- Email: [email protected]
(Bi0.15La0.27Sr0.53)(Co0.25Fe0.75)O3-δ perovskite: A novel cathode material for intermediate temperature solid oxide fuel cells
Journal of Power Sources 334 (2016) 137e145
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
(Bi0.15La0.27Sr0.53)(Co0.25Fe0.75)O3-d perovskite: A novel cathode material for intermediate temperature solid oxide fuel cells Deni S. Khaerudini a, c, Guoqing Guan a, b, *, Peng Zhang a, Pairuzha Xiaoketi a, Xiaogang Hao d, Zhongde Wang d, Yutaka Kasai e, Abuliti Abudula a, ** a
Graduate School of Science and Technology, Hirosaki University, 1-Bunkyocho, Hirosaki 036-8560, Japan North Japan Research Institute for Sustainable Energy (NJRISE), Hirosaki University, 2-1-3 Matsubara, Aomori 030-0813, Japan Research Center for Physics, Indonesian Institute of Sciences, Gd. 440 Kawasan Puspiptek Serpong, Tangerang Selatan 15314, Banten, Indonesia d Department of Chemical Engineering, Taiyuan University of Technology, Taiyuan 030024, China e Industrial Research Institute, Aomori Prefectural Industrial Technology Research Center, 4-11-6, Second Tonyamachi, Aomori 030-0113, Japan b c
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
(Bi0.15La0.27Sr0.53)x(Co0.25Fe0.75)O3-d is implemented as cathode in IT-SOFC. BiLSCF0.9 is chemically compatible with LSGM and shows the best performance. A low polarization resistance of 0.039 U cm2 for BiLSCF0.9 is achieved at 700 C. SOFC with BiLSCF0.9 shows the maximum power of 0.66 W cm2 at 700 C. Polarization loss is minimized by optimization of compositions and structure.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 June 2016 Received in revised form 15 September 2016 Accepted 7 October 2016
Perovskite oxides (Bi0.15La0.27Sr0.53)x(Co0.25Fe0.75)O3-d (BiLSCFx, x ¼ 0.8, 0.9, 1.0, 1.1) have been synthesized by solid state reaction and evaluated as a novel cathode material for intermediate-temperature solid oxide fuel cells (IT-SOFCs). The effects of A-site variations on lattice structure, calcination temperature, oxygen desorption and electrochemical properties of BiLSCFx are investigated. This kind of material has perfectly cubic structure based on the Pm-3m space group whose lattice size increases with x, which is thermally stable after calcination and shows desirable chemical compatibility with La0.8Sr0.2Ga0.8Mg0.2O3-d electrolyte at 1150 C for 8 h under air atmosphere. Among those A-site variations, it is found that BiLSCF0.9 demonstrates the best cathode performance. It has the minimum polarization resistance value of 0.039 U cm2 at 700 C and a-oxygen desorbed about 0.031 mmol g1, indicating a good reactivity and strong adsorbate of O2. The single cell with BiLSCF0.9 cathode delivers a power density of 0.66 W cm2 at 700 C with humidified H2 (~3% H2O) as the fuel and ambient air as the
Keywords: IT-SOFC BiLSCF Cathode Oxygen reduction reaction Electrochemical performance
* Corresponding author. North Japan Research Institute for Sustainable Energy (NJRISE), Hirosaki University, 2-1-3 Matsubara, Aomori 030-0813, Japan. ** Corresponding author. E-mail addresses: [email protected] (G. Guan), [email protected] (A. Abudula). http://dx.doi.org/10.1016/j.jpowsour.2016.10.026 0378-7753/© 2016 Elsevier B.V. All rights reserved.
138
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
oxidant. In addition, the cell shows sufficient stability with ~9% degradation over 75 h at 600 C. It indicates that BiLSCF0.9 is a promising candidate for application as cathode material in IT-SOFCs. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Solid oxide fuel cells (SOFCs) are the electrochemical power generation devices that convert chemical energy into electrical energy directly, providing several advantages of fuel flexibility, low pollutant emission, environmental friendliness, high power density, and high efficiency [1]. Recently, long-term durability has been the major concern for the commercialization of SOFC systems. One of methods to mitigate this issue is to decrease the operation temperature to the low or intermediate temperature range (500e700 C) through the application of advanced electrolyte and/ or electrode materials. That is to develop intermediate temperature-SOFCs (IT-SOFCs) since it is potential to enhance the stability and economic feasibility of SOFCs at lower operating temperature. However, at low and intermediate temperature range, the kinetics for oxygen transfer in cathode becomes a major limiting step for good performance [2]. The effective oxygen reduction reaction (ORR) area in the strontium-doped lanthanum manganite (LSM) with perovskite structure has proven to be unfavorable at low temperatures, (223 U cm2 at 650 C) [3] and (1100 U cm2 at 650 C) [4], when compared with those doped by cobalt, chromium and iron oxides [5e7], since LSM has significantly inferior kinetics for oxygen transfer at the electrolyte interface [5]. The perovskite LaCoO3 is also recognized as a candidate material on IT-SOFC cathodes, and especially the aliovalent element substituted LaCoO3 such as La12 xSrxCoO3 shows promising electrochemical activity (0.2 U cm at 600 C) towards the oxygen reduction reaction [8]. B-site doping with various transition metals has been investigated extensively and the composition La0.6Sr0.4Co0.2Fe0.8O3-d is often referred as one of functional cathodes due to its suitable combination of performance, stability, reliability, thermo-mechanical and chemical compatibility with other cell components. Recently, Bi replacement of A-site element in these perovskite materials has stimulated the exploration of their application as the cathode in IT-SOFC [9e11]. Especially, by Bi substitution, different electronic configuration could be generated since Bi3þ has higher polarizability owing to its 6s2 lone pair and lower basicity in contrast to those observed in Ba2þ (Ba1-xSrxFe1-yCoyO3-d) [12] as well as La3þ based materials. Thus, it is expected to increase the structural stability, offer higher concentration and mobility of oxygen vacancies at the reduced temperatures, and promote the catalytic activity in oxygen dissociation process, which is often the rate limiting step in oxygen reduction in IT-SOFC cathode [13]. In our previous study [14], LSCF50-BiSCF50 composite material was made by mixing self-made Bi0.3Sr0.7Co0.3Fe0.7O3-d (BiSCF3737) powder and commercial (La0.6Sr0.4)0.9Co0.2Fe0.8O3±d (LSCF) powder and used as cathode material of IT-SOFC. A significant performance improvement, specifically on oxygen dissociation process, was found with almost one order of magnitude higher than BiSCF3737 electrode and two orders of magnitude higher than LSCF electrode. Based on these results, in this study, we tried to synthesize a single phase material containing the same element contents in LSCF50BiSCF50 composite as a novel cathode material for IT-SOFCs. To the best our knowledge, BiLSCFx materials have not been reported to date. It is expected to find some better catalytic activity and structural stability than those in the previous composite
configuration. Here, we explored the structure, electrochemical properties, oxygen desorption and chemical reactivity of such single-phase cathodes with the perovskite framework, i.e., single phase (Bi0.15La0.27Sr0.53)x (Co0.25Fe0.75)O3-d compounds with x ¼ 0.8, 0.9, 1.0, 1.1, called as BiLSCF0.8, BiLSCF0.9, BiLSCF1.0 and BiLSCF1.1, respectively. Their performances in the electrolyte-supported SOFCs operating at temperatures ranged from 550 to 700 C were further investigated. 2. Experimental 2.1. Preparation of cathode materials The
original
powder of
(Bi0.15La0.27Sr0.53)x(Co0.25Fe0.75)O3-
d (x ¼ 0.8, 0.9, 1.0, 1.1) was synthesized using solid state reaction
route from the high-purity starting chemicals of Bi(NO3)3$5H2O, La(NO3)3$6H2O, Sr(NO3)2, Co(NO3)2$6H2O, Fe(NO3)3$9H2O, and Ce(NO3)3$6H2O (99.9%, Wako, Japan) with exactly stoichiometric molar ratio. These powders were mixed and ball-milled at a rotation speed of 300 rpm for 6 h in ethanol media. The resulting precursor solution was evaporated at 80 C for 24 h and thoroughly ground in an agate mortar and subsequently calcined at different temperatures from 1050 to 1150 C for 8 h in air to obtain fine BiLSCF powders. Similarly, Ce0.9Gd0.1O1.95 (GDC) as a composite anode material was also prepared by the solid state reaction from the appropriate proportion of Ce and Gd nitrates (99.9%, Wako, Japan) mixed with ethanol by ball milling. Then the mixture was evaporated and calcined at 900 C for 5 h in air to gain the GDC powder. The commercial NiO (Soekawa Chemicals, Japan) and La0.8Sr0.2Ga0.8Mg0.2O3-d (LSGM, FCM, USA) powders were used as received for the preparations of anode and electrolyte, respectively. To assess the phase reaction, the chemical compatibility of the BiLSCF cathode and LSGM electrolyte was investigated with X-ray diffraction (XRD) analysis by sintering the mixed powders in a weight ratio of 1:1 at the optimum calcination condition for 8 h. 2.2. Microstructural and morphological characterizations A Rigaku SmartLab X-Ray Diffractometer with Cu-Ka (l ¼ 1.5405 Å) radiation source was used to examine the phase composition and lattice structure of the sample. Powder samples were scanned with 2q ranging from 20 to 80 with a 5 /min scan rate. Structural refinement was fitted using the PDXL2.0 Rigaku Data Analysis Software package. A Hitachi SU8010 scanning electron microscope (SEM) equipped with a Horiba Scientific energy dispersive spectrometer (EDS) analyzer was used to investigate the microstructure of electrode and elemental distribution across cathode-electrolyte interface and the possible secondary phase which cannot be detected by XRD. 2.3. Oxygen reducibility test Oxygen desorption properties of different samples as a function of temperature were monitored by a Belcat oxygen temperatureprogrammed desorption (O2-TPD) catalyst analyzer (Belcat, Bel Japan, Inc.). In general, about 0.1 g of powder sample was placed in a quartz fixed-bed microreactor with 8 mm inner diameter. Before
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
