Evaluation of Ba-deficient PrBa_1−xFe₂O_5+δ oxides as cathode materials for intermediate-temperature solid oxide fuel cells

Abstract

Cobalt-free double perovskite oxides are promising cathode materials in intermediate-temperature solid oxide fuel cells, and often suffer from low activity in oxygen reduction reactions. Here, we report on Ba deficiency as an effective strategy to enhance the electrochemical performance of cobalt-free double perovskite PrBaFe₂O_5+δ, which is related to the formation and redistribution of oxygen vacancies in double perovskite. The effect of Ba deficiency on crystal structure, surface properties, oxygen content, electrical conductivity, thermal expansion, chemical compatibility with gadolinium-doped ceria (Gd_0.1Ce_0.9O_1.95) electrolyte, microstructure, and electrochemical performance of PrBa_1−xFe₂O_5+δ (x = 0.00–0.03) was evaluated systematically. Our preliminary results suggest that Ba deficiency is a feasible means to tailor the physico- and electrochemical properties of cobalt-free double perovskite, and that PrBa_0.97Fe₂O_5+δ is a potential cathode material for intermediate-temperature solid oxide fuel cells.

---

1. Introduction

The energy crisis and environmental problems caused primarily by the current widespread and inefficient use of fossil fuels has intensified the need for new energy conversion devices. Among available technologies capable of converting stored chemical energy to usable electrical energy, solid oxide fuel cells (SOFCs) offer an unmatched working efficiency, low gaseous pollutant emissions, and excellent fuel flexibility, and are considered to represent a key component of future energy solutions.^1,2 One strategy that has been proposed for the commercialization of SOFCs, while achieving enhanced durability and cost reduction, is to lower the working temperature to an intermediate temperature (IT) range of 600–800 °C. At these lower temperatures, however, increased polarization resistance at the cathode that results from a high oxygen reduction reaction (ORR) activation energy, becomes the primary factor that determines the overall cell performance. Thus, the exploration of novel cathode materials with high catalytic activity for ORR is of significant importance in IT-SOFC development.

In recent years, double perovskites with general formula LnBaCo₂O_5+δ (Ln = lanthanide) have been proposed as potential cathode materials for IT-SOFCs owing to their rapid oxygen surface exchange, bulk diffusion kinetics, and high electrical conductivities. These oxides possess 112-type perovskite structures with alternating stacking layers of [CoO₂]–[LnO_δ]–[CoO₂]–[BaO] along the c-axis. This periodic structure is expected to reduce the oxygen ion transport activation energy along the ab plane significantly and has been reported to enhance the oxygen transport rate effectively compared with transport rates exhibited by disordered counterparts of these materials.^3,4 The structural, thermal, electrical, and electrochemical properties of double perovskites have thus been studied extensively.^5–10 Unfortunately, as for other cobalt-based perovskites, the thermal expansion coefficient (TEC) of LnBaCo₂O_5+δ from 30–900 °C (such as 23.4 × 10⁻⁶ °C⁻¹ for PrBaCo₂O_5+δ)⁸ is much larger than that of commonly used electrolytes (10–13 × 10⁻⁶ °C⁻¹)¹¹ and sealing materials (11–14 × 10⁻⁶ °C⁻¹).¹² This leads to a significant thermal mismatch that can produce interfacial stresses between the cathode and other components, and lead to cell cracking and degradation. The volatility of these compounds at elevated temperatures and the relatively high price of cobalt also limit the practical application of LnBaCo₂O_5+δ as cathode material for IT-SOFCs. Substituting other, more readily available transition metals such as iron and nickel for cobalt may provide a solution to these challenges.^13–19 As an example, monotonic decreases in TEC of LnBaCo_2−xFe_xO_5+δ (Ln = Nd or Gd) with increasing Fe content were observed by Kim et al.¹⁷ and are attributed to the increased strength of the Fe–O bond compared with the Co–O bond and the reduced contribution of the low- to high-spin transition of Co³⁺ to TEC. The thermal and chemical compatibilities between PrBaCo_2−xFe_xO_5+δ and Sm_0.2Ce_0.8O_1.90 electrolyte were reported to be improved by Fe substitution for Co.¹³ For this reason, various series of cobalt-free double perovskites have been proposed, such as PrBaFe₂O_5+δ,^13–16 NdBaFe₂O_5+δ,^15,17 and GdBaFe₂O_5+δ,^15,18,19 as potential cathode materials for IT-SOFCs. However, since Fe ions are less conductive and more stable than Co ions, cobalt-free double perovskites often experience significant decreases in electrical conductivity and electrochemical catalytic activity in the ORR compared with cobalt-based double perovskites such as LnBaCo₂O_5+δ. Therefore, it is important to establish feasible strategies to enhance the electrochemical performance of cobalt-free double perovskites.

