Structure and electrochemical properties of Sm0.5Sr0.5Co1 − xFexO3 − δ cathodes for solid oxide fuel cells

Structure and electrochemical properties of Sm0.5Sr0.5Co1 − xFexO3 − δ cathodes for solid oxide fuel cells

Solid State Ionics 177 (2006) 901 – 906 www.elsevier.com/locate/ssi Structure and electrochemical properties of Sm0.5Sr0.5Co1 − xFexO3 − δ cathodes f...

262KB Sizes 0 Downloads 28 Views

Solid State Ionics 177 (2006) 901 – 906 www.elsevier.com/locate/ssi

Structure and electrochemical properties of Sm0.5Sr0.5Co1 − xFexO3 − δ cathodes for solid oxide fuel cells Hong Lv ⁎, Yu-ji Wu, Bo Huang, Bin-yuan Zhao, Ke-ao Hu State Key Laboratory of Metal Matrix Composites, Shanghai Jiaotong University, Shanghai 200030, P. R. China Received 23 October 2005; received in revised form 14 January 2006; accepted 21 January 2006

Abstract Crystal structure, thermal expansion coefficient, electrical conductivity and cathodic polarization of compositions in the system Sm0.5Sr0.5 Co1 − xFexO3 − δ with 0 ≤ x ≤ 0.9 were studied as function of Co / Fe ratio and temperature, in air. Two phases, including an Orthorhombic symmetry for 0 ≤ x ≤ 0.4 and a cubic symmetry for 0.5 ≤ x ≤ 0.9, were observed in samples of Sm0.5Sr0.5Co1 − xFexO3 − δ at room temperature. The adjustment of thermal expansion coefficient (TEC) to electrolyte, which is one of the main problems of SSC, could be achieved to lower TEC values with more Fe substitution. High electrical conductivity above 100S/cm at 800°C was obtained for all specimens, so they could be good conductors as cathodes of IT-SOFC. The polarization behavior of SSCF as a function of Fe content was evaluated by means of AC impedance using LSGM electrolyte. It was discovered that the Area Specific Resistance (ASR) of SSCF increased as the amount of substitution of Fe for Co increased. When the amount of Fe reached to 0.4, the highest ASR was obtained and then the resistance started decreasing above that. The electrode with a composition of Sm0.5Sr0.5Co0.2Fe0.8O3 − δ showed high catalytic activity for oxygen reduction operating at temperature ranging from 700 to 800°C. © 2006 Elsevier B.V. All rights reserved. Keywords: Solid oxide fuel cell; Cathode; Iron-doped Sm0.5Sr0.5CoO3 − δ; Polarization resistance

1. Introduction Solid oxide fuel cell (SOFC) is an all solid device that converts the chemical energy of gaseous such as hydrogen and natural gas to electricity through electrochemical processes. SOFC, being an electrochemical device, has unique advantages over the traditional power generation technologies. SOFCs combine the benefits of environmentally benign power generation with fuel flexibility. However, the necessity for high operating temperatures (900–1000°C) results in high costs and materials compatibility challenges [1]. As a consequence, significant effort has been devoted to the development of intermediate-temperature (500–800°C) SOFCs. A key obstacle to reduced-temperature operation of SOFCs is the poor activity of traditional cathode materials for electrochemical reduction of oxygen in this temperature regime. Currently, Sr-doped LaMnO3 is commonly used as the cathode material for the SOFCs operating at high temperatures ⁎ Corresponding author. Tel.: +86 21 62933751; fax: +86 21 62822012. E-mail address: [email protected] (H. Lv). 0167-2738/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.ssi.2006.01.038