each run, the sample was treated in O2 with a flow rate of 20 cm3 min1 at 500 C for 1 h to confirm that O2 can be adsorbed and activated completely across the bulk structure of the sample as those reported in other work [15]. Then, it followed by cooling to RT in O2 and argon purging with a flow rate of 50 cm3 min1 for 30 min. The purpose of argon purging was to remove gas-phase oxygen in the system. The sample was then heated from RT to 900 C with a heating rate of 10 C min1. The effluent was monitored by a thermal conductivity detector (TCD). The amount of O2 desorbed from the sample was quantified by calibrating the peak area against that of a standard O2 pulse. 2.4. Fabrication and testing of cells Electrochemical performance of the as-prepared cathode was evaluated with both of the symmetrical cells and electrolyte supported single cells. Dense LSGM as the electrolyte had a diameter of ~20 mm, which was prepared through a dry pressing method at 250 MPa and sintered at 1450 C for 10 h. Symmetrical cells were employed to investigate the electrochemical impedance behaviour. The as-synthesized powders were mixed with 80 wt% organic additives (a-terpineol þ ethyl cellulose) solution to form slurry. The electrode slurry was tape-casted onto both sides of the LSGM pellet with a thickness of ~0.5 mm. After drying, the electrode was sintered at 1150 C for 2 h in air to form the symmetrical cell. Single cell studied in this work was LSGM electrolyte-supported (~300 mm thickness) cell. The anode layer consisting of 40 wt% GDC þ 60 wt% NiO composite was prepared by tape casting the anode slurry onto the LSGM electrolyte and sintered at 1350 C for 2 h. Subsequently, cathode slurry was casted on the opposite side of the electrolyte, followed by sintering with the similar conditions as in the symmetrical cells to form the single cell. The effective area of electrode was 0.78 cm2. Pt paste/mesh was used as the current collector for all electrochemical evaluations. According to our previous and other work [14,16], the Pt current collector should have no obvious influence on the electrode catalytic activity. The electrode polarization resistance of the symmetrical cell was performed using an electrochemical workstation composed of a Solartron 1255B frequency response analyzer and SI-1287 galvanostat. The cell was measured at the temperatures ranged from 550 to 700 C (50 C intervals) under air atmosphere. The performance of cathode was further examined in the single fuel cell configuration. The cell was sealed on the top of an alumina support tube by glass ring. The anode was firstly reduced in situ from NiO/ GDC to Ni/GDC cermet anodes. 100 cm3 min1 humidified H2 (ca. 3% H2O) and 50 cm3 min1 ambient air were employed as the fuel and oxidant, respectively. The measurements were also carried out at temperatures ranged from 550 to 700 C (50 C intervals). Performance stability test was performed under a constant voltage of 0.52 V for 75 h. The performance characteristics of the cell and electrochemical impedance spectra were measured under opencircuit conditions over a frequency range from 0.1 Hz to 1 MHz with an AC perturbation of 10 mV and fitted using Z-view software. 3. Results and discussion 3.1. Phase structure The phase purity and composition of (Bi0.15La0.27Sr0.53)1.0 (Co0.25Fe0.75)O3-d (only the pattern of BiLSCF1.0 as an example is shown here) after calcined at 1050, 1100, and 1150 C for 8 h in air was confirmed by XRD. As shown in Fig. 1a, at 1050 and 1100 C, some impurity phases are found in the XRD patterns. Increasing the calcination temperature not only sharpens the detected diffraction peaks, but also increases the degree of crystallinity. It is obvious
139
that only reflections corresponding to the perovskite phase can be detected for the cathode calcined at 1150 C, which is the optimum sintering temperature for the preparation of BiLSCF cathode, and in this case, the single phase compound can be formed. Fig. 1b shows XRD patterns of samples of (Bi0.15La0.27 Sr0.53)x(Co0.25Fe0.75)O3 d (x ¼ 0.8, 0.9, 1.0, 1.1) calcined at 1150 C for 8 h. One can see that all the samples are crystallized as the cubic perovskite structure in the space group of Pm-3m. Table S1y shows the space groups and lattice parameters of these compounds. One can see that various A-site compositions have no effect on the formation of the pure cubic symmetry perovskite phase. With the increasing of x value on Asite, the XRD diffraction peak slightly shifts to the left as confirmed by the magnified main peak (110) of BiLSCF sample, indicating a lattice expansion. A simple way to attempt to explain the expansion is to consider the expansion of A/B distance as the variation in the averaged ionic radius of A-site cation compositions [17]. For simplicity, it is assumed that the oxide crystallizes in a cubic perovskite-type structure without any displacement of atoms, and the lattice volume is governed by the averaged ionic radius of the cation. It is also supposed that there is no preferred orientation in the specimen. This expansion also has potential advantages for the creation of oxygen vacancies [18]. Furthermore, the incorporation of Bi into A-site cation sublattice may further generate a large concentration of oxygen vacancies, due to the charge balance as shown in the following Kroger-Vink notation: ðLa;SrÞðCo;FeÞO3
Bi2 O3 !V::o þ 2Bi0La þ 3Oxo Mobility of oxygen ion is increased and subsequently results in the increase in oxygen ion conductivity, however, this can also lead to the loss of other desirable properties, such as the electrical conductivity [19]. Deficiency/excess of A-site can therefore be used to tune the conductivity (mixed ionic-electronic conductivity) of the material, which will be addressed later. Furthermore, lattice expansion is also an important factor in the design or selection of cathode material. The surface microstructure and elemental composition of the powder BiLSCF1.0 were characterized by SEM/EDS analysis. As shown in Fig. 1c and d, after calcined at 1150 C for 8 h, the grain size distribution is in the range of 1e4 mm. EDS analysis results reveal that no any impurities are included. The inset table in Fig. 1d shows the measured elemental composition. Obviously, the Bi:La:Sr:Co:Fe atomic ratio is 0.1498:0.2721:0.5320:0.2568:0.7500 which is very close to the nominal stoichiometric composition of 0.15:0.27:0.53: 0.25:0.75, indicating that the stoichiometry in the solid is relatively homogenous. To obtain the detail information about the crystal structure of BiLSCF, Rietveld refinement was applied. Here, only the result on the BiLSCF0.9 was selected because it has the highest catalytic activity (as indicated in the next section). The calculated and experimental XRD profiles of BiLSCF0.9 are shown in Fig. 2, which is also in agreement with the calculated Goldschmidt tolerance factor (t) about 0.976, indicating that the cubic structure is formed. The reliability factors obtained from the Rietveld refinement are Rp z 4.64, Rwp z 5.58, Rexp z 2.04, RBragg z 2.61, and c2 z 2.38, demonstrating the good agreement between the calculated and observed data. Based on ABO3 model, A-site atoms of Bi/La/Sr sit at cube corner (0, 0, 0) position, B-site atoms of Fe/Co sit at body center (1/2, 1/2, 1/2) position, and O atoms sit at face centered (0, 0, 1/2) position with random oxygen vacancy, as a result of highly polarizable of Bi3þ ions caused by its 6s2 lone pair electrons and weak BieO bonds structure, in a single oxygen site. Generally, such a high oxygen vacancy is beneficial for the free movement of oxygen ions and supplies the active sites for the oxygen adsorption, dissociation and diffusion at the surface of cathode. Therefore, this
140
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
Fig. 1. (a) XRD patterns of (Bi0.15La0.27Sr0.53)(Co0.25Fe0.75)O3-d (BiLSCF) powders after calcinations at 1050, 1100 and 1150 C for 8 h and X-ray diffraction patterns of the respective powders in the range of 25 2q 40 , (b) XRD of BiLSCF (x ¼ 0.8e1.1) powders calcined at 1150 C for 8 h in air and the right Fig. shows the magnified (110) peak of the respective powders, (c) SEM micrograph and (d) EDS analysis of BiLSCF1.0 after calcined at 1150 C for 8 h in air. The inset in (d) shows the table of comparison of EDS results with the targeted values.
BiLSCF and LSGM appear, and no significant impurities are identified under the detection limit from the XRD patterns. These indicate that no reaction occurred between BiLSCF and LSGM upon 1150 C for 8 h, therefore the effect of impurities on the electrochemical performance from sintering process is negligible. 3.2. Effect of cathode composition ratio on oxygen reduction reaction
Fig. 2. Rietveld fit of XRD data of as-prepared (Bi0.135La0.243Sr0.477)(Co0.25Fe0.75)O3d (BiLSCF0.9) powders. The series with blue thick marks corresponds to the allowed Bragg reflections. Inset Fig. shows unit cell of the crystal structure of BiLSCF0.9, the O site is partially occupied by oxygen vacancy. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
material should perform superior electrochemical performance. In general, the interface reaction between electrolyte and cathode is highly detrimental for the cell performance due to the increase of interfacial polarization resistance. Here, Fig. S1y shows XRD patterns of BiLSCF1.0, LSGM and BiLSCFx þ LSGM powders prepared by the direct calcinations of the mixture of BiLSCF and LSGM powders in 50:50 wt% at 1150 C for 8 h in air. After calcination, it shows that only those peaks corresponding to original
The oxygen reduction reaction (ORR) activity of the BiLSCF (x ¼ 0.8, 0.9, 1.0, 1.1) cathodes was quantified by electrochemical impedance spectroscopy (EIS) using symmetrical cells with LSGM electrolyte as support (~0.5 mm thickness). In order to properly compare the intrinsic material activity, measurements are carried out when the porous cathode morphologies of each cell variations are similar. From Fig. 3, it can be seen clearly that the grain size and pore morphology of all cathodes are comparable. Here, the variations in impedance behaviour can therefore be addressed mainly due to the composition. Fig. 4 shows Nyquist plots between 550 and 700 C in air at open circuit condition. The area specific resistance (ASR) of the cathode is obtained from the difference between the low-frequency and high-frequency intercepts on the real axis and represents the polarization resistance (Rp) of the cathode. The best fit equivalent circuit for the analysis of the impedance data is shown in the inset of Fig. 4b with two parallel connected resistances (R) and a constant phase element (CPE) in series. In this model, the ohmic resistance is set at zero (omitted) in the Nyquist plots as it is mainly contributed by the electrolyte. The highfrequency arc (RHF), specifically observed from the mid-frequency (103-102 Hz), corresponds to the ionic transport resistance [20];
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
141
Fig. 3. Cross-sectional SEM morphologies of cathodes on LSGM electrolyte with (a) BiLSCF0.8, (b) BiLSCF0.9, (c) BiLSCF1.0 and (d) BiLSCF1.1 in symmetrical fuel cells.
Fig. 4. Nyquist plots of impedance spectra for (a) BiLSCF0.8, (b) BiLSCF0.9, (c) BiLSCF1.0 and (d) BiLSCF1.1 cathodes tested at a temperature range from 550 to 700 C in air. The inset in (b) is the equivalent circuit used for all data fitting.
