Cathodic behavior of La0.8Sr0.2Co1 − xMnxO3 − δ perovskite oxide on YSZ electrolyte for intermediate temperature-operating solid oxide fuel cells

Cathodic behavior of La0.8Sr0.2Co1 − xMnxO3 − δ perovskite oxide on YSZ electrolyte for intermediate temperature-operating solid oxide fuel cells

Available online at www.sciencedirect.com Solid State Ionics 179 (2008) 1465 – 1469 www.elsevier.com/locate/ssi Cathodic behavior of La0.8Sr0.2Co1 −...

664KB Sizes 1 Downloads 94 Views

Available online at www.sciencedirect.com

Solid State Ionics 179 (2008) 1465 – 1469 www.elsevier.com/locate/ssi

Cathodic behavior of La0.8Sr0.2Co1 − xMnxO3 − δ perovskite oxide on YSZ electrolyte for intermediate temperature-operating solid oxide fuel cells Changbo Lee, Seung-Wook Baek, Joongmyeon Bae ⁎ Department of Mechanical Engineering, Korea Advanced Institute of Science and Technology, 373-1 Guseong-Dong, Yuseong-Gu, Daejeon 305-701, Republic of Korea Received 15 July 2007; received in revised form 4 January 2008; accepted 14 January 2008

Abstract A study of the cathode properties of an intermediate temperature-operating solid oxide fuel cell (IT-SOFC) was carried out using La0.8Sr0.2Co1 − x MnxO3 − δ (LSCM). LSCM-8246 had the lowest impedance on the 8 mol% yttria-stabilized zirconia (YSZ) electrolyte among the various compositions investigated in this study. A Ce0.9Gd0.1O2 − δ (CGO) interlayer between LSCM and YSZ was introduced to inhibit the formation of resistive phases. Cathode impedance was decreased by approximately three times compared to that shown with non-interlayer samples. Investigation of the oxygen partial pressure dependence of LSCM-8246 on CGO-layered YSZ showed that the adsorbed oxygen ionization is the rate-determining step of the cathode reaction. The cathode material and structure can be applied to IT-SOFC at 700 °C because the area-specific resistance is sufficiently low, leading to a cell power density higher than 0.5 W/cm2 at 0.7 V. © 2008 Elsevier B.V. All rights reserved. Keywords: Solid oxide fuel cell; Cathode; Oxygen reduction reaction; Surface exchange; Diffusion

1. Introduction The overpotential of cathodes in solid oxide fuel cells (SOFC) occupies a considerable proportion of the overall voltage loss. The polarization loss strongly depends on the kinetics of the oxygen reduction reaction (Eq. (1)) at the triple phase boundaries (TPB) of the gas–electrode–electrolyte. TPB needs to be enhanced in terms of its material composition and microstructure for fuel cells capable of high performance levels. 1 O2 þ 2e þ VO•• YOxO 2

ð1Þ

The La1 − ySryMnO3 − δ (LSM) cathode has been used as a typical cathode material on the 8 mol% Y2O3-stabilized ZrO2 ⁎ Corresponding author. Postal address: Department of Mechanical Engineering, Korea Advanced Institute of Science and Technology, 373-1 GuseongDong, Yuseong-Gu, Daejeon 305-701, Republic of Korea. Tel.: +82 42 869 3045; fax: +82 42 869 8207. E-mail addresses: [email protected] (C. Lee), [email protected] (S.-W. Baek), [email protected] (J. Bae). 0167-2738/$ - see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.ssi.2008.01.009

(YSZ) electrolyte as this material has low chemical reactivity and good thermal expansion coefficient (TEC) compatibility with the YSZ electrolyte. However, lowering of the operating temperature of SOFC below 700 °C increases the overpotential significantly with LSM due to its relatively low reaction rate of oxygen reduction in that temperature range. On the other hand, the La1 − ySryCoO3 − δ (LSC) cathode, known as a mixed ionicelectronic conductor (MIEC), provides a higher reaction rate of oxygen reduction with an enlargement of the reaction sites. However, undesirable resistive phases such as La2Zr2O7 or SrZrO3 easily form at the interface between the cathode and the electrolyte when the LSC cathode is directly applied onto the YSZ electrolyte. The practical properties of LSC from thermal cycling and/or long-term operation become degraded due to the mismatch of the thermal expansion and the chemical reactivity between YSZ and LSC. These issues have been extensively investigated in an effort to overcome the problems of cathodes on the YSZ electrolyte. Areas of investigation include (1) the Mn-containing cathode/YSZ composite type [1–7]; (2) the Cocontaining cathode/Ce1 − yGdyO2 − δ (CGO) composite type [8,9]; (3) and the Co-containing cathode/CGO interlayer structure