(900∼1000 °C) because of its high catalytic activity for oxygen reduction and thermal and chemical compatibility with the YSZ electrolyte at SOFCs operating temperatures [2]. While LSM has shown promising performance for SOFCs operating at temperature above 800 °C, its performance decreases rapidly as the operating temperature decreases, for example, the LSM overpotential at 1000°C is 1 Ω cm− 2 but increases to 2000Ω cm− 2 at 500°C [3]. Sm0.5Sr0.5CoO3 (SSC) is regarded as one of the most promising cathode materials for IT-SOFCs because SSC has shown a higher ionic conductivity than that of LSM, similar performance to La0.6Sr0.4Co0.2Fe0.8O3, exchange parameters higher than LSC and LSM [4,5], and to be particularly compatible with GDC and LSGM [6]. However, the thermal expansion coefficient (TEC) values are very larger than 20 (× 10− 6 °C− 1), which could produce stress at the interface during thermal cycling [11]. Further investigations should be done to improve the compatibility with the electrolyte. S. Balagopal et al. have pointed out that doping in B-site (replacing Co) could decrease the TEC [7]. It is found that the electronic conductivity of LSCF can exceed 300 S/cm at 750 °C [8] and has more compatible TEC than LSC [9,10]. Doping with

902

H. Lv et al. / Solid State Ionics 177 (2006) 901–906

Fig. 1. The scheme of a symmetric two-electrode measurement cell.

Fe into Co-site in SSC maybe lower the TEC, without resulting in a great decrease in electrochemical properties and there is little published information relating the structural, thermal and electrical properties of Sm0.5Sr0.5Co1 − xFexO3 − δ (0 ≤ x ≤ 0.9) (SSCF). In this paper, we investigated the effects of Fe doping of Sm0.5Sr0.5CoO3 − δ (0 ≤ x ≤ 0.9). Their crystal structure, thermal expansion coefficient, electrical conductivity and cathodic polarization were studied from the viewpoint of intermediate temperature SOFCs. 2. Experimental Sm0.5Sr0.5Co1 − xFexO3 − δ (x = 0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9) were prepared using traditional solid-state reaction method. The precursors were Sm2O3 (N99.99%), SrCO3 (N99.9%), Co3O4 (N99%), and Fe2O3 (N 99%). They were calcined in air for 2h at 800°C and 300 °C, respectively, to remove the absorbed moisture. The powders were mixed in absolute alcohol by ball mill for 24 h, then calcined at 1000°C for 10 h followed by repeated grinding and calcining until complete reaction and uniform composition were achieved. The products were ground, pressed into pellets and sintered in air at 1200 °C for 5 h. The phases were identified using X-ray diffraction (XRD, D8 Discover GADDS), diffractometer using monochromated CuKα radiation at the scanning speed of 0.5°/min. Structural parameters were refined from XRD data using least-squares procedure. La0.8Sr0.2Ga0.8Mg0.2O3 (LSGM) using as the electrolyte was prepared by traditional solid-state reaction method. Stoichio-

