSBCF-x were calculated by jade 6.0. The calculated parameters are shown in Table 1. The calculation lattice of BSCF agrees well with the values determined by previous studies.^13,18 From the X-ray diffraction patterns, the main diffraction peaks of BSBCF-x shift gradually to higher angles at a low Bi doping concentration compared to that of the perovskite BSCF oxide. In the right part of Fig. 1, the magnified image shows an obvious shift of 2θ. The reason for the change was due to the successful introduction of Bi³⁺ into the lattice of BSCF. The pre-sintering temperature of 800 °C is the key factor to reduce the volatilization for doping Bi³⁺, because Bi could be volatile when the sintering temperature is higher than the melting point of cubic Bi₂O₃ (825 °C) in an air environment.¹⁹

Fig. 1 X-ray diffraction patterns for BSBCF-x (x = 0, 0.1, 0.125, 0.15).

Table 1 The lattice parameters of (Ba_0.5Sr_0.5)_1−xBi_xCo_0.8Fe_0.2O_3−δ

x in (Ba0.5Sr0.5)1−xBixCo0.8Fe0.2O3−δ	Structure	A (Å)
0	Cubic	4.0199
0.1	Cubic	3.9980
0.125	Cubic	3.9824
0.15	Cubic	4.0701

---

With the substitution of the larger Ba²⁺ (0.175 nm) and Sr²⁺ (0.158 nm) for smaller Bi³⁺ (0.103 nm) in the A-site,^20,21 the lattice volume of BSCF cathode decreased. As the substitution concentration increased, the reduction rate of the lattice volume slowed down compared to BSBCF-0.1 for BSBCF-0.125. The lattice volume and the average ionic radius accordingly increased when the addition further increased the redundancies of Bi³⁺ entered into the B-site of the perovskite structure.²² The diffraction peaks moved to the direction of small angle of BSBCF-0.15. This indicates that the element of Bi could enter the B-site of BSCF when the concentration of x is larger than 0.1.

3.2 SEM-EDS analysis

To illustrate the distribution state of elements, SEM-EDS micro-region composition analysis was taken for the unground BSBCF-x powders. For comparative analysis, BSCF, BSBCF-0.1 and BSBCF-0.15 were performed. As shown in Fig. 2(a)–(c), the powders are easily sintered as the Bi concentration increases and grains of large size are observed. To study the distribution of elements, the element mapping mode was selected at random on the surface of the three powders. The element composition of BSBCF-x are calculated and summarized in Table 2. The chemical composition of BSCF is as follows: Ba 7.34%, Sr 5.80%, Co 7.75%, Fe 2.09%. The BSBCF-0.1 exhibits a composition of Ba 10.30%, Sr 11.10%, Bi 1.64%, Co 17.53%, Fe 5.28%. The ratios of A-site to B-site of two (BSCF and BSBCF-0.1) perovskite oxides are close to 1. In the EDS spectrums (Fig. 2(d)), it is easy to distinguish the peak position of Bi between BSCF and BSBCF-0.1 at 2.42–2.53 keV. The binding energies of all elements are shown in Table 3. The peak of Bi is also evident in the BSBCF-0.15 spectrum. However, the total concentration of Ba and Sr equals to the sum of Sr and Co for BSBCF-0.15. Whether the Bi was added into the A-site or B-site, it will change the balance of stoichiometric ratio. Therefore, we conclude that Bi could enter the A-site of the perovskite structure when the concentration is smaller than 0.1. Otherwise, it would enter the B-site of the perovskite structure.

Fig. 2 SEM-EDS micrographs of the BSBCF-x powders: (a) BSCF; (b) BSBCF-0.1; (c) BSBCF-0.15; (d) a representative EDS spectrum of the BSBCF-x powders.