The introduction of A-site cationic deficiencies is a useful means of modifying the physical and chemical properties of perovskite oxides.^8,20–26 For example, Liu et al.²⁰ and Yang et al.²¹ observed gradual increases in electrical conductivity and catalytic activity for ORR as the extent of Ba deficiency in Ba_1−xCo_0.7Fe_0.2Ni_0.1O_3−δ was increased from x = 0.00 to 0.15. In addition, maximum power density values of 1062 mW cm⁻² (600 °C) and 1139 mW cm⁻² (650 °C) were obtained for Ba_1−xCo_0.7Fe_0.2Ni_0.1O_3−δ (x = 0.10) oxide. Our group has found that A-site Ba-deficient cobalt-based double perovskites LnBaCo₂O_5+δ (Ln = La or Pr) show a significantly enhanced electrochemical performance.^8,22,23 For example, Ba-deficient PrBa_1−xCo₂O_5+δ (x = 0.08) oxide exhibits a polarization resistance from 600–700 °C that is approximately half that of the stoichiometric oxide PrBaCo₂O_5+δ. This effect likely originates from the generation of additional oxygen vacancies in Ba-deficient oxides that facilitate two-dimensional layer diffusion of oxygen along the ab plane and interlayer diffusion along the c-axis.^8,22,24,25 A-site cation deficiency is expected to be an effective means to improve the electrochemical performance of cobalt-based double perovskites. However, since the physical and chemical properties of oxides are determined by their compositions at the A and B sites, no reports exist to data on the effects of A-site deficiencies on the properties of cobalt-free double perovskite LnBaFe₂O_5+δ with regard its application as cathode material for IT-SOFCs.

We studied the effects of Ba deficiency on crystal structure, oxygen non-stoichiometry, electrical conductivity, thermal expansion, and electrochemical performance of PrBa_1−xFe₂O_5+δ. Preliminary results suggest that cobalt-free double perovskite PrBa_1−xFe₂O_5+δ, in which the Ba deficiency is x = 0.03, is a promising cathode material for IT-SOFCs.

2. Experimental

Cobalt-free double perovskites PrBa_1−xFe₂O_5+δ (x = 0.00–0.10) were synthesized via a combined ethylene diamine tetraacetic acid (EDTA)/citric acid sol–gel method that has been described in detail in our previous publications.^8,22 Briefly, the desired amount of metal nitrites (Pr(NO₃)₃·6H₂O, Ba(NO₃)₂, and Fe(NO₃)₂·6H₂O, all analytical grade) was dissolved in an EDTA/NH₃·H₂O solution (∼pH 6) with heating and stirring, after which a citric acid/NH₃·H₂O solution (∼pH 6) was added at a total metal ion :EDTA : citric acid molar ratio of 1 : 1 : 2. The resulting solution was evaporated at 80 °C over several hours to obtain a viscous gel. The obtained gels were pre-calcined at 250 °C and then calcined in air at 1200 °C for 7 h to generate a double perovskite structure.

The phase structures of the as-synthesized PrBa_1−xFe₂O_5+δ (x = 0.00–0.10) powders and calcined PrBa_1−xFe₂O_5+δ–gadolinium-doped ceria (Gd_0.1Ce_0.9O_1.95, GDC) mixtures were characterized by X-ray diffraction (XRD; Rigaku D/Max 2400) from a 2θ of 20–80°. Oxygen contents of the PrBa_1−xFe₂O_5+δ (x = 0.00–0.03) oxides at various temperature were characterized by iodometric titration method^27,28 and thermo-gravimetric analysis (TGA; Netzsch STA449C). Before the iodometric titration and TGA measurements, the PrBa_1−xFe₂O_5+δ (x = 0.00–0.03) powders were vacuum dried at 100 °C for 4 h to remove the surface absorbed gases (i.e. H₂O, CO₂ etc.). The TGA was conducted in air, fabricated by mixing high purity N₂ and O₂ in the volume ratio of 79 : 21, from room temperature to 900 °C at a heating rate of 2 °C min⁻¹. Conductivity data were acquired in air from 300–800 °C using a direct current (DC) four-electrode method. The chemical state of the transition metal ions was characterized by a multifunctional imaging electron spectrometer (Thermo ESCALAB 250XI) with a monochromatized microfocused Al Kα (1486.6 eV) radiation source. Thermal expansion data were obtained in air using a dilatometer (Netzsch DIL 402 PC) with a heating rate of 2 °C min⁻¹ from 30–900 °C. The electrochemical performance of the PrBa_1−xFe₂O_5+δ (x = 0.00–0.03) oxides were characterized by electrochemical impedance spectra, using GDC as the electrolyte in a PrBa_1−xFe₂O_5+δ/GDC/PrBa_1−xFe₂O_5+δ symmetrical cell configuration. Dense GDC disks (diameter: ∼10 mm, thickness: ∼1 mm) were prepared by pressing GDC powder uniaxially at 200 MPa followed by sintering at 1280 °C for 4 h. Slurries that were prepared by mixing the cathode powders with α-terpineol and ethyl cellulose were screen-printed onto both sides of the GDC disks and the disks were then calcined at 1000 °C for 2 h to obtain porous cathode layers. Electrochemical impedance spectra data for these symmetrical cells were collected using a multi-channel potentiostat/galvanostat (VMP2) with integrated frequency response analyzer (Biologic) by applying an open circuit voltage from 0.1 Hz to 0.1 MHz. Morphologies of the post-test cells were analyzed by field emission scanning electron microscopy (FESEM, JSM-7001F). The specific surface area and pore size distribution of the cathode layers were measured by Brunauer–Emmett–Teller (BET) adsorption–desorption isotherms of N₂ technology with an automatic analyzer (Nova 3000e, Quantachrome). Single cells with the configuration of Ni-YSZ/YSZ/GDC/PrBa_1−xFe₂O_5+δ were fabricated by screen printed cathode slurry onto the commercial Ni-YSZ/YSZ/GDC half cells (NingBo SOFCMAN Energy Technology Co., Ltd) and calcined at 1100 °C for 2 h. The as-obtained single cells were sealed on one side of alumina tube with the help of silver paste (DAD-87, Shanghai Institute of Synthesized Risen) and then assembled to the home-made cell test system. The current–voltage (I–V) and current–power (I–P) curves were collected by a Solartron 1287 potentiostat with humidified H₂ (∼100 mL min⁻¹) as fuel and flowing air as oxide.