metric amounts of La2O3 (N99.99% purity), SrCO3 (N 99.9% purity), Ga2O3 (N 99.99% purity) and MgO (N99.99% purity) were intimately mixed in an agate mortar with the aid of absolute alcohol for 24 h and then calcined at 950°C for 10 h. The calcined powders were crushed using agate mortar and pestle and ball milled in absolute alcohol for another 24h. The resulting fine powder was dried and uniaxially pressed into pellet under 100 MPa. After being sintered at 1500°C for 5 h, the pellets were obtained with diameter of 20mm and thickness of 0.6 ± 0.1 mm. The polarization resistance of SSCF cathode was measured in the two-electrode symmetric cell configuration under air [16]. Electrolyte-supported symmetric cells for impedance studies were prepared by screen printing. The slurry of SSCF, which was ground and mixed with isopropyl alcohol, was printed onto both sides of an LSGM electrolyte disk, followed by calcination at 1150°C for 2h under stagnant air. After being sintered, the resulting electrode areas are 1 cm2. Pt mesh (80 meshes) was attached to the electrode surfaces using Pt paste as the current collector. The scheme of electrolytesupported symmetric cell was shown in Fig. 1. The electrochemical activities of the SSCF cathodes were characterized by the electrochemical impedance spectroscopy (EIS), using a solartron 1260 frequency response analyzer at open circuit. The applied frequency was in the range of 10 mHz to 100KHz at five points per frequency decade with the signal amplitude of 20 mV. The EIS was measured in the temperature range from 500 to 800°C increments of 50 °C in the air. The impedance spectra were analyzed by the equivalent circuit of the program Zview. Cross-section of the cathode was observed by scanning electron microscopy (SEM, PHLIPS 515). The electrical conductivity of the Sm0.5Sr0.5Co1 − xFexO3 − δ materials was measured using the standard four-probe DC method. Rods of SSCF were sintered at 1200°C for 5h in air. Rectangular bars with approximate dimensions of 5 × 5 × 20mm3 were obtained from the sintered rods. Pt lead was attached to the rod with Pt paste and fired at 1000°C for 30min to obtain a firm bonding and good electrical contact between the Pt leads and the sample. Measurements were performed from room temperature to 900°C with a heating rate of 5°C/min. Thermal expansion measurement was performed on rectangular specimens (5 × 5 × 20mm3) from room temperature to 900°C with a heating rate of 5 °C/min using quartz as a reference. Measurements of YSZ and LSGM were also carried out for comparisons.

Table 1 The cell parameters of Sm0.5Sr0.5Co1 − xFexO3 − δ ceramics at room temperature SSCF

Crystal structure

a/Å

b/Å

c/Å

Cell volume/Å3

x=0 x = 0.1 x = 0.2 x = 0.3 x = 0.4 x = 0.5 x = 0.6 x = 0.7 x = 0.8 x = 0.9

Orthorhombic Orthorhombic Orthorhombic Orthorhombic Orthorhombic Cubic Cubic Cubic Cubic Cubic

5.3591 ± 0.0093 5.3722 ± 0.0090 5.3972 ± 0.0003 5.4112 ± 0.0024 5.4103 ± 0.0039 3.8267 ± 0.0008 3.8554 ± 0.0023 3.8539 ± 0.0004 3.8547 ± 0.0006 3.8739 ± 0.0009

5.3927 ± 0.0053 5.3885 ± 0.0093 5.3981 ± 0.0003 5.4026 ± 0.0025 5.4113 ± 0.0041

7.5766 ± 0.0107 7.5887 ± 0.0096 7.6283 ± 0.0003 7.6536 ± 0.0026 7.6453 ± 0.0042

218.96 219.68 222.25 223.75 223.83 56.04 57.26 57.24 57.28 58.14

H. Lv et al. / Solid State Ionics 177 (2006) 901–906

903

3. Results and discussion 3.1. Crystal structure and the cell parameters The lattice parameters for Sm0.5Sr0.5Co1 − xFexO3 − δ (0 ≤ x ≤ 0.9, sintered in air at 1200°C for 5 h) are summarized in Table 1. Sm0.5Sr0.5CoO3 has been studied by Tu et al. [11] the parameters of Sm0.5Sr0.5CoO3 (a = 5.367, b = 5.406, c = 7.588) they reported are the same as our results. Two phases are observed in sample of Sm0.5Sr0.5Co1 − xFexO3 − δ at room temperature, the perovskite phase of SSCF has an Orthorhombic symmetry for 0 ≤ x ≤ 0.4 and a cubic symmetry for 0.5 ≤ x ≤ 0.9. It also can be observed that the initial volume of both phases increases at room temperature with increasing iron content, as may result from the relative values of ionic radii of iron and cobalt cations in octahedral coordination. The ion size of Co3+ (r = 0.61 Å) is somewhat smaller than that of Fe3+ (r = 0.645 Å). In this paper, detailed crystal information has not been obtained, but the characteristics pertaining to use as an electrode for SOFC are the main focus. 3.2. Thermal expansion coefficient Thermal expansion curves in Fig. 2 show gradual increases in the high temperature region for most compositions, but they deviate from linearity as indicated by the plots of YSZ and LSGM. The observed abnormal expansion in samples at high temperatures is due to the loss of lattice oxygen [12]. Thermal expansion coefficients (TEC's) calculated from these curves in Fig. 2 are plotted as function of Fe content in Fig. 3. The TEC for compositions has been found to decrease with increasing Fe content. The adjustment of thermal expansion rate to electrolyte, which is one of the main problems of SSC, could be achieved to lower TEC values with more Fe substitution. At x = 0–0.3 composition, the TEC does not show remarkable decrease and it displays decreasing obviously from 0.4 to 0.9. Thermal compatibility with the electrolytes has been improved