Table 2 Composition of the detected elements of the random area on the surfaces of the BSBCF-x samples (Fig. 2(a)–(c))

Element	Atomic (%)
	BSCF	BSBCF-0.1	BSBCF-0.15
O K	77.02	54.15	54.90
Fe K	2.09	5.28	5.37
Co K	7.75	17.53	16.28
Sr K	5.80	11.10	12.12
Ba L	7.34	10.30	9.09
Bi L	—	1.64	2.25
Total	100	100	100

---

Table 3 Binding energies of elements (Ba, Sr, Co, Fe, Bi, O)

Element	Binding energies (keV)
Ba	3.945	4.466	4.828	5.157	5.531	5.797
Sr	1.582	1.649	1.807	14.165	15.836	16.085
Co	0.678	0.694	0.776	6.93	7.65	—
Fe	0.628	0.705	6.404	7.058	—	—
Bi	1.883	2.424	2.526	2.736	9.42	10.839
O	0.525	—	—	—	—	—

---

3.3 Thermal expansion analysis

Fig. 3(a) shows the thermal expansion behaviors of BSBCF-x (x = 0, 0.1, 0.125, 0.15) samples measured in the range from room temperature to 800 °C in air. The TEC of BSCF is similar to the results reported by Li.¹³ High densities of the sample minimize the influence of pore in our TEC test. The average TECs of BSBCF-x are shown in Table 4. It is found that the TEC of BSBCF-0.1 is lower than that of BSCF. The lower TEC of the cathode indicates the better integration of BSBCF-0.1 to the LDC electrolyte. Even though the average TEC is similar, but the curve is significantly different, especially under working temperature below 650 °C. The relative length variation is much less for BSBCF-0.1 than that of BSCF, especially over 450 °C. The TEC slightly reduces with a little content of Bi³⁺ addition. It could be mainly attributed to the decrease in the average ionic radius. When the addition of Bi³⁺ is further increased, the relative change is much more steady in the range of about 250 °C to 450 °C. The material expansion is larger than BSCF below 500 °C and more stable at a high temperature. One of the reasons is that some of Bi³⁺ entered into the B-site of the perovskite oxide. The additional oxygen vacancies can be generated by A-site deficiency.²³ Another reason is that the average ionic radius could be enlarged by the substitution of Bi³⁺ (0.103 nm) for Co⁴⁺ (0.053) and Fe⁴⁺ (0.059).²² When the temperature is elevated, the lattice oxygen becomes more active and is released from the lattice. The variable tendency of TEC for BSBCF-x presents the clear change in Fig. 3(b). The turning points of the thermal expansion curve gradually shift by oxygen emission. This indicates that the stabilization of the BSCF cathode material with Bi³⁺ substitution is increased steadily with a small value of TEC and less variation from room temperature to the working temperature for SOFCs.

Fig. 3 Thermal expansion curves (a) and TECs (b) for BSBCF-x (x = 0, 0.1, 0.125, 0.15).

Table 4 TEC values of (Ba_0.5Sr_0.5)_1−xBi_xCo_0.8Fe_0.2O_3−δ

Sample	TEC (10^−6 K^−1)		Reference
	25–450 °C	25–800 °C
BSCF	12.61	17.85	This study
BSBCF-0.1	12.16	17.46	This study
BSBCF-0.125	13.53	17.06	This study
BSBCF-0.15	14.72	18.53	This study
BSCF	—	19.7	13
BSCF	—	19	18

---

3.4 Electrochemical behaviors

In our experiment, we found that the relative density of cathode is related to the addition of Bi. Cathode materials sintered more easily when the volatile Bi was added steadily. High density indicates low gas diffusion. In order to evaluate the electrochemical performance, the low TEC cathode material of BSBCF-0.1 was selected to fabricate the cell of BSBCF-0.1|LDC|Ni + LDC, and the cell of BSCF|LDC|Ni + LDC was also fabricated for comparison. Fig. 4 shows the microstructure of the fracture surface of BSBCF-0.1 cathode sintered at 1100 °C on the LDC electrolyte. From the image, it can be clearly seen that the electrode has a porous structure. The optimized temperature for the cell fabrication was found to be between 1000 and 1150 °C.

Fig. 4 SEM micrograph of the intersection of BSBCF-0.1 cathode on LDC electrolyte.