3. Results and discussion

3.1 Phase structure

XRD patterns of the PrBa_1−xFe₂O_5+δ (x = 0.00–0.10) powders after calcined at 1200 °C for 7 h in air are shown in Fig. 1. As can be seen from Fig. 1a and from the data in Table 1, highly crystalline, layered perovskite phases with an orthorhombic lattice geometry and Pmmm space group were obtained for each PrBa_1−xFe₂O_5+δ oxide. The structural parameters of the cation stoichiometric PrBaFe₂O_5+δ are comparable with previously reported results.^13,16 The diffraction peaks were observed to shift to higher angles (Fig. 1b) with increase in Ba deficiency (x), which suggests lattice shrinkage (Table 1). A Pr₆O₁₁ impurity phase (JCPDS# 42-1121) can be identified based on minor diffraction peaks that appear at approximately 28° in samples for which x = 0.05–0.10, which indicates that the value of the A-site Ba deficiency fraction (x) was limited to approximately 0.03 in the PrBa_1−xFe₂O_5+δ oxides. Therefore, evaluation hereafter in this paper focuses on samples with x = 0.00 and 0.03.

Fig. 1 Whole (a) and magnified (b) XRD patterns of PrBa_1−xFe₂O_5+δ (x = 0.00–0.10) powders calcined at 1200 °C for 7 h in air.

Table 1 Lattice parameters of PrBa_1−xFe₂O_5+δ (x = 0.00–0.10) oxides

Samples	Space group	a (Å)	b (Å)	c (Å)	V (Å^3)
x = 0.00	Pmmm	3.952	3.979	7.869	123.740
x = 0.03	Pmmm	3.932	3.943	7.839	121.535
x = 0.05	Pmmm	3.927	3.929	7.850	121.119
x = 0.10	Pmmm	3.900	3.921	7.829	119.720

---

3.2 X-ray photoelectron spectroscopy analysis

To identify the surface properties of PrBaFe₂O_5+δ and PrBa_0.97Fe₂O_5+δ, the chemical state of elemental O, Pr, and Fe was measured by X-ray photoelectron spectroscopy (XPS) and normalized with a C 1s peak to 284.6 eV. Fig. 2 shows the O 1s XPS spectra and fitting curves of PrBaFe₂O_5+δ and PrBa_0.97Fe₂O_5+δ. Three distinguishable peaks with binding energies from 528.5–533.0 eV are visible. The peak at 528.5 eV is characterized by lattice O²⁻ ions (O_L) of the perovskite structure, whereas the other two peaks with binding energy of 530.9 and 533.0 eV probably originate from carbonate structures (O_C) and hydroxyl groups (O_H).^29–32 The appearance of peaks at 530.9 and 533.0 eV indicates the reactivity of PrBaFe₂O_5+δ and PrBa_0.97Fe₂O_5+δ with CO₂ and H₂O when exposed to ambient atmosphere, respectively. A similar phenomenon was observed in other perovskite-based materials and is reported to be closely related to the existence of alkaline-earth elements in the materials.^29,31,32

Fig. 2 O 1s core-level spectra and fitted lines of PrBa_1−xFe₂O_5+δ at room temperature: (a) x = 0.00 and (b) x = 0.03.

Fig. 3 shows Pr 3d core-level spectra and corresponding fitting curves of PrBaFe₂O_5+δ and PrBa_0.97Fe₂O_5+δ samples at room temperature. The main peaks with binding energy of 932.2–932.5 eV and 952.9–953.1 eV can be assigned to Pr⁴⁺ 3d_5/2 and 3d_3/2, whereas the peak pairs located at ∼928.0/933.6 eV and 949.0/956.4 eV originate from Pr³⁺ 3d_5/2 and 3d_3/2.^29,33 A similar result was also observed by Jin et al. in double perovskite PrBaCo_2/3Fe_2/3Cu_2/3O_5+δ³⁰ and indicated the coexistence of mixed valences of Pr⁴⁺ and Pr³⁺ in PrBaFe₂O_5+δ and PrBa_0.97Fe₂O_5+δ samples. Pr⁴⁺ is the dominant chemical state of Pr in both samples as is characterized by the main spectrum at 932.2–932.5 eV and 952.9–953.1 eV.