1466

C. Lee et al. / Solid State Ionics 179 (2008) 1465–1469

[9–12]. The cathode properties of La0.8Sr0.2Co1 − xMnxO3 − δ (LSCM) on the YSZ electrolyte are investigated in this study. 2. Experimental 2.1. Cathode preparation La0.8Sr0.2Co1 − xMnxO3 − δ (LSCM) powder compositions of which the x values were 0, 0.2, 0.5, 0.6, 0.8 and 1 were prepared via the glycine nitrate process (GNP). Crystal phases were obtained by calcining the powders at 1000 °C for 1 h in an ambient atmosphere. The chemical form is presented by an abbreviated character representing the material composition followed by the molar ratio. For example, La0.8Sr0.2MnO3 − δ is expressed as LSM-82, and La0.8Sr0.2Co0.5Mn0.5O3 − δ is expressed as LSCM-8255.

Fig. 2. Temperature dependence of ASR of LSCM cathodes on YSZ electrolyte.

2.2. Electrolyte preparation YSZ electrolyte pellets were prepared from YSZ fine powder (Tosoh TZ-8Y). The powders were uniaxially pressed in a circular mold at 2 MPa. The green pellets were sintered at 1500 °C for 4 h in air. Dense YSZ pellets with a diameter of approximately 26 mm and a thickness of 2 mm were obtained. CGO-layered YSZ electrolyte pellets were prepared to investigate the role of the CGO interlayer. Ce0.9Gd0.1O2 − δ (CGO-91) slurry was coated onto both sides of the YSZ pellets using a tapecasting process, and the coated pellets were sintered at 1300 °C for 2 h. 2.3. Half-cell preparation and electrochemical measurements Cathode materials were circularly screen-printed with a diameter of 10 mm in the center of both sides of the YSZ and CGO-layered YSZ pellets. The cathodes were sintered at 1200 °C for 1 h, producing symmetrical half cells. The thicknesses of the cathode and CGO interlayer were observed to be approximately 30–40 µm and 5–10 µm, respectively, according to a scanning electron microscope (SEM). For example, a half-cell structure composed of LSCM-8246, CGO, and YSZ is shown in Fig. 1. AC

Fig. 1. SEM image of a half cell composed of LSCM-8246, CGO, and YSZ.

impedance was measured for each sample at 100 °C intervals from 500 °C to 900 °C in air. Impedance of the LSCM-8246 cathode on CGO-layered YSZ was measured at 50 °C intervals in the same temperature region for various oxygen partial pressures of 0.0002, 0.002, 0.02, 0.2, and 1 atm to observe the oxygen reduction reaction mechanism. 3. Results The area-specific resistance (ASR, [Ω cm2]) property of the LSCM cathode on the YSZ electrolyte was acquired as a function of the temperature and cathode composition (Fig. 2). ASR decreased as the Co content increased from LSM-82 initially due to the good oxygen reduction kinetics of Co-containing perovskite oxides. However, ASR increased conversely as the Co content increased past the turning point for LSCM-8246. LSCM8246 had the lowest impedance on the YSZ electrolyte. The ASR property was improved by introducing a CGO interlayer between the Co-containing cathode and YSZ electrolyte (Fig. 3). However, the property was not improved when the CGO electrolyte was used instead of the CGO-layered YSZ electrolyte. The ASR values for the same cathode material were three times lower at a high-temperature region above 700 °C and one order of magnitude lower at a low-temperature region below 700 °C after changing the electrolyte from YSZ to CGOlayered YSZ. This occurred mainly as a result of the inhibition of the undesired chemical reaction between the cathodes and YSZ electrolytes via the CGO interlayer. The activation energy of LSCM-8246 decreased from 1.8 eV to 1.4 eV and that of LSC-82 decreased from 2.0 eV to 1.6 eV by changing the electrolyte from YSZ to CGO-layered YSZ or CGO. The activation energies of each cathode on the CGO-layered YSZ electrolyte were identical to those on the CGO electrolyte, while the ASRs of each cathode on the CGO-layered YSZ electrolyte were lower than those on the CGO electrolyte. The ASR values of LSCM-8246 on CGO-layered YSZ were measured as a function of the temperature and oxygen partial pressure (Fig. 4). The ASR increased and the activation energy decreased as the oxygen partial pressure decreased. Both the

C. Lee et al. / Solid State Ionics 179 (2008) 1465–1469

1467

Fig. 3. Temperature dependence of ASR of LSCM-8246 and LSC-82 cathodes on YSZ, CGO, and CGO-layered YSZ electrolytes.