As illustrated in Fig. 5, the electrochemical performances (I–V and power density curves) of the single fuel cell were measured at 600 °C and 650 °C. The open circuit voltage (OCV) is about 0.82 V at a temperature of 600 °C for two single cells corresponding to the CeO₂-based electrolyte. A slight decrease in the OCV was found with an elevation in the work temperature.

Fig. 5 Cell voltage and power density as functions of current density from Cathode|LDC|Ni-LDC fuel cells. Cathode (a) BSCF; cathode (b) BSBCF-0.1. Measured in humidified H₂ and air at 600 °C and 650 °C (anode H₂ flow rate = 50 ml min⁻¹, cathode exposed to air).

From the curve of output power density, the maximum power densities were about 220 and 270 mW cm⁻² at 600 and 650 °C, respectively, for the cell with the BSBCF-0.1 cathode (Fig. 5(a)), whereas the peak power densities were 100 mW cm⁻² (600 °C) and 130 mW cm⁻² (650 °C) for the cell constructed with the BSCF cathode (Fig. 5(b)). The output power density is larger than Zr-doped Ba_0.5Sr_0.5(Co_0.6Zr_0.2)Fe_0.2O_3−δ cathode material for SOFC (139 mW cm⁻² at 650 °C).⁹ However, the power density is lower than the electrode-supported fuel cell because of the thin electrolyte.⁷ Therefore, the electrode-supported structure would be taken to enhance the cell performance in the future work.

The relatively lower output power density is related to the following reasons: one of the reasons is that the mixed conductivity of the CeO₂-based electrolyte increased the internal electric energy consumption of the cell. Another is that the low current density is possibly unfavourable to high power density. Fig. 6 shows the electrochemical impedance spectra of BSCF and BSBCF-0.1 fuel cells at 650 °C. The standard impedance spectra of cells consist of three parts: the real axis intercept, two arcs and Warburg resistance. The curve intercept with the real axis at high frequency is attributed to the ohm loss. The two arcs mainly consist of cathode electrochemical reaction. Anode reaction arc always buries into a larger arc because of its higher activity than that of the cathode. Warburg resistance is related to the diffusion of gas in the electrode. According to the impedance spectra, ohm resistance values were almost the same between BSCF and BSBCF-0.1. However, the BSCF fuel cell has a larger flat arc, which consists of the diffusion and reaction of gas in the electrode, than that of BSBCF-0.1. The large resistance depressed the efficiency of the fuel cell. Moreover, the low current densities were observed for the two cells constructed by the BSCF or BSBCF-0.1 cathode materials. However, the maximum output power density of the cell constructed by the BSBCF-0.1 cathode material is better than BSCF. This indicates that the electrochemical reaction activity of BSCF cathode was enhanced by the Bi addition. The whole effect could be mainly related to the increasing density of oxygen vacancies per unit by the addition of small ionic radius of Bi³⁺ to the A-site of the perovskite structure.²¹ The effect of doped Bi ion on the oxygen vacancies resulted from the lattice shirking for BSBCF-0.1. Ding et al. ascribed the better electrochemical performance of the A-site La doped perovskite material to the improved oxygen adsorption and desorption.²⁴

Fig. 6 Electrochemical impedance spectra of BSCF and BSBCF-0.1 fuel cells at 650 °C.

4. Conclusion

A new series of cathode materials (Ba_0.5Sr_0.5)_1−xBi_xCo_0.8Fe_0.2O_3−δ (x = 0, 0.1, 0.125, 0.15) have been prepared by a citric–nitrate process. Volatile Bi was successfully added to the perovskite oxide of BSCF. When the content of Bi was less than 0.1, the element could mainly enter the A-site of the perovskite structure. The average small ionic radius strengthened the stability of BSBCF-0.1 lattice with small expansion. As the content further increased, Bi could enter the B-site. The temperature of oxygen loss is steadily elevated as the Bi content is increased. A relatively lower thermal expansion coefficient of BSBCF-0.1 cathode material would be more suitable for the ceria-base electrolyte. Moreover, the output power density of the cell, which was constructed with the BSBCF-0.1 cathode, is 270 mW cm⁻² at 650 °C. Therefore, BSBCF-0.1 is a promising cathode material for IT-SOFCs.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (51302102) and Huaibei Normal University Youth Research Project (Grant no: 2014xq001).