Fig. 3 Pr 3d core-level spectra and fitted lines of PrBa_1−xFe₂O_5+δ at room temperature: (a) x = 0.00 and (b) x = 0.03.

Fig. 4 shows the Fe 2d core-level spectra and corresponding fitting curves of samples of PrBaFe₂O_5+δ and PrBa_0.97Fe₂O_5+δ at room temperature. Double peaks at 711.2 and 724.2 eV represent Fe⁴⁺ 2d_3/2 and 2d_1/2, whereas the other set of peaks with binding energies of 709.6 and 722.9 eV represent Fe³⁺ 2d_3/2 and 2d_1/2, respectively.³³ The satellite peak with binding energy of 717.5 eV can be ascribed to the Fe³⁺ species.^30,33 No peak related to the Fe²⁺ species was observed in the XPS spectra, which indicates that the valences of Fe in both samples are 3+ and 4+.

Fig. 4 Fe 2p core-level spectra and fitted lines of PrBa_1−xFe₂O_5+δ at room temperature: (a) x = 0.00 and (b) x = 0.03.

Based on the fitting results of the XPS spectra for O 1s, Pr 3d, and Fe 2p, the integrated intensity ratios of O_L/(O_C + O_H), Pr⁴⁺/Pr³⁺, and Fe⁴⁺/Fe³⁺ were calculated and are given in Table 2. Within the experimental error, no detectable changes are observed for O_L/(O_C + O_H) and Pr⁴⁺/Pr³⁺. Compared with the parent perovskite (PrBaFe₂O_5+δ), a slight decrease in Fe⁴⁺ ratio was observed for the PrBa_0.97Fe₂O_5+δ sample, which suggests that the introduction of Ba deficiency to the cobalt-free double perovskite results in a reduction rather than an oxidation of transition metal (Fe). To maintain the electrical neutrality of the oxide, a reduction in positive charge can be balanced by the formation of additional oxygen vacancies at the PrBa_0.97Fe₂O_5+δ surface.

Table 2 Percentage contribution from core electrons for O, Pr, and Fe

Samples	OL	OC + OH	Pr^4+	Pr^3+	Fe^4+	Fe^3+
x = 0.00	43.0%	57.0%	74.9%	25.1%	48.3%	51.7%
x = 0.03	43.0%	57.0%	74.8%	25.2%	47.7%	52.3%

---

3.3 Oxygen content

Oxygen contents of the PrBa_1−xFe₂O_5+δ (x = 0.00–0.03) oxides at various temperature in air were characterized by iodometric titration method and thermo-gravimetric analysis technique,^6,28 and presented in Fig. 5. As depicted in this figure, three distinguishable stages, the oxygen content kept constant with higher temperature (25–230 °C), absorption of oxygen in the surrounding atmosphere with higher temperature (230–340 °C), and release of lattice oxygen with higher temperature (340–900 °C), were observed for both samples in the measured temperature range. For the fist stage, no detectable changes in oxygen contents were observed for both samples. Similar phenomenon were observed in other perovskite oxides,^6,28 and can be attributed to the relatively low temperatures for which the oxygen ions are expect to “frozen” in the lattice of the PrBa_1−xFe₂O_5+δ (x = 0.00–0.03) oxides. For the second stage, slightly increases in oxygen contents were observed for both samples with higher temperature which properly comes from the difference in the gases oxygen content of the surrounding atmosphere between the muffler furnace (where the samples were synthesized) and the TGA furnace (where the samples were tested). Gradually decreases in oxygen content with higher temperature were obtained in the third stage which properly related to the thermal induced release of lattice oxygen.^6,28 Furthermore, at a given temperature, PrBa_1−xFe₂O_5+δ (x = 0.03) exhibits much lower oxygen contents than that of the parent oxide (PrBaFe₂O_5+δ) in the measured temperature range indicating the formation of oxygen vacancies as the charge compensation for the Ba-deficiency in the cobalt-free double perovskites. It is worth noting that the formation of additional oxygen vacancies is expected to be beneficial in promoting the catalytic activity for ORR.³⁴

Fig. 5 Oxygen content (5 + δ) of PrBa_1−xFe₂O_5+δ (x = 0.00–0.03) oxides measured in air.