and the low-frequency (101-101 Hz) arc (RLF) represents the surface diffusion of oxygen species including the oxygen adsorption and dissociation processes [21]. From the impedance spectra at 700 C, except for BiLSCF1.1, it can be seen clearly that the high-frequency responses of cathodes show significantly higher contribution, about 70%, than the lowfrequency ones, as shown by the representative of impedance plot in the inset of Fig. 4c. Meanwhile, BiLSCF1.1 shows the opposite
behaviour under the same condition; where the high-frequency responses only contribute approximately 17% of the total resistance, as shown in the inset of Fig. 4d. It seems that the impedance behaviour of BiLSCF1.1 can be rationalised mainly by the excess of Bi on A-site which probably provides more dominant ionic contribution than other A-site values. However, at temperature lower than 700 C, all of cathodes show similar behaviour. The performance of high-frequency is significantly lower than that of
142
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
Table 1 Deconvoluted high- and low-frequency polarization resistance (U cm2) of BiLSCF0.8, BiLSCF0.9, BiLSCF1.0 and BiLSCF1.1 in air at different temperatures from 550 to 700 C. Cathode
T ( C)
Rp (U cm2)
RHF (U cm2)
RLF (U cm2)
BiLSCF0.8
550 600 650 700 550 600 650 700 550 600 650 700 550 600 650 700
0.359 0.160 0.085 0.048 0.219 0.108 0.058 0.039 0.248 0.118 0.064 0.041 0.798 0.451 0.239 0.137
0.071 0.039 0.021 0.037 0.036 0.020 0.011 0.031 0.059 0.031 0.021 0.025 0.105 0.069 0.036 0.023
0.288 0.121 0.063 0.012 0.182 0.088 0.047 0.008 0.189 0.087 0.043 0.016 0.693 0.381 0.203 0.114
BiLSCF0.9
BiLSCF1.0
BiLSCF1.1
low-frequency, as listed in Table 1; indicating that the incorporation of Bi into cathode structure is more effective on ionic transfer at the temperature lower than 700 C, and the high-frequency performance is controlled mainly by the appropriate Bi incorporation on A-site cation composition. Among all the cathodes investigated, BiLSCF0.9 performs well not only for the charge transfer at high-frequency process but also for the oxygen adsorption and dissociation at the low frequency process. The appropriate composition of Bi in the cathode structure provides more channels to transport oxygen ions via increasing the concentration of total oxygen vacancies. This will be explained further on the separated Section in 3.3. Hence, the oxygen reduction can occur much more at three-phase boundaries, and as a result, more oxygen ions may diffuse into the LSGM electrolyte to improve the whole oxygen diffusion. Fig. S2ay shows typical Arrhenius plots of the total interfacial polarization resistance (Rp) values obtained from the data of Fig. 4 as a function of reciprocal temperature, together with activation energies (Ea) of BiLSCF (x ¼ 0.8, 0.9, 1.0, 1.1) cathodes. It clearly shows that the Rp value decreases significantly with the increase in temperature, which can be attributed to the enhanced catalytic activity at higher temperature. Such temperature-dependent behaviour is typical for the mixed conducting perovskite [22]. While BiLSCF1.1 cathode exhibits an Ea value of 0.92 eV, BiLSCF0.9 provides a lower value of 0.8 eV that is deemed more favourable for obtaining high performance at lower temperatures, as also can be evidenced by their Rp values. As shown in Table 1, the Rp values of all prepared cathodes, BiLSCF0.8, BiLSCF0.9, BiLSCF1.0 and BiLSCF1.1, are 0.048, 0.039, 0.041 and 0.137 U cm2 at 700 C, respectively, in which BiLSCF0.9 delivers the lowest Rp values at all temperatures tested. Moreover, it is also comparable and lower than other reported work on high-performance perovskite cathodes such as La0.3Sr0.7Ti0.4Co0.6O3-d (0.093 U cm2 at 700 C) [23], La0.6Sr0.4Fe0.8Co0.2O3-d (0.1 U cm2 at 700 C) [24], La1.7Ca0.3Ni0.7Cu0.3O4þd (1.45 U cm2 at 650 C) [25], BaCo0.7Fe0.2 Nb0.1O3-d (0.12 U cm2 at 700 C) [26], Ba0.9Co0.7Fe0.2Nb0.1O32 d (0.07 U cm at 700 C) [27], SrSc0.175Nb0.025Fe0.8O3-d (0.040 U cm-2 at 700 C) [28], La0.4875Ca0.0125Ce0.5O2-d thin filmed on LaxSr12 xCoyFe1-yO3-d (0.184 U cm at 700 C) [29], La0.6Sr0.4Co0.2Fe0.8O32 Pd (<0.05 U cm at 700 C) [30], indicating its potential for application in IT-SOFCs. The electrode polarization resistances (Rp) of BiLSCF cathodes obtained from the impedance spectra as functions of the value of x and temperature are plotted in Fig. S2by. For A-site deficient
samples, the polarization resistances slightly decrease with the increasing of x from 0.8 to 0.9 at the temperatures ranged between 550 and 700 C and the minimum Rp is obtained when x ¼ 0.9. With further increase of x value to 1.0, the Rp of BiLSCF cathode begins to increase. Obviously, when x is 1.1, the Rp increases rapidly to several times higher than those of A-site deficient and stoichiometry samples under the same test condition. This can be explained by the significant difference between low-frequency arc (~15%) and high-frequency arc (~85%) contributions of BiLSCF1.1 when compared with other BiLSCF compositions, as shown in Fig. 4. It indicates that ionic contribution in this cathode is probably much more dominant than electronic properties. At higher operation temperature, the value of x corresponding to the lowest Rp is slightly lower than that at relatively lower temperature. These results demonstrate that the polarization resistance of BiLSCF is dependent not only on the A-site nonstoichiometry with either Asite excess or deficiency and the operating temperature, but also on gas transport to and away from the reaction zone. At higher operating temperature, the reaction is highly activated; on the contrary, at relatively low temperature, gas transport might be limited. 3.3. Oxygen sorption analysis The interaction between the perovskite surface and adsorbed molecules, for instance, the sorption of oxygen on the oxygen ion vacancies, can be further evaluated using temperature programmed desorption of oxygen (O2-TPD) technique. The increase of oxygen vacancies concentration has ever confirmed by the reduction of a-oxygen amount desorbed in the low temperature range in the O2-TPD experiment [31]. The O2-TPD profiles could provide information on the behaviour of lattice oxygen. Fig. 5 depicts the O2-TPD profiles of BiLSCF (x ¼ 0.8, 0.9, 1.0, 1.1) powders after exposure to O2 for 1 h at 500 C. All samples exhibit two desorption peaks (a and b oxygen) which are closely related to the formation of oxygen vacancies and the desorption of oxygen from the surface lattice, which are dependent on the composition as well as the lattice structure of the material. The a desorption, locating at a strong peak from ca. 300e600 C, is ascribed to the oxygen vacancies formed, originated from Oads located at oxygen vacancies
Fig. 5. O2-TPD profiles of (a) BiLSCF0.8, (b) BiLSCF0.9, (c) BiLSCF1.0 and (d) BiLSCF1.1with argon as sweep gas.
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
143
Fig. 6. Typical SEM morphologies for the cell after the operating test: (a) cross section view of electrolyte-supported Ni-GDCjLSGMjBiLSCF0.9 cell configuration, (b,e) cross section views of BiLSCF0.9/LSGM interface, (c) surface view of BiLSCF0.9 cathode (d) cross section view of Ni-GDC/LSGM interface.
[32]. The a-oxygen is related to the formation of Co4þ ions or positive holes. Here, the desorption of a-oxygen may be expressed as:
Oxo þ 2h_ /V::o þ 1=2O2 On the other hand, the b-oxygen, originated from the surface oxygen lattice (O2 latt) desorbs at a commencement temperature of approximately 800 C, is accompanied by the reduction of Co ion to a lower valence state, described as:
Oxo þ 2CoxCo /V::o þ 2Co0Co þ 1=2O2 For all BiLSCF samples, the peaks of a desorption exhibit much stronger than those of b desorption, indicating that a desorption is more effective to the A-site variations than b desorption. As such, its intensity can provide a rough estimation of the amount of oxygen vacancies on the surface. Here, the area of the a-oxygen desorption peak of BiLSCF1.1 is the largest, followed by BiLSCF0.8, BiLSCF1.0 and then BiLSCF0.9. The amounts of a-oxygen adsorbed on BiLSCF1.1, BiLSCF0.8, BiLSCF1.0 and BiLSCF0.9 are estimated to be 0.059, 0.045, 0.038 and 0.031 mmol g1, respectively, indicating that the oxygen vacancy concentration decreases in the order of BiLSCF0.9 > BiLSCF1.0 > BiLSCF0.8 > BiLSCF1.1. These results also confirm that the best cathode material with a good reactivity and the strongest adsorbate of O2 is BiLSCF0.9. In contrast, the increased b-oxygen can be desorbed with the increase in the concentration of A-site. One can see that the occurrence of b-oxygen desorption is also influenced by the stoichiometry of the A-site ions. However, in this study, the symmetrical cell performance tests are limited only up to 700 C. Therefore, only the a-oxygen desorption property is considered here. Also, it should be noted that the formation of oxygen vacancies starts at ~300 C, and in this case, the electrons can be generated as the following way [33]:
Oo / V::o þ 2e þ 1=2O2 Furthermore, the O2-TPD results also indicate that the oxygen vacancy strongly correlates with the oxygen kinetics in high-
frequency process of BiLSCF cathode working at intermediate temperature range and it is believed to contribute for enhancing the formation of more oxygen vacancies on the surface, as illustrated in Fig. S3y which should be benefit for the ORR process.
3.4. Performance and stability test To further evaluate the cathodic properties, here, the cell based on BiLSCF0.9 cathode with the highest ORR activity was selected, and the LSGM electrolyte-supported single cell with the configuration of Ni-GDCjLSGMjBiLSCF0.9 was constructed. The cell microstructure after the performance and stability test are provided in Fig. 6. The thickness of LSGM electrolyte is about 300 mm (Fig. 6a). The electrodes still maintain porous structure and good connection with the LSGM electrolyte (Fig. 6bee), indicating the good thermal compatibility between electrolyte and electrodes. The BiLSCF0.9 cathode layer, consisting of sub-micron grains (as shown in the magnified morphology in Fig. 6b and c), has an average thickness of about 40 mm (Fig. 6e). The thickness of anode is almost the same as the cathode layer (Fig. 6d). The good connection between cathode-electrolyte, as well as anode-electrolyte, and proper porous structure of the cathode should be benefit for the oxygen diffusion and charge conduction during the electrocatalytic process. It should be noted that a higher porosity is observed clearly in the cathode than in the anode (Fig. 6d and e). Despite the oxidation of hydrogen in the anode is kinetically more favourable than oxygen reduction in the cathode, insufficient porosity of anode can lead to higher anodic activation polarization. This should affect the resistance at high-frequency processes, especially at low temperatures. However, in this study, we mainly consider the properties of BiLSCF cathode. The element diffusion on the interface between BiLSCF and LSGM was further investigated by SEM/EDS line-scan analysis. As shown in Fig. S4ay, no discernible BiLSCF elements diffusion toward LSGM layer is found. One can see that all the major elements are well distributed among the respective layers of BiLSCF and LSGM. This also confirms the good compatibility of BiLSCF/LSGM as in XRD
144
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
values are 1.090, 1.098, 1.106 and 1.115 V at 700, 650, 600 and 550 C, respectively which are very close to the theoretical values (for example, ~1.11 V at 700 C) under the similar condition, indicating that the gas leakage is negligible. It is known that the electronic conduction in the lanthanum gallate-based material occurs by the movement of small polaron via hopping [34], which is a thermally activated process. As such, the contribution of electronic conduction in the electrolyte to the efficiency decreases since any electrons passing through the electrolyte will not contribute to the electrical current in the external circuit. This will affect the cell performance, which is decreased with the decreasing of operating temperature. In this study, the maximum power densities achieved are 0.66, 0.56, 0.44 and 0.31 W cm2 at 700, 650, 600 and 550 C, respectively. Fig. 7b shows the impedance spectra of the cells with BiLSCF0.9 cathode measured under open circuit condition during the performance test. Herein, the polarization and ohmic resistances are separated from the total resistance values as determined from the impedance spectra (inset Fig. 7b) at different operating temperatures. At 700 C, the ohmic contribution (electrolyte resistance) to the total resistance is 58.6%. With the decrease in operating temperature, the ohmic contribution becomes comparatively smaller (Ea-ohmic ¼ 0.45 eV) than the polarization resistance (Ea polarization ¼ 0.73 eV). At 550 C, the relative ohmic contribution to the total resistance is reduced to 42.7%. Considering that the thickness of LSGM electrolyte used in the single-cell is about 300 mm, the resulted cell performance is acceptable. In fact, the improvement in cell performance can be realized by the decrease in electrolyte thickness (to reduce the ohmic resistance). Below 600 C, the polarization contribution to the total resistance becomes greater, and the catalytic activity of both electrodes could be decreased. It should be noted that the reaction rate of hydrogen oxidation at the anode also decreases with the decrease of the operating temperature. As such, the cell performance at low temperature should be mainly dominated by the electrode polarization resistance, which can be reduced by improving the electrode morphology and structure. As shown in Fig. 7c, despite the decline in the first 12 h, which probably due to the activating process, the current density at an output voltage of 0.52 V decreases only about 9% of its original value at the end of 75 h testing, indicating that BiLSCF0.9 has sufficient stability during the operation. Therefore, it has the potential to be used as a cathode material for IT-SOFCs. 4. Conclusions
Fig. 7. (a) Cell voltage (left) and power density (right) as a function of the current density for a cell based on LSGM electrolyte (with 300 mm thickness) supported cell with BiLSCF0.9 cathode, which was measured in humidified H2 as the fuel and stationary air as the oxidant; (b) Resistance contributions (ohmic resistance, polarization resistance and total resistance) as determined from the impedance spectra (inset Fig.) under open circuit conditions during performance test of single cell based on BiLSCF0.9 cathode at 550, 600, 650 and 700 C; (c) Life durability with a constant output voltage of 0.52 V for the cell based on BiLSCF0.9 cathode at 600 C.