ASR and the activation energy increased as the temperature decreased. The activation energy was saturated at 0 eV at the lowest oxygen partial pressure in the high-temperature region, as shown in Fig. 4. The ASR properties at different frequency regions were observed for various oxygen partial pressures and various temperatures. The impedance, which is the basis for calculating the ASR, can be divided into three regions set apart by the frequency level. These are the high-frequency impedance (N103 Hz), medium-frequency impedance (~ 1–103 Hz), and low-frequency impedance (b 1 Hz) groups. The inverse of the ASR classified by the frequency can be arranged as a function of the oxygen partial pressure (Fig. 5). Medium-frequency impedance dominated the overall impedance at higher oxygen partial pressures at higher temperatures, and dominated as well at lower oxygen partial pressures at lower temperatures. Lowfrequency impedance became important at lower oxygen partial

Fig. 5. Oxygen partial pressure dependence of reciprocal ASR, classified by impedance frequency, of LSCM-8246 on CGO-layered YSZ.

Fig. 4. Temperature dependence of ASR of LSCM-8246 on CGO-layered YSZ for various oxygen partial pressures.

pressures at higher temperatures. Medium-frequency impedance and high-frequency impedance jointly dominated at higher oxygen partial pressures at lower temperatures.

1468

C. Lee et al. / Solid State Ionics 179 (2008) 1465–1469

4. Discussion LSCM-8246 is the best composition for the ASR property on the YSZ electrolyte among the various cathode compositions investigated here. This result is due to the contradictory relationship of the properties of the oxygen reduction reaction kinetics, the TEC, and the chemical reactivity between the cathode and the electrolyte. The number of oxygen vacancies can be increased by adding Co in a certain crystal structure to satisfy the electroneutrality condition. An increase in the number of oxygen vacancies in a cathode material leads to better oxygen reduction kinetics. However, the adhesion property with the electrolyte worsens as the number of vacancies increases because the TEC of the cathode increases. The TEC of LSM-82 is in an acceptable range of (11.2–11.7) × 10− 6/°C [13–17], compared to (10.3–11.0) × 10− 6/°C for YSZ [17–22]. However, the TEC of LSC-82 has a considerably high value of 17.2 × 10− 6/°C [23], which is enough to cause serious delamination or cracking at the interface of the electrolyte and the cathode. This TEC property was investigated for all experimental compositions from room temperature to 1000 °C to confirm the acceptability with the neighboring electrolyte (Fig. 6). The TEC of LSCM-8246 is lower than that of LSC82 and similar to that of LSCF-6428, which is a commercial product (Rhodia). Another issue is the chemical reactivity between the cathode and electrolyte. Cathodes containing the Co element react with YSZ electrolyte to form resistive phases. Intensity peaks for resistive phases were easily observed for the combinations of LSCM-8255 and YSZ by an X-ray diffraction (XRD) analysis. Resistive phases cause ohmic and activation polarization losses to increase. Judging from the above, LSCM8246 had the lowest impedance, compared to the other compositions because the positive and negative effects of the oxygen reduction reaction kinetics, the TEC, and the chemical reactivity were optimized. Co-containing cathodes are generally used for non-reactive electrolytes such as CGO. The ASR property of Co-containing cathodes on CGO is superior to that on YSZ. The CGO interlayer between the Co-containing cathode and the YSZ elec-

Fig. 7. Reaction order with respect to the impedance frequency.