3.4 Electrical conductivity

Fig. 6 summarizes the electrical conductivities of samples in air at various temperatures. The oxygen ion conductivity of this kind of perovskite oxide has been reported to be much lower than the electron conductivity,¹⁰ such that the electrical conductivity described herein can be assumed to represent primarily electronic conductivity. As shown in Fig. 6, all samples exhibit an initial increase in electrical conductivity from 300 to 425–450 °C, followed by a downward trend in higher temperatures. This results from thermally-induced lattice oxygen release and the formation of oxygen vacancies that hinders the mobility of electronic conduction carriers. Maximum electrical conductivities of 36 and 32 S cm⁻¹ were obtained for PrBaFe₂O_5+δ and PrBa_0.97Fe₂O_5+δ, respectively. These values are expected to be acceptable for SOFC applications since they are similar to those of previously reported cathode materials such as Ba_0.5Sr_0.5Co_0.8Fe_0.2O_3−δ,^35,36 Ba_0.5Sr_0.5Fe_0.8Cu_0.2O_3−δ,³⁷ and Sm_0.5Sr_0.5Fe_0.8Cu_0.2O_3−δ,³⁸ which have comparable electrical conductivities and exhibited good electrochemical performance. As depicted in Fig. 6, PrBa_0.97Fe₂O_5+δ exhibiter a lower conductivity than the parent material within the measured temperature range, which suggests that the introduction of a Ba deficiency results in a decrease in electron conductivity in the cobalt-free double perovskite. A similar trend was observed for the cobalt-based double perovskites LaBa_1−xCo₂O_5+δ²² and cubic perovskites La_0.6Sr_0.4−xCo_0.2Fe_0.8O_3−δ (x = 0.0–0.2),¹⁰ which is ascribed to the formation of additional oxygen vacancies, allows them to act as an electron-capture defect, and affects electron conductivity negatively, with higher Ba/Sr deficiency. Generally, both additional holes and oxygen vacancies will form simultaneously in the mixed ion-electron conducting perovskites at higher A-site cationic deficiencies as charge compensation mechanisms, both of which oppositely affect the electrical conductivity. As indicated by the previously studied oxygen content results, introduction of Ba deficiency will result in an obvious increase in oxygen vacancy concentration for PrBa_1−xFe₂O_5+δ (x = 0.00–0.03) oxides. Therefore, the presently observed decrease in electrical conductivity with higher x from 0.00 to 0.03 properly related to the formation of oxygen vacancies which is the main charge compensation mechanism for the Ba deficiency.

Fig. 6 Electrical conductivity of PrBa_1−xFe₂O_5+δ (x = 0.00–0.03) oxides measured in air.

3.5 Thermal expansion behavior

The thermal compatibility between cathode material and other components is another important factor that relates to the working stability of SOFCs, since a large difference in TEC between these components will introduce interfacial stress during thermal cycling and result in cracks and cell degradation. Therefore, the thermal expansion behavior of PrBaFe₂O_5+δ and PrBa_0.97Fe₂O_5+δ was characterized by dilatometry measurements over 30–900 °C. The obtained thermal expansion curves and calculated TEC values are presented in Fig. 7. Both samples exhibit a gradual expansion with higher temperature that can be ascribed to nonharmonic atomic vibrations, dependent on electrostatic attraction forces within the lattice, and chemical expansion induced by the thermal reduction of iron ions to the lower valence state.^21,26 An average TEC value as low as 16.1 × 10⁻⁶ °C⁻¹ was obtained for the PrBaFe₂O_5+δ oxide from 30–900 °C, which is comparable to that reported by Zhao et al.¹³ and is much lower than that of the corresponding cobalt-based double perovskite PrBaCo₂O_5+δ (23.4 × 10⁻⁶ °C⁻¹).⁸ This suggests that complete substitution of cobalt by iron can decrease the thermal mismatch between the double perovskite and GDC electrolyte significantly. A Ba deficiency results in a smaller TEC (15.2 × 10⁻⁶ °C⁻¹) compared with the parent sample. Similar trends have been reported by Zhou et al.²⁶ for (Ba_0.5Sr_0.5)_1−xCo_0.8Fe_0.2O_3−δ and by Hansen et al.³⁹ for (La_0.6Sr_0.4)_1−xFe_0.8Co_0.2O_3−δ perovskites, and may be ascribed to the increased electrostatic attractions from lattice shrinkage that accompanies higher levels of Ba deficiency (Fig. 1).

Fig. 7 Thermal expansion curves of PrBa_1−xFe₂O_5+δ (x = 0.00–0.03) oxides measured in air.

3.6 Chemical compatibility with GDC electrolyte

The chemical compatibility between cathode and various electrolytes is key with regards the long-term working stability of SOFCs, since these two components may react to form insulating phases at elevated temperatures.¹³ Therefore, chemical compatibilities between double perovskites and the GDC electrolyte were characterized by mixing materials at a 1 : 1 mass ratio followed by calcined at 1000 °C for 2 h. The resulting mixtures were characterized by XRD to assess any reactivity. As an example of a typical result, the XRD pattern of a calcined PrBa_0.97Fe₂O_5+δ–GDC powder is presented in Fig. 8 with those of pure PrBa_0.97Fe₂O_5+δ and GDC powders provided for comparison. The diffraction peaks that result from the PrBa_0.97Fe₂O_5+δ–GDC mixture can be indexed to a simple combination of PrBa_0.97Fe₂O_5+δ and GDC without any evidence of impurities and with no noticeable peak shifts, which suggests that the PrBa_1−xFe₂O_5+δ oxides are chemically compatible with the GDC electrolyte at 1000 °C.

Fig. 8 XRD patterns of as-synthesized PrBa_0.97Fe₂O_5+δ, commercial GDC, and PrBa_0.97Fe₂O_5+δ–GDC mixtures after being heat-treated at 1000 °C for 2 h.