Fig. 2. Linear thermal expansion curves for Sm0.5Sr0.5Co1 − xFexO3 − δ.

Fig. 3. Thermal expansion coefficient of Sm0.5Sr0.5Co1 − xFexO3 − δ.

at x ≥ 0.8 evidently. The same phenomena have been observed in La0.8Sr0.2Co1 − yFeyO3 and Nd0.7Sr0.3Co1 − yFeyO3 − δ as Fe solution content increases [12,14]. L.-W. Tai et al. explained that the lattice expansion associated with the formation of oxygen vacancies: (1) the repulsion force arising between those mutually exposed cation when oxygen ions are extracted from the lattice; (2) the increase in cation size due to the reduction of the Fe and Co ions from higher to lower valences, which must occur concurrently with the creation of oxygen vacancies in order to maintain electrical neutrality. Therefore, the loss of lattice oxygen decreased with increasing Fe content, which leads to the decrease of TEC [12,13]. 3.3. Electrical conductivity Fig. 4 shows the electrical conductivity plots (logσ) versus inverse temperature measured from room temperature to

Fig. 4. Temperature dependence of the electrical conductivity for Sm0.5Sr0.5 Co1 − xFexO3 − δ.

904

H. Lv et al. / Solid State Ionics 177 (2006) 901–906

Fig. 5. Composition dependence of the electrical conductivity for Sm0.5Sr0.5 Co1 − xFexO3 − δ at 800°C.

900 °C, the electrical conductivities of SSCF decrease as the Fe content increases from x = 0 to x = 0.9 (except x = 0.6). In Sm0.5Sr0.5Co1 − xFexO3 − δ systems, most of the samples show the semi-conducting behavior and the electrical conductivity of each composition increases with temperature through a maximum, then decreases. The same phenomena have been observed in La0.8Sr0.2Co1 − yFeyO3, the p-type electrical conductivity of LSCF decreases with Fe content because oxygen vacancies formed in these oxides at high temperatures resulted in a reduction in the concentration of electronic charge carriers [12]. The observed unique temperature dependence of electrical conductivity was attributed to several factors including: charge disproportionation of Co ions, ionic compensation by the formation of oxygen vacancies at higher temperatures, and preferential electronic compensation to form Fe4+ rather than Co4+. The composition dependence of electrical conductivity at 800 °C is shown in Fig. 5. High electrical conductivity has been obtained for all specimens. Although conductivity decreases with the enhancement of Fe content from the order value of 3 (σ = 103 S/cm, x = 0) to 2(σ = 102 S/cm, x = 0.9), all the values are higher than LSM. So they can be better conductors as cathodes of IT-SOFC than LSM.