trolyte inhibits the formation of resistive phases. The LSCM8246 cathode on CGO-layered YSZ had the lowest ASR, as shown in Fig. 3, because the formation of resistive phases was prevented and because there were more reaction sites. The particle size of the CGO-91 interlayer was nearly twice as large compared to that of LSCM-8246, as shown in Fig. 1. This implies that LSCM-8246 particles can seep into the pores of the CGO interlayer during the fabrication of the half cell. There are some other possibilities for the improvement of the ASR property, such as a change in the ionic transfer behavior by changing the electrolyte of YSZ to CGO. The activation energies of LSCM-8246 and LSC-82 on CGO-layered YSZ or CGO were 1.4 eV and 1.6 eV against 1.8 eV and 2.0 eV on the YSZ electrolyte, respectively. One possible explanation for this is that the rate-determining step was changed from the oxygen ion transfer into resistive phases to the oxygen surface exchange on the cathode. The inhibition of the formation of resistive phases by the CGO interlayer improves the ASR property and decreases the activation energy. However, the values of the activation energies do not provide much information regarding the oxygen reduction reaction mechanism because the activation energies of the oxygen surface exchange kinetics for LSCM-8237, LSCM-8255, and LSC-82 are known to be 1.7, 1.2, and 1.3 eV, respectively [24,25], which implies a doubtful connection between the oxygen surface exchange kinetics and the ASR property. The oxygen partial pressure experiment provides specific information concerning the oxygen reduction reaction behavior. Reciprocal ASR values were related to the oxygen partial pressure because reciprocal ASR is proportional to the current flux at a given overpotential. The reaction order (m) can be calculated using the oxygen partial pressure dependence of the current flux or the reciprocal ASR with respect to the impedance frequency region from Fig. 5, with reference to previous studies [26,27] (Fig. 7). ASR1 ~POm2

Fig. 6. TEC of selected cathodes (LSCM series and LSCF-6428) and electrolytes (CGO and YSZ).

ð2Þ

The oxygen reduction reaction step shown below was adopted because the surface diffusion of the adsorbed oxygen is

C. Lee et al. / Solid State Ionics 179 (2008) 1465–1469

thought to prevail to transport oxygen due to the relatively low oxygen ion diffusion coefficient of LSCM oxides [24,25]. O2 ðgasÞYO2 ðgasÞ;

m¼1

ð3Þ

O2 ðgasÞY2Oad ; m ¼ 1

ð4Þ

Oad YOTPB ; m ¼ 1=2

ð5Þ

OTPB þ e YO TPB ; m ¼ 3=8

ð6Þ

 2 O TPB þ e YOTPB ; m ¼ 1=8

ð7Þ

•• x O2 TPB þ VO YOO ;

ð8Þ

m¼0

Comparing the oxygen reduction reaction step (Eqs. (3)–(8)) with Fig. 7, the following can be deduced. The reciprocal ASR for medium frequencies is approximately proportional to 3/8 power of the oxygen partial pressure. The 3/8 value represents oxygen ionization as a rate-determining step for the oxygen reduction reaction (Eq. (6)). The reciprocal ASR for low frequencies is proportional to 1 power, which implies that lowfrequency impedance is governed by oxygen gas diffusion in the porous cathode (Eq. (3)). The reciprocal ASR for high frequencies is proportional to 0 power, which implies that highfrequency impedance is governed by oxygen ion transfers into the electrolyte (Eq. (8)). The oxygen gas diffusion effect becomes important as the temperature increases and the oxygen partial pressure decreases, as shown in Fig. 5(a). This is also confirmed from the ASR of nearly 0 eV at higher temperatures at lower oxygen partial pressures shown in Fig. 4. Note that the temperature effect on the oxygen gas diffusion is insignificant, whereas the oxygen ion diffusion or the oxygen surface exchange kinetics is affected strongly by temperature, following the Arrhenius equation-type behavior [28]. Oxygen ionization dominates the ASR at higher temperatures at higher oxygen partial pressures, as shown in Fig. 5(a), or at lower temperatures at lower oxygen partial pressures, as shown in Fig. 5(c). Both oxygen ionization and the oxygen ion transfer into the electrolyte become equivalently the main factors determining the ASR at lower temperatures at higher oxygen partial pressures, as shown in Fig. 5(c). 5. Conclusions The composition of the La0.8Sr0.2Co1 − xMnxO3 − δ structure as a cathode showed excellent performances on a YSZ electrolyte for intermediate-temperature solid oxide fuel cells. LSCM-8246 was superior to the other compositions tested in this study because the cathodic properties of the LSCM-8246 of the oxygen reduction reaction kinetics, the TEC, and the chemical reactivity with YSZ were optimized. A CGO-91 interlayer was introduced between the LSCM-8246 cathode and the YSZ electrolyte in order to inhibit the serious degradation