3.7 Polarization resistance

The electrochemical activity of the PrBa_1−xFe₂O_5+δ (x = 0.00–0.03) oxides were characterized via EIS measurements with the symmetrical cell configuration PrBa_1−xFe₂O_5+δ/GDC/PrBa_1−xFe₂O_5+δ. Because the measured area specific resistances (ASRs) are microstructure-sensitive, the structural properties of the cells after EIS testing were characterized by SEM and BET adsorption–desorption isotherms of N₂ technology. Similar microstructures were observed for all the samples and a typical PrBa_0.97Fe₂O_5+δ/GDC interface is presented in Fig. 9a. The cathode layer, with an average thickness of approximately 16 μm, possesses a porous microstructure with uniformly distributed micro-to-nanometer sized pores that favor gas diffusion through the cathode. Good binding between the PrBa_0.97Fe₂O_5+δ cathode and the GDC electrolyte is also observed in this image, which is an important factor with regard to avoiding interfacial polarization resulting from poor interface contact. Similar specific surface area and pore size distribution were observed for both cathode layers (Fig. 9b). These features suggest the reliability for directly comparison of electrochemical activity of both samples through the presently obtained ASR values.

Fig. 9 (a) Typical cross-sectional SEM image of the PrBa_0.97Fe₂O_5+δ/GDC interface calcined at 1000 °C for 2 h in air. (b) Pore size distribution curves for both samples.

To assess the effect of Ba deficiency on the electrochemical performance of cobalt-free double perovskites, impedance spectra of the PrBa_1−xFe₂O_5+δ (x = 0.00–0.03) oxides were acquired at 650 and 700 °C, as shown in Fig. 10. In these spectra, the ohmic resistances arising from the GDC electrolyte and lead wires have been normalized to zero for clarity. The introduction of Ba deficiency results in an obvious decrease in area specific resistance (ASR) for the layered oxides. For example, an ASR value as low as 0.119 Ω cm² was obtained at 700 °C for PrBa_0.97Fe₂O_5+δ, which is approximately 12% lower than that of the parent PrBaFe₂O_5+δ (0.136 Ω cm² at 700 °C), and indicates that the A-site Ba deficiency is an effective strategy to enhance the electrochemical performance of the cobalt-free double perovskites PrBa_1−xFe₂O_5+δ. These values are lower than those reported previously for cobalt-free cathode materials, including SmBaFe₂O_5+δ (0.196 Ω cm² at 700 °C),¹⁵ La_0.6Sr_0.4Ni_0.2Fe_0.8O_3−δ (0.5 Ω cm² at 700 °C),⁴⁰ and Ba_0.5Sr_0.5Zn_0.2Fe_0.8O_3−δ (0.23 Ω cm² at 700 °C),⁴¹ and are comparable or higher than those of cobalt-based perovskite cathodes such as PrBaCo_2/3Fe_2/3Cu_2/3O_5+δ (0.144 Ω cm² at 700 °C),⁴² PrBaCo_1.6Fe_0.4O_5+δ (0.13 Ω cm² at 700 °C),¹⁶ and GdBaCo_2/3Fe_2/3Cu_2/3O_5+δ (0.165 Ω cm² at 700 °C).⁴³ The formation and redistribution of oxygen vacancies are two possible reasons for the reduction in ASR. In terms of oxygen vacancy formation, both the oxygen content (Fig. 5) and electrical results (Fig. 6) obtained for the PrBa_1−xFe₂O_5+δ oxides indicate that the formation of oxygen vacancies is the main charge compensation mechanism associated with A-site Ba deficiency with x values from 0.00 to 0.03. The presence of additional oxygen vacancies has been reported to be favorable for the ORR at the surface of the mixed ion-electronic conducting PrBa_1−xFe₂O_5+δ and at the PrBa_1−xFe₂O_5+δ/GDC interfaces, and also promotes oxygen ion diffusion through the cathode.³⁴ The decrease in ASR values with A-site cation deficiency is therefore understandable. In terms of oxygen vacancy redistribution, the XRD patterns in Fig. 1 show that PrBaFe₂O_5+δ has a similar layered structure to that of the cobalt-based double perovskites LnBaCo₂O_5+δ (Ln = lanthanide), in which the A-site cations Lnⁿ⁺/Ba²⁺ are ordered along the c axis and oxygen vacancies are restricted to the LnO and CoO layers. This structure will exhibit fast oxygen ion diffusion along the ab plane, whereas diffusion along the c axis will be forbidden.^44,45 Ba deficiencies in the cobalt-free double perovskites PrBa_1−xFe₂O_5+δ are expected to function in a manner similar to that proposed in our previous work on cobalt-based double perovskites LnBa_1−xCo₂O_5+δ (Ln = Pr and La).^8,22 The deficiency breaks the Ba–O bonds, creates helpful additional oxygen vacancies as charge carriers for oxygen ions in the Ba–O layers, increases the possibility of oxygen ion diffusion along the c axis, shortens the oxygen diffusion length through the cathode, and results in a decrease in ASR. This hypothesis requires further confirmation by theoretical simulation and experimental measurements.