analysis (Fig. S1y). The performance of BiLSCF0.9 cathode in SOFC was examined using the LSGM electrolyte support with humidified H2 (3% H2O) as a fuel and ambient air as an oxidant in a temperature range of 550e700 C. Fig. 7a shows the I-V curves and the corresponding power density. Open circuit voltage (OCV) of the cell increases with the decreasing of operating temperature and the obtained OCV
A novel BiLSCFx (x ¼ 0.8, 0.9, 1.0, 1.1) perovskite oxide is successfully designed and synthesized for the first time and can be applied as a potential cathode material for IT-SOFCs. XRD analysis reveals that a pure phase with a cubic perovskite structure is obtained by calcinations at 1150 C for 8 h. Furthermore, the desirable chemical compatibility with LSGM is also realized at this condition. It is found that A-site variation has a significant effect on the electrochemical performance of the BiLSCF oxides. Especially, BiLSCF0.9 with A-site deficiency exhibits excellent molecular oxygen dissociation and electron charge transfer properties on the three phase boundary. As a result, the ASR of the BiLSCF0.9 electrode is low and a high power density of single cell fabricated by using it as cathode is obtained. The maximum power density of LSGM electrolyte (with 300 mm thickness)-supported single cell with BiLSCF0.9 as the cathode reaches the value of 0.66 W cm2 at 700 C by using air as the oxidant. Furthermore, the relatively stable performance is obtained, in which the current density at an output voltage of 0.52 V decreases only ~9% of its original value at the end of 75 h testing. It is expected that BiLSCF0.9 can be applied as a promising candidate material for IT-SOFC cathode.
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
Acknowledgements This work is supported by Aomori City Government, Japan. Deni S. Khaerudini gratefully acknowledges the scholarship from the MEXT of Japan. The authors also thank Professor Takeshi Kubota at NJRISE for his technical support on experiments. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpowsour.2016.10.026. References [1] B.C.H. Steele, A. Heinzel, Materials for fuel-cell technologies, Nature 414 (2001) 345e352. [2] A. Chroneos, B. Yildiz, A. Tarancon, D. Parfitt, J.A. Kilner, Oxygen diffusion in solid oxide fuel cell cathode and electrolyte materials: mechanistic insights from atomistic simulations, Energy Environ. Sci. 4 (2011) 2774e2789. [3] E.P. Murray, T. Tsai, S.A. Barnett, Oxygen transfer processes in (La,Sr)MnO3/ Y2O3-stabilized ZrO2 cathodes: an impedance spectroscopy study, Solid State Ion. 110 (1998) 235e243. [4] J.M. Ralph, C. Rossignol, R. Kumar, Cathode materials for reduced-temperature SOFCs, J. Electrochem. Soc. 150 (2003) A1518eA1522. [5] S. Badwal, Stability of solid oxide fuel cell components, Solid State Ion. 143 (2001) 39e46. [6] P. Hernandez-Fernandez, S. Rojas, P. Ocon, J.L.G. de la Fuente, P. Terreros, J.L. Garcia-Fierro, An opening route to the design of cathode materials for fuel cells based on PtCo nanoparticles, Appl. Catal. B Environ. 77 (1e2) (2007) 19e28. [7] J. Piao, K. Sun, N.Q. Zhang, X.B. Chen, S. Xu, D.R. Zhou, Preparation and characterization of Pr1-xSrxFeO3 cathode material for immediate temperature solid oxide fuel cells, J. Power Sources 172 (2007) 633e640. [8] C. Peters, A. Weber, E. Ivers-Tiffee, Nanoscaled ( La0.5Sr0.5 ) CoO3 d thin film cathodes for SOFC application at 500 C < T < 700 C, J. Electrochem. Soc. 155 (2008) B730eB737. [9] M. Li, Y. Wang, Y.L. Wang, F. Chen, C.R. Xia, Bismuth doped lanthanum ferrite perovskites as novel cathodes for intermediate-temperature solid oxide fuel cells, ACS Appl. Mater. Interfaces 6 (2014) 11286e11294. [10] A. Wedig, R. Merkle, B. Stuhlhofer, H.U. Habermeier, J. Maier, E. Heifets, Fast oxygen exchange kinetics of pore-free Bi1xSrxFeO3d thin films, Phys. Chem. Chem. Phys. 13 (2011) 16530e16533. [11] Y.J. Niu, J. Sunarso, F.L. Liang, W. Zhou, Z.H. Zhu, Z.P. Shao, A comparative study of oxygen reduction reaction on Bi- and La-doped SrFeO3-d perovskite cathodes, J. Electrochem. Soc. 158 (2011) B132eB138. [12] L. Wang, R. Merkle, J. Maier, Surface kinetics and mechanism of oxygen incorporation into Ba1-xSrxCoyFe1-yO3-d SOFC microelectrodes, J. Electrochem. Soc. 157 (2010) B1802eB1808. [13] D.S. Khaerudini, G. Guan, P. Zhang, X. Hao, A. Abudula, Prospects of oxide ionic conductivity bismuth vanadate-based solid electrolytes, Rev. Chem. Eng. 30 (6) (2014) 539e551. [14] D.S. Khaerudini, G. Guan, P. Zhang, X. Hao, Z. Wang, C. Xue, Y. Kasai, A. Abudula, Performance assessment of Bi0.3Sr0.7Co0.3Fe0.7O3-d-LSCF composite as cathode for intermediate-temperature solid oxide fuel cells with La0.8Sr0.2Ga0.8Mg0.2O3-d electrolyte, J. Power Sources 298 (2015) 269e279.
145
[15] J.G. Deng, L. Zhang, H.X. Dai, H. He, C.T. Au, Hydrothermally fabricated singlecrystalline strontium-substituted lanthanum manganite microcubes for the catalytic combustion of toluene, J. Mol. Catal. A Chem. 299 (2009) 60e67. ndez-Díaz, New [16] R. Martínez-Coronado, J.A. Alonso, A. Aguadero, M.T. Ferna SrMo1xCrxO3d perovskites as anodes in solid-oxide fuel cells, Int. J. Hydrogen Energy 39 (2013) 4067e4073. [17] S. McIntosh, J.F. Vente, W.G. Haije, D.H.A. Blank, H.J.M. Bouwmeester, Oxygen stoichiometry and chemical expansion of Ba0.5Sr0.5Co0.8Fe0.2O3d measured in situ diffraction, Chem. Mater. 18 (2006) 2187e2913. [18] X. Yong, J. Song, S.B. Adler, Thermal and chemical expansion of Sr-doped lanthanum cobalt oxide (La1-xSrxCoO3-d, Chem. Mater. 17 (17) (2005) 4537e4546. [19] M. Mori, Y. Hiei, Thermal expansion behavior of titanium-doped La(Sr)CrO3 solid oxide fuel cell interconnects, J. Am. Ceram. Soc. 84 (11) (2001) 2573e2578. [20] J. Pena-Martınez, D. Marrero-Lopez, J.C. Ruiz-Morales, P. Nunez, C. SanchezBautista, A.J. Dos Santos-Garcia, J. Canales-Vazquez, On Ba0.5Sr0.5Co1-yFeyO3d (y ¼0.1e0.9) oxides as cathode materials for La0.9Sr0.1Ga0.8Mg0.2O2.85 based IT-SOFCs, Int. J. Hydrogen Energy 34 (2009) 9486e9495. [21] D. Chen, R. Ran, K. Zhang, J. Wang, Z. Shao, Intermediate-temperature electrochemical performance of a polycrystalline PrBaCo2O5þd cathode on samarium-doped ceria electrolyte, J. Power Sources 188 (2009) 96e105. [22] N. Grunbaum, L. Mogni, F. Prado, A. Caneiro, Phase equilibrium and electrical conductivity of SrCo0.8Fe0.2O3d, J. Solid State Chem. 177 (2004) 2350e2357. [23] Z.H. Du, H.L. Zhao, Y.N. Shen, L. Wang, M.Y. Fang, K. Swierczek, K. Zheng, Evaluation of La0.3Sr0.7Ti1-xCoxO3 as a potential cathode material for solid oxide fuel cells, J. Mater. Chem. A 2 (2014) 10290e10299. [24] B. Philippeau, F. Mauvy, C. Mazataud, S. Fourcade, J.C. Grenier, Comparative study of electrochemical properties of mixed conducting Ln2NiO4 þ d (Ln ¼ La, Pr and Nd) and La0.6Sr0.4Fe0.8Co0.2O3-d as SOFC cathodes associated to Ce0.9Gd0.1O2- d, La0.8Sr0.2Ga0.8Mg0.2O3-d and La9Sr1Si6O26.5 electrolytes, Solid State Ion. 249e250 (2013) 17e25. [25] Y.N. Shen, H.L. Zhao, K. Swierczek, Z.H. Du, Z.X. Xie, Lattice structure, sintering behavior and electrochemical performance of La1.7Ca0.3Ni1-xCuxO4þd as cathode material for intermediate temperature solid oxide fuel cell, J. Power Sources 240 (2013) 759. [26] Z.B. Yang, C.H. Yang, B. Xiong, M.F. Han, F.L. Chen, BaCo0.7Fe0.2Nb0.1O3-d as cathode material for intermediate temperature solid oxide fuel cells, J. Power Sources 196 (2011) 9164e9168. [27] Z.B. Yang, C.H. Yang, C. Jin, M.F. Han, F.L. Chen, Ba0.9Co0.7Fe0.2Nb0.1O3-d as cathode material for intermediate temperature solid oxide fuel cells, Electrochem. Commun. 13 (2011) 882e885. [28] G.H. Chen, Y. Wang, J. Sunarso, F.L. Liang, H.P. Wang, A new scandium and niobium co-doped cobalt-free perovskite cathode for intermediatetemperature solid oxide fuel cells, Energy 95 (2016) 137e143. [29] M. Liu, D. Ding, K. Blinn, X. Li, L. Nie, M. Liu, Enhanced performance of LSCF cathode through surface modification, Int. J. Hydrogen Energy 37 (2012) 8613e8620. [30] M. Sahibzada, S.J. Benson, R.A. Rudkin, J.A. Kilner, Pd-promoted La0.6Sr0.4Co0.2Fe0.8O3 cathodes, Solid State Ion. 113e115 (1998) 285e290. [31] Y. He, X. Zhu, W. Yang, The role of A-site ion nonstoichiometry in the oxygen absorption properties of Sr1þxCo0.8Fe0.2O3 oxides, AlChE J. 57 (2011) 87e95. [32] Y. Teraoka, M. Yoshimatsu, N. Yamazoe, T. Seiyama, Chem. Lett. (1984) 893e896. [33] J.I. Jung, S.T. Misture, D.D. Edwards, The electronic conductivity of Ba0.5Sr0.5CoxFe1xO3d (BSCF: x ¼ 0 ~ 1.0) under different oxygen partial pressures, J. Electroceram. 24 (2010) 261e269. [34] T. Ishihara, H. Furutani, T. Yamada, Y. Takita, Oxide ion conductivity of double doped lanthanum gallate perovskite type oxide, Ionics 3 (1997) 209e213.