1469

expected by resistive phases generated by material reactions. The ASR and activation energy levels were considerably decreased when using the CGO-91 interlayer. The reaction orders of LSCM-8246 were revealed to be 0 for high-frequency impedance, 3/8 for medium-frequency impedance, and 1 for low-frequency impedance through the oxygen partial pressure dependence. The LSCM-8246 cathode on CGO-layered YSZ showed the lowest impedance of 0.14 Ω cm2 at 700 °C. Acknowledgements This work is an outcome of the projects of the Best Lab program of the Ministry of Commerce, Industry and Energy (MOCIE) and the Brain Korea 21 (BK21) program of the Ministry of Education and Human Resources Development (MOE). The authors greatly appreciate their financial support. References [1] K. Yamahara, T.Z. Sholklapper, C.P. Jacobson, S.J. Visco, L.C. De Jonghe, Solid State Ionics 176 (2005) 1359. [2] Y. Ji, J.A. Kilner, M.F. Caloran, Solid State Ionics 176 (2005) 937. [3] A. Barbucci, M. Viviani, P. Carpanese, D. Vladikova, Z. Stoynov, Electrochimica Acta 51 (8–9) (2006) 1641. [4] A. Barbucci, P. Carpanese, G. Cerisola, M. Viviani, Solid State Ionics 176 (2005) 1753. [5] A. Barbucci, R. Bozzo, G. Cerisola, P. Costamagna, Electrochimica Acta 47 (13–14) (2002) 2183. [6] M.J. Jørgensen, M. Mogensen, Journal of the Electrochemical Society 148 (5) (2001) A433. [7] M.J. Jørgensen, S. Primdahl, M. Mogensen, Electrochimica Acta 44 (24) (1999) 4195. [8] E. Perry Murray, M.J. Sever, S.A. Barnett, Solid State Ionics 148 (2002) 27. [9] W.G. Wang, M. Mogensen, Solid State Ionics 176 (2005) 457. [10] M. Shiono, K. Kobayashi, T.L. Nguyen, K. Hosoda, T. Kato, K. Ota, M. Dokiya, Solid State Ionics 170 (2004) 1. [11] H. Uchida, S. Arisaka, M. Watanabe, Solid State Ionics 135 (2000) 347. [12] S. Charojrochkul, K.-L. Choy, B.C.H. Steele, Solid State Ionics 121 (2000) 107. [13] M. Mori, Solid State Ionics 174 (2004) 1. [14] M. Mori, Y. Hiei, N.M. Sammes, G.A. Tompsett, Journal of the Electrochemical Society 147 (4) (2000) 1295. [15] A. Hammouche, E. Siebert, A. Hammou, Materials Research Bulletin 24 (1989) 367. [16] A. Poirson, P. Decorse, G. Caboche, L.C. Doufour, Solid State Ionics 99 (1997) 287. [17] Y. Sakaki, Y. Takeda, A. Kato, N. Imanishi, O. Yamamoto, M. Hattori, M. Iio, Y. Esaki, Solid State Ionics 118 (1999) 187. [18] H.U. Anderson, Solid State Ionics 52 (1992) 33. [19] M. Mori, T. Abe, H. Itoh, O. Yamamoto, G.Q. Shen, Y. Takeda, N. Imanishi, Solid State Ionics 123 (1999) 113. [20] F. Tietz, Ionics 5 (1999) 129. [21] N.Q. Minh, Journal of the American Ceramic Society 76 (3) (1993) 563. [22] M. Mogensen, T. Lindegaard, U.R. Hansen, G. Mogensen, Journal of the Electrochemical Society 141 (8) (1994) 2122. [23] X. Chen, J. Yu, S.B. Adler, Chemistry of Materials 17 (2005) 4537. [24] R.A. De Souza, J.A. Kilner, Solid State Ionics 106 (1998) 175. [25] R.A. De Souza, J.A. Kilner, Solid State Ionics 126 (1999) 153. [26] J-D. Kim, G-D. Kim, J-W. Moon, Y-I. Park, H-W. Lee, K. Kobayashi, M. Nagai, C-E. Kim, Solid State Ionics 143 (2001) 379. [27] F.H. van Heuveln, H.J.M. Bouwmeester, Journal of The Electrochemical Society 144 (1997) 134. [28] C. Lee, J. Bae, Journal of Power Sources, 176 (2008) 62.