Fig. 10 Impedance spectra of PrBa_1−xFe₂O_5+δ/GDC/PrBa_1−xFe₂O_5+δ symmetrical cells measured at 650 °C (a) and 700 °C (b) in air. Numbers in these plots correspond to logarithm of frequency.

3.8 Single cell performance

Fig. 11 presents the I–V and I–P curves of the PrBa_1−xFe₂O_5+δ based anode support single cells with the configuration of Ni-YSZ/YSZ/GDC/PrBa_1−xFe₂O_5+δ. The single cells were fabricated and tested with the same procedure to ensure the comparability of the presently obtained electrochemical performance data. The PrBa_0.97Fe₂O_5+δ based single cell exhibits the maximum power densities of 329, 256, 183 and 118 mW cm⁻² at 800, 750, 700 and 650 °C, respectively. These values are much higher than that of PrBaFe₂O_5+δ based single cell which indicate that the introduction of proper amount of Ba deficient can efficiently enhance the electrochemical performance of the cobalt-free double perovskite PrBa_1−xFe₂O_5+δ. PrBa_0.97Fe₂O_5+δ is a potential cathode material for IT-SOFCs.

Fig. 11 I–V and I–P curves of the (a) PrBaFe₂O_5+δ and (b) PrBa_0.97Fe₂O_5+δ based single cells.

4. Conclusions

Cobalt-free double perovskite PrBa_1−xFe₂O_5+δ (x = 0.00–0.10) were synthesized and evaluated as a cathode material for IT-SOFCs with respect to the crystal structure, surface properties, oxygen content, electrical conductivity, thermal expansion, chemical compatibility with GDC electrolyte, microstructure, and electrochemical performance. A layered structure with c approximately equal to 2a ≈ 2b and gradual shrinkage in the lattice with higher Ba deficiency (x) was observed for PrBa_1−xFe₂O_5+δ oxides. Minor Pr₆O₁₁ phase was detected for samples for which x ≥ 0.05. Ba deficiency has little influence on the chemical state of O and Pr, but results in a slight reduction in transition metal (Fe) at surfaces of the sample. Iodometric titration and TGA results show that Ba-deficient PrBa_0.97Fe₂O_5+δ exhibits a lower oxygen content than the parent PrBaFe₂O_5+δ at given temperature. An obvious decrease in electrical conductivity was obtained with higher x from 300–800 °C, which indicates that the formation of additional oxygen vacancies was the main charge compensation mechanism for Ba deficiency in the cobalt-free double perovskite PrBa_1−xFe₂O_5+δ. A slight decrease in TEC results from a Ba deficiency from 30–900 °C. The catalytic activity for the ORR of the cobalt-free double perovskites can be enhanced by introducing the appropriate amount of Ba deficiency, and is characterized by a decreased ASR compared with the parent oxide. ASR values as low as 0.119 Ω cm² were obtained for PrBa_0.97Fe₂O_5+δ at 700 °C. The PrBa_0.97Fe₂O_5+δ based single cell exhibits the maximum power densities of 329 and 183 mW cm⁻² at 800 and 700 °C, respectively. These results suggest PrBa_0.97Fe₂O_5+δ is a potential cathode material for IT-SOFCs.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (51402127, 51474037, 51474113), the Natural Science Foundation of Jiangsu Province (SBK201341159), the China Postdoctoral Science Foundation (2014M550269, 2015T80524) and the Senior Talent Foundation of Jiangsu University (13JDG009).