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
(Bi0.15La0.27Sr0.53)(Co0.25Fe0.75)O3-d perovskite: A novel cathode material for intermediate temperature solid oxide fuel cells Deni S. Khaerudini a, c, Guoqing Guan a, b, *, Peng Zhang a, Pairuzha Xiaoketi a, Xiaogang Hao d, Zhongde Wang d, Yutaka Kasai e, Abuliti Abudula a, ** a
Graduate School of Science and Technology, Hirosaki University, 1-Bunkyocho, Hirosaki 036-8560, Japan North Japan Research Institute for Sustainable Energy (NJRISE), Hirosaki University, 2-1-3 Matsubara, Aomori 030-0813, Japan Research Center for Physics, Indonesian Institute of Sciences, Gd. 440 Kawasan Puspiptek Serpong, Tangerang Selatan 15314, Banten, Indonesia d Department of Chemical Engineering, Taiyuan University of Technology, Taiyuan 030024, China e Industrial Research Institute, Aomori Prefectural Industrial Technology Research Center, 4-11-6, Second Tonyamachi, Aomori 030-0113, Japan b c
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
(Bi0.15La0.27Sr0.53)x(Co0.25Fe0.75)O3-d is implemented as cathode in IT-SOFC. BiLSCF0.9 is chemically compatible with LSGM and shows the best performance. A low polarization resistance of 0.039 U cm2 for BiLSCF0.9 is achieved at 700 C. SOFC with BiLSCF0.9 shows the maximum power of 0.66 W cm2 at 700 C. Polarization loss is minimized by optimization of compositions and structure.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 15 June 2016 Received in revised form 15 September 2016 Accepted 7 October 2016
Perovskite oxides (Bi0.15La0.27Sr0.53)x(Co0.25Fe0.75)O3-d (BiLSCFx, x ¼ 0.8, 0.9, 1.0, 1.1) have been synthesized by solid state reaction and evaluated as a novel cathode material for intermediate-temperature solid oxide fuel cells (IT-SOFCs). The effects of A-site variations on lattice structure, calcination temperature, oxygen desorption and electrochemical properties of BiLSCFx are investigated. This kind of material has perfectly cubic structure based on the Pm-3m space group whose lattice size increases with x, which is thermally stable after calcination and shows desirable chemical compatibility with La0.8Sr0.2Ga0.8Mg0.2O3-d electrolyte at 1150 C for 8 h under air atmosphere. Among those A-site variations, it is found that BiLSCF0.9 demonstrates the best cathode performance. It has the minimum polarization resistance value of 0.039 U cm2 at 700 C and a-oxygen desorbed about 0.031 mmol g1, indicating a good reactivity and strong adsorbate of O2. The single cell with BiLSCF0.9 cathode delivers a power density of 0.66 W cm2 at 700 C with humidified H2 (~3% H2O) as the fuel and ambient air as the
Keywords: IT-SOFC BiLSCF Cathode Oxygen reduction reaction Electrochemical performance
* Corresponding author. North Japan Research Institute for Sustainable Energy (NJRISE), Hirosaki University, 2-1-3 Matsubara, Aomori 030-0813, Japan. ** Corresponding author. E-mail addresses: [email protected] (G. Guan), [email protected] (A. Abudula). http://dx.doi.org/10.1016/j.jpowsour.2016.10.026 0378-7753/© 2016 Elsevier B.V. All rights reserved.
138
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
oxidant. In addition, the cell shows sufficient stability with ~9% degradation over 75 h at 600 C. It indicates that BiLSCF0.9 is a promising candidate for application as cathode material in IT-SOFCs. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Solid oxide fuel cells (SOFCs) are the electrochemical power generation devices that convert chemical energy into electrical energy directly, providing several advantages of fuel flexibility, low pollutant emission, environmental friendliness, high power density, and high efficiency [1]. Recently, long-term durability has been the major concern for the commercialization of SOFC systems. One of methods to mitigate this issue is to decrease the operation temperature to the low or intermediate temperature range (500e700 C) through the application of advanced electrolyte and/ or electrode materials. That is to develop intermediate temperature-SOFCs (IT-SOFCs) since it is potential to enhance the stability and economic feasibility of SOFCs at lower operating temperature. However, at low and intermediate temperature range, the kinetics for oxygen transfer in cathode becomes a major limiting step for good performance [2]. The effective oxygen reduction reaction (ORR) area in the strontium-doped lanthanum manganite (LSM) with perovskite structure has proven to be unfavorable at low temperatures, (223 U cm2 at 650 C) [3] and (1100 U cm2 at 650 C) [4], when compared with those doped by cobalt, chromium and iron oxides [5e7], since LSM has significantly inferior kinetics for oxygen transfer at the electrolyte interface [5]. The perovskite LaCoO3 is also recognized as a candidate material on IT-SOFC cathodes, and especially the aliovalent element substituted LaCoO3 such as La12 xSrxCoO3 shows promising electrochemical activity (0.2 U cm at 600 C) towards the oxygen reduction reaction [8]. B-site doping with various transition metals has been investigated extensively and the composition La0.6Sr0.4Co0.2Fe0.8O3-d is often referred as one of functional cathodes due to its suitable combination of performance, stability, reliability, thermo-mechanical and chemical compatibility with other cell components. Recently, Bi replacement of A-site element in these perovskite materials has stimulated the exploration of their application as the cathode in IT-SOFC [9e11]. Especially, by Bi substitution, different electronic configuration could be generated since Bi3þ has higher polarizability owing to its 6s2 lone pair and lower basicity in contrast to those observed in Ba2þ (Ba1-xSrxFe1-yCoyO3-d) [12] as well as La3þ based materials. Thus, it is expected to increase the structural stability, offer higher concentration and mobility of oxygen vacancies at the reduced temperatures, and promote the catalytic activity in oxygen dissociation process, which is often the rate limiting step in oxygen reduction in IT-SOFC cathode [13]. In our previous study [14], LSCF50-BiSCF50 composite material was made by mixing self-made Bi0.3Sr0.7Co0.3Fe0.7O3-d (BiSCF3737) powder and commercial (La0.6Sr0.4)0.9Co0.2Fe0.8O3±d (LSCF) powder and used as cathode material of IT-SOFC. A significant performance improvement, specifically on oxygen dissociation process, was found with almost one order of magnitude higher than BiSCF3737 electrode and two orders of magnitude higher than LSCF electrode. Based on these results, in this study, we tried to synthesize a single phase material containing the same element contents in LSCF50BiSCF50 composite as a novel cathode material for IT-SOFCs. To the best our knowledge, BiLSCFx materials have not been reported to date. It is expected to find some better catalytic activity and structural stability than those in the previous composite
configuration. Here, we explored the structure, electrochemical properties, oxygen desorption and chemical reactivity of such single-phase cathodes with the perovskite framework, i.e., single phase (Bi0.15La0.27Sr0.53)x (Co0.25Fe0.75)O3-d compounds with x ¼ 0.8, 0.9, 1.0, 1.1, called as BiLSCF0.8, BiLSCF0.9, BiLSCF1.0 and BiLSCF1.1, respectively. Their performances in the electrolyte-supported SOFCs operating at temperatures ranged from 550 to 700 C were further investigated. 2. Experimental 2.1. Preparation of cathode materials The
original
powder of
(Bi0.15La0.27Sr0.53)x(Co0.25Fe0.75)O3-
d (x ¼ 0.8, 0.9, 1.0, 1.1) was synthesized using solid state reaction
route from the high-purity starting chemicals of Bi(NO3)3$5H2O, La(NO3)3$6H2O, Sr(NO3)2, Co(NO3)2$6H2O, Fe(NO3)3$9H2O, and Ce(NO3)3$6H2O (99.9%, Wako, Japan) with exactly stoichiometric molar ratio. These powders were mixed and ball-milled at a rotation speed of 300 rpm for 6 h in ethanol media. The resulting precursor solution was evaporated at 80 C for 24 h and thoroughly ground in an agate mortar and subsequently calcined at different temperatures from 1050 to 1150 C for 8 h in air to obtain fine BiLSCF powders. Similarly, Ce0.9Gd0.1O1.95 (GDC) as a composite anode material was also prepared by the solid state reaction from the appropriate proportion of Ce and Gd nitrates (99.9%, Wako, Japan) mixed with ethanol by ball milling. Then the mixture was evaporated and calcined at 900 C for 5 h in air to gain the GDC powder. The commercial NiO (Soekawa Chemicals, Japan) and La0.8Sr0.2Ga0.8Mg0.2O3-d (LSGM, FCM, USA) powders were used as received for the preparations of anode and electrolyte, respectively. To assess the phase reaction, the chemical compatibility of the BiLSCF cathode and LSGM electrolyte was investigated with X-ray diffraction (XRD) analysis by sintering the mixed powders in a weight ratio of 1:1 at the optimum calcination condition for 8 h. 2.2. Microstructural and morphological characterizations A Rigaku SmartLab X-Ray Diffractometer with Cu-Ka (l ¼ 1.5405 Å) radiation source was used to examine the phase composition and lattice structure of the sample. Powder samples were scanned with 2q ranging from 20 to 80 with a 5 /min scan rate. Structural refinement was fitted using the PDXL2.0 Rigaku Data Analysis Software package. A Hitachi SU8010 scanning electron microscope (SEM) equipped with a Horiba Scientific energy dispersive spectrometer (EDS) analyzer was used to investigate the microstructure of electrode and elemental distribution across cathode-electrolyte interface and the possible secondary phase which cannot be detected by XRD. 2.3. Oxygen reducibility test Oxygen desorption properties of different samples as a function of temperature were monitored by a Belcat oxygen temperatureprogrammed desorption (O2-TPD) catalyst analyzer (Belcat, Bel Japan, Inc.). In general, about 0.1 g of powder sample was placed in a quartz fixed-bed microreactor with 8 mm inner diameter. Before
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