References

1. T. Hibino, A. Hashimoto, T. Inoue, J. I. Tokuno, S. I. Yoshida and M. Sano, Science, 2000, 288, 2031–2033.
2. E. D. Wachsman and K. T. Lee, Science, 2011, 334, 935–939.
3. A. Ecija, K. Vidal, A. Larrañaga, A. Martínez-Amesti, L. Ortega-San-Martín and M. I. Arriortua, Solid State Ionics, 2013, 235, 14–21.
4. W. Araki and J. Malzbender, Solid State Ionics, 2013, 233, 67–72.
5. T. Wei, Y. H. Huang, L. Jiang, J. H. Yang, R. Zeng and J. B. Goodenough, RSC Adv., 2013, 3, 2336–2340.
6. S. Park, S. Choi, J. Shin and G. Kim, RSC Adv., 2014, 4, 1775–1781.
7. Z. Shao and S. M. Haile, Nature, 2004, 431, 170–173.
8. F. S. Baumann, J. Fleig, H. U. Habermeier and J. Maier, Solid State Ionics, 2006, 177, 3187–3191.
9. X. Meng, B. Meng, X. Tan, N. Yang and Z. F. Ma, Mater. Res. Bull., 2009, 44, 1293–1297.
10. L. Bi, E. Fabbri and E. Traversa, Solid State Ionics, 2012, 214, 1–5.
11. S. Li, Z. Lü, B. Wei, X. Huang, J. Miao, G. Cao, R. Zhu and W. Su, J. Alloys Compd., 2006, 426, 408–414.
12. Z. Li, B. Wei, Z. Lü, Y. Zhang, K. Chen, J. Miao and W. Su, Ceram. Int., 2012, 38, 3039–3046.
13. S. Li, Z. Lü, X. Huang and W. Su, Solid State Ionics, 2008, 178, 1853–1858.
14. Y. Niu, W. Zhou, J. Sunarso, L. Ge, Z. Zhu and Z. Shao, J. Mater. Chem., 2010, 20, 9619–9622.
15. Z. Shao, W. Yang, Y. Cong, H. Dong, J. Tong and G. Xiong, J. Membr. Sci., 2000, 172, 177–188.
16. D. C. Li, M. J. Chao, J. M. Yu, M. Y. Li, J. J. Zhang and E. J. Liang, Mater. Lett., 2012, 86, 171–173.
17. C. Niedrig, S. Taufall, M. Burriel, W. Menesklou, S. F. Wagner, S. Baumann and E. Ivers-Tiffee, Solid State Ionics, 2011, 197, 25–31.
18. S. McIntosh, J. F. Vente, W. G. Haije, D. H. A. Blank and H. J. M. Bouwmeester, Chem. Mater., 2006, 18, 2187–2193.
19. A. M. Azad, S. Larose and S. A. Akbar, J. Mater. Sci., 1994, 29, 4135–4151.
20. S. V. Moharil, B. S. Nagrare and S. P. S. Shaikh, Int. J. Hydrogen Energy, 2012, 37, 5208–5215.
21. Y. J. Niu, W. Zhou, J. Sunarso, L. Ge, Z. Zhu and Z. Shao, J. Mater. Chem., 2010, 20, 9619–9622.
22. Z. P. Shao, G. X. Xiong, Y. Cong and W. S. Yang, J. Membr. Sci., 2000, 164, 167–176.
23. W. L. Li, K. Lu and Z. B. Xia, J. Power Sources, 2013, 237, 119–127.
24. X. F. Ding, X. Kong, J. J. Jiang, C. Cui and X. X. Guo, Mater. Res. Bull., 2010, 45, 1271–1277.

---