References

1. E. D. Wachsman and K. T. Lee, Science, 2011, 334, 935–939.
2. B. C. H. Steel and A. Heinzel, Nature, 2001, 414, 345–352.
3. A. A. Taskin, A. N. Lavrov and Y. Ando, Appl. Phys. Lett., 2005, 86, 091910.
4. A. A. Taskin, A. N. Lavrov and Y. Ando, Appl. Phys. Lett., 2005, 86, 134414.
5. A. Tarancon, M. Burriel, J. Santiso, S. J. Skinner and J. A. Kilner, J. Mater. Chem., 2010, 20, 3799–3813.
6. J.-H. Kim and A. Manthiram, J. Electrochem. Soc., 2008, 155, B385–B390.
7. E. Chavez, M. Mueller, L. Mogni and A. Caneiro, J. Phys.: Conf. Ser., 2009, 167, 012043.
8. S. L. Pang, X. N. Jiang, X. N. Li, Q. Wang and Z. X. Su, J. Power Sources, 2012, 204, 53–59.
9. S. L. Pang, X. N. Jiang, X. N. Li, Z. X. Su, H. X. Xu and Q. L. Xu, et al., Int. J. Hydrogen Energy, 2012, 37, 6836–6843.
10. K. Zhang, L. Ge, R. Ran, Z. P. Shao and S. M. Liu, Acta Mater., 2008, 56, 4876–4889.
11. Y. J. Niu, W. Zhou, J. Sunarso, L. Ge, Z. H. Zhu and Z. P. Shao, J. Mater. Chem., 2010, 20, 9619–9622.
12. Q. J. Zhou, T. Wei, S. Q. Guo, X. L. Qi, R. F. Ruan and Y. Li, et al., Ceram. Int., 2012, 38, 2899–2903.
13. L. Zhao, J. C. Shen, B. B. He, F. L. Chen and C. R. Xia, Int. J. Hydrogen Energy, 2011, 36, 3658–3665.
14. X. J. Xue and H. P. Ding, J. Power Sources, 2010, 195, 7038–7041.
15. D. J. Chen, F. C. Wang, H. G. Shi, R. Ran and Z. P. Shao, Electrochim. Acta, 2012, 78, 466–474.
16. J. Zou, J. Park, B. Kwak, H. Yoon and J. Chung, Solid State Ionics, 2012, 206, 112–119.
17. Y. N. Kim, J.-H. Kim and A. Manthiram, J. Power Sources, 2010, 195, 6411–6619.
18. X. J. Xue and H. P. Ding, J. Power Sources, 2010, 195, 4718–4721.
19. X. J. Xue and H. P. Ding, J. Power Sources, 2010, 195, 4139–4142.
20. Z. Liu, L. Z. Cheng and M. F. Han, J. Power Sources, 2011, 196, 868–871.
21. Z. B. Yang, M. F. Han, P. Y. Zhu, F. Zhao and F. L. Chen, Int. J. Hydrogen Energy, 2011, 36, 9162–9168.
22. S. L. Pang, X. N. Jiang, X. N. Li, H. X. Xu, L. Jiang and Q. L. Xu, et al., J. Power Sources, 2013, 240, 54–59.
23. S. L. Pang, X. N. Jiang, X. N. Li, Q. Wang, Z. X. Su and Q. Y. Zhang, Int. J. Hydrogen Energy, 2012, 37, 3998–4001.
24. F. F. Dong, M. Ni, Y. B. Chen, D. J. Chen, M. O. Tadé and Z. P. Shao, J. Mater. Chem. A, 2014, 2, 20520–20529.
25. J. P. Wang, F. C. Meng, T. Xia, Z. Shi, J. Lian and C. B. Xu, et al., Int. J. Hydrogen Energy, 2014, 39, 18392–18404.
26. W. Zhou, Z. P. Shao, W. Q. Jin and N. P. Xu, J. Power Sources, 2008, 182, 24–31.
27. E.-L. Rautama, P. Boullay, A. K. Kundu, V. Caignaert, V. Pralong and M. Karppinen, et al., J. Mater. Chem., 2008, 20, 2742–2750.
28. S. L. Pang, X. N. Jiang, X. N. Li, Q. Wang and Q. Y. Zhang, Mater. Chem. Phys., 2012, 131, 642–646.
29. M. Ghaffari, M. Shannon, H. Hui, O. K. Tan and A. Irannejad, Surf. Sci., 2012, 606, 670–677.
30. F. J. Jin, Y. Shen, R. Wang and T. M. He, J. Power Sources, 2013, 234, 244–251.
31. H. Falcόn, J. A. Barbero, G. Araujo, M. T. Casais, M. J. Martínez-Lope and J. A. Alonso, et al., Appl. Catal., B, 2004, 53, 37–45.
32. L. L. Zhang, Q. J. Zhou, Q. He and T. M. He, J. Power Sources, 2001, 195, 6356–6366.
33. F. J. Jin, H. W. Xu, W. Long, Y. Shen and T. M. He, J. Power Sources, 2013, 243, 10–18.
34. S. B. Adler, Chem. Rev., 2004, 104, 4791–4843.
35. Z. P. Shao and S. M. Haile, Nature, 2004, 431, 170–173.
36. S. L. Pang, X. N. Jiang and X. N. Li, Ferroelectrics, 2015, 478, 26–39.
37. L. Zhao, B. B. He, X. Z. Zhang, R. R. Peng, G. Y. Meng and X. Q. Liu, J. Power Sources, 2010, 195, 1859–1861.
38. Y. H. Ling, L. Zhao, B. Lin, Y. C. Dong, X. Z. Zhang and G. Y. Meng, et al., Int. J. Hydrogen Energy, 2010, 35, 6905–6910.
39. K. K. Hansen and K. V. Hansen, Solid State Ionics, 2007, 178, 1379–1384.
40. G. Y. Zhu, X. H. Fang, C. R. Xia and X. Q. Liu, Ceram. Int., 2005, 31, 115–119.
41. B. Wei, Z. Lu, X. Q. Huang, M. L. Liu, N. Li and W. H. Su, J. Power Sources, 2008, 176, 1–8.
42. F. J. Jin, Y. Shen, R. Wang and T. M. He, J. Power Sources, 2013, 234, 244–251.
43. S. H. Jo, P. Muralidharan and D. K. Kim, Electrochem. Commun., 2009, 11, 2085–2088.
44. D. Parfitt, A. Chroneos, A. Tarancón and J. A. Kilner, J. Mater. Chem., 2011, 21, 2183–2186.
45. M. Burriel, J. P. Martínez, R. J. Chater, S. Fearn, A. V. Berenov and S. J. Skinner, et al., Chem. Mater., 2012, 24, 613–621.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra19555a

---