each run, the sample was treated in O2 with a flow rate of 20 cm3 min1 at 500 C for 1 h to confirm that O2 can be adsorbed and activated completely across the bulk structure of the sample as those reported in other work [15]. Then, it followed by cooling to RT in O2 and argon purging with a flow rate of 50 cm3 min1 for 30 min. The purpose of argon purging was to remove gas-phase oxygen in the system. The sample was then heated from RT to 900 C with a heating rate of 10 C min1. The effluent was monitored by a thermal conductivity detector (TCD). The amount of O2 desorbed from the sample was quantified by calibrating the peak area against that of a standard O2 pulse. 2.4. Fabrication and testing of cells Electrochemical performance of the as-prepared cathode was evaluated with both of the symmetrical cells and electrolyte supported single cells. Dense LSGM as the electrolyte had a diameter of ~20 mm, which was prepared through a dry pressing method at 250 MPa and sintered at 1450 C for 10 h. Symmetrical cells were employed to investigate the electrochemical impedance behaviour. The as-synthesized powders were mixed with 80 wt% organic additives (a-terpineol þ ethyl cellulose) solution to form slurry. The electrode slurry was tape-casted onto both sides of the LSGM pellet with a thickness of ~0.5 mm. After drying, the electrode was sintered at 1150 C for 2 h in air to form the symmetrical cell. Single cell studied in this work was LSGM electrolyte-supported (~300 mm thickness) cell. The anode layer consisting of 40 wt% GDC þ 60 wt% NiO composite was prepared by tape casting the anode slurry onto the LSGM electrolyte and sintered at 1350 C for 2 h. Subsequently, cathode slurry was casted on the opposite side of the electrolyte, followed by sintering with the similar conditions as in the symmetrical cells to form the single cell. The effective area of electrode was 0.78 cm2. Pt paste/mesh was used as the current collector for all electrochemical evaluations. According to our previous and other work [14,16], the Pt current collector should have no obvious influence on the electrode catalytic activity. The electrode polarization resistance of the symmetrical cell was performed using an electrochemical workstation composed of a Solartron 1255B frequency response analyzer and SI-1287 galvanostat. The cell was measured at the temperatures ranged from 550 to 700 C (50 C intervals) under air atmosphere. The performance of cathode was further examined in the single fuel cell configuration. The cell was sealed on the top of an alumina support tube by glass ring. The anode was firstly reduced in situ from NiO/ GDC to Ni/GDC cermet anodes. 100 cm3 min1 humidified H2 (ca. 3% H2O) and 50 cm3 min1 ambient air were employed as the fuel and oxidant, respectively. The measurements were also carried out at temperatures ranged from 550 to 700 C (50 C intervals). Performance stability test was performed under a constant voltage of 0.52 V for 75 h. The performance characteristics of the cell and electrochemical impedance spectra were measured under opencircuit conditions over a frequency range from 0.1 Hz to 1 MHz with an AC perturbation of 10 mV and fitted using Z-view software. 3. Results and discussion 3.1. Phase structure The phase purity and composition of (Bi0.15La0.27Sr0.53)1.0 (Co0.25Fe0.75)O3-d (only the pattern of BiLSCF1.0 as an example is shown here) after calcined at 1050, 1100, and 1150 C for 8 h in air was confirmed by XRD. As shown in Fig. 1a, at 1050 and 1100 C, some impurity phases are found in the XRD patterns. Increasing the calcination temperature not only sharpens the detected diffraction peaks, but also increases the degree of crystallinity. It is obvious
139
that only reflections corresponding to the perovskite phase can be detected for the cathode calcined at 1150 C, which is the optimum sintering temperature for the preparation of BiLSCF cathode, and in this case, the single phase compound can be formed. Fig. 1b shows XRD patterns of samples of (Bi0.15La0.27 Sr0.53)x(Co0.25Fe0.75)O3 d (x ¼ 0.8, 0.9, 1.0, 1.1) calcined at 1150 C for 8 h. One can see that all the samples are crystallized as the cubic perovskite structure in the space group of Pm-3m. Table S1y shows the space groups and lattice parameters of these compounds. One can see that various A-site compositions have no effect on the formation of the pure cubic symmetry perovskite phase. With the increasing of x value on Asite, the XRD diffraction peak slightly shifts to the left as confirmed by the magnified main peak (110) of BiLSCF sample, indicating a lattice expansion. A simple way to attempt to explain the expansion is to consider the expansion of A/B distance as the variation in the averaged ionic radius of A-site cation compositions [17]. For simplicity, it is assumed that the oxide crystallizes in a cubic perovskite-type structure without any displacement of atoms, and the lattice volume is governed by the averaged ionic radius of the cation. It is also supposed that there is no preferred orientation in the specimen. This expansion also has potential advantages for the creation of oxygen vacancies [18]. Furthermore, the incorporation of Bi into A-site cation sublattice may further generate a large concentration of oxygen vacancies, due to the charge balance as shown in the following Kroger-Vink notation: ðLa;SrÞðCo;FeÞO3
Bi2 O3 !V::o þ 2Bi0La þ 3Oxo Mobility of oxygen ion is increased and subsequently results in the increase in oxygen ion conductivity, however, this can also lead to the loss of other desirable properties, such as the electrical conductivity [19]. Deficiency/excess of A-site can therefore be used to tune the conductivity (mixed ionic-electronic conductivity) of the material, which will be addressed later. Furthermore, lattice expansion is also an important factor in the design or selection of cathode material. The surface microstructure and elemental composition of the powder BiLSCF1.0 were characterized by SEM/EDS analysis. As shown in Fig. 1c and d, after calcined at 1150 C for 8 h, the grain size distribution is in the range of 1e4 mm. EDS analysis results reveal that no any impurities are included. The inset table in Fig. 1d shows the measured elemental composition. Obviously, the Bi:La:Sr:Co:Fe atomic ratio is 0.1498:0.2721:0.5320:0.2568:0.7500 which is very close to the nominal stoichiometric composition of 0.15:0.27:0.53: 0.25:0.75, indicating that the stoichiometry in the solid is relatively homogenous. To obtain the detail information about the crystal structure of BiLSCF, Rietveld refinement was applied. Here, only the result on the BiLSCF0.9 was selected because it has the highest catalytic activity (as indicated in the next section). The calculated and experimental XRD profiles of BiLSCF0.9 are shown in Fig. 2, which is also in agreement with the calculated Goldschmidt tolerance factor (t) about 0.976, indicating that the cubic structure is formed. The reliability factors obtained from the Rietveld refinement are Rp z 4.64, Rwp z 5.58, Rexp z 2.04, RBragg z 2.61, and c2 z 2.38, demonstrating the good agreement between the calculated and observed data. Based on ABO3 model, A-site atoms of Bi/La/Sr sit at cube corner (0, 0, 0) position, B-site atoms of Fe/Co sit at body center (1/2, 1/2, 1/2) position, and O atoms sit at face centered (0, 0, 1/2) position with random oxygen vacancy, as a result of highly polarizable of Bi3þ ions caused by its 6s2 lone pair electrons and weak BieO bonds structure, in a single oxygen site. Generally, such a high oxygen vacancy is beneficial for the free movement of oxygen ions and supplies the active sites for the oxygen adsorption, dissociation and diffusion at the surface of cathode. Therefore, this
140
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
Fig. 1. (a) XRD patterns of (Bi0.15La0.27Sr0.53)(Co0.25Fe0.75)O3-d (BiLSCF) powders after calcinations at 1050, 1100 and 1150 C for 8 h and X-ray diffraction patterns of the respective powders in the range of 25 2q 40 , (b) XRD of BiLSCF (x ¼ 0.8e1.1) powders calcined at 1150 C for 8 h in air and the right Fig. shows the magnified (110) peak of the respective powders, (c) SEM micrograph and (d) EDS analysis of BiLSCF1.0 after calcined at 1150 C for 8 h in air. The inset in (d) shows the table of comparison of EDS results with the targeted values.
BiLSCF and LSGM appear, and no significant impurities are identified under the detection limit from the XRD patterns. These indicate that no reaction occurred between BiLSCF and LSGM upon 1150 C for 8 h, therefore the effect of impurities on the electrochemical performance from sintering process is negligible. 3.2. Effect of cathode composition ratio on oxygen reduction reaction
Fig. 2. Rietveld fit of XRD data of as-prepared (Bi0.135La0.243Sr0.477)(Co0.25Fe0.75)O3d (BiLSCF0.9) powders. The series with blue thick marks corresponds to the allowed Bragg reflections. Inset Fig. shows unit cell of the crystal structure of BiLSCF0.9, the O site is partially occupied by oxygen vacancy. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
material should perform superior electrochemical performance. In general, the interface reaction between electrolyte and cathode is highly detrimental for the cell performance due to the increase of interfacial polarization resistance. Here, Fig. S1y shows XRD patterns of BiLSCF1.0, LSGM and BiLSCFx þ LSGM powders prepared by the direct calcinations of the mixture of BiLSCF and LSGM powders in 50:50 wt% at 1150 C for 8 h in air. After calcination, it shows that only those peaks corresponding to original
The oxygen reduction reaction (ORR) activity of the BiLSCF (x ¼ 0.8, 0.9, 1.0, 1.1) cathodes was quantified by electrochemical impedance spectroscopy (EIS) using symmetrical cells with LSGM electrolyte as support (~0.5 mm thickness). In order to properly compare the intrinsic material activity, measurements are carried out when the porous cathode morphologies of each cell variations are similar. From Fig. 3, it can be seen clearly that the grain size and pore morphology of all cathodes are comparable. Here, the variations in impedance behaviour can therefore be addressed mainly due to the composition. Fig. 4 shows Nyquist plots between 550 and 700 C in air at open circuit condition. The area specific resistance (ASR) of the cathode is obtained from the difference between the low-frequency and high-frequency intercepts on the real axis and represents the polarization resistance (Rp) of the cathode. The best fit equivalent circuit for the analysis of the impedance data is shown in the inset of Fig. 4b with two parallel connected resistances (R) and a constant phase element (CPE) in series. In this model, the ohmic resistance is set at zero (omitted) in the Nyquist plots as it is mainly contributed by the electrolyte. The highfrequency arc (RHF), specifically observed from the mid-frequency (103-102 Hz), corresponds to the ionic transport resistance [20];
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
141
Fig. 3. Cross-sectional SEM morphologies of cathodes on LSGM electrolyte with (a) BiLSCF0.8, (b) BiLSCF0.9, (c) BiLSCF1.0 and (d) BiLSCF1.1 in symmetrical fuel cells.
Fig. 4. Nyquist plots of impedance spectra for (a) BiLSCF0.8, (b) BiLSCF0.9, (c) BiLSCF1.0 and (d) BiLSCF1.1 cathodes tested at a temperature range from 550 to 700 C in air. The inset in (b) is the equivalent circuit used for all data fitting.