overall ohmic resistances including the electrolyte resistance, electrode ohmic resistance, lead resistance and contact resistance between cell and Pt mesh. L is the inductance, which could be due to the platinum current/voltage probes or the high frequency phase shift of electrochemical equipment. At the low frequency phase, L has little effect on impedance spectra. In the temperature range over 400°C, electrolyte impedance that appeared in high-frequency domain could not be observed with the exception of resistor part, so the difference between realaxes intercepts of the impedance plot is considered to be the electrode polarization resistance (Rp). Because the electrical conductivity of SSCF shows high values (σ ≥ 100S/cm) and the electrode average thickness is 15 ± 1.5μm, the electrode ohmic resistances are so small as to be omitted. In Fig. 7, the real-axes value which meets the left side of semi-circle corresponds to Rs of the equivalent circuit in Fig. 8. Fig. 7 shows the difference of the Rs at the same temperature, which mainly attributes to the different thickness of electrolyte disk, the different length of the lead and the different contact resistance between cell and Pt mesh. From Fig. 7, it can be known that the polarization resistance decreases between the electrolyte and electrode as the temperature increases. Fig. 9 shows the Area Specific Resistance (ASR) at 800°C according to the change of Fe doping mole fraction. Rp of Sm0.5Sr0.5 Co1 − xFexO3 − δ is the fitting result of equivalent circuit by Zview software from EIS (Fig. 7). ASR is one half of the polarization resistance and corrected for area since the impedance is measured on symmetric cells [15]. As the amount of substitution of Fe for Co increases, the ASR increases. When the amount of Fe reaches to 0.4, the highest ASR is obtained and then the resistance starts decreasing above that. This tendency exists at various temperatures, which can be shown in Fig. 10. Higher ASRs are found due to the lower catalytic activity for oxygen reduction. LSCF systems indicate that cobalt makes an important contribution to the electrochemical reaction with cathode. The substitution of Fe on the Co-site makes the overpotential increase in the composition range from 0 to 1 in

3.4. AC impedance measurement The microstructure of SSC/LSGM has been shown in Fig. 6. SSC sintered in 1150°C for 2 h shows a structure with a reasonable porosity and well-necked particles. LSGM sintered in 1500 °C for 5h shows a densified structure. The interface combines well between electrode and electrolyte. Electrode thickness was determined from SEM micrograph; the average thickness of SSC is 15 ± 1.5μm. The electrode thickness of SSCF with different Fe content is also the same with SSC. The typical impedance spectroscopies for SSCF cathode at various temperatures (at 550, 650, 750, 800°C) are shown in Fig. 7, the impedance spectra are evaluated by fitting impedance data with the equivalent circuit shown in Fig. 8, where Rs is the

Fig. 6. SEM photograph of cross-section of SSC/LSGM.

H. Lv et al. / Solid State Ionics 177 (2006) 901–906

905

Fig. 7. Impedance spectroscopy for Sm0.5Sr0.5Co1 − xFexO3 − δ (x = 0, 0.2, 0.6, 0.8) cathode in the air at various temperature.

Ln1 − xSrxCo1 − yFeyO3 − δ (Ln = Pr, Nd, Gd; x = 0.2, 0.3) [14], which is different from the result of our investigation. Ln1 − x SrxCo1 − yFeyO3 − δ have a single-phase Orthorhombic perovskite structure as Fe solution content increase, but two phases

Fig. 8. Equivalent circuit of the solid oxide with Fe doping mole fraction at 800°C.

Fig. 9. Area specific resistance of SSCF/LSGM electrolyte/electrode system.

906

H. Lv et al. / Solid State Ionics 177 (2006) 901–906

conductors as cathodes of IT-SOFC. Polarization resistance of SSCF was obtained by AC impedance. As the amount of substitution of Fe for Co increased, the ASR increased. When the amount of Fe reached to 0.4, the highest ASR was obtained and then the resistance started decreasing above that. The Sm0.5Sr0.5Co0.2Fe0.8O3 − δ electrode showed high catalytic activity for oxygen reduction operating at temperature from 700 to 800 °C. Acknowledgements This work is supported by Instrumental analysis center of Shanghai Jiaotong University. References Fig. 10. Temperature dependence of the polarization for SSCF measured over a temperature range of 500–800 °C in air.