and the low-frequency (101-101 Hz) arc (RLF) represents the surface diffusion of oxygen species including the oxygen adsorption and dissociation processes [21]. From the impedance spectra at 700 C, except for BiLSCF1.1, it can be seen clearly that the high-frequency responses of cathodes show significantly higher contribution, about 70%, than the lowfrequency ones, as shown by the representative of impedance plot in the inset of Fig. 4c. Meanwhile, BiLSCF1.1 shows the opposite
behaviour under the same condition; where the high-frequency responses only contribute approximately 17% of the total resistance, as shown in the inset of Fig. 4d. It seems that the impedance behaviour of BiLSCF1.1 can be rationalised mainly by the excess of Bi on A-site which probably provides more dominant ionic contribution than other A-site values. However, at temperature lower than 700 C, all of cathodes show similar behaviour. The performance of high-frequency is significantly lower than that of
142
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
Table 1 Deconvoluted high- and low-frequency polarization resistance (U cm2) of BiLSCF0.8, BiLSCF0.9, BiLSCF1.0 and BiLSCF1.1 in air at different temperatures from 550 to 700 C. Cathode
T ( C)
Rp (U cm2)
RHF (U cm2)
RLF (U cm2)
BiLSCF0.8
550 600 650 700 550 600 650 700 550 600 650 700 550 600 650 700
0.359 0.160 0.085 0.048 0.219 0.108 0.058 0.039 0.248 0.118 0.064 0.041 0.798 0.451 0.239 0.137
0.071 0.039 0.021 0.037 0.036 0.020 0.011 0.031 0.059 0.031 0.021 0.025 0.105 0.069 0.036 0.023
0.288 0.121 0.063 0.012 0.182 0.088 0.047 0.008 0.189 0.087 0.043 0.016 0.693 0.381 0.203 0.114
BiLSCF0.9
BiLSCF1.0
BiLSCF1.1
low-frequency, as listed in Table 1; indicating that the incorporation of Bi into cathode structure is more effective on ionic transfer at the temperature lower than 700 C, and the high-frequency performance is controlled mainly by the appropriate Bi incorporation on A-site cation composition. Among all the cathodes investigated, BiLSCF0.9 performs well not only for the charge transfer at high-frequency process but also for the oxygen adsorption and dissociation at the low frequency process. The appropriate composition of Bi in the cathode structure provides more channels to transport oxygen ions via increasing the concentration of total oxygen vacancies. This will be explained further on the separated Section in 3.3. Hence, the oxygen reduction can occur much more at three-phase boundaries, and as a result, more oxygen ions may diffuse into the LSGM electrolyte to improve the whole oxygen diffusion. Fig. S2ay shows typical Arrhenius plots of the total interfacial polarization resistance (Rp) values obtained from the data of Fig. 4 as a function of reciprocal temperature, together with activation energies (Ea) of BiLSCF (x ¼ 0.8, 0.9, 1.0, 1.1) cathodes. It clearly shows that the Rp value decreases significantly with the increase in temperature, which can be attributed to the enhanced catalytic activity at higher temperature. Such temperature-dependent behaviour is typical for the mixed conducting perovskite [22]. While BiLSCF1.1 cathode exhibits an Ea value of 0.92 eV, BiLSCF0.9 provides a lower value of 0.8 eV that is deemed more favourable for obtaining high performance at lower temperatures, as also can be evidenced by their Rp values. As shown in Table 1, the Rp values of all prepared cathodes, BiLSCF0.8, BiLSCF0.9, BiLSCF1.0 and BiLSCF1.1, are 0.048, 0.039, 0.041 and 0.137 U cm2 at 700 C, respectively, in which BiLSCF0.9 delivers the lowest Rp values at all temperatures tested. Moreover, it is also comparable and lower than other reported work on high-performance perovskite cathodes such as La0.3Sr0.7Ti0.4Co0.6O3-d (0.093 U cm2 at 700 C) [23], La0.6Sr0.4Fe0.8Co0.2O3-d (0.1 U cm2 at 700 C) [24], La1.7Ca0.3Ni0.7Cu0.3O4þd (1.45 U cm2 at 650 C) [25], BaCo0.7Fe0.2 Nb0.1O3-d (0.12 U cm2 at 700 C) [26], Ba0.9Co0.7Fe0.2Nb0.1O32 d (0.07 U cm at 700 C) [27], SrSc0.175Nb0.025Fe0.8O3-d (0.040 U cm-2 at 700 C) [28], La0.4875Ca0.0125Ce0.5O2-d thin filmed on LaxSr12 xCoyFe1-yO3-d (0.184 U cm at 700 C) [29], La0.6Sr0.4Co0.2Fe0.8O32 Pd (<0.05 U cm at 700 C) [30], indicating its potential for application in IT-SOFCs. The electrode polarization resistances (Rp) of BiLSCF cathodes obtained from the impedance spectra as functions of the value of x and temperature are plotted in Fig. S2by. For A-site deficient
samples, the polarization resistances slightly decrease with the increasing of x from 0.8 to 0.9 at the temperatures ranged between 550 and 700 C and the minimum Rp is obtained when x ¼ 0.9. With further increase of x value to 1.0, the Rp of BiLSCF cathode begins to increase. Obviously, when x is 1.1, the Rp increases rapidly to several times higher than those of A-site deficient and stoichiometry samples under the same test condition. This can be explained by the significant difference between low-frequency arc (~15%) and high-frequency arc (~85%) contributions of BiLSCF1.1 when compared with other BiLSCF compositions, as shown in Fig. 4. It indicates that ionic contribution in this cathode is probably much more dominant than electronic properties. At higher operation temperature, the value of x corresponding to the lowest Rp is slightly lower than that at relatively lower temperature. These results demonstrate that the polarization resistance of BiLSCF is dependent not only on the A-site nonstoichiometry with either Asite excess or deficiency and the operating temperature, but also on gas transport to and away from the reaction zone. At higher operating temperature, the reaction is highly activated; on the contrary, at relatively low temperature, gas transport might be limited. 3.3. Oxygen sorption analysis The interaction between the perovskite surface and adsorbed molecules, for instance, the sorption of oxygen on the oxygen ion vacancies, can be further evaluated using temperature programmed desorption of oxygen (O2-TPD) technique. The increase of oxygen vacancies concentration has ever confirmed by the reduction of a-oxygen amount desorbed in the low temperature range in the O2-TPD experiment [31]. The O2-TPD profiles could provide information on the behaviour of lattice oxygen. Fig. 5 depicts the O2-TPD profiles of BiLSCF (x ¼ 0.8, 0.9, 1.0, 1.1) powders after exposure to O2 for 1 h at 500 C. All samples exhibit two desorption peaks (a and b oxygen) which are closely related to the formation of oxygen vacancies and the desorption of oxygen from the surface lattice, which are dependent on the composition as well as the lattice structure of the material. The a desorption, locating at a strong peak from ca. 300e600 C, is ascribed to the oxygen vacancies formed, originated from Oads located at oxygen vacancies
Fig. 5. O2-TPD profiles of (a) BiLSCF0.8, (b) BiLSCF0.9, (c) BiLSCF1.0 and (d) BiLSCF1.1with argon as sweep gas.
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
143
Fig. 6. Typical SEM morphologies for the cell after the operating test: (a) cross section view of electrolyte-supported Ni-GDCjLSGMjBiLSCF0.9 cell configuration, (b,e) cross section views of BiLSCF0.9/LSGM interface, (c) surface view of BiLSCF0.9 cathode (d) cross section view of Ni-GDC/LSGM interface.
[32]. The a-oxygen is related to the formation of Co4þ ions or positive holes. Here, the desorption of a-oxygen may be expressed as:
Oxo þ 2h_ /V::o þ 1=2O2 On the other hand, the b-oxygen, originated from the surface oxygen lattice (O2 latt) desorbs at a commencement temperature of approximately 800 C, is accompanied by the reduction of Co ion to a lower valence state, described as:
Oxo þ 2CoxCo /V::o þ 2Co0Co þ 1=2O2 For all BiLSCF samples, the peaks of a desorption exhibit much stronger than those of b desorption, indicating that a desorption is more effective to the A-site variations than b desorption. As such, its intensity can provide a rough estimation of the amount of oxygen vacancies on the surface. Here, the area of the a-oxygen desorption peak of BiLSCF1.1 is the largest, followed by BiLSCF0.8, BiLSCF1.0 and then BiLSCF0.9. The amounts of a-oxygen adsorbed on BiLSCF1.1, BiLSCF0.8, BiLSCF1.0 and BiLSCF0.9 are estimated to be 0.059, 0.045, 0.038 and 0.031 mmol g1, respectively, indicating that the oxygen vacancy concentration decreases in the order of BiLSCF0.9 > BiLSCF1.0 > BiLSCF0.8 > BiLSCF1.1. These results also confirm that the best cathode material with a good reactivity and the strongest adsorbate of O2 is BiLSCF0.9. In contrast, the increased b-oxygen can be desorbed with the increase in the concentration of A-site. One can see that the occurrence of b-oxygen desorption is also influenced by the stoichiometry of the A-site ions. However, in this study, the symmetrical cell performance tests are limited only up to 700 C. Therefore, only the a-oxygen desorption property is considered here. Also, it should be noted that the formation of oxygen vacancies starts at ~300 C, and in this case, the electrons can be generated as the following way [33]:
Oo / V::o þ 2e þ 1=2O2 Furthermore, the O2-TPD results also indicate that the oxygen vacancy strongly correlates with the oxygen kinetics in high-
frequency process of BiLSCF cathode working at intermediate temperature range and it is believed to contribute for enhancing the formation of more oxygen vacancies on the surface, as illustrated in Fig. S3y which should be benefit for the ORR process.
3.4. Performance and stability test To further evaluate the cathodic properties, here, the cell based on BiLSCF0.9 cathode with the highest ORR activity was selected, and the LSGM electrolyte-supported single cell with the configuration of Ni-GDCjLSGMjBiLSCF0.9 was constructed. The cell microstructure after the performance and stability test are provided in Fig. 6. The thickness of LSGM electrolyte is about 300 mm (Fig. 6a). The electrodes still maintain porous structure and good connection with the LSGM electrolyte (Fig. 6bee), indicating the good thermal compatibility between electrolyte and electrodes. The BiLSCF0.9 cathode layer, consisting of sub-micron grains (as shown in the magnified morphology in Fig. 6b and c), has an average thickness of about 40 mm (Fig. 6e). The thickness of anode is almost the same as the cathode layer (Fig. 6d). The good connection between cathode-electrolyte, as well as anode-electrolyte, and proper porous structure of the cathode should be benefit for the oxygen diffusion and charge conduction during the electrocatalytic process. It should be noted that a higher porosity is observed clearly in the cathode than in the anode (Fig. 6d and e). Despite the oxidation of hydrogen in the anode is kinetically more favourable than oxygen reduction in the cathode, insufficient porosity of anode can lead to higher anodic activation polarization. This should affect the resistance at high-frequency processes, especially at low temperatures. However, in this study, we mainly consider the properties of BiLSCF cathode. The element diffusion on the interface between BiLSCF and LSGM was further investigated by SEM/EDS line-scan analysis. As shown in Fig. S4ay, no discernible BiLSCF elements diffusion toward LSGM layer is found. One can see that all the major elements are well distributed among the respective layers of BiLSCF and LSGM. This also confirms the good compatibility of BiLSCF/LSGM as in XRD
144
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
values are 1.090, 1.098, 1.106 and 1.115 V at 700, 650, 600 and 550 C, respectively which are very close to the theoretical values (for example, ~1.11 V at 700 C) under the similar condition, indicating that the gas leakage is negligible. It is known that the electronic conduction in the lanthanum gallate-based material occurs by the movement of small polaron via hopping [34], which is a thermally activated process. As such, the contribution of electronic conduction in the electrolyte to the efficiency decreases since any electrons passing through the electrolyte will not contribute to the electrical current in the external circuit. This will affect the cell performance, which is decreased with the decreasing of operating temperature. In this study, the maximum power densities achieved are 0.66, 0.56, 0.44 and 0.31 W cm2 at 700, 650, 600 and 550 C, respectively. Fig. 7b shows the impedance spectra of the cells with BiLSCF0.9 cathode measured under open circuit condition during the performance test. Herein, the polarization and ohmic resistances are separated from the total resistance values as determined from the impedance spectra (inset Fig. 7b) at different operating temperatures. At 700 C, the ohmic contribution (electrolyte resistance) to the total resistance is 58.6%. With the decrease in operating temperature, the ohmic contribution becomes comparatively smaller (Ea-ohmic ¼ 0.45 eV) than the polarization resistance (Ea polarization ¼ 0.73 eV). At 550 C, the relative ohmic contribution to the total resistance is reduced to 42.7%. Considering that the thickness of LSGM electrolyte used in the single-cell is about 300 mm, the resulted cell performance is acceptable. In fact, the improvement in cell performance can be realized by the decrease in electrolyte thickness (to reduce the ohmic resistance). Below 600 C, the polarization contribution to the total resistance becomes greater, and the catalytic activity of both electrodes could be decreased. It should be noted that the reaction rate of hydrogen oxidation at the anode also decreases with the decrease of the operating temperature. As such, the cell performance at low temperature should be mainly dominated by the electrode polarization resistance, which can be reduced by improving the electrode morphology and structure. As shown in Fig. 7c, despite the decline in the first 12 h, which probably due to the activating process, the current density at an output voltage of 0.52 V decreases only about 9% of its original value at the end of 75 h testing, indicating that BiLSCF0.9 has sufficient stability during the operation. Therefore, it has the potential to be used as a cathode material for IT-SOFCs. 4. Conclusions
Fig. 7. (a) Cell voltage (left) and power density (right) as a function of the current density for a cell based on LSGM electrolyte (with 300 mm thickness) supported cell with BiLSCF0.9 cathode, which was measured in humidified H2 as the fuel and stationary air as the oxidant; (b) Resistance contributions (ohmic resistance, polarization resistance and total resistance) as determined from the impedance spectra (inset Fig.) under open circuit conditions during performance test of single cell based on BiLSCF0.9 cathode at 550, 600, 650 and 700 C; (c) Life durability with a constant output voltage of 0.52 V for the cell based on BiLSCF0.9 cathode at 600 C.
analysis (Fig. S1y). The performance of BiLSCF0.9 cathode in SOFC was examined using the LSGM electrolyte support with humidified H2 (3% H2O) as a fuel and ambient air as an oxidant in a temperature range of 550e700 C. Fig. 7a shows the I-V curves and the corresponding power density. Open circuit voltage (OCV) of the cell increases with the decreasing of operating temperature and the obtained OCV
A novel BiLSCFx (x ¼ 0.8, 0.9, 1.0, 1.1) perovskite oxide is successfully designed and synthesized for the first time and can be applied as a potential cathode material for IT-SOFCs. XRD analysis reveals that a pure phase with a cubic perovskite structure is obtained by calcinations at 1150 C for 8 h. Furthermore, the desirable chemical compatibility with LSGM is also realized at this condition. It is found that A-site variation has a significant effect on the electrochemical performance of the BiLSCF oxides. Especially, BiLSCF0.9 with A-site deficiency exhibits excellent molecular oxygen dissociation and electron charge transfer properties on the three phase boundary. As a result, the ASR of the BiLSCF0.9 electrode is low and a high power density of single cell fabricated by using it as cathode is obtained. The maximum power density of LSGM electrolyte (with 300 mm thickness)-supported single cell with BiLSCF0.9 as the cathode reaches the value of 0.66 W cm2 at 700 C by using air as the oxidant. Furthermore, the relatively stable performance is obtained, in which the current density at an output voltage of 0.52 V decreases only ~9% of its original value at the end of 75 h testing. It is expected that BiLSCF0.9 can be applied as a promising candidate material for IT-SOFC cathode.
D.S. Khaerudini et al. / Journal of Power Sources 334 (2016) 137e145
Acknowledgements This work is supported by Aomori City Government, Japan. Deni S. Khaerudini gratefully acknowledges the scholarship from the MEXT of Japan. The authors also thank Professor Takeshi Kubota at NJRISE for his technical support on experiments. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jpowsour.2016.10.026. References [1] B.C.H. Steele, A. Heinzel, Materials for fuel-cell technologies, Nature 414 (2001) 345e352. [2] A. Chroneos, B. Yildiz, A. Tarancon, D. Parfitt, J.A. Kilner, Oxygen diffusion in solid oxide fuel cell cathode and electrolyte materials: mechanistic insights from atomistic simulations, Energy Environ. Sci. 4 (2011) 2774e2789. [3] E.P. Murray, T. Tsai, S.A. Barnett, Oxygen transfer processes in (La,Sr)MnO3/ Y2O3-stabilized ZrO2 cathodes: an impedance spectroscopy study, Solid State Ion. 110 (1998) 235e243. [4] J.M. Ralph, C. Rossignol, R. Kumar, Cathode materials for reduced-temperature SOFCs, J. Electrochem. Soc. 150 (2003) A1518eA1522. [5] S. Badwal, Stability of solid oxide fuel cell components, Solid State Ion. 143 (2001) 39e46. [6] P. Hernandez-Fernandez, S. Rojas, P. Ocon, J.L.G. de la Fuente, P. Terreros, J.L. Garcia-Fierro, An opening route to the design of cathode materials for fuel cells based on PtCo nanoparticles, Appl. Catal. B Environ. 77 (1e2) (2007) 19e28. [7] J. Piao, K. Sun, N.Q. Zhang, X.B. Chen, S. Xu, D.R. Zhou, Preparation and characterization of Pr1-xSrxFeO3 cathode material for immediate temperature solid oxide fuel cells, J. Power Sources 172 (2007) 633e640. [8] C. Peters, A. Weber, E. Ivers-Tiffee, Nanoscaled ( La0.5Sr0.5 ) CoO3 d thin film cathodes for SOFC application at 500 C < T < 700 C, J. Electrochem. Soc. 155 (2008) B730eB737. [9] M. Li, Y. Wang, Y.L. Wang, F. Chen, C.R. Xia, Bismuth doped lanthanum ferrite perovskites as novel cathodes for intermediate-temperature solid oxide fuel cells, ACS Appl. Mater. Interfaces 6 (2014) 11286e11294. [10] A. Wedig, R. Merkle, B. Stuhlhofer, H.U. Habermeier, J. Maier, E. Heifets, Fast oxygen exchange kinetics of pore-free Bi1xSrxFeO3d thin films, Phys. Chem. Chem. Phys. 13 (2011) 16530e16533. [11] Y.J. Niu, J. Sunarso, F.L. Liang, W. Zhou, Z.H. Zhu, Z.P. Shao, A comparative study of oxygen reduction reaction on Bi- and La-doped SrFeO3-d perovskite cathodes, J. Electrochem. Soc. 158 (2011) B132eB138. [12] L. Wang, R. Merkle, J. Maier, Surface kinetics and mechanism of oxygen incorporation into Ba1-xSrxCoyFe1-yO3-d SOFC microelectrodes, J. Electrochem. Soc. 157 (2010) B1802eB1808. [13] D.S. Khaerudini, G. Guan, P. Zhang, X. Hao, A. Abudula, Prospects of oxide ionic conductivity bismuth vanadate-based solid electrolytes, Rev. Chem. Eng. 30 (6) (2014) 539e551. [14] D.S. Khaerudini, G. Guan, P. Zhang, X. Hao, Z. Wang, C. Xue, Y. Kasai, A. Abudula, Performance assessment of Bi0.3Sr0.7Co0.3Fe0.7O3-d-LSCF composite as cathode for intermediate-temperature solid oxide fuel cells with La0.8Sr0.2Ga0.8Mg0.2O3-d electrolyte, J. Power Sources 298 (2015) 269e279.
145
[15] J.G. Deng, L. Zhang, H.X. Dai, H. He, C.T. Au, Hydrothermally fabricated singlecrystalline strontium-substituted lanthanum manganite microcubes for the catalytic combustion of toluene, J. Mol. Catal. A Chem. 299 (2009) 60e67. ndez-Díaz, New [16] R. Martínez-Coronado, J.A. Alonso, A. Aguadero, M.T. Ferna SrMo1xCrxO3d perovskites as anodes in solid-oxide fuel cells, Int. J. Hydrogen Energy 39 (2013) 4067e4073. [17] S. McIntosh, J.F. Vente, W.G. Haije, D.H.A. Blank, H.J.M. Bouwmeester, Oxygen stoichiometry and chemical expansion of Ba0.5Sr0.5Co0.8Fe0.2O3d measured in situ diffraction, Chem. Mater. 18 (2006) 2187e2913. [18] X. Yong, J. Song, S.B. Adler, Thermal and chemical expansion of Sr-doped lanthanum cobalt oxide (La1-xSrxCoO3-d, Chem. Mater. 17 (17) (2005) 4537e4546. [19] M. Mori, Y. Hiei, Thermal expansion behavior of titanium-doped La(Sr)CrO3 solid oxide fuel cell interconnects, J. Am. Ceram. Soc. 84 (11) (2001) 2573e2578. [20] J. Pena-Martınez, D. Marrero-Lopez, J.C. Ruiz-Morales, P. Nunez, C. SanchezBautista, A.J. Dos Santos-Garcia, J. Canales-Vazquez, On Ba0.5Sr0.5Co1-yFeyO3d (y ¼0.1e0.9) oxides as cathode materials for La0.9Sr0.1Ga0.8Mg0.2O2.85 based IT-SOFCs, Int. J. Hydrogen Energy 34 (2009) 9486e9495. [21] D. Chen, R. Ran, K. Zhang, J. Wang, Z. Shao, Intermediate-temperature electrochemical performance of a polycrystalline PrBaCo2O5þd cathode on samarium-doped ceria electrolyte, J. Power Sources 188 (2009) 96e105. [22] N. Grunbaum, L. Mogni, F. Prado, A. Caneiro, Phase equilibrium and electrical conductivity of SrCo0.8Fe0.2O3d, J. Solid State Chem. 177 (2004) 2350e2357. [23] Z.H. Du, H.L. Zhao, Y.N. Shen, L. Wang, M.Y. Fang, K. Swierczek, K. Zheng, Evaluation of La0.3Sr0.7Ti1-xCoxO3 as a potential cathode material for solid oxide fuel cells, J. Mater. Chem. A 2 (2014) 10290e10299. [24] B. Philippeau, F. Mauvy, C. Mazataud, S. Fourcade, J.C. Grenier, Comparative study of electrochemical properties of mixed conducting Ln2NiO4 þ d (Ln ¼ La, Pr and Nd) and La0.6Sr0.4Fe0.8Co0.2O3-d as SOFC cathodes associated to Ce0.9Gd0.1O2- d, La0.8Sr0.2Ga0.8Mg0.2O3-d and La9Sr1Si6O26.5 electrolytes, Solid State Ion. 249e250 (2013) 17e25. [25] Y.N. Shen, H.L. Zhao, K. Swierczek, Z.H. Du, Z.X. Xie, Lattice structure, sintering behavior and electrochemical performance of La1.7Ca0.3Ni1-xCuxO4þd as cathode material for intermediate temperature solid oxide fuel cell, J. Power Sources 240 (2013) 759. [26] Z.B. Yang, C.H. Yang, B. Xiong, M.F. Han, F.L. Chen, BaCo0.7Fe0.2Nb0.1O3-d as cathode material for intermediate temperature solid oxide fuel cells, J. Power Sources 196 (2011) 9164e9168. [27] Z.B. Yang, C.H. Yang, C. Jin, M.F. Han, F.L. Chen, Ba0.9Co0.7Fe0.2Nb0.1O3-d as cathode material for intermediate temperature solid oxide fuel cells, Electrochem. Commun. 13 (2011) 882e885. [28] G.H. Chen, Y. Wang, J. Sunarso, F.L. Liang, H.P. Wang, A new scandium and niobium co-doped cobalt-free perovskite cathode for intermediatetemperature solid oxide fuel cells, Energy 95 (2016) 137e143. [29] M. Liu, D. Ding, K. Blinn, X. Li, L. Nie, M. Liu, Enhanced performance of LSCF cathode through surface modification, Int. J. Hydrogen Energy 37 (2012) 8613e8620. [30] M. Sahibzada, S.J. Benson, R.A. Rudkin, J.A. Kilner, Pd-promoted La0.6Sr0.4Co0.2Fe0.8O3 cathodes, Solid State Ion. 113e115 (1998) 285e290. [31] Y. He, X. Zhu, W. Yang, The role of A-site ion nonstoichiometry in the oxygen absorption properties of Sr1þxCo0.8Fe0.2O3 oxides, AlChE J. 57 (2011) 87e95. [32] Y. Teraoka, M. Yoshimatsu, N. Yamazoe, T. Seiyama, Chem. Lett. (1984) 893e896. [33] J.I. Jung, S.T. Misture, D.D. Edwards, The electronic conductivity of Ba0.5Sr0.5CoxFe1xO3d (BSCF: x ¼ 0 ~ 1.0) under different oxygen partial pressures, J. Electroceram. 24 (2010) 261e269. [34] T. Ishihara, H. Furutani, T. Yamada, Y. Takita, Oxide ion conductivity of double doped lanthanum gallate perovskite type oxide, Ionics 3 (1997) 209e213.