are observed in sample of Sm0.5Sr0.5Co1 − xFexO3 − δ at room temperature, the perovskite phase of SSCF has an Orthorhombic symmetry for 0 ≤ x ≤ 0.4 and a cubic symmetry for 0.5 ≤ x ≤ 0.9. The change of crystal structure in SSCF systems may lead to the difference, which catalytic activity changes as substitution of Fe on the Co-site increases, between SSCF and LSCF. Meanwhile, the sintering activity of the materials is not the same when the composition is different, and has a strong influence on the electrode microstructure, which also influences the performance of the cathode. However, these causes are still not well understood and further study is necessary. 4. Conclusion Sm0.5Sr0.5Co1 − xFexO3 − δ (0 ≤ x ≤ 0.9) (SSCF) were studied for their crystal structure, thermal expansion rate, electrical conductivity and cathodic polarization. Two phases were observed in sample of Sm0.5Sr0.5Co1 − xFexO3 − δ at room temperature, the perovskite phase of SSCF has an Orthorhombic symmetry for 0 ≤ x ≤ 0.4 and a cubic symmetry for 0.5 ≤ x ≤ 0.9. The adjustment of thermal expansion rate to electrolyte, which is one of the main problems of SSC, could be achieved to lower TEC values with more Fe substitution. High electrical conductivity was obtained for all specimens and they demonstrated above 100 S/cm at 800 °C, so they could be good

[1] N.P. Brandon, S. Skinner, B.C.H. Steele, Annu. Rev. Mater. Res. 33 (2003) 183. [2] S.C. Singhal, Solid State Ionics 135 (2000) 305. [3] V. Doshj, Rajiv Richards, J. Carter, X. Wang, M. Krumpelt, J. Electrochem. Soc. 146 (1999) 1273. [4] H. Fukunaga, M. Koyama, N. Takahashi, C. Wen, K. Yamada, Solid State Ionics 132 (2000) 279. [5] S. Wang, T. Kato, T. Honda, T. Kancko, N. Iwashita, SOFC VII, vol. 607, 2001. [6] T. Ishihara, T. Shibayama, M. Honda, H. Nishiguchi, Y. Takita, Mater. Res. Soc. Symp. Proc. 575 (2000) 238. [7] S. Balagopal, I. Bay, J. Hartvigsen, S. Elangovan, Intermediate Temperature Solid Oxide Fuel Cell Development-SECA Core Technology Program Review Meeting, Advance Materials and Electrochemical Technologies, 2002. [8] L.-W. Tai, M.M. Nasrallah, H.U. Anderson, in: S.C. Singhal, H. Iwahara (Eds.), Proc. 3rd Int. Symp. Solid Oxide Fuel Cells, the Electrochemical Society Proceedings Series, Gordon and Breach Science Publishers, Pennington, NJ, 1999, p. 241. [9] J.M. Ralph, J.A. Kilner, B.C.H Steele, Mater. Res. Soc. Symp. Proc. 575 (2001) 309. [10] S. Wang, T. Kato, S. Nagata, T. Honda, T. Kaneko, N. Iwashita, M. Dokiya, Solid State Ionics 146 (2002) 203. [11] H.Y. Tu, Y. Takeda, N. Imanishi, O. Yamamoto, Solid State Ionics 100 (1997) 283. [12] L.-W. Tai, M.M. Nasrallah, H.U. Anderson, et al., Solid State Ionics 76 (1995) 259. [13] L.-W. Tai, M.M. Nasrallah, H.U. Anderson, et al., Solid State Ionics 76 (1995) 273. [14] L. Qiu, T. Ichikawa, A. Hirano, N. Imanishi, Y. Takeda, Solid State Ionics 158 (2003) 55. [15] M. Guillodo, P. Vernoux, J. Fouletier, Solid State Ionics 127 (2000) 99. [16] Z.P. Shao, S.M. Haile, Nature 431 (2004) 170.